The Cambridge Handbook of Consciousness

This page intentionally left blank P1: KAE 0521857430pre CUFX049/Zelazo 0 521 85743 0 printer: cupusbw March 3, 2007 16:33

The Cambridge Handbook of Consciousness is the (CFI), focuses on the mechanisms underly- first of its kind in the field, and its appearance ing typical and atypical development of exec- marks a unique time in the history of intellectual utive function – the conscious self-regulation inquiry on the topic. After decades during which of thought, action, and emotion. In September consciousness was considered beyond the scope 2007, he will assume the Nancy M. and John of legitimate scientific investigation, conscious- L. Lindhal Professorship at the Institute of Child ness re-emerged as a popular focus of research Development, University of Minnesota. toward the end of the last century, and it has remained so for nearly 20 years. There are now Morris Moscovitch is the Max and Gianna Glass- so many different lines of investigation on con- man Chair in Neuropsychology and Aging in sciousness that the time has come when the field the Department of Psychology at the University may finally benefit from a book that pulls them of Toronto. He is also a Senior Scientist at the together and, by juxtaposing them, provides a Rotman Research Institute of Baycrest Centre comprehensive survey of this exciting field. for Geriatric Care. His research focuses on the neuropsychology of memory in humans but also Philip David Zelazo is Professor of Psychology addresses attention, face recognition, and hemi- at the University of Toronto, where he holds spheric specialization in young and older adults, a Canada Research Chair in Developmental and in people with brain damage. Neuroscience. He is also Co-Director of the Sino-Canadian Centre for Research in Child Evan Thompson is Professor of Philosophy at the Development, Southwest University, China. He University of Toronto. He is the author of Mind was Founding Editor of the Journal of Cogni- in Life: Biology, Phenomenoloy, and the Sciences of tion and Development. His research, which is Mind and Colour Vision: A Study in Cognitive Sci- funded by the Natural Sciences and Engineer- ence and the Philosophy of Perception. He is also ing Research Council (NSERC) of Canada, the the co-author of The Embodied Mind: Cognitive Canadian Institutes of Health Research (CIHR), Science and Human Experience. He is a former and the Canadian Foundation for Innovation holder of a Canada Research Chair.

i P1: KAE 0521857430pre CUFX049/Zelazo 0 521 85743 0 printer: cupusbw March 3, 2007 16:33

ii P1: KAE 0521857430pre CUFX049/Zelazo 0 521 85743 0 printer: cupusbw March 3, 2007 16:33

The Cambridge Handbook of Consciousness

Edited by Philip David Zelazo, Morris Moscovitch and Evan Thompson University of Toronto

iii CAMBRIDGE UNIVERSITY PRESS Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo

Cambridge University Press The Edinburgh Building, Cambridge CB2 8RU, UK Published in the United States of America by Cambridge University Press, New York www.cambridge.org Information on this title: www.cambridge.org/9780521857437

This publication is in copyright. Subject to statutory exception and to the provision of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published in print format 2007

ISBN-13 978-0-511-28923-1 eBook (EBL) ISBN-10 0-511-28923-5 eBook (EBL)

ISBN-13 978-0-521-85743-7 hardback ISBN-10 0-521-85743-0 hardback

ISBN-13 978-0-521-67412-6 paperback ISBN-10 0-521-67412-3 paperback

Cambridge University Press has no responsibility for the persistence or accuracy of urls for external or third-party internet websites referred to in this publication, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate. P1: KAE 0521857430pre CUFX049/Zelazo 0 521 85743 0 printer: cupusbw March 3, 2007 16:33

To the memory of Francisco J. Varela (7 September 1946–28 May 2001) –ET

To my growing family: Jill, Elana, David, Leora, and Ezra Meir –MM

For Sam, and the next iteration – PDZ

And a special dedication to Joseph E. Bogen (13 July 1926–22 April 2005)

v P1: KAE 0521857430pre CUFX049/Zelazo 0 521 85743 0 printer: cupusbw March 3, 2007 16:33

vi P1: KAE 0521857430pre CUFX049/Zelazo 0 521 85743 0 printer: cupusbw March 3, 2007 16:33

Contents

List of Contributors page xi b. computational approaches to consciousness 1. Introduction 1 6. Artiﬁcial Intelligence and Consciousness 117 part i Drew McDermott THE COGNITIVE SCIENCE OF CONSCIOUSNESS 7. Computational Models of Consciousness: A Taxonomy a. philosophy and Some Examples 151 2 . A Brief History of the Philosophical Ron Sun and Stan Franklin Problem of Consciousness 9 William Seager c. cognitive psychology 3. Philosophical Theories of 8. Cognitive Theories of Consciousness: Contemporary Consciousness 177 Western Perspectives 35 Katharine McGovern and Uriah Kriegel Bernard J. Baars 4. Philosophical Issues: 9 Phenomenology 67 . Behavioral, Neuroimaging, and Neuropsychological Approaches to Evan Thompson and Dan Zahavi Implicit Perception 207 5 . Asian Perspectives: Indian Theories Daniel J. Simons, Deborah E. of Mind 89 Hannula, David E. Warren, and Georges Dreyfus and Evan Thompson Steven W. Day

vii P1: KAE 0521857430pre CUFX049/Zelazo 0 521 85743 0 printer: cupusbw March 3, 2007 16:33

viii contents

10. Three Forms of Consciousness in g. anthropology/social Retrieving Memories 251 psychology of consciousness Henry L. Roediger III, Suparna 20. Social Psychological Approaches Rajaram, and Lisa Geraci to Consciousness 555 11. Metacognition and Consciousness 289 John A. Bargh Asher Koriat 21. The Evolution of Consciousness 571 12. Consciousness and Control Michael C. Corballis of Action 327 22. The Serpent’s Gift: Evolutionary Carlo Umilt`a Psychology and Consciousness 597 Jesse M. Bering and David F. d. linguistic considerations Bjorklund

13. Language and Consciousness 355 23. Anthropology of Consciousness 631 Wallace Chafe C. Jason Throop and Charles D. Laughlin 14. Narrative Modes of Consciousness h. psychodynamic approaches to and Selfhood 375 consciousness Keith Oatley 24. Motivation, Decision Making, and e. developmental psychology Consciousness: From Psychodynamics to Subliminal 15 . The Development Priming and Emotional Constraint 405 of Consciousness Satisfaction 673 Philip David Zelazo, Helena Hong Drew Westen, Joel Weinberger, and Gao, and Rebecca Todd Rebekah Bradley

f. alternative states of part ii consciousness THE NEUROSCIENCE OF 16. States of Consciousness: Normal CONSCIOUSNESS and Abnormal Variation 435 a. neurophysiological mechanisms J. Allan Hobson of consciousness

17. Consciousness in Hypnosis 445 25. Hunting the Ghost: Toward a John F. Kihlstrom Neuroscience of Consciousness 707 Petra Stoerig 18. Can We Study Subjective Experiences Objectively? 26. Neurodynamical Approaches to First-Person Perspective Consciousness 731 Approaches and Impaired Diego Cosmelli, Jean-Philippe Subjective States of Awareness Lachaux, and Evan Thompson in Schizophrenia 481 Jean-Marie Danion and Caroline b. neuropsychological aspects of Huron consciousness: disorders and neuroimaging 19. Meditation and the Neuroscience of Consciousness: An Introduction 499 27. The Thalamic Intralaminar Nuclei 775 Antoine Lutz, John D. Dunne, and and the Property of Consciousness Richard J. Davidson Joseph E. Bogen P1: KAE 0521857430pre CUFX049/Zelazo 0 521 85743 0 printer: cupusbw March 3, 2007 16:33

contents ix

28. The Cognitive Neuroscience of d. social neuroscience of Memory and Consciousness 809 consciousness Scott D. Slotnick and Daniel 30. Consciousness: Situated and Social 863 L. Schachter Ralph Adolphs

c.affectiveneuroscience part iii of consciousness QUANTUM APPROACHES TO 29. The Affective Neuroscience CONSCIOUSNESS of Consciousness: Higher- 31. Quantum Approaches to Order Syntactic Thoughts, Consciousness 881 Dual Routes to Emotion Henry Stapp and Action, and Consciousness 831 Author Index 909 Edmund T. Rolls Subject Index 939 P1: KAE 0521857430pre CUFX049/Zelazo 0 521 85743 0 printer: cupusbw March 3, 2007 16:33

x P1: KAE 0521857430pre CUFX049/Zelazo 0 521 85743 0 printer: cupusbw March 3, 2007 16:33

List of Contributors

Ralph Adolphs, PhD Joseph E. Bogen, MD (Deceased) Department of Psychology Formerly of University of Southern California California Institute of Technology and the University of California, Los Angeles HSS 228-77 Pasadena, CA 91125 USA Rebekah Bradley E-mail: [email protected] Department of Psychiatry and Behavioral Bernard J. Baars, PhD Sciences Emory University The Neurosciences Institute 1462 Clifton Road 10640 John Jay Hopkins Drive Atlanta, GA 30322 USA San Diego, CA 92121 USA E-mail: [email protected] E-mail: [email protected] John A. Bargh, PhD Wallace Chafe, PhD Department of Psychology Department of Linguistics Yale University University of California, Santa Barbara 2 Hillhouse Avenue Santa Barbara, CA 93106 USA P.O. Box 208205 E-mail: [email protected] New Haven, CT 06520-8205 USA E-mail: [email protected] Michael C. Corballis, PhD Jesse M. Bering, PhD Department of Psychology University of Auckland Institute of Cognition and Culture Private Bag 92019 Queen’s University, Belfast Auckland 1020 NEW ZEALAND 4 Fitzwilliam Street E-mail: [email protected] Belfast, Northern Ireland BT71NN E-mail: [email protected] Diego Cosmelli, PhD David F. Bjorklund, PhD Centro de Estudios Neurobiologicos´ Department of Psychology Departomento de Psiquiatroa´ Florida Atlantic University P. Universidad Catolica´ de Chile Boca Raton, FL 33431-0091 USA Marcoleto 387, 2 ◦ piso E-mail: [email protected] Santiago, Chile

xi P1: KAE 0521857430pre CUFX049/Zelazo 0 521 85743 0 printer: cupusbw March 3, 2007 16:33

xii list of contributors

(Also: Laboratoire de neurosciences Cognitives Campus Box 1125 et Imagerie Cer´ ebrale´ (LENA) St. Louis, MO 63130-4899 USA 47 Bd de l’Hopital,ˆ 75651 Paris FRANCE) E-mail: [email protected] E-mail: [email protected] Deborah E. Hannula Jean-Marie Danion, MD Psychology Department 405 INSERM Unite´ University of Illinois Hopitalˆ Civil de Strasbourg – Clinique 603 E. Daniel Street, Room 807 Psychiatrique Champaign, IL 61820 USA 1 ◦ 426 place de l’Hopitalˆ – BP n E-mail: [email protected] 67091 STRASBOURG Cedex FRANCE E-mail: jean-marie.danion@chru strasbourg.fr J. Allan Hobson, MD Richard J. Davidson, PhD Massachusetts Mental Health Center 12 W. M. Keck Laboratory for Functional Brain Psychiatry, S 74 Imaging and Behavior Fenwood Road 02115 Waisman Center Boston, MA USA University of Wisconsin-Madison E-mail: allan [email protected] 1500 Highland Avenue Caroline Huron, MD, PhD Madison, WI 53703-2280 USA 0117 E-mail: [email protected] INSERM Service Hopitalo-Universitaire de Sante´ Mentale Steven W. Day, BSc et Therapeuthique´ Department of Psychology Hopitalˆ Sainte-Anne University of Illinois Universite´ Paris V 603 East Daniel Street Pavillon Broca Champaign, IL 61820 USA 2 ter rue d’Alesia´ 75014 Georges Dreyfus, PhD Paris FRANCE E-mail: [email protected] Department of Religion Williams College John F. Kihlstrom, PhD E14 Stetson Hall Department of Psychology, MC 1650 Williamstown, MA 01267 USA University of California, Berkeley E-mail: [email protected] Tolman Hall 3210 John D. Dunne, PhD Berkeley, CA 94720-1650 USA Department of Religion E-mail: [email protected] Emory University Mailstop: 1535/002/1AA Asher Koriat, PhD 537 Kilgo Circle Department of Psychology Atlanta, GA 30322 USA University of Haifa E-mail: [email protected] Haifa 31905 ISRAEL E-mail: [email protected] Stan Franklin, PhD Institute for Intelligent Systems Uriah Kriegel, PhD The University of Memphis Department of Philosophy Memphis, TN 38152 USA Social Science Bldg. Rm 213 E-mail: [email protected] P.O. Box 210027 Helena Hong Gao, PhD Tucson, AZ 85721-0027 USA School of Humanities and Social Sciences E-mail: [email protected] Nanyang Technological University Jean-Philippe Lachaux Singapore 639798 INSERM – Unite´ 280 E-mail: [email protected] Centre Hospitalier Le Vinatier Lisa Geraci, PhD Batimentˆ 452 Department of Psychology 95 Boulevard Pinel Washington University 69500 BRON, FRANCE One Brookings Drive E-mail: [email protected] P1: KAE 0521857430pre CUFX049/Zelazo 0 521 85743 0 printer: cupusbw March 3, 2007 16:33

list of contributors xiii

Charles D. Laughlin, PhD Daniel L. Schachter, PhD Department of Sociology and Anthropology Department of Psychology Carleton University Harvard University 125 Colonel By Drive Cambridge, MA 02138 USA Ottawa,ONK1S 5B6 CANADA E-mail: [email protected] William Seager, PhD Antoine Lutz, PhD Department of Philosophy W. M. Keck Laboratory for Functional Brain University of Toronto at Scarborough Imaging and Behavior 265 Military Trail Waisman Center Scarborough, ON M1C 1A4 CANADA University of Wisconsin-Madison E-mail: [email protected] 1500 Highland Avenue Madison, WI 53703-2280 USA Daniel J. Simons, PhD E-mail: [email protected] Psychology Department University of Illinois Drew McDermott, PhD 603 E. Daniel Street, Room 807 Department of Computer Science Champaign, IL 61820 USA Yale University E-mail: [email protected] 208285 P.O. Box Scott D. Slotnick New Haven, CT 06520-8285 USA Department of Psychology E-mail: [email protected] Boston College Katharine McGovern, PhD McGuinn Hall Chestnut Hill, MA 02467 USA California Institute of Integral Studies E-mail: [email protected] 1453 Mission Street San Francisco, CA 94103 USA Henry Stapp, PhD E-mail: [email protected] Lawrence Berkeley National Lab Physics Division Keith Oately, PhD 1 Cyclotron Road Mail Stop 50A-5101 Department of Human Development and Berkeley, CA 94720-8153 USA Applied Psychology E-mail: [email protected] Ontario Institute for Studies in Petra Stoerig, PhD Education/University of Toronto Institute of Physiological Psychology 252 Bloor Street West Heinrich-Heine-University Toronto, ON M5S 1V6 CANADA Dusseldorf¨ D-40225 GERMANY E-mail: [email protected] E-mail: [email protected] Suparna Rajaram, PhD Ron Sun, PhD Department of Psychology Cognitive Science Department SUNY at Stony Brook Rensselaer Polytechnic Institute Stony Brook, NY 11794-2500 USA 110 Eighth Street, Carnegie 302A E-mail: [email protected] Troy, NY 12180 USA E-mail: [email protected] Henry L. Roediger III, PhD Evan Thompson, PhD Department of Psychology, Box 1125 Department of Philosophy Washington University University of Toronto One Brookings Drive 15 King’s College Circle St. Louis, MO 63130-4899 USA Toronto, ON M5S 3H7 CANADA E-mail: [email protected] E-mail: [email protected] Edmund T. Rolls, PhD C. Jason Throop, PhD University of Oxford Department of Anthropology Department of Experimental Psychology University of California, Los Angeles South Parks Road 341 Haines Hall Oxford OX13UD ENGLAND Los Angeles, CA 90095 USA E-mail: [email protected] E-mail: [email protected] P1: KAE 0521857430pre CUFX049/Zelazo 0 521 85743 0 printer: cupusbw March 3, 2007 16:33

xiv list of contributors

Rebecca Todd, BA Drew Westen, PhD Department of Human Development and Department of Psychology Applied Psychology Emory University Ontario Institute for Studies in 532 N. Kilgo Circle Education/University of Toronto Atlanta, GA 30322 USA 252 Bloor Street West E-mail: [email protected] Toronto, Ontario M5S 1V6 CANADA E-mail: [email protected] Dan Zahavi, PhD Carlo Umilta,` PhD Danish National Research Foundation Dipartimeto di Psicologia Generale Center for Subjectivity Research Universita di Padova Kobmagergade 46 via 8 Febbraio, 2-35122 Padova ITALY DK-1150 Copenhagen K DENMARK E-mail: [email protected] E-mail: [email protected] David E. Warren, BSc Philip David Zelazo, PhD Department of Psychology University of Illinois Department of Psychology 603 E. Daniel Street University of Toronto 100 Champaign, IL 61820 USA St. George Street 5 3 3 E-mail:[email protected] Toronto, ON M S G CANADA E-mail: [email protected] Joel Weinberger, PhD (After September 2007: Derner Institute Institute of Child Development Adelphi University University of Minnesota Box 701 51 East River Road Garden City, NY 11530 USA Minneapolis, MN 55455 USA E-mail:[email protected] E-mail: [email protected]) P1: KAE 0521857430pre CUFX049/Zelazo 0 521 85743 0 printer: cupusbw March 3, 2007 16:33

The Cambridge Handbook of Consciousness

xv P1: KAE 0521857430pre CUFX049/Zelazo 0 521 85743 0 printer: cupusbw March 3, 2007 16:33

xvi P1: JzG 0521857430c01 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:36

CHAPTER 1 Consciousness: An Introduction

Philip David Zelazo, Morris Moscovitch, and Evan Thompson

The Cambridge Handbook of Consciousness of this agreement are many, and it would brings together leading scholars from around be beyond the scope of this Introduction to the world who address the topic of con- do more than highlight a few (for further sciousness from a wide variety of perspec- discussion, see Umilta` & Moscovitch, 1994). tives, ranging from philosophical to anthro- One of the most obvious is the so-called cog- pological to neuroscientific. This handbook nitive revolution in psychology and the sub- is the first of its kind in the field, and its sequent emergence of cognitive science as an appearance marks a unique time in the his- interdisciplinary enterprise. Whereas previ- tory of intellectual inquiry on the topic. ously psychologists sought to describe law- After decades during which consciousness ful relations between environmental stimuli was considered beyond the scope of legiti- and behavioral responses, in the mid-1950s mate scientific investigation, consciousness or so they began to trace the flow of infor- re-emerged as a popular focus of research mation through a cognitive system, viewing during the latter part of the last century and the mind as a kind of computer program. it has remained so for more than 20 years. It eventually became clear, however, that by Indeed, there are now so many different lines focusing on the processing of information – of investigation on consciousness that the the kind of thing a computer can do – time has come when the field may finally psychology left out most of what really benefit from a book that pulls them together matters to us as human beings; as conscious and, by juxtaposing them, provides a com- subjects, it left us cold. The cognitive revo- prehensive survey of this exciting field. lution opened the door to the study of such By the mid-1990s, if not earlier, it was topics as attention and memory, and some widely agreed that one could not get a full time later, consciousness came on through. appreciation of psychological phenomena – The pre-1990s tendency to avoid discus- for example, of perception or memory – sions of consciousness, except in certain con- without distinguishing between conscious texts (e.g., in phenomenological philoso- and unconscious processes. The antecedents phy and psychoanalytic circles), may have 1 P1: JzG 0521857430c01 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:36

2 the cambridge handbook of consciousness

been due, in part, to the belief that con- over, some of these same atypical phenom- sciousness necessarily was a kind of ghost in ena (e.g., blindsight) have also been demon- the machine – one that inevitably courted strated in non-human animals, raising the the awful specter of dualism. Since then, possibility that consciousness is not associ- however, our ontological suppositions have ated exclusively with human beings. evolved, and this evolution may be a conse- A third prominent contribution to the quence of the growing trend toward interdis- current state of affairs is the development ciplinary investigation – seen, for example, in of new techniques that have made it pos- the emergence of cognitive science and neu- sible to treat consciousness in a more rig- roscience as coherent fields. The transdisci- orous and scientifically respectable fashion. plinary perspective afforded by new fields Foremost among these is the development may have engendered an increased open- of neuroimaging techniques that allow us to ness and willingness to explore problems correlate performance and subjective expe- that earlier were deemed too difficult to rience with brain function. These techniques address. Certainly, it provided the means include electrophysiological methods, such that made these problems seem soluble. as magneto-encephalography (MEG), and Indeed, precisely because consciousness is various types of functional neuroimaging, such a difficult problem, progress in solv- including functional magnetic resonance ing it probably depends on a convergence of imaging (fMRI). The analytic sophistication ideas and methodologies: We are unlikely to of these technologies is growing rapidly, as arrive at an adequate understanding of con- is the creation of new technologies that will sciousness in the absence of a transdiscipli- expand our capabilities to look into the brain nary perspective. more closely and seek answers to questions Clinical sciences, and in particular neu- that now seem impossible to address. ropsychology, also played a prominent role There is currently considerable interest in in helping usher in a new willingness to exploring the neural correlates of conscious- tackle the problem of consciousness. Vari- ness. There is also a growing realization, ous unusual syndromes came to light in the however, that it will not be possible to latter half of the 20th century, and these make serious headway in understanding con- syndromes seemed to demand an explana- sciousness without confronting the issue of tion in terms of consciousness. Blindsight is how to acquire more precise descriptive a good example: In this syndrome, patients first-person reports about subjective expe- with lesions to the occipital lobe of the rience (Jack & Roepstorff, 2003, 2004). brain are phenomenologically blind, but can Psychologists, especially clinical psycholo- nonetheless perform normally on a number gists and psychotherapists, have grappled of visual tasks. Another example is amne- with this issue for a long time, but it has sia, in which people who are phenomeno- gained new prominence thanks to the use logically amnesic as a result of damage to of neuroimaging techniques. Here one guid- medial temporal lobes or the diencephalon ing idea is that it may be possible to recover can acquire, retain, and recover informa- information about the highly variable neu- tion without awareness. Similar examples ral processes associated with conscious- emerged in other domains, and it soon ness by collecting more precise, trial-by- became clear that processes under con- trial first-person reports from experimental scious control complement, or compete participants. with, unconscious processes in the control If ever it was possible to do so, cer- of cognition and behavior. These issues are tainly serious students of the mind can also beginning to play a major role in the no longer ignore the topic of conscious- rigorous, scientific analysis of psychopathol- ness. This volume attempts to survey the ogy, the one field in which concerns with the major developments in a wide range of role of conscious and unconscious processes intellectual domains to give the reader an have played a steady role since Freud. More- appreciation of the state of the field and P1: JzG 0521857430c01 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:36

consciousness: an introduction 3

where it is heading. Despite our efforts to or levels of consciousness within domains of provide a comprehensive overview of the functioning, across domains, across species, field, however, there were several unavoid- and/or across the lifespan?), but there has able omissions. Though we had hoped to also been considerable progress. We hope include chapters on psychedelic drugs and this collection serves a useful function by on split-brain research, in the end we were helping readers see both how far we have unable to obtain these chapters in time. come in understanding consciousness and Readers interested in the latest scientific how far we have to go. writing on drugs and consciousness may wish to see Benny Shanon’s (2002) book on ayahuasca. Michael Gazzaniga’s (1998) Acknowledgments book, The Mind’s Past, provides an accessible overview of work on split-brain research The editors would like to thank Phil Laugh- and its implications for subjective experi- lin, formerly of CUP, who encouraged us to ence. We note, too, that although we were prepare this volume, and Armi Macaballug able to cover philosophical approaches to and Mary Cadette, who helped us during the consciousness from a variety of cultural final production phases. Dana Liebermann perspectives, including Continental phe- provided valuable assistance as we planned nomenology and Asian philosophy (particu- the volume, and Helena Hong Gao helped us larly Buddhism), there were inevitably oth- pull the many chapters together; we are very ers that we omitted. We apologize for these grateful to them both. We would also like unfortunate gaps. to thank the contributors for their patience The volume is organized mainly around during the editorial process (the scope of this a broad (sometimes untenable) distinction volume threatened, at times, to turn this pro- between cognitive scientific approaches and cess into an editorial nightmare . . . ). Finally, neuroscientific approaches. Although we are we note with sadness the death of Joseph mindful of the truly transdisiplinary nature Bogen, one of the pioneers in research on of contemporary work on consciousness, we consciousness. We regret that he was unable believe this distinction may be useful for to see his chapter in print. readers who wish to use this handbook as an advanced textbook. For example, readers who want a course in consciousness from a References cognitive science perspective might concentrate on Chapters 2–24. Readers approaching the topic from the perspective of neu- Gazzaniga, M. S. (1998). The mind’s past. Berke- roscience might emphasize Chapters 25–31. ley, CA: University of California Press. A more sociocultural course could include Jack, A. & Roepstorff, A. (Eds.) (2003). Trusting Chapters 2–4, 13–15, 19–24, and 31. More the subject? The use of introspective evidence in focused topical treatments are also possible. cognitive science. Vol. 1.Thorverton, UK: Imprint Academic. For example, a course on memory might 2004 include Chapters 6–8, 10, 18, and 29. Jack, A. & Roepstorff, A. (Eds.) ( ). Trust- ing the subject? The use of introspective evi- The topic of consciousness is relevant to dence in cognitive science. Vol. 2 .Thorverton, all intellectual inquiry – indeed, it is the UK: Imprint Academic. foundation of this inquiry. As the chap- Shanon, B. (2002). The antipodes of the mind: ters collected here show, individually and Charting the phenomenology of the ayahuasca together, by ignoring consciousness, one experience. New York: Oxford University Press. places unnecessary constraints on our under- Umilta,` C. & Moscovitch, M. (Eds.). (1994). Con- standing of a wide range of phenomena – scious and nonconscious information processing: and risks grossly distorting them. Many mys- Attention and Performance XV: Conscious and teries remain (e.g., what are the neural sub- nonconscious processes in cognition. Cambridge, strates of consciousness? are there varieties MA: MIT/Bradford Press. P1: JzG 0521857430c01 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:36

4 P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

Part I THE COGNITIVE SCIENCE OF CONSCIOUSNESS

5 P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

6 P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

A. Philosophy

7 P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

8 P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

CHAPTER 2 A Brief History of the Philosophical Problem of Consciousness

William Seager

Abstract ur-material; and increasingly sophisticated forms of materialism which, despite fail- The problem of consciousness, generally ing to resolve the problem of consciousness, referred to as the mind-body problem seemed to fit best with the scientific view of although this characterization is unfortu- the world and eventually came to dominate nately narrow, has been the subject of philo- thinking about the mind in the 20th century. sophical reflection for thousands of years. This chapter traces the development of this problem in Western philosophy from the I. Forms of Consciousness time of the ancient Greeks to the middle of the 20th century. The birth of sci- The term ‘consciousness’ possesses a huge ence in the 17th century and its subse- and diverse set of meanings. It is not even quent astounding success made the problem obvious that there is any one ‘thing’ that of mind particularly acute, and produced a all uses of the term have in common which host of philosophical positions in response. could stand as its core referent (see Wilkes These include the infamous interactionist 1988). When we think about conscious- dualism of Descartes and a host of dual- ness we may have in mind highly complex ist alternatives forced by the intractable mental activities, such as reflective self- problem of mind-matter interaction; a vari- consciousness or introspective conscious- ety of idealist positions which regard mind ness, of which perhaps only human beings as ontologically fundamental; emergentist are capable. Or we may be thinking about theories which posit entirely novel enti- something more purely phenomenal, perhaps ties, events, and laws which ‘grow’ out of something as apparently simple and uni- the material substrate; panpsychist, double tary as a momentary stab of pain. Paradig- aspect, and ‘neutral monist’ views in which matic examples of consciousness are the per- both mind and matter are somehow reflec- ceptual states of seeing and hearing, but tions of some underlying, barely knowable the nature of the consciousness involved is 9 P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

10 the cambridge handbook of consciousness

actually complex and far from clear. Are 2001, following James 1890/1950, ch. 9), the conscious elements of perception made which refers to the background of aware- up only of raw sensations from which we ness which sets the context for experience. construct objects of perception in a quasi- An example is our sense of orientation or intellectual operation? Or is perceptual con- rightness in a familiar environment (consider sciousness always of ‘completed’ objects the change in your state of consciousness with their worldly properties? when you recognize someone’s face who The realm of consciousness is hardly at first appeared to be a stranger). Moods exhausted by its reflective, introspective, or present another form of fringe conscious- perceptual forms. There is distinctively emo- ness, with clear links to the more overtly tional consciousness, which seems to nec- conscious emotional states but also clearly essarily involve both bodily feelings and distinct from them. some kind of cognitive assessment of them. But I think there is a fundamental com- Emotional states require a kind of evalua- monality to all these different forms of con- tion of a situation. Does consciousness thus sciousness. Consciousness is distinctive for include distinctive evaluative states, so that, its subjectivity or its first-person character. for example, consciousness of pain would There is ‘something it is like’ to be in a con- involve both bodily sensations and a conscious state, and only the conscious subject scious sense of aversion? Linked closely with has direct access to this way of being (see emotional states are familiar, but nonethe- Nagel 1974). In contrast, there is nothing it less rather peculiar, states of consciousness is like to be a rock, no subjective aspect to an that are essentially other directed, notably ashtray. But conscious beings are essentially empathy and sympathy. We visibly wince different in this respect. The huge variety in when others are hurt and almost seem to the forms of consciousness makes the prob- feel pain ourselves as we undergo this unique lem very complex, but the core problem of kind of experience. consciousness focuses on the nature of sub- Philosophers argue about whether all jectivity. thinking is accompanied by or perhaps even A further source of complexity arises constituted out of sensory materials (images from the range of possible explanatory tar- have been the traditional favorite candidate gets associated with the study of conscious- material), and some champion the idea of ness. One might, for instance, primarily a pure thought-consciousness independent focus on the structure or contents of con- of sensory components. In any event, there sciousness. These would provide a valid is no doubt that thought is something that answer to one legitimate sense of the ques- often happens consciously and is in some tion, What is consciousness? But then again, way different from perception, sensation, or one might be more interested in how con- other forms of consciousness. sciousness comes into being, either in a Another sort of conscious experience is developing individual or in the universe at closely associated with the idea of conscious large. Or one might wonder how conscious- thought but not identical to it: epistemolog- ness, seemingly so different from the purely ical consciousness, or the sense of certainty objective properties of the material world or doubt we have when consciously enter- studied by physics or chemistry, fits in with taining a proposition (such as ‘2 + 3 = 5’ the overall scientific view of the world. To or ‘the word ‘eat’ consists of three letters’). address all these aspects of the problem of Descartes famously appealed to such states consciousness would require volumes upon of consciousness in the ‘method of doubt’ volumes. The history presented in this chap- (see his Meditations 1641/1985). ter focuses on what has become perhaps the Still another significant if subtle form central issue in consciousness studies, which of consciousness has sometimes been given is the problem of integrating subjectivity the name ‘fringe’ consciousness (see Mangan into the scientific view of the world. P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

a brief history of the philosophical problem of consciousness 11

II. The Nature of the Problem tion by natural selection took hold and ultimately led at some time, somewhere to the Despite the huge range of diverse opinion, first living beings that could feel – pain and I think it is fair to say that there is now pleasure, want and fear – and could expe- something of a consensus view about the ori- rience sensations of light, sound, or odors. gin of consciousness, which I call here the The mainstream view sees this radical devel- mainstream view. It is something like the fol- opment as being conditioned by the evolu- lowing. The world is a purely physical sys- tion of neurological behavior control systems tem created some 13 billion years ago in the in co-evolutionary development with more prodigious event that Fred Hoyle labeled the capable sensory systems. Consciousness thus big bang. Very shortly after the big bang emerged as a product of increasing biological the world was in a primitive, ultra-hot, and complexity, from non-conscious precursors chaotic state in which normal matter could composed of non-conscious components. not exist, but as the system cooled the famil- Here we can raise many of the cen- iar elements of hydrogen and helium, as tral questions within the problem of con- well as some traces of a few heavier ele- sciousness. Imagine we were alien exo- ments, began to form. Then very interesting biologists observing the Earth around the things started to happen, as stars and galax- time of the emergence of consciousness. ies quickly evolved, burned through their How would we know that certain organ- hydrogen fuel, and went nova, in the process isms were conscious, while other organisms creating and spewing forth most of the ele- were not? What is it about the conscious ments of the periodic table into the increas- organisms that explains why they are ingly rich galactic environments. conscious? Furthermore, the appearance of There was not a trace of life, mind, or conscious beings looks to be a development consciousness throughout any of this pro- that sharply distinguishes them from their cess. That was to come later. The mainstream precursors, but the material processes of view continues with the creation of plan- evolution are not marked by such radical etary systems. At first these systems were discontinuities. To be sure, we do find strik- poor in heavier elements, but after just a ing differences among extant organisms. The few generations of star creation and destruc- unique human use of language is perhaps tion there were many Earth-like planets scat- the best example of such a difference, but tered through the vast – perhaps infinite – of course the apes exhibit a host of related, expanse of galaxies, and indeed some potentially precursor abilities, as do human 7 or 8 billion years after the big bang, the beings who lack full language use. Thus Earth itself formed along with our solar we have possible models of at least some system. aspects of our prelinguistic ancestors which We do not yet understand it very well, suggest the evolutionary path that led to but whether in a warm little pond, around language. a deeply submerged hydrothermal vent, But the slightest, most fleeting spark of amongst the complex interstices of some feeling is a full-fledged instance of conscious- clay-like matrix, as a pre-packaged gift from ness which entirely differentiates its posses- another world, or in some other way of sor from the realm of the non-conscious. which we have no inkling, conditions on the Note here a dissimilarity to other biologi- early Earth somehow enabled the special – cal features. Some creatures have wings and though entirely in accord with physical others do not, and we would expect that law – chemistry necessary for the beginnings in the evolution from wingless to winged of life. there would be a hazy region where it just But even with the presence of life or would not be clear whether or not a cer- proto-life, consciousness still did not grace tain creature’s appendages would count as the Earth. The long, slow processes of evolu- wings or not. Similarly, as we consider the P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

12 the cambridge handbook of consciousness

evolutionary advance from non-conscious to phrase, there is ‘something it is like’ to be a conscious creatures, there would be a range conscious entity? of creatures about which we would be unclear whether they were conscious or not. But in this latter case, there is a fact whether III. Ancient Hints or not the creatures in that range are feeling anything, however dimly or weakly, whereas Of course, the mainstream view has not we do not think there must be a fact about long been mainstream, for the problem of whether a certain appendage is or is not a consciousness cannot strike one at all until wing (a dim or faint feeling is 100% a kind of a fairly advanced scientific understanding consciousness, but a few feathers on a fore- of the world permits development of the limb is not a kind of wing). It is up to us materialism presupposed by the mainstream whether to count a certain sort of appendage view. A second necessary condition is sim- as a wing or not – it makes no difference, so ply the self-recognition that we are con- to speak, to the organism what we call it. scious beings possessing a host of mental But it is not up to us to decide whether or attributes. And that conception has been not organism X does or does not enjoy some around for a long time. Our ancestors ini- smidgen of consciousness – it either does or tiated a spectacular leap in conceptual tech- it does not. nology by devising what is nowadays called Lurking behind these relatively empiri- folk psychology. The development of the cal questions is a more basic theoretical, concepts of behavior explaining states such or metaphysical, issue. Given that crea- as belief and desire, motivating states of tures capable of fairly complex behavior pleasure and pain, and information-laden were evolving without consciousness, why states of perceptual sensation, as well as is consciousness necessary for the contin- the complex links amongst these concepts, ued evolution of more complex behavior? is perhaps the greatest piece of theorizing Just as wings are an excellent solution to ever produced by human beings. The power the problem of evolving flight, brains (or and age of folk psychology are attested by more generally nervous systems) are won- the universal animism of preliterate peo- derful at implementing richly capable sen- ples and the seemingly innate tendencies sory systems and coordinated behavior con- of very young children to regard various trol systems. But why should these brains natural or artificial processes as exemplify- be conscious? Although perhaps of doubting agency (see, among many others, Bloom ful coherence, it is useful to try to imag- 2004; Gergeley et al. 1995; Perner 1991). The ine our alien biologists as non-conscious persistence of the core mentalistic notions beings. Perhaps they are advanced machines of goal and purpose in Aristotle’s proto- well programmed in deduction, induction, scientific but highly sophisticated theoriz- and abduction. Now, why would they ever ing also reveals the powerful hold these con- posit consciousness in addition to, or as a cepts had, and have, on human thought. But feature of, complex sensory and behavioral to the extent that mentalistic attributes are control systems? As Thomas Huxley said, regarded as ubiquitous, no special problem ‘How it is that anything so remarkable as of relating the mental to the non-mental a state of consciousness comes about as a realm can arise, for there simply is no such result of irritating nervous tissue, is just as realm. unaccountable as the appearance of Djin But interesting hints of this problem arise when Aladdin rubbed his lamp’ (1866, 8, early on in philosophy, as the first glim- 210). We might, rather fancifully, describe merings of a naturalistic world view occur. this core philosophical question about con- A fruitful way to present this history is sciousness as how the genie of conscious- in terms of a fundamental divergence in ness gets into the lamp of the brain, or thought that arose early and has not yet died why, to use Thomas Nagel’s (1974) famous out in current debate. This is the contrast P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

a brief history of the philosophical problem of consciousness 13

between emergence and panpsychism. The and instead advanced the view that ‘every- mainstream view accepts emergence: mind thing is in everything’. Anaxagoras explained or consciousness appeared out of non- the obvious contrary appearance by a conscious precursors and non-conscious ‘principle of dominance and latency’ (see components (note there is both a synchronic Mourelatos 1986), which asserted that some and diachronic sense of emergence). Panpsy- qualities were dominant in their contri- chism is the alternative view that emergence bution to the behaviour and appearance is impossible and mind must be already and of things. However, Anaxagoras’s views on always present, in some sense, throughout mind are complex because he apparently the universe (a panpsychist might allow that regarded it as uniquely not containing any mind emerges in the trivial sense that the measure of other things and thus not fully universe may have been created out of noth- in accord with his mixing principles. Per- ing and hence out of ‘non-consciousness’; haps this can be interpreted as the asser- the characteristically panpsychist position tion that mind is ontologically fundamen- here would be that consciousness must have tal in a special way; Anaxagoras did seem been created along with whatever other to believe that everything has some por- fundamental features of the world were tion of mind in it while refraining from put in place at the beginning). Of course, the assertion that everything has a mind this divergence transcends the mind-body (even this is controversial; see Barnes 1982, problem and reflects a fundamental differ- 405 ff.). ence in thinking about how the world is On the other hand, Empedocles, an structured. almost exact contemporary of Anaxagoras, The Presocratic philosophers who flour- favoured an emergentist account based upon ished some 2,500 years ago in the Mediter- the famous doctrine of the four elements: ranean basin were the first in the West earth, air, fire and water. All qualities were to conceive of something like a scientific to be explicated in terms of ratios of these approach to nature, and it was their con- elements. The overall distribution of the ception that eventually led to what we call elements, which were themselves eternal science. Although their particular theories and unchangeable, was controlled by ‘love were understandably crude and often very and strife’, whose operations are curiously fanciful, they were able to grasp the idea that reminiscent of some doctrines of mod- the world could be viewed as composed out ern thermodynamics, in a grand cyclically of elemental features, whose essential char- dynamic universe. It is true that Empedo- acterization might be hidden from human cles is sometimes regarded as a panpsy- senses and which acted according to constant chist because of the universal role of love and universal principles or laws. and strife (see Edwards 1967, for exam- The Presocratics immediately recognized ple), but there seems little of the mental in the basic dilemma: either mind (or, more Empedocles’s conceptions, which are rather generally, whatever apparently macroscopic, more like forces of aggregation and dis- high-level, or non-fundamental property is aggregation, respectively (see Barnes 1982, at issue) is an elemental feature of the world, 308 ff.). or it somehow emerges from, or is condi- The purest form of emergentism was pro- tioned by, such features. If one opts for emer- pounded by the famed atomist Democritus gence, it is incumbent upon one to at least (c. 460–370 bce). His principle of emer- sketch the means by which new features gence was based upon the possibility of emerge. If one opts for panpsychism (thus multi-shaped, invisibly tiny atoms interlock- broadly construed), then one must account ing to form an infinity of more complex for the all too obviously apparent total lack structures. But Democritus, in a way echo- of certain features at the fundamental level. ing Anaxagoras and perhaps hinting at the For example, Anaxagoras (c. 500–425 bce) later distinction between primary and sec- flatly denied that emergence was possible ondary properties, had to admit that the P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

14 the cambridge handbook of consciousness

qualities of experience (what philosophers provides a basic, and perpetually influen- nowadays call qualia, the subjective fea- tial, tri-component-based psychological the- tures of conscious experience) could not be ory (see Republic, Book 4, Plato 1961). These accounted for in this way and chose, ulti- facets of his thought illustrate the two basic mately unsatisfactorily, to relegate them to aspects of the problem of consciousness: the non-existence: ‘sweet exists by convention, ontological question and the issue of how bitter by convention, in truth only atoms mind is structured. Plato’s primary motiva- and the void’. Sorely missed is Democritus’s tion for accepting a dualist account of mind account of how conventions themselves – and body presumably stems from the doc- the consciously agreed upon means of com- trine of the forms. These are entities which mon reference to our subjective responses – in some way express the intrinsic essence emerge from the dancing atoms (thus, the of things. The form of circle is that which ideas of Democritus anticipate the reflex- our imperfect drawings of circles imitate and ive problem of modern eliminativist mate- point to. The mind can grasp this form, even rialists [e.g., Churchland 1981] who would though we have never perceived a true cir- enjoin us to consciously accept a view which cle, but only more or less imperfect approx- evidently entails that there is no such thing imations. The ability of the mind to com- as conscious acceptance of views – see mune with the radically non-physical forms Chapter 3). suggests that mind itself cannot be physical. What is striking about these early strug- In the Phaedo, Plato (putting words in the gles about the proper form of a scientific mouth of Socrates) ridicules the reduction- understanding of the world is that the mind ist account of Anaxogoras which sees human and particularly consciousness keep rising as action as caused by local physical events. In special problems. It is sometimes said that its place, the mind is proposed as the final the mind-body problem is not an ancient (i.e., teleological) cause of action, merely philosophical issue on the basis that sensa- conditioned or constrained by the physical: tions were complacently regarded as bod- ‘if it were said that without such bones and ily phenomena (see Matson 1966), but it sinews and all the rest of them I should not does seem that the problem of conscious- be able to do what I think is right, it would ness was vexing philosophers 2,500 years be true. But to say that is because of them ago, and in a form redolent of contem- that I do what I am doing, and not through porary worries. Also critically important is choice of what is best – although my actions the way that the problem of conscious- are controlled by mind – would be a very lax ness inescapably arises within the context of and inaccurate form of expression’ (Phaedo, developing an integrated scientific view of 98b ff.). the world. In general, Plato’s arguments for dualism The reductionist strain in the Presocratics are not very convincing. Here’s one. Life was not favoured by the two giants of Greek must come from death, because otherwise, philosophy, Plato and Aristotle, despite their as all living things eventually die, everything own radical disagreements about how the would eventually be dead. Life can come world should be understood. Plato utterly from death only if there is a distinct ‘compo- lacked the naturalizing temperament of the nent’, responsible for something being alive, Presocratic philosophers, although he was that persists through the life-death-life cycle. well aware of their efforts. He explicitly criti- That persistent component is soul or mind cizes Anaxagoras’s efforts to provide natural- (Phaedo 72c-d). Another argument which istic, causal explanations of human behavior Plato frequently invokes (or presupposes in (see Phaedo, Plato 1961). other argumentation) is based on reincarna- Of course, Plato nonetheless has a signifi- tion. If we grant that reincarnation occurs, cant role in the debate because he advances it is a reasonable inference that something positive arguments in favour of the thesis persists which is what is reincarnated. This that mind and body are distinct. He also is a big ‘if’ to modern readers of a scientific P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

a brief history of the philosophical problem of consciousness 15

bent, but the doctrine of reincarnation was see Nelson 1990; see also Nussbaum and widespread throughout ancient times and Putnam 1992). is still taken seriously by large numbers Yet there are elements of Aristotle’s of people. The kernel of a more powerful account that are not very naturalistic. Early argument for dualism lurks here as well, in the De Anima Aristotle raises the pos- which was deployed by Descartes much later sibility that the relation between the body (see below). and the mind is analogous to that between Aristotle is famously more naturalisti- sailor and ship, which would imply that cally inclined than Plato (Raphael’s School of mind is independent of body. Later Aristotle Athens shows Plato pointing upward to the apparently endorses this possibility when he heavens while Aristotle gestures downward discusses, notoriously obscurely, the ‘active to Earth as they stare determinedly at each intellect’ – the ‘part’ of the soul capable of other). But Aristotle’s views on mind are rational thought (De Anima, bk. 3, chs. 4– complex and obscure; they are certainly not 5). Aristotle clearly states that the active straightforwardly reductionist (the soul is intellect is separable from body and can not, for example, a particularly subtle kind of exist without it. For Aristotle, like Plato, the matter, such as fire). Aristotle’s metaphysics problematic feature of mind was its capac- deployed a fundamental distinction between ity for abstract thought and not conscious- matter and form, and any object necessar- ness per se, although of course these thinkers ily instantiates both. A statue of a horse were implicitly discussing conscious thought has its matter, bronze, and its form, horse. and had no conception of mind apart from Aristotle is not using Plato’s conception of consciousness. form here. The form of something is not an Discussion of one particular, and highly other-world separate entity, but something interesting if perennially controversial, fea- more like the way in which the matter of ture of consciousness can perhaps be traced something is organized or structured. Nor by to Aristotle. This is the self-intimating or matter does Aristotle mean the fundamental self-representing nature of all conscious physical stuff we refer to by that word; mat- states. Many thinkers have regarded it as ter is whatever relatively unstructured stuff axiomatic that one could not be in a con- is ‘enformed’ to make an object (English scious state without being aware of that retains something of this notion in its use state, and Aristotle makes some remarks that of matter to mean topic), so bronze is the suggest he may belong to this school of matter of a statue, but soldiers would be the thought. For example, in Book Three of De matter of an army. Objects can differ in mat- Anima Aristotle presents, rather swiftly, the ter, but agree in form (two identical pictures, following regress argument: one on paper and another on a computer Since we perceive that we see and hear, screen) or vice versa. More abstractly, Aris- it is necessarily either by means of the totle regarded life as the form of plants and seeing that one perceives that one sees animals and named the form of living things or by another [perception]. But the same soul (‘the form of a natural body having life [perception] will be both of the seeing and of potentially within it’ 1984, De Anima, bk. 2, the colour that underlies it, with the result ch. 1). Aristotle’s views have some affinity that either two [perceptions] will be of the both with modern biology’s conception of same thing, or it [the perception] will be of life and the doctrine of psychophysical func- itself. Further, if the perception of seeing is tionalism insofar as he stresses that soul is a different [perception], either this will pro- not a separate thing requiring another onto- ceed to infinity or some [perception] will be logical realm, but also cannot be reduced to of itself; so that we ought to posit this in the first instance. mere matter because its essential attribute is function and organization (for a close and The passage is somewhat difficult to inter- skeptical look at the link between Aristo- pret, even in this translation from Vic- tle’s philosophy and modern functionalism tor Caston (2002) which forms part of an P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

16 the cambridge handbook of consciousness

intricate (and controversial) mixed philo- certain mental state, they too have mental sophical, exegetical, and linguistic argument states. In addition, he anticipated certain key in favor of the view that Aristotle accepted features of Descartes’ dualistic account of a self-representational account of conscious human beings, including Descartes’ famous states which possessed unique phenomenal argument from his conscious self-awareness properties. Aristotle’s argument appears to to the certainty of his own existence (City of be that if it is essential to a conscious state God, Bk. 11, Ch. 21) and the idea that mind that it be consciously apprehended then con- and body, although ontologically entirely scious states must be self-representing on distinct, somehow are united in the human pain of an infinite regress of states, each person. Here Augustine also broaches one of representing (and hence enabling conscious the key puzzles of Cartesian dualism where apprehension of) the previous state in the he admits the ‘mode of union’ by which series. The crucial premise that all mental bodies and spirits are bound together to states must be conscious is formally neces- become animals is ‘beyond the comprehen- sary for the regress. Modern representational sion of man’ (City of God, Bk. 21, Ch. 10). accounts of consciousness which accept that Although we see here that Augustine did conscious states are self-intimating, such not agree with Descartes in denying minds as the Higher Order Thought theory, can to animals, we can also note the complete block the regress by positing non-conscious lack of any idea that this mystery poses any thoughts which make lower order thoughts special problem for our understanding of the conscious by being about them (see Seager natural world (see O’Daly 1987 for a detailed 1999, Chapter 3, and see Chapters 3 and 4, discussion of Augustine’s philosophy of this volume). mind). In fact, the tenets of Christian dogma, eventually wedded to a fundamentally Aris- IV. The Scientific Revolution totelian outlook, conspired to suppress any idea that consciousness or mind could be, Although the philosophy of the Middle Ages should be, or needed to be explained in natu- was vigorous and compendious, the prob- ralistic terms. It was the scientific revolution lem of fitting consciousness into the natu- of the 16th and 17th centuries that forced the ral world did not figure prominently (for an problem into prominence. argument, following in the tradition of Mat- Galileo’s distinction between primary son 1966, that the medievals’ views on the and secondary properties, crucial for the nature of sensation precluded the recogni- development of science insofar as it freed sci- tion of at least some versions of the mind- ence from a hopelessly premature attempt body problem, see King 2005). There were to explain complex sensible qualities in many acute studies of human psychology mechanical terms, explicitly set up an oppo- and innovative theoretical work on the con- sition between matter and consciousness: ‘I tent and structure of consciousness and cog- think that tastes, odors, colors, and so on are nition. Of special note is the 4th-century no more than mere names so far as the object philosopher and Church Father, St. Augus- in which we place them is concerned, and tine (354–430 ce). His writings exhibit that they reside only in the consciousness. important insights into the phenomenology Hence if the living creature were removed of consciousness, especially with regard to all these qualities would be wiped away and the experience of time, will, and the self annihilated’ (1623/1957, 274). The welcome (see especially Confessions and On Free Will; consequence is that if there are therefore no 400/1998, 426/1998). He was one of the first colors in the world then science is free to philosophers to address the problem of other ignore them. That was perhaps good tactics minds, arguing on the basis of introspection in Galileo’s time, but it was a strategic time and analogy that because others behave as bomb waiting to go off when science could he behaves when he is aware of being in a no longer delay investigating the mind itself. P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

a brief history of the philosophical problem of consciousness 17

The mind-body problem in its modern But Descartes is most remembered and form is essentially the work of a single genius, reviled for his insistence upon the strict Rene´ Descartes (1596–1650), who reformed separation of mind and body which, we the way we think about mind and conscious- are enjoined to believe, required sunder- ness, leaving us with a set of intuitions that ing the world itself into radically dis- persist to this day. To take just one topi- tinct realms, thereby fundamentally splitting cal example, the basic idea behind the fic- human beings from nature (including their tional technology of the Matrix films is thor- own), denigrated emotion in favour of rea- oughly Cartesian: what we experience is not son, and inspired a lack of respect for animals directly related to the state of the environ- and nature in general. Why was Descartes a ment, but is instead the result of a com- dualist? Some have suggested that Descartes plex function – involving essential sensory lacked the courage to follow his science to its and cognitive mediation based upon neural logical and materialist conclusion (the fate of systems – from the environment to our cur- Galileo is said to have had a strong effect on rent state of consciousness. Thus two brains him, or it may be that Descartes really had that are in identical states ought to be in no wish to harm the Catholic church). But the same state of consciousness, no matter Descartes did have arguments for his dual- what differences there are in their respec- ism, some of which still have supporters. tive environments. It now seems intuitively These arguments also set out one of the basic obvious that this is correct (so contemporary strategies of anti-materialism. philosophers make exotic and subtle argu- To show that mind and body are distinct, ments against it) and that, to take another it will suffice to show that mind has some stock philosophical example, a brain in a vat, property that matter lacks. The general prin- if kept alive in an appropriate chemical bath ciple here, which is that of the alibi, was cod- and if fed proper input signals into its sev- ified by another 17th-century philosopher, ered nerve endings (cleverly coupled to the Gottfried Leibniz (1646–1716) and is now output of the brain’s motor output nerves), known as Leibniz’s Law: if x has a property would have experiences which could be which y lacks, then x and y are not iden- indistinguishable from, say, those you are tical. Descartes argued, for example, that having at this very moment. This thought although matter is extended in space, mind experiment reveals another of the reforma- takes up no space at all. Thus, they could tions of philosophy instituted by Descartes: not be identical. It certainly does seem odd the invention of modern epistemology, for to ask how many cubic centimeters my mind how could you know that you are not such a takes up (does a broad mind take up more brain in a vat. space than a narrow one?). But it is not obvi- Descartes was of course also one of the ous that this question is anything more than creators of the scientific revolution, pro- merely a feature of the conventional way viding seminal efforts in mathematics and we think about minds. An analogy would physics. But he also saw with remarkable be an argument that machines cannot think prevision the outlines of neuropsychology. because they are not alive; there is no par- With no conception of how the nervous ticular reason to think that the heretofore system actually works and instead deploy- constant and evident link between life and ing a kind of hydraulic metaphor, Descartes thought represents anything more than a envisioned nerve-based sensory and cogni- kind of accident in the way minds happened tive systems and a kind of network struc- to be created. In any event, this strategy is still ture in the brain, even – anticipating Hebb – at the core of the problem of consciousness. suggesting that connections in the brain One current line of argument, for example, are strengthened through associated activa- contends that consciousness has a kind of tion. His notorious discussion of animals as first-person subjectivity (the ‘what it is like’ machines can be seen as the precursor of a to experience something), whereas matter materialist account of cognition. is purely third-person objective – hence P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

18 the cambridge handbook of consciousness

consciousness and matter must be funda- incomprehensible. The Princess allowed her- mentally different phenomena. self less than fully satisfied with this reply. Descartes, in the sixth of his Meditations It was also noticed that Descartes’ dual- (1641/1985), also invented an astonishingly ism conflicted with the emerging under- novel kind of argument for dualism. The standing of the conservation of certain argument is couched in theological terms, physical quantities. Descartes himself only but that was merely for purposes of clarity accepted that the total amount, but not and forcefulness (in the 17th century, using direction, of motion was conserved. Thus God to explain one’s argument was impec- the mind’s ability to wiggle the pineal gland cable rhetoric). Descartes asked us to con- (where Descartes posited the seat of the sider whether it was at least possible that God soul) would redirect motion without vio- could destroy one’s body while leaving one’s lating natural law. But it was soon discov- mind intact. If it was possible then of course ered that it was momentum – or directed God could perform the feat if He wished. motion – that is conserved, and thus the But nothing can be separated from itself! So mind-induced motion of the pineal gland if it is merely possible that God could sunder would indeed contradict the laws of nature mind from body, then they must already be (one might try to regard this as a fea- different things. So, anyone who thinks that, ture rather than a bug, because at least it say, a consciousness persisting after bodily makes Descartes’ theory empirically testable death is even so much as a bare possibil- in principle). ity already thinks that consciousness is not In addition to the ontological aspect of a physical phenomenon. This argument is his views, Descartes had some interesting valid, but it has a little flaw: how do we insights into the phenomenological side of know that what we think is possible is truly consciousness. For Descartes, the elements so? Many are the mathematicians labouring of conscious experience are what he called to prove theorems which will turn out to ‘ideas’ (Descartes pioneered the modern use be unprovable (think of the centuries-long of this term to stand for mental items), and effort to square the circle) – what do they every idea possesses two kinds of reality: think they are doing? Nonetheless, it is a formal and objective. The formal reality highly interesting revelation that the mere of something is simply what it is in itself, possibility of dualism (in the sense consid- whereas the objective reality is what, if any- ered here) entails that dualism is true. thing, it represents (so, the formal reality of a Cartesian dualism also included the doc- picture of a horse is paper and paint; a horse trine of mind-body interaction. This seems is the objective reality). Though Descartes like common sense: when someone kicks is often pilloried as one who believed that me, that causes me to feel pain and anger, we are only ever conscious of our own ideas, and then it is my anger that makes me it is far from clear that this is Descartes’ kick them back. Causation appears to run position. It is possible to read him instead from body to mind and back again. But as as a precursor of modern representational soon as Descartes propounded his theory of theories of consciousness [see Chapter 3], mind, this interaction was seen to be deeply in which it is asserted that, although con- problematic. One of Descartes’ aristocratic sciousness essentially involves mental rep- female correspondents, the Princess Elisa- resentation, what we are conscious of is beth of Palatine, asked the crucial question: not the representations themselves but their “How can the soul of man determine the content (rather in the way that although spirits of the body, so as to produce voluntary we must use words to talk about things, actions (given that the soul is only a thinking we are not thereby always talking about substance)?” (from a letter of May 1643). It’s words). Descartes says that ‘there cannot be a fair question and Descartes’ only answer any ideas which do not appear to represent was that the mind-body union was instituted some things . . . ’ (Meditation 3), and perhaps and maintained by God and was humanly this suggests that even in cases of illusion P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

a brief history of the philosophical problem of consciousness 19

Descartes’ view was that our experience is position seemed dominant, but the mate- of the representational content of the ideas rialists, like the early mammals scrabbling and that we do not, as it were, see our own under the mighty dinosaurs, were to have ideas. their day. Finally, because Descartes is often misrep- Early materialists had to face more than resented as denigrating bodily feelings and an intellectual struggle, because their doc- emotions in favour of pure reason, it is worth trine stood in clear contradiction with fun- pointing out that he developed a sophisti- damental beliefs endorsed by the Chris- cated account of the emotions which stresses tian church, and many thinkers have been both their importance and the importance charged with softening their views to avoid of the bodily feelings which accompany ecclesiastical censure. One such is Pierre them (1649/1985). Descartes – perhaps con- Gassendi (1592–1655), who espoused an tra Aristotle – strenuously denied that the updated version of ancient Epicurean atom- mind was ‘in’ the body the way a pilot is ism, but who added immortal and immate- in a ship, for the intimate connection to the rial souls to the dance of the atoms. The souls body and the host of functionally significant were responsible, in a familiar refrain, for our feelings which the body arouses in the mind higher intellectual abilities. On the material- in the appropriate circumstances meant that ist core of such a view, nature is ultimately the mind-body link was not a mere commu- composed of tiny, indivisible, and indestruc- nication channel. Descartes declared instead tible physical particles whose interactions that the mind and body formed a ‘substan- account for all the complexity and behaviour tial union’ and that emotional response was of organized matter. Gassendi asserted that essential to cognition. the ‘sentient soul’, as opposed to the imma- Despite the fact that if one is willing terial ‘sapient soul’, was a material compo- to endorse a dualism of mind and body nent of animals and humans, composed of an then Descartes’ interactive version seems especially subtle, quick-moving type of mat- to be the most intuitively reasonable, the ter which is capable of forming the system difficulties of understanding how two such of images we call imagination and percep- entirely distinct realms could causally inter- tion (Gassendi also endorsed the empiricist act created an avid market for alternative principle that all ideas are based on prior theories of the mind-body relation. Two sensory experience). These are literally lit- broad streams of theory can be discerned, tle images in the brain. Of course, there which I label, not altogether happily, ide- is a problem here: who is looking at these alist and materialist. Idealists regard mind images? What good does it do to postu- or consciousness as the fundamental exis- late them? For Descartes, the experience of tent and deny the independent existence sensory perception or imagination is simi- of the material world; its apparent reality larly dependent upon corporeal imagery, but is to be explained as a function of mental- because the visual experience is a mental ity. Materialists cannot follow such a direct act, there really is someone to appreciate the route, for they have great difficulty in out- brain’s artwork. (Descartes in fact tried to right denying the existence of mind and use the imagistic quality of certain experi- generally content themselves with in some ences as an argument for the existence of way identifying it with features of matter. material objects, because real images need The asymmetry in these positions is inter- a material substrate in which they are real- esting. Idealists can easily assert that the ized – but Descartes concluded that this material world is all illusory. Materialists fall argument was far from conclusive.) A subtle into paradox if they attempt the same strat- distinction here may have directed philoso- egy – for the assertion that mind is illu- phers’ thinking away from this worry. This sory presupposes the existence of illusions, is the difference between what are nowa- which are themselves mental entities. For days called substance and property dualism. a long time (centuries, I mean) the idealist Descartes is a substance dualist (hence also P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

20 the cambridge handbook of consciousness

a property dualist, but that is a rather triv- appetites or motivations as motions, but says ial consequence of his view). Substance in that pleasure and pain are the appearances general was understood as that which could of these motions (1651/1998, pt. 1, ch. 6). It exist independently (or perhaps requiring would seem that ‘appearance’ is Hobbes’s only the concurrence of God). Matter was term for something like phenomenal con- thus a substance, but properties of matter sciousness, and he seems to be saying that were not themselves substantial, for proper- such consciousness is caused by motions in ties require substance in which to be instan- the brain but is not identical to them, which tiated. According to Descartes, mind is a sec- of course flatly contradicts his claim that ond kind of substance, with, naturally, its motion can only produce motion. Though own set of characteristically mental prop- obviously Hobbes is not clear about this erties. Thus one basic form of materialism problem, we might anachronistically char- involves merely the denial of mental sub- acterize him as a substance materialist who stance, and the early materialists were keen is also a property dualist. to make this aspect of their views clear. But In any case, materialism was very far from denial of substance dualism leaves open the the generally favoured opinion, and the per- question of the nature of mental properties ceived difficulties of Descartes’ substance or attributes (consciousness can be regarded dualism led instead to a series of inventive as a feature of the brain, but is no less myste- alternatives to interactive substance dual- rious for being labeled a property of a phys- ism, the two most important being those of ical object). Baruch de Spinoza (1632–1677) and Leib- The problem is clearer in the work of niz. In an austerely beautiful if forbidding another early materialist, Thomas Hobbes work, the Ethics (1677/1985), Spinoza laid (1588–1679) who, entranced by the new out a theory which perhaps, logically, ought science inaugurated by Galileo, declared to have been that of Descartes. Spinoza notes that absolutely everything should be expli- that substance is that which exists indepen- cable in terms of the motions of mat- dently of all other things, and thus there can ter and the efficient causal interaction of be only one ‘maximal’ substance: God. If material contact. Eventually coming to con- that is so, then matter and mind can only sider the mind, Hobbes pursues motion be features of the God-substance (Spinoza into the brain to account for sensory called them attributes and asserted there phenomena: ‘the cause of sense is the were an infinite number of them, although external body . . . which presses the organ we are only aware of two). Spinoza’s theory proper to each sense . . . which pressure, by is an early form of what came to be called the mediation of the nerves . . . continues ‘dual aspect theory’, which asserts that mind inwards to the brain . . . ’ (1651/1998, pt. 1, and matter are mere aspects of some under- ch. 1). Hobbes goes out of his way to stress lying kind of thing of which we have no clear that there is nothing immaterial, occult, or apprehension. Particular material or men- supernatural here; there is just the vari- tal individuals (as we would say) are mere ous ways that physical events influence our modifications of their parent attributes (so material sense organs: ‘neither in us that are your mind is a kind of short-lived ripple in pressed are they anything else but divers the attribute of mind and your body a small motions; for motion produceth nothing but disturbance in the material attribute). The motion’ (1651/1998, pt. 1, ch. 1). But then attributes are a perfect reflection of their Hobbes makes a curious remark: speaking underlying substance, but only in terms of of these ‘divers motions’ in the brain he says, one aspect (very roughly like having both ‘but their appearance to us is fancy, the same a climatographic and topographic map of waking that dreaming’. However, he else- the same territory). Thus Spinoza believed where states that ‘all fancies are motions that the patterns within any attribute would within us’ (1651/1998, pt. 1, ch. 3). Com- be mirrored in all the others; in particu- pounding the confusion he also describes our lar, mind and body would be synchronized P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

a brief history of the philosophical problem of consciousness 21

automatically and necessarily. This explains in themselves radically non-spatial and per- the apparent linkage between mind and haps even non-temporal (Leibniz’s vision of body – both are merely aspects of the same space and time emerging from some more underlying substance – while at the same elementary systems of relations has always time preserving the causal completeness of been tempting, if hard to fathom, and now each realm. In the illustrative scholium to fuels some of the most advanced physics on proposition seven of book two of the Ethics the planet). (1677/1985) Spinoza writes, ‘A circle exist- However, Leibniz does not see the ing in nature and the idea of the exist- monadic substances as having both men- ing circle, which is also in God, are one tal and material aspects. Leibniz’s mon- and the same thing . . . therefore, whether ads are fundamentally to be conceived we conceive nature under the attribute mentalistically; they are in a way mental- of Extension, or under the attribute of istic automatons moving from one percep- Thought . . . we shall find one and the same tual or apperceptual state to another, all order, or one and the same connection of exactly according to a God-imposed prede- causes . . . ’. On the downside, Spinoza does fined rule. The physical world is a kind of log- have to assume that every physical event has ical construction out of these mental states, a corresponding mental event, and he is thus one which meets various divinely insti- a kind of panpsychist. Even worse (from a tuted constraints upon the relation between 17th-century point of view) Spinoza’s view those aspects matching what we call ‘mate- is heretical, because it sees God as being lit- rial objects’ with those we call ‘states of erally in everything and thus as a material consciousness’ – Leibniz called this the pre- thing not separate from the world. established harmony, and it is his explana- Leibniz never wrote down his meta- tion for the appearance of mind-body inter- physical system in extensive detail (he was action. So Leibniz’s view is one that favours doubtless too busy with a multitude of the mental realm; that is, it is at bottom other projects, such as inventing calculus, a kind of idealism as opposed to Spinoza’s rediscovering binary arithmetic, building the many aspect theory. first calculating machines, and writing end- As we shall see, Leibniz’s vision here had less correspondence and commentary, not a much greater immediate impact on subse- to mention his day job of legal counsel quent philosophy than Spinoza’s. An impor- and historian to the Hanoverian house of tant difference between the two theories is Brunswick), but his views can be recon- that, unlike Spinoza, Leibniz can maintain a structed from the vast philosophical writ- distinction between things that have minds ings he left us. They can be caricatured, or mental attributes from those that do not, in part, as Spinoza’s with an infinite num- despite his panpsychism. This crucial dis- ber of substances replacing the unique God- tinction hinges on the difference between substance. These substances Leibniz called a ‘mere aggregate’ and what Leibniz some- monads (see Leibniz 1714/1989). Because times calls an ‘organic unity’ or an organ- they are true substances, and hence can ism. Each monad represents the world – in exist independently of any other thing, and all its infinite detail – from a unique point because they are absolutely simple, they can- of view. Consider a heap of sand. It cor- not interact with each other in any way responds to a set of monads, but there is (nonetheless they are created by God, who no monad which represents anything like a is one of them – here Spinoza seems rather point of view of the heap. By contrast, your more consistent than Leibniz). Yet each body also corresponds to a set of monads, but monad carries within it complete informa- one of these monads – the so-called domi- tion about the entire universe. What we call nant monad – represents the point of view of space and time are in reality sets of rela- the system which is your living body. (There tions amongst these monads (or, better, the presumably are also sub-unities within you, information which they contain), which are corresponding to organized and functionally P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

22 the cambridge handbook of consciousness

unified physiological, and hence also psycho- V. The Idealist Turn logical, subsystems.) Organisms correspond to a hierarchically ordered set of monads; In some way, Leibniz represents the culmi- mere aggregates do not. This means that nation of the tradition of high metaphysics: there is no mental aspect to heaps of sand the idea that reason could reveal the ulti- as such, even though at the most fundamen- mate nature of things and that this nature tal level mind pervades the universe. is radically different from that suggested by One last point: you might wonder why common sense. But his model of the material you, a monad that represents every detail world as mere appearance was taken to its of the entire universe, seem so relatively logical next step by the, at least superficially, ignorant. The answer depends upon another anti-metaphysical Immanuel Kant (1724– important aspect of the conception of men- 1804). In Kant (see especially 1781/1929)we tality. Leibniz allows that there are uncon- see the beginning of the idealism which in scious mental states. In fact, almost all men- one form or another dominated philosophy tal states are unconscious, and low-level for more than a century afterward. monads never aspire to consciousness (or Once mind is established as the sole real- what Leibniz calls apperception). You are ity, the problem of consciousness and all the aware, of course, only of your conscious other traditional problems of relating matter mental states, and these represent a literally to mind, virtually disappear. The problem infinitesimal fraction of the life of your mind, that now looks big and important is in a way most of which is composed of consciously the inverse of the problem of consciousness: imperceptible petite perceptions (it is galling how exactly is the material world which to think that somewhere within each of our we evidently experience to be constructed minds lies the invisible answers to such ques- out of pure and seemingly evanescent con- tions as whether there are advanced civiliza- sciousness. Two modes of response to this tions in the Andromeda galaxy, but there problem can be traced that roughly divide it is). the thinkers of the British Isles (forgive For Leibniz the material world is, fun- me for including Ireland here) from those damentally, a kind of illusion, but one of of continental Europe, although the geo- a very special kind. What Leibniz calls graphic categorization becomes increasingly ‘well grounded’ phenomena are those that misleading as we enter the 20th century. are in some way directly represented in Very crudely, these modes of idealism can every monad. Imagine aerial photographs be characterized respectively as phenome- of downtown Toronto taken from a vari- nalism (material objects are ‘permanent pos- ety of altitudes and angles. The same build- sibilities of sensation’) and transcendental ings appear in each photograph, though their idealism (a system of material objects repre- appearance is more or less different. But, for sented in experience is a necessary condition example, sun flares caused by the camera for coherent experience and knowledge). lens will not appear in every picture. The There were, of course, materialists lurk- buildings would be termed well grounded, ing about in this period, though they were the sun flare an illusion. So Leibniz can pro- nowhere near the heart of philosophical vide a viable appearance-reality distinction progress; in fact they were frequently not that holds in the world of matter (though it philosophers at all, and quite a number came is tricky, because presumably the illusions of from the ranks of intellectually inclined any one monad are actually reflected in all medical doctors. One such was Julien de La monads – hence the weasel word ‘directly’ Mettrie (1709–1751) who outraged Europe, above). Nonetheless, it is the domain of con- or at least enough of France to require a sciousness which is fundamental and, in the retreat to Berlin, with his L’Homme machine end, the totality of reality, with the physical (1748/1987) (see also the slightly earlier world being merely a kind of construction L’Histoire naturelle de l’âme; 1745). In this out of the mental. brisk polemical work, La Mettrie extends P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

a brief history of the philosophical problem of consciousness 23

the Cartesian thesis that animals are ‘mere’ disconnection between experience and the machines to include the human animal. material world enshrined in Descartes’ dual- But of note here is the same reluctance ism. But what if the material world was to shed all reference to the specialness of somehow really a feature of the realm of the mind that we observed in earlier mate- consciousness for which we obviously seem rialists. La Mettrie is willing to deny that to have infallible access? For example, sup- there are immaterial mental substances, but pose, as did George Berkeley (1685–1753), describes matter as having three essential that material objects are nothing but ordered attributes: extension, motion, and conscious- sequences of perceptions (1710/1998). We ness.InL’Histoire naturelle de l’âme (1745), know by introspection that we have percep- La Mettrie makes the interesting points that tions and that they obey certain apparent the intrinsic nature of matter is utterly mys- laws of succession. Under Berkeley’s identifi- terious to us and that the attribution of cation we thereby know that there are mate- mental properties to it should be no less rial objects and the epistemological crisis is strange than the attribution of extension and resolved. motion, in the sense that we understand On the other side of the English Chan- what it is about matter itself that supports nel, Kant was investigating the deeper prob- extension no better – that is not at all – than lem of how we could know that our percep- we understand how it can or cannot support tions really do follow law-governed patterns mental properties. This idea has remained an which guarantee that they can be interpreted important, if somewhat marginalized, part in terms of a scientifically explicable mate- of the debate about the relation between rial world. Kant accepted Leibniz’s view that mind and matter. Although not always very all we could possibly have knowledge of are clear about their own positions, most mate- constructions out of subjective experience rialists or quasi-materialists of the period, and that any distinction between reality and such as John Toland (1670–1722), Paul-Henri appearance within the realm of perception D’Holbach (1723–1789) (see Holbach 1970), and scientific investigation would have to Joseph Priestley (1733–1804) (see Priest- be based upon some set of relations hold- ley 1975), and Pierre-Jean-George Caba- ing amongst our experiences. He added the nis (1757–1808), agreed on the approach remarkable idea that these relations were that denies substance dualism while allow- a reflection of the structure of the mind ing that matter may have properties that itself – concepts of space, time, and causa- go beyond motion and extension, promi- tion are necessary conditions for the exis- nent among which are the various mental tence of experience of an ‘external world’ attributes. and are ‘discovered’ in that world because The tide of philosophy was, however, run- they pre-exist in the mind. There is no rea- ning in favour of idealism. A central reason son at all to suppose that they reflect some for this was independent of the problem of deeper reality beyond appearances. But they consciousness, but stemmed from the epis- are a necessary condition for having coherent temological crisis brought about by Carte- experience at all and hence will and must be sian philosophy (itself but a partial reflection discovered in the world which is a construct of the general cultural upheaval occasioned out of such experience. Kant called this style by the scientific revolution). Descartes had of reasoning transcendental argumentation. argued that the true nature of the world was In one sense, however, Kant was an impure quite unlike that apparently revealed by idealist. He allowed for the existence of the the senses, but that reality could be dis- ‘thing-in-itself’: the necessarily unknowable, covered by the ‘light of reason’. Unfortu- mind-independent basis of experience. In nately, although everyone took to heart the this respect, Kant’s philosophy is somewhat skeptical challenge to conventional wisdom, like Spinoza’s, save of course that Spinoza Descartes’ positive arguments convinced was fully confident that reason could reveal hardly anybody. The core problem was the something of the thing-in-itself. Idealists P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

24 the cambridge handbook of consciousness

who followed Kant, such as A. Schopen- view of life and its conflict with any material- hauer (1788–1869) and the absolute idealist ist account can be traced back at least to the G. Hegel (1770–1831) and many other conti- 17th century. Another of our philosophically nental philosophers, as well as the later fol- inclined physicians, Nehemiah Grew (1641– lowers of Hegel, such as the British philoso- 1712), who helped found scientific botany pher F. Bradley (1846–1924), espoused purer and was secretary of the Royal Society in forms of idealism (see Bradley 1987/1966; 1677, quaintly put the problem thus (see Hegel 1812/1969; Schopenhauer 1819/1966). Garrett 2003), perhaps not atypically con- Kant’s hypothesis that it was a ‘transcenden- fusing the issues of life and consciousness: tal’ condition for the very possibility of intro- spectible experience to be lawfully ordered The Variety of the Mixture, will not suf- led to a huge philosophical industry focused, fice to produce Life . . . Nor will its being to put it crudely, on the mind’s contribu- mechanically Artificial. Unless the Parts of a Watch, set, as they ought to be, together; tion to the structures we find in the external may be said to be more Vital, than when world (an industry that eventually leads into they lye in a confused Heap. Nor its being the postmodern desconstructionist ghetto). Natural. There being no difference, between But this industry, by its nature, did not face the Organs of Art and Nature; saving, the problem of consciousness as defined here that those of Nature are most of all Arti- and so is not a main player in the drama of ficial. So that an Ear, can no more hear, this chapter. by being an Organ; than an Artificial Ear would do . . . And although we add the Auditory nerves to the Ear, the Brain to VI. Evolution and Emergence the Nerves, and the Spirits to the Brain; yet is it still but adding Body to Body, Art to Subtility, and Engine or Art to Art: Instead, as we now enter the heart of the 19 Which, howsoever Curious, and Many; th century, two crucial non-philosophical can never bring Life out of themselves, nor developments transformed the problem: the make one another to be Vital. (Grew 1701, rise of Darwinism in biology and, drawn 33) from the brow of philosophy itself, the beginning of scientific psychology. Above Vitalism flourished in the 19th century and all else, the evolutionary theory of Charles persisted into the 20th, notably in the writ- Darwin (1809–1882) promised to unify the ings of Hans Driesch (1867–1941), who had simple with the complex by suggesting discovered that fragments of sea urchin some way that mere atoms could, guided embryos would develop into normal sea only by the laws of physics, congregate urchins, contrary to then-current mecha- into such complex forms as plants, animals, nist theory (indeed it is hard to understand and even human beings. This led immedi- how a few of the parts of a machine would ately to two deep questions: what is life go on to operate exactly as did the origi- and how does matter organized via natural nal whole machine). Vitalists thus assumed selection acquire consciousness? These are there must be some special added feature to both questions about emergence, for it cer- living things which accounted for the abil- tainly appears, if evolution be true, that life ity to organize and reorganize even in the springs forth from the lifeless and conscious- face of such assaults. It was the unfortu- ness appears in beings evolved from non- nately delayed development of Mendelian conscious ancestors composed of utterly ‘information-based’ genetics which sug- non-conscious parts. The first question led gested the answer to Driesch’s paradox and to the vitalism controversy, which bears led to the successful integration of evolution some analogy to the problem of conscious- and heredity. ness. Vitalists contended that there was For our purposes, the decline of vitalism something more to life than mere material not only provides a cautionary tale but also organization: a vital spark, or elan´ vital. This highlights an important disanalogy between P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

a brief history of the philosophical problem of consciousness 25

the problems of life and consciousness. Life But as noted above, in the most basic was seen to be problematic from the mate- sense of the term, consciousness seems to rialist point of view because of what it could be an all-or-nothing affair. No non-conscious do, as in Driesch’s sea urchins. It seemed precursor state seems to give the slightest hard to explain the behavioral capacities of hint that consciousness would be its evolu- living things in terms of non-organic science. tionary successor. The tiniest spark of feeling Perhaps conscious beings, as living things, and the weakest and most obscure sensation present the same problem. But this difficulty are fully states of consciousness. Thus the was ultimately swept away with the rise of emergence of consciousness at some point genetics as an adjunct to evolutionary theory. in evolutionary history appears to be an However, in addition to and independent of intrusion of true novelty at odds with the the puzzle of behaviour, consciousness has smoothly evolving complexity of organisms. an internal or subjective aspect, which life, as William Clifford (1845–1879), a tragically such, utterly lacks. What is especially prob- short-lived philosophical and mathematical lematic about consciousness is the question genius (he anticipated general relativity’s of why or how purely material systems could unification of gravity with geometry and pre- become such that there is ‘something it is dicted gravitational waves), put the problem like’ to be them. thus: Another aspect of Darwinism played directly into the mind-matter debate. ...we cannot suppose that so enormous a Darwin himself, and for a long time all jump from one creature to another should Darwinists, was a committed gradualist and have occurred at any point in the process assumed that evolution worked by the long of evolution as the introduction of a fact entirely different and absolutely separate and slow accumulation of tiny changes, with from the physical fact. It is impossible for no infusions of radically new properties at anybody to point out the particular place in any point in evolutionary history. Applying the line of descent where that event can be gradualism to the mind, Darwin went out supposed to have taken place (1874/1886, of his way to emphasize the continuity in 266). the mental attributes of animals and humans (see Darwin 1874). So, although Darwinism provided great sup- Gradualism has its difficulties, which have port for and impetus to the materialist vision long been noted and persist to this day in talk of the world, within it lurked the old, and of punctuated equilibrium (see Eldredge & still unresolved, problem of emergence. Gould 1972) and so-called irreducible com- Perhaps it was time to tackle the mind plexity (Behe 1998, for example). The evolu- directly with the tools of science. During the tion of the eye was seen as very hard for evo- 19th century, psychology broke away from lution to explain even by Darwin himself: philosophy to become a scientific discipline ‘To suppose that the eye . . . could have in its own right. Despite the metaphysi- been formed by natural selection, seems, I cal precariousness of the situation, no one freely confess, absurd in the highest degree’ had any doubt that there was a correspon- (1859/1967, 167). Of course, the idea that a dence between certain physical states and fully formed eye could appear as the result mental states and that it ought to be pos- of one supremely lucky mutational accident sible to investigate that correspondence sci- is truly absurd and is not what is at issue entifically. The pseudo science of phrenol- here. But Darwin went on to give some basis ogy was founded on reasonably acceptable for how the evolution of the eye was pos- principles by Franz Gall (1758–1828) with sible, and there are nowadays sophisticated the aim of correlating physical attributes of accounts of how complex multi-part organs the brain with mental faculties, of which can evolve, as well as compelling theories of Gall, following a somewhat idiosyncratic the evolution of particular organs, such as system of categorization, counted some two the eye (see e.g., Dawkins 1995). dozen, including friendship, amativeness, P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

26 the cambridge handbook of consciousness

and acquisitiveness. True, the categoriza- tinguish the study of the structure of con- tion used is quaint and bizarrely ‘high-level’ sciousness from the question of the ultimate and Gall’s shortcut methodology of infer- nature of consciousness and its place in the ring brain structure from bumps on the skull natural world. The pioneers of psychology dubious (to say the least), but the core idea were, so to speak, officially interested in the retains vigorous life in today’s brain imag- structure of consciousness, both its intro- ing studies and the theory of mental/brain spectible experiential structure and its struc- modules. As D. B. Klein said, ‘Gall gave tural link to physical conditions (both inter- wrong answers to good questions’ (Klein, nal and external to the body). 1970, 669). The growth of interest in these questions Throughout the 19th century, one of the can also be seen in more purely philosoph- primary activities of psychological science ical work in the rise of the phenomeno- and even the main impetus for its creation logical movement, although of course the was discovering correlations between psy- philosophers were not particularly inter- chological states and physical conditions, ested in investigating correlations between either of the environment of the subject or mental and material conditions but rather brain anatomy discovered via postmortem focused on the internal structure of pure investigation (but not of course brain states, consciousness. Phenomenology was fore- which were entirely inaccessible to 19th- shadowed by Franz Brentano (1838–1917), century science). Unlike in the quackery of who in a highly influential work, Psychology phrenology, genuine and profound advances from the Empirical Standpoint (1874/1973, were made. Following foundational work 121 ff.), advanced the view that mental on the physical basis of sensation by Her- states were self-intimating coupled with an mann Helmholtz (1821–1894), also famed updated version of Aristotle’s regress argu- for introducing the hypothesis that uncon- ment (Brentano rather generously credits scious inference accounts for many aspects Aristotle for his whole line of thought here). of cognition and perception, important dis- Brentano also reminded philosophers of coveries included the connection between a feature of mental states which had almost certain brain regions and linguistic ability in been forgotten since it had first been noted the work of Paul Broca (1824–1880); sem- in the Middle Ages (though Descartes’ inal studies of stimulus strength and the notion of objective reality is closely related). introspected intensity of sensation by Gus- Brentano labeled this feature intentionality, tav Fechner (1803–1887), who coined the which is the ‘directedness’ or ‘aboutness’ of phrase ‘psycho-physical law’ (1946), and the at least many mental states onto a content, creation of the first psychological labora- which may or may not refer to an exist- tory devoted to such studies by Wilhelm ing object. If I ask you to imagine a uni- Wundt (1832–1920), who also developed corn, you are easily able to do so, despite the first distinctive research methodology the fact that there are no unicorns. Now, of psychology – that of introspectionism what is your thought about? Evidently not (1892/1894). any real unicorn, but neither is your thought From the point of view of the problem about the image of a unicorn produced in of consciousness these developments point your imagination or even just the idea of a to a bifurcation in the issue. Almost all the unicorn. For if I asked you to think about thinkers associated with the birth of psy- your image or idea of a unicorn you could chology endorsed some form of idealism as do that as well, but it would be a differ- the correct metaphysical account of mind ent thought, and a rather more complex and matter, and none of prominence were one. One way to think about this is to say materialists. They were nonetheless keen on that any act of imagination has a certain studying what we would call the neural bases representational content, and imagining a of consciousness and never questioned the unicorn is simply the having of a partic- legitimacy of such studies. It is useful to dis- ular unicorn-content (in the ‘appropriate’ P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

a brief history of the philosophical problem of consciousness 27

way as well, for imagination must be distin- of bile. Just which part of the brain gener- guished from other content-bearing mental ates these ‘secretions’ and how do they man- acts). The consciousness involved in such an age to possess representational or phenome- act of imagination is the presentation of that nal content? Nonetheless, the metaphor was content to your mind. This is not to say that effective. It was approved by Darwin him- you are aware of your mental state whenever self (a closet materialist), who (privately) you imagine, but rather it is through having endorsed its repetition by John Elliotson such a state that you are conscious of what (1791–1868) – physician, phrenologist, mes- the state represents, although Brentano him- merist, and the so-called strongest materi- self held that any conscious state presented alist of the day (see Desmond and Moore itself as well as its content to the subject. 1994, 250 ff.). In one of his private notebooks The failure to notice the intentional aspect Darwin modified the metaphor in an inter- of consciousness had bedeviled philosophy, esting way, writing, ‘Why is thought, being leading to a plethora of theories of thought a secretion of brain, more wonderful than and perception that left us in the awkward gravity as a property of matter’? This composition of never being aware of anything but ment is striking because it clarifies how the our own mental states. metaphor implicitly suggests that it is a brute Brentano went so far as to declare inten- fact that brains produce thought, just as it is a tionality the mark of the mental, the unique brute fact that matter is associated with grav- property that distinguished the mental from itation. Note also how the power to gravi- the physical. Of course, many other things, tate seems remote from matter’s core prop- such as pictures, words, images on television, erties of extension, exclusion, and mass and, electronic computation, and so on, have rep- at least in the Newtonian view, provides the resentational content, but arguably these all almost miraculous ability to affect all things get their content derivatively, via a men- instantaneously at a distance. Nevertheless, tal interpretation. Uniquely mental or not, the essential emptiness of the metaphor did intentionality poses an extremely difficult not go unremarked. William James (1842– question: how is it that mental states (or 1910) wrote ‘the lame analogy need hardly anything else) can acquire representational be pointed out . . . we know of nothing con- content? Perhaps if one accepts, as so many nected with liver and kidney activity which of the thinkers of this period did, that mind can be in the remotest degree compared is the bedrock reality, then one can accept with the stream of thought that accompanies that it is simply a brute fact, an essential the brain’s material secretions’ (1890/1950, property of mentality, that it carry represen- 102–3). tational content. No explanation of this basic Leaving aside once again this metaphys- fact can be given in terms of anything sim- ical issue, workers focused on the structure pler or more fundamental. and meaning of the contents of conscious- However, if one aspires to a materialist ness along with their empirically deter- account of mind, then one cannot avoid this minable relationship to a host of internal and issue. A frequent metaphor which materi- external material conditions. I have referred alists of the time appealed to was that of to early scientific psychology above, but biological secretion, perhaps first explicitly I would also put Sigmund Freud (1856– articulated by Cabanis in his Rapports du 1939) in this group. Although an advo- physique et du moral de l’homme (1802/1981), cate of materialism, his theory of mind was who proclaimed that the brain secretes focused on psychological structure, rather thought as the liver secretes bile. As it stands, than explications of how matter gives rise this is little more than a declaration of loy- to mind. In philosophy, this emphasis even- alty to the materialist viewpoint, for we tually led to the birth of a philosophical expect there should be an explication of the viewpoint explicitly dedicated to investi- process and nature of such secretion just as gating the inner structure of consciousness: there is such an account of the production the phenomenology of Edmund Husserl P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

28 the cambridge handbook of consciousness

(1859–1938; see Chapter four). In the In a notebook he wrote that ‘the constitu- newly scientific psychology under the guid- tion of reality which I am making for is of ance of Wundt, introspection became the the psychic type’ (see Cooper 1990). paradigmatic research methodology, raising This lack of clarity may arise from the such fundamental questions as whether all epistemological asymmetry between our thought was necessarily accompanied by, or apprehension of mind and matter. We seem even constituted out of, mental imagery. to have some kind of direct access to the for- Unfortunately, this methodology suffered mer – when we feel a pain there is an occur- from inherent difficulties of empirical verifi- rence, at least some properties of which are cation and inter-observer objectivity, which made evident to us. We do not seem to have eventually brought it into disrepute, proba- any similarly direct awareness of the nature bly overall to the detriment of scientific psy- of matter. Thus the avowed neutrality of chology. neutral monism tends to slide towards some Though James decried the gross kind of panpsychism. From another point of metaphor of consciousness as a brain view, the asymmetry encourages the associ- secretion, he introduced one of the most ation of some forms of phenomenalism with potent and durable metaphors, that of the neutral monism. stream of consciousness. In his remarkably For example, the highly influential British compendious work, The Principles of Psy- philosopher John Stuart Mill (1806–1873) chology, which remains to this day full of endorsed a phenomenalism which regarded fresh insight, James devoted a chapter to material objects as ‘permanent possibilities the stream of thought in which he noted of sensation’. This allows for the interpo- that ongoing consciousness is continuous, sition of a something we know not which meaning ‘without breach, crack or division’ lurking behind our sensations (what might (1890/1950, 237) and that, by contrast, ‘the be called ‘unsensed sensibilia’ – see Mill breach from one mind to another is perhaps 1865/1983 and Wilson 2003), but the seem- the greatest breach in nature’. James of ingly unbridgeable gap between this ur- course allowed that there were noticeable matter and our perceptual experiences cre- gaps in one’s stream of consciousness, but ates a constant pressure to replace it with these are peculiar gaps such that we sense entirely mental sequences of sensations. To that both sides of the gap belong together in be sure, intuition suggests that material some way; he also noted that the stream is objects exist unperceived, but this ‘exis- a stream of consciousness and that unnoticed tence’ can, perhaps, be analysed in terms temporal gaps – which are perfectly con- of dispositions to have certain sensations ceivable – are simply not part of the stream. under certain mentalistically defined condi- Throughout his writings James exhibits a tions. Furthermore, as our relation with the keen and durable interest in the structure unknowable basis of matter is entirely men- and contents of the stream of consciousness, talistic, why not accept that the primal mate- even delving enthusiastically into mystical rial is itself mental (a view which can lead and religious experience. either back to idealism or to some form Along with virtually all psychological of panpsychism)? Bertrand Russell (1872– researchers of the time, James was no mate- 1970) devoted great effort to developing rialist. His metaphysics of mind is complex Mill’s phenomenalism as a kind of neutral and somewhat obscure, wavering between monism (see Russell 1927) in which what a neutral monism and a form of panpsy- we call matter has intrinsic mental proper- chism (see Stubenberg 2005). James heaped ties with which we are directly acquainted scorn (and powerful counter-arguments) in experience – thus, Russell’s seemingly upon crude forms of ‘molecular’ panpsy- bizarre remark that when a scientist exam- chism, what he called the ‘mind dust’ the- ines a subject’s brain he is really observing a ory (see 1890/1950, ch. 5), but his monism part of his own brain (for an updated defense leaned decidedly towards the mental pole. of a Russellian position see Lockwood 1991). P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

a brief history of the philosophical problem of consciousness 29

His one-time collaborator, Alfred North Mill attempted a compendious classifica- Whitehead (1861–1947), pursued the alter- tion of scientific law, two forms of which native panpsychist option in a series of works he called ‘homopathic’ and ‘heteropathic’. culminating in the dense and obscurely Homopathic laws are ones in which the written Process and Reality (1929). Roughly resultant properties of a system are the mere speaking Whitehead proposed a radical additive results of the properties of the sys- reform of our conception of the fundamen- tem’s components. For example, the laws tal nature of the world, placing events (or of motion are homopathic: the motion of items that are more event-like than thing- an object is the result of all the forces act- like) and the ongoing process of their cre- ing on the object, and the resultant force ation as the core feature of the world, rather is simply the vector addition of each sep- than the traditional triad of matter, space, arate force. Heteropathic laws are ones in and time. His panpsychism arises from the which the resultant properties are not sim- idea that the elementary events that make ply the sum of the properties of the compo- up the world (which he called occasions) par- nents. It was George Lewes (1817–1878)– take of mentality in some often extremely now best remembered as the consort of attenuated sense, metaphorically expressed George Elliot – who coined the term ‘emer- in terms of the mentalistic notions of cre- gent’ to refer to heteropathic effects (he used ativity, spontaneity, and perception. White- ‘resultant’ to mean those features which Mill head’s position nicely exposes the difficulty called homopathic effects). Here it is impor- in maintaining a pure neutral monism. Mat- tant to distinguish the more general notion of ter must have some underlying intrinsic homopathic effects from what is sometimes nature. The only intrinsic nature we seem to called part-whole reductionism. The latter be acquainted with is consciousness. Thus it may well be false of the world: there are rea- is tempting to simplify our metaphysics by sonable arguments that some physical prop- assigning the only known intrinsic nature to erties are non-local and perhaps in some way matter. We thus arrive at panpsychism rather holistic (both general relativity and quan- than neutral monism (for an introduction to tum mechanics can be invoked to support Whitehead’s philosophy of mind see Griffin these contentions). But the crucial ques- 1998). tion about homopathic versus heteropathic Such high metaphysical speculations, effects is whether the fundamental physi- though evidently irresistible, seem far from cal state of the world, along with the basic the common-sense view of matter which physical laws, determines all other, higher- was more or less enshrined in the world view level properties and laws. If not, we have true of 19th-century science, which began then emergence. to fund the rapid and perpetual develop- The emergentists postulated that con- ment of technology we are now so famil- sciousness was a heteropathic effect, or iar with, and which greatly added to the emergent property, of certain complex social prestige of science. If we take the sci- material systems (e.g., brains). Emergentism entific picture seriously – and it came to may seem no more than extravagant meta- seem irresponsible not to – then the cen- physical speculation, except that at the time tral mystery of consciousness becomes that it was widely conceded that there were of the integration of mind with this scientific excellent candidate emergent properties in viewpoint. This is the modern problem of areas other than consciousness. That is, it consciousness, which bypasses both idealist seemed there were independent grounds metaphysics and phenomenalistic construc- for endorsing the concept of emergence, tionism. which – thus legitimated – could then be But how could such integration be fruitfully applied to the mind-body prob- achieved? An important line of thought lem. The primary example of supposedly begins with some technical distinctions of uncontentious emergent properties were Mill. In his System of Logic (1843/1963) those of chemistry. It was thought that, for P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

30 the cambridge handbook of consciousness

example, the properties of water could not novo calculation of some simple chemical be accounted for in terms of the proper- features based on the quantum mechani- ties of oxygen and hydrogen and the laws cal description of atomic components). Of of nature which governed atomic-level phe- course, this does not demonstrate that there nomena. Emergentists, of which two promi- is no real emergence in the world, but with- nent ones were Conwy Lloyd Morgan (1852– out any uncontentious example of it, and 1936) (see Morgan 1923) and C. D. Broad with the growing ability of physics to pro- (1887–1971), recognized that the complex- vide seemingly complete accounts of the ity of the interactions of the components basic structure of the world, the emergen- of a system could present the appearance tist position was devastated (see McLaughlin of emergence when there was none. Broad 1992). (1925) liked to imagine a ‘mathematical On the mainstream view articulated in archangel’ who knew the laws of nature as Section II, emergentism seems no less meta- they applied at the submergent level, knew physically extravagant than the other posi- the configuration of the components, and tions we have considered. Emergentism suffered from no cognitive limitations about espouses a form of property dualism and deducing the consequences of this informa- postulates that the novel emergent prop- tion. If the archangel could figure out that erties of a system would have distinctive water would dissolve salt by considering only causal powers, going beyond those deter- the properties of oxygen, hydrogen, sodium, mined solely by the basic physical fea- and chlorine as well as the laws which gov- tures of the system (seemingly courting vio- erned their interaction at the atomic level, lation of a number of basic conservation then this aspect of H2 O would fail to be an laws). emergent property. Not that the science of psychology pro- Thus the emergentists would have scoffed vided a more palatable alternative. Early at current popular examples of emergence, in the 20th century, the introspectionist such as John Conway’s Game of Life and methodology as well as the sophisticated chaotic dynamical systems. Such examples sensitivity to the issues raised by con- represent nothing more than ‘epistemologi- sciousness by such psychologists as James cal emergence’. Cognitive and physical limi- disappeared with the rise of a soulless tations – albeit quite fundamental ones – on behaviourism that at best ignored the mind computational power and data acquisition and at worst denied its very existence. It prevent us (or our machines) from deducing took until halfway through the 20th cen- the high-level properties of complex sys- tury before philosophy and psychology grap- tems, but this is not a metaphysical bar- pled with the problem of the mind in new rier. The mathematical archangel could fig- ways. In psychology, the so-called cognitive ure out the effect of the butterfly’s flight on revolution made an appeal to inner mental future weather. processes and states legitimate once again But the emergentists believed that the (see Neisser 1967 for a classic introduction world really did contain non-epistemological to cognitive psychology), although scientific emergence; in fact, it was virtually every- psychology largely steered clear of the issue where. They regarded the world as an of consciousness until near the end of the hierarchical cascade of emergent features 20th century. built upon other, lower level emergent fea- In philosophy, the 1950s saw the begin- tures. Unfortunately for them, their linchpin ning of a self-conscious effort to under- example, chemistry, was the masterpiece of stand the mind and, eventually, conscious- the new quantum mechanics of the 1920s, ness as physical through and through in which basically provided new laws of nature essentially scientific terms. This was part of which opened the door – in principle – to a broader movement based upon the doc- the deduction of chemical properties from trine of scientific realism, which can be atomic states (nowadays we even have de roughly defined as the view that it is science P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

a brief history of the philosophical problem of consciousness 31

that reveals the ultimate nature of real- Bloom, P. (2004). Descartes’s baby: How ity, rather than philosophy or any other the science of child development explains non-empirical domain of thought. Applied what makes us human. New York: Basic to the philosophy of mind and conscious- Books. ness, this led to the rise of the identity the- Bradley, F. (1966). Appearance and reality (2nd ory (see e.g., Smart 1959, or for a more ed). Oxford: Clarendon Press. (Original work 1897 penetrating and rather prescient account, published ) 1973 Feigl 1958). This was a kind of turning Brentano, Franz C. ( ). Psychology from an point in philosophy, in which it was self- empirical standpoint. London: Routledge and Kegan Paul. (Original work published 1874.) consciously assumed that because physical 1925 science should be the basis for our beliefs Broad, C. D. ( ). The mind and its place in nature. London: Routledge and Kegan about the ultimate nature of the world, phi- Paul. losophy’s job in this area would henceforth Cabanis, P. (1802/1981). Rapports du physique et be to show how the mind, along with every- du moral de l’homme (On the relations between thing else, would smoothly fit into the sci- the physical and moral aspects of man; G. Mora, entific picture of the world. It embraced Ed. & M. Saidi, Trans). Paris: Crapart, Caille et what I earlier called the mainstream view Ravier. and began with high optimism that a scien- Caston, V. (2002). Aristotle on consciousness. tific outlook would resolve not just the prob- Mind, 111(444), 751–815. lem of consciousness but perhaps all philo- Clifford, William K. (1874). Body and mind. Fort- sophical problems. But subsequent work has nightly Review, December. (Reprinted in Leslie revealed just how extraordinarily difficult it Stephen & Frederick Pollock (Eds.), Lectures is to fully explicate the mainstream view, and essays, 1876, London: Macmillan) especially with regard to consciousness. In Churchland, Paul (1981). Eliminative materialism the face of unprecedented expansion in our and propositional attitudes. Journal of Philoso- technical abilities to investigate the workings phy, 78, 67–90. of the brain and an active and explicit scien- Cooper, W. E. (1990). William James’s theory tific, as well as philosophical, effort to under- of mind. Journal of the History of Philosophy, stand how, within the mainstream view, con- 28(4), 571–93. sciousness emerges, we find that the ultimate Darwin, C. (1967). The origin of species by natural problem remains intractable and infinitely selection. London: Dent (Everyman’s Library). 1859 fascinating. (Originally published .) Darwin, C. (1874). The descent of man. London: Crowell. 1995 References Dawkins, R. ( ). River out of Eden. New York: Basic Books. Descartes, R. (1985). Meditations on first phi- 1984 Aristotle ( ). The complete works of Aristotle (J. losophy. In J. Cottingham, R. Stoothoff, & D. Barnes, Ed.). Princeton: Princeton University Murdoch (Eds.), The philosophical writings of Press. Descartes. Cambridge: Cambridge University Augustine (1998). Confessions. Oxford: Oxford Press. (Originally published 1641.) University Press. Descartes, R. (1985). The passions of the soul. Augustine (1998). City of God. Cambridge: Cam- In J. Cottingham, R. Stoothoff, & D. Mur- bridge University Press. doch (Eds.), The philosophical writings of Barnes, J. (1982). The Presocratic philosophers. Descartes. Cambridge: Cambridge University London: Routledge and Kegan Paul. Press. (Originally published 1649.) Behe, M. (1998). Darwin’s black box: The biochem- Desmond, A., & Moore, J. (1994). Darwin: The ical challenge to evolution. New York: Free Press. life of a tormented evolutionist. New York: Berkeley, G. (1998). A treatise concerning the prin- Norton. ciples of human knowledge. Oxford: Oxford Edwards, P. (1967). Panpsychism. In P. Edwards University Press. (Original work published (Ed.), The encyclopedia of philosophy (Vol. 5). 1710.) New York: Macmillan. P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

32 the cambridge handbook of consciousness

Eldredge, N. & Gould, S. (1972). Punctuated Klein, D. (1970). A history of scientific psychology. equilibria: An alternative to phyletic gradual- New York: Basic Books. ism. In T. Schopf (Ed.), Models in paleobiology. La Mettrie, J. (1745). L’histoire naturelle de l’âme. San Francisco: Freeman Cooper. The Hague. Fechner, G. (1946). The religion of a scientist (W. La Mettrie, J. (1987). L’homme machine / Man a Lowrie, Ed. & Trans). New York: Pantheon. Machine. La Salle, IL: Open Court. (Originally Feigl, H. (1958). The mental and the physical. In published 1748.) H. Feigl, M. Scriven, & G. Maxwell (Eds.), Min- Leibniz, G. (1989). Monadology. In R. Ariew & nesota studies in the philosophy of science: Vol. 2 . D. Garber (Eds. & Trans.), G. W.Leibniz: Philo- Concepts, theories and the mind-body problem. sophical essays. Indianapolis: Hackett. (Origi- Minneapolis: University of Minnesota Press. nally published 1714.) Galileo, G. (1957). The assayer. In D. Stillman Lockwood, M. (1991). Mind, brain and the quan- (Ed.), Discoveries and opinions of Galileo. New tum. Oxford: Blackwell. 1623 York: Doubleday. (Originally published .) Mangan, B. (2001). Sensation’s ghost: The Garrett, B. (2003). Vitalism and teleology in the non-sensory ‘fringe’ of consciousness. Psy- natural philosophy of Nehemiah Grew, British che, 7(18). http://psyche.cs.monash.edu.au/v7/ Journal for the History of Science 36(1), 63– psyche-7–18-mangan.html. 81 . Matson, W. (1966). Why isn’t the mind-body Gergeley, G, Z., Nadasdy,´ G. Csibra, & Biro,´ problem ancient? In P. Feyerabend & G. S. (1995). Taking the intentional stance at 12 Maxwell (Eds.), Mind, matter and method: months of age. Cognition, 56(2), 165–93. Essays in philosophy and science in honor of Grew, N. (1701). Cosmologia sacra: or a dis- Herbert Feigl. Minneapolis: University of Min- course of the universe as it is the creature and nesota Press. kingdom of God. London: Rogers, Smith and McLaughlin, B. (1992). The rise and fall of British Walford. emergentism. In A. Beckermann, H. Flohr, & J. Griffin, D. (1998). Unsnarling the world knot: Con- Kim (Eds.), Emergence or reduction? Berlin: De sciousness, freedom and the mind-body problem. Gruyter. Berkeley: University of California Press. Mill, J. (1963). A system of logic. In J. Robson (Ed.), Hegel, G. (1969). The science of logic (A. Miller, Collected works of John Stuart Mill (Vols. 7 & 8). Trans.). London: Allen and Unwin. (Originally Toronto: University of Toronto Press. (Origi- published 1812.) nally published 1843.) Hobbes, T. (1998). Leviathan (J. Gaskin, Ed.). Mill, J. (1963). An examination of Sir William Oxford: Oxford University Press. (Originally Hamilton’s philosophy, In J. Robson (Ed.), published 1651.) Collected works of John Stuart Mill (Vol. 9). Holbach, Baron d’ (1970). Système de la nature; Toronto: University of Toronto Press. (Origi- ou, Des lois du monde physique et du monde nally published 1865.) moral (published under the pseudonym J. Morgan, C. (1923). Emergent evolution. London: Mirabaud) (The system of nature: or, laws of the Williams and Norgate. moral and physical world; H. Robinson, Trans.). Mourelatos, Alexander (1986). Quality, struc- (Originally published 1770.) ture, and emergence in later Pre-Socratic phi- Huxley, T. (1866). Lessons in elementary physiol- losophy. Proceedings of the Boston Area Collo- ogy. London: Macmillan. quium in Ancient Philosophy, 2 , 127–94. James, W.(1950). The principles of psychology (Vol. Nagel, T. (1974). What is it like to be a bat? Philo- 1). New York: Dover. (Originally published sophical Review, 83, 435–50. 1890.) Neisser, U. (1967). Cognitive psychology. New Kant, I. (1929). Critique of pure reason (N. Kemp York: Appleton-Century-Crofts. Smith, Ed. & Trans.). New York: St. Macmillan. Nelson, J. (1990). Was Aristotle a functionalist? (Originally published 1781.) Review of Metaphysics, 43, 791–802. King, P. (2005). Why isn’t the mind-body prob- Nussbaum, M. C., & Putnam, H. (1992). lem mediaeval? In H. Lagerlund & O. Pluta Changing Aristotle’s mind. In M. Nussbaum (Eds.), Forming the mind – Conceptions of body & A. Rorty (Eds.), Essays on Aristotle’s and soul in late medieval and early modern phi- De Anima (pp. 27–56). Oxford: Clarendon losophy. Berlin and New York: Springer Verlag. Press. P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

a brief history of the philosophical problem of consciousness 33

O’Daly, G. (1987). Augustine’s philosophy of mind. The philosophy of mind. Englewood Cliffs, NJ: London: Duckworth. Prentice-Hall. Perner, J. (1991). Understanding the repre- Spinoza, B. (1677/1985). Ethics. In E. Curly (Ed. sentational mind. Cambridge, MA: MIT & Trans.), The collected works of Spinoza (Vol. 1). Press. Princeton: Princeton University Press. Plato (1961). Plato: Collected dialogues (E. Hamil- Stubenberg, L. (2005). ‘Neutral Monism’, ton & H. Cairns, Eds.). Princeton: Princeton The Stanford Encyclopedia of Philosophy University Press. (Spring 2005 Edition), Edward N. Zalta (ed.) Priestley, J. (1975). Disquisitions relating to matter http://plato.stanford.edu/. and spirit. New York: Arno Press. (Originally Toland, J. (1704). Letters to Serena, London: published 1777.) Bernard Lintot. Russell, B. (1927). The analysis of matter. London: Whitehead, A. (1929). Process and reality: An Kegan Paul. essay in cosmology. New York: Macmillan. Schopenhauer, A. (1966). The world as will Wilkes, K. (1988). Yishi, duh, um, and con- and representation (E. Payne, Trans.). New sciousness. In A. Marcel & E. Bisiach (Eds.), York: Dover Books. (Originally published Consciousness in contemporary science. Oxford: 1819.) Oxford University Press. Seager, W. (1999). Theories of consciousness, New Wilson, F. (2003). John Stuart Mill. In E. Zalta York: Routledge. (Ed.), Stanford encyclopedia of philosophy. Seager, W. (2005). Panpsychism. In E. Zalta http://plato.stanford.edu/archives/fall2003/ (Ed.), Stanford encyclopedia of philosophy, entries/mill http://plato.stanford.edu/archives/sum2005/ Wundt, W.(1894). Vorlesungenuber ¨ die Menschen- entries/panpsychism. und Thierseele (Lectures on human and ani- Smart, J. (1959). Sensations and brain processes. mal psychology; J. Creighton & E. Titchener, Philosophical Review, 68, 141–56. Reprinted in Trans.). Hamburg: L. Voss. (Originally pub- slightly revised form in V. Chappell (ed.), lished 1892.) P1: JzG 0521857430c02 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:3

34 P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

CHAPTER 3 Philosophical Theories of Consciousness: Contemporary Western Perspectives

Uriah Kriegel

Abstract With each approach, I will present in order (i) the leading account of consciousness This chapter surveys current approaches to along its line, (ii) the case for the approach, 1 consciousness in Anglo-American analytic and (iii) the case against the approach. I will philosophy. It focuses on five approaches, not issue a final verdict on any approach, to which I will refer as mysterianism, dual- though by the end of the chapter it should ism, representationalism, higher-order mon- be evident where my own sympathies lie. itoring theory, and self-representationalism. Before starting, let us draw certain dis- With each approach, I will present in order tinctions that may help fix our ideas for the (i) the leading account of consciousness discussion to follow. The term “conscious- along its line, (ii) the case for the approach, ness” is applied in different senses to dif- and (iii) the case against the approach. I will ferent sorts of things. It is applied, in one not issue a final verdict on any approach, sense, to biological species, as when we say though by the end of the chapter it should something like “Gorillas are conscious, but be evident where my own sympathies lie. snails are not”; in a different sense, to individual organisms or creatures, as when we say “Jim is conscious, but Jill is comatose”; and in a third sense, to particular mental states, Introduction: The Concept events, and processes, as when we say “My of Consciousness thought about Vienna is conscious, but Jim’s belief that there are birds in China is not.” This chapter surveys current approaches to To distinguish these different senses, we may consciousness in Anglo-American analytic call the first species consciousness, the second philosophy. It focuses on five approaches, creature consciousness, and the third state con- to which I will refer as mysterianism, dual- sciousness.2 ism, representationalism, higher-order mon- There appear to be certain concep- itoring theory, and self-representationalism. tual connections among these three senses, 35 P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

36 the cambridge handbook of consciousness

such that they may be analyzable in functions.5 The sense is that an insight of a terms of one another. Plausibly, species completely different order would be needed consciousness is analyzable in terms of to make scientific theories, and indeed sci- creature consciousness: a species S is species- ence itself, at all relevant to our understand- conscious just in case a prototypical spec- ing of phenomenal consciousness. Some sort imen of S is creature-conscious. Creature of conceptual breakthrough, which would consciousness may in turn be analyzable in enable us to conceive of the problem of con- terms of state consciousness: a creature C sciousness in new ways, is required. This is is creature-conscious just in case C has (or where philosophical theories of conscious- is capable of having) mental states that are ness come into the picture.6,7 state-conscious. If so, state consciousness is the most fundamental notion of the three. State consciousness is itself ambiguous as Mysterianism between several senses. If Jim tacitly believes that there are birds in China, but never con- Some philosophers hold that science cannot sciously entertained this belief, whereas Jill and will not, in fact, help us understand con- often contemplates consciously the fact that sciousness. So-called mysterianists hold that there are birds in China, but is not doing so the problem of consciousness – the prob- right now, there is a sense of “conscious” in lem of how there could be something like which we may want to say that Jim’s belief is phenomenal consciousness in a purely nat- unconscious whereas Jill’s is conscious. Let ural world – is not a problem we are capa- 3 us call this availability consciousness. By ble (even in principle) of solving. Thus con- contrast, there is a sense of “conscious” in sciousness is a genuine mystery, not merely which a mental state is conscious when and a prima facie mystery that we may one day only when there is something it is like for the demystify. 4 subject – from the inside – to have it. Thus, We may introduce a conceptual distinc- when I take in a spoonful of honey, there is tion between two kinds of mysterianism – an a very specific – sweet, smooth, honey-ish, ontological one and an epistemological one. if you will – way it is like for me to have the According to ontological mysterianism, con- resulting conscious experience. Let us call sciousness cannot be demystified because it this phenomenal consciousness. is an inherently mysterious (perhaps super- Some of the leading scientific theories natural) phenomenon.8 According to episte- of consciousness – such as Baars’ (1988, mological mysterianism, consciousness is in 1997) Global Workspace Theory and Crick no way inherently mysterious, and a greater and Koch’s (1990, 2003) synchrony-based mind could in principle demystify it – but “neurobiological theory” – shed much light it just so happens that we humans lack the on availability consciousness and neighbor- cognitive capacities that would be required. ing notions. But there is a persistent feel- Epistemological mysterianism has actu- ing that they do not do much to explain ally been pursued by contemporary West- phenomenal consciousness. Moreover, there ern philosophers. The most comprehensive is a widespread sense that there is some- development of the view is offered in Colin thing principled about the way in which McGinn’s (1989, 1995, 1999, 2004) writ- they fail to do so. One way to bring out this ings. We now turn to an examination of his feeling is through such philosophers’ con- account. cepts as the explanatory gap (Levine, 1983)or the hard problem (Chalmers, 1995). Accord- McGinn’s Mysterianism ing to Chalmers, for instance, the problems of explaining the various cognitive func- McGinn’s theory of consciousness has two tions of conscious experiences are the “easy central tenets. First, the phenomenon of con- problems” of consciousness; the “hard prob- sciousness is in itself perfectly natural and lem” is that of understanding why there in no way mysterious. Second, the human should be something it is like to execute these mind’s conceptual capacities are too poor P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 37

to demystify consciousness. That is, McGinn induction. From the fact that all the theories is an epistemological mysterianist: he does of consciousness we have come up with to not claim that the world contains, in and date are hopelessly unsatisfactory, it should of itself, insoluble mysteries, but he does not be concluded that our future theories contend that we will never understand con- will be the same. It may well be that a thou- sciousness. sand years hence we will look back with At the center of McGinn’s theory is the amusement at the days of our ignorance and concept of cognitive closure. McGinn (1989, self-doubt. p. 529) deﬁnes cognitive closure as follows: However, McGinn’s main argument for “A type of mind M is cognitively closed with his position is not the inductive argument respect to a property P (or a theory T) if and just sketched. Rather, it is a deductive argu- only if the concept-forming procedures at ment based on consideration of our cog- M’s disposal cannot extend to a grasp of P (or nitive constitution. The argument revolves an understanding of T).”9 To be cognitively around the claim that we do not have a sin- closed to X is thus to lack the procedure for gle mechanism, or faculty, that can access concept formation that would allow one to both consciousness and the brain. Our access form the concept of X. to consciousness is through the faculty of To illustrate the soundness and appli- introspection. Our access to the brain is cability of the notion of cognitive clo- through the use of our senses, mainly vision. sure, McGinn adduces the case of animal But unfortunately, the senses do not give us minds and their constitutive limitations. access to consciousness proper, and intro- As James Joyce writes in A Portrait of spection does not give us access to the brain the Artist as a Young Man, rats’ minds proper. Thus, we cannot see with our eyes do not understand trigonometry. Likewise, what it is like to taste chocolate. Nor can we snails do not understand quantum physics, taste with our buds what it is like to taste and cats do not understand market eco- chocolate. We can, of course, taste choco- nomics. Why should humans be spared this late. But we cannot taste the feeling of tast- predicament? As a natural, evolved mecha- ing chocolate. The feeling of tasting choco- nism, the human mind must have its own late is something we encounter only through limitations. One such limitation, McGinn introspection. But alas, introspection fails to suggests, may be presented by the phe- give us access to the brain. We cannot intro- nomenon of consciousness. spect neurons, and so could never introspect Interestingly, McGinn does not claim that the neural correlates of consciousness. we are cognitively closed to consciousness Using the term “extrospective” to denote itself. Rather, his claim is that we are cog- the access our senses give us to the world, nitively closed to that property of the brain McGinn’s argument may be formulated as responsible for the production of conscious- follows: ness. As someone who does not wish to por- 1 tray consciousness as inherently mysterious, ) We can have introspective access to con- McGinn is happy to admit that the brain sciousness but not to the brain; has the capacity to somehow produce con- 2) We can have extrospective access to the scious awareness. But how the brain does so is brain but not to consciousness; something he claims we cannot understand. 3) We have no accessing method that is both Our concept-forming procedures do extend introspective and extrospective; there- to a grasp of consciousness, but they do not fore, extend to a grasp of the causal basis of con- 4) We have no method that can give us sciousness in the brain. access to both consciousness and the brain. The Master Argument for Mysterianism As we can see, the argument is based on con- A natural reaction to McGinn’s view is that siderations that are much more principled it may be based upon an overly pessimistic than a simple pessimistic induction from P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

38 the cambridge handbook of consciousness

past theories. Dismayed as we may be by kinds of states a person or organism may be the prospects of mysterianism, we must not in: brain states on the one hand and con- confuse McGinn’s position for sheer despair. scious states on the other. Instead, we must contend with the argument Recall that McGinn’s mysterianism is of just formulated. the epistemological variety. The epistemo- Some materialists would contest the first logical claim now appears to be conditional premise. Paul Churchland (1985) has repeat- upon an ontological claim, namely dualism. edly argued that we will one day be able So at the end of the day, as far as the ontology to directly introspect the neurophysiolog- of consciousness is concerned, McGinn is a ical states of our brains. Perception and straightforward dualist. The plausibility of introspection are theory-laden, according to his (epistemological) mysterianism depends, Churchland, and can therefore be funda- to that extent, on the plausibility of (onto- mentally changed when the theory they are logical) dualism. In the next section, we con- laden with is changed.10 Currently, our intro- sider the plausibility of dualism. spective practice is laden with a broadly Before doing so, let us raise one more diffi- Cartesian theory of mind. But when we culty for mysterianism, and in particular the mature enough scientifically, and when the notion of cognitive closure. It is, of course, right neuroscientific theory of consciousness undeniable that rats do not understand makes its way to our classroom and living trigonometry. But observe that trigonomet- room, this will change and we (or rather ric problems do not pose themselves to rats our distant offspring) will start thinking (Dennett, 1995,pp.381–383). Indeed, it about ourselves in purely neurophysiological is precisely because rats do not understand categories. trigonometry that trigonometric problems Other materialists may deny the second do not pose themselves to rats. For rats to premise of the argument. As long as brain grapple with trigonometric problems, they states are considered to be merely correlates would have to understand quite a bit of of conscious states, the claim that the con- trigonometry. Arguably, it is a mark of gen- scious states cannot be perceived extrospec- uine cognitive closure that certain questions tively is plausible. But according to material- do not even pose themselves to the cogni- ists, conscious states will turn out to be iden- tively closed. The fact that certain questions tical with the brain states in question, rather about consciousness do pose themselves to than merely correlated therewith. If so, per- humans may therefore indicate that humans ceiving those brain states would just be per- are not cognitively closed to consciousness ceiving the conscious states.11 To assume that (or more accurately to the link between con- we cannot perceive the conscious states is to sciousness and the brain).14 beg the question against the materialist.

Dualism The Case Against Mysterianism To repeat the last point, McGinn appears Traditionally, approaches to the ontology of to assume that conscious states are caused mind and consciousness have been divided by brain states. His argument does not go into two main groups: monism and dual- through if conscious states are simply iden- ism. The former holds that there is one kind tical to brain states. In other words, the of stuff in the world; the latter that there argument does not go through unless any are two.15 Within monism, there is a further identity of conscious states with brain states distinction between views that construe the is rejected.12 But such rejection amounts single existing stuff as material and views to dualism. McGinn is thus committed to that construe it as immaterial; the former dualism.13 On the view he presupposes, the are materialist views, the latter idealist.16 conscious cannot be simply identiﬁed with Descartes framed his dualism in terms of the physical. Rather, there are two different two different kinds of substance (where a P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 39

substance is something that can in principle At the same time, Chalmers does not exist all by itself). One is the extended sub- take phenomenal properties to be accidental stance, or matter; the other is the thinking or random superpositions onto the physical substance, or mind. A person, on this view, world. On the contrary, he takes them to be is a combination of two different objects:a causally grounded in physical laws. That is, body and a soul. A body and its correspond- instantiations of phenomenal properties are ing soul “go together” for some stretch of caused by instantiations of physical proper- time, but being two separate objects, their ties, and they are so caused in accordance existence is independent and can therefore with strict laws of nature.21 come apart.17 This means that phenomenal conscious- Modern dualism is usually of a more sub- ness can be explained in physical terms. It is tle sort, framed not in terms of substances just that the explanation will not be a reduc- (or stuffs), but rather in terms of properties. tive explanation, but rather a causal expla- The idea is that even though there is only nation. To explain an event or phenomenon one kind of stuff or substance, there are two causally is to cite its cause, that is, to say kinds of properties, mental and physical, and what brought it about or gave rise to it.22 neither can be reduced to the other.18 This According to Chalmers, one could in princi- is known as property dualism. A particularly ple explain the instantiation of phenomenal cautious version of property dualism claims properties by citing their physical causes. that although most mental properties are A full theory of consciousness would reducible to physical ones, conscious or phe- uncover and list all the causal laws that gov- nomenal properties are irreducible. ern the emergence of phenomenal properties from the physical realm. And a full description of nature and its behavior would Chalmers’ Naturalistic Dualism have to include these causal laws on top For many decades, dualistic arguments were of the causal laws obtained by “ultimate treated mainly as a challenge to a physical- physics.”23 ist worldview, not so much as a basis for Chalmers himself does not attempt to a non-physicalist alternative. Thus dualism detail many of these laws. But he does pro- was not so much an explanation or account pose a pair of principles to which we should of consciousness, but rather the avoidance expect such laws to conform. These are of one. This state of affairs has been recti- the “structural coherence” principle and the ﬁed in the past decade or so, mainly through “organizational invariance” principle. The the work of David Chalmers (1995, 1996, former concerns the sort of direct avail- 2002a). ability for global control that conscious Chalmers’ theory of consciousness, which states appear to exhibit, the latter the sys- he calls naturalistic dualism, is stronger than tematic correspondence between a system’s ordinary dualism, in that it claims not only functional organization and its phenomenal that phenomenal properties are not identical properties.24 to physical properties, but also that they fail to supervene – at least with metaphysical or The Case for Dualism logical necessity19 – on physical properties.20 We tend to think, for instance, that biological The best-known arguments in favor of prop- properties necessarily supervene on physical erty dualism about consciousness are so- properties, in the sense that two systems can- called epistemic arguments. The two main not possibly differ in their biological proper- ones are Frank Jackson’s (1984) “Knowl- ties if all their physical properties are exactly edge Argument” and Thomas Nagel’s (1974) similar. But according to Chalmers, phenom- “what is it like” argument. Both follow a sim- enal properties are different: two systems ilar pattern. After describing a situation in can be exactly the same physically, but have which all physical facts about something are different phenomenal properties. known, it is shown that some knowledge is P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

40 the cambridge handbook of consciousness

still missing. It is then inferred that the miss- tion of a new piece of knowledge one cannot ing knowledge must be knowledge of non- infer the existence of a new fact – and that physical facts. is precisely the inference made in the above The Knowledge Argument proceeds as dualist arguments.27,28 follows. Suppose a baby is kept in a black- A different argument for dualism that is and-white environment, so that she never widely discussed today is Chalmers’ (1996) has color experiences. But she grows to argument from the conceivability of zom- become an expert on color and color vision. bies. Zombies are imaginary creatures that Eventually, she knows all the physical facts are physically indistinguishable from us but about color and color vision. But when she lack consciousness. We seem to be able to sees red for the ﬁrst time, she learns some- conceive of such creatures, and Chalmers thing new: she learns what it is like to see red. wants to infer from this that materialism is That is, she acquires a new piece of knowl- false. The argument is often caricatured as edge. Since she already knew all the physical follows: facts, this new piece of knowledge cannot be 1 knowledge of a physical fact. It is therefore ) Zombies are conceivable; knowledge of a non-physical fact. So, the fact 2) If As are conceivable, then As are (meta- 29 thereby known (what it is like to see red) is physically) possible; therefore, a non-physical fact. 3) Zombies are possible; but, Nagel’s argument, although more obscure 4) Materialism entails that zombies are not in its original presentation, can be “format- possible; therefore, ted” along similar lines. We can know all the 5) Materialism is false. physical facts about bats without knowing what it is like to be a bat. It follows that the Or, more explicitly formulated: knowledge we are missing is not knowledge 1 of a physical fact. Therefore, what it is like ) For any physical property P, it is conceiv- to be a bat is not a physical fact. able that P is instantiated but conscious- These arguments have struck many mate- ness is not; rialists as suspicious. After all, they infer an 2) For any pair of properties F and G, if it ontological conclusion from epistemological is conceivable that F is instantiated when premises. This move is generally suspicious, G is not, then it is (metaphysically) pos- but it is also vulnerable to a response that sible that F is instantiated when G is not; emphasizes what philosophers call the inten- therefore, sionality of epistemic contexts.25 This has 3) For any physical property P, it is possible been the main response among materialists that P is instantiated and consciousness is (Loar, 1990; Tye, 1986). The claim is that not; but, the Knowledge Argument’s protagonist does 4) If a property F can be instantiated when not learn a new fact when she learns what it property G is not, then F does not super- is like to see red, but rather learns an old vene on G;30 therefore, fact in a new way; and similarly for the bat 5 26 ) For any physical property P,consciousness student. does not supervene on P. Consider knowledge that the evening star glows and knowledge that the morning star To this argument it is objected that the sec- glows. These are clearly two different pieces ond premise is false, and the conceivability of knowledge. But the fact thereby known of something does not entail its possibility. is one and the same – the fact that Venus Thus, we can conceive of water not being glows. Knowledge that this is what it is like H2 O, but this is in fact impossible; Escher to see red and knowledge that this is the neu- triangles are conceivable, but not possible.31 ral assembly stimulated by the right wave- The zombie argument is more subtle length may similarly constitute two separate than this, however. One way to get at the pieces of knowledge that correspond to only real argument is this.32 Let us distinguish one fact being known. So from the acquisi- between the property of being water and P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 41

the property of appearing to be water, or water or being conscious do exist, “unnatu- being apparent water.33 For a certain quan- ral” properties do not, and appearance prop- 36 tity of stuff to be water, it must be H2 O. erties are unnatural in the relevant sense. But for it to appear to be water, it need To avoid this latter objection, a dualist only be clear, drinkable, liquid, and so on – may proceed to develop the argument dif- or perhaps only strike normal subjects as ferently, claiming that in the case of con- clear, drinkable, liquid, etc. Now, although sciousness, there is no distinction between the unrestricted principle that conceivabil- appearance and reality (Kripke, 1980). This ity entails possibility is implausible, a version would amount to the claim that the property of the principle restricted to what we may of being conscious is identical to the prop- call appearance properties is quite plausible. erty of appearing to be conscious. The con- Thus, if we can conceive of apparent water ceivability argument then goes like this: not being H2 O, then it is indeed possible that 1) For any physical property P, it is conceiv- apparent water should not be H2 O. Once the restricted principle is accepted, able that P is instantiated but apparent there are two ways a dualist may proceed. consciousness is not; The zombie argument seems to be captured 2) For any pair of properties F and G, such more accurately as follows:34 that F is an appearance property, if it is conceivable that F is instantiated when G 1) For any physical property P, it is conceiv- is not, then it is (metaphysically) possi- able that P is instantiated but apparent ble that F is instantiated when G is not; consciousness is not; therefore, 2) For any pair of properties F and G, such 3) For any physical property P, it is possible that F is an appearance property, if it is that P is instantiated when apparent con- conceivable that F is instantiated when G sciousness is not; but, is not, then it is (metaphysically) possi- 4) If property F can be instantiated when ble that F is instantiated when G is not; property G is not, then F does not super- therefore, vene on G; therefore, 3 ) For any physical property P, it is possible 5) For any physical property P,apparent con- that P is instantiated when apparent consciousness does not supervene on P; but, sciousness is not; but, 6) Consciousness = apparent consciousness; 4 ) If a property F can be instantiated when therefore, property G is not, then F does not (meta- 7) For any physical property P,consciousness physically) supervene on G; therefore, does not supervene on P. 5) For any physical property P, apparent consciousness does not (metaphysically) Materialists may reject this argument by supervene on P. denying that there is no distinction between appearance and reality when it comes to con- A materialist might want to reject this sciousness (the sixth premise). 2 argument by denying Premise (the res- The debate over the plausibility of the tricted conceivability-possibility principle). various versions of the zombie argument Whether the restricted principle is true is continues. A full critical examination is something we cannot settle here. Note, how- impossible here. Let us move on, then, ever, that it is surely much more plausible to consideration of the independent case than the corresponding unrestricted princi- against dualism. ple, and it is the only principle that the argument for dualism really needs. The Case against Dualism Another way the argument could be rejected is by denying the existence of such The main motivation to avoid dualism con- properties as apparent water and appar- tinues to be the one succinctly worded ent consciousness.35 More generally, perhaps, by Smart (1959,p.143) almost a half- while “natural” properties such as being century ago: “It seems to me that science is P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

42 the cambridge handbook of consciousness

increasingly giving us a viewpoint whereby that epiphenomenalism is merely counter- organisms are able to be seen as physico- intuitive, but does not face serious argumen- chemical mechanisms: it seems that even tative challenges. This is not particularly sat- the behavior of man himself will one day be isfying, however: all arguments must come explicable in mechanistic terms.” It would be to an end, and in most of philosophy, the curious if consciousness stood out in nature end is bound to be a certain intuition or as the only property that defied reduc- intuitively compelling claim. As intuitions tive explanation in microphysical terms. go, the intuition that consciousness is not More principled arguments aside, this simple epiphenomenal is very strong. observation seems to be the chief motivat- The second prong is more interesting. ing force behind naturalization projects that Chalmers notes that physics characterizes attempt to reductively explain conscious- the properties to which it adverts in purely ness and other recalcitrant phenomena. relational terms – essentially, in terms of As I noted above, against traditional dual- the laws of nature into which they enter. ists it was common to present the more The resulting picture is a network of inter- methodological argument that they do not related nodes, but the intrinsic character of in fact propose any positive theory of con- the thus interrelated nodes remains opaque. sciousness, but instead rest content with It is a picture that gives us what Bertrand arguing against existing materialist theories, Russell once wittily called “the causal skele- and that this could not lead to real progress in ton of the world.” Chalmers’ suggestion is the understanding of consciousness. Yet, this that phenomenal properties may constitute charge cannot be made against Chalmers, the intrinsic properties of the entities whose who does propose a positive theory of relational properties are mapped out by consciousness. physics. At least this is the case with intrin- Chalmers’ own theory is open to more sic properties of obviously conscious entities. substantial criticisms, however. In particu- As for apparently inanimate entities, their lar, it is arguably committed to epiphenom- intrinsic properties may be crucially similar enalism about consciousness, the thesis that to the phenomenal properties of conscious conscious states and events are causally inert. entities. They may be, as Chalmers puts it, As Kim (1989a,b, 1992) has pointed out, “protophenomenal” properties. it is difficult to find causal work for non- Although intriguing, this suggestion has supervenient properties. Assuming that the its problems. It is not clear that physics physical realm is causally closed (i.e., that indeed gives us only the causal skeleton of every instantiation of a physical property the world. It is true that physics characterizes has as its cause the instantiation of another mass in terms of its causal relations to other physical property), non-supervenient prop- properties. But it does not follow that the erties must either (i) have no causal effect property thus characterized is nothing but a on the physical realm or (ii) causally overde- bundle of causal relations. More likely, the termine the instantiation of certain physical relational characterization of mass is what properties.37 But because pervasive overde- fixes the reference of the term “mass,” but termination can be ruled out as implausi- the referent itself is nonetheless an intrinsic ble, non-supervenient properties must be property. The bundle of causal relations is causally inert vis-a-vis` the physical world. the reference-fixer, not the referent. On this However, the notion that consciousness is view of things, although physics character- causally inert, or epiphenomenal, is extremely izes mass in causal terms, it construes mass counter-intuitive: we seem to ourselves to not as the causing of effects E, but rather as act on our conscious decisions all the time the causer (or just the cause) of E. It con- and at will. strues mass as the relatum, not the relation. In response to the threat of epiphenom- Furthermore, if physics did present us enalism, Chalmers pursues a two-pronged with the causal skeleton of the world, then approach.38 The first prong is to claim physical properties would turn out to be P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 43

epiphenomenal (or nearly so). As Block have my conscious experience of the sky is (1990b) argued, functional properties – just a matter of my experience’s representa- properties of having certain causes and tion of the blue sky. The phenomenal char- effects – are ultimately inert, because an acter of my experience can be identified with effect is always caused by its cause, not by (an aspect of) its representational content.42 its causing. So if mass was the causing of E, This would be a theoretically happy rather than the cause of E, then E would result, since we have a fairly good notion as not be caused by mass. It would be caused, to how mental representation may be itself rather, by the protophenomenal property reductively explained in terms of informa- that satisfies the relational characterization tional and/or teleological relations between attached to mass in physics.39 The upshot neurophysiological states of the brain and is that if mass was the causing of E, rather physical states of the environment.43 The than the cause of E, mass would not have the reductive strategy here is two-stepped, then: causal powers we normally take it to have. first reduce phenomenal properties to repre- More generally, if physical properties were sentational properties, then reduce represen- nothing but bundles of causal relations, they tational properties to informational and/or would be themselves causally inert.40 other physical properties of the brain. Chalmers faces a dilemma, then: either he violates our strongly held intuitions Tye’s PANIC Theory regarding the causal efficacy of phenomenal properties, or he violates our strongly held Not every mental representation is con- intuitions regarding the causal efficacy of scious. For this reason, a representational physical properties. Either way, half his account of consciousness must pin down world is epiphenomenal, as it were. In any more specifically the kind of representation event, as we saw above, the claim that phys- that would make a mental state conscious. ical properties are merely bundles of causal The most worked-out story in this genre is relations – which therefore call for the postu- probably Michael Tye’s (1992, 1995, 2000, lation of phenomenal and protophenomenal 2002) “PANIC Theory.”44 properties as the putative causal relata – is The acronym “PANIC” stands for Poised, implausible. Abstract, Non-conceptual, Intentional Con- Problems concerning the causal efficacy tent. So for Tye, a mental representation of phenomenal properties will attach to qualifies as conscious when, and only when, any account that portrays them as non- its representational content is (a) inten- supervenient upon, or even as non-reducible tional, (b) non-conceptual, (c) abstract, and to, physical properties. These problems are (d) poised. What all these qualifiers mean is less likely to rear their heads for reduc- not particularly important, though the prop- tive accounts of consciousness. Let us erties of non-conceptuality and poise are turn, then, to an examination of the main worth pausing to explicate.45 reductive accounts discussed in the current The content of a conscious experience is literature. non-conceptual in that the experience can represent properties for which the subject lacks the concept. My conscious experience Representationalism of the sky represents the sky not simply as being blue, but as being a very specific shade According to the representational theory of blue, say blue17. And yet if I am presented of consciousness – or for short, representa- a day later with two samples of very simi- tionalism – the phenomenal properties of lar shades of blue, blue17 and blue18, I will conscious experiences can be reductively be unable to recognize which shade of blue explained in terms of the experiences’ rep- was the sky’s. This suggests that I lack the 41 resentational properties. Thus, when I look concept of blue17. If so, my experience’s rep- 46 up at the blue sky, what it is like for me to resentation of blue17 is non-conceptual. P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

44 the cambridge handbook of consciousness

The property of poise is basically a func- It appears, then, that when you pay attentional role property: a content is poised tion to your experience, the only thing you when it is ready and available to make a become aware of is which features of the direct impact on the formation of beliefs and external sky your experience represents.In desires. Importantly, Tye takes this to dis- other words, the only introspectively acces- tinguish conscious representation from, say, sible properties of conscious experience are blindsighted representations. A square can its representational properties. be represented both consciously and blind- The transparency of experience provides sightedly. But only the conscious representa- a straightforward argument for representation is poised to make a direct impact on the tionalism. The argument may be laid out as beliefs that the subject subsequently forms. follows: PANIC theory is supposed to cover not only conscious perceptual experiences but 1) The only introspectively accessible prop- also all manners of phenomenal experi- erties of conscious experience are its rep- ence: somatic, emotional, and so on. Thus, a resentational properties; toothache experience represents tissue dam- 2) The phenomenal character of conscious age in the relevant tooth, and does so inten- experience is given by its introspectively tionally, non-conceptually, abstractly, and accessible properties; therefore, with poise.47 3) The phenomenal character of conscious experience is given by its representational The Master Argument for properties. Representationalism The first premise is the thesis of The main motivation for representational- transparency; the second one is intended ism may seem purely theoretical: it holds the as a conceptual truth (about what we promise of a reductive explanation of con- mean by “phenomenal”). The conclusion is sciousness in well-understood informational representationalism. and/or teleological terms. Perhaps because Another version of the argument from of this, however, the argument that has been transparency, one that Tye employs, centers most influential in making representational- on the idea that rejecting representationalism popular is a non-theoretical argument, ism in the face of transparency would require one that basically rests on a phenomenolog- one to commit to an “error theory.”48 This ical observation. This is the observation of version may be formulated as follows: the so-called transparency of experience. It has been articulated in a particularly influential 1) The phenomenal properties of conscious manner by Harman (1990), but goes back at experience seem to be representational least to Moore (1903). properties; Suppose you have a conscious experience 2 of the blue sky. Your attention is focused ) It is unlikely that the phenomenal prop- on the sky. You then decide to turn your erties of conscious experience are radi- attention away from the sky and onto your cally different from what they seem to be; experience of the sky. Now your attention therefore, is no longer focused on the sky, but rather 3) It is likely that the phenomenal proper- on the experience thereof. What are you ties of conscious experience are represen- aware of? It seems that you are still aware tational properties. of the blueness of the sky. Certainly you are not aware of some second blueness, which Here the transparency thesis is again the attaches to your experience rather than to first premise. The second premise is the the sky. You are not aware of any intermedi- claim that convicting experience of massive ary blue quality interposed between yourself error is to be avoided. And the conclusion is and the sky. representationalism. P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 45

The Case against Representationalism Tye, 2000). To be sure, you do not necessarily possess the concept of subtending a visual Most of the arguments that have been mar- angle. But recall that the content of expe- shaled against representationalism are argu- rience can be construed as non-conceptual. ments by counter-example. Scenarios of So your experience can represent the two varying degrees of fancifulness are adduced, trees to subtend different visual angles within which allegedly (i) a conscious experi- out employing the concept of subtend- ence has no representational properties, or ing a visual angle. Thus a representational (ii) two possible experiences with different difference is matched to the phenomenal phenomenal properties have the same repre- difference. sentational properties, or (iii) inversely, two Perhaps the most prominent alleged possible experiences with the same phenom- counter-example is Block’s (1990a) Inverted enal properties have different representa- Earth case. Inverted Earth is an imaginary tional properties. For want of space, I present planet just like Earth, except that every only one representative scenario from each object there has the color complementary to category. the one it has here. We are to imagine that a Block (1996) argues that phosphene expe- subject is clothed with color-inverting lenses riences are non-representational. These can and shipped to Inverted Earth unbeknownst be obtained by rubbing one’s eyes long to her. The color inversions due to the lenses enough so that when one opens them again, and to the world cancel each other out, one “sees” various light bits floating about. so that her phenomenal experiences remain Such experiences do not represent any exter- the same. But externalism about represen- nal objects or features, according to Block. tational contents ensures that the represen- In response, Tye (2000) claims that such tational content of her experiences eventu- experiences do represent – it is just that they ally change.49 Her bluish experiences now misrepresent. They misrepresent there to be represent a yellow sky. When her sky expe- small objects with phosphorescent surfaces riences on Inverted Earth are compared to floating around the subject’s head. her earthly sky experience, it appears that A long-debated case in which phenome- the two groups are phenomenally the same nal difference is accompanied by represen- but representationally different. tational sameness is due to Peacocke (1983). This case is still being debated in the Suppose you stand in the middle of a mostly literature, but there are two representa- empty road. All you can see in front of you tionalist strategies for accommodating it. are two trees. The two trees, A and B, have One is to argue that the phenomenal char- the same size and shape, but A is twice as far acter also changes over time on Inverted from you as B. Peacocke claims that, being Earth (Harman, 1990); the other is to devise aware that the two trees are the same size, accounts of representational content that you represent to yourself that they have the make the representational content of the same properties. And yet B “takes up more subject’s experiences remain the same on of your visual field” than A, in a way that Inverted Earth, externalism notwithstanding makes you experience the two trees differ- (Tye, 2000).50 ently. There is phenomenal difference with- There may be, however, a more princi- out representational difference. pled difficulty for representationalism than Various responses to this argument have the myriad counter-examples it faces.51 Rep- been offered by representationalists. Perhaps resentationalism seems to construe the phe- the most popular is that although you rep- nomenal character of conscious experiences resent the two trees to have the same size purely in terms of the sensuous qualities they properties, you also represent them to have involve. But arguably there is more to phe- certain different properties – for example, nomenal character than sensuous quality. In B is represented to subtend a larger visual particular, there seems to be a certain mine- angle than A (DeBellis, 1991; Harman, 1990; ness, or for-me-ness, to them. P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

46 the cambridge handbook of consciousness

One way to put it is as follows (Kriegel, Higher-Order Monitoring Theory 2005a; Levine, 2001; Smith, 1986). When I have my conscious experience of the blue One theory of consciousness from analytic sky, there is a bluish way it is like for me philosophy that can be interpreted as target- to have my experience. A distinction can ing subjective character is the higher-order be drawn between two components of this monitoring theory (HOMT). According to “bluish way it is like for me”: the bluish com- HOMT, what makes a mental state con- ponent, which we may call qualitative char- scious is the fact that the subject is aware acter, and the for-me component, which we of it in the right way. It is only when the may call subjective character. We may con- subject is aware (in that way) of a mental strue phenomenal character as the comp- state that the state becomes conscious.55 resence of qualitative and subjective charac- HOMT tends to anchor consciousness in ter. This subjective character, or for-me-ness, the operation of a monitoring device. This is certainly an elusive phenomenon, but it device monitors and scans internal states and is present in every conscious experience. events and produces higher-order represen- Indeed, its presence seems to be a condi- tations of some of them.56 When a mental tion of any phenomenality: it is hard to state is represented by such a higher-order make sense of the idea of a conscious expe- representation, it is conscious. So a mental rience that does not have this for-me-ness state M of a subject S is conscious when, to it. If it did not have this for-me-ness, it and only when, S has another mental state, would be a mere subpersonal state, a state M∗, such that M∗ is an appropriate represen- that takes place in me but is not for me tation of M. The fact that M∗ represents M in the relevant sense. Such a subpersonal guarantees that there is something it is like state seems not to qualify as a conscious for S to have M.57 experience. Observe that, on this view, what con- The centrality of subjective character (as fers conscious status on M is something out- construed here) to consciousness is some- side M, namely, M∗. This is HOMT’s reduc- thing that has been belabored in the phe- tive strategy. Neither M nor M∗ is conscious nomenological tradition (see Chapter 4; in and of itself, independently of the other Zahavi, 1999). The concept of prereﬂec- state. It is their coming together in the right tive self-consciousness – or a form of self- way that yields consciousness.58 awareness that does not require focused Versions of the HOMT differ mainly in and explicit awareness of oneself and one’s how they construe the monitoring device current experience, but is rather built into and/or the representations it produces. The that very experience – is one that ﬁgures most seriously worked out version is prob- centrally in almost all phenomenological ably David Rosenthal’s (1986, 1990, 2002a, 52 accounts of consciousness. But it has been b). Let us take a closer look at his “higher- somewhat neglected in analytic philosophy order thought” theory. of mind.53 The relative popularity of representation- Rosenthal’s Higher-Order Thought alism attests to this neglect. While a repre- Theory sentationalist account of sensuous qualities – what we have called qualitative character – According to Rosenthal, a mental state is may turn out to win the day (if the all- conscious when its subject has a suitable eged counter-examples can be overcome), it higher-order thought about it.59 The higher- would not provide us with any perspective order state’s being a thought is supposed on subjective character.54 Therefore, even if to rule out, primarily, its being a quasi- representationalism turns out to be a satis- perceptual state. factory account of qualitative character, it is There is a long tradition, hailing from unlikely to be a satisfactory account of phe- Locke, of construing the monitoring device nomenal consciousness proper. as analogous in essential respects to a sense P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 47

organ (hence as being a sort of “inner sense”) It is worth noting that according to Rosen- and accordingly as producing mental states thal the second-order representation is nor- that are crucially similar to perceptual rep- mally an unconscious state. To be sure, it need resentations and that may to that extent not necessarily be: in the more introspective, be called “quasi-perceptual.” This sort of or reflective, episodes of our conscious life, “higher-order perception theory” is cham- the second-order state becomes itself con- pioned today by Armstrong (1968, 1981) scious. It is then accompanied by a third- and Lycan (1987, 1996). Rosenthal believes order state, one that represents its occur- that this is a mistake and that the higher- rence in a suitable way. When I explicitly order states that confer consciousness are not introspect and dwell on my conscious expe- analogous to perceptual representations.60 rience of the sky, there are three separate Rather, they are intellectual, or cognitive, states I am in: the (first-order) experience, a states – that is, thoughts. (second-order) awareness of the experience, Another characteristic of thoughts – in and a (third-order) representation of that addition to being non-perceptual – is their awareness. When I stop introspecting and being assertoric. An assertoric state is one turn my attention back to the sky, however, that has a thetic, or mind-to-world, direc- the third-order state evaporates, and con- tion of fit.61 This is to be contrasted with sequently the second-order state becomes states (such as wanting, hoping, disapprov- unconscious again. In any event, at any one ing, etc.) that have primarily a telic, or world- time the subject’s highest-order state, the to-mind, direction of fit.62 A third character- one that confers consciousness on the chain istic of thoughts – at least the kind suitable of lower-order states “below” it, is uncon- for conferring consciousness – is that they scious.68 are occurrent mental states.63 In summary, Rosenthal’s central thesis Crucially, a suitable higher-order thought is that a mental state is conscious just in would also have to be non-inferential, in that case the subject has a non-perceptual, non- it could not be the result of a conscious infer- inferential, assertoric, de se, occurrent repre- ence from the lower-order state (or from any sentation of it. This account of consciousness other state, for that matter).64 To be sure, is not intended as an account of introspec- the thought is formed through some process tive or reflective consciousness, but of regu- of information processing, but that process lar, everyday consciousness. must be automatic and unconscious. This is intended to reflect the immediacy, or at least The Master Argument for Higher-Order felt immediacy, of our awareness of our con- Monitoring Theory scious states.65 The fact that my experience The master argument for the higher-order of the sky has for-me-ness entails that I am monitoring approach to consciousness has somehow aware of its occurrence; but not been succinctly stated by Lycan (2001): any sort of awareness would do – very mediated forms of awareness cannot confer con- 1) A mental state M of subject S is conscious scious status on their objects. when, and only when, S is aware of M in One last characteristic Rosenthal ascribes the appropriate way; to the “suitable” higher-order representation 2) Awareness of X requires mental represen- is that it represents the lower-order state tation of X; therefore, as a state of oneself. Its content must be, as 66 3) M is conscious when, and only when, S this is sometimes put, de se content. So ∗ ∗ has a mental state M , such that M rep- the content of the higher-order represen- resents M in the appropriate way. tation of my conscious experience of the sky is not simply something like “this bluish Although the second premise is by no means experience is occurring,” but rather some- trivial, it is the first premise that has been thing like “I myself am having this bluish the bone of contention in the philosophical experience.”67 literature (see, e.g., Dretske, 1993). P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

48 the cambridge handbook of consciousness

One can defend the claim that conscious side world. The claim is rather that we are states are states we are aware of having sim- at least peripherally aware of every conscious ply as a piece of conceptual analysis – as a state we are in.69 As long as M is conscious, platitude reflecting the very meaning of the S is aware, however dimly and inattentively, word “conscious” (Lycan, 1996). To my ear, of M. Once S’s awareness of M is extin- this sounds right: a mental state of which guished altogether, M drops into the realm of the subject is completely unaware is a sub- the unconscious. This seems highly plausible personal, and therefore unconscious, state. on both conceptual and phenomenological To some, however, this seems plainly grounds.70 false. When I have an experience of the sky, I am attending to the sky, they stress, not to The Case against Higher-Order myself and my internal goings-on. By con- Monitoring Theory sequence, I am aware of the sky, not of my experience of the sky. I am aware through my Several problems for the monitoring the- experience, not of my experience. ory have been continuously debated in the This objection seems to rely, however, philosophical literature. I focus here on what on an unwarranted assimilation of awareness I take to be the main three.71 and attention. There is a distinction to be The first is the problem of animal and made between attentive awareness and inat- infant consciousness. It is intuitively plausi- tentive awareness. If S attends to X and not ble to suppose that cats, dogs, and human to Y, it follows that S is not attentively aware neonates are conscious, that is, they have of Y, but it does not follow that S is com- conscious states; but it appears empirically pletely unaware of Y. For S may still be inat- implausible that they should have second- tentively aware of Y. order representations (Lurz, 1999). The Consider straightforward visual aware- problem is particularly acute for Rosenthal’s ness. The distinction between foveal vision account, since it is unlikely that these crea- and peripheral vision means that our visual tures can have thoughts, and moreover of the awareness at any one time has a periphery complex form, “I myself am enjoying this as well as a focal center. Right now, I am milk.” (visually) focally aware of my laptop, but There are two ways to respond to this also (visually) peripherally aware of an ash- objection. One is to deny that having such tray at the far corner of my desk. A similar higher-order representations requires a level distinction applies to perceptual awareness of sophistication of an order unlikely to in other modalities: I am now (auditorily) be found in (say) cats. Thus, Rosenthal focally aware of Duke Ellington’s voice and (2002b) claims that whereas adult human (auditorily) peripherally aware of the air higher-order thoughts tend to be conceptu- conditioner’s hum overhead. ally structured and employ a rich concept There is no reason to think that a similar of self, these are not necessary features of distinction would not apply to higher-order such thoughts. There could be higher-order awareness. In reflective moods I may be thoughts that are conceptually simple and focally aware of my concurrent experiences employ a rudimentary concept of self, one and feelings, but on other occasions I am just that consists merely in the ability to distin- peripherally aware of them. The former is guish oneself from anything that is not one- an attentive form of second-order awareness, self. It may well turn out that worms, wood- the latter an inattentive one. Again, from the peckers, or even day-old humans lack even fact that it is inattentive it would be falla- this level of conceptual sophistication – in cious to infer that it is no awareness at all. which case we would be required to deny When it is claimed that conscious states them consciousness – but it is unlikely that are states we are aware of, the claim is not cats, dogs, and year-old humans lack them. that we are focally aware of every conscious The second possible line of response is state we are in. That is manifestly false: the to dismiss the intuition that animals, such focus of our attention is mostly on the out- as cats, dogs, and even monkeys, do in P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 49

fact have conscious states. Thus, Carruthers There are several ways higher-order mon- (1998, 1999) claims that there is a signifi- itoring theorists may respond to this objec- cant amount of projection that takes place tion. Let us briefly consider three possible when we ascribe conscious states to, say, responses. our pets. In reality there is very little evi- First, they may claim that when M∗ is dence to suggest that they have not only targetless, the property of being conscious, perceptual and cognitive states but also although not instantiated by M, is instanti- conscious ones. ated by M∗. But as we saw above, according Both lines of response offer some hope to their view, M∗ is normally unconscious. to the defender of higher-order monitor- So to say that M∗ instantiates the property ing, but also implicate the theory in certain of being conscious would be to say that it is, counter-intuitive and prima facie implausi- in the normal case, both conscious and not ble claims. Whether these could somehow conscious – which is incoherent.75 be neutralized, or accepted as outweighed by Second, they may claim that the prop- the theoretical benefits of HOMT, is some- erty of being conscious is, in reality, not thing that is very much under debate. a property of the discrete state M, but Perhaps more disturbing is the problem rather attaches itself to the compound of of so-called targetless higher-order thoughts M and M∗.76 But this will not work either, (or more generally, representations). When because HOMT would then face the fol- someone falsely believes that the almond lowing dilemma. Either the compound state tree in the backyard is blooming again, there M + M∗ is a state we are aware of having, are two ways he or she may get things wrong: or it is not. If it is not, then HOMT is false, (i) it may be that the backyard almond tree since it claims that conscious states are states is not blooming, or (ii) it may be that there we are aware of having. If it is, then accord- is no almond tree in the backyard (blooming ing to the theory it must be represented by a or not). Let us call a false belief of type (ii) third-order mental state, M∗∗, in which case a targetless thought. HOMT gets into trou- the same problem would recur when M∗∗ is ble when a subject has a targetless higher- targetless. order thought (Byrne, 1997).72 Suppose at a Third, they may claim that there are time t subject S thinks (in the suitable way) no targetless higher-order representations. that she has a throbbing toothache, when in But even if this can be shown to be the reality she has no toothache at all (throb- actual case (and it is hard to imagine how bing or not). According to HOMT, what it this would be done), we can surely con- is like for S at t is the way it is like to have a ceive of counterfactual situations in which throbbing toothache, even though S has no targetless higher-order representations toothache at t. In other words, if S has an do occur.77 M∗ that represents M when in reality there A third problem for the HOMT is its is no M,73 S will be under the impression treatment of the epistemology of conscious- that she is in a conscious state when in real- ness (Goldman, 1993b; Kriegel, 2003b). Our ity she is not. (She is not in a conscious state knowledge that we are in a conscious state because M does not exist, and it is M that is is first-person knowledge, knowledge that is supposed to bear the property of being con- not based on inference from experimental, scious.) Moreover, on the assumption that a or theoretical, or third-personal evidence. person is conscious at a time t only if she But if HOMT were correct, what would has at least one conscious state at t,74 this make our conscious states conscious is (nor- would entail that when a subject harbors mally) the occurrence of some unconscious a targetless higher-order misrepresentation, state (i.e., the higher-order representation), she is not conscious, even though it feels to so in order to know that we are in a conscious her as though she is. This is a highly counter- state we would need to know of the occur- intuitive consequence: we want to say that a rence of that unconscious state. But knowl- person cannot be under the impression that edge of unconscious states is necessarily the- she is conscious when she is not. oretical and third-personal, since we have P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

50 the cambridge handbook of consciousness

no direct acquaintance with our unconscious popularity, is what we may call the “self- states. representational theory.” According to this Another way to put the argument is this. view, mental states are conscious when, and How does the defender of HOMT know only when, they represent their own occur- that conscious states are states of which we rence (in the right way). Thus, my conscious are aware? It does not seem to be some- experience of the blue sky represents both thing she knows on the basis of experimenta- the sky and itself – and it is in virtue of repre- tion and theorization. Rather, it seems to be senting itself that it is a conscious experience. intuitively compelling, something that she Historically, the most thorough devel- knows on the basis of first-person acquain- opment and elucidation of the self-rep- tance with her conscious states. But if resentational theory is Brentano’s (1874). HOMT were correct, it would seem that Through his work, the view has had a sig- that knowledge would have to be purely nificant influence in the phenomenologi- theoretical and third-personal. So construed, cal tradition. But apart from a couple of this “epistemic argument” against HOMT exceptions – Lehrer (1996, 1997) and Smith may be formulated as follows: (1986, 1989) come to mind – the view had enjoyed virtually no traction in Anglo- 1) If HOMT were correct, our awareness of American philosophy. Recently, however, our conscious states would normally be versions of the view, and close variations an unconscious state; that is, on it, have been defended by a number of 2) We do not have non-theoretical, first- philosophers.78 person knowledge of our unconscious Rather than focus on any one particular states; therefore, account of consciousness along these lines, I 3) If HOMT were correct, we would not now survey the central contributions to the have non-theoretical, first-person knowl- understanding of consciousness in terms of edge of the fact that we are aware of our self-representation. conscious states; but, 4) We do have non-theoretical, first-person Varieties of Self-Representational Theory knowledge of the fact that we are aware of our conscious states; therefore, Brentano held that every conscious state is intentionally directed at two things. It is 5) HOMT is incorrect. primarily directed at whatever object it is The upshot of the argument is that the about, and it is secondarily directed at itself. awareness of our conscious states must in the My bluish sky experience is directed pri- normal case be itself a conscious state. This is marily at the sky and secondarily at itself. something that HOMT cannot allow, how- In more modern terminology, a conscious ever, since within its framework it would state has two representational contents: an lead to infinite regress. The problem is to other-directed (primary) content and a self- reconcile the claim that conscious states are directed (secondary) content. Thus, if S states we are aware of having with the notion consciously fears that p, S’s fear has two that we have non-theoretical knowledge of contents: the primary content is p, the sec- this fact. ondary content is itself, the fear that p. The distinction between primary intentionality and secondary intentionality is presumably intended to capture the differ- The Self-Representational Theory ence (discussed above) between attentive or of Consciousness focal awareness and inattentive or peripheral awareness.79 One approach to consciousness that has a Caston (2002) offers an interesting gloss venerable tradition behind it, but has only on this idea in terms of the type/token dis- very recently regained a modest degree of tinction. For Caston, S’s conscious fear that p P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 51

is a single token state that falls under two sep- thesis – and this seems to amount to the arate state types: the fear-that-p type and the claim that M is conscious just in case it rep- awareness-of-fear-that-p type. The state has resents itself (constitutes a representation of two contents, arguably, precisely in virtue of itself). But the point is that there are other, falling under two types. weaker constitutive relations that fall short Brook and Raymont (2006) stress that of full identity. the self-representational content of the con- One such relation is the part-whole rela- scious state is not simply that the state tion. Accordingly, one version of the view, occurs, but rather that it occurs within one- the one defended by Gennaro (1996, 2006), self – that it is one’s own state. Just as Rosen- holds that M∗ is a part of M; another version, thal construed the content of higher-order apparently put forth by Kobes (1995), holds states as “I myself am having that state,” so that M is part of M∗; and yet another ver- Brook and Raymont suggest that the full self- sion, Van Gulick’s (2001, 2006), holds that representational content of conscious states M is conscious when it has two parts, one of is something like “I myself am herewith hav- which represents the other. ing this very state.”80 In Van Gulick’s “higher-order global For Brentano and his followers, the self- states theory,” S’s fear that p becomes con- directed element in conscious states is an scious when the fear and S’s awareness of the aspect of their intentionality, or content. fear are somehow integrated into a single, In David Woodruff Smith’s (1986, 2004) unified state. This new state supersedes its “modal account,” by contrast, the self- original components, though, in a way that directed element is construed not as an makes it a genuine unity, rather than a sum aspect of the representational content, but of two parts, one of which happens to rep- rather as an aspect of the representational resent the other. The result is a state that, if attitude (or mode). When S consciously it does not represent itself, does something fears that p, it is not in virtue of figuring very close to representing itself.84 in its own secondary content that the fear is conscious. Indeed, S’s fear does not have a secondary content. Its only content is p. The Master Argument for the The “reflexive character” of the fear, as Smith Self-Representational Theory puts it, is rather part of the attitude S takes The basic argument for the self-represen- toward p. Just as the attitudes toward p can tational approach to consciousness is that vary from fear, hope, expectation, and so it is the only way to accommodate the on, so they can vary between self-directed notion that conscious states are states we are or “reflexive” fear and un-self-directed or aware of without falling into the pitfalls of “irreflexive” fear. S’s fear that p is conscious, HOMT. on this view, because S takes the attitude of The argument can be organized, then, as 81,82 self-directed fear toward p. a disjunctive syllogism that starts from the One way in which the self-represen- master argument for HOMT, but then goes tational thesis can be relaxed to make a sub- beyond it: tler claim is the following. Instead of claiming that a mental state M of a subject S is 1) A mental state M of subject S is con- conscious just in case M represents itself, the scious when, and only when, S is aware thesis could be that M is conscious just in of M; case S has an M∗ that is a representation of 2 M and there is a constitutive, non-contingent ) Awareness of X requires mental represen- relation between M and M∗.83 One consti- tation of X; therefore, tutive relation is of course identity. So one 3) M is conscious when, and only when, S ∗ ∗ version of this view would be that M is con- has a mental state M , such that M rep- scious just in case M is identical with M∗ – resents M. this is how Hossack (2002) formulates his 4) Either M∗ = MorM∗ = M; P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

52 the cambridge handbook of consciousness

5) There are good reasons to think that it is at all. The same problem does not arise for not the case that M∗ = M; therefore, self-representing states, however: although 6) There are good reasons to think that it is M could in principle misrepresent itself to be the case that M∗ = M; therefore, F when in reality it is not F, it could not possi- 7) Plausibly, M is conscious when, and only bly misrepresent itself to be F when in reality when, M is self-representing. it does not exist at all. For if it did not exist it could not represent anything, itself included. The fourth premise could also be formu- Thus the problem of targetless higher-order lated as “either M∗ and M do not enter- representations has no bite against the self- tain a constitutive, non-contingent relation, representational theory. or they do,” with appropriate modifications These are already two major problems in Premises 5 and 6 to suit. The conclu- that affect gravely the plausibility of HOMT, sion of the relevantly modified argument but do not apply to the self-representational would then be the thesis that M is conscious theory. They make a strong prima facie case when, and only when, S has a mental state for the fifth premise above. The fourth M∗, such that (i) M∗ represents M and (ii) premise is a logical truism, and the first and there is a constitutive, non-contingent rela- second ones were defended above. So the tion between M and M∗. argument appears to go through. The fallacy in the master argument for HOMT is the supposition that if S is aware Problems for the Self-Representational of M, then S must be so aware in virtue of Theory being in a mental state that is numerically different from M. This supposition is brought One problem that does persist for the self- to the fore and rejected in the argument just representational theory is the problem of sketched. animal consciousness. The ability to have The case for the fifth premise consists in self-representing states presumably requires all the reasons to be suspicious of HOMT, as all the conceptual sophistication that the elaborated in the previous section, although ability to have higher-order monitoring it must also be shown that the same prob- states does (since the self-representational lems do not bedevil the self-representational content of a conscious state is the same as theory as well. the representational content that a higher- Consider first the epistemic argument. order state would have), and perhaps even We noted that HOMT fails to account for greater sophistication.85 the non-theoretical, first-person knowledge Another problem is the elucidation and we have of the fact that we are aware viability of the notion of self-representation. of our conscious states. This is because it What does it mean for a mental state to construes this awareness as (normally) an represent itself, and what sort of mecha- unconscious state. The self-representational nism could subserve the production of self- theory, by contrast, construes this awareness representing states? There is something at as a conscious state, since it construes the least initially mysterious about the notion awareness as the same state, or part of the of a self-representing state that needs to be state, of which one is thereby aware. So the confronted. self-representational theory, unlike HOMT, In fact, one might worry that there are can provide for the right epistemology of principled reasons why self-representation consciousness. is incompatible with any known natural- Consider next the problem of target- ist account of mental representation. These less higher-order representations. Recall, the accounts construe mental representation as problem ensues from the fact that M∗ could some sort of natural relation between brain in principle misrepresent not only that M states and world states. Natural relations, as is F when in reality M is not F, but also opposed to conceptual or logical ones, are that M is F when in reality there is no M based on causality and causal processes. But P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 53

causality is an anti-reflexive relation, that ture with no conscious awareness whatso- is, a relation that nothing can bear to itself. ever who harbors mental states that rep- Thus no state can bring about its own occur- resent themselves. Thus Chalmers’ zombie rence or give rise to itself. The argument can argument can be run in a particularized ver- be formulated as follows: sion directed specifically against the self- representational theory.90 1) Mental representation involves a causal relation between the representation and the represented; Conclusion: Directions for 2) The causal relation is anti-reflexive; Future Research therefore, 3) No mental state can cause itself; and Much of the philosophical discourse on therefore, consciousness is focused on the issue of 4) No mental state can represent itself. reducibility. As we just saw, the zombie argument and other dualist arguments can be The basic idea is that there is no naturalist tailored to target any particular reductive account of mental representation that could account of consciousness. This debate holds allow for self-representing mental represen- great intrinsic importance, but it is impor- tations. tant to see that progress toward a scientific Even more fundamentally, one may worry explanation of consciousness can be made whether the appeal to self-representation without attending to it. really explains consciousness. Perhaps self- All three reductive approaches to con- representation is a necessary condition for sciousness we considered – the represen- consciousness, but why think it is also a suf- tational, higher-order monitoring, and self- ficient condition? A sentence such as “this representational theories – can readily be very sentence contains six words” is self- refashioned as accounts not of consciousness representing, but surely there is nothing it itself, but of the emergence base (or causal is like to be that sentence.86 basis) of consciousness. Instead of claiming One may respond to this last point that that consciousness is (or is reducible to) phys- what is required for consciousness is intrinsic ical structure P, the claim would be that or original self-representation, not derivative consciousness emerges from (or is brought self-representation.87 Sentences and linguis- about by) P. To make progress toward the tic expressions do not have any representa- scientific explanation of consciousness, we tional content in and of themselves, indepen- should focus mainly on what the right phys- dently of being interpreted. But plausibly, ical structure is – what P is. Whether P is mental states do.88 The same goes for self- consciousness itself or only the emergence representational content: sentences and lin- base of consciousness is something we can guistic expressions may be derivatively self- set aside for the purposes of scientific expla- representing, but only mental states can be nation. If it turns out that P is conscious- non-derivatively self-representing. A more ness itself (as the reductivist holds), then we accurate statement of the self-representation will have obtained a reductive explanation of theory is therefore this: A mental state M of a consciousness; if it turns out that P is only subject S is conscious when, and only when, the emergence base of consciousness (as the M is non-derivatively self-representing. dualist holds), then we will have obtained Still, self-representing zombies are read- a causal explanation of consciousness. But ily conceivable. It is quite easy to imagine both kinds of explanation are bona fide sci- unconscious mental states in our own cog- entific explanations. nitive system – say, states formed early on In other words, philosophers could use- in visual processing – that represent them- fully reorganize their work on conscious- selves without thereby being conscious.89 ness around a distinction between two sep- Furthermore, it is easy to imagine a crea- arate issues or tasks. The first task is to P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

54 the cambridge handbook of consciousness

devise a positive account of the physical case against. Obviously, space and other lim- (or more broadly, natural) correlate of con- itations do not allow me to present the full sciousness, without prejudging whether it story on each of these approaches. will constitute a reduction base or merely 2. The distinction between creature conscious- an emergence base. Work along these lines ness and state consciousness is due to Rosen- will involve modifying and refining the rep- thal (1986). resentational, higher-order monitoring, and 3. Availability consciousness as construed here is self-representational theories and/or devis- very similar to the notion of access conscious- ing altogether novel positive accounts. The ness as defined by Block (1995). There are second task is to examine the a priori and certain differences, however. Block defines a posteriori cases for reducibility. Work here access consciousness as the property a mental state has when it is poised for free use will probably focus on the issue of how much by the subject in her reasoning and action can be read off of conceivability claims, control. It may well be that a mental state as well as periodic reconsideration of the is availability-conscious if and only if it is intuitive plausibility of such claims in light access-conscious. For a detailed discussion of of newer and subtler positive accounts of the relation between phenomenal conscious- 91 consciousness. ness and access consciousness, see Kriegel Another front along which progress can (2006b). certainly be made is tightening the connec- 4. It is debatable whether thoughts, beliefs, tion between the theoretical and experi- desires, and other cognitive states can at all mental perspectives on consciousness. Ulti- be conscious in this sense. I remain silent on mately, one hopes that experiments could be this issue here. For arguments that they can 1993 designed that would test well-defined empir- be conscious, see Goldman ( a), Horgan 2002 1998 ical consequences of philosophical (or more and Tienson ( ), and Siewert ( ). generally, purely theoretical) models of con- 5. The terms “easy problems” and “hard prob- sciousness. This would require philosophers lem” are intended as mere labels, not as to be willing to put forth certain empiri- descriptive. Thus it is not suggested here that understanding any of the functions of con- cal speculations, as wild as these may seem, sciousness is at all easy in any significant sense. based on their theories of consciousness, and Any scientist who has devoted time to the experimental scientists to take interest in study of consciousness knows how outstand- the intricacies of philosophical theories in ing the problems in this field are. These terms an attempt to think up possible ways to test are just a terminological device designed to them. bring out the fact that the problem of why All in all, progress in our understand- there is something it feels like to undergo a ing of consciousness and the outstanding conscious experience appears to be of a dif- methodological and substantive challenges it ferent order than the problems of mapping presents has been quite impressive over the out the cognitive functions of consciousness. past two decades. The central philosophical 6. This is so even if phenomenal consciousness issues are today framed with a clarity and does not turn out to have much of a func- precision that allow a corresponding level of tional significance in the ordinary cognitive life of a normal subject – as some (Libet, 1985; clarity and precision in our thinking about Velmans, 1992; Wegner, 2002) have indeed consciousness. Even more happily, there is argued. no reason to suppose that this progress will 7. In the course of the discussion I avail myself come to a halt or slow down in the near 92 of philosophical terminology that may not future. be familiar to the non-philosophically trained reader. However, I have tried to recognize all the relevant instances and such and include an Notes endnote that provides a standard explication of the terminology in question. 1. More accurately, I present central aspects of 8. No major philosopher holds this view, to my the main account, the case in favor, and the knowledge. P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 55

9. Many of the key texts discussed in this chap- is that if a conceptual scheme is power- ter are conveniently collected in Block et al. ful enough to frame a problem it should (1997). Here, and in the rest of the chapter, I be powerful enough to frame the solution. refer to the reprint in that volume. Whether the correct solution will actually 10. This is what Churchland often discusses be framed is of course anyone’s guess. But under the heading of the “plasticity of mind” the problem cannot be a constitutive limita- (see especially Churchland, 1979). tion on concept formation mechanisms. (For a more detailed development of this line of cri- 11 . It may not be perceiving those brain states tique, see Kriegel, 2004a.) There is a counter- as brain states. But it will nonetheless be a example of this sort of claim, however. Cer- matter of perceiving the brain states. tain problems that can be framed within the 12. The view – sometimes referred to as emer- theory of rational numbers cannot be solved gentism – that consciousness is caused by the within it; the conceptual machinery of irra- brain, or causally emerges from brain activ- tional numbers must be brought in to solve ity, is often taken by scientists to be material- these problems. It might be claimed, how- ist enough. But philosophers, being interested ever, that this sort of exception is limited to in the ontology rather than genealogy of conformal systems and does not apply to theo- sciousness, commonly take it to be a form of ries of the natural world. Whether this claim dualism. If consciousness cannot be shown to is plausible is something I do not adjudicate be itself material, but only caused by matter, here. then consciousness is itself immaterial, as the 15. Monism divides into two subgroups: mate- dualist claims. At the same time, the position rialist monism, according to which the only implicit in scientists’ work is often that what kind of stuff there is is matter, and idealist is caused by physical causes in accordance monism, according to which the stuff in ques- with already known physical laws should be tion is some sort of mindstuff. immediately considered physical. This position, which I have called elsewhere inclusive 16. Idealism is not really considered a live option materialism (Kriegel, 2005b), is not unreason- in current philosophical discussions, although able. But the present chapter is dedicated to it is defended by Foster (1982). I do not discuss philosophers’ theories of consciousness, so I set it here. it aside. 17. Such coming-apart happens, for Descartes, 13. It should be noted that McGinn himself has upon death of the physical body. We should repeatedly claimed that his position is not note that Cartesian substance dualism drew dualist. Nonetheless others have accused him much of its motivation from religious con- of being committed to dualism (e.g., Brueck- siderations, partly because it provided for the ner and Berukhim, 2003). There is no doubt survival of the soul. The main difﬁculty his- that McGinn does not intend to commit to torically associated with it is whether it can dualism. In a way, his position is precisely account for the causal interaction between that, because of our cognitive closure we can- the mind and the body. not even know whether materialism or dual- 18. So property dualism is compatible with sub- ism is true. Yet it is a fair criticism to suggest stance monism. Unlike Descartes and other that McGinn is committed to dualism despite old-school dualists, modern dualists for the himself because his argument for mysterian- most part hold that there is only one kind of ism would not go through unless dualism was stuff, or substance, in the world – matter. But true. matter has two different kinds of properties – 14. More generally, it is curious to hold, as material and immaterial. McGinn does, that an organism’s concept- 19. A kind of property F supervenes on a kind forming procedures are powerful enough to of property G with logical necessity – or for frame a problem, without being powerful short logically supervenes on them – just in enough to frame the solution. To be sure, case two objects differing with respect to their the wrong solution may be framed, but this F properties without differing with respect would suggest not that the conceptual capa- to their G properties would be in contraven- bilities of the organism are at fault, but rather tion of the laws of logic. A kind of prop- that the organism made the wrong turn some- erty F supervenes on a kind of property G where in its reasoning. The natural thought with metaphysical necessity – or for short P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

56 the cambridge handbook of consciousness

metaphysically supervenes on them – just in tial generalization (the inference from “a is F” case it is impossible for two objects to dif- to “there is an x, such that x is F”) and sub- fer with respect to their F properties with- stitution of co-referential terms salva veritate out differing with respect to their G prop- (the inference from “a is F” and “a = b” to “b is erties. Philosophers debate whether there is F”). Epistemic contexts – contexts involving a difference between the two (logical and the ascription of knowledge – are intensional metaphysical supervenience). That debate in this sense. does not concern us here. 26. Another popular materialist response to these 20. This stronger claim will require a stronger arguments is that what is being gained is argument. The claim that phenomenal prop- not new knowledge, but rather new abil- erties are not identical to physical proper- ities (Lewis, 1990; Nemirow, 1990). Upon ties could be established through the now being released from her room, the Knowledge familiar argument from multiple realizability Argument’s protagonist does not acquire new (Putnam, 1967). But multiple realizability knowledge, but rather a new set of abilities. does not entail failure of supervenience. To And likewise what we lack with respect to obtain the latter, Chalmers will have to appeal what it is like to be a bat is not any particular to a different argument, as we will see in the knowledge, but a certain ability – the ability next subsection. to imagine what it is like to be a bat. But from 21. As a consequence, phenomenal properties do the acquisition of a new ability one can surely supervene on physical properties with nomo- not infer the existence of a new fact. logical necessity, even though they do not 27. Materialists reason that because what it is supervene with metaphysical or logical neces- like to see red is identical to a neurophysi- sity. A kind of property F supervenes on a kind ological fact about the brain, and ex hypoth- of property G with nomological (or natural) esi the Knowledge Argument’s protagonist necessity – or for short nomologically super- knows the latter fact, she already knows the venes on them – just in case two objects differ- former. So she knows the fact of what it is like ing with respect to their F properties without to see red, but not as a fact about what it is differing with respect to their G properties like to see red. Instead, she knows the fact of would be in contravention of laws of nature. what it is like to see red as a fact about the 22. So causal explanation is the sort of explana- neurophysiology of the brain. What happens tion one obtains by citing the cause of the when she comes out of her room is that she explanandum. For discussions of the nature comes to know the fact of what it is like to of causal explanation, see (e.g., Lewis, 1993). see red as a fact about what it is like to see red. That is, she learns in a new way a fact 23. The latter will govern only the causal inter- she already knew in another way. The same action among physical events. They will not applies to knowledge of what it is like to be a cover causal interaction between physical and bat: we may know all the facts about what it phenomenal, non-physical events. These will is like to see a bat, and still gain new knowl- have to be covered by a special and new set edge about bats, but this new knowledge will of laws. present to us a fact we already know in a way 24. In Baars’ (1988, 1997) Global Workspace The- we do not know it yet. ory, consciousness is reductively explained in 28. It could be responded by the dualist that terms of global availability. In a functionalist some pieces of knowledge are so different that 1981 1991 theory such as Dennett’s ( , ), con- the fact known thereby could not possibly sciousness is reductively explained in terms of turn out to be the same. Knowledge that the functional organization. Chalmers’ position is evening star is glowing and knowledge that that neither theory can explain consciousness the morning star is glowing are not such. But reductively, though both may ﬁgure as part consider knowledge that justice is good and of the causal explanation of it. These the- knowledge that banana is good. The dual- ories are not discussed in the present chap- ist could argue that these are such different ter, because they are fundamentally psycho- pieces of knowledge that it is impossible that logical (rather than philosophical) theories of the facts thereby known should turn out to consciousness. be one and the same. The concepts of evening 25. A linguistic context is intensional if it disal- star and morning star are not different enough lows certain inferences, in particular existen- to exclude the possibility that they pick out P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 57

the same thing, but the concepts of justice not mean, however, that the property is actu- and banana are such that it cannot possibly ally instantiated by any actual object.) But on be the case that justice should turn out to be a sparse conception of property – one that the same thing as bananas. rejects the latitudinous assumption – there 29. The kind of possibility we are concerned with may not be appearance properties at all. here, and in the following presentation of 36. The notion of a natural property is hard to variations on this argument, is not practical pin down and is the subject of philosophi- possibility, or even a matter of consistency cal debate. The most straightforward way of with the laws of nature. Rather it is possibil- understanding natural properties is as proper- ity in the widest possible sense – that of con- ties that figure in the ultimate laws of nature sistency with the laws of logic and the very (Armstrong, 1978; Fodor, 1974). essence of things. This is what philosophers 37. That is, they would have their causal effi- refer to as metaphysical possibility. cacy restricted to bringing about physi- 30. The modal force of this supervenience claim cal events and property-instantiations that is concordant with that of the claim in already have independent sufficient causes Premise 2; that is, that of metaphysical (and that would therefore take place any- necessity. way, regardless of the non-supervenient prop- 31. The reason it is impossible is that there is erties. (This is the second option of the no such thing as contingent identity, according dilemma.) to the official doctrine hailing from Kripke. 38. This is the strategy in Chalmers (1996). Since all identity is necessary, and necessity is Later on, Chalmers (2002a) embraces a three- cashed out as truth in all possible worlds, it pronged approach, the third prong consisting follows that when a = b in the actual world, in accepting causal overdetermination. a = b in all possible worlds, that is, a is neces- 39. When a cause C causes an effect E, C’s caus- sarily identical to b. ing of E may have its own (mostly acciden- 32 . The interpretation I provide is based on cer- tal) effects (e.g., it may surprise an observer 1996 131 tain key passages in Chalmers ( ,pp. – who did not expect the causing to take place), 134 ), but I cast the argument in terms that are but E is not one of them. This is because mine, not Chalmers’. E is caused by C, not by C’s causing of E. 33. I mean the property of apparent water to be Dretske (1988) distinguished between trig- more or less the same as the property philoso- gering causes and structuring causes, the lat- phers often refer to as “watery stuff” (i.e., ter being causes of certain causal relations the property of being superficially (or to the (such as C’s causing of E), and offers an naked eye) the same as water – clear, drink- account of structuring causes. But this is an able, liquid, etc.). account of the causes of causal relations, not 34. Chalmers (1996, 132) writes, “...theprimary of their effects. To my knowledge, there is intension [of “consciousness”] determines a no account of the effects of causal relations, perfectly good property of objects in possible mainly because these seem to be chiefly acci- worlds. The property of being watery stuff dental. [or apparent water] is a perfectly reasonable 40. Or at least they would be nearly epiphenom- property, even though it is not the same as the enal, having no causal powers except perhaps property of being H2 O. If we can show that to bring about some accidental effects of the there are possible worlds that are physically sort pointed out in the previous endnote. identical to ours but in which the properly 41. By “representational properties” it is meant introduced by the primary intension is lack- properties that the experience has in virtue ing, then dualism will follow [italics added].” of what it represents – not, it is important to 35. Our discussion so far has presupposed a “lat- stress, properties the experience has in virtue itudinous” approach to properties, according of what does the representing. In terms of the to which there is a property that corresponds distinction between vehicle and content, rep- to every predicate we can come up with. resentational properties are to be understood (Thus, if we can come up with the predi- as content properties rather than vehicular cate, “is a six-headed space lizard or a flying properties. We can also make a distinction cow,” then there is the property of being a six- between two kinds of vehicular properties: headed space lizard or a flying cow. This does those that are essential to the vehicling of P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

58 the cambridge handbook of consciousness

the content and those that are not. (Block’s Shoemaker (1994a, b, 1996, 2002) and Thau (1996) distinction between mental paint and (2002). Some of these versions are impor- mental latex (later, “mental oil”) is supposed tantly different from Tye’s, not only in detail to capture this distinction.) There is a sense in but also in spirit. This is particularly so which a view according to which phenomenal with regard to Shoemaker’s view (as well properties are reductively accountable for in as Lurz’s). For a limited defense and elab- terms of vehicular properties essential to the oration of Shoemaker’s view, see Kriegel vehicling is representational, but the way the (2002a, b). In what way this defense is term “representationalism” is used in current limited will become evident at the end of discussions of consciousness, it does not qual- this section. ify as representationalism. A view of this sort 45. The properties of intentionality and abstract- 1989 is defended, for instance, by Maloney ( ), ness are fairly straightforward. The for- but otherwise lacks a vast following. I do not mer is a matter of intensionality; that is, discuss it here. the disallowing of existential generalizations 42. By the “phenomenal character” of a mental and truth-preserving substitutions of co- state at a time t I mean the set of all phenom- referential terms. The second is a matter of enal properties the state in question instanti- the features represented by experience not ates at t. By “representational content” I mean being concrete entities (this is intended to whatever the experience represents. (Experi- make sense of misrepresentation of the same ences represent things, in that they have cer- features, in which case no concrete entity is tain accuracy or veridicality conditions: con- being represented). ditions under which an experience would be 46. This line of thought can be resisted on a said to get things right.) number of scores. First, it could be argued 43. See Dretske (1981, 1988) for the most thor- that I do have a short-lived concept of blue17 , oughly worked out reductive account of men- which I possess more or less for the dura- tal representation in informational and tele- tion of my experience. Second, it could be 1981 ological terms. According to Dretske ( ), claimed that although I do not possess the every event in the world generates a certain descriptive concept “blue17 ,” I do possess the amount of information (in virtue of exclud- indexical concept “this shade of blue,” and ing the possibility that an incompatible event that it is the latter concept that is deployed can take place). Some events also take place in my experience’s representational content. only when other events take place as well, Be that as it may, the fact that conscious and this is sometimes dictated by the laws of experiences can represent properties that nature. Thus it may be a law of nature that the subject cannot recognize across rela- dictates that an event type E1 is betokened tively short stretches of time is signiﬁcant only when event type E2 is betokened. When enough. Even if we do not wish to treat them this is the case, E1 is said to be nomically depen- as non-conceptual, we must treat them at dent upon E2 , and the tokening of E1 carries least as “sub-recognitional.” Tye’s modiﬁed the information that E2 has been betokened. claim would be that the representational con- Or more accurately, the tokening of E1 carries tent of experience is poised, abstract, sub- the information generated by the tokening of recognitional, intentional content. E2 . Some brain states bear this sort of rela- 47 tion to world states: the former come into . To be sure, it does not represent the tissue being, as a matter of law, only when the latter damage as tissue damage, but it does repre- do (i.e., the former are nomically dependent sent the tissue damage. Since the represen- upon the latter). Thus, a certain type of brain tation is non-conceptual, it certainly cannot state may be tokened only when it rains. This employ the concept of “tissue damage.” brain state type would thus carry the informa- 48. An error theory is a theory that ascribes tion that it rains. An informational account of a widespread error in commonsense beliefs. mental representation is based on this idea: The term was coined by J. L. Mackie (1977). that a brain state can represent the fact that it Mackie argued that values and value judg- rains by carrying information about it, which ment are subjective. Oversimplifying the it does in virtue of nomically depending on it. dialectic, a problem for this view is that such a 44. Other representational theories can be found judgment as “murder is wrong” appears to be, in Byrne (2001), Dretske (1995), Lurz (2003), and is commonly taken to be, objectively true. P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 59

In response Mackie embraced what he termed 54. There are versions of representationalism that an error theory: that the common view of may be better equipped to deal with the sub- moral and value judgments is simply one huge jective character of experience. Thus, accord- mistake. ing to Shoemaker’s (2002) version, a men- 49. Externalism about representational content, tal state is conscious when it represents a or “content externalism” for short, is the subject-relative feature, such as the disposi- thesis that the representational content of tion to bring about certain internal states in experiences, thoughts, and even spoken state- the subject. It is possible that some kind of ments is partially determined by objects out- for-me-ness can be accounted for in this man- side the subject’s head. Thus, if a person’s ner. It should be noted, however, that this interactions with watery stuff happen to is not one of the considerations that moti- be interactions with H2 O, and another per- vate Shoemaker to develop his theory the way son’s interactions with watery stuff happen he does. to be interactions with a superficially simi- 55. Rosenthal prefers to put this idea as follows: lar stuff that is not composed of H2 O, then conscious states are states we are conscious even if the two persons cannot tell apart of. He then draws a distinction between con- H2 O and the other stuff and are unaware sciousness and consciousness of – intransi- of the differences in the molecular com- tive and transitive consciousness (Rosenthal, position of the watery stuff in their envi- 1986, 1990). To avoid unnecessary confusion, ronment, the representational contents of I state the same idea in terms of awareness- their respective water thoughts (as well as of, rather than consciousness-of. But the idea water pronouncements and water experi- is the same. It is what Rosenthal calls some- ences) are different (Putnam, 1975). Or so times the “transitivity principle” (e.g., Rosen- externalists claim. thal, 2000): a mental state is intransitively 50. Another option is to go internalist with conscious only if we are transitively conscious respect to the representational content that of it. determines the phenomenal properties of 56. The representation is “higher-order” in the conscious experiences. With the recent sense that it is a representation of a repre- advent of credible account of narrow con- sentation. In this sense, a first-order represen- tent (Chalmers, 2002b, Segal, 2000), it is now tation is a representation of something that is a real option to claim that the phenomenal not itself a representation. Any other repre- properties of experience are determined by sentation is higher-order. experience’s narrow content (Kriegel, 2002a; 57. More than that, according to Rosenthal Rey, 1998). However, it may turn out that (1990), for instance, the particular way it is this version of representationalism will not like for S to have M is determined by the par- be as well supported by the transparency of ticular way M∗ represents M. Suppose S tastes experience. an identical wine in 1980 and in 1990. During 51. For one such line of criticism, on which I do the 1980s, however, S had become a wine con- not elaborate here, see Kriegel (2002c). noisseur. Consequently, wines she could not 1980 1990 52. Elsewhere, I construe this form of pre- distinguish at all in strike her in as reflective self-consciousness as what I call worlds apart. That is, during the eighties she intransitive self-consciousness. Intransitive self- acquired a myriad of concepts for very spe- consciousness is to be contrasted with transi- cific and subtle wine tastes. It is plausible to tive self-consciousness. The latter is ascribed claim that what it is like for S to taste the 1990 in reports of the form “I am self-conscious wine in is different from what it was 1980 of my thinking that p,” whereas the former like for her to taste it in – even though is ascribed in reports of the form “I am self- the wines’ own flavors are identical. Arguably, consciously thinking that p.” For details see the reason for the difference in what it is like Kriegel (2003b, 2004b). to taste the wine is that the two wine-tasting experiences are accompanied by radically dif- 53. Part of this neglect is justified by the thesis ferent higher-order representations of them. that the for-me-ness of conscious experiences This suggests, then, that the higher-order rep- is an illusory phenomenon. For an argument resentation not only determines that there is for the psychological reality of it, see Kriegel something it is like for S to have M, but also (2004b). what it is like for S to have M. P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

60 the cambridge handbook of consciousness

58. I do not mean the term “yield” in a causal 64. Earlier on, Rosenthal (1990) required that sense here. The higher-order monitoring the- the higher-order thought be not only ory does not claim that M∗’s represent- non-inferential but also non-observational. ing of M somehow produces,orgives rise This latter requirement was later dropped to, M’s being conscious. Rather, the claim (Rosenthal, 1993). is conceptual: M’s being conscious consists 65. A person may come to believe that she is in,oris constituted by,M∗’s representing ashamed about something on the strength of of M. her therapist’s evidence. And yet the shame 59. Other versions of the higher-order thought state is not conscious. In terms of the ter- view can be found in Carruthers (1989, 1996), minology introduced in the introduction, the Dennett (1969, 1991), and Mellor (1978). state may become availability-conscious, but not phenomenally conscious. This is why 60 1990 739 40 . Rosenthal ( ,pp. – ) claims that it is the immediacy of awareness is so crucial. essential to a perceptual state that it has a sen- Although the person’s second-order belief sory quality, but the second-order represen- constitutes an awareness of the shame state, it tations do not have sensory qualities and are is not a non-inferential awareness, and there- 2001 therefore non-perceptual. Van Gulick ( ) fore not immediate awareness. details a longer and more thorough list of 66. De se content is content that is of one- features that are characteristic of perceptual self, or more precisely, of oneself as oneself. states and considers which of them is likely Castaneda (1966), who introduced this term, to be shared by the higher-order represen- ˜ also claimed that de se content is irreducible tations. His conclusion is that some are and to any other kind of content. This latter claim some are not. is debatable and is not part of the official 61 . The notion of direction of fit has its origins in higher-order thought theory. the work of Anscombe (1957), but has been 67. Rosenthal’s (1990,p.742) argument for this developed in some detail and put to extensive requirement is the following. My awareness work mainly by Searle (1983). The idea is that of my bluish experience is an awareness of mental states divide into two main groups, the that particular experience, not of the general cognitive ones (paradigmatically, belief) and type of experience it is. But it is impossi- the conative ones (paradigmatically, desire). ble to represent a mental state as particu- The former are such that they are supposed to lar without representing in which subject it make the mind fit the way the world is (thus occurs. Therefore, the only way the higher- “getting the facts right”), whereas the latter order thought could represent my experience are such that they are supposed to make the in its particularity is if it represented it as world fit the way the mind is (a change in the occurring in me. world is what would satisfy them). 68. This is necessary to avert infinite regress. If 62. Kobes (1995) suggests a version of higher- the higher-order state was itself conscious, order monitoring theory in which the higher- it would have to be itself represented by a order representation has essentially a telic yet higher-order state (according to the the- direction of fit. But Rosenthal construes it as ory) and so the hierarchy of states would having only a thetic one. go to infinity. This is problematic on two 63. Carruthers (1989, 1996, 2000), and probably scores. Firstly, it is empirically implausible, also Dennett (1969, 1991), attempt to account and perhaps impossible, that a subject should for consciousness in terms of merely tacit entertain an infinity of mental states when- or dispositional higher-order representations. ever conscious. Secondly, if a mental state’s But these would not do, according to Rosen- being conscious is explained in terms of thal. The reason for this is that a merely dis- another conscious states, the explanation is positional representation would not make the “empty,” inasmuch as it does not explain con- subject aware of her conscious state, but only sciousness in terms of something other than disposed to being aware of it, whereas the cen- consciousness. tral motivation behind the higher-order mon- 69. This claim can be made on phenomenological itoring view is the fact that conscious states grounds, instead of on the basis of conceptual are states we are aware of having (Rosenthal analysis. For details, see Kriegel (2004b). 1990,p.742). 70. To repeat, the conceptual grounds are the fact that it seems to be a conceptual truth that P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 61

conscious states are states we are aware of 2002, 2006), Kobes (1995), Kriegel (2003a, having. This seems to be somehow inherent 2005, 2006a), and Van Gulick (2001, 2004). in the very concept of consciousness. 79. For fuller discussion of Brentano’s account, 71. There are other arguments that have been see Caston (2002), Kriegel (2003a), Smith leveled against the higher-order monitoring (1986, 1989) Thomasson (2000), and Zahavi theory, or specific versions thereof, which I (1998, 2004). do not have the space to examine. For argu- 80. So the self-representational content of con- ments not discussed here, see Block (1995), scious states is de se content. There are places Caston (2002), Dretske (1995), Guzeldere where Brentano seems to hold something like (1995), Kriegel (2006a), Levine (2001), Nat- this as well. See also Kriegel (2003a). 1993 1988 1999 soulas ( ), Rey ( ), Seager ( ), and 81. For more on the distinction between content 1998 Zahavi and Parnas ( ). and attitude (or mode), see Searle (1983). 72. The argument has also been made by Caston For a critique of Smith’s view, see Kriegel (2002), Levine (2001), and Seager (1999). For (2005a). a version of the argument directed at higher- 82. A similar account would be that conscious order perception theory (and appealing to states are not conscious in virtue of stand- higher-order misperceptions), see Neander ing in a certain relation to themselves, but 1998 ( ). this is because their secondary intentionality ∗ 73. Note that M does not merely misrepresent should be given an adverbial analysis. This M to be F when in reality M is not F, but is not to say that all intentionality must be misrepresents M to be F when in reality there treated adverbially. It may well be that the is no M at all. primary intentionality of conscious states is a 74. This would be a particular version of the sup- matter of their standing in a certain informa- position we made at the very beginning of this tional or teleological relation to their primary chapter, by way of analyzing creature con- objects. Thus, it need not be the case that S’s sciousness in terms of state consciousness. conscious fear that p involves S’s fearing p- 75. Furthermore, if M∗ were normally conscious, ly rather than S’s standing in a fear relation the same problem would arise with the third- to the fact that p. But it is the case that S’s order representation of it (and if the third- awareness of her fear that p involves being order representation were normally con- aware fear-that-p-ly rather than standing in scious, the problem would arise with the an awareness relation to the fear that p.To fourth-order state). To avert infinite regress, my knowledge, nobody holds this view. the higher-order monitoring theorist must 83. A constitutive, non-contingent relation is a somewhere posit an unconscious state, and relation that two things do not just happen to when she does, she will be unable to claim entertain, but rather they would not be the that that state instantiates the property of things they are if they did not entertain those being conscious when it misrepresents. relations. Thus A’s relation to B is constitutive 76. This appears to be Rosenthal’s latest stance if bearing it to B is part of what constitutes A’s on the issue (in conversation). being what it is. Such a relation is necessary 77. There are surely other ways the higher-order rather than contingent, since there is no pos- monitoring theorist may try to handle the sible world in which A does not bear it to B – problem of targetless higher-order represen- for in such a world it would no longer be A. tations. But many of them are implausible, 84. Elsewhere, I have defended a view similar in and all of them complicate the theory con- key respects to Van Gulick’s – see Kriegel siderably. One of the initial attractions of the (2003a, 2005, 2006a). theory is its clarity and relative simplicity. 85. Indeed, the problem may be even more press- Once it is modified along any of the lines ing for a view such as the higher-order global sketched above, it becomes significantly less states theory. For the latter requires not only clear and simple. To that extent, it is consid- the ability to generate higher-order contents, erably less attractive than it initially appears. but also the ability to integrate those with the 78. See Brook and Raymont (2006), Caston right lower-order contents. (2002), Hossack (2002), Kriegel (2003b), and 86. For a more elaborate argument that self- Williford (2006). For the close variation, see representation may not be a sufficient condi- Carruthers (2000, 2006), Gennaro (1996, tion for consciousness, one that could provide P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

62 the cambridge handbook of consciousness

a reductive explanation of it, see Levine Armstrong, D. M. (1978). A theory of univer- (2001, Ch. 6). sals, Vol. 2. Cambridge: Cambridge University 87. I am appealing here to a distinction defended, Press. e.g., by Cummins (1979), Dretske (1988), Armstrong, D. M. (1981). What is consciousness? and Searle (1992). Grice noted that some In D. M. Armstrong (Ed.), The nature of mind things that exhibit aboutness of meaningful- (pp. 55–67). Ithaca, NY: Cornell University ness, such as words, traffic signs, and arrows, Press. do so only on the assumption that someone Anscombe, G. E. M. (1957). Intention. Oxford: interprets them to have the sort of meaning Blackwell. they have. But these acts of interpretation Baars, B. (1988). A cognitive theory of consciousness. are themselves contentful, or meaningful. So New York: Cambridge University Press. their own meaning must be either derived by Baars, B. (1997). In the theater of consciousness: The further interpretative acts or be intrinsic to workspace of the mind. Oxford: Oxford Univer- them and non-derivative. Grice’s claim was sity Press. that thoughts and other mental states have 1990 an aboutness all their own, independently of Block, N. J. ( a). Inverted earth. Philosophical 52 79 any interpretation. Perspective, 4, – . 1990 88. This is denied by Dennett (1987), who claims Block, N. J. ( b). Can the mind change that all intentionality is derivative. the world? In G. Boolos (Ed.), Meaning and method: Essays in honor of Hilary Putnam.New 89. One might claim that such states are York: Cambridge University Press. less clearly conceivable when their self- 1995 representational content is fully specified. Block, N. J. ( ). On a confusion about the Thus, if the content is of the form, “I myself function of consciousness. Behavioral and Brain 227 247 am herewith having this very bluish experi- Sciences, 18, – . ence,” it is less clearly the case that one can Block, N. J. (1996). Mental paint and mental latex. conceive of an unconscious state having this Philosophical Issues, 7, 19–50. content. Block, N. J., Flanagan, O., & Guzeldere, G. (Eds.). 90. The conceivability of unconscious self- (1997). The nature of consciousness: Philosophi- representing states may not be proof of their cal debates. Cambridge, MA: MIT Press. possibility, but it is evidence of their possibil- Brentano, F. (1973). Psychology from empirical ity. It is therefore evidence against the self- standpoint (O. Kraus, Ed. & A. C. Rancurello, representational theory. D. B. Terrell, & L. L. McAlister, Trans.). 91. The reductivist may claim that zombies with London: Routledge and Kegan Paul. (Original the same physical properties we have are con- work published 1874) ceivable only because we are not yet in a posi- Brook, A., & Raymont, P. (2006). A unified tion to focus our mind on the right physical theory of consciousness. Cambridge, MA: MIT structure. As progress is made toward identi- Press. fication of the right physical structure, it will Brueckner, A., & Berukhim, E. (2003). McGinn become harder and harder to conceive of a on consciousness and the mind-body problem.” zombie exhibiting this structure but lacking In Q. Smith & D. Jokic (Eds.), Consciousness: all consciousness. New philosophical perspectives. Oxford: Oxford 92. For comments on an earlier draft of this University Press. chapter, I would like to thank George Byrne, D. (1997). Some like it HOT: Conscious- Graham, David Jehle, Christopher Mal- ness and higher order thoughts. Philosophical oney, Amie Thomasson, and especially David Studies, 86, 103–129. Chalmers. Byrne, A. (2001). Intentionalism defended. Philo- sophical Review, 110, 199–240. Carruthers, P. (1989). Brute experience. Journal References of Philosophy, 85, 258–269. Carruthers, P. (1996). Language, thought, and con- Armstrong, D. M. (1968). A materialist theory of sciousness. Cambridge: Cambridge University the mind. New York: Humanities Press. Press. P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 63

Carruthers, P. (1998). Natural theories of con- Dennett, D. C. (1991). Consciousness explained. sciousness. European Journal of Philosophy, 6, Cambridge, MA: MIT Press. 203–222. Dennett, D. C. (1995). Darwin’s dangerous idea. Carruthers, P. (1999). Sympathy and subjectiv- New York: Simon and Schuster. ity. Australasian Journal of Philosophy, 77, 465– Dretske, F. I. (1981). Knowledge and the ﬂow of 482. information. Oxford: Clarendon. Carruthers, P. (2000). Phenomenal consciousness. Dretske, F. I. (1988). Explaining behavior. Cam- New York: Cambridge University Press. bridge MA: MIT Press. Carruthers, P. (2006). Conscious experience ver- Dretske, F. I. (1993). Conscious experience. sus conscious thought. In U. Kriegel & K. Willi- Mind, 102, 263–283. ford (Eds.), Consciousness and self-reference. Dretske, F. I. (1995). Naturalizing the mind. Cam- Cambridge, MA: MIT Press. bridge MA: MIT Press. 1966 Castaneda,˜ H.-N. ( ). ‘He’: A study in the Fodor, J. A. (1974). Special sciences. Synthese, 28, 130 157 logic of self-consciousness. Ratio, 8, – . 97–115. 2002 Caston, V. ( ). Aristotle on consciousness. Foster, J. (1982). The case for idealism. London: 751 815 Mind, 111, – . Routledge. 1995 Chalmers, D. J. ( ). Facing up to the prob- Gennaro, R. J. (1996). Consciousness and self- lem of consciousness. Journal of Consciousness consciousness. Philadelphia: John Benjamin. Studies, 2 , 200–219. Gennaro, R. J. (2002). Jean-Paul Sartre and the 1996 Chalmers, D. J. ( ). The conscious mind. HOT theory of consciousness. Canadian Jour- Oxford: Oxford University Press. nal of Philosophy, 32, 293–330. 2002 Chalmers, D. J. ( a). Consciousness and its Gennaro, R. J. (2006). Between pure self- place in nature. In D. J. Chalmers (Ed.), Philos- referentialism and (extrinsic) HOT theory. In ophy of mind. Oxford: Oxford University Press. U. Kriegel & K. Williford (Eds.), Conscious- Chalmers, D. J. (2002b). The components of con- ness and self-reference. Cambridge, MA: MIT tent. In D. J. Chalmers (Ed.), Philosophy of Press. mind. Oxford: Oxford University Press. Goldman, A. (1993a). The psychology of folk psy- Churchland, P. M. (1979). Scientiﬁc realism and chology. Behavioral and Brain Sciences, 16, 15– the plasticity of mind. New York: Cambridge 28. University Press. Goldman, A. (1993b). Consciousness, folk psy- Churchland, P. M. (1985). Reduction, qualia, and chology, and cognitive science. Consciousness the direct introspection of brain states. Journal and Cognition, 2 , 364–383. 8 28 of Philosophy, 82, – . Grice, P. (1957). Meaning. Philosophical Review, Crick, F., & Koch. C. (1990). Towards a neurobio- 66, 377–388. logical theory of consciousness. Seminars in the Guzeldere, G. (1995). Is consciousness the per- 263 275 Neurosciences, 2 , – . ception of what passes in one’s own mind? Crick, F., & Koch. C. (2003). A framework for In T. Metzinger (Ed.), Conscious experience. consciousness. Nature Neuroscience, 6, 119–126. Padborn: Schoeningh-Verlag. Cummins, R. (1979). Intention, meaning, and Harman, G. (1990). The intrinsic quality of truth conditions. Philosophical Studies, 35, experience. Philosophical Perspectives, 4, 31– 345–360. 52. DeBellis, M. (1991). The representational con- Horgan, T., & Tienson, J. (2002). The intention- tent of musical experience. Philosophy and Phe- ality of phenomenology and the phenomenol- nomenological Research, 51, 303–324. ogy of intentionality. In D. J. Chalmers (Ed.), Dennett, D. C. (1969). Consciousness and content. Philosophy of mind. Oxford: Oxford University London: Routledge. Press. Dennett, D. C. (1981). Towards a cognitive the- Hossack, K. (2002). Self-knowledge and con- ory of consciousness. In D. C. Dennett (Ed.), sciousness. Proceedings of the Aristotelian Soci- Brainstorms. Brighton: Harvester. ety, 163–181. Dennett, D. C. (1987). The intentional stance. Jackson, F. (1984). Epiphenomenal qualia. Philo- Cambridge, MA: MIT Press. sophical Quarterly, 34, 147–152. P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

64 the cambridge handbook of consciousness

Kim, J. (1989a). The myth of nonreductive mate- Kripke, S. (1980). The identity thesis. Reprinted rialism. Proceedings and Addresses of the Amer- in N. J. Block, O. Flanagan, & G. Guzeldere ican Philosophical Association, 63, 31–47. (Eds.), (1997). The nature of consciousness: Kim, J. (1989b). Mechanism, purpose, and Philosophical debates. Cambridge, MA: MIT explanatory exclusion. Philosophical Perspec- Press. tives, 3, 77–108. Lehrer, K. (1996). Skepticism, lucid content, and Kim, J. (1992). Multiple realization and the the metamental loop. In A. Clark, J. Ezquerro, metaphysics of reduction. Philosophy and Phe- & J. M. Larrazabal (Eds.), Philosophy and cogni- nomenological Research, 52, 1–26. tive science. Dordrecht: Kluwer. Kobes, B. W. (1995). Telic higher-order thoughts Lehrer, K. (1997). Self-trust: A study of rea- and Moore’s paradox. Philosophical Perspec- son, knowledge, and autonomy. Oxford: Oxford tives, 9, 291–312. University Press. Kriegel, U. (2002a). Phenomenal content. Erken- Levine, J. (1983). Materialism and qualia: The ntnis, 57, 175–198. explanatory gap. Pacific Philosophical Quar- terly, 64, 354–361. Kriegel, U. (2002b). Emotional content. Con- 2001 sciousness and Emotion, 3, 213–230. Levine, J. ( ). Purple haze: The puzzle of consciousness. Oxford: Oxford University Press. Kriegel, U. (2002c). PANIC theory and the prospects for a representational theory of phe- Lewis, D. K. (1990). What experience teaches. In nomenal consciousness. Philosophical Psychol- W. G. Lycan (Ed.), Mind and cognition. Oxford: ogy, 15, 55–64. Blackwell. 1993 Kriegel U. (2003a). Consciousness, higher-order Lewis, D. K. ( ). Causal explanation. In content, and the individuation of vehicles. Syn- D.-H. Ruben (Ed.), Explanation. Oxford: these, 134, 477–504. Oxford University Press. Kriegel, U. (2003b). Consciousness as intransi- Libet, B. (1985). Unconscious cerebral initiative tive self-consciousness: Two views and an argu- and the role of conscious will in voluntary ment. Canadian Journal of Philosophy, 33, 103– action. Behavioral and Brain Sciences, 8, 529– 132. 566. Kriegel, U. (2004a). The new mysterianism and Loar, B. (1990). Phenomenal states. Philosophical the thesis of cognitive closure. Acta Analytica, Perspectives, 4, 81–108. 18, 177–191. Lurz, R. (1999). Animal consciousness. Journal of Kriegel, U. (2004b). Consciousness and self- Philosophical Research, 24, 149–168. consciousness. The Monist 87, 185–209. Lurz, R. (2003). Neither HOT nor COLD: An Kriegel, U. (2005a). Naturalizing subjective alternative account of consciousness. Psyche, 9. character. Philosophy and Phenomenological Lycan, W. G. (1987). Consciousness. Cambridge, Research, 71, 23–57. MA: MIT Press. 2005 Kriegel, U. ( b). Review of Jeffrey Gray, Lycan, W. (1996). Consciousness and experience. Consciousness: creeping up on the hard problem. Cambridge, MA: MIT Press. Mind, 114, 417–421. Lycan, W. G. (2001). A simple argument for 2006 Kriegel, U. ( a). The same-order mon- a higher-order representation theory of con- itoring theory of consciousness In U. sciousness. Analysis, 61, 3–4. Kriegel & K. Williford (Eds.), Conscious- McGinn, C. (1989). Can we solve the mind-body ness and self-reference. Cambridge, MA: MIT problem? Mind, 98, 349–366. Press. McGinn, C. (1995). Consciousness and space. Kriegel, U. (2006b). The concept of conscious- Journal of Consciousness Studies, 2 , 220– ness in the cognitive sciences: Phenomenal con- 30. sciousness, access consciousness, and scientific 1999 practice. In P. Thagard (Ed.), Handbook of the McGinn, C. ( ). The mysterious flame. Cam- philosophy of psychology and cognitive science. bridge, MA: MIT Press. Amsterdam: North-Holland. McGinn, C. (2004). Consciousness and its objects. Kriegel, U., & Williford, K. (eds.). (2006). Oxford: Oxford University Press. Self-representational approaches to consciousness. Mackie, J. L. (1977). Ethics: Inventing right and Cambridge, MA: MIT Press. wrong. New York: Penguin. P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

philosophical theories of consciousness: contemporary western perspectives 65

Maloney, J. C. (1989). The mundane matter of the Macmillan encyclopedia of cognitive science. New mental language.New York: Cambridge Univer- York: Macmillan. sity Press. Seager, W. (1999). Theories of consciousness. Mellor, D. H. (1978). Conscious belief. Proceed- London: Routledge. ings of the Aristotelian Society, 78, 87–101. Searle, J. R. (1983). Intentionality: An essay in the Moore, G. E. (1903). The refutation of ideal- philosophy of mind. New York: Cambridge Uni- ism. In G. E. Moore (Ed.), Philosophical papers. versity Press. London: Routledge and Kegan Paul. Searle, J. R. (1992). The rediscovery of mind. Nagel, T. (1974). What is it like to be a bat? Philo- Cambridge, MA: MIT Press. sophical Review, 83, 435–450. Segal, G. (2000). A slim book on narrow content. Natsoulas, T. (1993). What is wrong with Cambridge, MA: MIT Press. appendage theory of consciousness? Philosoph- Shoemaker, S. (1994a). Phenomenal character. 137 154 ical Psychology, 6, – . Nous, 28, 21–38. 1998 Neander, K. ( ). The division of phenomenal Shoemaker, S. (1994b). Self-knowledge and labor: A problem for representational theories ‘inner sense.’ Lecture III: The phenomenal of consciousness. Philosophical Perspectives, 12, character of experience. Philosophy and Phe- 411 434 – . nomenological Research, 54, 291–314. 1990 Nemirow, L. ( ). Physicalism and the cogni- Shoemaker, S. (1996). Colors, subjective reactive role of acquaintance. In W. G. Lycan (Ed.), tions, and qualia. Philosophical Issues, 7, 55–66. Mind and cognition. Oxford: Blackwell. Shoemaker, S. (2002). Introspection and phe- Peacocke, C. (1983). Sense and content. Oxford: nomenal character. In D. J. Chalmers (Ed.), Clarendon. Philosophy of mind. Oxford: Oxford University 1967 Putnam, H. ( ). The nature of mental states. Press. In D. M. Rosenthal (Ed.), The nature of mind. Siewert, C. P. (1998). The signiﬁcance of conscious- Oxford: Oxford University Press. ness. Princeton, NJ: Princeton University Press. Putnam, H. (1975). The meaning of ‘meaning.’ In Smart, J. J. C. (1959). Sensations and brain pro- H. Putnam (Ed.), Mind, language, and reality, cesses. Philosophical Review, 68, 141–156. New York: Cambridge University Press. Smith, D. W. (1986). The structure of (self-) con- Rey,G.(1988). A question about consciousness. sciousness. Topoi, 5, 149–156. In H. Otto & J. Tueidio (Eds.), Perspectives on mind. Norwell: Kluwer Academic Publishers. Smith, D. W. (1989). The circle of acquaintance. Rey, G. (1998). A narrow representationalist Dordrecht: Kluwer Academic Publishers. account of qualitative experience. Philosophi- Smith, D. W. (2004). Return to consciousness. cal Perspectives, 12, 435–457. In D. W. Smith (Ed.), Mind world. New York: Rosenthal, D. M. (1986). Two concepts of Cambridge University Press. consciousness. Philosophical Studies, 94, 329– Thau, M. (2002). Consciousness and cognition. 359. Oxford: Oxford University Press. Rosenthal, D. M. (1990). A theory of con- Thomasson, A. L. (2000). After Brentano: A one- sciousness. ZiF Technical Report 40, Bielﬁeld, level theory of consciousness. European Journal Germany. of Philosophy, 8, 190–209. Rosenthal, D. M. (1993). Thinking that one Tye, M. (1986). The subjective qualities of expe- thinks. In M. Davies & G. W. Humphreys rience. Mind, 95, 1–17. (Eds.), Consciousness: Psychological and philo- Tye, M. (1992). Visual qualia and visual content. sophical essays. Oxford: Blackwell. In T. Crane (ed.), The contents of experience. Rosenthal, D. M. (2000). Consciousness and New York: Cambridge University Press. metacognition. In D. Sperber (Ed.), Metarep- Tye, M. (1995). Ten problems of consciousness. resentation. Oxford: Oxford University Press. Cambridge, MA: MIT Press. Rosenthal, D. M. (2002a). Explaining conscious- Tye, M. (2000). Consciousness, color, and content. ness. In D. J. Chalmers (Ed.), Philosophy of Cambridge, MA: MIT Press. mind. Oxford: Oxford University Press. Tye, M. (2002). Visual qualia and visual content Rosenthal, D. M. (2002b). Consciousness and revisited. In D. J. Chalmers (Ed.), Philosophy of higher-order thoughts. In L. Nadel (Ed.), mind. Oxford: Oxford University Press. P1: JzG 0521857430c03 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 4:21

66 the cambridge handbook of consciousness

Van Gulick, R. (1993). Understanding the phe- Wegner, D. M. (2002). The illusion of conscious nomenal mind: Are we all just armadillos? will. Cambridge, MA: MIT Press. Reprinted in N. J. Block, O. Flanagan, & Williford, K. W.(2006). The self-representational G. Guzeldere (Eds.), (1997). The nature structure of consciousness. In U. Kriegel & of consciousness: Philosophical debates. Cam- K. Williford (Eds.), Consciousness and self- bridge, MA: MIT Press. reference. Cambridge, MA: MIT Press. Van Gulick, R. (2001). Inward and upward – Zahavi, D. (1998). Brentano and Husserl on self- reflection, introspection needed and self- awareness. Etudesˆ Phénomènologiques, 27(8), awareness. Philosophical Topics, 28, 275– 127–169. 305. Zahavi, D. (1999). Self-awareness and alterity. Van Gulick, R. (2006). Mirror mirror – is that Evanston, IL: Northwestern University Press. all? In U. Kriegel & K. Williford (Eds.), Con- Zahavi, D. (2004). Back to Brentano? Journal of sciousness and self-reference. Cambridge, MA: Consciousness Studies, 11, 66–87. MIT Press. Zahavi, D., & Parnas, J. (1998). Phenomenal con- Velmans, M. (1992). Is human information pro- sciousness and self-awareness: A phenomeno- cessing conscious? Behavioral and Brain Sci- logical critique of representational theory. Jour- ences, 14, 651–669. nal of Consciousness Studies, 5, 687–705. P1: JzG 0521857430c04 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:56

CHAPTER 4 Philosophical Issues: Phenomenology

Evan Thompson and Dan Zahavi∗

Abstract and developed by numerous other philosophers throughout the 20th century. This Current scientific research on conscious- chapter provides an introduction to this tra- ness aims to understand how conscious- dition and its way of approaching issues ness arises from the workings of the brain about consciousness. We first discuss some and body, as well as the relations between features of phenomenological methodology conscious experience and cognitive pro- and then present some of the most impor- cessing. Clearly, to make progress in these tant, influential, and enduring phenomeno- areas, researchers cannot avoid a range of logical proposals about various aspects of conceptual issues about the nature and consciousness. These aspects include inten- structure of consciousness, such as the tionality, self-awareness and the first-person following: What is the relation between perspective, time-consciousness, embodi- intentionality and consciousness? What is ment, and intersubjectivity. We also high- the relation between self-awareness and light a few ways of linking phenomenology consciousness? What is the temporal struc- and cognitive science in order to suggest ture of conscious experience? What is it some directions that consciousness research like to imagine or visualize something, and could take in the years ahead. how is this type of experience different from perception? How is bodily experience related to self-consciousness? Such issues Introduction have been addressed in detail in the philosophical tradition of phenomenology, inau- Contemporary Continental perspectives on gurated by Edmund Husserl (1859–1938) consciousness derive either whole or in part from phenomenology, the philosophical tradition inaugurated by Edmund Husserl ∗ Order of authors was set alphabetically, and each (1859–1938). This tradition stands as one author did equal work. of the dominant philosophical movements 67 P1: JzG 0521857430c04 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:56

68 the cambridge handbook of consciousness

of the last century and includes major lowing topics relevant to cognitive science 20th-century European philosophers, nota- and the philosophy of mind: intentionality, bly Martin Heidegger, Jean-Paul Sartre, and self-awareness and the first-person perspec- Maurice Merleau-Ponty, as well as impor- tive, time-consciousness, embodiment, and tant North American and Asian exponents intersubjectivity. (Moran, 2000). Considering that virtually all of the leading figures in 20th-century German and French philosophy, including Method Adorno, Gadamer, Habermas, Derrida, and Foucault, have been influenced by phe- Phenomenology grows out of the recogni- nomenology, and that phenomenology is tion that we can adopt, in our own first- both a decisive precondition and a constant person case, different mental attitudes or interlocutor for a whole range of subse- stances toward the world, life, and experi- quent theories and approaches, including ence. In everyday life we are usually straight- existentialism, hermeneutics, structuralism, forwardly immersed in various situations and deconstruction, and post-structuralism, projects, whether as specialists in one or phenomenology can be regarded as the another form of scientific, technical, or prac- cornerstone of what is often (but some- tical knowledge or as colleagues, friends, and what misleadingly) called Continental members of families and communities. In philosophy. addition to being directed toward more-or- The phenomenological tradition, like any less particular, ‘thematic’ matters, we are other philosophical tradition, spans many also overall directed toward the world as different positions and perspectives. This the unthematic horizon of all our activities point also holds true for its treatments and (Husserl, 1970,p.281). Husserl calls this atti- analyses of consciousness. Like analytic phi- tude of being straightforwardly immersed losophy, phenomenology offers not one but in the world ‘the natural attitude’, and he many accounts of consciousness. The fol- thinks it is characterized by a kind of unre- lowing discussion, therefore, is by neces- flective ‘positing’ of the world as something sity selective. Husserl’s analyses are the existing ‘out there’ more or less indepen- main reference point, and the discussion dently of us. focuses on what we believe to be some of The ‘phenomenological attitude’, on the the most important, influential, and endur- other hand, arises when we step back from ing proposals about consciousness to have the natural attitude, not to deny it, but emerged from these analyses and their sub- to investigate the very experiences it com- sequent development in the phenomenolog- prises. If such an investigation is to be gen- ical tradition.1 uinely philosophical, then it must strive to Furthermore, in recent years a new cur- be critical and not dogmatic, and therefore rent of phenomenological philosophy has cannot take the na¨ıve realism of the nat- emerged in Europe and North America, one ural attitude for granted. Yet to deny this that goes back to the source of phenomenol- realistic attitude would be equally dogmatic. ogy in Husserl’s thought, but addresses The realistic ‘positing’ of the natural atti- issues of concern to contemporary analytic tude must rather be suspended, neutral- philosophy of mind, philosophy of psy- ized, or put to one side, so that it plays chology, and cognitive science (see Peti- no role in the investigation. In this way, tot, Varela, Pachoud, & Roy, 1999, and we can focus on the experiences that sus- the new journal Phenomenology and the tain and animate the natural attitude, but Cognitive Sciences). This important current in an open and non-dogmatic manner. We of phenomenological research also informs can investigate experience in the natural atti- our discussion.2 Accordingly, after introduc- tude without being prejudiced by the natural ing some features of the phenomenological attitude’s own unexamined view of things. method of investigation, we focus on the fol- This investigation should be critical and P1: JzG 0521857430c04 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:56

philosophical issues: phenomenology 69

not dogmatic, shunning metaphysical and and thus as strict relational correlates of our scientific prejudices. It should be guided by experience. what is actually given to experience, rather The phenomenological reduction, in its than by what we expect to find given our full sense, is a rich mode of analysis compris- theoretical commitments. ing many steps. Two main ones are crucial. Yet how exactly is such an investigation The first leads back from the natural attitude to proceed? What exactly are we supposed to the phenomenological attitude by neu- to investigate? Husserl’s answer is decep- tralizing the realistic positing of the natural tively simple: Our investigation should turn attitude and then orienting attention toward its attention toward the givenness or appear- the disclosure or appearance of reality to us. ance of reality; that is, it should focus on The second leads from this phenomenolog- the way in which reality is given to us in ical attitude to a more radical kind of philo- experience. We are to attend to the world sophical attitude. Put another way, this step strictly as it appears, the world as it is phe- leads from phenomenology as an empirical, nomenally manifest. Put another way, we psychological attitude (phenomenological should attend to the modes or ways in which psychology) to phenomenology as a ‘tran- things appear to us. We thereby attend to scendental’ philosophical attitude (transcen- things strictly as correlates of our experience, dental phenomenology). and the focus of our investigation becomes ‘Transcendental’ is used here in its the correlational structure of our subjec- Kantian sense to mean an investigation tivity and the appearance or disclosure of concerned with the modes or ways in which the world.3 objects are experienced and known, as well The philosophical procedure by which as the a priori conditions for the possi- this correlational structure is investigated bility of such experience and knowledge. is known as the phenomenological reduction. Husserl casts these two aspects of transcen- ‘Reduction’ in this context does not mean dental inquiry in a specific form, which is replacing or eliminating one theory or model clearly related to but nonetheless different in favour of another taken to be more funda- from Kant’s (see Steinbock, 1995,pp.12– mental. It signifies rather a ‘leading back’ (re- 15). Two points are important here. First, ducere) or redirection of thought away from transcendental phenomenology focuses not its unreflective and unexamined immersion on what things are, but on the ways in in experience of the world to the way in which things are given. For Husserl, this which the world manifests itself to us. To means focusing on phenomena (appear- redirect our interest in this way does not ances) and the senses or meanings they mean we doubt the things before us or some- have for us and asking how these mean- how try to turn away from the world to ingful phenomena are ‘constituted’. ‘Con- look elsewhere. Things remain before us, but stitution’ does not mean fabrication or we envisage them in a new way; namely, creation; the mind does not fabricate the strictly as they appear to us. Thus, every- world. To constitute, in the technical phe- day things available to our perception are nomenological sense, means to bring to not doubted or considered as illusions when awareness, to present, or to disclose. The they are ‘phenomenologically reduced’, but mind brings things to awareness; it discloses instead are envisaged and examined simply and presents the world. Stated in a classi- and precisely as perceived (and similarly for cal phenomenological way, the idea is that remembered things as remembered, imag- objects are disclosed or made available to ined things as imagined, and so on). In other experience in the ways they are thanks words, once we adopt the phenomenologi- to how consciousness is structured. Things cal attitude, we are interested not in what show up, as it were, having the features they things are in themselves, in some na¨ıve, do, because of how they are disclosed and mind-independent, or theory-independent brought to awareness, given the structure of sense, but rather in exactly how they appear, consciousness. P1: JzG 0521857430c04 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:56

70 the cambridge handbook of consciousness

Such constitution is not apparent to us perceived, judged, or evaluated – we are led in everyday life, but requires systematic back to the intentional structures to which analysis to discern. Consider, for exam- these modes of appearance are correlated. ple, our experience of time. Our sense of We are led to the intentional acts of presen- the present moment as both simultaneously tation, perception, judgement, and evalua- opening into the immediate future and slip- tion and thereby to the subject (or subjects), ping away into the immediate past depends in relation to whom the object as appearing on the formal structure of our conscious- must necessarily be understood. ness of time. The present moment mani- Through the phenomenological attitude fests as having temporal breadth, as a zone we thus become aware of the givenness of or span of actuality, instead of as an instan- the object. Yet the aim is not simply to taneous flash, because of the way our con- focus on the object exactly as it is given, but sciousness is structured. Second, to address also on the subjective side of consciousness. this constitutional problem of how mean- We thereby become aware of our subjec- ingful phenomena are brought to awareness tive accomplishments, specifically the kinds or disclosed, transcendental phenomenology of intentionality that must be in play for any- tries to uncover the invariant formal princi- thing to appear as it does. When we inves- ples by which experience necessarily oper- tigate appearing objects in this way, we also ates in order to be constitutive. A funda- disclose ourselves as ‘datives of manifesta- mental example of this type of principle tion’ (Sokolowski, 2000), as those to whom is the ‘retentional-protentional’ structure of objects appear. time-consciousness, which we discuss in the As a procedure of working back, as it later section, Temporality and Inner Time- were, from the objects of experience, as Consciousness. given to perception, memory, imagination, The purpose of the phenomenological and so on, to the acts whereby one is aware of reduction, therefore, contrary to many mis- these objects – acts of perceiving, remember- understandings, is neither to exclude the ing, imagining, and so on – the phenomeno- world from consideration nor to com- logical reduction has to be performed in mit one to some form of methodolog- the first person. As with any such proce- ical solipsism. Rather, its purpose is to dure, it is one thing to describe its gen- enable one to explore and describe the eral theoretical character and another to spatiotemporal world as it is given. For describe it pragmatically, the concrete steps Husserl, the phenomenological reduction is by which it is carried out. The main method- meant as a way of maintaining this radi- ical step crucial for the phenomenological difference between philosophical reflec- cal reduction Husserl calls the epoché. This tion on phenomenality and other modes term comes originally from Greek skepti- of thought. cism, where it means to refrain from judge- Henceforth, we are no longer to consider ment, but Husserl adopted it as a term for the worldly object na¨ıvely; rather, we are to the ‘suspension’, ‘neutralization’, or ‘brack- focus on it precisely as a correlate of expe- eting’ of both our natural ‘positing’ attitude rience. If we restrict ourselves to that which (see above) and our theoretical beliefs or shows itself (whether in straightforward per- assertions (scientific or philosophical) about ception or a scientific experiment), and if ‘objective reality’. From a more concrete we focus specifically on that which tends and situated first-person perspective, how- to be ignored in daily life (because it is so ever, the epoche´ can be seen as a prac- familiar) – namely, on phenomenal mani- ticed mental gesture of shifting one’s atten- festation as such, the sheer appearances of tion to how the object appears and thus to things – then we cannot avoid being led one’s experiencing of the object: “Literally, back (re-ducere) to subjectivity. Insofar as we the epoche´ corresponds to a gesture of sus- are confronted with the appearance of an pension with regard to the habitual course object – that is, with an object as presented, of one’s thoughts, brought about by an P1: JzG 0521857430c04 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:56

philosophical issues: phenomenology 71

interruption of their continuous flowing ... nologists such as Heidegger and Merleau- As soon as a mental activity, a thought Ponty, who took up and then recast in anchored to the perceived object alone, their own ways the method of the phe- turns me away from the observation of the nomenological reduction (see Heidegger, perceptual act to re-engage me in the per- 1982; Merleau-Ponty, 1962). ception of the object, I bracket it” (Depraz, For this reason, one new current in phe- 1999,pp.97–98). The aim of this bracketing nomenology aims to develop more explic- is to return one’s attention to the act of expe- itly the pragmatics of the epocheasa´ riencing correlated to the object, thereby ‘first-person method’ for investigating con- sustaining the phenomenological reduction: sciousness (Depraz, 1999; Depraz, Varela, “in order that the reduction should always be & Vermersch, 2003; Varela & Shear 1999). a living act whose freshness is a function of its This pragmatic approach has also com- incessant renewal in me, and never a simple pared the epoche´ to first-person meth- and sedimented habitual state, the reflective ods in other domains, such as contem- conversion [of attention] has to be operative plative practice (Depraz et al., 2003), at every instant and at the same time per- and has explored the relevance of first- manently sustained by the radical and vig- person methods for producing more refined ilant gesture of the epoché” (Depraz, 1999, first-person reports in experimental psychol- p. 100). ogy and cognitive neuroscience (Lutz & One can discern a certain ambivalence Thompson, 2003; Varela, 1996). This latin the phenomenological tradition regard- ter endeavour is central to the research ing the theoretical and practical or exis- programme known as ‘neurophenomenol- tential dimensions of the epoche.´ On the ogy’, introduced by Francisco Varela (1996, one hand, Husserl’s great concern was to 1999) and developed by other researchers establish phenomenology as a new philo- (Lloyd, 2002, 2003; Lutz & Thompson, sophical foundation for science, and so the 2003; Rainville, 2005; Thompson, in press; epoche´ in his hands served largely as a Thompson, Lutz, & Cosmelli, 2005; see also critical tool of theoretical reason.4 On the Chapters 19 and 26). other hand, because Husserl’s theoretical project was based on a radical reappraisal of experience as the source of meaning Intentionality and knowledge, it necessitated a constant return to the patient, analytic description of Implicit in the foregoing treatment of phe- lived experience through the phenomeno- nomenological method is the phenomenological reduction. This impulse generated logical concept of intentionality. According a huge corpus of careful phenomenologi- to Husserlian phenomenology, conscious- cal analyses of numerous different dimen- ness is intentional, in the sense that it ‘aims sions and aspects of human experience – toward’ or ‘intends’ something beyond itself. the perceptual experience of space (Husserl, This sense of intentional should not be con- 1997), kinesthesis and the experience of fused with the more familiar sense of hav- one’s own body (Husserl, 1989, 1997), ing a purpose in mind when one acts, which time-consciousness (Husserl, 1991), affect is only one kind of intentionality in the (Husserl, 2001), judgement (Husserl, 1975), phenomenological sense. Rather, intention- imagination and memory (Husserl, 2006), ality is a generic term for the pointing- and intersubjectivity (Husserl, 1973), to beyond-itself proper to consciousness (from name just a few. Nevertheless, the epoche´ the Latin intendere, which once referred to as a practical procedure – as a situated prac- drawing a bow and aiming at a target). tice carried out in the first person by the Phenomenologists distinguish different phenomenologist – has remained strangely types of intentionality. In a narrow sense, neglected in the phenomenological litera- intentionality is defined as object-direct- ture, even by so-called existential phenome- edness. In a broader sense, which covers P1: JzG 0521857430c04 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:56

72 the cambridge handbook of consciousness

what Husserl (2001,p.206) and Merleau- sense or meaning, that mediates the inten- Ponty (1962, p. xviii) called ‘operative tional relation between the mental act and intentionality’ (see below), intentionality the object. On this view, consciousness is is deﬁned as openness toward otherness directed toward the object by means of the (or alterity). In both cases, the empha- noema and thus achieves its openness to sis is on denying that consciousness is the world only in virtue of the represen- self-enclosed. tational noema. Object-directedness characterizes almost According to the rival non-represen- all of our experiences, in the sense that in tationalist interpretation, the noema is not having them we are exactly conscious of any intermediate, representational entity; something. We do not merely love, fear, see, the noema is the object itself, but the object or judge; we love, fear, see, or judge some- considered phenomenologically; that is, pre- thing. Regardless of whether we consider a cisely as experienced. In other words, perception, a thought, a judgement, a fan- the object-as-intended is the object-that-is- tasy, a doubt, an expectation, a recollection, intended, abstractly and phenomenologically and so on, these diverse forms of conscious- considered; namely, in abstraction from ness are all characterized by the intending the realistic positing of the natural atti- of an object. In other words, they cannot tude and strictly as experientially given. be analyzed properly without a look at their The noema is thus graspable only in a objective correlates; that is, the perceived, phenomenological or transcendental atti- the doubted, the expected, and so forth. The tude. This view rejects the representation- converse is also true: The intentional object alism of the former view. Consciousness cannot be analyzed properly without a look is intrinsically self-transcending and accord- at its subjective correlate, the intentional act. ingly does not achieve reference to the world Neither the intentional object nor the men- in virtue of intermediate ideal entities that tal act that intends it can be understood apart bestow intentionality upon it. Experiences from the other. are intrinsically intentional (see Searle, 1983, Phenomenologists call this act-object for a comparable claim in the analytic tra- relation the ‘correlational structure of dition). Their being is constituted by being intentionality’. ‘Correlational’ does not of something else. It would take us too far mean the constant conjunction of two terms aﬁeld to review the twists and turns of this that could be imagined to exist apart, but debate, so we simply state for the record that the necessary structural relation of mental for a variety of reasons we think the rep- act and intended object. Object-directed resentationalist interpretation of the noema intentional experiences necessarily comprise is mistaken and the non-representationalist these two inseparable poles. In Husserlian interpretation is correct (see Zahavi, 2003, phenomenological language these two poles 2004). are known as the ‘noema’ (the object as We have been considering object-directed experienced) and the ‘noesis’ (the mental act intentionality, but many experiences are that intends the object). not object-directed – for example, feelings There has been a huge amount of schol- of pain and nausea, and moods, such as arly discussion about the proper way to inter- anxiety, depression, and boredom. Philoso- pret the Husserlian notion of the noema (see phers whose conception of intentionality Drummond, 2003, for an overview). The is limited to object-directedness deny that discussion concerns the relation between such experiences are intentional (e.g., Searle, the object-as-intended (the noema) and the 1983). Phenomenologists, however, in dis- object-that-is-intended (the object itself )– the tinguishing between intentionality as object- wine bottle-as-perceived (as felt and seen) directedness and intentionality as openness, and the bottle itself. According to the rep- have a broader conception. It is true that resentationalist interpretation, the noema is pervasive moods, such as sadness, boredom, a type of representational entity, an ideal nostalgia, and anxiety, must be distinguished P1: JzG 0521857430c04 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:56

philosophical issues: phenomenology 73

from intentional feelings, such as the desire The object can be given more or less directly; for an apple or the admiration for a particular that is, it can be more or less present. One person. Nevertheless, moods are not without can also speak of different epistemic lev- a reference to the world. They do not enclose els. The lowest and most empty way in us within ourselves, but are lived through as which the object can appear is in the sig- pervasive atmospheres that deeply influence nitive acts. These (linguistic) acts certainly the way the world is disclosed to us. Moods, have a reference, but apart from that the such as curiosity, nervousness, or happiness, object is not given in any fleshed-out man- disclose our embeddedness in the world ner. The pictorial acts have a certain intu- and articulate or modify our existential itive content, but like the signitive acts, they possibilities. As Heidegger argued, moods, intend the object indirectly. Whereas the sig- rather than being merely attendant phenom- nitive acts intend the object via a contingent ena, are fundamental forms of disclosure: representation (a linguistic sign), the picto- “Mood has always already disclosed being- rial acts intend the object via a representa- in-the-world as a whole and first makes postion (picture) that bears a certain similar- sible directing oneself toward something” ity or projective relation to the object. It is (Heidegger, 1996,p.129). only the perception that gives us the object What about pain? Sartre’s classic analy- directly. This is the only type of intention sis in Being and Nothingness (1956) is illu- that presents the object in its bodily presence minating in this case. Imagine that you are (leibhaftig). sitting late a night trying to finish reading a Recollection and imagination are two book. You have been reading most of the day other important forms of object-directed and your eyes hurt. How does this pain orig- intentionality (empathy is a third: see the inally manifest itself? According to Sartre, section, Intersubjectivity). These types are not initially as a thematic object of reflec- mediated; that is, they intend their objects tion, but by influencing the way in which by way of other, intermediate mental activ- you perceive the world. You might become ities, rather than directly, as does percep- restless and irritated and have difficulties tion. In recollection, I remember the within focusing and concentrating. The words ering oak (the object itself ) by means of on the page might tremble and quiver. The re-presenting (reproducing or re-enacting) pain is not yet apprehended as an inten- a past perception of the oak. In imagina- tional object, but that does not mean that tion, I can either imagine the withering oak it is either cognitively absent or uncon- (the actual tree), or I can imagine a non- scious. It is not yet reflected-upon as a psy- existent oak in the sense of freely fantasizing chic object, but given rather as a vision- a different world. Either way, imagination in-pain, as an affective atmosphere that involves re-presenting to myself a possi- influences your intentional interaction with ble perceptual experience of the oak. Yet, the world. in imagination, the assertoric or ‘positing’ Another important part of the phe- character of this (re-presented) perceptual nomenological account of intentionality is experience is said to be ‘neutralized’, for the distinction among signitive (linguis- whereas an ordinary perceptual experience tic), pictorial, and perceptual intentionali- posits its object as actually there (regard- ties (Husserl, 2000). I can talk about a with- less of whether the experience is veridi- ering oak, I can see a detailed drawing of the cal), imagination does not. In recollection, oak, and I can perceive the oak myself. These by contrast, this assertoric or positing fea- different ways to intend an object are not ture of the experience is not neutralized, unrelated. According to Husserl, there is a but remains in play, because the percep- strict hierarchical relation among them, in tion reproduced in the memory is repre- the sense that they can be ranked accord- sented as having actually occurred in the ing to their ability to give us the object as past. Husserl thus describes perception and directly, originally, and optimally as possible. recollection as positional (assertoric) acts, P1: JzG 0521857430c04 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:56

74 the cambridge handbook of consciousness

whereas imagination is non-positional (non- intentional activity; it consists in responding assertoric; see Husserl, 2006, and Bernet, to or paying attention to that which is affect- Kern, & Marbach, 1993, for an overview). ing us passively. Even receptivity understood We thus arrive at another crucial dis- as a mere ‘I notice’ presupposes a prior tinction, the distinction between intentional ‘affection’ (meaning one’s being affectively acts of presentation (Gegenwärtigung) and of influenced or perturbed, not a feeling of fond- re-presentation (Vergegenwärtigung). Accord- ness). Whatever becomes thematized (even ing to standard usage in the analytic phi- as a mere noticing) must have been already losophy of mind and cognitive science, the affecting and stimulating one in an unheeded term ‘representation’ applies to any kind manner. Affectivity, however, is not a mat- of mental state that has intentional content ter of being affected by an isolated, undif- (‘intentional content’ and ‘representational ferentiated sense impression. If something is content’ being used synonymously). In phe- to affect us, impose itself on us, and arouse nomenological parlance, on the other hand, our attention, it must be sufficiently strong. ‘re-presentation’ applies only to those types It must be more conspicuous than its sur- of mental acts that refer to their objects roundings, and it must stand out in some way by way of intermediate mental activity, through contrast, heterogeneity, and differ- as in remembrance, imagination (imaging ence. Thus, receptivity emerges from within and fantasy), and pictorial consciousness a passively organized and structured field of (looking at a picture). Perception, by con- affectivity.5 trast, is not re-presentational, but presen- In summary, explicit, object-directed tational, because the object-as-experienced intentional experience arises against the (the intentional object or objective correlate background of a precognitive, operative of the act) is ‘bodily present’ or there ‘in flesh intentionality, which involves a dynamic and blood’ (regardless of whether the per- interplay of affectivity and receptivity and ceptual experience turns out to be veridical constitutes our most fundamental way of or not). being open to the world. Perceptual intentionality can be further differentiated into, on the one hand, a thematic, explicit, or focal object-directed Phenomenal Consciousness mode of consciousness and, on the other and Self-Awareness hand, a non-reflective tacit sensibility, which constitutes our primary openness to the In contemporary philosophy of mind the world. This tacit sensibility, called ‘opera- term ‘phenomenal consciousness’ refers to tive [fungierende] intentionality’, functions mental states that have a subjective and prereflectively, anonymously, and passively, experiential character. In Nagel’s words, without being engaged in any explicit cog- for a mental state to be (phenomenally) nitive acquisition. In this context it is conscious is for there to be something important to distinguish between activity it is like for the subject to be in that and passivity. One can be actively taking a state (Nagel, 1979). Various notions are position in acts of comparing, differentiat- employed to describe the properties char- ing, judging, valuing, wishing, and so on. As acteristic of phenomenal consciousness – Husserl (2001) points out, however, when- qualia, sensational properties, phenomenal ever one is active, one is also passive, because properties, and the subjective character to be active is to react to something that has of experience – and there is considerable affected one. Every kind of active position- debate about the relation between these taking presupposes a prior and passive properties and other properties of mental being affected. states, such as their representational con- Following Husserl one step further in his tent or their being cognitively accessible analysis, we can distinguish between receptiv- to thought and verbal report (‘access con- ity and affectivity. Receptivity is taken to be sciousness’). The examples used in these the first, lowest, and most primitive type of discussions are usually bodily sensations, P1: JzG 0521857430c04 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:56

philosophical issues: phenomenology 75

such as pain, or perceptual experiences, est of the German idealists, as there is an such as the visual experience of colour. experiential difference between expecting Much less frequently does one find discus- and doubting that Denmark will win the sion of the subjective character of emo- 2010 FIFA World Cup. What it is like to be tion (feelings, affective valences, moods), in one type of intentional state is different to say nothing of memory, mental imagery, from what it is like to be in another type of or thought. intentional state. Similarly, each of the dif- According to Husserl, however, the phe- ferent intentional matters has its own phenomenal aspect of experience is not lim- nomenal character. To put it differently, a ited to sensory or even emotional states, change in the intentional matter will entail but also characterizes conscious thought. a change in what it is like to undergo the In his Logical Investigations, Husserl (2000) experience in question. (This does not entail, argues that conscious thoughts have expe- however, that two experiences differing in riential qualities and that episodes of con- what it is like to undergo them cannot intend scious thought are experiential episodes. the same object, nor that two experiences Every intentional experience possesses two alike in this respect must necessarily intend different, but inseparable ‘moments’ (i.e., the same object.) These experiential differ- dependent aspects or ingredients): (i) Every ences, Husserl argues, are not simply sensory intentional experience is an experience of a differences.6 specific type, be it an experience of judging, In summary, every phenomenally con- hoping, desiring, regretting, remembering, scious state, be it a perception, an emotion, a affirming, doubting, wondering, fearing, and recollection, an abstract belief, and so forth, so on. Husserl calls this aspect the intentional has a certain subjective character, a certain quality of the experience. (ii) Every inten- phenomenal quality, corresponding to what tional experience is also directed at or about it is like to live through or undergo that state. something. He calls this aspect the inten- This is what makes the mental state in ques- tional matter of the experience. Clearly, the tion phenomenally conscious. same quality can be combined with different This experiential quality of conscious matters, and the same matter can be com- mental states, however, calls for further elu- bined with different qualities. It is possible cidation. Let us take perceptual experience to doubt that ‘the inflation will continue’, as our starting point. Whereas the object doubt that ‘the election was fair’, or doubt of my perceptual experience is intersubjec- that ‘one’s next book will be an international tively (publicly) accessible, in the sense that bestseller’, precisely as it is possible to deny it can in principle be given to others in the that ‘the lily is white’, to judge that ‘the lily same way it is given to me, the case is dif- is white’, or to question whether ‘the lily ferent with my perceptual experience itself. is white’. Husserl’s distinction between the Whereas you and I can both perceive one intentional matter and the intentional qual- and the same cherry, each of us has his or her ity thus bears a certain resemblance to the own distinct perception of it, and we can- contemporary distinction between proposi- not share these perceptions, precisely as we tional content and propositional attitudes cannot share each other’s pains. You might (though it is important to emphasize that certainly realize that I am in pain and even Husserl by no means took all intentional empathize with me, but you cannot actu- experiences to be propositional in nature; ally feel the pain the same way I do. This see Husserl, 1975). point can be formulated more precisely by Nevertheless – and this is the central saying that you have no access to the first- point – Husserl considered these cognitive personal givenness of my experience. This differences to be also experiential differences. first-personal quality of experience leads to Each of the different intentional qualities the issue of self and self-awareness. has its own phenomenal character. There is When one is directly and non-inferent- an experiential difference between affirm- ially conscious of one’s own occurrent thou- ing and denying that Hegel was the great- ghts, perceptions, feelings, or pains, they are P1: JzG 0521857430c04 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:56

76 the cambridge handbook of consciousness

characterized by a first-personal givenness In summary, any (object-directed) con- that immediately reveals them as one’s own. scious experience, in addition to being of This first-personal givenness of experiential or about its intentional object is prereflec- phenomena is not something incidental to tively manifest to itself. To use another for- their being, a mere varnish the experiences mulation, transitive phenomenal conscious- could lack without ceasing to be experiences. ness (consciousness-of) is also intransitive On the contrary, it is their first-personal self-consciousness (see Chapter 3). Intran- givenness that makes the experiences subjec- sitive self-consciousness is a primitive form tive. To put it differently, their first-personal of self-consciousness in the sense that (i) givenness entails a built-in self-reference, it does not require any subsequent act of a primitive experiential self-referentiality. reflection or introspection, but occurs simul- When I am aware of an occurrent pain, per- taneously with awareness of the object; (ii) ception, or thought from the first-person does not consist in forming a belief or mak- perspective, the experience in question is ing a judgement; and (iii) is passive in the given immediately and non-inferentially as sense of being spontaneous and involun- mine. I do not first scrutinize a specific per- tary. According to some phenomenologists ception or feeling of pain and then identify (e.g., Merleau-Ponty, 1962), this tacit self- it as mine. awareness involves a form of non-objective Accordingly, self-awareness cannot be bodily self-awareness, an awareness of one’s equated with reflective (explicit, thematic, lived body (Leib) or embodied subjectiv- introspective) self-awareness, as claimed by ity, correlative to experience of the object some philosophers and cognitive scientists. (see the section, Embodiment and Percep- On the contrary, reflective self-awareness tion). The roots of such prereflective bod- presupposes a prereflective (implicit, tacit) ily self-awareness sink to the passive and self-awareness. Self-awareness is not some- anonymous level of the interplay between thing that comes about only at the moment receptivity and affectivity constitutive of I realize that I am (say) perceiving the ‘operative intentionality’ (see the section, Empire State Building, or realize that I am Intentionality). the bearer of private mental states, or refer Phenomenology thus corroborates cer- to myself using the first-person pronoun. tain proposals about consciousness com- Rather, it is legitimate to speak of a primitive ing from neuroscience. Such theorists as but basic type of self-awareness whenever I Panksepp (1998a, b) and Damasio (1999) am acquainted with an experience from a have argued that neuroscience needs to first-person perspective. If the experience in explain both how the brain enables us to question, be it a feeling of joy, a burning experience the world outside us and how thirst, or a perception of a sunset, is given it “also creates a sense of self in the act of in a first-personal mode of presentation to knowing . . . how each of us has a sense of me, it is (at least tacitly) given as my expe- ‘me’” (Parvizi & Damasio, 2001,pp.136– rience and can therefore count as a case of 137). In phenomenological terms, this sec- self-awareness. To be aware of oneself is con- ond issue concerns the primitive sense of ‘I- sequently not to apprehend a pure self apart ness’ belonging to consciousness, known as from the experience, but to be acquainted ‘ipseity’ (see also Chapter 19). As a num- with an experience in its first-personal mode ber of cognitive scientists have emphasized, of presentation; that is, from ‘within’. Thus, this core of self-awareness in conscious- the subject or self referred to is not some- ness is fundamentally linked to bodily pro- thing standing opposed to, apart from, or cesses of life-regulation, emotion, and affect, beyond experience, but is rather a feature or such that cognition and intentional action function of its givenness. Or to phrase it differ- are emotive (Damasio, 1999; Freeman, ently, it is this first-personal givenness of the 2000; Panksepp 1998a,b). A promising line experience that constitutes the most basic of collaboration between phenomenology form of selfhood (Zahavi, 1999, 2005). and affective-cognitive neuroscience could P1: JzG 0521857430c04 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:56

philosophical issues: phenomenology 77

therefore centre on the lived body as a integration is temporal in nature. Thus, time- way of deepening our understanding of consciousness must be regarded as a formal subjectivity and consciousness (Thompson, condition of possibility for the perception of 2007). any object. Yet, how must this experiential process be structured for identity or unity over time to be possible? Temporality and Inner Phenomenological analyses point to the Time-Consciousness ‘width’ or ‘depth’ of the ‘living present’ of consciousness: Our experience of temporal Why must an investigation of conscious- enduring objects and events, as well as our ness inevitably confront the issue of time? experience of change and succession, would There are many reasons, of course, but in be impossible were we conscious only of that this section we focus on two main ones. which is given in a punctual now and were First, experiences do not occur in isola- our stream of consciousness composed of a tion. The stream of consciousness comprises series of isolated now-points, like a string an ensemble of experiences that are uni- of pearls. According to Husserl (1991), the fied both at any given time (synchronically) basic unit of temporality is not a ‘knife- and over time (diachronically); therefore, we edge’ present, but a ‘duration-block’ (to bor- need to account for this temporal unity and row William James’s words; see James, 1981, continuity. In addition, we are able not only p. 574) i.e., a temporal field that comprises to recollect earlier experiences and recog- all three temporal modes of present, past, nize them as our own but also to perceive and future. Just as there is no spatial object enduring (i.e., temporally extended) objects without a background, there is no experi- and events; hence, we need to account for ence without a temporal horizon. We can- how consciousness must be structured for not experience anything except on the back- there to be such experiences of coherence ground of what it succeeds and what we and identity over time. Second, our present anticipate will succeed it. We can no more cognitive activities are shaped and influ- conceive of an experience empty of future enced conjointly by both our past experi- than one empty of past. ences and our future plans and expectations. Three technical terms describe this tem- Thus, if we are to do justice to the dynamic poral form of consciousness. There is (i) character of cognition, we cannot ignore the a ‘primal impression’ narrowly directed role of time. toward the now-phase of the object. The In a phenomenological context, the term primal impression never appears in isola- ‘temporality’ does not refer to objective, cos- tion and is an abstract component that by mic time, measured by an atomic clock, or itself cannot provide us with a perception to a merely subjective sense of the passage of a temporal object. The primal impres- of time, although it is intimately related sion is accompanied by (ii) a ‘retention’, to the latter. Temporality, or ‘inner time- which provides us with a consciousness of consciousness’, refers to the most fundamen- the just-elapsed phase of the object, and tal, formal structure of the stream of con- by (iii) a ‘protention’, which in a more-or- sciousness (Husserl, 1991). less indefinite way intends the phase of the To introduce this idea, we can consider object about to occur. The role of the proten- what phenomenologists call ‘syntheses of tion is evident in our implicit and unreflec- identity’ in the flow of experience. If I move tive anticipation of what is about to happen around a tree to gain a fuller appreciation as experience progresses. That such antici- of it, then the tree’s different profiles – its pation belongs to experience is illustrated front, sides, and back – do not appear as by the fact that we would be surprised if disjointed fragments, but as integrated fea- (say) the wax figure suddenly moved or if tures belonging to one and the same tree. the door we opened hid a stone wall. It The synthesis that is a precondition for this makes sense to speak of surprise only in light P1: JzG 0521857430c04 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:56

78 the cambridge handbook of consciousness

of anticipation, and because we can always and cognitive science. Work by Francisco be surprised, we always have a horizon of Varela (1999) in particular has shown anticipation. The concrete and full struc- that phenomenological analyses of time- ture of all lived experience is thus primal consciousness can be profitably linked to impression-retention-protention. Although the neurodynamical accounts of the brain pro- specific experiential contents of this struc- cesses associated with the temporal flow of ture from moment to moment progressively conscious experience (see also Chapter 26 change, at any given moment this threefold and Thompson, 2007). This linkage between structure is present (synchronically) as a uni- phenomenology and neurodynamics is cen- fied whole. This analysis provides an account tral to the research programme of neurophe- of the notion of the specious present that nomenology, mentioned above. improves on that found in William James, C. D. Broad, and others (see Gallagher, 1998). Embodiment and Perception It is important to distinguish retention and protention, which are structural features Conscious experience involves one’s body. of any conscious act, from recollection and Yet what exactly is the relationship between expectation, understood as specific types the two? It is obvious that we can perceive of mental acts. There is a clear difference our own body by (say) visually inspecting between, on the one hand, retaining notes our hands. It is less obvious that our bod- that have just sounded and protending notes ily being constitutes our subjectivity and the about to sound while listening to a melody, correlative modes or ways in which objects and, on the other hand, remembering a past are given to us. holiday or looking forward to the next vaca- The phenomenological approach to the tion. Whereas recollection and expectation role of the body in its constitution of sub- presuppose the work of retention and projective life is closely linked to the analysis tention, protention and retention are intrin- of perception. Two basic points about per- sic components of any occurrent experi- ception are important here: (i) the inten- ence one might have. Unlike recollection tional objects of perceptual experience are and expectation, they are passive (involun- public spatiotemporal objects (not private tary) and automatic processes that take place mental images or sense-data); and (ii) such without our active or deliberate contribu- objects are always given only partially to per- tion. Finally, they are invariant structural fea- ception and can never present themselves tures that make possible the temporal flow in their totality. On the one hand, percep- of consciousness as we know and experience tion purports to give us experience of public it. In other words, they are a priori condi- things, not private mental images. On the tions of possibility of there being ‘syntheses other hand, whatever we perceive is always of identity’ in experience at all. perceived in certain ways and from a cer- Husserl’s analysis of the structure of inner tain perspective. We see things, for instance, time-consciousness serves a double purpose. as having various spatial forms and visible It is meant to explain not only how we can qualities (lightness, colour, etc.), and we are be aware of objects with temporal extension able to distinguish between constancy and but also how we can be aware of our own variation in appearance (the grass looks uni- stream of experiences. To put it differently, formly green, but the shaded part looks dark, Husserl’s investigation is meant to explain whereas the part in direct sunlight looks not only how we can be aware of temporally light). We see only one profile of a thing at extended units but also how consciousness any given moment, yet we do not see things unifies itself across time. as mere façades, for we are aware of the pres- Like bodily self-awareness, temporality ence of the other sides we do not see directly. and time-consciousness are rich in potential We do not perceive things in isolation; we for collaborative study by phenomenology see them in contexts or situations, in which P1: JzG 0521857430c04 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:56

philosophical issues: phenomenology 79

they relate to and depend on each other and flight, the bird is given in conjunction with on dimensions of the environment in multi- a sensing of my eye and head movements; farious ways. when I touch the computer keys, the keys These invariant characteristics of per- are given in conjunction with a sensing of ception presuppose what phenomenologists my finger movements. Husserl’s 1907 lec- call the lived body (Leib). Things are percep- tures on Thing and Space (Husserl, 1997) dis- tually situated in virtue of the orientation cuss how this relation between perception they have to our perceiving and moving and kinaesthesis (including proprioception) bodies. To listen to a string quartet by is important for the constitution of objects Schubert is to enjoy it from a certain per- and space. To perceive an object from a cer- spective and standpoint, be it from the tain perspective is to be aware (tacitly or street, in the gallery, or in the first row. If prereflectively) that there are other coexist- something appears perspectivally, then the ing but absent profiles of the object. These subject to whom it appears must be spa- absent profiles stand in certain sensorimo- tially related to it. To be spatially related tor relations to the present profile: They to something requires that one be embod- can be made present if one carries out cer- ied. To say that we perceive only one pro- tain movements. In other words, the profiles file of something while being aware of other are correlated with kinaesthetic systems of possible profiles means that any profile we possible bodily movements and positions. If perceive points beyond itself to further pos- one moves this way, then that aspect of the sible profiles. Yet this reference of a given object becomes visible; if one moves that profile beyond itself is equally a reference way, then this aspect becomes visible. to our ability to exchange this profile for In Husserl’s terminology, every perspec- another through our own free movement tival appearance is kinaesthetically moti- (tilting our head, manipulating an object in vated. In the simple case of a motionless our hands, walking around something, etc.). object, for instance, if the kinaesthetic expe- Co-given with any profile and through any rience (K1) remains constant, then the per- sequence of profiles is one’s lived body as ceptual appearance (A1) remains constant. the ‘zero point’ or absolute indexical ‘here’, If the kinaesthetic experience changes (K1 in relation to which any appearing object is becomes K2 ), then the perceptual appear- oriented. One’s lived body is not co-given ance changes in correlation with it (A1 as an intentional object, however, but as becomes A2 ). There is thus an interde- an implicit and practical ‘I can’ of move- pendency between kinaesthetic experiences ment and perception. We thus rejoin the and perceptual appearances: A given appear- point made earlier (in the section, Phenom- ance (A1) is not always correlated with the enality and Self-Awareness) that any object- same kinaesthetic experience (e.g., K1), but directed (transitive) intentional experience it must be correlated with some kinaes- involves a non-object-directed (intransitive) thetic experience or other. Turning now to self-awareness, here an intransitive bod- the case of perceptual space, Husserl argues ily self-awareness. In short, every object- that different kinaesthetic systems of the experience carries with it a tacit form of self- body imply different perceptual achieve- experience. ments with regard to the constitution of The role of bodily self-experience in space. One needs to distinguish among the perception can be phenomenologically oculomotor systems of one eye alone and of described in much greater detail. One two eyes together, the cephalomotor system important topic is the role it plays in the of head movements, and the system of the constitution (i.e., the bringing to awareness whole body as it moves towards, away from, or disclosure) of both objects and space for and around things. These kinaesthetic sys- perception. Perspectival appearances of the tems are hierarchically ordered in relation object bear a certain relation to kinaesthetic to the visual field: The cephalomotor visual situations of the body. When I watch a bird in field contains a continuous multiplicity of P1: JzG 0521857430c04 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:56

80 the cambridge handbook of consciousness

oculomotor fields; the egocentric field of et al., 1997, for an exception). For exam- the body as a whole contains a contin- ple, neuropsychologists Milner and Goodale uous multiplicity of cephalomotor fields. write in their influential book, The Visual This hierarchy also reflects a progressive Brain in Action: “For most investigators, disclosure of visual space: The eyes alone the study of vision is seen as an enter- give only a two-dimensional continuum; prise that can be conducted without any head movements expand the field into a reference whatsoever to the relationship spherical plane at a fixed distance (like a between visual inputs and motor outputs. planetarium); and movement of the body This research tradition stems directly from as a whole introduces distance, depth, and phenomenological intuitions that regard three-dimensional structure. It is the linkage vision purely as a perceptual phenomenon” between the kinaesthetic system of whole- (Milner & Goodale 1995,p.13). It can be body movements (approaching, retreating, seen from our discussion in this section, and circling) and the corresponding per- however, that it is important to distinguish ceptual manifold of profiles or perspectival between uncritical common-sensical intu- appearances that fully discloses the three- itions and the critical examination of percep- dimensional space of visual perception. tual experience found in the phenomeno- Insofar as the body functions as the zero- logical tradition. The intuitions Milner and point for perception and action (i.e., consid- Goodale target do not belong to phe- ered in its function as a bodily first-person nomenology. On the contrary, Husserl’s perspective), the body recedes from experi- and Merleau-Ponty’s analyses of the rela- ence in favour of the world. My body sup- tion between perception and kinaesthesis plies me with my perspective on the world, clearly indicate that perception is also a and thus is first and foremost not an object on motor phenomenon. Indeed, these analy- which I have a perspective. In other words, ses anticipate the so-called dynamic sen- bodily awareness in perception is not in the sorimotor approach to perception (Hurley, first instance a type of object-consciousness, 1998; Hurley & Noe,¨ 2003;Noe,¨ 2004; but a type of non-transitive self-awareness O’Regan & Noe,¨ 2001). Rather than look- (see the section on Phenomenal and Self- ing to the intrinsic properties of neural awareness and Chapter 3). Although one can activity to understand perceptual experi- certainly experience one’s body as an object ence, this approach looks to the dynamic (e.g., in a mirror), bodily self-awareness is sensorimotor relations among neural activ- more fundamentally an experience of one’s ity, the body, and the world. This approach body as a unified subjective field of percep- has so far focused mainly on the phe- tion and action. A full account of bodily nomenal qualities of perceptual experience, experience thus reveals the body’s double or but has yet to tackle the perceptual con- ambiguous character as both a subjectively stitution of space, intransitive bodily self- lived body (Leib) and a physical (spatiotem- awareness, or the relationship between per- poral) objective body (Körper). ception and affectively motivated attention, The phenomenological analyses of embo- all long-standing topics in phenomenology. diment and perception summarized in this Further development of the dynamic sen- section are relevant to current trends in sorimotor approach might therefore bene- cognitive science. In recent years cog- fit from the integration of phenomenological nitive scientists have increasingly chal- analyses of embodiment and perception (see lenged the classical view that perception Thompson, 2007). and action are separate systems. Although phenomenologists have long emphasized the constitutive role of motor action Intersubjectivity in perceptual experience, cognitive scientists often seem unaware of this impor- For many philosophers, the issue of inter- tant body of research (but see Rizzolatti subjectivity is equated with the ‘problem of P1: JzG 0521857430c04 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:56

philosophical issues: phenomenology 81

other minds’: How can one know the men- that relate mental states to each other and tal states of others, or even that there are any to observable behaviour. According to the other minds at all (see Dancy, 1985,pp.67– simulation-theory, on the other hand, ‘mind- 68)? One classical attempt to deal with this reading’ depends not on the possession of a problem takes the form of trying to justify tacit psychological theory, but on the ability our belief in other minds on the basis of the to mentally ‘simulate’ another person – to following argument from analogy: The only use the resources of one’s own mind to cre- mind I have direct access to is my own. My ate a model of another person and thereby access to the mind of another is always medi- identify with that person, projecting oneself ated by my perception of the other’s bodily imaginatively into his or her situation (Gold- movements, which I interpret as intentional man, 2000). In either case, intersubjectivity behaviour (i.e., as behaviour resulting from is conceptualized as a cognitively medi- internal mental states). But what justifies me ated relation between two otherwise iso- in this interpretation? How can the percep- lated subjects. Both theories take intersub- tion of another person’s bodily movements jective understanding to be a matter of how provide me with information about his one represents unobservable, inner mental mind, such that I am justified in viewing states on the basis of outward behaviour his movements as intentional behaviour? In (what they disagree about is the nature of the my own case, I can observe that I have representations involved). Thus, both theo- experiences when my body is causally influ- ries foster a conception of the mental as an enced and that these experiences frequently inner realm essentially different from out- bring about certain actions. I observe that ward behaviour. other bodies are influenced and act in sim- Phenomenologists do not frame the issue ilar manners, and I therefore infer by anal- of intersubjectivity in this way, for they ogy that the behaviour of foreign bodies is reject the presuppositions built into the associated with experiences similar to those problem of other minds (see Zahavi, 2001a, I have myself. Although this inference does 2001b). Two presuppositions in particular not provide me with indubitable knowledge are called into question. The first is that about others, it gives me reason to believe in one’s own mind is given to one as a soli- their existence and to interpret their bodily tary and internal consciousness. The prob- movements as meaningful behaviour. lem with this assumption is that our ini- This way of conceptualizing self and other tial self-acquaintance is not with a purely can also be discerned, to varying degrees, in internal, mental self, for we are embodied certain approaches to social cognition in cog- and experience our own exteriority, includ- nitive science. Thus, both certain versions ing our bodily presence to the other. The of the ‘theory-theory’ (e.g., Gopnik, 1993) second assumption is that, in perceiving the and the ‘simulation-theory’ (e.g., Goldman, other, all we ever have direct access to is 2000) have crucial features in common the other’s bodily movements. The prob- with the traditional argument from anal- lem with this assumption is that what we ogy. According to the theory-theory, nor- directly perceive is intentional or meaningful mal human adults possess a common-sense behaviour – expression, gesture, and action – or folk-psychological ‘theory of mind’ that not mere physical movement that gets inter- they employ to explain and predict human preted as intentional action as a result of behaviour. Advocates of the theory-theory inference. Thus, on the one hand, one’s own consider this folk-psychological body of subjectivity is not disclosed to oneself as knowledge to be basically equivalent to a a purely internal phenomenon, and on the scientific theory: Mental states are unob- other hand, the other’s body is not dis- servable entities (like quarks), and our attri- closed as a purely external phenomenon. bution of them to each other involves Put another way, both the traditional prob- causal-explanatory generalizations (compa- lem of other minds and certain cognitive- rable in form to those of empirical science) scientific conceptions of ‘mind-reading’ rest P1: JzG 0521857430c04 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:56

82 the cambridge handbook of consciousness

on a deeply problematic conception of the their bodily behaviour and then inferring or mind as essentially inner, the body as essen- hypothesizing that their behaviour is caused tially outer, and intentional behaviour as aris- by experiences or inner mental states sim- ing from a purely contingent and causal con- ilar to those that apparently cause similar nection between these two spheres. behaviour in us. Rather, in empathy, we Phenomenological treatments of inter- experience the other directly as a person, as subjectivity start from the recognition that, an intentional being whose bodily gestures in the encounter with the other, one is faced and actions are expressive of his or her expe- neither with a mere body nor a hidden psy- riences or states of mind (for further discus- che, but with a unified whole. This unified sion, see Thompson, 2001, 2005, 2007). whole is constituted by the expressive relation Phenomenological investigations of inter- between mental states and behaviour, a rela- subjectivity go beyond intentional analy- tion that is stronger than that of a mere con- ses of empathy, however, in a variety of tingent, causal connection, but weaker than ways (see Zahavi, 2001b). Another approach that of identity (for clearly not every men- acknowledges the existence of empathy, but tal state need be overtly expressed). In other insists that our ability to encounter oth- words, expression must be more than simply ers cannot simply be taken as a brute fact. a bridge supposed to close the gap between Rather, it is conditioned by a form of alter- inner mental states and external bodily ity (otherness) internal to the embodied self. behaviour; it must be a direct manifesta- When my left hand touches my right, or tion of the subjective life of the mind (see when I perceive another part of my body, Merleau-Ponty, 1962, Part One, Chapter 6). I experience myself in a manner that antici- Thus, one aspect of the phenomenologi- pates both the way in which an other would cal problem of intersubjectivity is to under- experience me and the way in which I stand how such manifestation is possible. would experience an other. My bodily self- Phenomenologists insist that we need to exploration thus permits me to confront begin from the recognition that the body of my own exteriority. According to Husserl the other presents itself as radically different (1989), this experience is a crucial pre- from any other physical entity, and accord- condition for empathy: It is precisely the ingly that our perception of the other’s bod- unique subject-object status of the body, ily presence is unlike our perception of phys- the remarkable interplay between ipseity ical things. The other is given in its bodily (I-ness) and alterity characterizing body- presence as a lived body according to a dis- awareness, that provides me with the means tinctive mode of consciousness called empa- of recognizing other embodied subjects. thy (see Husserl, 1989; Stein, 1989). Empa- Still another line of analysis goes one thy is a unique form of intentionality, in step further by denying that intersubjectiv- which one is directed towards the other’s ity can be reduced to any factual encounter lived experiences. Thus, any intentional act between two individuals, such as the face- that discloses or presents the other’s sub- to-face encounter (see Zahavi, 2001a, b). jectivity from the second-person perspec- Rather, such concrete encounters presup- tive counts as empathy. Although empathy, pose the existence of another, more fun- so understood, is based on perception (of damental form of intersubjectivity that is the other’s bodily presence) and can involve rooted a priori in the very relation between inference in difficult or problematic situ- subjectivity and world. Heidegger’s (1996) ations (where one has to work out how way of making this point is to describe another person feels about something), it is how one always lives in a world perme- not reducible to some additive combination ated by references to others and already of perception and inference. furnished with meaning by others. Husserl The phenomenological conception of (1973) and Merleau-Ponty (1962) focus on empathy thus stands opposed to any the- the public nature of perceptual objects. ory according to which our primary mode The subject is intentionally directed towards of understanding others is by perceiving objects whose perspectival appearances bear P1: JzG 0521857430c04 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:56

philosophical issues: phenomenology 83

witness to other possible subjects. My per- coming to seem outdated, as this volume ceptual objects are not exhausted in their itself indicates. Conferences on conscious- appearance for me; each object always pos- ness (such as the biannual ‘Towards a Sci- sesses a horizon of coexisting profiles, which ence of Consciousness’ conference held in although momentarily inaccessible to me, Tucson, Arizona, and the annual meetings could be perceived by other subjects. The of the Association for the Scientific Study of perceptual object as such, through its per- Consciousness) now routinely include col- spectival givenness, refers, as it were, to other loquia informed by phenomenology along- possible subjects, and is for that very reason side cognitive science and analytic philoso- already intersubjective. Consequently, prior phy. In 2001 there appeared a new journal, to any concrete perceptual encounter with Phenomenology and the Cognitive Sciences. another subject, intersubjectivity is already Other journals, such as Consciousness and present as co-subjectivity in the very struc- Cognition and the Journal of Consciousness ture of perception. Studies, include articles integrating phe- Finally, there is a deep relation bet- nomenological, cognitive-scientific, and ana- ween intersubjectivity so understood and lytic approaches to consciousness. Given objectivity. My experience of the world these developments, the prospects for coop- as objective is mediated by my experi- eration and exchange among these traditions ence of and interaction with other world- in the study of consciousness now look very engaged subjects. Only insofar as I experi- promising. To this end, in this chapter we ence that others experience the same objects have called attention to a number of related as myself do I really experience these objects areas in which there is significant poten- as objective and real. To put this point in tial for collaborative research – intention- phenomenological language, the objectiv- ality, self-awareness, temporality, embodi- ity of the world is intersubjectively con- ment and perception, and intersubjectivity. stituted (i.e., brought to awareness or dis- We have also sketched a few ways of link- closed). This is an idea not foreign to ing phenomenology and cognitive science in Anglo-American philosophy, as the follow- these areas in order to suggest some direc- ing remark by Donald Davidson indicates: tions such research could take in the years “A community of minds is the basis of ahead. knowledge; it provides the measure of all things. It makes no sense to question the adequacy of this measure, or to seek a more ulti- Notes mate standard” (Davidson 2001,p.218). 1. For a recent discussion of the unity of the phenomenological tradition, see Zahavi (2006). Conclusion 2. An important forerunner of the current interest in the relation between phenomenology Phenomenology and analytic philosophy and cognitive science is the work of Hubert are the two most influential philosophi- Dreyfus (1982). Dreyfus has been a pioneer in cal movements of the twentieth century. bringing the phenomenological tradition into Unfortunately, their relationship in the past the heartland of cognitive science through was not one of fruitful cooperation and his important critique of artificial intelligence exchange, but ranged from disregard to out- (Dreyfus, 1991) and his groundbreaking stud- right hostility. To the extent that cogni- ies on skillful knowledge and action (Dreyfus, 2002 1982 tive science (especially in North America) ; Dreyfus & Dreyfus, ). Yet this work is also marked by a peculiar (mis)interpretation has been informed by analytic philoso- and use of Husserl. Dreyfus presents Husserl’s phy of mind, this attitude was at times phenomenology as a form of representational- perpetuated between phenomenology and ism that anticipates cognitivist and computa- cognitive science. tional theories of mind. He then rehearses Hei- In recent years, however, this state of degger’s criticisms of Husserl thus understood affairs has begun to change and is rapidly and deploys them against cognitivism and P1: JzG 0521857430c04 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:56

84 the cambridge handbook of consciousness

artificial intelligence. Dreyfus reads Husserl for two main reasons: First, the given in the largely through a combination of Heideg- phenomenological sense is not non-intentional ger’s interpretation and a particular ana- sense-data, but the phenomenal world as dis- lytic philosophical (Fregean) reconstruction of closed by consciousness. Second, the phenom- one aspect of Husserl’s thought (the repre- enality of the world is not understandable sentationalist interpretation of the noema). apart from the constitutive relation that con- Thus, Husserlian phenomenology as Dreyfus sciousness bears to it. For recent discussions presents it to the cognitive science and ana- of some of these issues see Botero (1999) and lytic philosophy of mind community is a prob- Roy (2003). lematic interpretive construct and should not 4. This sense of the epoche´ is well put by the be taken at face value. For a while Dreyfus’s noted North American and Indian phenome- interpretation functioned as a received view in nologist, J. N. Mohanty (1989,pp.12–13): “I this community of Husserl’s thought and its need not emphasize how relevant and, in fact, relation to cognitive science. This interpreta- necessary is the method of phenomenolog- tion has since been seriously challenged by a ical epoche´ for the very possibility of gen- number of Husserl scholars and philosophers uine description in philosophy. It was Husserl’s (see Zahavi, 2003, 2004, for further discussion; genius that he both revitalized the descrip- see also Thompson, 2007). These studies have tive method for philosophy and brought to argued that (i) Husserl does not subscribe to a the forefront the method of epoche,´ without representational theory of mind; (ii) Husserl is which one cannot really get down to the job. not a methodological solipsist (see the section The preconceptions have to be placed within on methodology; (iii) Husserl does not assimi- brackets, beliefs suspended, before philosophy late all intentionality to object-directed inten- can begin to confront phenomena as phenom- tionality (see the section on intentionality); ena. This again is not an instantaneous act of (iv) Husserl does not treat the ‘background’ of suspending belief in the world or of directing object-directed intentional experiences as sim- one’s glance towards the phenomena as phe- ply a set of beliefs understood as mental rep- nomena, but involves a strenuous effort at rec- resentations (see the section on intentionality); ognizing preconceptions as preconceptions, at and (v) Husserl does not try to analyze the ‘life- unraveling sedimented interpretations, at get- world’ into a set of sedimented background ting at presuppositions which may pretend to assumptions or hypotheses (equivalent to a be self-evident truths, and through such pro- system of frames in artificial intelligence). In cesses aiming asymptotically at the prereflec- summary, although Dreyfus is to be credited tive experience.” for bringing Husserl into the purview of cog- 5. We can discern here another reason for not nitive science, it is important to go beyond interpreting Husserl as a philosopher who his interpretation and to re-evaluate Husserl’s relies on any simple or straightforward notion relationship to cognitive science on the basis of an uninterrupted given in experience: Pas- of a thorough assessment of his life’s work. sive affection is not the reception of simple and This re-evaluation is already underway (see unanalyzable sense impressions, but has a field Petitot, Varela, Pachoud & Roy, 1999) and structure. can be seen as part of a broader reappro- 6. When we think a certain thought, the thinking priation of phenomenology in contemporary will often be accompanied by a non-vocalized thought. utterance or aural imagery of the very string 3. Does Husserl thereby succumb to the so-called of words used to express the thought. At the philosophical myth of the given? This is a same time, the thought will also frequently difficult and complicated question, and space evoke certain mental images. It could be argued prevents us from addressing it here. There is that the phenomenal qualities encountered not one but several different notions of the in abstract thought are constituted by such ‘given’ in philosophy, and Husserl’s thought imagery. Husserl makes clear in his Logical developed considerably over the course of his Investigations, however, that this attempt to life, such that he held different views at dif- deny that thought has any distinct phenome- ferent times regarding what might be meant nality beyond such sensorial and imagistic phe- by it. Suffice it to say that it is a mistake to nomenality is problematic. There is a marked label Husserl as a philosopher of the given in difference between what it is like to imag- the sense originally targeted by Wilfrid Sellars, ine aurally a certain string of meaningless P1: JzG 0521857430c04 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:56

philosophical issues: phenomenology 85

noise, and what it is like to imagine aurally sided polygon’, and ‘a myriagon is a many sided the very same string while understanding and polygon’, although the imagery that accompa- meaning something by it (Husserl, 2000, I., nies both thoughts might be indistinguishable. pp. 193–194, II., p. 105). Because the phenom- Thus, as Husserl concludes, although imagery enality of the sensory content is the same in might function as an aid to the understanding, both cases, the phenomenal difference must it is not what is understood; it does not con- be located elsewhere, namely, in the think- stitute the meaning of the thought (Husserl, ing itself. The case of homonyms and syn- 2000, I., p. 208). onyms also demonstrates that the phenomenality of thinking and the phenomenality of aural imagery can vary independently of each References other. As for the attempt to identify the phenomenal quality of thought with the phenomenal quality of visualization, a similar argument Bernet, R., Kern, I., & Marbach, E. (1993). can be employed. Two different thoughts, say, An introduction to Husserlian phenomenology. ‘Paris is the capital of France’, and ‘Parisians Evanston, IL: Northwestern University Press. regularly consume baguettes’, might be accom- Botero, J.-J. (1999). The immediately given as panied by the same visualization of baguettes, ground and background. In J. Petitot, F. J. but what it is like to think the two thoughts Varela, B. Pachoud, & J.-M. Roy (Eds.), Nat- remains different. Having demonstrated this uralizing phenomenology: Issues in contemporary much, Husserl then proceeds to criticize the phenomenology and cognitive science (pp. 440– view according to which the imagery actually 463). Stanford, CA: Stanford University Press. constitutes the very meaning of the thought – Damasio, A. R. (1999). The feeling of what that to understand what is being thought happens: Body and emotion in the making of con- is to have the appropriate ‘mental image’ sciousness. New York: Harcourt Brace. 2000 before one’s inner eye (Husserl, , I., Dancy, J. (1985). An introduction to contemporary 206 209 pp. – ). The arguments he employs bear epistemology. Oxford: Basil Blackwell. striking resemblance to some of the ideas Davidson, D. (2001). Subjective, intersubjective, that were subsequently used by Wittgenstein objective. Oxford: Oxford University Press. (1999)inhisPhilosophical Investigations: (i) 1999 From time to time, the thoughts we are think- Depraz, N. ( ). The phenomenological reducing, for instance ‘every algebraic equation of tion as praxis. Journal of Consciousness Studies, 95 110 uneven grade has at least one real root’, will in 6, – . fact not be accompanied by any imagery what- Depraz, N., Varela, F. J., & Vermersch, P. (2003). soever. If the meaning were actually located in On becoming aware: A pragmatics of expe- the ‘mental images’, the thoughts in question riencing. Amsterdam and Philadelphia: John would be meaningless, but this is not the case. Benjamins Press. (ii) Frequently, our thoughts, for instance ‘the Dreyfus, H. (1982). Introduction. In H. Dreyfus & horrors of World War I had a decisive impact H. Harrison (Eds.), Husserl, intentionality and on post-war painting’, will in fact evoke cer- cognitive science. Cambridge, MA: MIT Press. tain visualizations, but visualizations of quite Dreyfus, H. (1991). What computers still can’t do. unrelated matters. To suggest that the mean- Cambridge, MA: MIT Press. ings of the thoughts are to be located in such Dreyfus, H. (2002). Intelligence without repre- images is absurd. (iii) Furthermore, the fact sentation – Merleau-Ponty’s critique of mental that the meaning of a thought can remain representation. Phenomenology and the Cogni- the same although the accompanying imagery tive Sciences, 1, 367–383. varies also precludes any straightforward iden- 1982 tiﬁcation. (iv) An absurd thought, like the Dreyfus, H., & Dreyfus, S. ( ). Mind over thought of a square circle, is not meaning- machine. New York: Free Press. less, but cannot be accompanied by a matching Drummond, J. J. (2003). The structure of inten- image (a visualization of a square circle being tionality. In D. Welton (Ed.), The new Husserl: impossible in principle). (v) Finally, referring A critical reader (pp. 65–92). Bloomington: to Descartes’ famous example in the Medita- Indiana University Press. tions, Husserl points out that we can easily dis- Freeman, W. J. (2000). Emotion is essential tinguish thoughts like ‘a chiliagon is a many to all intentional behaviors. In M. D. Lewis P1: JzG 0521857430c04 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:56

86 the cambridge handbook of consciousness

& I. Granic (Eds.), Emotion, development, and tal logic (A. J. Steinbock, Trans.) Dordrecht: self-organization. Dynamic systems approaches Kluwer Academic Publishers. 209 235 to emotional development (pp. – ). Husserl, E. (2006). Phantasy, image consciousness, Cambridge: Cambridge University Press. and memory (1898–1925) (J. B. Brough, Trans.). Gallagher, S. (1998). The inordinance of time. Dordrecht, The Netherlands: Springer. Evanston, IL: Northwestern University Press. James, W. (1981). The principles of psychology. Goldman, A. I. (2000). Folk psychology and men- Cambridge, MA: Harvard University Press. 4 25 tal concepts. Protosociology, 14, – . Lloyd, D. (2002). Functional MRI and the study Gopnik, A. (1993). How we know our minds: of human consciousness. Journal of Cognitive The illusion of first-person knowledge of inten- Neuroscience, 14, 818–831. tionality. Behavioral and Brain Sciences, 16, Lloyd, D. (2003). Radiant cool. A novel theory of 1 14 – . consciousness. Cambridge, MA: MIT Press. 1982 Heidegger, M. ( ). The basic problems of phe- Lutz, A., & Thompson, E. (2003). Neurophe- nomenology (A. Hofstadter, Trans.). Blooming- nomenology: Integrating subjective experience ton, IN: Indiana University Press. and brain dynamics in the neuroscience of con- 1996 Heidegger, M. ( ). Being and time (J. Stam- sciousness. Journal of Consciousness Studies, 10, baugh, Trans). Albany, NY: State University of 31–52. New York Press. Merleau-Ponty, M. (1962). Phenomenology of per- 1998 Hurley, S. L. ( ). Consciousness in action. ception (C. Smith, Trans.) London: Routledge Cambridge, MA: Harvard University Press. Press. 2003 Hurley, S. L., & Noe,¨ A. ( ). Neural plasticity Milner, A. D., & Goodale, M. A. (1995). The and consciousness. Biology and Philosophy, 18, visual brain in action. New York: Oxford Uni- 131 168 – . versity Press. Husserl, E. (1970). The crisis of European sciences Mohanty, J. N. (1989). Transcendental phenomeno- and transcendental phenomenology (D. Carr, logy. Oxford: Basil Blackwell. Trans.). Evanston, IL: Northwestern University Moran, D. (2000). Introduction to phenomenology. Press. London: Routledge Press. 1973 Husserl, E. . Zur Phänomenologie der Intersub- Nagel, T. (1979). What is it like to be a bat? In jektivität, Dreiter Teil: 1929–1935. Husserliana T. Nagel (Ed.), Mortal questions (pp. 165–180). 15 (Vol. ). The Hague: Martinus Nijhoff. New York: Cambridge University Press. 1975 Husserl, E. ( ). Experience and judgment (J. S. Noe,¨ A. (2004). Action in perception. Cambridge, Churchill, Trans.) Evanston, IL: Northwestern MA: MIT Press. University Press. O’Regan, J. K., & Noe,¨ A. (2001). A sensori- 1987 Husserl, E. ( ). Aufsätze und Vorträge (1911– motor account of vision and visual conscious- 1921). Husserliana XXV. Dordrecht: Martinus ness. Behavioral and Brain Sciences, 24, 939– Nijhoff. 1031. Husserl, E. (1989). Ideas pertaining to a pure phe- Panskepp, J. (1998a). Affective neuroscience: The nomenology and to a phenomenological philoso- foundations of human and animal emotions. phy, second book (R. Rojcewicz & A. Schuwer, Oxford: Oxford University Press. Trans). Dordrecht: Kluwer Academic Publish- Panskepp, J. (1998b). The periconscious sub- ers. strates of consciousness: Affective states and Husserl, E. (1991). On the phenomenology of the the evolutionary origins of self. Journal of Con- consciousness of internal time (1893–1917) (J. B. sciousness Studies, 5, 566–582. Brough, Trans.) Dordrecht: Kluwer Academic Parvizi, J., & Damasio, A. (2001). Consciousness Publishers. and the brainstem. Cognition, 79, 135–159. 1997 Husserl, E. ( ). Thing and space: Lectures of Petitot, J., Varela, F. J., Pachoud, B., & Roy, J.-M. 1907 (R. Rojcewicz, Trans.) Dordrecht: Kluwer (Eds.). (1999). Naturalizing phenomenology: Academic Publishers. Issues in contemporary phenomenology and cogni- Husserl, E. (2000). Logical investigations I–II (J. tive science. Stanford, CA: Stanford University N. Findley, Trans.). London: Routledge Press. Press. Husserl, E. (2001). Analyses concerning passive Rainville, P. (2005). Neurophenom´ enologie´ des and active synthesis. Lectures on transcenden- etats´ et des contenus de conscience dans P1: JzG 0521857430c04 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:56

philosophical issues: phenomenology 87

l’hypnose et l’analgesie´ hypnotique. Th´eolog- (Eds.), Cognition and the brain: The philosophy ique, 12, 15–38. and neuroscience movement (pp. 40–97). New Rizzolatti, G. L., Fadiga, L., Fogassi, L., & Gallese, York: Cambridge University Press. V. (1997). The space around us. Science, 277, Varela, F. J. (1996). Neurophenomenology: A 190–191. methodological remedy for the hard prob- Roy, J.-M.(2003). Phenomenological claims and lem. Journal of Consciousness Studies, 3, 330– the myth of the given. In E. Thompson (Ed.), 350. The problem of consciousness: New essays in Varela, F. J. (1999). The specious present: A neu- phenomenological philosophy of mind. Cana- rophenomenology of time consciousness. In J. dian Journal of Philosophy, Suppl. Vol. 29 Petitot, F. J. Varela, B. Pachoud, & J.-M. Roy (pp. 1–32). Calgary, AL: University of Alberta (Eds.), Naturalizing phenomenology: Issues in Press. contemporary phenomenology and cognitive sci- Sartre, J.-P. (1956). Being and nothingness (H. ence (pp. 266–314). Stanford, CA: Stanford Barnes, Trans). New York: Philosophical University Press. Library. Varela, F. J., & Shear, J. (1999). The view from Searle, J. R. (1983). Intentionality: An essay in within. First-person approaches to the study of the philosophy of mind. Cambridge: Cambridge consciousness. Thorverton, UK: Imprint Aca- University Press. demic. Stein, E. (1989). On the problem of empathy (W. Wittgenstein, L. (1999). Philosophical investiga- Stein, Trans.) Washington, DC: ICS Publica- tions (3d ed.; G. E. M. Anscombe, Trans.) tions. Englewood Cliffs, NJ: Prentice Hall. Steinbock, A. (1995). Home and beyond: Genera- Zahavi, D. (1999). Self-awareness and alterity. tive phenomenology after Husserl. Evanston, IL: A phenomenological investigation. Evanston, IL: Northwestern University Press. Northwestern University Press. Sokolowski, R. (2000). An introduction to phe- Zahavi, D. (2001a). Husserl and transcendental nomenology. Cambridge: Cambridge University intersubjectivity. Athens, OH: Ohio University Press. Press. Thompson, E. (2001). Empathy and conscious- Zahavi, D. (2001b). Beyond empathy: Phenomen- ness. In E. Thompson (Ed.), Between ourselves: ological approaches to intersubjectivity. Jour- Second-person issues in the study of conscious- nal of Consciousness Studies, 8, 151–167. ness (pp. 1–32). Thorverton, UK: Imprint Zahavi, D. (2003). Husserl’s phenomenology. Stan- Academic. ford, CA: Stanford University Press. Thompson, E. (2005). Empathy and human Zahavi, D. (2004). Husserl’s noema and the experience. In J. D. Proctor (Ed.), Science, reli- internalism-externalism debate. Inquiry, 47, gion, and the human experience (pp. 261–285). 42–66. New York: Oxford University Press. Zahavi, D. (2005). Subjectivity and selfhood: Inves- Thompson, E. (2007). Mind in life: Biology, phe- tigating the ﬁrst-person perspective. Cambridge, nomenology, and the sciences of mind. Cam- MA: MIT Press. bridge, MA: Harvard University Press. Zahavi, D. (2006). The phenomenological tra- Thompson, E., Lutz, A., & Cosmelli, D. (2005). dition. In D. Moran (Ed.), Routledge compan- Neurophenomenology: An introduction for ion to twentieth-century philosophy. London: neurophilosophers. In A. Brook & K. Akins Routledge. P1: JzG 0521857430c04 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 3:56

88 P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

CHAPTER 5 Asian Perspectives: Indian Theories of Mind

Georges Dreyfus and Evan Thompson

Abstract Introduction

This chapter examines Indian views of the In discussing Asian views of mind and con- mind and consciousness, with particular sciousness, we must start from the realiza- focus on the Indian Buddhist tradition. To tion that this topic presents insurmountable contextualize Buddhist views of the mind, challenges. The diversity of Asian cultures we first provide a brief presentation of some from China to India to Iran is so great that of the most important Hindu views, particu- it is impossible to find coherent ways to dis- larly those of the S¯am. khya school. Whereas cuss the mental concepts of these cultures this school assumes the existence of a real over and above listing these conceptions transcendent self, the Buddhist view is that and noting their differences. Hence, rather mental activity and consciousness function than chart a territory that hopelessly extends on their own without such a self. We focus our capacities, we have chosen to exam- on the phenomenological and epistemologi- ine Indian views of the mind, with a spe- cal aspects of this no-self view of the mind. cial focus on the Indian Buddhist tradition, We first discuss the Buddhist Abhidharma which can be traced back to the first cen- and its analysis of the mind in terms of turies after the life of Siddhartha Gautama, awareness and mental factors. The Abhid- the Buddha (566–483 bce), and which con- harma is mainly phenomenological; it does tinued to develop in India through the 7th not present an epistemological analysis of and 8th centuries ce. This approach allows the structure of mental states and the way us to present a more grounded and coherent they relate to their objects. To cover this view of the mind as conceived in the Indian topic we turn to Dharmak¯ırti, one of the philosophical tradition and to indicate some main Buddhist epistemologists, who offers areas of interest that this tradition offers a comprehensive view of the types of cogni- to cognitive scientists and philosophers tion and their relation to their objects. of mind.

89 P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

90 the cambridge handbook of consciousness

In talking about the mind, it is impor- ily amount to a classical mind-body dualism tant to define the term, for it is far from (of the sort found in Descartes’ Meditations unambiguous. In most Indian traditions, the or Plato’s Phaedo). Moreover, although they mind is neither a brain structure nor a mech- agree in rejecting the materialist view, they anism for treating information. Rather, mind strongly disagree in their presentations of the is conceived as a complex cognitive process mind. consisting of a succession of related mental In this chapter, we focus mostly on the states. These states are at least in principle Buddhist tradition, exploring some of its phenomenologically available; that is, they views of the mind. One of the most salient can be observed by attending to the way features of this tradition is that its accounts in which we experience feeling, perceiving, of the mind and consciousness do not posit thinking, remembering, and so on. Indian the existence of a self. According to this tra- thinkers describe these mental states as cog- dition, there is no self, and mental activity nizing (jña¯) or being aware (buddh) of their cannot be understood properly as long as object. Thus, the mind is broadly conceived one believes in a self. The Hindu tradition, by traditional Indian thinkers as constituted by contrast, maintains that mental life does by a series of mental states that cognize involve a permanent self. Thus, to contextu- their objects. alize Buddhist views of the mind, we begin This general agreement breaks down with a brief presentation of some of the most quickly, however, when we turn to a more important Hindu views. We then present the detailed analysis of the nature and struc- Buddhist Abhidharma and its analysis of the ture of the mind, a topic on which vari- mind in terms of awareness and mental fac- ous schools entertain vastly different views. tors. Traditionally, the Abhidharma makes Some of these disagreements relate to the up one of the ‘three baskets’ into which ontological status of mental states and the Buddhists divide their scriptures – Sutra or way they relate to other phenomena, partic- sayings of the Buddha, Vinaya or monas- ularly physical ones. Such disagreements are tic discipline, and Abhidharma, which sys- related to well-known ideas in the Western tematizes Buddhist teachings in the form of tradition, particularly the mind-body dual- detailed analyses of experience. In examin- ism that has concerned Western philoso- ing the Abhidharma, we examine the ways phy since Descartes. But many of the views in which this tradition analyzes the differ- entertained by Indian thinkers are not eas- ent functions of the mind without positing ily mapped in Western terms, as we see in the existence of a self. These analyses are in this chapter. certain ways reminiscent of those in cogni- Most Indian thinkers do not consider the tive science that aim to account for cognitive ontological status of mental states to be a processing without invoking a homunculus particularly difficult question, for most of or ‘little man’ inside the head who over- them accept that there is an extra-physical sees the workings of the mind (or merely reality. Among all the schools, only the passively witnesses the results; see Varela, Materialist, the C¯arv¯aka, reduces the men- Thompson, & Rosch, 1991, for further distal to physical events. For its proponents, cussion of this parallel). The Abhidharma, mental states do not have any autonomous however, is phenomenological; its concern is ontological status and can be completely to discern how the mind works as evidenced reduced to physical processes. They are just by experience (but especially by mentally properties of the body, much like the ine- disciplined and refined contemplative expe- briating property of beer is a property of rience). Although thus it is also epistemo- beer. Most other thinkers reject this view logical, the Abhidharma does not present forcefully and argue that the mind can any developed epistemological analysis of neither be eliminated nor reduced to the the structure of mental states and the way material. Their endorsement of an extra- they relate to their objects so as to produce physical reality does not, however, necessar- knowledge. To cover this topic we turn to P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

asian perspectives: indian theories of mind 91

Dharmak¯ırti (c. 600 ce), one of the main mind has been largely adopted in the Hindu Buddhist epistemologists, who offers a com- tradition and beyond.1 prehensive view of the types of cognition The S¯am. khya approach rests on a dual- and their relation to their objects. istic metaphysics built on the opposi- The phenomenological analyses con- tion between material primordial nature tained in the Abhidharma and the epis- (pradhana¯ ) or materiality (prakr.ti) and a 2 temological analyses of Dharmak¯ırti offer spiritual self (atman¯ ) or person (purus.a). significant resources for cognitive scientists Nature is the universal material substratum and philosophers of mind in their efforts to out of which all phenomena other than the gain a better understanding of consciousness. self emerge and evolve. These phenomena, These analyses also constitute the theoretical which make up the world of diversity, are framework for the ways in which the Bud- physical transformations of the three quali- dhist tradition conceives of meditation and ties (gun. a) that compose primordial nature. mental training, both with regard to the phe- These three qualities are sattva (trans- nomenology of contemplative mental states parency, buoyancy), rajas (energy, activity), and the epistemology of the types of knowl- and tamas (inertia, obstruction). They are edge that these states are said to provide. principles or forces, rather than building Given the increasing scientific interest in the blocks. All material phenomena, including physiological correlates and effects of med- the intellect and organs of perception, are itation and their relation to consciousness understood to be made up of a combina- (see Chapter 19), it is important for the sci- tion of these three principles. The one prin- entific community to appreciate the phe- ciple not included in this constant process of nomenological and philosophical precision transformation is the self, which is perma- with which these states are conceptualized nent, non-material, and conscious or aware. in the Buddhist tradition. The self is also described as the conscious presence that witnesses the transformations of nature, but does not participate in them. Self and Mental States: As such it is passive, though it witnesses the AS¯am. khya View experiences deriving from the transformations of the world of diversity.3 One of the most important views of the Although the S¯am. khya analysis of mind mind in the Hindu tradition is found in the is dualistic, it does not fit within classical S¯am. khya school. Traditionally this school is mind-body dualism. For the S¯am. khya, the said to have been founded by the philoso- mind involves a non-material spiritual ele- pher Kapila, a legendary figure who may ment, namely the self. The self, however, is have lived as early as the 7th century not the same as the mind. Rather, the self is bce, but the earliest S¯am. khya text we pos- the mere presence to or pure witnessing of sess dates from the 3rd century ce. The the mental activities involved in the ordi- S¯am. khya tradition is one of the six classical nary awareness of objects. This pure wit- schools of Hindu philosophy (Nyaya, Vais- nessing, untainted by the diversity of the esika, S¯am. khya, Yoga, Purva Mimamsa, and material world, is not sufficient for men- Vedanta). ¯ Its influence extends to the other tal activities, for mental activities are rep- schools, particularly the Ved¯anta school, resentational or semantic and require more which later became especially important in than passive mirroring. Mental activity is the the development of Hindu thought. The apprehension of an object, and this activity S¯am. khya was in fact less a school proper than requires active engagement with objects and a way of thinking based on the categoriza- the formation of ideas and concepts neces- tion of reality. It was crucial in the forma- sary for purposeful action in the world. The tion of Indian philosophical thinking before self cannot account for such activity, how- and after the start of the Common Era, and ever, because it is changeless and hence pas- hence it is unsurprising that its view of the sive. To account for our cognitive activities, P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

92 the cambridge handbook of consciousness

we therefore need other elements that par- The intellect’s reflecting the self and tak- ticipate in the world of diversity. Because ing on the form of an object are not, how- any element that participates in the world ever, sufficient to fully determine experi- of change must emerge out of primordial ence. To become fully cognitive, experience materiality and hence be material, it fol- requires the formation of subjective and lows that the analysis of mental states can- objective poles. Experience needs to be the not be limited to their spiritual dimension experience of a particular individual appre- (the self), but must also involve material hending a particular object. The formation elements. Hence, for the S¯am. khya, mental of the subjective pole is the function of the activity requires the cooperation of the two ‘ego-sense’ (aham. kara¯ ), the sense of individ- fundamental types of substance that make ual subjectivity or selfhood tied to embodi- up the universe, passive consciousness and ment. This sense colors most of our expe- material nature. riences, which involve a sense of being a Having described the S¯am. khya meta- subject opposed to an object. The determi- physics, we can now sketch its influential nation of the objective pole, on the other analysis of mental activity.4 This analysis hand, is the function of ‘mentation’ (manas), starts with buddhi, which is usually trans- which oversees the senses and whose spe- lated as ‘the intellect’ and is the ability to cial function is discrimination. This function distinguish and experience objects. This abil- allows mentation to serve as an intermediary ity provides the prereflective and presubjec- between the intellect and the senses. Men- tive ground out of which determined mentation organizes sensory impressions and tal states and their objects arise; it is also the objects and integrates them into a temporal locus of all the fundamental predispositions framework created by memories and expec- that lead to these experiences. The intellect tations. In this way, our experience of objects emerges out of primordial matter and therein the world is created. fore is active, unlike the non-material and Although the dualistic metaphysics asso- passive self. The self is described metaphor- ciated with this view was rejected in the ically as a light, for it passively illuminates history of Indian philosophy, the S¯am. khya objects, making it possible for the intellect model of the mind was taken over by other to distinguish them. The intellect operates Hindu schools. It serves as a foundation in a representational way by taking on the of the philosopher Patanjali’s˜ (c. 2nd cen- form of what is known. This representational tury bce) Yoga view of mind, which is 5 ability works in two directions – toward the similar to the S¯am. khya. The Yoga view conscious and uninvolved self and toward also rests on the opposition between pas- the objects. The intellect, thanks to its qual- sive self and active mental activities (citta), a ity of clarity and transluscence (sattva), takes rubric under which intellect, ego-sense, and on the form of the self by reflecting it. As a mentation are grouped. Similarly, Iamkara´ result, it seems as if the self experiences the (788–820 ce), who savaged the dualism of diversity of objects, when it is actually the the S¯am. khya, took over its model of the intellect that undergoes these experiences, mind in his Advaita Ved¯anta, emphasizing the self being the mere witness of them. This the contrast between the transcendence of ability of the intellect to usurp the func- the self and the mental activities of the tion of consciousness helps the intellect in ‘inner sense’ (antah. karava) belonging to the 6 its apprehension of objects, for by itself the person. Hence, the S¯am. khya view can be intellect is active but unconsious. Awareness taken as representative of the Hindu view of of objects arises only when the intellect takes the mind, especially in its emphasis on the on the light of the self and reflects it on difference between a passive witnessing con- objects, much like pictures are created when sciousness and mental activity. light is projected onto a film. In this way, the According to this view, as we have seen, intellect becomes able to take on the form mental events come about through the con- of the object and thus to discern it. junction of two heterogeneous factors – a P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

asian perspectives: indian theories of mind 93

transcendent self and a diversity of men- the diversity of opinions and debates that has tal activities. It is a basic presupposition of animated this tradition. the Hindu tradition that mental life involves The object of the Abhidharma is to ana- a permanent self. Yet because mental life lyze both the realm of sentient experience also undeniably involves change, it cannot and the world given in such experience into be reduced to this single, motionless factor of its components in language that avoids the the self; hence the need for the complicated postulation of a unified subject. This anal- analysis briefly summarized here. This ten- ysis concerns the whole range of phenom- sion in accounts of the mind and conscious- ena, from material phenomena to nirvana ness between identity and change, unity and (the state of enlightenment, understood as diversity, is of course also prevalent through- the direct realization of the nature of reality, out Western philosophy and persists in cog- including especially the lack of any essen- nitive science. We turn now to the Bud- tial self and the consequent liberation from dhist tradition, which presents a different suffering). For example, there are elaborate perspective on this issue. discussions of the four primary and four secondary elements that make up matter (see de la Vallee´ Poussin, 1971,I:22). There are also The Abhidharma Tradition lengthy treatments of the nature, scope, and and its View of the Mind types of soteriological practices prescribed by the Buddhist tradition, a central focus The Buddhist tradition is based on the oppo- of the Abhidharma. But a large part of the site view of no-self (anatman¯ ). For the Bud- Abhidharmic discourse focuses on the anal- dhists, there is no self, and hence mental ysis of mental phenomena and their various activity is not in the service of such an entity, components. It is this part of the Abhid- but rather functions on its own. In short, harma that we examine in this chapter. for the Buddhists there is no self that is In considering experience, the Abhid- aware of the experiences one undergoes or harma proceeds in a rather characteristic the thoughts one has. Rather the thoughts way that may be disconcerting for new- themselves are the thinker, and the experi- comers, but reflects its historical origin as ences the experiencer. mnemonic lists of elements abstracted from How, then, do Buddhists explain the the Buddha’s discourses. For each type of complexities of the mind? How do they phenomenon considered, the Abhidharma explain mental regularities if there is no analyzes it into its basic elements (dharma), central controller to oversee the whole lists these elements, and groups them into process? the appropriate categories (examples are For an answer, we turn to the Abhid- given below). The study of the Abhidharma harma, one of the oldest Buddhist tradi- thus often revolves around the consideration tions, which can be traced back to the first of series of extended lists. centuries after the Buddha (566–483 bce). In elaborating such lists of components First elaborated as lists,7 the Abhidharma of experience and the world given in expe- contains the earlier texts in which Bud- rience, the Abhidharma follows the cen- dhist concepts were developed and hence is tral tenets of Buddhist philosophy, in par- the source of most philosophical develop- ticular the twin ideas of non-substantiality ments in Indian Buddhism. But the Abhid- and dependent origination. According to this harma is not limited to this role as a source philosophy, the phenomena given in experi- of Buddhist philosophical development. It ence are not unitary and stable substances, remained a vital focus of Buddhist thought but complex and fleeting formations of basic and kept evolving, at least until the 7th or elements that arise in dependence on com- 8th century ce. In this chapter, we focus on plex causal nexuses. Such non-substantiality two Indian thinkers from the 4th or 5th cen- is particularly true of the person, who is not tury ce, Asanga˙ and Vasubandhu, and ignore a substantial self, but a changing construct P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

94 the cambridge handbook of consciousness

dependent on complex configurations of It should be clear from this preliminary mental and material components. This anal- characterization that in elaborating a the- ysis, which is diametrically opposed to the ory of the mind the Abhidharma relies pri- S¯am. khya view, is not just limited to the per- marily on what we would call a first-person son, but is applied to other objects. approach. It is by looking directly at experi- All composite things are thus analyzed as ence that we gain an understanding of mind, being constituted of more basic elements. not by studying it as an object and attending Moreover, and this point is crucial, these to its external manifestations. This approach basic elements should not be thought of of the Ahbhidharma is not unlike that of as reified or stable entities, but as dynam- such Western thinkers as James, Brentano, ically related momentary events instanta- and Husserl, who all agree that the study neously coming into and going out of exis- of the mind must be based on attention to tence. Thus, when the Abhidharma analyzes experience (see Chapter 4). This approach matter as being made up of basic compo- is well captured by James’s famous claim nents, it thinks of those components not that in the study of the mind, “Introspec- as stable particles or little grains of matter, tive Observation is what we have to rely on but rather as fleeting material events, com- first and foremost and always” (James, 1981, ing into and going out of existence depend- p. 185). ing on causes and conditions. Similarly, the As James himself recognizes, how- mind is analyzed into its basic components; ever, first-person observation of the mind, namely, the basic types of events that although it might seem a straightforward make up the complex phenomenon we call enterprise, is not a simple affair and raises ‘mind’. numerous questions. What does it mean to This Abhidharmic analysis is not just observe the mind? Who observes? What is philosophical but it also has practical import. being observed? Is the observation direct Its aim is to support the soteriological prac- or mediated? In addition to these difficult tices that the Buddhist tradition recom- epistemological issues (some of which we mends. The lists of material and mental take up in the next section), there are also events are used by practitioners to inform questions about the reliability of observa- and enhance their practices. For example, tion. We are all able to certain degrees to the list of mental factors we examine shortly observe our own minds, but it is clear that is a precious aid to various types of medi- our capacities to do so differ. Whose obser- tation, providing a clear idea of which fac- vations are to be considered reliable? This tors need to be developed and which are to question is significant for the Abhidharmists, be eliminated. In this way, the Abhidharma who may include in their data not only functions not just as the source of Buddhist ordinary observations but also the observa- philosophy but also informs and supports tions of trained meditators. This inclusion the practices central to this tradition. of observation based on contemplative men- In the Abhidharma the mind is conceived tal training and meditative experience marks as a complex cognitive process consisting of an important difference between the Abhid- a succession of related momentary mental harma and James, as well as other Western states. These states are phenomenologically phenomenologists. Nevertheless, the degree available, at least in principle: They can be to which meditative experience is relevant observed by turning inwardly and attending to Buddhist theories of the mind is not a to the way we feel, perceive, think, remem- straightforward matter, as we see shortly. ber, and so on. When we do so, we notice The comparison between the Abhid- a variety of states of awareness, and we also harma and James goes further, however, than notice that these states change rapidly. It is their reliance on an introspective method. these mental states arising in quick succes- They also share some substantive similari- sion that the Abhidharma identifies as being ties, the most important of which is per- the basic elements of the mind. haps the idea of the stream of consciousness. P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

asian perspectives: indian theories of mind 95

For the Abhidharma, mental states do not is the 64th part of the time necessary to arise in isolation from each other. Rather, click one’s fingers or blink an eye (see de la each state arises in dependence on preceding Vallee´ Poussin, 1991,pp.70–71). Although moments and gives rise to further moments, these measurements differ, one could argue thus forming a mental stream or continuum that given the imprecision of premodern (santana,¯ rgyud), much like James’s ‘stream measurement, there is a rough agreement of thought’. This metaphor is also found in between these accounts, which present a the Buddhist tradition in which the Bud- moment of awareness as lasting for about dha is portrayed as saying, “The river never 1/100th of a second. This is already signifi- stops: there is no moment, no minute, no cantly faster than pyschophysical and elec- hour when the river stops: in the same way, trophysiological estimates of the duration of the flux of thought” (de la Vallee´ Poussin, a moment of awareness as being on the order 1991,p.69, translation from the French by of 250 milliseconds or a quarter of a sec- Dreyfus). ond (see Poppel,¨ 1988; Varela, Thompson, Unsurprisingly, there are also significant & Rosch, 1991,pp.72–79). But consider the differences between James and the Abhid- claim made by a Theravada Abhidharma harma. One difference of interest to context that “in the time it takes for lightning to temporary research is the issue of whether flash or the eyes to blink, billions of mind- mental states arise in continuity or not (see moments can elapse” (Bodhi, 1993,p.156). Varela, Thompson, & Rosch, 1991,pp.72– The time scale in this account, which is stan- 79). James’s view is well known: “Con- dard in the Theravada tradition, is faster by sciousness does not appear to itself chopped many orders of magnitude. up in bits” (James, 1981,p.233). Although This dramatic discrepancy alerts us to the content of consciousness changes, we some of the difficulties of accounts based on experience these changes as smooth and observation. For whom are we to believe? continuous, without any apparent break. On which tradition should we rely? More- The Abhidharma disagrees, arguing that over, we cannot but wonder about the although the mind is rapidly changing, its sources of these differences. Do they derive transformation is discontinuous. It is only from the observations of meditators, or are to the untrained observer that the mind they the results of theoretical elaborations? appears to flow continuously. According It is hard to come to a definitive conclusion, to the Abhidharma, a deeper observation but it seems reasonable to believe that these reveals that the stream of consciousness is accounts are not simply empirical observa- made up of moments of awareness, moments tions, but largely theoretical discussions, per- that can be introspectively individuated and haps supplemented by observation reports. described. Hence one must be cautious and not assume Several Abhidharma texts even offer that these texts reflect empirical findings. measurements of this moment, measure- Although some may, they are mostly the- ments one would expect to be based on oretical elaborations, which cannot be taken empirical observation. Yet such claims are at face value, but require critical interpre- problematic, for different Abhidharma tra- tation. Finally, another Abhidharma text ditions make claims that at times are strik- seems to muddy the waters further by claim- ingly at odds with one another. For exam- ing that the measure of a moment is beyond ple, the Mahavibhas¯.a¯, an important text the understanding of ordinary beings. Only from the first centuries of the Common Era, enlightened beings can measure the dura- states that there are 120 basic moments in tion of a moment (de la Vallee´ Poussin, 1991, an instant. The text further illustrates the p. 73). Thus it is not surprising that we are duration of an instant by equating it to the left wondering! time needed by an average spinner to grab According to the Abhidharma, the men- a thread. Not at all, argues another text: tal episodes that compose a stream of con- This measurement is too coarse. A moment sciousness take as their objects either real or P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

96 the cambridge handbook of consciousness

fictional entities. This object-directed char- tal states, but goes much further in its anal- acter of mind has been called ‘intentionality’ ysis, breaking down each mental state into by Western philosophers, such as Brentano its components. According to the Abhid- and Husserl. Brentano claimed that inten- harma schema, which is to our knowledge tionality is an essential feature of conscious- unique, each mental state is analyzed as ness and proposed it as a criterion of the having two aspects: (i) the primary factor mental. All acts of awareness are directed of awareness (citta), whose function is to toward or refer to an object, regardless of be aware of the object, and (ii) mental fac- whether this object is existent or not. We tors (caitesika), whose function is to qual- cannot think, wish, or dread unless our ify this awareness by determining its qual- mind is directed toward something thought itative nature as pleasant or unpleasant, about, wished for, or dreaded, which thus focused or unfocused, calm or agitated, pos- appears to the mind. Therefore, to be aware itive or negative, and so on. The philosopher is for something to appear to the mind. The Vasubandhu (c. 4th or 5th century ce), one Abhidharma seems to share this view, hold- of the great Abhidharmists, explains this dis- ing that every moment of cognition relates to tinction between awareness and mental fac- particular objects, and hence it assumes that tors as follows: intentionality and consciousness are inseparable.8 Cognition or awareness apprehends the The Abhidharma also holds that this thing itself, and just that; mental factors stream of consciousness is not material. It or dharmas associated with cognition such as sensation, etc., apprehend special char- is associated with the body during this life- acteristics, special conditions (de la Vallée time, but will come to exist in dependence Poussin, 1971,I:30).9 on other bodies after the death of this body. It is crucial to recognize, however, that the The basic insight is that mental states have immaterial stream of consciousness is not a two types of cognitive functions – (1) aware- soul in the Platonic or Cartesian sense, but an ness and (2) cognitive and affective engage- impersonal series of mental events. Buddhist ment and characterization. The mental state philosophers do not believe in an ontology of is aware of an object. For example, the sense substances – that reality comprises the exis- of smell is aware of a sweet object. But men- tence of independent entities that are the tal states are not just states of awareness. subjects of attributes or properties. Rather, They are not passive mirrors in which objects they argue that reality is made up of events are reflected. Rather, they actively engage consisting of a succession of moments. Thus, their objects, apprehending them as pleasant mind and matter are not substances, but or unpleasant, approaching them with par- evanescent events, and mental and mate- ticular intentions, and so forth. For example, rial events interact in a constantly ongo- a gustatory cognition of a sweet object is not ing and fluctuating process. Moreover, Bud- just aware of the sweet taste but also appre- dhist philosophers partake of the general hends the object as pleasant, distinguishes Indian reluctance to separate the mental and certain qualities such as its texture, and so the material. Hence they do not hold that on. It also categorizes the object as being the divide between the material and men- (say) one’s favorite Swiss chocolate. Such tal spheres is absolute. Nevertheless, for the characterization of the object is the func- Buddhists, in contrast to the Sam¯ . khya, there tion of the mental factors. We now describe is a sharp divide between the mental, which this distinction between the primary factor of is intentional and conscious, and other ele- awareness and mental factors in more detail. ments. In this respect, Buddhists are perhaps the closest among Indian philosophers to a The Primary Factor of Awareness classical mind-body dualism. The Abhidharma, however, does not stop The primary factor of awareness (citta)is at a view of the mind as a succession of men- also described as vijñana¯ , a term often P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

asian perspectives: indian theories of mind 97

translated as consciousness or cognitive aware- stances, such as fainting, that its presence ness. It is the aspect of the mental state that can be noticed or at least inferred. This con- is aware of the object. It is the very activity sciousness contains all the basic habits, ten- of cognizing the object, not an instrument dencies, and propensities (including those in the service of an agent or self (which, that persist from one life to the next) accu- as we have seen, the Buddhist philosophers mulated by the individual. It thus provides argue is nonexistent). This awareness merely a greater degree of continuity than manifest discerns the object, as in the above exam- cognitive awareness on its own. ple where one apprehends the taste of what The store-consciousness is mistaken by turns out to be one’s favorite Swiss choco- the afflictive mentation as being a self. In this late. Thus Vasubandhu speaks of awareness way one’s core inborn sense of self is formed. as the “bare apprehension of each object” (de From a Buddhist point of view, however, this la Vallee´ Poussin, 1971,I:30). sense of self is fundamentally mistaken. It is In most Abhidharma systems, there are a mental imposition of unity where there is six types of awareness: five born from the five in fact only the arising of a multiplicity of physical senses (sight, hearing, smell, taste, interrelated physical and mental events. The and touch) and mental cognition. Each type sense of control belonging to one’s sense of of sensory cognition is produced in depen- self is thus largely illusory. There is really dence on a sensory basis (one of the five nobody in charge of the physical and men- physical senses) and an object. This aware- tal processes, which arise according to their ness arises momentarily and ceases imme- own causes and conditions, not our whims. diately, to be replaced by another moment The mind is not ruled by a central unit, but of awareness, and so on. The sixth type by competing factors whose strength varies of awareness is mental. It is considered a according to circumstances. sense by the Abhidharma, like the five phys- Thus Asanga,˙ allegedly Vasubandhu’s ical senses, though there are disagreements half-brother, posits as many as eight types about its basis (see Guenther, 1976,pp.20– of consciousness, a doctrine usually asso- 30). ciated with a particular Buddhist school, Some Abhidharma texts, such as Asanga’s˙ the Yog¯ac¯ara. This school contains many (Rahula, 1980), argue that these six types of interesting insights, without which there is consciousness do not exhaust all the possi- no complete understanding of the depth ble forms of awareness. To this list Asanga˙ of Buddhist views of the mind, but there adds two types of awareness: the store- is not space to discuss these insights here. consciousness (alaya-vij˜¯ nana,¯ kun gzhi rnam Let us simply point out that there are shes) and afflictive mentation (klis. †a-manas, some interesting similarities between the 10 nyon yid; Rahula, 1980,p.17). The idea Yogac¯ ¯ ara and the S¯am. khya views. The of a store-consciousness is based on a dis- store-consciousness, in acting as the holder tinction between the six types of awareness, of all the potentialities accumulated by which are all described as manifest cognitive an individual, is not unlike the intellect awareness (pravr.tti-vijñana,¯ ’jug shes), and a (buddhi), whereas the afflictive mentation more continuous and less manifest form of seems similar to the ego-sense (aham. kara¯ ). awareness, which is the store-consciousness. Furthermore, mental cognition does not This awareness is invoked to answer the fol- seem too different from mentation (manas). lowing objection: If there is no self and the These similarities indicate the reach of the mind is just a succession of mental states, S¯am. khya model, even in a tradition whose then how can there be any continuity in our basic outlook is radically different. mental life? Asanga’s˙ answer is that there is a more continuous form of consciousness, Mental Factors which is still momentary, but exists at all times. Because it is subliminal, we usually Mental states are not just states of awareness; do not notice it. It is only in special circum- they also actively engage their objects, P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

98 the cambridge handbook of consciousness

qualifying them as pleasant or unpleasant, (lovingkindness), wisdom, equanimity, approaching them with a particular attitude, and non-harmfulness (compassion). r and so on. Mental factors, which are aspects Six root-afflictions: attachment, anger, of the mental state that characterize the ignorance, pride, negative doubt, and mis- object of awareness, account for this engage- taken view. r ment. In other words, whereas conscious- Twenty branch-afflictions: belligerence, ness makes known the mere presence of the vengefulness, concealment, spite, jeal- object, mental factors make known the par- ousy, avarice, pretense, dissimulation, ticulars of the content of awareness, defining self-satisfaction, cruelty, self-regarding the characteristics and special conditions of shamelessness, other-regarding shame- its object. They qualify the apprehension of lessness, inconsideration, mental dullness, the object as being pleasant or unpleasant, distraction, excitement, lack of confi- attentive or distracted, peaceful or agitated, dence/faith, laziness, lack of conscien- and so forth. tiousness, and forgetfulness. The translation of these elements of the mind (caitesika)asfactors is meant to cap- The nature of this complex typology ture the range of meanings that the Abhid- becomes clearer when one realizes that these harma associates with this term. The rela- six groups can be further reduced to three. tion between cognitive awareness and men- The first three groups contain all the neu- tal factors is complex. At times the Abhid- tral factors. They are the factors that can be harma construes this relation diachronically present in any mental state, whether positive as being causal and functional. Factors cause or negative. Hence these factors are neither the mind to apprehend objects in particu- positive nor negative in and of themselves. lar ways. At other times, the Abhidharma The next three groups are different. These seems to emphasize a synchronic perspec- factors are ethically determined. The eleven tive in which cognitive awareness and men- virtuous factors are positive in that they do tal factors coexist and cooperate in the same not compel us toward attitudes that lead to cognitive task.11 suffering. They leave us undisturbed, open In accordance with its procedure, the to encounter reality with a more relaxed and Abhidharma studies mental factors by listing freer outlook. The twenty-six afflictive fac- them, establishing the ways in which they tors, on the other hand, disturb the mind, arise and cease, and grouping them in the creating frustration and restlessness. They appropriate categories. Each Abhidharma are the main obstacles to the life of the good tradition has a slightly different list. Here we as understood by the Buddhist tradition. The follow a list of 51 mental factors distributed very presence of these factors marks the in 6 groups.12 The mental typology presented mental state as virtuous or afflictive. Thus in this list has a number of interesting fea- it is clear that the Abhidharma typology is tures in relation to more familiar Western explicitly ethical. philosophical and scientific typologies: This presentation also offers interesting r insights concerning the cognitive functions Five omnipresent factors: feeling, discern- of the mind. In particular, the analysis of the ment, intention, attention, and contact r five omnipresent factors – feeling, discern- Five determining factors: aspiration, ment, intention, attention, and contact – appreciation, mindfulness, concentra- shows some of the complexities of Abhid- tion, and intelligence r harmic thinking. These five are described Four variable factors: sleep, regret, inves- as omnipresent because they are present in tigation, and analysis every mental state. Even in a subliminal state r Eleven virtuous factors: confidence/faith, such as the store-consciousness these five self-regarding shame, other-regarding factors are present. The other factors are not shame, joyful effort, pliability, consci- necessary for the performance of the most entiousness, detachment, non-hatred minimal cognitive task (the apprehension of P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

asian perspectives: indian theories of mind 99

an object, however dimly and indistinctly). on the basis of visual clues. For the Abhid- Hence they are not present in all mental harma, however, this question strikes deeper, states, but only in some. because several meditative states in the Bud- One striking feature of this list is the dhist tradition are described as signless (ani- pre-eminent place of feeling (vedana,¯ tshor mitta, mthan med).13 Can the mind in these ba) as the first of the factors. This empha- states identify its object without making sis reflects the fundamental outlook of the distinctions? Or is it the case that even in the tradition, which views humans as being first case of signless states the mind still makes and foremost sentient. But it also reflects a distinctions, although they are not linguistic distinctive view of the cognitive realm that or even conceptual? In a short chapter such emphasizes the role of spontaneous value as this one, we cannot delve into this issue, attribution. For the Abhidharma, a mental despite its relevance to the dialogue between state is not only aware of an object but at Buddhism and the sciences of mind. the same time it also evaluates this object. Other factors are also significant. Inten- This evaluation is the function of the feel- tion (cetana,¯ sems pa) is a central and ing tone that accompanies the awareness and omnipresent factor, which determines the experiences of the object as either pleasant, moral (not ethical) character of the men- unpleasant, or neutral. This factor is central tal state. Every mental state approaches its in determining our reactions to the events object with an intention, a motivation that we encounter, because, for the most part, may be evident to the person or not. This we do not perceive an object and then feel intention determines the moral nature of the good or bad about it out of considerate judgmental state, whether it is virtuous, non- ments. Rather, evaluation is already built virtuous, or neutral. This factor is associ- into our experiences. We may use reflection ated with the accomplishment of a goal and to come to more objective judgments, but hence is also thought of as a focus of organi- those mostly operate as correctives to our zation for the other factors. spontaneous evaluations. Also important is the role of attention Feeling is not the only important factor, (manasikara,¯ yid la byed pa), another one of however. A mental state involves not only the five omnipresent factors. It is the ability awareness and feeling but also discernment of the mind to be directed to an object. A (sam. jnã,¯ ’du shes also often translated as per- contemporary commentator explains atten- ception or recognition). This factor involves tion this way: “Attention is the mental fac- the mind’s ability to identify the object by tor responsible for the mind’s advertence to distinguishing it from other objects. This the object, by virtue of which the object is concept of discernment presents some diffi- made present to consciousness. Its character- culties, however. In its most elaborate form, istic is the conducting of the associated men- discernment is based on our semiotic ability tal states [i.e., factors] to the object. Its func- to make distinctions, mostly through linguis- tion is to yoke the associated mental states tic signs. But for the Abhidharma, the mind’s [i.e., factors] to the object” (Bodhi, 1993, ability to identify objects is not limited to lin- p. 81). Every mental state has at least a min- guistic distinctions, however important they imal amount of focus on its object; hence may be. Infants and non-human animals are attention is an omnipresent factor. understood to have the ability to make dis- Attention needs to be distinguished from tinctions, although they do not use sym- two other related factors. The first is concen- bolic thinking. Are these prelinguistic cogni- tration (samadhi,¯ ting nge ’dzin), the ability tions nevertheless semiotic? Do they involve of the mind to dwell on its object single- non-linguistic signs, or do they make dis- pointedly. The second is mindfulness (smr.ti, tinctions without the use of signs? It seems dran pa, also translated as recollection), plausible to argue that some of these states which is the mind’s ability to keep the object involve non-linguistic signs, as in the case in focus without forgetting, being distracted, of visual cognitions that distinguish objects wobbling, or floating away from the object. P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

100 the cambridge handbook of consciousness

Both abilities are not present in every men- ousy, avarice, pretense, dissimulation, tal state. Concentration differs from atten- self-satisfaction, cruelty, self-regarding tion in that it involves the ability of the shamelessness, other-regarding shame- mind not just to attend to an object but lessness, inconsideration, mental dullness, also to sustain this attention over a period of distraction, excitement, lack of confi- time. Similarly, mindfulness is more than the dence/faith, laziness, lack of conscien- simple attending to the object. It involves tiousness, and forgetfulness. the capacity of the mind to hold the object in its focus, preventing it from slipping Here again we notice that this list away in forgetfulness. Hence both factors, contains factors that look quite differ- which are vital to the practice of Buddhist ent. Some factors such as ignorance are meditation (see Chapter 19), are included clearly cognitive, whereas others such as among the determining factors. They are anger and jealousy are more affective. They present only when the object is apprehended are grouped together because they are with some degree of clarity and sustained afflictive: They trouble the mind, mak- focus. ing it restless and agitated. They also The factors discussed so far are mainly compel and bind the mind, preventing cognitive, but the Abhidharma list also one from developing more positive atti- includes mental factors we would describe as tudes. This afflictive character may be obvi- emotions. Consider the ethically determined ous in the case of attachment and jeal- factors, starting with the eleven virtuous ousy, which directly lead to dissatisfaction, ones: confidence/faith, self-regarding shame, frustration, and restlessness. Ignorance – other-regarding shame, joyful effort, plia- that is, our innate and mistaken sense of self – bility, conscientiousness, detachment, non- is less obviously afflictive, but its role is hatred (lovingkindness), wisdom, equanim- nonetheless central here, because it brings ity, and non-harmfulness (compassion). about the other more obviously afflictive We would describe several of these fac- factors. tors, such as lovingkindness and compassion, Although there are many elements in the as emotions. These two factors belong to typology of mental factors that we can iden- what we would characterize as the affec- tify as emotions (anger, pride, jealously, lov- tive domain, although here they are under- ingkindness, and compassion), there is no stood not with regard to their affectivity, but category that maps onto our notion of emo- rather in relation to their ethical character.14 tion. Most of the positive factors are not Hence they are grouped with other fac- what we would call emotions, and although tors, such as wisdom and conscientious- most of the negative factors are affective, ness, that are more cognitive than affec- not all are. Hence it is clear that the Abhid- tive. For the Abhidharma all these factors harma does not recognize the notion of emo- are grouped together. They are all positive tion as a distinct category of a mental typol- in that they promote well-being and free- ogy. There is no Abhidharma category that dom from the inner compulsions that lead to can be used to translate our concept of emo- suffering. tion, and similarly our concept of emotion is The afflictive factors, on the other hand, difficult to use to translate the Abhidharma are precisely those that lead to suffering. terminology. Rather than opposing rational They are by far the most numerous group and irrational elements of the psyche, or and are clearly a major focus of this typology: cognitive and emotive systems of the mind r (or brain), the Abhidharma emphasizes the Six root-afflictions: attachment, anger, distinction between virtuous and afflictive ignorance, pride, negative doubt, and mis- mental factors. Thus, our familiar Western taken view. distinction between cognition and emotion r Twenty branch-afflictions: belligerence, simply does not map onto the Abhidharma vengefulness, concealment, spite, jeal- typology. Although the cognition/emotion P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

asian perspectives: indian theories of mind 101

distinction has recently been called into is the nature of valid cognition (praman¯ . a) question by some scientists (see Chapter 29 and what are its types? Hindu thinkers tend and Damasio, 1995), it remains central to present a realist theory, which liberally to most of contemporary cognitive science allows a diversity of instruments of valid and philosophy of mind. The Abhidharma cognition. For example, the S¯am. khya asserts typology offers a different approach, one that there are three types of valid sources in which mental factors are categorized of knowledge: perception (pratyaks.a), infer- according to their ethical character. This ence (anumana¯ ), and verbal testimony typology could prove fruitful for psychol- (sabda´ ). The Ny¯aya, perhaps the most ogists and social and affective neuroscien- important Hindu logico-epistemological tra- tists interested in studying the biobehav- dition, added a fourth type of valid cog- ioral components of human well-being (see nition, analogy (upamana¯ ). This fourfold Goleman, 2003). typology provided the most authoritative The analyses of mental factors we have epistemological typology in India. Buddhist reviewed indicate the complexity, sophis- epistemology, however, rejects these typolo- tication, and uniqueness of the Abhid- gies and offers a more restrictive view, lim- harma mental typology. For this reason, the iting knowledge to inference and percep- Abhidharma is often called, somewhat mis- tion. It is in its examination of inference leadingly, ‘Buddhist psychology’.15 Yet the as a source of knowledge that the Buddhist Abhidharma analysis does not answer all tradition analyzes reasoning, in particular the questions raised by the Buddhist view the conditions necessary for the forma- of the mind as lacking a real self. In par- tion of sound reasons and all their pos- ticular, it leaves out the issue of the cog- sible types. Hence this tradition is often nitive or epistemic structure of the mental described, also somewhat misleadingly, as states that make up the stream of con- ‘Buddhist logic’.16 sciousness. To examine this issue, we turn The interpretation of the word praman¯ . a to another Indian Buddhist tradition, the is itself a topic of debate among Buddhist logico-epistemological tradition of Dign¯aga and Hindu thinkers. For the latter, this word, and Dharmak¯ırti (see Dreyfus, 1997; Dunne, in accordance with its grammatical form, 2004). refers to ‘means of valid cognition’. This understanding also accords with the basic view of this school that knowledge is owned Buddhist Epistemology by a subject, the self, to whom knowledge is ultimately conveyed. For example, the This tradition was started by Dign¯aga around Ny¯aya asserts that knowledge is a quality 500 ceand was expanded significantly more of the self. It is only when I become con- than a century later by Dharmak¯ırti, the scious of something that I can be said to focus of our analysis. Its contribution was know it. This view is energetically rejected the explicit formulation of a complete Bud- by Dharmak¯ırti, who follows the classical dhist logical and epistemological system. Buddhist line that there is no knowing self, The importance of this system in India can only knowledge. Hence, praman¯ . a should not be seen in the continuous references to it be taken in an instrumental sense, but as by later Buddhist thinkers and the numer- referring to the knowledge-event, the word ous attacks it received from orthodox Hindu itself being then interpreted as meaning thinkers. It gradually came to dominate the valid cognition. This type of cognition is in Indian Buddhist tradition, even eclipsing the turn defined as that cognition that is non- Abhidharma as the prime focus of intellec- deceptive (avisam. vadi¯ -jnãna¯ ): tual creativity. The concern of this tradition is the nature Valid cognition is that cognition [that of knowledge. In the Indian context, this is] non-deceptive (avisam. v¯adi). Non- issue is formulated as this question: What deceptiveness [consists] in the readiness P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

102 the cambridge handbook of consciousness

[for the object] to perform a function appears to it and in the process reveals the (Dharmak¯ırti, Commentary on Valid object that is apprehended. In this way, Cognition II: 1, translated by Dreyfus, in each mental state cognizes its object. But as Miyasaka, 1971–2 ). an epistemologist, Dharmak¯ırti investigates issues left out by the Abhidharma, tackling This statement emphasizes that praman¯ . a questions that are central to any philosophi- is not the instrument that a knowing self cal exploration of the mind. In this chapter, uses to know things. There is no separate we examine some of these questions. First, knowing subject, but just knowledge, which we consider Dharmak¯ırti’s analysis of the is praman¯ . a. According to this account, nature of cognitive events. We examine his a cognition is valid if, and only if, it is view of the mind as apprehending represen- non-deceptive. Dharmak¯ırti in turn inter- tations of external objects, rather than the prets non-deceptiveness as consisting of an objects themselves, and the consequences object’s readiness to perform a function that that this view has for the issue of whether the relates to the way it is cognized. For example, mind is inherently reflexive (self-revealing the non-deceptiveness of a fire is its dispo- and self-aware). We also examine Dhar- sition to burn, and the non-deceptiveness of mak¯ırti’s theory of perception, as well as its perception is its apprehension as burning. some of his views on the nature of concep- This perception is non-deceptive because it tuality and its relation to language. Finally, practically corresponds to the object’s own we revisit the issue of intentionality, showing causal dispositions, contrary to the appre- the complexity of this notion and attempting hension of the fire as cold. to disentangle its several possible meanings The scope of the discussion of praman¯ . a, within the context of a Buddhist account of however, is not limited to the analysis of the mental. knowledge, but constitutes a veritable philosophical method used in investigating other The Reflexive Nature of Mental Events philosophical and even metaphysical topics. All pronouncements about the world and We commonly assume that we have unprob- our ways of knowing it must rest on some lematic access to our environment through attested forms of knowledge, such as percep- our senses. Even casual first-person investi- tion and inference, if they are to be taken gation shows, however, that such access may seriously. No one can simply claim truth, well not be the case. There are cases of per- but must be able to establish statements ceptual illusions, and even when we are not by pinning down their epistemic supports. deceived, the perceptions of individuals vary The advantage of this method is that it pro- greatly. Hence philosophy cannot take for vides intertraditional standards of validation granted the common-sense view of percep- and the development of a relatively neutral tual knowledge. Many Western philosophers framework within which philosophical and have argued that our perceptual knowledge metaphysical claims can be assessed, with- goes well beyond the sensible experiences out regard to religious or ideological back- that give rise to it. Although this claim is grounds. This procedure is different from the debatable, we cannot assume without exam- Abhidharmic approach, which presupposes ination that we understand the way in which Buddhist ideas and vocabulary. cognition apprehends its objects. In analyzing the mind, Dharmak¯ırti starts In thinking about the nature of cognition, from the same view of mind as the Abhid- Dharmak¯ırti relies crucially on the concept harma. Mind is made up of momentary men- of aspect (ak¯ ara¯ ), a notion that goes back tal states that arise in quick succession. Each to the S¯am. khya, but has been accepted by moment of consciousness comes to be and several other schools. The idea behind this disappears instantaneously, making a place position, which is called in Indian philos- for other moments of awareness. Moreover, ophy sak¯ arav¯ ada¯ (‘assertion of aspect’), is each moment apprehends the object that that cognition does not apprehend its object P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

asian perspectives: indian theories of mind 103

nakedly, but rather through an aspect, which the beholding of an internal representation. is the reflection or imprint left by the object From one side, consciousness has an exter- on the mind. For example, a visual sense con- nally oriented feature, called the objective sciousness does not directly perceive a blue aspect (grahy¯ ak¯ ara¯ ). This feature is the form color, but captures the likeness of blue as that a mental state assumes under the influ- imprinted on cognition. Thus, to be aware ence of an external object. The second side of an object does not mean apprehending is the internal knowledge of our own men- this object directly, but having a mental tal states. It is called the subjective aspect state that has the form of this object and (grahak¯ ak¯ ara¯ ), the feature that ensures that being cognizant of this form. The aspect is we are aware of the objective aspect, the rep- the cognitive form or epistemic factor that resentation of the object. These two parts allows us to distinguish mental episodes and do not exist separately. Rather, each mental differentiate among our experiences. With- state consists of both and hence is necessar- out aspects, we could not distinguish, for ily reflexive (aware of itself in being aware instance, a perception of blue from a percep- of its object). tion of yellow, for we do not perceive yellow The necessary reflexivity of consciousness directly. The role of the aspect is thus cru- is understood by Dharmak¯ıirti and his fol- cial in Dharmak¯ırti’s system, for it explains lowers as a particular type of perception a key feature of consciousness: Conscious- called self-cognition (svasam. vedana). Self- ness is not the bare seeing that direct real- cognition can be compared to what Western ism and common sense suppose, but rather philosophers call apperception; namely, the the apprehension of an aspect that repre- knowledge that we have of our own men- sents this object in the field of conscious- tal states. It is important to keep in mind, ness. The aspect is not external to conscious- however, that apperception does not imply ness. It is not only the form under which an a second and separate cognition directed external object presents itself to conscious- toward a given mental state of which one ness but also the form that consciousness is thereby aware. For Dharmak¯ırti, apper- assumes when it perceives its object. Thus ception is not introspective or reflective, for an aspect is a representation of objects in it does not take inner mental states as its consciousness, as well as the consciousness objects. It is instead the self-cognizing fac- that sees this representation. tor inherent in every mental episode, which The implication of this analysis is that provides us with a non-thematic awareness perception is inherently reflexive. Aware- of our mental states. For Dharmak¯ırti, reflex- ness takes on the form of an object and ivity is a necessary consequence of his analy- reveals that form by assuming it. Thus, in sis of perception, according to which a sub- the process of revealing external things, cog- jective aspect beholds an objective aspect nition reveals itself. This view of cogni- that represents the external object within tion as ‘self-luminous’ (svayam praka´¯sa) and the field of consciousness. Self-cognition is self-presencing is not unique to Dign¯aga, nothing over and above this beholding. its first Buddhist propounder, or to Dhar- Self-cognition is the intuitive presence mak¯ırti, his follower. It is also accepted that we feel we have toward our own men- by other thinkers, particularly the Hindu tal episodes. We may not be fully aware of Vedantins, ¯ who identify consciousness as all the aspects and implications of our expe- the self and describe it as being ‘only riences, but we do seem to keep track of known to itself’ (svayaµvedya) and ‘self- them. Tibetan scholars express this idea by effulgent’ (svayamprabha; see Gupta 1998, saying that there is no person whose men- 2003; Mayeda, 1979/1992,pp.22, 44). For tal states are completely hidden to him- or Dign¯aga and Dharmak¯ırti, however, the herself. This limited self-presence is not due inherently reflexive character of conscious- to a metaphysical self, but to self-cognition. ness is not a consequence of its transcen- Because apperception does not rely on rea- dent and pure nature, but of its consisting of soning, it is taken to be a form of perception. P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

104 the cambridge handbook of consciousness

Apperception does not constitute, however, mental states. For Dharmak¯ırti, as for the a separate reflective or introspective cogni- Abhidharma, suffering and happiness are tion. Otherwise, the charge that the notion not external to consciousness, but integral of apperception opens an infinite regress to our awareness of external objects. Our would be hard to avoid. perceptions arise with a certain feeling-tone, Dharmak¯ırti’s ideas are not unlike those be it pleasant, unpleasant, or neutral; this Western philosophers who have argued that feeling-tone is a function of the presence consciousness implies self-consciousness of the mental factor of feeling as described (see Chapters 3 and 4). Such philosophers by the Abhidharma. This feeling needs to include (despite their otherwise vast dif- be noticed, however; otherwise we would ferences) Aristotle, Descartes, Locke, Kant, not be aware of how the apprehension Husserl, and Sartre (see Wider, 1997, of the object feels. Because this noticing pp. 7–39). According to Locke, a person is cannot be the function of another men- conscious of his or her own mental states. tal state without incurring the problem of He defines consciousness as “the perception an infinite regress, it must be the mental of what passes in a man’s mind” (Essay state apprehending the external object that Concerning Human Understanding II: ii, becomes aware at the same time of the 19). Leibniz, in his New Essays Concerning feeling. This conclusion indicates, for Dhar- Human Understanding (II: i, 19), criticizes mak¯ırti, the dual nature of mental states. Locke, pointing out that this view leads to In a single mental state, two aspects can an infinite regress, for if every cognitive act be distinguished: (1) the objective aspect, implies self-awareness, self-knowledge must the representation of the external object in also be accompanied by another awareness, consciousness, and (2) the subjective aspect, and so on ad infinitum. This regress arises, the apprehension of this appearance or however, only if knowledge of one’s mental self-cognition. states is assumed to be distinct from knowl- For Dharmak¯ırti, a mental state thus has edge of external objects. This assumption two functions. It apprehends an external is precisely what Dharmak¯ırti denies. A object (alambana¯ ) and beholds itself. The consciousness is aware of itself in a non-dual apprehension of an external object is not way that does not involve the presence of direct, but results from the causal influence a separate awareness of consciousness. The of the object, which induces cognition to cognizing person simply knows that he or experience (anubhava) the object’s repre- she cognizes without the intervention of a sentation. Hence, mind does not experience separate perception of the cognition. This an external object, but beholds an internal knowledge is the function of apperception, representation that stands for an external which thus provides an element of certainty object. Cognition cannot be reduced to a with respect to our mental states. Apper- process of direct observation, but involves ception does not necessarily validate these a holding of an inner representation. This states, however. For example, one can take beholding is not, however, an apprehension oneself to be seeing water without knowing in the usual sense of the word, for the two whether that seeing is veridical. In this case, aspects of a single mental episode are not one knows that one has an experience, but separate. It is an ‘intimate’ contact, a direct one does not know that one knows. The experiencing of the mental state by itself determination of the validity of a cognition through which we become aware of our is not internal or intrinsic to that cogni- mental states at the same time as we per- tion, but is to be established by practical ceive things. investigation. Several arguments are presented by Dhar- Theory of Perception mak¯ırti to establish the reflexive nature of consciousness.17 One of his main arguments This view of cognition as bearing only indi- concerns the nature of suffering and hap- rectly on external objects has obvious con- piness as it reveals the deeper nature of sequences for the theory of perception. The P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

asian perspectives: indian theories of mind 105

theory of perception is an important element ring to argue on the basis of the commonsen- of Dharmak¯ırti’s epistemology, for we have sical assumption that external objects exist. access to external reality first and foremost His theory of perception thus has a pecu- through perception, the primary valid cog- liar two-tiered structure, in which he pre- nition. But this access is not as unprob- supposes the existence of external objects, lematic as one might think. Although it which he then ultimately rejects to pro- might seem commonsensical that percep- pound a form of idealism. tion results from our encounter with the Among these two tiers, the one Dharm- world, in reality consciousness does not ak¯ırti most often refers to is the Sautr¯antika directly cognize the object, but only indi- representationalist theory of perception. rectly cognizes it. For Dharmak¯ırti, as we According to this view, consciousness does have seen, the mind has direct access only not have direct access to external objects, to the representational aspect caused by the but grasps objects via the intermediary of object; the object itself remains inaccessi- an aspect caused by and similar to an exter- ble to consciousness. The similarity between nal object. He sometimes replaces this view object and aspect – and hence between by a Yog¯ac¯ara view, which holds that inter- object and consciousnesss, the aspect being nal impressions are not produced by exter- the cognitive form of the object that stands nal objects, but by internal tendencies. This for the object in the field of consciousness – shift into full-blown idealism allows Dhar- is the crucial element in this causal the- mak¯ırti to bypass the difficulties involved ory of perception. This similarity ensures in explaining the relation between internal that perception is not locked up in its own perceptions and external objects. Because appearances, as conceptions are. Conscious- there are no external objects, the problem ness is not in direct contact with the exter- of the relation between internal impres- nal world, but only with an internal impressions and external objects does not arise. At sion caused by the external object. Hence this level, his philosophy of perception can the external object remains hidden, though be described as phenomenalist, for it holds not completely. that there is no external object outside of When pressed by these problems, Dhar- aspects. mak¯ırti sometimes shifts between the views Another major feature of Dharmak¯ırti’s of two different Buddhist philosophical account is his sharp separation between schools, using one perspective to bypass perception and conception, a separation problems that arise in the other. These two enshrined in his definition of perception as views are the Sautr¯antika theory of per- the cognition that is unmistaken (abhranta¯ ) ception, which is representationalist in the and free from conceptions (kalpanapod¯ .ha) ways just described, and the Yog¯ac¯ara the- (Commentary on Valid Cognition, III: 300 ory, which is idealist and denies that there cd). Because perception is unmistaken and is anything outside of consciousness. Fol- conception is mistaken, perception must be lowing Dign¯aga’s example and his strategy free from conception. This analysis of per- of ascending scales of philosophical analysis, ception differs sharply from the dominant Dharmak¯ırti holds that the Yog¯ac¯ara theory account in India, the epistemological realis truer and hence higher on the scale of anal- ism of the Ny¯aya school and its assertion of ysis. This theory denies that there are any the existence of a determinate (savikalpaka) external objects over and above the direct form of perception. For the Ny¯aya, percep- objects of perception. Thus its view of per- tion does not stop with the simple taking ception is phenomenalist: It reduces external in of sensory stimuli, but also involves the objects to interpreted mental data, but such ability to categorize this input. Although data are no longer taken to stand for external we may start with a first moment of inde- objects (because it is now held that nothing terminate perception, in which we merely exists outside of consciousness). This theory, take in external reality, we do not stop there however, is counter-intuitive, and so Dhar- but go on to formulate perceptual judg- mak¯ırti refers to it only occasionally, prefer- ments. Moreover, and this is the crux of the P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

106 the cambridge handbook of consciousness

question, these judgments are for the Ny¯aya sions, which by themselves are not forms fully perceptual. They are not mistaken of knowledge but become so only when conceptual overlays, but true reflections of they are integrated within our categorical reality. schemes. For example, when we are hit on This commonsensical view of perception the head, we first have an impression. We is not acceptable to Dharmak¯ırti, for it leads just have a sensation of pain, which is not to an unenviable choice: either accept the by itself cognitive. This sensation becomes reality of the abstract entities necessary for cognitive when it becomes integrated into a the articulation of the content of percep- conceptual scheme, in which it is explained tion or reject the possibility of an unmis- as being an impact on a certain part of our taken cognition. Because neither possibil- body due to certain causes. It is only then ity is acceptable for Dharmak¯ırti, he holds that the impression of being hit becomes that perception can only be non-conceptual. fully intentional. Prior to this cognitive inte- There is no determinate perception, for the gration, the impression, or to speak Dhar- judgments induced by perception are not mak¯ırti’s language, the aspect, does not yet perceptual, but are just conceptual superim- represent anything in the full sense of the positions. They do not reflect the individual word. It only becomes so when interpreted reality of phenomena, but instead address conceptually. their general characteristics. Because those This view of perception agrees with Dhar- are only constructs, the cognitions that con- mak¯ırti’s analysis of the validity of cogni- ceive them cannot be true reflections of real- tions, which consists in their being ‘non- ity. Hence for perception to be undistorted deceptive’, a term interpreted in practical in a universe of particulars, it must be totally terms. Cognitions are valid if, and only if, free from conceptual elaborations. This posi- they have the ability to lead us toward suc- tion implies a radical separation between cessful practical actions. In the case of per- perception, which merely holds the object as ception, however, practical validity is not as it is in the perceptual ken, and interpretation straightforward as one might think. Achiev- of this object, which introduces conceptual ing practical purposes depends on correctly constructs into the cognitive process. describing the objects we encounter. It is not This requirement that perception be non- enough to see an object that is blue; we must conceptual is the cornerstone of the Bud- also see it as being blue. To be non-deceptive, dhist theory of perception. But it creates a cognition depends on the appropriate iden- problems for Dharmak¯ırti. It would seem tification of the object as being this or that given his privileging of perception he that. Perceptions, however, do not identify should hold an empiricist view, according their objects, for they are not conceptual. to which perception boils down to a bare They cannot categorize their objects, but encounter with reality and knowledge is only hold them without any determination. given to the senses. Dharmak¯ırti should hold Categorization requires conceptual thought the view that the aspects through which we under the form of a judgment. Such a judg- come to perceive reality are fully represen- ment subsumes its object under an appro- tational like Locke’s ideas, that they stand priate universal, thereby making it part of for external objects, and that their appre- the practical world where we deal with long- hension is in and of itself cognitive. Dhar- lasting entities that we conceive of as parts mak¯ırti’s view of perception, however, is of a determined order of things. For exam- more complex, for he shares with Sellars ple, we sense a blue object that we catego- (1956) the recognition that knowledge, even rize as blue. The perceptual aspect (the blue at the perceptual level, does not boil down aspect) is not yet a representation in the full to an encounter with reality, but requires sense of the word, because its apprehension, active categorization. We do not know things the perception of blue, is not yet cognitive. by sensing them, for perception does not It is only when it is interpreted by a concep- deliver articulated objects, but only impres- tion that the aspect becomes a full-fledged P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

asian perspectives: indian theories of mind 107

intentional object standing for an external given’. Locke, for example, holds that con- object. Hence, Dharmak¯ırti’s account of cepts and words are linked through associ- perception leads us to realize the impor- ation. The word ‘tree’ acquires its meaning tance of categorical interpretation in the for- by becoming connected with the idea tree, mation of perceptual knowledge, a position which is the mental image of a tree. Hence that is not without problems for his system, for Locke the representation of the tree is given his emphasis on the primacy and non- not formed through language, but is given conceptuality of perception. Nevertheless, to sensation (Dharmak¯ırti’s perception). We the merit of this analysis is that it disentan- understand a tree as a tree through mere gles the processes through which we come acquaintance with its representation with- to know the world, explaining the role of out recourse to concepts. Dharmak¯ırti’s phi- perception as a way to contact the world losophy is quite different, for it emphasizes while emphasizing the role of conceptual the constitutive and constructive nature of categorization in the formation of practical language. This conception of language is well knowledge. captured by one of Dharmak¯ırti’s deﬁnitions of thought:

Thought and Language Conceptual cognition is that consciousness In examining thought (kalpana¯), Dhar- in which representation (literally, appearance) is fit to be associated which words mak¯ırti postulates a close association with (Ascertainment of Valid Cognition 40: language. In fact, the two can be considered 6–7, in Vetter, 1966). equivalent from an epistemological point of view. Language signifies through conceptual Thought identifies its object by associat- mediation in the same way that thought con- ing the representation of the object with a ceives of things. The relation between the word. When we conceive of an object we do two also goes the other way: We do not first not apprehend it directly, but through the understand things independently of linguis- mediation of its aspect. Mediation through tic signs and then communicate this under- an aspect also occurs with perception, but standing to others. Dharmak¯ırti recognizes a here the process of mediation is different. cognitive import to language; through lan- In the case of perception there is a direct guage we identify the particular things we causal connection between the object and encounter, and in this way we integrate the its representation, but no such link exists object into the meaningful world we have for thought. There is no direct causal link constructed. The cognitive import of lan- between the object and thought, but rather guage is particularly obvious in the acqui- an extended process of mediation in which sition of more complex concepts. In these linguistic signs figure prominently. cases, it is clear that there is nothing in For Dharmak¯ırti, the starting point of this experience that could possibly give rise to process is our encounter with a variety of these concepts without language. Without objects that we experience as being simi- linguistic signs thought cannot keep track of lar or different. We construct concepts in things to any degree of complexity. Dhar- association with linguistic signs to capture mak¯ırti also notes that we usually remember this sense of experienced similarity and dif- things by recollecting the words associated ference. This linguistic association creates a with those things. Thus concepts and words more precise concept in which the represen- mutually depend on each other. tations are made to stand for a commonality This close connection between thought that the objects are assumed to possess. For and language, inherited from Dign¯aga, example, we see a variety of trees and appre- differentiates Dharmak¯ırti from classical hend a similarity between these objects. At empiricists, such as Locke and modern this level, our mental representations have sense-data theorists, who believe in what yet to yield a concept of tree. The con- Sellars (1956) describes as the ‘myth of the cept of tree is formed when we connect our P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

108 the cambridge handbook of consciousness

representations with a socially formed and plex topic is beyond the scope of this chap- communicated sign and assume that they ter. Suffice it to say that the apoha theory stand for a treeness that we take individ- is a way to explain how language signifies ual trees to share. In this way experiences in a world of individuals. Linguistic mean- give rise to mental representations, which ing poses a particularly acute problem for are transformed into concepts by associa- Dign¯aga and Dharmak¯ırti, for they are com- tion with a linguistic sign. The formation mitted to a connotationist view of language, of a concept consists of the assumption that in which sense has primacy over reference. mental representations stand for an agreed- Such a view, however, is difficult to hold in upon imagined commonality. Thus concepts a nominalist ontology that disallows abstract come to be through the conjunction of the entities, such as meaning.18 experience of real objects and the social pro- The apoha theory tries to solve this cess of language acquisition. Concept forma- conundrum by arguing that language does tion is connected to reality, albeit in a medi- not describe reality positively through uni- ated and highly indirect way. versals, but negatively by exclusion. Lan- But concept formation is also mistaken, guage is primarily meaningful, but this does according to this view. A concept is based not mean that there are real senses. Rather, on the association of a mental representa- we posit agreed-upon fictions that we con- tion with a term that enables the repre- struct for the sake of categorizing the world sentation to stand for a property assumed according to our purposes. Thus ‘cow’ does to be shared by various individuals. In not describe Bessie through the mediation of Dharmak¯ırti’s nominalist world of indi- a real universal (cowness), but by excluding viduals, however, things do not share a a particular (Bessie) from the class of non- common property; rather, the property is cow. Matilal describes Dign¯aga’s view this projected onto them. The property is man- way: ufactured when a representation is made to stand for an assumed commonality, which a Each name, as Dignaga¯ understands, variety of individuals are mistakenly taken dichotomizes the universe into two: those to to instantiate. Hence this property is not which the name can be applied and those real; it is merely a pseudo-entity superim- to which it cannot be applied. The function of a name is to exclude the object from the posed (adhyaropa¯ ) on individual realities. class of those objects to which it cannot be This property is also not reducible to a gen- applied. One might say that the function of eral term. In other words, the commonality a name is to locate the object outside of the that we project onto things does not reside class of those to which it cannot be applied in using the same term to designate discrete (Matilal, 1971,p.45). individuals. Upon analyzing the notion of sameness of terms, we realize that identifying Although linguistic form suggests that we individual terms as being the same presup- subsume an individual under a property, poses the concept of sameness of meaning, analysis reveals that words merely exclude in relation to which the individual terms can objects from being included in a class to be identified. Thus commonality is not due which they do not belong. The function simply to a term, but requires the forma- of a name is to locate negatively an object tion of concepts on the basis of the mistaken within a conceptual sphere. The impression imputation of commonality onto discrete that words positively capture the nature of individuals. objects is misleading. What does it mean, however, for a This theory was immediately attacked by concept to be based on an assumed Hindu thinkers, such as Kum¯arila and Uddy- commonality? Here Dharmak¯ırti’s theory otakara, who raised strong objections. One must be placed within its proper context, of them was that this theory is counter- the apoha or exclusion theory of language, intuitive, because we do not perceive which was created by Dign¯aga. This com- ourselves to eliminate non-cows when we P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

asian perspectives: indian theories of mind 109

conceive of cows. Dharmak¯ırti’s theory of This grounding in perception ensures that, concept formation is in many ways an although conception is mistaken in the way attempt to answer these attacks. It argues reviewed above, it is neither baseless nor ran- that the apoha theory is not psychological, dom and hence can lead to the formation of but epistemological. In conceiving of objects concepts that will be attuned to the causal we do not directly eliminate other objects, capacities of particulars. but instead rely on a representation that is made to stand in for an assumed common- Dharmak¯ırti and Abhidharma: ality shared by several particulars. It is this Intentionality Revisited fictional commonality that is the result of an exclusion. There is nothing over and above Dharmak¯ırti’s analysis has in certain respects particulars, which are categorized on the a great deal of continuity with the Abhid- basis of their being excluded from what they harma. Both view the mind as constituted by are not. The concept that has been formed a succession of mental states in accordance in an essentially negative way is projected with their ontological commitments, which onto real things. In the process of making privilege the particular over the general. judgments such as ‘this is a tree’, the real Reality is made up of a plurality of elements differences that exist between the different (here moments of awareness), and general- trees come to be ignored and the similarities ity, when it is not a figment of our imag- are reified into a common universal property, ination, is at best the result of aggregation. which is nothing but a socially agreed-upon This emphasis on the particular derives from fiction. the central tenets of the Buddhist tradition; The eliminative nature of thought and namely, non-substantiality and dependent language is psychologically revealed when origination. In Dharmak¯ırti’s epistemologi- we examine the learning process. The word cal approach, this emphasis expresses itself ‘cow’, for instance, is not learned only in valuing perception over conception, and through a definition, but by a process of in the problematic but necessary coopera- elimination. We can give a definition of tion between the two forms of cognition. We ‘cow’, but the definition works only if its ele- do not come to know things by merely comments are known already. For example, we ing across them, but by integrating them into can define cows as animals having dewlaps, our conceptual schemes on the basis of our horns, and so on (the traditional definition of experiences. ‘cow’ in Indian philosophy). But how do we One question raised by this analysis con- know what counts as a dewlap? Not just by cerns intentionality. The Abhidharma tra- pointing to the neck of a cow, but by elimi- dition had assumed all along that cogni- nating the cases that do not fit. In this way, tions were intentional, but did not provide we establish a dichotomy between those ani- a systematic analysis of intentionality. Dhar- mals that fit, and other animals or things mak¯ırti fills this gap, analyzing the way in that do not, and on the basis of this negative which various types of cognition bear on dichotomy we construct a fictive property, their objects. But because he makes a sharp cowness. This construction is not ground- distinction between perception and concep- less, however, but proceeds through an indi- tion, his analysis does not yield a single rect causal connection with reality. Concepts concept of intentionality, but on the con- are not formed a priori, but elaborated as trary leads us to realize that this central a result of experiences. Dharmak¯ırti’s solu- notion may have to be understood in mul- tion to the problem of thought and mean- tiple ways. The cognitive process starts with ing is thus to argue that in a world bereft our encounter with the world through per- of real abstract entities (properties), there ceptions, but this encounter is not enough are only constructed intensional (linguistic) to bring about knowledge. Only when we pseudo-entities, but that this construction are able to integrate the objects deliv- is based on experience; that is, perception. ered through the senses into our categorical P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

110 the cambridge handbook of consciousness

schemes can we be said to know them in the the flow of thoughts gradually reappears, full sense of the word. Hence, if we under- and we are able to make judgments about stand intentionality as cognitive – that is, as what we saw during our meditation. One is pertaining to knowledge – we may well have then also able to make a clear differentiation to agree with Dharmak¯ırti that perception between the products of thoughts and the is not in and of itself fully intentional. Only bare delivery of the senses and to distinguish when perception is coordinated with con- cognitive from phenomenal intentionality. ception does it become intentional; hence it The analysis of intentionality, however, can be said to be intentional only in a derived may have to go even further to account for sense of the word. Perception is not in and of all the forms of cognition known to Buddhist itself cognitive, but only inasmuch as it has traditions. We alluded above to the Abhid- the ability to induce conceptual interpreta- harmic idea of a store-consciousness, a sub- tions of its objects. This does not mean, how- liminal form of cognition that supports all ever, that perception is completely blank or the propensities, habits, and tendencies of a purely passive. It has an intentional func- person. Although such a store-consciousness tion, that of delivering impressions that we is usually asserted by the Yog¯ac¯ara to sup- take in and organize through our conceptual port their idealist view, it is known to other schemes. Hence, perception can be said to traditions under other names and hence has have a phenomenal intentionality, which to be taken seriously within a Buddhist acc- may be revealed in certain forms of medi- ount of the mind, regardless of the par- tative experiences. ticular views that are associated with it. Dharmak¯ırti alludes to such experiences But given the particularities of this form when he describes a form of meditation, in of consciousness, its integration within a which we empty our mind without closing Buddhist view of the mind is not without it completely to the external world (Com- problems. The difficulties come from the mentary on Valid Cognition III: 123–5,in fact that the store-consciousness does not Miyasaka 1971–2). In this state of liminal seem to have cognitive or even phenomenal awareness, things appear to us but we do not intentionality. Because it does not capture identify them. We merely let them be. When any feature, it cannot be said to know its we come out of this stage, the usual concep- object, like conceptions. Because it is sub- tual flow returns, and with it the concep- liminal, it is difficult to attribute to it a phe- tualization that allows us to identify things nomenal content able to induce categoriza- as being this or that. This experience shows, tion, like perceptions. How then can it be Dharmak¯ırti argues, that identification is not intentional? perceptual, but is due to conceptualization. To respond to this question would neces- In such a state, perception takes place but sitate an analysis that goes well beyond the not conceptualization. Hence, perception is purview of this chapter. Several avenues are a non-conceptual sensing onto which inter- open to us. We could argue that the store- pretations are added. consciousness is not intentional and hence Due to the speed of the mental process, that intentionality is not the defining char- the untrained person cannot differentiate acteristic of the mental, but only of certain conceptual from non-conceptual cognitions. forms of cognitions. We would then be faced It is only on special occasions, such as in some with the task of explaining the nature of the form of meditation, that a clear differentia- mental in a way that does not presuppose tion can be made. There, the flow of thought intentionality. Or we could extend the con- gradually subsides, and we reach a state in cept of intentionality, arguing that the store- which there is a bare sensing of things. In consciousness is not intentional in the usual this state, what we call shapes and colors are cognitive or phenomenal senses of the word, seen barely (i.e., as they are delivered to our but rather that its intentionality consists in senses without the adjunctions of concep- its having a dispositional ability to generate tual interpretations). When one gradually more explicit cognitive states. Some Western emerges from such a non-conceptual state, phenomenologists, notably Husserl and P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

asian perspectives: indian theories of mind 111

Merleau-Ponty, distinguish ‘object directed ing rich phenomenological investigations of intentionality’ from ‘operative intentional- various aspects of human cognition and ity’ (see Chapter 4). Whereas the former exploration of various levels and types of is what we usually mean by intentional- meditative consciousness. This tradition also ity, the latter is a non-reﬂective tacit sensi- shows, however, that it would be na¨ıve to bility, a spontaneous and involuntary level take these investigations of consciousness that makes us ready to respond cognitively as being objectively given or established. and affectively to the world, though it is Rather, they are accounts of experience that not by itself explicitly cognitive. This most are often intertwined with doctrinal formu- basic form of intentionality is important lations and hence are open to critique, revi- in explaining our openness to the world. sion, and challenge, like any other human It also seems an interesting avenue for interpretation. Indeed, these formulations exploring the cognitive nature of the store- need to be taken seriously and examined consciousness. with the kind of critical spirit and rigorous philosophical thinking exhibited by Dhar- mak¯ırti. Only then, can we do justice to the Conclusion insights of this tradition.

We can now see the richness and the complexities of the Indian Buddhist analyses of Glossary the nature of the mind and consciousness. The Abhidharma provides the basis of these Sam¯ . khya analyses, with its view of the mind as a Pradhana:¯ primordial nature or prakrti, stream of moments of consciousness and its . materiality. The primordial substance distinction between the primary factor of out of which the diversity of phe- awareness and mental factors. This tradition nomena arise. It is composed of three also emphasizes the intentional nature of qualities (guna): sattva (transparency, consciousness, the ability of consciousness to . buoyancy), rajas (energy, activity), and be about something else. As we have seen, tamas (inertia, obstruction). They are however, this concept is far from self-evident the principles or forces whose combi- and needs further philosophical clarifica- nation produces mental and material tion. This clarification is one of the impor- phenomena. tant tasks of Dharmak¯ırti’s philosophy. In accomplishing this task, Dharmak¯ırti criti- Atman: spiritual self or purus.a, per- cally explores the variety of human cogni- son. The non-material spiritual ele- tions, distinguishing the conceptual from the ment that merely witnesses the men- perceptual modes of cognition and empha- tal activities involved in the ordinary sizing the constructed nature of the for- awareness of objects. mer and its close connection with language. Buddhi: usually translated as ‘the intel- Yet, as we have also seen, this philoso- lect’. It has the ability to distinguish and phy is not always able to account for all experience objects. This ability pro- the insights of the Abhidharma, particu- vides the prereflective and presubjec- larly those concerning the deeper layers of tive ground out of which determined consciousness. mental states and their objects arise. It When we look at the Indian Buddhist is also the locus of all the fundamental tradition, we should not look for a unified predispositions that lead to these expe- and seamless view of the mind. Like any riences. other significant tradition, Indian Buddhist Aham. kara:¯ egoity or ego-sense. This is philosophy of mind is plural and animated the sense of individual subjectivity or by debates, questions, and tensions. This rich selfhood tied to embodiment, which tradition has a great deal to offer contem- gives rise to the subjective pole of porary mind science and philosophy, includ- cognition. P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

112 the cambridge handbook of consciousness

Manas: mentation. It oversees the senses there is no manifest mental activity, it and discriminates between objects. By also provides a sense of continuity for serving as an intermediary between the Theravada school, which asserts its the intellect and the senses, menta- existence. tion organizes sensory impressions and Klis.ta-manas: afflictive mentation. This objects and integrates them into a tem- is the inborn sense of self that arises poral framework created by memories from the apprehension of the store- and expectations. consciousness as being a self. From a Citta: mental activities or antah. karan.a, Buddhist point of view, however, this internal organ. This is the grouping of sense of self is fundamentally mistaken. buddhi, aham. kara¯ , and manas. It is a mental imposition of unity where Praman¯ . a: instrument of valid cognition of there is in fact only the arising of a mul- the self. The S¯am. khya recognizes three tiplicity of interrelated physical and such instruments: perception, infer- mental events. ence, and testimony. The Ny¯aya adds praman¯ . a: valid cognition. Not the instru- a fourth one, analogy. ment of a self but the knowledge-event itself. There are only two types of valid Buddhist cognition admissible in Buddhist epistemology, pratyaks.a, perception, and Citta: primary factor of awareness or anumana¯ , inference. vijñana¯ , consciousness. It is the aspect Svasamvedana: self-cognition. This is the of the mental state that is aware of the . limited but intuitive presence that we object, or the bare apprehension of the feel we have toward our own men- object. It is the awareness that merely tal episodes, which is due not to the discerns the object, the activity of cog- presence of a metaphysical self but to nizing the object. the non-thematic reflexive knowledge Caitesika: mental factor. Mental factors that we have of our own mental states. are aspects of the mental state that Because self-cognition does not rely on characterize the object of awareness reasoning, it is taken to be a form of and account for its engagement. In perception. It does not constitute, how- other words, whereas consciousness ever, a separate reflective or introspec- makes known the mere presence of the tive cognition. Otherwise, the charge object, mental factors make known the that the notion of apperception opens particulars of the content of awareness, an infinite regress would be hard to defining the characteristics and special avoid. conditions of its object. Alaya-vijñana:¯ store-consciousness. This continuously present subliminal con- Notes sciousness is posited by some of the Yogac¯ ¯ ara thinkers to provide a sense 1. Presenting the S¯am. khya view in a few lines of continuity in the person over time. is problematic given its evolution over a long It is the repository of all the basic period of time, an evolution shaped by the habits, tendencies, and propensities addition of numerous refinements and new (including those that persist from one analyses in response to the critiques of Bud- dhists and Ved¯antins. For a quick summary, life to the next) accumulated by the see Mahalingam (1977). For a more detailed individual. examination, see Larson and Bhattacharya Bhavan. ga citta: life-constituent con- (1987). sciousness. Although this conscious- 2. Contrary to Ved¯anta, the S¯am. khya holds that ness is not said to be always present and there are many individual selves rather than arises only during the moments where a universal ground of being such as Brahman. P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

asian perspectives: indian theories of mind 113

3. The notion of a pure and passive ‘witness 14. For a discussion of whether compassion and consciousness’ is a central element of many lovingkindness, seen from a Buddhist point of Hindu views about consciousness (see Gupta, view, are emotions, see Dreyfus (2002). 1998, 2003). 15. For a brief but thoughtful discussion of the 4. For a thoughtful discussion of this view of the idea of Buddhism as a psychology, see Gomez mind, see Schweizer (1993). (2004). 5. Numerous translations of Patanjali’s˜ Yoga 16. For discussion of the characteristics of Indian Sutras are available in English. logic, see Matilal (1985) and Barlingay (1975). 6. For discussion of the Advaita Ved¯anta view of On Buddhist logic, see Kajiyama (1966). For consciousness, see Gupta (2003, Chapter 5). an analysis of Dharmak¯ırti’s philosophy, see For a philosophical overview of Advaita Dreyfus (1997) and Dunne (2004). Vedanta, ¯ see Deutsch (1969). 17. For a detailed treatment of Dharmak¯ırti’s 7. For a glimpse of the origins of the Abhid- arguments and their further elaboration in the harma, see Gethin (1992). Tibetan tradition, see Dreyfus (1997,pp.338– 341 400 415 8. For Husserl, by contrast, not all conscious- , – ). ness is intentional in the sense of being object- 18. For more on this difficult topic, see Dreyfus directed. See Chapter 4 and the final section (1997) and Dunne (2004). of this chapter. 9. All quotations from this work are translated from the French by G. Dreyfus. References 10. See Rahula (1980,p.17). Although the Theravada Abhidharma does not recognize Barlingay, S. S. (1975). A modern introduction to a distinct store-consciousness, its concept of Indian logic. Delhi: National. bhava¸cga citta, the life-constituent conscious- Bodhi, B. (Ed.). (1993). A comprehensive manual ness, is similar. For a view of the complexities of Abhidharma. Seattle, WA: Buddhist Publica- of the bhava¸cga, see Waldron (2003,pp.81– tion Society. 87). Damasio, A. (1995). Descartes’ error: Emotion, 11. They are then said to be conjoined (sam- reason and the human brain. New York: Harper payutta, mtshungs ldan), in that they are Perennial. simultaneous and have the same sensory basis, 1971 the same object, the same aspect or way of de la Vallee´ Poussin, L. ( ). L’Abhidharmako˙sa apprehending this object, and the same sub- de Vasubandhu. Bruxelles: Institut Belge des stance (the fact that there can be only one Hautes Etudes Chinoises. representative of a type of consciousness and de la Vallee´ Poussin, L. (1991). Notes sur le mental factor at the same time). See Waldron moment ou ksana des bouddhistes. In H. S. (2003,p.205). Prasad (Ed.), Essays on time. Delhi: Sri Satguru. 12. This list, which is standard in the Tibetan Deutsch, E. (1969). Advaita Vedanta: A philo- tradition, is a compilation based on Asaçga’s sophical reconstruction. Honolulu: University Abhidharma–samuccaya. It is not, how- Press of Hawaii. ever, Asaçga’s own list, which contains 52 Dreyfus, G. (1997). Recognizing reality: Dhar- items (Rahula 1980,p.7). For further dis- mak¯ırti’s philosophy and its Tibetan interpreta- cussion, see Napper (1980) and Rabten tions. Albany, NY: State University of New (1978/1992). For the lists of some of the York Press. other traditions, see Bodhi (1993,pp.76– Dreyfus, G. (2002). Is compassion an emotion? A 79) and de la Vallee´ Poussin (1971, II: 150– cross-cultural exploration of mental typologies. 178). In R. Davidson & A. Harrington (Eds.), Visions 13. Although some of these states may be sote- of compassion: Western scientists and Tibetan riologically significant and involve the abil- Buddhists examine human nature (pp. 31–45). ity to transcend duality, not all need be. The Oxford: Oxford University Press. practice of concentration can involve signless Dunne, J. D. (2004). Foundations of Dhar- meditative states, and so too does the practice makjr¯ıti’s philosophy. Boston: Wisdom. of some of the so-called formless meditative Gethin, R. (1992). The M¯atrik¯as: Memorization, states. mindfulness and the list. In J. Gyatso (Ed.), In P1: JzG 0521857430c05 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 5, 2007 5:25

114 the cambridge handbook of consciousness

the mirror of memory (pp. 149–172). Albany, NY: Miyasaka, Y. (Ed.) (1971–2). Pramanavarttika- State University of New York Press. karika. Acta Indologica 2. Goleman, D. (2003). Destructive emotions. A sci- Napper, E. (1980). Mind in Tibetan Buddhism. entiﬁc dialogue with the Dalai Lama. New York: Ithaca, NY: Snow Lion. Bantam. Poppel,¨ E. (1988). Mindworks: Time and conscious Gomez, L. (2004). Psychology. In R. Buswell experience. Boston: Harcourt Brace Jovanovich. (Ed.), Encyclopedia of Buddhism (pp. 678–692). Rabten, G. (1992). The mind and its functions. Mt. New York: MacMillan. Pelerin:´ Rabten Choeling. Original work pub- Guenther, H. (1976). Philosophy and psychology lished 1978. in the Abhidharma. Berkeley, CA: Shambala Rahula, W. (1980). Le Compendium de la Super- 1976 Press, . Doctrine de Asa¸cga. Paris: Ecole Franc¸aise Gupta, B. (1998). The disinterested witness. A d’Extreme-Orient.ˆ fragment of Advaita Vedanta phenomenology. Schweizer, P. (1993). Mind/consciousness Dual- Evanston, IL: Northwestern University Press. ism in Samkhya-Yoga philosophy. Philoso- Gupta, B. (2003). Cit. consciousness. New Delhi: phy and Phenomenological Research, 53, 845– Oxford University Press. 859. 1981 James, W. ( ). Principles of psychology. Cam- Sellars, W. (1956). Empiricism and the philoso- bridge, MA: Harvard University Press. phy of mind. In H. Feigl & M. Scriven (Eds.), Kajiyama, Y.(1966). Introduction to Buddhist logic. Minnesota studies in the philosophy of science. Kyoto: Kyoto University. Vol. 1: The foundations of science and the con- Larson, J., & Bhattacharya, R. S. (1987). Encyclo- cepts of psychology and psychoanalysis (pp. 253– pedia of Indian philosophies: Sam¯ . khya, A dualist 329). Minneapolis, MN: University of Min- tradition in Indian philosophy. Delhi: Motilal. nesota Press. Mahalingam, I. (1977). S¯am. khya-Yoga. In B. Carr Varela, F. J., Thompson, E., & Rosch, E. & I. Mahalingam (Eds.), Companion encyclo- (1991). The embodied mind: Cognitive science pedia of Asian philosophy. London: Routledge and human experience. Cambridge, MA: MIT Press. Press. Matlilal, B. K. (1971). Epistemology, logic, and Vetter, T. (1966). Dharmak¯ırti’s Pramanavinis- grammar in Indian philosophical analysis. The cayah 1. Kapitel: Pratyaksam. Vienna: Osterrei-¨ Hague: Mouton. chische Akademie der Wissenschaften. Matilal, B. K. (1985). Logic, language, and reality. Waldron, W. (2003). The Buddhist unconscious. Delhi: Matilal Banarsidas. London: Routledge Press. Mayeda, S. (1992). A thousand teaching: The Wider, K. (1997). The bodily nature of Upade˙sashasri¯. Albany, NY: State University consciousness: Sartre and contemporary philos- of New York Press. Original work published ophy of mind. Ithaca, NY: Cornell University 1979. Press. P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

B. Computational Approaches to Consciousness

115 P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

116 P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

CHAPTER 6 Artiﬁcial Intelligence and Consciousness

Drew McDermott

Abstract Introduction

Consciousness is only marginally relevant Computationalism is the theory that the to artificial intelligence (AI), because to human brain is essentially a computer, most researchers in the field other prob- although presumably not a stored-program, lems seem more pressing. However, there digital computer like the kind Intel makes. have been proposals for how consciousness Artificial intelligence (AI) is a field of com- would be accounted for in a complete computer science that explores computational putational theory of the mind from such models of problem solving, where the prob- theorists as Dennett, Hofstadter, McCarthy, lems to be solved are of the complex- McDermott, Minsky, Perlis, Sloman, and ity of those solved by human beings. An Smith. One can extract from these spec- AI researcher need not be a computa- ulations a sketch of a theoretical synthe- tionalist because he or she might believe sis, according to which consciousness is the that computers can do things that brains property a system has by virtue of mod- do non-computationally. However, most AI eling itself as having sensations and mak- researchers are computationalists to some ing free decisions. Critics such as Harnad extent, even if they think digital computers and Searle have not succeeded in demolish- and brains-as-computers compute things in ing a priori this or any other computational different ways. When it comes to the prob- theory, but no such theory can be verified lem of phenomenal consciousness, however, or refuted until and unless AI is successful the AI researchers who care about the prob- in finding computational solutions to diffi- lem and believe that AI can solve it are a tiny cult problems, such as vision, language, and minority, as shown in this chapter. Nonethe- locomotion. less, because I count myself in that minority,

117 P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

118 the cambridge handbook of consciousness

I do my best here to survey the work of its the word “conscious,” but I am talking here members and defend a version of the theory about the “Hard Problem” (Chalmers, 1996), that I think represents that work fairly well. the problem of explaining how it is that a Perhaps calling computationalism a the- physical system can have vivid experiences ory is not exactly right here. One might with seemingly intrinsic “qualities,” such as prefer calling it a working hypothesis, the redness of a tomato or the spiciness of assumption, or dogma. The evidence for a taco. These qualities usually go by their computationalism is not overwhelming, and Latin name, qualia. We all know what we some even believe it has been refuted by a are talking about when we talk about sensa- priori arguments or empirical evidence. But, tions, but they are notoriously undefinable. in some form or other, the computational- We all learn to attach a label such as “spicy” ist hypothesis underlies modern research in to certain tastes, but we really have no idea cognitive psychology, linguistics, and some whether the sensation of spiciness to me is kinds of neuroscience. That is, there would the same as the sensation of spiciness to you. not be much point in considering formal or Perhaps tacos produce my “sourness” in computational models of mind if it turned you, and lemons produce my “spiciness” in out that most of what the brain does is not you.2 We would never know, because you computation at all, but, say, some quantum- have learned to associate the label “sour” mechanical manipulation (Penrose, 1989). with the quale of the experience you have Computationalism has proven to be a fer- when you eat lemons, which just happens to tile working hypothesis, although those who be very similar to the quale of the experi- reject it typically think of the fertility as sim- ence I have when I eat tacos. We can’t just ilar to that of fungi or of pod people from tell each other what these qualia are like; outer space. the best we can do is talk about compar- Some computationalist researchers be- isons. But we agree on such questions as, Do lieve that the brain is nothing more than a tacos taste more like Szechuan chicken or computer. Many others are more cautious more like lemons? I focus on this problem and distinguish between modules that are because other aspects of consciousness raise quite likely to be purely computational (e.g., no special problem for computationalism, as the vision system) and others that are less opposed to cognitive science generally. likely to be so, such as the modules, or princi- The purpose of consciousness, from an ples of brain organization, that are responsi- evolutionary perspective, is often held to ble for creativity or for romantic love. There have something to do with the allocation and is no need, in their view, to require that organization of scarce cognitive resources. absolutely everything be explained in terms For a mental entity to be conscious is for it of computation. The brain could do some to be held in some globally accessible area things computationally and other things by (Baars, 1988, 1997). AI has made contribu- different means, but if the parts or aspects tions to this idea, in the form of specific ideas of the brain that are responsible for these about how this global access works, going various tasks are more or less decoupled, under names such as the “blackboard model” we could gain significant insight into the (Hayes-Roth, 1985), or “agenda-based con- pieces that computational models are good trol” (Currie & Tate, 1991). One can evalu- for and could then leave the other pieces to ate these proposals by measuring how well some other disciplines, such as philosophy they work or how well they match human and theology.1 behavior. But there does not seem to be any Perhaps the aspect of the brain that is philosophical problem associated with them. most likely to be exempt from the compu- For phenomenal consciousness, the sit- tationalist hypothesis is its ability to pro- uation is very different. Computationalism duce consciousness; that is, to experience seems to have nothing to say about it, simply things. There are many different meanings of because computers do not have experiences. P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

artiﬁcial intelligence and consciousness 119

I can build an elaborate digital climate- how the leaders of our field view the fol- control system for my house, which keeps lowing problem: Create a computer or pro- its occupants at a comfortable temperature, gram that has “phenomenal consciousness,” but the climate-control system never feels that is, the ability to experience things. By overheated or chilly. Various physical mech- “experience” here I mean “qualitative expe- anisms implement its temperature sensors in rience,” the kind in which the things one senses seem to have a definite but indescrib- various rooms. These sensors produce signals able quality, the canonical example being that go to units that compute whether to “looking red” as opposed to “looking green.” turn on the furnace or the air conditioner. Anyway, please choose from the following The result of these computations causes possible resolutions of this problem: switches to close so that the furnace or air conditioner does actually change state. We 1. The problem is just too uninteresting can see the whole path from temperature compared to other challenges 2 . The problem is too ill defined to be sensing to turning off the furnace. Every step interesting; or, the problem is only can be seen to be one of a series of straight- apparent, and requires no solution forward physical events. Nowhere are you 3. It’s an interesting problem, but AI has tempted to invoke conscious sensation as an nothing to say about it effect or element of the causal chain. 4. AI researchers may eventually solve it, This is the prima facie case against com- but will require new ideas putationalism, and a solid one it seems to 5. AI researchers will probably solve it, be. The rest of this chapter is an attempt to using existing ideas dismantle it. 6. AI’s current ideas provide at least the outline of a solution 7. My answer is not in the list above. Here An Informal Survey it is: . . . Of course, I don’t mean to exclude other Although one might expect AI researchers branches of cognitive science; when I say to adopt a computationalist position on most “AI” I mean “AI, in conjunction with other issues, they tend to shy away from ques- relevant disciplines.” However, if you think tions about consciousness. AI has often been neuroscientists will figure out phenomenal accused of being over-hyped, and the only consciousness, and that their solution will way to avoid the accusation, apparently, is to entail that anything not made out of neu- be so boring that journalists stay away from rons cannot possibly be conscious, then you. As the field has matured and as a flock choose option 3. Because this topic is of passionate interest to a minority, and quickly of technical problems have become its focus, becomes annoying to many others, please it has become easier to bore journalists. The direct all follow up discussion to fellows- last thing most serious researchers want is [email protected]. Directions for subscrib- to be quoted on the subject of computation ing to this mailing list are as follows: . . . and consciousness. Thanks for your time and attention. To get some kind of indication of what positions researchers take on this issue, I con- Of the approximately 207 living Fellows, ducted an informal survey of Fellows of the I received responses from 34. The results are American Association for Artificial Intelli- as indicated in Table 6.1. gence (AAAI) in the summer of 2003. I sent Of those who chose category 7 (None of the following e-mail to all of them: the above) as answer, here are some of the Most of the time AI researchers don’t con- reasons why: cern themselves with philosophical ques- r tions, as a matter of methodology and per- “Developing an understanding of the haps also opinion about what is ultimately basis for conscious experience is a central, at stake. However, I would like to find out long-term challenge for AI and related P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

120 the cambridge handbook of consciousness

Table 6.1. Results of survey of AAAI fellows pessimists (category 3) were definitely in the minority. 1 Problem uninteresting 3% 2a Ill-defined 11% 19% 2b Only apparent 8% 3 AI silent 7% Research on Computational Models 4 Requires new ideas 32% of Consciousness 5 AI will solve it as is 3% 6 Solution in sight 15% In view of the shyness about conscious- 7 21 None of the above % ness shown by serious AI researchers, it is Percentages indicate fraction of the 34 who responded. not surprising that detailed proposals about phenomenal consciousness from this group disciplines. It’s unclear at the present time should be few and far between. whether new ideas will be needed. . . . ” r “If two brains have isomorphic computa- Moore/Turing Inevitability tion then the ‘qualia’ must be the same. Qualia must be just another aspect of One class of proposals can be dealt with computation – whatever we say of qualia fairly quickly. Hans Moravec, in a series of 1988 1999 must be a property of the computation books ( , ), and Raymond Kurzweil 1999 viewed as computation.” ( ) have more or less assumed that r continuing progress in the development of “There are two possible ways (at least) of faster, more capable computers will cause solving the problem of phenomenal con- computers to equal and then surpass humans sciousness, ‘explaining what conscious- in intelligence and that computer consciousness is’ and ‘explaining consciousness ness will be an inevitable consequence. The away.’ It sounds like you are looking for only argument offered is that the computers a solution of the first type, but I believe will talk as though they are conscious; what the ultimate solution will be of the second more could we ask? type.” r I believe a careful statement of the argu- “The problem is ill-defined, and always ment might go like this: will be, but this does not make it uninteresting. AI will play a major role in 1. Computers are getting more and more solving it.” powerful. 2 If Table 6.1 seems to indicate no particular . This growing power allows computers to pattern, just remember that what the data do tasks that would have been considered show is that the overwhelming majority (173 infeasible just a few years ago. It is rea- of 207) refused to answer the question at all. sonable to suppose, therefore, that many Obviously, this was not a scientific survey, things we think of as infeasible will even- and the fact that its target group contained a tually be done by computers. disproportionate number of Americans per- 3. Pick a set of abilities such that if a system haps biased it in some way. Furthermore, had them we would deal with it as we the detailed responses to my questions indi- would a person. The ability to carry on a cated that respondents understood the terms conversation must be in the set, but we used in many different ways. But if 84%of can imagine lots of other abilities as well: AAAI Fellows don’t want to answer, we can skill in chess, agility in motion, visual per- infer that the questions are pretty far from spicacity, and so forth. If we had a talk- those that normally interest them. Even the ing robot that could play poker well, we 34 who answered include very few optimists would treat it the same way we treated (if we lump categories 5 and 6 together), any real human seated at the same table. although about the same number (cate- 4. We would feel an overwhelming impulse gories 1 and 2) thought the problem didn’t to attribute consciousness to such a robot. really need to be solved. Still, the outright If it acted sad when losing money or made P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

artiﬁcial intelligence and consciousness 121

whimpering sounds when it was dam- they will eventually be able to carry aged, we would respond as we would to out any task we set them. In particular, a human who was sad or in pain. progress in carrying on conversations has 4 5. This kind of overwhelming impulse is our been dismal. r only evidence that a creature is conscious. Even if a computer could carry on a con- In particular, it is the only real way we versation, that would not tell us anything can tell that people are conscious. There- about whether it really was conscious. r fore, our evidence that the robot was con- Overwhelming impulses are not good scious would be as good as one could indicators for whether something is true. have. Therefore the robot would be con- The majority of people have an over- scious, or be conscious for all intents and whelming impulse to believe that there is purposes. such a thing as luck, so that a lucky person has a greater chance of winning at roulette I call this the “Moore/Turing inevitability” than an unlucky person. The whole gam- argument because it relies both on Moore’s bling industry is based on exploiting the 1965 Law (Moore, ), which predicts expo- fact that this absurd theory is so widely nential progress in the power of computers, believed. and on a prediction about how well future programs will do on the “Turing Test,” pro- I come back to the second of these objec- posed by Alan Turing (1950) as a tool for tions in the section on Turing’s test. The oth- 3 rating the intelligence of a computer. Tur- ers I am inclined to agree with. ing thought all questions about the actual intelligence (and presumably degree of con- Hofstadter, Minsky, and McCarthy sciousness) of a computer were too vague or mysterious to answer. He suggested a behav- Richard Hofstadter touches on the prob- iorist alternative. Let the computer carry on lem of consciousness in many of his writ- a conversation over a teletype line (or via ings, especially the material he contributed an instant-messaging system, we would say to Hofstadter and Dennett (1981). Most of he today). If a savvy human judge could not dis- what he writes seems to be intended to stim- tinguish the computer’s conversational abil- ulate or tantalize one’s thinking about the ities from those of a real person at a rate bet- problem. For example, in Hofstadter (1979) ter than chance, then we would have some there is a chapter (reprinted in Hofstadter measure of the computer’s intelligence. We & Dennett, 1981) in which characters talk to could use this measure instead of insisting on an anthill. The anthill is able to carry on a measuring the computer’s real intelligence, conversation because the ants that compose or actual consciousness. This argument has it play roughly the role neurons play in a a certain appeal. It certainly seems that if brain. Putting the discussion in the form of a technology brings us robots that we can- vignette allows for playful digressions on var- not help treating as conscious, then in the ious subjects. For example, the anthill offers argument about whether they really are con- the anteater (one of the discussants) some scious the burden of proof will shift, in the of its ants, which makes vivid the possibility public mind, to the party-poopers who deny that “neurons” could implement a negotia- that they are. But so what? You can’t win tion that ends in their own demise. an argument by imagining a world in which It seems clear from reading this story that you’ve won it and declaring it inevitable. Hofstadter believes that the anthill is con- The anti-computationalists can make sev- scious, and therefore one could use inte- eral plausible objections to the behavioral- grated circuits rather than ants to achieve the inevitability argument: same end. But most of the details are left out. In this as in other works, it’s as if he wants to r Just because computers have made invent a new, playful style of argumentation, impressive strides does not mean that in which concepts are broken up and tossed P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

122 the cambridge handbook of consciousness

together into so many conﬁgurations that the carries through the representations of many original questions one might have asked get things in W∗, especially to M itself. Hence, a shunted aside. If you’re already convinced man’s model of himself is bipartite, one part by the computational story, then this con- concerning his body as a physical object ceptual play is delightful. If you’re a skeptic, and the other accounting for his social and I expect it can get a bit irritating. psychological experience. I put Marvin Minsky in this category as This is why dualism is so compelling. In well, which perhaps should be called “Those particular, Minsky accounts for free will by who don’t take consciousness very seriously supposing that it develops from a “strong 1968 as a problem.” He wrote a paper in primitive defense mechanism” to resist or 1968 (Minsky, b) that introduced the concept deny compulsion. of self-model, which I argue is central to the computational theory of consciousness. If one asks how one’s mind works, he notices areas where it is (perhaps incor- ∗ To an observer B, an object A is a model of rectly) understood, that is, where one rec- ∗ an object A to the extent that B can use A ognizes rules. One sees other areas where to answer questions that interest him about he lacks rules. One could ﬁll this in by A....IfAistheworld, questions for A are postulating chance or random activity. But ∗ experiments. A is a good model of A, in B’s this too, by another route, exposes the self ∗ view, to the extent that A ’s answers agree to the . . . indignity of remote control. We with those of A, on the whole, with respect resolve this unpleasant form of M∗∗ by pos- to the questions important to B. When a tulating a third part, embodying a will or man M answers questions about the world, spirit or conscious agent. But there is no then (taking on ourselves the role of B) we structure in this part; one can say nothing attribute this ability to some internal mech- meaningful about it, because whenever a ∗ anism W inside M. regularity is observed, its representation is transferred to the deterministic rule region. This part is presumably uncontroversial. The will model is thus not formed from ∗ But what is interesting is that W , however it a legitimate need for a place to store def- appears, will include a model of M himself, inite information about one’s self; it has the M∗. In principle, M∗ will contain a model singular character of being forced into the of W∗, which we can call W∗∗. M can use model, willy-nilly, by formal but essentially W∗∗ to answer questions about the way he content-free ideas of what the model must (M) models the world. One would think that contain. ∗∗ ∗ ∗∗ M (the model of M in W ) would be One can quibble with the details, but used to answer questions about the way M the conceptual framework offers a whole models himself, but Minsky has a somewhat ∗∗ new way of thinking about consciousness by different take: M is used to answer gen- showing that introspection is mediated by eral questions about himself. Ordinary ques- models. There is no way for us to penetrate tions about himself (e.g., how tall he is) are ∗ through them or shake them off, so we must answered by M , but very broad questions simply live with any “distortion” they intro- about his nature (e.g., what kind of a thing duce. I put “distortion” in quotes because it he is, etc.) are answered, if at all, by descrip- ∗∗ ∗ is too strong a word. The concepts we use tive statements made by M about M . to describe our mental lives were developed Now, the key point is that the accuracy of ∗ ∗∗ over centuries by people who all shared the M and M need not be perfect. same kind of mental model. The distortions are built in. For instance, there is no inde- A man’s model of the world has a distinctly bipartite structure: One part is concerned pendent notion of “free will” beyond what with matters of mechanical, geometrical, we observe by means of our self-models. We physical character, while the other is associ- cannot even say that free will is a dispens- ated with things like goals, meanings, social able illusion, because we have no way of matters, and the like. This division of W∗ getting rid of it and living to tell the tale. P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

artiﬁcial intelligence and consciousness 123

Minsky’s insight is that to answer many ques- detailed and complete as one can imag- tions about consciousness we should focus ine, it will still make sense to suppose that more on the models we use to answer the there could be insentient beings that exem- questions than on the questions themselves. plify that description. That is, it is pos- Unfortunately, in that short paper, and sible that there could be a behaviourally in his later book The Society of Mind (Min- indiscernible but insentient simulacrum of a human cognizer: a zombie. sky, 1986), Minsky throws off many interesting ideas, but refuses to go into the depth The plausibility of this picture is that that many of them deserve. He has a lot to it does indeed seem that an intricate dia- say about consciousness in passing, such as gram of the hardware and software of a how Freudian phenomena might arise out of robot would leave consciousness out, just as the “society” of subpersonal modules that he with the computer-controlled heating sys- takes the human mind to be. But there is no tem described in the introduction to this solid proposal to argue for or against. chapter. One could print the system descrip- John McCarthy has written a lot on what tion on rose-colored paper to indicate that he usually calls “self-awareness” (McCarthy, the system was conscious, but the color of 1995b). However, his papers are mostly the paper would play no role in what it actu- focused on robots’ problem-solving capaci- ally did. The problem is that in imagining a ties and how they would be enhanced by the zombie one tends at first to forget that the ability to introspect. An important example zombie would say exactly the same things is the ability of a robot to infer that it doesn’t non-zombies say about their experiences. It know something (such as whether the Pope would be very hard to convince a zombie is currently sitting or lying down). This may that it lacked experience, which means, as be self-awareness, but the word “awareness” far as I can see, that we might be zombies, at here is used in a sense that is quite separate which point the whole idea collapses. from the notion of phenomenal conscious- Almost everyone who thinks the idea is ness that is our concern here. coherent sooner or later slips up the way McCarthy (1995a) specifically addresses Moody does: They let the zombie figure out the issue of “zombies,” philosophers’ term that it is a zombie by noticing that it has for hypothetical beings who behave exactly no experience. By hypothesis, this is some- as we do but do not experience anything. thing zombies cannot do. Moody’s paper is This paper is a reply to an article by remarkable only in how obvious the slip-up Todd Moody (1994) on zombies. He lists in it is. some introspective capacities it would be good to give to a robot (“ . . . Observing its Consider, for example, the phenomenon of goal structure and forming sentences about dreaming. Could there be a cognate con- it. . . . Observing how it arrived at its current cept in zombie-English? How might we beliefs. . . . ”). Then he concludes abruptly: explain dreaming to them? We could say that dreams are things that we experience Moody isn’t consistent in his descrip- while asleep, but the zombies would not be tion of zombies. On page 1 they behave able to make sense[z] of this.5 like humans. On page 3 they express puzzlement about human consciousness. Of course, zombies would talk about their Wouldn’t a real Moody zombie behave as dreams (or dreams[z]?) exactly as we do; though it understood as much about con- consult the intricate system diagram to ver- sciousness as Moody does? ify this. McCarthy’s three-sentence reply is just I tend to agree with McCarthy that the about what Moody’s paper deserves. But idea of a zombie is worthless, in spite of its meanwhile philosophers such as Chalmers initial plausibility. Quoting Moody: (1996) have written weighty tomes based Given any functional [ = , more or less, on the assumption that zombies make sense. computational] description of cognition, as McCarthy is not interested in refuting them. P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

124 the cambridge handbook of consciousness

Similarly, McCarthy (1990b) discusses to the good. It’s clear that any computa- when it is legitimate to ascribe mental prop- tionalist theory must eventually explain the erties to robots. In some ways his treatment mechanism of the first-person view in terms is more formal than that of Dennett, which of third-person components. The third per- I discuss below. But he never builds on this son is that which you and I discuss and there- theory to ask the key question: Is there more fore must be observable by you and me, and to your having a mental state than having by other interested parties, in the same way. that state ascribed to you? In other words, the term “third-person data” is just another way of saying “scientific data.” If there is to be a scientific explanation of the Daniel Dennett first person, it will surely seem more like an Daniel Dennett is not a researcher in artifi- “explaining away” than a true explanation. cial intelligence, but a philosopher of mind An account of how yonder piece of meat or and essayist in cognitive science. Nonethe- machinery is conscious will almost certainly less, he is sympathetic to the AI project invoke the idea of the machinery playing a and bases his philosophy on computational trick on itself, the result of which is for it premises to a great degree. The models of to have a strong belief that it has a special mind that he has proposed can be considered first-person viewpoint. to be sketches of a computational model One of Dennett’s special skills is using and therefore constitute one of the most vivid images to buttress his case. He invented ambitious and detailed proposals for how AI the phrase “Cartesian Theater” to describe might account for consciousness. the hypothetical place in the brain where the Dennett’s (1969) Ph.D. dissertation pro- self becomes aware of things. He observes posed a model for a conscious system. It con- that belief in the Cartesian Theater is deep- tains the sort of block diagram that has since seated and keeps popping up in philosophi- become a standard feature of the theories of cal and psychological writings, as well as in such psychologists as Bernard Baars (1988, common-sense musings. We all know that 1997), although the central working arena is there is a lot going on the brain that is pre- designed to account for introspection more conscious or subconscious. What happens than for problem-solving ability. when a train of events becomes conscious? In later work, Dennett has not built upon According to the view Dennett is ridicul- this model, but, in a sense, has been rebuild- ing, to bring it to consciousness is to show ing it from the ground up. The result has it on the screen in the Cartesian Theater. been a long series of papers and books, rich When presented this way, the idea does with insights about consciousness, free will, seem silly, if for no other reason than that and intentionality. Their very richness makes there is no plausible homunculus to put it hard to extract a brisk theoretical state- in the audience. What’s interesting is how ment, but that is my aim. hard it is to shake this image. Just about Dennett has one overriding methodolog- all theorists of phenomenal consciousness at ical principle, to be distrustful of introspec- some point distinguish between “ordinary” tion. This position immediately puts him and “conscious” events by making the latter at odds with such philosophers as Nagel, be accessible to . . . what, exactly? The sys- Searle, and McGinn, for whom the first- tem as a whole? Its self-monitoring modules? person point of view is the alpha and omega One must tread very carefully to keep from of consciousness. On his side Dennett has describing the agent with special access as many anecdotes and experimental data that the good old transcendental self, sitting alone show how wildly inaccurate introspection in the Cartesian Theater. can be, but his view does leave him open to To demolish the Cartesian Theater, Den- the charge that he is ruling out all the com- nett uses the tool of discovering or invent- petitors to his theory from the start. From a ing situations in which belief in it leads to computationalist’s vantage point, this is all absurd conclusions. Many of these situations P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

artiﬁcial intelligence and consciousness 125

are experiments set up by psychology Language, for Dennett, is very impor- researchers. Most famous are the experi- tant, but not because it is spoken by the ments by Libet (1985), whose object was homunculus in the Cartesian Theater. If you to determine exactly when a decision to leave it out, who is speaking? Dennett’s make a motion was made. What emerged answer is certainly bold: In a sense, the lan- from the experiments was that at the point guage speaks itself. We take it for granted where subjects think they have made the that speaking feels like it emanates from our decision, the neural activity preparatory to “transcendental self” or, less politely, from the motion has already been in progress for the one-person audience in the Theater. hundreds of milliseconds. Trying to make Whether or not that view is correct now, it sense of these results using the homuncu- almost certainly was not correct when lan- lar models leads to absurdities. (Perhaps the guage began. In its original form, language choice causes effects in the person’s past?) was an information-transmission device used But it is easy to explain them if you make a by apes whose consciousness, if similar to more inclusive picture of what’s going on in ours in any real respect, would be about a subject’s brain. Libet and others tended to the same as a chimpanzee’s today. Messages assume that giving a subject a button to push expresssed linguistically would be heard by when the decision had been made provided one person and, for one reason or another, a direct route to . . . that pause again . . . the be passed to others. The messages’ chance subject’s self, perhaps? Or perhaps the guy of being passed would depend, very roughly, in the theater? Dennett points out that the on how useful their recipients found them. neural apparatus required to push the but- The same mechanism has been in oper- ton is part of the overall brain system. Up ation ever since. Ideas (or simple patterns to a certain resolution, it makes sense to ask unworthy of the name “idea” – advertising someone, “When did you decide to do X?” jingles, for instance) tend to proliferate in But it makes no sense to try to tease off a proportion to how much they help those subsystem of the brain and ask it the same who adopt them or in proportion to how question, primarily because there is no sub- well they tend to stifle competing ideas – system that embodies the “will” of the whole not unlike what genes do. Dennett adopts system. Dawkins’ (1976) term meme to denote a lin- Having demolished most of the tradi- guistic pattern conceived of in this way. One tional model of consciousness, Dennett’s key meme is the idea of talking to oneself; next goal is to construct a new one, and when it first popped up, it meant literally here he becomes more controversial, and in talking out loud and listening to what was places more obscure. A key component is said. Although nowadays we tend to view human language. It is difficult to think about talking to oneself as a possible symptom of human consciousness without pondering the insanity, we have forgotten that it gives our ability of a normal human adult to say what brains a whole new channel for its parts to they are thinking. There are two possible communicate with each other. If an idea – a views about why it should be the case that pattern of activity in the brain – can reach the we can introspect so easily. One is that we linguistic apparatus, it gets translated into evolved from animals that can introspect, so a new form, and, as it is heard, gets trans- naturally when language evolved one of the lated back into a somewhat different pat- topics it was used on was the contents of our tern than the one that started the chain of introspections. The other is that language events. Creatures that start to behave this plays a more central role than that; with- way start to think of themselves in a new out language, we would not be conscious at light, as someone to talk to or listen to. Self- all, at least full-bloodedly. Dennett’s view is modeling, according to Dennett (and Jaynes, the second. He has little to say about ani- 1976) starts as modeling this person to whom mal consciousness, and what he does say is we are talking. There is nothing special about disparaging. this kind of model; it is as crude as most of P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

126 the cambridge handbook of consciousness

the models we make. But memes for self- is at bottom a property only of mental rep- modeling have been some of the most suc- resentations. There seem to be many kinds cessful in the history (and prehistory) of of “aboutness” in the world; for instance, humankind. To a great degree, they make there are books about lions, but items such us what we are by giving us a model of who as books can be about a topic only if they we are that we then live up to. Every child are created by humans using language and must recapitulate the story Dennett tells, as writing systems to capture thoughts about he or she absorbs from parents and peers all that topic. Books are said to have derived the ways to think of oneself, as a being with intentionality, whereas people have original free will, sensations, and a still small voice or intrinsic intentionality. inside. Computers seem to be textbook cases of The theory has one striking feature: It physical items whose intentionality, if any, assumes that consciousness is based on lan- is derived. If one sees a curve plotted on guage and not vice versa. For that matter, a computer’s screen, the surest way to ﬁnd it tends to assume that for consciousness to out what it’s about is to ask the person who come to be, there must be in place a sub- used some program to create it. In fact, that’s stantial infrastructure of perceptual, motor, the only way. Digital computers are syntac- and intellectual skills. There may be some tic engines par excellence. Even if there is an linguistic abilities that depend on conscious- interpretation to be placed on every step of ness, but the basic ability must exist before a computation, this interpretation plays no and independent of consciousness. role in what the computer does. Each step This conclusion may be fairly easy to is produced purely by operations dependent accept for the more syntactic aspects of lan- on the formal structure of its inputs and prior guage, but it is contrary to the intuitions of state at that step. If you use TurboTax to many when it comes to semantics. Know- compute your income taxes, then the num- ing what a sentence means requires know- bers being manipulated represent real-world ing how the sentence relates to the world. If quantities, and the number you get at the I am told, “There is a lion on the other side end represents what you actually do owe to of that bush,” I have to understand that “that the tax authorities. Nonetheless, TurboTax bush” refers to a particular object in view, I is just applying formulas to the numbers. It have to know how phrases like “other side “has no idea” what they mean. of” work, and I have to understand what “a This intuition is what Dennett wants to lion” means so that I have a grasp of just what defeat, as should every other researcher who I’m expecting to confront. Furthermore, it’s expects a theory of consciousness based on hard to see how I could know what these AI. There is really no alternative. If you words and phrases meant without knowing believe that people are capable of original that I know what they mean. intentionality and computers are not, then Meditating in this way on how mean- you must believe that something will be ing works, the late 19th-century philoso- missing from any computer program that pher Franz Brentano developed the notion tries to simulate humans. That means that of intentionality, the power that mental rep- human consciousness is fundamentally dif- resentations seem to have of pointing to – ferent from machine consciousness, which “being about” – things outside of, and arbi- means that a theory of consciousness based trarily far from, the mind or brain containing on AI is radically incomplete. those representations. The ability of some- Dennett’s approach to the required one to warn me about that lion depends demolition job on intrinsic intentional- on that person’s sure-footed ability to rea- ity is to focus on the prelinguistic, non- son about that animal over there, as well as introspective case. In a way, this is changing on our shared knowledge about the species the subject fairly radically. In the introspec- Panthera leo. Brentano and many philoso- tive set-up, we are talking about elements phers since have argued that intentionality or aspects of the mind that we are routinely P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

artiﬁcial intelligence and consciousness 127

acquainted with, such as words and images. because it is easy to take him as saying that In the non-introspective case, it’s not clear all intentionality is observer-relative. This that those elements or aspects are present would be almost as bad as maintaining a dis- at all. What’s left to talk about if we’re not tinction between original and derived inten- talking about words, “images,” or “thoughts”? tionality, because it would make it hard to We will have to shift to talking about neu- see how the process of intentionality attribu- rons, chips, firing rates, bits, pointers, and tion could ever get started. Presumably my other “subpersonal” entities and events. It’s intuition that I am an intentional system is not clear at all whether these things are indubitable, but on what could it be based? even capable of exhibiting intentionality. It seems absurd to think that this opin- Nonetheless, showing that they are is a key ion is based on what others tell me, but it tactic in Dennett’s attack on the problem of seems equally absurd that I could be my own consciousness (see especially Appendix A of observer. Presumably to be an observer you Dennett, 1991b). If we can define what it is have to be an intentional system (at least, if for subpersonal entities to be intentional, we your observations are to be about anything). can then build on that notion and recover Can I bootstrap my way into intentionality the phenomenal entities we (thought we) somehow? If so, how do I tell the successful started with. “Original” intentionality will bootstrappers from the unsuccessful ones? turn out to be a secondary consequence of A computer program with an infinite loop, what I call impersonal intentionality. endlessly printing, “I am an intentional sys- Dennett’s approach to the problem is to tem because I predict, by taking the inten- call attention to what he calls the intentional tional stance, that I will continue to print this stance, a way of looking at systems in which sentence out,” would not actually be claim- we impute beliefs and goals to them sim- ing anything, let alone something true. ply because there’s no better way to explain Of course, Dennett does not mean for what they’re doing. For example, if you’re intentionality to be observer-relative, even observing a good computer chess program though many readers think he does. (To in action, and its opponent has left himself take an example at random from the Inter- vulnerable to an obvious attack, then one net, the online Philosopher’s Magazine,inits feels confident that the program will embark “Philosopher of the Month” column in April, on that attack. This confidence is not based 2003 (Douglas & Saunders, 2003), writes, on any detailed knowledge of the program’s “Dennett suggests that intentionality is not actual code. Even someone who knows the so much an intrinsic feature of agents, rather, program well won’t bother trying to do a it is more a way of looking at agents.”) Den- tedious simulation to make a prediction that nett has defended himself from this misinter- the attack will occur, but will base their pre- pretation more than once (Dennett, 1991a). diction on the fact that the program almost I come back to this issue in my attempt never misses an opportunity of that kind. If at a synthesis in the section, “A Synthetic you refuse to treat the program as though it Summary.” had goals, you will be able to say very little about how it works. The intentional stance Perlis and Sloman applies to the innards of the program as well. If a data structure is used by the program The researchers in this section, although to make decisions about some situation or they work on hard-headed problems in object S, and the decisions it makes are well artificial intelligence, do take philosophi- explained by assuming that one state of the cal problems seriously and have contributed data structure means that P is true of S, and substantial ideas to the development of the that another means P , then those states do computational model of consciousness. mean P and P . Donald Perlis’s papers build a case that It is perhaps unfortunate that Dennett consciousness is ultimately based on self- has chosen to express his theory this way, consciousness, but I believe he is using the P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

128 the cambridge handbook of consciousness

term “self-consciousness” in a misleading and arguments, which often depend on thought unnecessary way. Let’s start with his paper experiments, such as imagining cases where (Perlis, 1994), which I think lays out a very one is conscious but not of anything, or of as important idea. He asks, Why do we need a little as possible. The problem is that intro- dichotomy between appearance and reality? spective thought experiments are just not a The answer is, Because they could disagree very accurate tool. One may perhaps con- (i.e., because I could be wrong about what I clude that Perlis, although housed in a Com- think I perceive). For an organism to be able puter Science department, is not a thorough- to reason explicitly about this difference, it going computationalist at all. As he says, “I must be able to represent both X (an object conjecture that we may find in the brain spe- in the world) and quote-X, the representa- cial amazing structures that facilitate true tion of X in the organism itself. The latter is self-referential processes, and constitute a the “symbol,” the former the “symboled.” To primitive, bare or ur-awareness, an ‘I.’ I will my mind the most important consequence call this the amazing-structures-and-processes of this observation is that it must be possi- paradigm” (Perlis, 1997) (italics in original). ble for an information-processing system to It is not clear how amazing the “amazing” get two kinds of information out of its X- structures will be, but perhaps they will not recognizer: signals meaning “there’s an X” be computational. and signals meaning “there’s a signal mean- Aaron Sloman has written prolifi- ing ‘there’s an X.’” cally about philosophy and computation, Perlis takes a somewhat different tack. although his interests range far beyond He believes there can be no notion of our topic here. In fact, although he has appearance without the notion of appear- been interested in conscious control, both ance to someone. So the self-model cannot philosophically and as a strategy for orga- get started without some prior notion of self nizing complex software, he has tended to to model. shy away from the topic of phenomenal consciousness. His book The Computer When we are conscious of X, we are also Revolution in Philosophy (Sloman, 1978) has conscious of X in relation to ourselves: It is almost nothing to say about the subject, here, or there, or seen from a certain angle, and in many other writings the main point or thought about this way and then that. Indeed, without a self model, it is not clear he has to make is that the concept of con- to me intuitively what it means to see or sciousness covers a lot of different processes, feel something: it seems to me that a point which should be sorted out before hard of view is needed, a place from which the questions can be answered. However, in scene is viewed or felt, defining the place a few of his papers he has confronted the occupied by the viewer. Without something issue of qualia, notably (Sloman & Chrisley, along these lines, I think that a “neuronal 2003). I think the following is exactly box” would indeed “confuse” symbol and right: symboled: to it there is no external reality, and it has no way to “think” (consider Now suppose that an agent A . . . uses a alternatives) at all. Thus I disagree [with self-organising process to develop concepts Crick] that self-consciousness is a special for categorising its own internal virtual case of consciousness: I suspect that it is the machine states as sensed by internal mon- most basic form of all. itors. . . . If such a concept C is applied by Perlis continues to elaborate this idea in A to one of its internal states, then the only way C can have meaning for A is in rela- later publications. For example, “Conscious- tion to the set of concepts of which it is a ness is the function or process that allows a member, which in turn derives only from system to distinguish itself from the rest of the the history of the self-organising process in world. . . . To feel pain or have a vivid expe- A. These concepts have what (Campbell, rience requires a self” (Perlis, 1997) (ital- 1994) refers to as ‘causal indexicality’. This ics in original). I have trouble following his can be contrasted with what happens when P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

artiﬁcial intelligence and consciousness 129

A interacts with other agents in such a way ist, but he is also an antireductionist, which as to develop a common language for refer- places him in a unique category. Although ring to features of external objects. Thus it is clear in reading his work that he con- A could use ‘red’ either as expressing a siders consciousness to be a crucial topic, he private, causally indexical, concept refer- has been working up to it very carefully. His ring to features of A’s own virtual-machine early work (Smith, 1984) was on “reflection” states, or as expressing a shared concept in programming languages; that is, how and referring to a visible property of the surfaces of objects. This means that if two why a program written in a language could agents A and B have each developed con- have access to information about its own cepts in this way, then if A uses its causally subroutines and data structures. One might indexical concept Ca, to think the thought conjecture that reflection might play a key ‘I am having experience Ca’, and B uses role in a system’s maintaining a self-model its causally indexical concept Cb, to think and thereby being conscious. But since that the thought ‘I am having experience Cb’ early work Smith has moved steadily away the two thoughts are intrinsically private from straightforward computational topics and incommunicable, even if A and B actu- and toward foundational philosophical ones. ally have exactly the same architecture and Each of his papers seems to take tinier steps have had identical histories leading to the from first principles than the ones that have formation of structurally identical sets of concepts. A can wonder: ‘Does B have an gone before, so as to presuppose as lit- experience described by a concept related tle as humanly possible. Nonetheless, they to B as my concept Ca is related to me?’ often express remarkable insight. His paper But A cannot wonder ‘Does B have expe- (Smith, 2002) on the “Foundations of Com- riences of type Ca’, for it makes no sense puting” is a gem. (I also recommend Sloman for the concept Ca to be applied outside (2002) from the same collection [Scheutz, the context for which it was developed, 2002].) namely one in which A’s internal sensors One thing both Smith and Sloman classify internal states. They cannot classify argue is that Turing machines are mislead- states of B. ing as ideal vehicles for computationalism, This idea suggests that the point I casu- which is a point often missed by philoso- 1990 ally assumed at the beginning of this chap- phers. For example, Wilkes ( ) says that ter, that two people might wonder if they “ . . . computers (as distinct from robots) pro- experienced the same thing when they ate duce at best only linguistic and exclusively tacos, is actually incoherent. Our feeling that ‘cognitive’ – programmable – ‘behaviour’: the meaning is clear is due to the twist our the emphasis is on internal psychological self-models give to introspections of the kind processes, the cognitive ‘inner’ rather than Sloman and Chrisley are talking about. The on action, emotion, motivation, and sen- internal representation of the quale of red- sory experience.” Perhaps I’ve misunder- ness is purely local to A’s brain, but the self- stood him, but it’s very hard to see how model says quite the opposite – that objects this can be true, given that all interest- with the color are recognizable by A because ing robots are controlled by digital comput- they have that quale. The quale is made into ers. Furthermore, when computers and soft- an objective entity that might attach itself ware are studied isolated from their physical to other experiences, such as my encounters environments, it’s often for purely tactical with blue things or B’s experiences of red reasons (from budget or personnel limita- things. tions, or to avoid endangering bystanders). If we go all the way back to Winograd’s (1972) SHRDLU system, we find a simu- Brian Cantwell Smith lated robot playing the role of conversa- The last body of research to be examined in tionalist, not because Winograd thought real this survey is that of Brian Cantwell Smith. It robots were irrelevant, but precisely because is hard to dispute that he is a computational- he was thinking of a long-term project in P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

130 the cambridge handbook of consciousness

which an actual robot would be used. As The problem with this point of view is Smith (2002) says, that it is hard to bootstrap oneself out of In one way or another, no matter what what Smith calls the Criterion of Ultimate construal [of formality] they pledge alle- Concreteness: “No naturalistically palatable giance to, just about everyone thinks that theory of intentionality – of mind, com- computers are formal. . . . But since the putation, semantics, ontology, objectivity – outset, I have not believed that this is can presume the identify or existence of necessarily so. . . . Rather, what computers any individual object whatsoever” (1995. are . . . is neither more nor less than the full- p. 184). He tries valiantly to derive subjects fledged social construction and develop- and objects from prior . . . umm . . . “entities” ment of intentional artifacts. (Emphasis called s-regions and o-regions, but it is hard in original.) to see how he succeeds. In spite of its length The point he is trying to make (and it of 420 pages, the book claims to arrive at can be hard to find a succinct quote in no more than a starting point for a com- Smith’s papers) is that computers are always plete rethinking of physics, metaphysics, and connected to the world, whether they are everything else. robots or not, and therefore the meaning Most people will have a hard time follow- their symbols possess is more determined by ing Smith’s inquiry, not least because few those connections than by what a formal the- people agree on his opening premise, that ory might say they mean. One might want everyday ontology is broken and needs to to rule that the transducers that connect be fixed. I actually do agree with that, but I them to the world are non-computational think the problem is much worse than Smith (cf. Harnad, 1990), but there is no princi- does. Unlike him, I am reductionist enough pled way to draw a boundary between the to believe that physics is the science of “all two parts, because ultimately a computer is there is”; so how do objects emerge from a physical parts banging against other physical primordial superposition of wave functions? parts. As Sloman puts it, Fortunately, I think this is a problem for ...The view of computers as somehow everyone and has nothing to do with the 6 essentially a form of Turing machine . . . is problem of intentionality. If computation- simply mistaken. . . . [The] mathematical alists are willing to grant that there’s a glacier notion of computation . . . is not the primary over there, anyone should be willing to con- motivation for the construction or use of sider the computational theory of how sys- computers, nor is it particularly helpful in tems refer to glaciers. understanding how computers work or how to use them (Sloman, 2002). The point Smith makes in the paper cited A Synthetic Summary above is elaborated into an entire book, On the Origin of Objects (Smith, 1995). The In spite of the diffidence of most AI problem the book addresses is the basic researchers on this topic, I believe that there ontology of physical objects. The problem is a dominant position on phenomenal con- is urgent, according to Smith, because the sciousness among computationalists; it is basic concept of intentionality is that a sym- dominant in the sense that among the small bol S stands for an object X, but we have no population of those who are willing to take prior concept of what objects or symbols are. a clear position, this is more or less the posi- A geologist might see a glacier on a moun- tion they take. In this section I try to sketch tain, but is there some objective reason why that postion, pointing out the similarities the glacier is an object (and the group of and differences from the positions sketched stones suspended in it is not)? Smith believes in the preceding section. that all object categories are to some extent The idea in a nutshell is that phe- carved out by subjects (i.e., by information- nomenal consciousness is the property a processing systems like us and maybe some- computational system X has if X models day by robots as well). itself as experiencing things. To understand P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

artiﬁcial intelligence and consciousness 131

it, I need to explain the following three computer’s memory and claim that it con- things: tains a number. The critic then says that the mapping of states that causes this state to 1. what a computational system is; encode “55,000” is entirely arbitrary; there 2 . how such a system can exhibit intention- are an infinite number of ways of interpret- ality; and ing the state of the register, none of which is 3. that to be conscious is to model oneself as the “real” one in any sense. Therefore, all we having experiences. can talk about is the intended one. A notorious example given by John Searle (1992) exemplifies this kind of attack; he claims that The Notion of Computational System the wall of his office could be considered to Before we computationalists can really get be a computer under the right encoding of started, we run into the objection that the its states. word “computer” doesn’t denote the right The semantic prong is the observation, kind of thing to play an explanatory role discussed in the sections on Daniel Den- in a theory of any natural phenomenon. A nett and on Brian Cantwell Smith, that computer, so the objection goes, is an object even after we’ve agreed that the register that people7 use to compute things. With- state encodes “55,000,” there is no objec- out people to assign meanings to its inputs tive sense in which this figure stands for and outputs, a computer is just an overly “Jeanne D’Eau’s 2003 income in euros.” If complex electronic kaleidoscope, generating Jeanne D’Eau is using the EuroTax software a lot of pseudo-random patterns. We may package to compute her income tax, then interpret the output of a computer as a such semantic statements are nothing but prediction about tomorrow’s weather, but a convention adopted by her and the peo- there’s no other sense in which the comple who wrote EuroTax. In other words, the puter is predicting anything. A chess com- only intentionality exhibited by the program puter outputs a syntactically legal expres- is derived intentionality. sion that we can take to be its next move, To avoid these objections, we have to be but the computer doesn’t actually intend careful about how we state our claims. I have to make that move. It doesn’t intend any- space for only a cursory overview here; see thing. It doesn’t care whether the move is McDermott (2001) for a more detailed treat- actually made. Even if it’s displaying the ment. First, the idea of computer is prior to move on a screen or using a robot arm to the idea of symbol. A basic computer is any pick up a piece and move it, these out- physical system whose subsequent states are puts are just meaningless pixel values or predictable given its prior states. By “state” I drive-motor torques until people supply the mean “partial state,” so that the system can meaning. be in more than one state at a time. An encod- In my opinion, the apparent difficulty ing is a mapping from partial physical states of supplying an objective definition of syn- to some syntactic domain (e.g., numerals). tax and especially semantics is the most To view a system as a computer, we need serious objection to the computational the- two encodings, one for inputs and one for ory of psychology, and in particular to a outputs. It computes f(x) with respect to a computational explanation of phenomenal pair I, O of encodings if and only if putting consciousness. To overcome it, we need to it into the partial state encoding x under I come up with a theory of computation causes it to go into a partial state encoding (and eventually semantics) that is observer- f(x) under O. independent. A memory element under an encoding E is There are two prongs to this attack, one a physical system that, when placed into a syntactic and the other semantic. The syn- state s such that E(s) = x, tends to remain tactic prong is the claim that even the sym- in the set of states {s: E(s) = x} for a while. bols we attribute to computers are observer- A computer is then a group of basic relative. We point to a register in the computers and memory elements viewed P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

132 the cambridge handbook of consciousness

under a consistent encoding scheme, mean- niques and think of neural nets as one of ing merely that if changes of component 1’s them – entirely properly, in my opinion. I state cause component 2’s state to change, return to this subject in the section, “Sym- then the encoding of 1’s outputs is the same bol Grounding.” as the encoding of 2’s inputs. Symbol sites then appear as alternative possible stable Intentionality of Computational Systems regions of state space, and symbol tokens as chains of symbol sites such that the occu- I have described Dennett’s idea of the “inten- pier of a site is caused by the presence of the tional stance,” in which an observer explains occupier of its predecessor site. Space does a system’s behavior by invoking such inten- not allow me to discuss all the details here, tional categories as beliefs and goals. Den- but the point is clear: The notions of com- nett is completely correct that there is such puter and symbol are not observer-relative. a stance. The problem is that we some- Of course, they are encoding-relative, but times adopt it inappropriately. People used then velocity is “reference-frame-relative.” to think thunderstorms were out to get The encoding is purely syntactic, or even them, and a sign on my wife’s printer says, presyntactic, because we have said nothing “Warning! This machine is subject to break- about what syntax an encoded value has, down during periods of critical need.” What if any. We could go on to say more about could it possibly mean to say that a machine syntax, but one has the feeling that the demonstrates real intentionality when it is whole problem is a practical joke played so easy to indulge in a mistaken or merely by philosophers on naive AI researchers. metaphorical “intentional stance”? (“Let’s see how much time we can get them Let’s consider an example. Suppose to waste defining ‘computer’ for us, until someone has a cat that shows up in the they catch on.”) I direct you to McDermott kitchen at the time it is usually fed, meow- (2001) for more of my theory of syntax. The ing and behaving in other ways that tend to important issue is semantics, to which we attract the attention of the people who usu- now turn. ally feed it. Contrast that with the case of One last remark: The definitions above a robot that, whenever its battery is low, are not intended to distinguish digital from moves along a black trail painted on the analog computers, or serial from parallel floor that leads to the place where it gets ones. They are broad enough to include any- recharged, and, when it is over a large black thing anyone might ever construe as a com- cross that has been painted at the end of putational system. In particular, they allow the trail, emits a series of beeps that tend to neural nets (Rumelhart et al., 1986), both attract the attention of the people who usu- natural and artificial, to count as computers. ally recharge it. Some people might refuse Many observers of AI (Churchland, 1986, to attribute intentionality to either the cat 1988; Wilkes, 1990) believe that there is an or the robot and treat as purely metaphor- unbridgeable chasm between some classical, ical such comments as, “It’s trying to get digital, traditional AI and a revolutionary, to the kitchen [or recharging area],” or “It analog, connectionist alternative. The for- wants someone to feed [or recharge] it.” mer is the realm of von Neumann machines, They might take this position, or argue that the latter the realm of artificial neural net- it’s tenable, on the grounds that we have no works – “massively parallel” networks of sim- reason to suppose that either the cat or the ple processors (meant to mimic neurons), robot has mental states, and hence nothing which can be trained to learn different cat- with the kind of “intrinsic aboutness” that egories of sensory data (Rumelhart et al., people exhibit. High catologists8 are sure 1986). The “chasm” between the two is cats do have mental states, but the skep- less observable in practice than you might tic will view this as just another example of infer from the literature. AI researchers are someone falling into the metaphorical pit of omnivorous consumers of algorithmic tech- “as-if” intentionality. P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

artiﬁcial intelligence and consciousness 133

I believe, though, that even hard-headed certainly) only one combination of natural- low catologists think the cat is truly inten- language text and simple cipher that pro- tional, albeit in the impersonal way dis- duces the encrypted message. cussed in the section on Daniel Dennett. Unfortunately for this example, it in- They would argue that if you could open volves interpreting the actions of people. So up its brain you would find neural structures even if there is no observer-relativity from that “referred to” the kitchen or the path to the cryptanalyst’s point of view, the inten- it, in the sense that those structures became tionality in a message is “derived” accord- active in ways appropriate to the cat’s needs: ing to skeptics about the possible authentic They were involved in steering the cat to the intentionality of physical systems. kitchen and stopping it when it got there. A similar account would tie the meowing geology behavior to the event of getting food, medi- A geologist strives to find the best expla- ated by some neural states. We would then nation for how various columns and strata feel justified in saying that some of the neu- of rock managed to place themselves in the ral states and structures denoted the kitchen, positions in which they are found. A good or the event of being fed. explanation is a series of not-improbable The question is, are the ascriptions of events that would have transformed a plau- impersonal intentionality so derived arbi- sible initial configuration of rocks into what trary, or are they objectively true? It’s diffi- we see today. cult to make either choice. It feels silly saying In this case, there is no observer-relativity, that something is arbitrary if it takes consid- because there was an actual sequence of erable effort to figure it out, and if one is con- events that led to the current rock config- fident that if others independently under- uration. If two geologists have a profound took the same project they would reach disagreement about the history of a rock for- essentially the same result. But it also feels mation, they cannot both be right (as they odd to say that something is objectively true might be if disagreeing about the beauty of if it is inherently invisible. Nowhere in the a mountain range). Our normal expectation cat will you find labels that say “This means is that any two geologists will tend to agree X,” nor little threads that tie neural struc- on at least the broad outline of an explana- tures to objects in the world. One might tion of a rock formation and that as more want to say that the cat is an intentional data are gathered the areas of agreement system because there was evolutionary pres- will grow. sure in favor of creatures whose innards These examples are cases where, even were tied via “virtual threads” to their sur- though internal harmoniousness is how we roundings. I don’t like dragging evolution judge explanations, what we get in the end in because it’s more of a question stopper is an explanation that is true, independent than a question answerer. I prefer the con- of the harmoniousness. All we need to do is clusion that the reluctance to classify inten- allow for this to be true even though, in the tionality as objectively real simply reveals case of intentionality, even a time machine or an overly narrow conception of objective mind reader would not give us an indepen- reality. dent source of evidence. To help us accept A couple of analogies should help. this possibility, consider the fact that geologists can never actually get the entire story code breaking right. What they are looking at is a huge A code breaker is sure he or she has cracked structure of rock with a detailed microhis- a code when the message turns into mean- tory that ultimately accounts for the posi- ingful natural-language text. That’s because tion of every pebble. What they produce in there are an enormous number of possible the end is a coarse-grained history that talks messages and an enormous number of pos- only about large intrusions, sedimentary lay- sible ciphers, out of which there is (almost ers, and such. Nonetheless we say that it is P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

134 the cambridge handbook of consciousness

objectively true, even though the objects it scale effects, will go unhypothesized, and speaks of don’t even exist unless the account some parts of geologists’ attempts to make is true. It explains how a particular “intru- sense of what they see will be too incoher- sion” got to be there, but if geological theory ent to be true or false or even to refer to isn’t more or less correct, there might not anything. be such a thing as an intrusion; the objects The other remark is that it might be the might be parsed in a totally different way. case that the sheer size of the symbolic If processes and structures inside a cat’s systems inside people’s heads might make brain exhibit objectively real impersonal the impersonal intentionality story irrele- intentionality, then it’s hard not to accept vant. We don’t, of course, know much about the same conclusion about the robot trying the symbol systems used by human brains, to get recharged. It might not navigate the whether there is a “language of thought” way the cat does – for instance, it might have (Fodor, 1975) or some sort of connection- no notion of a place it’s going to, as opposed ist soup but clearly we can have beliefs that to the path that gets it there – but we see are orders of magnitude more complex than the same fit with its environment among the those of a cat or a robot (year-2006 model). symbol structures in its hardware or data. In If you walk to work, but at the end of the the case of the robot the hardware and soft- day absentmindedly head for the parking lot ware were designed, and so we have the extra to retrieve your car, what you will believe option of asking the designers what the enti- once you get there has the content, “My car ties inside the robot were supposed to denote. is not here.” Does this belief correspond to a But it will often happen that there is con- symbol structure in the brain whose pieces flict between what the designers intended include symbol tokens for “my car,” “here,” and what actually occurs, and what actually and “not”? We don’t know. But if anything occurs wins. The designers don’t get to say, like that picture is accurate, then assigning a “This boolean variable means that the robot meaning to symbols such as “not” is consider- is going through a door,” unless the vari- ably more difficult than assigning a meaning able’s being true tends to occur if and only to the symbols a cat or robot might use to if the robot is between two door jambs. If denote “the kitchen.” Nonetheless, the same the variable is correlated with something else basic story can still be told: that the symbols instead, then that’s what it actually means. mean what the most harmonious interpreta- It’s appropriate to describe what the roboti- tion says they mean. This story allows us to cists are doing as debugging the robot so that assign arbitrarily abstract meanings to sym- its actual intentionality matches their intent. bols like “not”; the price we pay is that for The alternative would be to describe the now all we have is an IOU for a holistic the- robot as “deranged” in the sense that it con- ory of the meanings inside our heads. tinuously acts in ways that are bizarre given what its data structures mean. Modeling Oneself as Conscious Two other remarks are in order. What the symbols in a system mean is dependent on I have spent a lot of time discussing inten- the system’s environment. If a cat is moved tionality because once we can establish the to a house that is so similar to the one it’s concept of an impersonal level of mean- familiar with that the cat is fooled, then the ing in brains and computers, we can intro- structures inside it that used to refer to the duce the idea of a self-model, a device that kitchen of house 1 now refer to the kitchen a robot or a person can use to answer ques- of house 2. And so forth; and there will of tions about how it interacts with the world. course be cases in which the denotation of This idea was introduced by Minsky almost a symbol breaks down, leaving no coherent forty years ago (Minsky, 1968a) and has since story about what it denotes, just as in the been explored by many others, including geological case an event of a type unknown Sloman (Sloman & Chrisley, 2003), McDer- to geology, but large enough to cause large- mott (2001), and Dennett (1991b). As I P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

artiﬁcial intelligence and consciousness 135

have mentioned, Dennett mixes this idea quences because many of A’s beliefs about with the concept of meme, but self-models itself will stem from the way its internal don’t need to be made out of memes. surrogates participate in mental models. We We start with Minsky’s observation that call the beliefs about a particular surrogate complex organisms use models of their envi- a self-model, but usually for simplicity I refer ronments to predict what will happen and to the self-model, as if all those beliefs are decide how to act. In the case of humans, pulled together into a single “database.” Let model making is taken for granted by psy- me state up front that the way things really chologists (Johnson-Laird, 1983); no one work is likely to be much more complex really knows what other animals’ capaci- and messy. Let me also declare that the self- ties for using mental models are. A mental model is not a Cartesian point of transcen- model is some sort of internal representation dence where the self can gaze at itself. It is of part of the organism’s surroundings that a resource accessible to the brain at various can be inspected, or even “run” in some way, points for several different purposes. so that features of the model can then be We can distinguish between exterior and transformed back into inferred or predicted interior self-models. The former refers to the features of the world. For example, suppose agent considered as a physical object, some- you’re planning to go grocery shopping, the thing with mass that might sink a lifeboat. skies are threatening rain, and you’re try- The latter refers to the agent considered as an ing to decide whether to take an umbrella. information-processing system. To be con- You enumerate the situations where the crete, let’s look at a self-model that arises umbrella might be useful and think about in connection with the use of any-time algo- whether on balance it will be useful enough rithms to solve time-dependent planning prob- to justify having to keep track of it. One such lems (Boddy & Dean, 1989). An any-time situation is the time when you emerge from algorithm is one that can be thought of as the store with a cartload of groceries to put an asynchronous process that starts with a in the car. Will the umbrella keep you or rough approximation to the desired answer your groceries dry?9 and gradually improves it; it can be stopped This definition is general (and vague) at any time, and the quality of the result enough to cover non-computational models, it returns depends on how much run time but the computationalist framework pro- it was given. We can apply this idea to vides an obvious and attractive approach planning robot behavior, in situations where to theorizing about mental models. In this the objective is to minimize the total time framework, a model is an internal computer required to solve the problem, which is set up to simulate something. The organ- equal to time (tP) to find a plan P + time ism initializes it, lets it run for a while, reads (tE(P)) to execute P. off its state, and interprets the state as a set If the planner is an any-time algorithm, of inferences that then guide behavior. In then the quality of the plan it returns the umbrella example, one might imagine a improves with tP. We write P(tP) to indi- physical simulation, at some level of resolu- cate that the plan found is a function of the tion, of a person pushing a cart and holding time allotted to finding it. Because quality an umbrella while rain falls. is execution time, we can refine that state- A mental model used by an agent A to ment and say that tE(P(tP)) decreases as tP decide what to do must include A itself, sim- increases. Therefore, to optimize ply because any situation A finds itself in will have A as one of its participants. If I tP + tE (P(tP )) am on a sinking ship, and trying to pick a lifeboat to jump into, predicting the num- we must find the smallest tP such that the ber of people on the lifeboat must not omit time gained by planning t longer than that the “+ 1” required to include me. This seem- would probably improve tE by less than t. ingly minor principle has far-reaching conse- The only way to find that optimal tP is to P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

136 the cambridge handbook of consciousness

have an approximate model of how fast be back.” We start by assuming that it is tE(P(tP)) changes as a function of tP. Such not conscious and then think about what it a model would no doubt reflect the law of would gain by having interior self-models of diminishing returns, so that finding the opti- a certain class. The starting assumption, that mal tP is an easy one-dimensional optimiza- it isn’t conscious, should be uncontroversial. tion problem. The important point for us is One thing such a robot might need is a that this model is a model of the planning way to handle perceptual errors. Suppose component of the robot, and so counts as an that it has a subroutine for recognizing chess- interior self-model. boards and chess pieces.10 For serious play Let me make sure my point is clear: Inte- only Staunton chess pieces are allowed, but rior self-models are no big deal. Any algo- you can buy a chessboard with pieces of rithm that outputs an estimate of something almost any shape; I have no doubt that plus an error range incorporates one. The Disney sells a set with Mickey and Min- mere presence of a self-model does not pro- nie Mouse as king and queen. Our robot, vide us some kind of mystical reflection zone we suppose, can correct for scale, lighting, where we can make consciousness pop out and other variations of the appearance of as an “emergent” phenomenon. This point is Staunton pieces, but just can’t “parse” other often misunderstood by critics of AI (Block, kinds of pieces. It could also be fooled by 1997; Rey, 1997) who attribute to computa- objects that only appeared to be Staunton tionalists the idea that consciousness is noth- chess pieces. ing but the ability to model oneself. In so Now suppose that the robot contained doing, they tend to muddy the water fur- some modules for improving its perfor- ther by saying that computationalists con- mance. It might be difficult to calibrate fuse consciousness with self-consciousness. I the perceptual systems of our chess-playing hope in what follows I can make these waters robots at the factory, especially because dif- a bit clearer. ferent owners will use them in different situ- Today’s information-processing systems ations. So we suppose that after a perceptual are not very smart. They tend to work in nar- failure a module we will call the perception row domains and outperform humans only tuner will try to diagnose the problem and in certain areas, such as chess and numerical change the parameters of the perceptual sys- computation, in which clear formal ground tem to avoid it in the future. rules are laid out in advance. A robot that can The perception tuner must have access to walk into a room, spy a chessboard, and ask if the inputs and outputs of the chess recogni- anyone wants to play is still far in the future. tion system and, of course, access to param- This state of affairs raises a huge obstacle eters that it can change to improve the for those who believe that consciousness is system’s performance. It must have a self- built on top of intelligence, rather than vice model that tells it how to change the param- versa, that obstacle being that everything eters to reduce the likelihood of errors. (The we say is hypothetical. It’s easy to counter “backpropagation” algorithm used in neural the computationalist argument. Just say, “I nets (Rumelhart et al., 1986) is an exam- think you’re wrong about intelligence pre- ple.) What I want to call attention to is ceding consciousness, but even if you’re right that the perception tuner interprets the out- I doubt that computers will ever reach the puts of the perceptual system in a rather level of intelligence required.” different way from the decision-making sys- To which I reply, Okay. But let’s sup- tem. The decision-making system interprets pose they do reach that level. We avoid them (to oversimplify) as being about the begging any questions by using my hypothet- environment; the tuning system interprets ical chess-playing robot as a concrete exam- them as being about the perceptual system. ple. We can imagine it being able to loco- For the decision maker, the output “Pawn mote, see chessboards, and engage in simple at x, y, z” means that there is a pawn at a conversations: “Want to play?” “Later.” “I’ll certain place. For the tuner, it means that P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

artiﬁcial intelligence and consciousness 137

the perceptual system says there is a pawn; how wrong-headed people can be about in other words, that there appears to be explaining patterns of events. But perhaps a pawn. all that is required is enough paranoia to Here is where the computationalist anal- avoid too many false negatives in predicting ysis of intentionality steps in. We don’t need catastrophes. to believe that either the decision maker The final step is to suppose that the robot or the tuner literally “thinks” that a symbol can ask fairly general questions about the structure at a certain point means a particu- operation of its perceptual and decision- lar thing. The symbol structure S means X if making systems. Actually, this ability is there is a harmonious overall interpretation closely tied to the ability to store episodic of the states of the robot in which S means memories. To remember something one X. The perceptual-tuner scenario suggests must have a notation to express it. Remem- that we can distinguish two sorts of access bering a motor skill might require storing to a subsystem: normal access and introspec- a few dozen numerical parameters (e.g., tive access. The former refers to the flow weights in neural networks, plus some of information that the subsystem extracts sequencing information). If this is correct, from the world (Dretske, 1981). The latter then, as argued above, learning a skill means refers to the flow of information it produces nudging these parameters toward optimal about the normal flow.11 For our robot, nor- values. Because this notation is so lean, it mal access gives it information about chess won’t support recording the episodes during pieces; introspective access gives it informa- which skill was enhanced. You may remem- tion about . . . what, exactly? A datum pro- ber your golf lessons, but those memories are duced by the tuner would consist of a desig- independent of the “memories,” encoded as nator of some part of the perceptual field numerical parameters, that manifest them- that was misinterpreted, plus information selves as an improved putt. Trying to think about how it was interpreted and how it of a notation in which to record an arbitrary should have been. We can think of this as episode is like trying to think of a formal being information about “appearance” vs. notation to capture the content of a Tol- “reality.” stoy novel. It’s not even clear what it would The next step in our story is to sup- mean to record an episode. How much detail pose that our robot has “episodic” memories; would there be? Would it always have to be that is, memories of particular events that from the point of view of the creature that occurred to it. (Psychologists draw distinc- recorded it? Such questions get us quickly tions between these memories and other into the realm of Knowledge Representa- kinds, such as learned skills [e.g., the mem- tion, and the Language of Thought (Fodor, ory of how to ride a bicycle] and abstract 1975). For that matter, we are quickly led knowledge [e.g., the memory that France is to the topic of ordinary human language, next to Germany], sometimes called seman- because the ability to recall an episode seems tic memory.) We take episodic memory for closely related to the abilities to tell about it granted, but presumably flatworms do with- and to ask about it. We are far from under- out it; there must be a reason why it evolved standing how language, knowledge represen- in some primates. One possibility is that tation, and episodic memory work, but it it’s a means to keep track of events whose seems clear that the mechanisms are tightly significance is initially unknown. If some- connected, and all have to do with what thing bad or good happens to an organ- sorts of questions the self-model can answer. ism, it might want to retrieve past occa- This clump of mysteries accounts for why sions when something similar happened Dennett’s (1991b) meme-based theory is so and try to see a pattern. It’s hard to say attractive. He makes a fairly concrete pro- why the expense of maintaining a com- posal that language came first and that the plex “database” would be paid back in terms evolution of the self-model was driven by of reproductive success, especially given the evolution of language. P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

138 the cambridge handbook of consciousness

Having waved our hands a bit, we can It comes down to this: “Don’t look over get back to discussing the ability of humans, there! The key move is over here, where you and presumably other intelligent creatures, weren’t looking!”14 Phenomenal conscious- to ask questions about how they work. We ness is not part of the mechanism of percep- will just assume that these questions are tion, but part of the mechanism of introspection asked using an internal notation reminiscent about perception. of human language, and then answered using It is easy to think that this theory is simi- a Minskyesque self-model. The key observa- lar to Perlis’s model of self-consciousness as tion is that the self-model need not be com- ultimately fundamental, and many philoso- pletely accurate or, rather, that there is a cer- phers have misread it that way. That’s why tain flexibility in what counts as an accurate the term “self-consciousness” is so mislead- answer, because what it says can’t be con- ing. Ordinarily what we mean by it is contradicted by other sources of information. If sciousness of self. But the self-model theory all people’s self-models say they have free of consciousness aims to explain all phenom- will, then free will can’t be anything but enal consciousness in terms of subpersonal whatever it is all people think they have. It modeling by an organism R of R’s own per- becomes difficult to deny that we have free ceptual system. Consciousness of self is just a will, because there’s no content to the claim particular sort of phenomenal consciousness, that we have it over and above what the cho- so the theory aims to explain it in terms of rus of self-models declare.12 modeling by R of R’s own perceptual system Phenomenal experience now emerges as in the act of perceiving R. In these last two the self-model’s answer to the question, sentences the word “self” does not appear What happens when I perceive something? except as part of the definiendum, not as part The answer, in terms of appearance, real- of the definiens. Whatever the self is, it is ity, and error, is accurate up to a point. not lying around waiting to be perceived; It’s when we get to qualia that the model the act of modeling it defines what it is to a ends the explanation with a just-so story. great extent. There is nothing mystical going It gives more useful answers on such ques- on here. When R’s only view of R is R∗,in tions as whether it’s easier to confuse green Minsky’s terminology, then it is no surprise and yellow than green and red, or what if terms occur in R∗ whose meaning depends to do when senses conflict, or what condi- at least partly on how R∗ fits into everything tions make errors more or less likely. But else R is doing, and in particular on how (the to questions such as, How do I know this natural-language equivalents of those) terms is red in the first place?, it gives an answer are used by a community of organisms to designed to stop inquiry. The answer is that which R belongs. red has this quality (please focus atten- I think the hardest part of this theory tion on the red object), which is intrinsi- to accept is that perception is normally cally different from the analogous quality not mediated, or even accompanied, by for green objects (now focus over here, if qualia. In the introduction to this chapter, I you don’t mind). Because red is “intrinsically invited readers to cast their eyes over a com- like . . . this,” there is no further question to plex climate-control system and observe the ask. Nor should there be. I can take steps to absence of sensation. We can do the same improve my classification of objects by color, exercise with the brain, with the same result. but there’s nothing I can do to improve my It just doesn’t need sensations to do its job. ability to tell red from green (or, more plau- But if you ask it, it will claim it does. A quale sibly, to tell two shades of red apart) once exists only when you look for it. I’ve obtained optimal lighting and viewing Throughout this section, I have tried to conditions.13 stay close to what I think is a consensus posi- The computationalist theory of phenom- tion on a computational theory of phenomenal consciousness thus ends up looking like enal consciousness. But I have to admit that a spoil-sport’s explanation of a magic trick. the endpoint to which I think we are driven P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

artiﬁcial intelligence and consciousness 139

is one that many otherwise fervent computa- tiva of “strong self-reference” is because the tionalists are reluctant to accept. There is no problem-solving system he’s imagining is not alternative conclusion on the horizon, just a an ordinary computer program, but a tran- wish for one, as in this quote from (Perlis, scendental self-contemplating mind – some- 1997): thing not really divided into modules at all, but actively dividing itself into time-shared ...Perhaps bare consciousness is in and of virtual modules as it shifts its attention from itself a self-distinguishing process, a process one aspect of its problem to another, then that takes note of itself. If so, it could still to a meta-layer, a meta-meta-layer, and so be considered a quale, the ur-quale, what it’s like to be a bare subject. . . . What might forth. If you bite the bullet and accept that this be? That is unclear. . . . all this meta-stuff, if it exists at all, exists only in the system’s self-model, then the need Perlis believes that a conscious system needs for strong self-reference, and the “ur-quale,” to be “strongly self-referring,” in that its goes away, much like the ether in the the- modeling of self is modeled in the very mod- ory of electromagnetism. So I believe, but I eling, or something like that. “Why do we admit that most AI researchers who take a need a self-contained self, where referring position probably share Perlis’s reluctance to stops? Negotiating one’s way in a complex let that ether go. world is a tough business. . . . ” He sketches a scenario in which Ralph, a robot, needs a new arm: The Critics

Suppose the new arm is needed within 24 AI has always generated a lot of controversy. hours. He cannot allow his decision-making The typical pattern is that some piece of about the best and quickest way to order research captures the public’s imagination, the arm get in his way, i.e., he must not allow it to run on and on. He can use as amplified by journalists, then the actual meta-reasoning to watch his reasoning so results don’t fit those public expectations, it does not use too much time, but then and finally someone comes along to chalk what is to watch his meta-reasoning? . . . up one more failure of AI research. Mean- He must budget his time. Yet the budget- while, often enough the research does suc- ing is another time-drain, so he must pay ceed, not on the goals hallucinated by the attention to that too, and so on in an infinite popular press, but on those the researchers regress . . . Somehow he must regard [all actually had in mind, so that the AI com- these modules] as himself, one (complex) munity continues to gain confidence that it system reasoning about itself, including is on the right track. Criticism of AI models that very observation. He must strongly of consciousness doesn’t fit this pattern. As self-refer: he must refer to that very referring so that its own time-passage can be I observed at the outset, almost no one in taken into account. (Emphasis in original.) the field is “working on” consciousness, and certainly there’s no one trying to write a con- It appears to me that two contrary intu- scious program. It is seldom that a journalist itions are colliding here. One is the hard- can make a breathless report about a robot headed computationalist belief that self- that will actually have experiences!!15 modeling is all you need for consciousness; Nonetheless, there has been an out- the other is the nagging feeling that self- pouring of papers and books arguing that modeling alone can’t quite get us all the mechanical consciousness is impossible and way. Yet when he tries to find an exam- that suggestions to the contrary are wasteful ple, he winds up with a mystical version of of research dollars and possibly even danger- the work by Boddy and Dean (1989) that ously dehumanizing. The field of “artificial I cited above as a prosaic example of self- consciousness” (AC) is practically defined modeling. It seems clear to me that the by writers who deny that such a thing is only reason Perlis needs the virtus dormi- possible. Much more has been written by P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

140 the cambridge handbook of consciousness

AC skeptics than by those who think it Turing’s Test is not necessarily relevant is possible. In this section I discuss some to the computational theory of conscious- of those criticisms and refute them as best ness. Few of the theorists discussed in this I can. chapter have invoked it as a methodologi- Due to space limitations, I try to focus cal tool. Where it comes in is when reliance on critiques that are specifically directed on it is attributed to computationalists. A at computational models of consciousness, critic will take the computationalist’s focus as opposed to general critiques of materi- on the third-person point of view as an alist explanation. For example, I pass over endorsement of behaviorism and then jump Jackson’s (1982) story about “Mary, the color to Turing’s Test as the canonical behaviorist scientist” who learns what red looks like. tool for deciding whether an entity is con- There are interesting things to say about scious. That first step, from “third-person” it,which I say in McDermott (2001), but to “behaviorist,” is illegitimate. It is, in fact, Jackson’s critique is not directed at, and somewhat ludicrous to accuse someone of doesn’t mention, computationalism in par- being a behaviorist who is so eager to open ticular. I also pass over the vast literature an animal up (metaphorically, that is) and on “inverted spectrum” problems, which is stuff its head with intricate block diagrams. a somewhat more complex version of the All the “third-personist” is trying to do is sour/spicy taco problem. stick to scientifically, that is, publicly, avail- Another class of critiques that I omit are able facts. This attempt is biased against the those whose aim is to show that computers first-person view, and that bias pays off by can never achieve human-level intelligence. eventually giving us an explanation of the As discussed in the section on research on first person. computational models of consciousness, I So there is no particular reason for a com- concede that if computers can’t be intel- putationalist to defend the Turing Test. It ligent then they can’t be conscious either. doesn’t particularly help develop theoreti- But our focus here is on consciousness, so cal proposals, and it gets in the way of think- the critics I try to counter are those who ing about intelligent systems that obviously specifically argue that computers will never can’t pass the test. Nonetheless, an objection be conscious, even if they might exhibit to computationalism raised in the section, intelligent behavior. One important group “Moore/Turing Inevitability,” does require of arguments this leaves out are those based an answer. That was the objection that even on Godel’s¨ proof that Peano arithmetic is if a computer could pass the Turing Test, incomplete (Nagel & Newman, 1958; Pen- this achievement wouldn’t provide any evi- rose, 1989, 1994). These arguments are dence that it actually was conscious. I dis- intended to show a limitation in the abili- agree with this objection on grounds that ties of computers to reason, not specifically should be clear at this point: To be con- a limitation on their ability to experience scious is to model one’s mental life in terms things; in fact, the connection between the of things like sensations and free decisions. two is too tenuous to justify talking about It would be hard to have an intelligent robot the topic in detail. that wasn’t conscious in this sense, because everywhere the robot went it would have to deal with its own presence and its own Turing’s Test decision making, and so it would have to Let’s start where the field started: with have models of its behavior and its thought Turing’s Test (Turing, 1950). As described processes. Conversing with it would be a earlier, it consists of a judge trying to dis- good way of finding out how it thought tinguish a computer from a person by carry- about itself; that is, what its self-models were ing on typed conversations with both. If the like. judge gets it wrong about 50% of the time, Keep in mind, however, that the Tur- then the computer passes the test. ing Test is not likely to be the standard P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

artiﬁcial intelligence and consciousness 141

method to check for the presence of con- mind the slow motion we could carry on a sciousness in a computer system, if we ever perfectly coherent conversation with him. need a standard method. A robot’s self- Searle goes on: model, and hence its consciousness, could be quite different from ours in respects that are Now the claims made by strong AI are that the programmed computer understands the impossible to predict given how far we are stories and that the program in some sense from having intelligent robots. It is also explains human understanding. But we are just barely possible that a computer not now in a position to examine these claims connected to a robot could be intelligent in light of our thought experiment. with only a very simple self-model. Sup- pose the computer’s job was to control the 1. As regards the first claim, it seems to me quite obvious in the example that I do traffic, waste management, and electric grid not understand a word of the Chinese of a city. It might be quite intelligent, but stories. I have inputs and outputs that hardly conscious in a way we could recog- are indistinguishable from those of the nize, simply because it wouldn’t be present native Chinese speaker, and I can have in the situations it modeled the way we are. any formal program you like, but I still It probably couldn’t pass the Turing Test understand nothing.... either. 2 . As regards the second claim, that the Somewhere in this thicket of possibilities program explains human understand- there might be an artificial intelligence with ing, we can see that the computer and its an alien form of consciousness that could program do not provide sufficient con- pretend to be conscious on our terms while ditions of understanding since the computer and the program are functioning, knowing full well that it wasn’t. It could then and there is no understanding. pass the TuringTest, wine tasting division, by faking it. All this shows is that there is a slight It’s hard to see what this argument has possibility that the TuringTest could be good to do with consciousness. The connection is at detecting intelligence and not so good at somewhat indirect. Recall that in the sec- detecting consciousness. This shouldn’t give tion, “Intentionality of Computational Sys- much comfort to those who think that the tems,” I made sure to discuss “impersonal” Turing Test systematically distracts us from intentionality, the kind a system has by the first-person viewpoint. If someone ever virtue of being a computer whose symbol builds a machine that passes it, it will cer- structures are causally connected to the envi- tainly exhibit intentionality and intelligence ronment so as to denote objects and states and almost certainly be conscious. There’s a of affairs in that environment. Searle abso- remote chance that human-style conscious- lutely refuses to grant that there is any such ness can be faked, but no chance that intel- thing as impersonal or subpersonal inten- ligence can be.16 tionality (Searle, 1992). The paradigm case of any mental state is always the conscious mental state, and he is willing to stretch men- The Chinese Room tal concepts only far enough to cover uncon- One of the most notorious arguments in the scious mental states that could have been debate about computational consciousness conscious (repressed desires, for instance). is Searle’s (1980) “Chinese Room” argument. Hence there is no understanding of Chi- It’s very simple. Suppose we hire Searle nese unless it is accompanied by a conscious (who speaks no Chinese) to implement a awareness or feeling of understanding. computer program for reading stories in Chi- If Searle’s stricture were agreed upon, nese and then answering questions about then all research in cognitive science would those stories. Searle reads each line of the cease immediately, because it routinely program and does what it says. He executes assumes the existence of non-conscious sym- the program about a million times slower bol processing to explain the results of than an actual CPU would, but if we don’t experiments.17 P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

142 the cambridge handbook of consciousness

Searle seems to have left an escape clause, lection of real weights. What, then, is the the notion of “weak AI”: ontological status of these symbols and data structures? If we believe that these symbols I ﬁnd it useful to distinguish what I will and the computational processes over them call ‘strong’ AI from ‘weak’ or ‘cautious’ are really present in the brain, and really AI. . . . According to weak AI, the principal value of the computer in the study of explain what the brain does, then we are the mind is that it gives us a very power- back to strong AI. But if we don’t believe ful tool. For example, it enables us to for- that, then why the hell are we simulating mulate and test hypotheses in a more rig- them? By analogy, let us compare strong vs. orous and precise fashion. But according weak computational meteorology. The for- to strong AI, the computer is not merely a mer is based on the belief that wind veloci- tool in the study of the mind; rather, the ties and air pressures really have something appropriately programmed computer really to do with how hurricanes behave. The latter is a mind, in the sense that computers given allows us to build “powerful tools” that per- the right programs can be literally said to form “computer simulations of [hurricanes’ understand and have other cognitive states physical] capacities,” and “formulate and test (Searle, 1980). hypotheses” about . . . something other than Many people have adopted this terminol- wind velocities and air pressures? ogy, viewing the supposed weak version of Please note that I am not saying that AI as a safe harbor in which to hide from crit- all cognitive scientists are committed to a icism. In my opinion, the concept of weak computationalist account of consciousness. AI is incoherent. Suppose someone writes I’m just saying that they’re committed to a a program to simulate a hurricane, to use a computationalist account of whatever it is common image. The numbers in the simula- they’re studying. If someone believes that tion denote actual or hypothetical air pres- the EPAM model (Feigenbaum & Simon, sures, wind velocities, and the like. The sim- 1984) accounts for human errors in memo- ulation embodies differential equations that rizing lists of nonsense syllables, they have to are held to be more or less true statements believe that structures isomorphic to the dis- about how wind velocities affect air pres- crimination trees in EPAM are actually to be sures and vice versa, and similarly for all the found in human brains. If someone believes other variables involved. Now think about that there is no computationalist account of “computer simulations of human cognitive intelligence, then they must also believe that capacities” (Searle’s phrase). What are the a useful computer simulation of intelligence analogues of the wind velocities and air pres- must simulate something other than symbol sures in this case? When we use the simu- manipulation, perhaps ectoplasm secretions. lations to “formulate and test hypotheses,” In other words, given our lack of any non- what are the hypotheses about? They might computational account of the workings of be about membrane voltages and currents in the mind, they must believe it to be point- neurons, but of course they aren’t, because less to engage in simulating intelligence at neurons are “too small.” We would have to all at this stage of the development of the simulate an awful lot of them, and we don’t subject. really know how they’re connected, and the There remains one opportunity for con- simulation would just give us a huge chunk fusion. No one believes that a simulation of predicted membrane currents anyway. So of a hurricane could blow your house off no one does that. Instead, they run simula- the beach. Why should we expect a simulations at a much higher level, at which sym- tion of a conscious mind to be conscious (or bols and data structures emerge. This is true expect a simulation of a mind to be a mind)? even for neural-net researchers, whose mod- Well, we need not expect that, exactly. If a els are much, much smaller than the real simulation of a mind is disconnected from an thing, so that each connection weight rep- environment, then it would remain a mere resents an abstract summary of a huge col- simulation. P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

artiﬁcial intelligence and consciousness 143

However, once the connection is made too much in speed, size, and energy con- properly, we confront the fact that a suffi- sumption. I use scare quotes here because ciently detailed simulation of computation I doubt that things will turn out to be that Ciscomputation C. This is a property of for- tidy. Nonetheless, no matter how the wires mal systems generally. As Haugeland (1985) work out, the point is that nothing other than observes, the difference between a game like computation need be involved in conscious- tennis and a game like chess is that the ness, which is what Strong AI boils down to. former involves moving a physical object, Weak AI boils down to a sort of “cargo cult” the ball, through space, whereas the latter whose rituals involve simulations of things involves jumping from one legal board posi- someone only guesses might be important tion to the next, and legal board positions are in some way. not physical entities. In tennis, one must hit Now that I’ve clarified the stakes, let’s a ball with certain prescribed physical prop- look at Searle’s argument. It is ridiculously erties using a tennis racket, which must also easy to refute. When he says that “the satisfy certain physical requirements. Chess claims made by strong AI are that the pro- requires only that the state of the game be grammed computer understands the sto- represented with enough detail to capture ries and that the program in some sense the positions of all the pieces.18 One can use explains human understanding,” he may be any 8 × 8 array as a board, and any col- right about the second claim (depending on lection of objects as pieces, provided they how literally you interpret “explains”), but are isomorphic to the standard board and he is completely wrong about the first claim, pieces. One can even use computer data that the programmed computer understands structures. So a detailed simulation of a good something. As McCarthy says, “The Chi- chess player is a good chess player, provided nese Room Argument can be refuted in one it is connected by some channel, encoded sentence: Searle confuses the mental qual- however you like, between its computations ities of one computational process, himself and an actual opponent with whom it is for example, with those of another pro- alternating moves. Whereas for a simulation cess that the first process might be inter- of a tennis player to be a tennis player, it preting, a process that understands Chinese, would have to be connected to a robot capa- for example” (McCarthy, 2000). Searle’s ble of tracking and hitting tennis balls. slightly awkward phrase “the programmed This property carries over to the simula- computer” gives the game away. Computers tion of any other process that is essentially and software continually break our histori- computational. So, if it happens that con- cally founded understanding of the identity sciousness is a computational phenomenon, of objects across time. Any computer user then a sufficiently faithful simulation of a has (too often) had the experience of not conscious system would be a conscious sys- knowing “whom” they’re talking to when tem, provided it was connected to the envi- talking to their program. Listen to a layper- ronment in the appropriate way. This point son try to sort out the contributions to their is especially clear if the computations in current state of frustration made by the question are somewhat modularizable, as e-mail delivery program, the e-mail read- might be the case for a system’s self-model. ing program, and the e-mail server. When The difference between a non-conscious you run a program you usually then talk tennis player and a conscious one might to it. If you run two programs at once you involve connections among its internal com- switch back and forth between talking to putational modules, and not the connections one and talking to the other.19 The phrase from there to its cameras and motors. There “programmed computer” makes it sound as would then be no difference between the if programming it changes it into something “consciousness module” and a detailed simu- you can talk to. The only reason to use such lation of that “module”; they would be inter- an odd phrase is because in the story Searle changeable, provided that they didn’t differ himself plays the role of the programmed P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

144 the cambridge handbook of consciousness

computer, the entity that doesn’t under- tional Systems,” symbols get their meanings stand. By pointing at the “human CPU” and by being causally connected to the world. shouting loudly, he hopes to distract us from Harnad doesn’t disagree with this, but he the abstract entity that is brought into exis- thinks that the connection must take the tence by executing the story-understanding special form of neural networks, natural or program. artiﬁcial.21 The inputs to the networks must We can state McCarthy’s argument be sensory transducers. The outputs are neu- vividly by supposing that two CPUs are rons that settle into different stable patterns involved, as they might well be. The story- of activation depending on how the trans- understanding program might be run on one ducers are stimulated. The possible stable for a while, then on the other, and so forth, patterns and the way they classify inputs as dictated by the internal economics of the are learned over time as the network is operating system. Do AI researchers imag- trained by its owner’s encounters with its ine that the ability to “understand” jumps surroundings. back and forth between the two CPUs? How does the hybrid system ﬁnd the If we replace the two CPUs by two peo- invariant features of the sensory projection ple, does Strong AI predict that the ability that make it possible to categorize and iden- to understand Chinese will jump back and tify objects correctly? Connectionism, with forth between the two people (McDermott, its general pattern-learning capability, seems 2001)? Of course not. to be one natural candidate (though there may well be others): Icons, paired with feedback indicating their names, could be pro- Symbol Grounding cessed by a connectionist network that learns In both of the preceding sections, it some- to identify icons correctly from the sample of times seems as if intentionality is the real confusable alternatives it has encountered, issue, or what Harnad (1990, 2001) calls by dynamically adjusting the weights of the the symbol-grounding problem. The prob- features and feature combinations that are lem arises from the idea of a disembodied reliably associated with the names in a way computer living in a realm of pure syntax, that (provisionally) resolves the confusion. which we discussed in the section on Brian It thereby reduces the icons to the invari- Cantwell Smith. Suppose that such a com- ant (confusion-resolving) features of the cat- puter ran a simulation of the battle of Water- egory to which they are assigned. The net loo. That is, we intend it to simulate that result is that the grounding of the name to battle, but for all we know there might be the objects that give rise to their sensory pro- another encoding of its states that would jections and their icons would be provided make it be a simulation of coffee prices in by neural networks (Harnad, 1990). Ecuador.20 What connects the symbols to The symbol-grounding problem, if it is the things they denote? In other words, what a problem, requires no urgent solution, as grounds the symbols? far as I can see. I think it stems from a This problem underlies some people’s basic misunderstanding about what com- concerns about the Turing Test and the Chi- putationalism is and what the alternatives nese Room because the words in the Tur- are. According to Harnad, “The predom- ing Test conversation might be considered inant approach to cognitive modeling is to be ungrounded and therefore meaningless still what has come to be called ‘compu- (Davidson, 1990) and the program and data tationalism’ . . . , the hypothesis that cogni- structures being manipulated by the human tion is computation. The more recent rival CPU John Searle seem also to be discon- approach is ‘connectionism’ . . . , the hypoth- nected from anything that could give them esis that cognition is a dynamic pattern of meaning. connections and activations in a ‘neural net’” As should be clear from the discussion (Harnad, 2001). Put this way, it seems clear in the section, “Intentionality of Computa- that neural nets would be welcome under P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

artiﬁcial intelligence and consciousness 145

computationalism’s “big tent,” but Harnad such that if X implies Y,” what is recorded spurns the invitation by imposing a series is a “relation . . . of temporal activation, such of fresh requirements. By “computation” that when X is presently active, Y is a cat- he means “symbolic computation,” which egorization that is potentially active next” consists of syntactic operations on “symbol (p. 31). But he remains a committed com- tokens.” Analog computation is ruled out. putationalist through this seemingly discon- Symbolic computation doesn’t depend on tinuous change. For instance, in discussing the medium in which it is implemented, how the new paradigm would actually work, just so long as it is implemented some- he writes, “The discussion of [insert detailed how (because the syntactic categories of proposal here] illustrates how the discipline the symbol tokens will be unchanged). And of implementing a process in a computer last, but certainly not least, “the symbols representation forces distinctions to be redis- and symbol manipulations in a symbol sys- covered and brings into question consistency tem [must be] systematically interpretable of the theory” (p. 44). (Fodor & Pylyshyn, 1988): They can be The moral is that we must be careful to assigned a semantics, and they mean some- distinguish between two ways in which com- thing (e.g., numbers, words, sentences, chess puters are used in psychological modeling: as moves, planetary motions, etc.).” The alter- implementation platform and as metaphor. native is “trivial” computation, which pro- The digital-computer metaphor might shed duces “uninterpretable formal gibberish.” light on why we have a single stream of As I argue in McDermott (2001), these consciousness (∼ von Neumann instruc- requirements have seldom been met by what tion stream?), why we can only remem- most people call “computational” systems. ber 7 ± 2 things (∼ size of our register The average computer programmer knows set?), and why we have trouble with deep nothing about formal semantics or system- center-embedded sentences like “The boy atic interpretability. Indeed, in my experi- the man the dog bit spanked laughed” (∼ ence it is quite difﬁcult to teach a program- stack overﬂow?). The metaphor may have mer about formal systems and semantics. had some potential in the 1950s, when cog- One must scrape away layers of prior con- nitive science was just getting underway, but ditioning about how to “talk” to comput- it’s pretty much run out of steam at this ers. Furthermore, as I write in the section, point. Clancey is correct to point out how “The Notion of a Computational System,” the metaphor may have affected cognitive few AI practitioners refuse to mix and match science in ways that seemed too harmless connectionist and symbolic programs. One to notice, but that in retrospect are hard to must be careful about how one interprets justify. For instance, the program counter in what they say about their practice. Clancey a computer makes pursuing a rigid list of (1999), in arguing for a connectionist archi- tasks easy. If we help ourselves to a program tecture, calls the previous tradition model- counter in implementing a cognitive model, ing the brain as a “wet” computer similar in we may have begged an important question important respects to the “dry” computers about how sequentiality is achieved in a par- we use as models. He argues that we should allel system like the brain. replace it with a particular connectionist What I argue is that the essence of com- architecture. As an example of the change putationalism is to believe that (a) brains are this would bring, he says (p. 30), “Cogni- essentially computers and (b) digital com- tive models have traditionally treated pro- puters can simulate them in all important cedural memory, including inference rules respects, even if they aren’t digital at all. (‘if X then Y’), as if human memory is just Because a simulation of a computation is a computer random-access memory. . . . ” He computation, the “digitality” of the digital proposes to “explore the hypothesis that a computer cancels out. If symbol grounding sequential association, such as an inference- is explained by some very special properties rule . . . , is a temporal relation of activation, of a massively parallel neural network of a P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

146 the cambridge handbook of consciousness

particular sort, then if that net can be sim- how the system’s senses work. The differ- ulated in real time on a cluster of parallel ence between appearance and reality arises workstations, the cluster becomes a virtual at this point, and allows the system to rea- neural net, which grounds symbols as well son about its errors in order to reduce the as a “real” one would. chance of making them. But the self-model Perhaps this is the place to mention the also serves to set boundaries to the questions paper by O’Brien and Opie (1999) that that it can answer. The idea of a sensory presents a “connectionist theory of phe- quale arises as a useful way of cutting off nomenal experience.” The theory makes a useless introspection about how things are basic assumption, that a digital simulation ultimately perceived and categorized. of a conscious connectionist system would Beyond this point it is hard to find consen- not be conscious. It is very hard to see sus between those who believe that the just- how this could be true. It’s the zombie so story the self-model tells its owner is all hypothesis, raised from the dead one more you need to explain phenomenal conscious- time. The “real” neural net is conscious, but ness, and those who think that something the simulated one, in spite of operating in more is needed. Frustratingly, we won’t be exactly the same way (plus or minus a little able to create systems and test hypothe- noise), would be experience-less – another ses against them in the foreseeable future, zombie lives. because real progress on creating conscious programs awaits further developments in enhancing the intelligence of robots. There Conclusions is no guarantee that AI will ever achieve the requisite level of intelligence, in which case The contribution of artificial intelligence to this chapter has been pretty much wasted consciousness studies has been slender so effort. far, because almost everyone in the field There are plenty of critics who don’t would rather work on better defined, less want to wait to see how well AI suc- controversial problems. Nonetheless, there ceeds, because they think they have argu- do seem to be common themes running ments that can shoot down the concept through the work of AI researchers that of machine consciousness or rule out cer- touches on phenomenal consciousness. Con- tain forms of it, right now. We examined sciousness stems from the structure of the three such arguments: the accusation that self-models that intelligent systems use to rea- AI is behaviorist on the subject of conscious- son about themselves. A creature’s models of ness, the “Chinese Room” argument, and the itself are like models of other systems, except symbol-grounding problem. In each case the for some characteristic indeterminacy about basic computationalist working hypothesis what counts as accuracy. To explain how survived intact: that the embodied brain is an information-processing system can have an “embedded” computer and that a reason- a model of something, there must be a prior ably accurate simulation of it would have notion of intentionality that explains why whatever mental properties it has, including and how symbols inside the system can refer phenomenal consciousness. to things. This theory of impersonal intentionality is based on the existence of harmonious match-ups between the states of the system Notes and states of the world. The meanings of symbol structures are what the match-ups 1. I would be tempted to say there is a spectrum say they are. from “weak” to “strong” computationalism to Having established that a system’s model reflect the different stances on these issues, of that very system is a non-vacuous idea, the but the terms “weak” and “strong” have been next step is to show that the model almost used by John Searle (1980) in a quite different certainly will contain ways of thinking about way. See the section on the “Chinese room.” P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

artiﬁcial intelligence and consciousness 147

2. I am taking this possibility seriously for now is the “designer.” But such basic concepts can’t because everyone will recognize the issue and depend on historical events arbitrarily that are its relationship to the nature of qualia. But I far in the past. 2003 follow Sloman & Chrisley ( ) in believ- 12. For the complete story on free will, see ing that cross-personal comparison of qualia McDermott (2001, Chapter 3). I referred makes no sense. See section on Perlis and to Minsky’s rather different theory above; 2001 Sloman and McDermott ( ). McCarthy champions his own version in 3. Turing actually proposed a somewhat differ- McCarthy & Hayes (1969). 1990 ent test. See Davidson ( ) for discussion. 13. One may view it as a bug that a concept, Nowadays this version is the one everyone qualia, whose function is to end introspec- works with. tive questioning, has stimulated so much 4. The Loebner Prize is awarded every year conversation! Perhaps if human evolution to the writer of a program that appears goes on long enough natural selection will “most human” to a panel of judges. You eliminate those of us who persist in talking can see how close the programs are get- about such things, especially while crossing ting to fooling anyone by going to its Web busy streets. site, http://www.loebner.net/Prizef/loebner- 14. Cf. Wittgenstein (1953): “The decisive move- prize.html. ment in the conjuring trick has been made, 5 . The “[z]” is used to flag zombie words whose and it was the very one we thought quite meanings must not be confused with normal innocent.” human concepts. 15. One occasionally hears news reports about 6 . Even more fortunate, perhaps, is the fact that attempts to build an artificial nose. When I few will grant that foundational ontology is a hear such a report, I picture a device that problem in the first place. Those who think measures concentrations of substances in the elementary particles invented us, rather than air. But perhaps the average person imag- vice versa, are in the minority. ines a device that “smells things,” so that, 7. Or intelligent aliens, but this is an irrelevant for example, the smell of a rotten egg would variation on the theme. be unpleasant for it. In any case, these news 8. By analogy with Christology in Christian the- reports seem not to have engendered much ology, which ranges from high to low depend- controversy so far. ing on how superhuman one believes Jesus to 16. I realize that many people, for instance be. Robert Kirk (1994), believe that in principle 9. For some readers this example will elicit fairly something as simple as a lookup table could detailed visual images of shopping carts and simulate intelligence. I don’t have space here umbrellas, and for those readers it’s plausible to refute this point of view, except to note that the images are part of the mental-model that in addition to the fact that the table machinery. But even people without much would be larger than the known universe and visual imagery can still have mental models take a trillion years to build, a computer car- and might still use them to reason about gro- rying on a conversation by consulting it would cery shopping. not be able to answer a question about what 10. I have two reasons for positing a chessboard- time it is. recognition subroutine instead of a general- 17. There is a popular belief that there is such purpose vision system that recognizes chess- a thing as “nonsymbolic” or “subsymbolic” boards and chess pieces in terms of more cognitive science, as practiced by those who “primitive” elements: (1) Many roboticists study artificial neural nets. As I mentioned in prefer to work with specialized perceptual the section, “The Notion of Computational systems, and (2) the qualia-like entities we System,” this distinction is usually unimpor- will predict will be different in content from tant, and the present context is an exam- human qualia, which reduces the chances of ple. The goal of neural-net researchers is jumping to conclusions about them. to explain conscious thought in terms of 11. Of course, what we’d like to be able to say unconscious computational events in neu- here is that normal access is the access it was rons, and as far as Searle is concerned, this designed to support, and for most purposes is just the same fallacy all over again (Searle, that’s what we will say, even when evolution 1990). P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

148 the cambridge handbook of consciousness

18. And a couple of other bits of information, Davidson, D. (1990). Turing’s test. In K. M. such as whether each player still has castling Said, W. Newton-Smith, R. Viale, & K. Wilkes as an option. (Eds.), Modelling the mind (pp. 1–11). Oxford: 19. Technically I mean “process” here, not “pro- Clarendon Press. gram.” McCarthy’s terminology is more accu- Dawkins, R. (1976). The selfish gene. Oxford: rate, but I’m trying to be intelligible by tech- Oxford University Press. nical innocents. Dennett, D. C. (1969). Content and consciousness. 20. I believe these particular examples (Water- London: Routledge & Kegan Paul. loo and Ecuador) were invented by someone Dennett, D. C. (1991a). Real patterns. Journal of other than me, but I have been unable to find Philosophy, 88, 27–51. the reference. Dennett, D. C. (1991b). Consciousness explained. 21. The fact that these are called “connectionist” Boston: Little, Brown. is a mere pun in this context – I hope. Douglas, G., & Saunders, S. (2003). Dan Dennett: Philosopher of the month. TPM Online: The Philosophers’ Magazine on the References internet. Retrieved August 2, 2005, from http://www.philosophers.co.uk/cafe/phil 2003 Baars, B. J. (1988). A cognitive theory of conscious- apr .htm. ness. New York: Guilford Press. Dretske, F. I. (1981). Knowledge and the flow of Baars, B. J. (1997). In the theater of consciousness: information. Cambridge, MA: MIT Press. The work space of the mind. New York: Oxford Feigenbaum, E. A., & Simon, H. A. (1984). University Press. Epam-like models of recognition and learning. 4 305 336 Block, N. (1997). On a confusion about a func- Cognitive Science, 8( ), – . tion of consciousness. In N. Block, O. Flana- Fodor, J. (1975). The language of thought. New gan,&G.Guzeldere¨ (Eds.), The nature of York: Thomas Y. Crowell. consciousness: Philosophical debates (pp. 375– Fodor, J., & Pylyshyn, Z. (1988). Connectionism 415). Cambridge, MA: MIT Press. and cognitive architecture: A critical analysis. Block, N., Flanagan, O., & Guzeldere,¨ G. In S. Pinker, & J. Mehler (Eds.), Connections (Eds.) (1997). The nature of consciousness: and symbols (pp. 3–72). Cambridge, MA: MIT Philosophical debates. Cambridge, MA: MIT Press. Press. Harnad, S. (1990). The symbol grounding prob- Boddy, M., & Dean, T. (1989). Solving time- lem. Physica D, 42, 335–346. dependent planning problems. Proceedings of Harnad, S. (2001). Grounding symbols in the the 11th International Joint Conference on Art- analog world with neural nets – a hybrid ficiail Intelligence, 979–984. model. Psycoloquy. Retrieved August 2, 2005, Campbell, J. (1994). Past, space and self. Cam- from http://psycprints.ecs.soton.ac.uk/archive/ bridge, MA: MIT Press. 00000163/. Chalmers, D. (1996). The conscious mind: In Haugeland, J. (1985). Artificial intelligence: The search of a fundamental theory. New York: very idea. Cambridge, MA: MIT Press. Oxford University Press. Hayes-Roth, B. (1985). A blackboard architecture Churchland, P. (1986). Neurophilosophy: To- for control. Artificial Intelligence, 26(3), 251– ward a unified science of the mind-brain. Cam- 321. bridge, MA: MIT Press. Hofstadter, D. R. (1979). Gödel, Escher, Bach: An Churchland, P. (1988). Matter and consciousness: eternal golden braid. New York: Basic Books. A contemporary introduction to the philosophy of Hofstadter, D. R., & Dennett, D. C. (1981). The mind. Cambridge, MA: MIT Press. mind’s I: Fantasies and reflections on self and Clancey, W. J. (1999). Conceptual coordination: soul. New York: Basic Books. How the mind orders experience in time. Mah- Jackson, F. (1982). Epiphenomenal qualia. Philo- wah, NJ: Erlbaum. sophical Quarterly, 32, 127–136. Currie, K., & Tate, A. (1991). O-plan: The Jaynes, J. (1976). The origins of consciousness in open planning architecture. Artificial Intelli- the breakdown of the bicameral mind. Boston: gence, 52(1), 49–86. Houghton Mifflin. P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

artiﬁcial intelligence and consciousness 149

Johnson-Laird, P. N. (1983). Mental models. Cam- Moravec, H. P. (1988, Summer). Sensor fusion in bridge, MA: Harvard University Press. certainty grids for mobile robots. AI Magazine, Kirk, R. (1994). Raw feeling: A philosophical 9, 61–74. account of the essence of consciousness. Oxford: Moravec, H. (1999). Robot: Mere machine to tran- Oxford University Press. scendent mind. New York: Oxford University Kurzweil, R. (1999). The age of spiritual machines: Press. When computers exceed human intelligence. New Nagel, E., & Newman, J. R. (1958). Goedel’s proof. York: Penguin Books. New York: New York University Press. Libet, B. (1985). Unconscious cerebral initiative O’Brien, G., & Opie, J. (1999). A connectionist and the role of conscious will in voluntary theory of phenomenal experience. Behavioral action. Behavioral and Brain Sciences, 8, 529– and Brain Sciences, 22, 127–148. 566. Penrose, R. (1989). The emperor’s new mind: Con- McCarthy, J. (1990b). Ascribing mental qualities cerning computers, minds, and the laws of physics. to machines. In J. McCarthy (Ed.), Formalizing New York: Oxford University Press. commonsense. Norwood, NJ: Ablex. Penrose, R. (1994). Shadows of the mind: A search McCarthy, J. (1995a). Todd Moody’s zombies. for the missing science of consciousness. New Journal of Consciousness Studies. Retrieved June York: Oxford University Press. 26, 2006, from http://www-formal.stanford. Perlis, D. (1994). An error-theory of conscious- edu/jmc/zombie/zombie.html. ness. Unpublished material, University of Mary- McCarthy, J. (1995b). Making robots con- land Computer Science, CS-TR-3324, College scious of their mental states. Proceedings of Park, MD. the Machine Intelligence Workshop. Retrieved Perlis, D. (1997). Consciousness as self-function. August 2, 2005, from http://www-formal. Journal of Consciousness Studies, 4(5/6), 509– stanford.edu/jmc/consciousness.html. 525. McCarthy, J. (2000). John Searle’s Chinese room Pinker, S., & Mehler, J. (1988). Connections and argument. Retrieved August 2, 2005, from symbols. Cambridge, MA: MIT Press. http://www-formal.stanford.edu/jmc/chinese. Rey,G.(1997). A question about consciousness. html. In N. Block, O. Flanagan, & G. Guzeldere¨ McCarthy, J. & Hayes, P. (1969). Some philo- (Eds.), The nature of consciousness: Philosophical sophical problems from the standpoint of arti- debates (pp. 461–482). Cambridge, MA: MIT ficial intelligence. In B. Meltzer, & D. Michie Press. (Eds.), Machine intelligence 4 (pp. 463–502). Rumelhart, D. E., McClelland, J. L., & the PDP Edinburgh: Edinburgh University Press. Research Group (1986). Parallel distributed McDermott, D. (2001). Mind and mechanism. processing: Explorations in the microstructure of Cambridge, MA: MIT Press. cognition. Cambridge, MA: MIT Press. Meltzer, B., & Michie, D. (Eds.). (1969). Machine Said, K. M., Newton-Smith, W., Viale, R., & intelligence 4. Edinburgh: Edinburgh University Wilkes, K. (1990). Modelling the mind. Oxford: Press. Clarendon Press. Minsky, M. (Ed.). (1968a). Semantic information Scheutz, M. (Ed.) (2002). Computationalism: processing. Cambridge, MA: MIT Press. New directions. Cambridge, MA: The MIT Minsky, M. (1968b). Matter, mind, and models. Press. In M. Minsky (Ed.), Semantic information pro- Searle, J. R. (1980). Minds, brains, and program. cessing (pp. 425–432). Cambridge, MA: MIT The Behavioral and Brain Sciences, 3, 417–424. Press. Searle, J. R. (1990). Is the brain’s mind a computer Minsky, M. (1986). The society of mind. New York: program? Scientific American, 262, 26–31. Simon and Schuster. Searle, J. R. (1992). The rediscovery of the mind. Moody, T. C. (1994). Conversations with zom- Cambridge, MA: MIT Press. bies. Journal of Consciousness Studies, 1(2), 196– Sloman, A. (1978). The computer revolution in phi- 200. losophy. Hassocks, Sussex: Harvester Press. Moore, G. E. (1965). Cramming more com- Sloman, A. (2002). The irrelevance of Tur- ponents onto integrated circuits. Electronics, ing machines to artificial intelligence. In M. 38(8), pp. 114–117. Scheutz, M, (Ed.), Computationalism: New P1: JzZ 0521857430c06 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:22

150 the cambridge handbook of consciousness

directions (pp. 87–127). Cambridge, MA: The lism: New directions (pp. 23–58). Cambridge, MIT Press. MA: The MIT Press. Sloman, A., & Chrisley, R. (2003). Virtual Turing, A. (1950). Computing machinery and machines and consciousness. Journal of Con- intelligence. Mind, 49, 433–460. sciousness Studies, 10(4–5), 6–45. Wilkes, K. (1990). Modelling the mind. In K. M. Smith, B. C. (1984). Reﬂection and semantics in Said, W. Newton-Smith, R. Viale, & K. Wilkes Lisp. Proceedings of the Conference on Principles (Eds.), Modelling the mind (pp. 63–82). Oxford: of Programming Languages, 11, 23–35. Clarendon Press. Smith, B. C. (1995). On the origins of objects. Cam- Winograd, T. (1972). Understanding natural lan- bridge, MA: MIT Press. guage. New York: Academic Press. Smith, B. C. (2002). The foundations of comput- Wittgenstein, L. (1953). Philosophical investiga- ing. In M. Scheutz, M, (Ed.), Computationa- tions. New York: MacMillan. P1: JzG 0521857430c07 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 20:23

CHAPTER 7 Computational Models of Consciousness: A Taxonomy and Some Examples

Ron Sun and Stan Franklin

Abstract Introduction

This chapter aims to provide an overview of In this chapter, we aim to present a short sur- existing computational (mechanistic) mod- vey and a brief evaluation of existing compu- els of cognition in relation to the study of tational (mechanistic) models of cognition consciousness, on the basis of psychologi- in relation to the study of consciousness. cal and philosophical theories and data. It The survey focuses on their explanations examines various mechanistic explanations of the difference between conscious and of consciousness in existing computational unconscious cognitive processes on the basis cognitive models. Serving as an example for of psychological and philosophical theo- the discussions, a computational model of ries and data, as well as potential practical the conscious/unconscious interaction, uti- applications. lizing the representational difference expla- Given the plethora of models, theories, nation of consciousness, is described brieﬂy. and data, we try to provide in this chapter an As a further example, a software agent overall (and thus necessarily sketchy) exam- model that captures another explanation ination of computational models of con- of consciousness (the access explanation of sciousness in relation to the available psy- consciousness) is also described. The discus- chological data and theories, as well as the sions serve to highlight various possibilities existing philosophical accounts. We come to in developing computational models of con- some tentative conclusions as to what a plau- sciousness and in providing computational sible computational account should be like, explanations of conscious and unconscious synthesizing various operationalized psycho- cognitive processes. logical notions related to consciousness.

151 P1: JzG 0521857430c07 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 20:23

152 the cambridge handbook of consciousness

We begin by examining some foun- On the other hand, the necessity of mech- dational issues concerning computational anistic explanations, according to the fore- approaches toward consciousness. Then, var- going definition of mechanistic processes, ious existing models and their explanations should be intuitively obvious to anyone who of the conscious/unconscious distinction are is not a dualist. If one accepts the universal- presented. After examining a particular ity of computation, then computation, in its model embodying a two-system approach, broadest sense, can be expected to include we look at one embodying a unified (one- the necessary conditions for consciousness. system) approach and then at a few addi- On the basis of such intuition, we need tional models. to provide an explanation of the computational/mechanistic basis of consciousness that answers the following questions. What Computational Explanations kind of mechanism leads to conscious proof Consciousness cesses, and what kind of mechanism leads to unconscious processes? What is the func- Work in the area of computational mod- tional role of conscious processes (Baars, eling of consciousness generally assumes 1988, 2002; Sun, 1999a, b)? What is the the sufficiency and the necessity of mech- functional role of unconscious processes? anistic explanations. By mechanistic expla- There have been many such explanations in nation, we mean any concrete computa- computational or mechanistic terms. These tional processes, in the broadest sense of computational or mechanistic explanations the term “computation.” In general, com- are highly relevant to the science of con- putation is a broad term that can be used sciousness as they provide useful theoretical to denote any process that can be realized frameworks for further empirical work. on generic computing devices, such as Tur- Another issue we need to address before ing machines (or even beyond if there is we move on to details of computational such a possibility). Thus, mechanistic expla- work is the relation between biological/phy- nations may utilize, in addition to standard siological models and computational mod- computational notions, a variety of other els in general. The problem with biologi- conceptual constructs ranging, for example, cally centered studies of consciousness in from chaotic dynamics (Freeman, 1995), to general is that the gap between phenomenol- “Darwinian” competition (Edelman, 1989), ogy and physiology/biology is so great that and to quantum mechanics (Penrose, 1994). something else may be needed to bridge it. (We leave out the issue of complexity Otherwise, if we rush directly into com- for now.) plex neurophysiological thickets (Edelman, In terms of the sufficiency of mechanistic 1989; Crick & Koch, 1990; Damasio et al., explanations, a general working hypothesis is 1990; LeDoux, 1992,), we may lose sight succinctly expressed by the following state- of the forests. Computation, in its broad- ment (Jackendoff, 1987): est sense, can serve to bridge the gap. It provides an intermediate level of explana- Hypothesis of computational sufficiency: tion in terms of processes, mechanisms, and every phenomenological distinction is caused by/supported by/projected from a functions and helps determine how various corresponding computational distinction. aspects of conscious and unconscious processes should figure into the architecture of For the lack of a clearly better alterna- the mind (Anderson & Lebiere, 1998; Sun, tive, this hypothesis remains a viable work- 2002). It is possible that an intermediate ing hypothesis in the area of computational level between phenomenology and physi- models of consciousness, despite various ology/neurobiology might be more apt to criticisms (e.g., Damasio, 1994; Edelman, capture fundamental characteristics of con- 1989; Freeman, 1995; Penrose, 1994; Searle, sciousness (Coward & Sun, 2004). This no- 1980). tion of an intermediate level of explanation P1: JzG 0521857430c07 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 20:23

computational models of consciousness: a taxonomy and some examples 153

has been variously expounded recently; for The difference lies in the two different example, in terms of virtual machines by ways of organizing knowledge – whether Sloman and Chrisley (2003). in an action-centered way (procedural knowledge) or in an action-independent way (declarative knowledge). Computa- Different Computational Accounts tionally, both types of knowledge are of Consciousness represented symbolically (using either symbolic semantic networks or symbolic 1 Existing computational explanations of the production rules). The semantic net- conscious/unconscious distinction may be works use parallel spreading activation categorized based on the following differ- (Collins & Loftus, 1975) to activate rel- ent emphases: (1) differences in knowledge evant nodes, and the production rules organization (e.g., the SN+PS view, to be compete for control through parallel detailed later), (2) differences in knowledge- matching and ﬁring. The models embody- processing mechanisms (e.g., the PS+SN ing this view have been used for model- view), (3) differences in knowledge content ing a variety of psychological tasks, espe- (e.g., the episode+activation view), (4) dif- cially skill learning tasks (Anderson, 1983, ferences in knowledge representation (e.g., Anderson & Lebiere, 1998). + 5 r the localist distributed view), or ( ) dif- The PS+SN view: an instance of the ferent processing modes of the same sys- explanations based on differences in tem (e.g., the attractor view or the threshold knowledge-processing mechanisms. As view). proposed by Hunt and Lansman (1986), Contrary to some critics, the debate the “deliberate” computational process of among these differing views is not analogous production matching and ﬁring in a pro- to a debate between algebraists and geome- duction system (PS), which is serial in ters in physics (which would be irrelevant). this case, is assumed to be a conscious It is more analogous to the wave vs. parti- process, whereas the spreading activation cle debate in physics concerning the nature computation (Collins & Loftus, 1975)in of light, which was truly substantive. Let us semantic networks (SN), which is mas- discuss some of the better known views con- sively parallel, is assumed to be an uncon- cerning computational accounts of the conscious process. The model based on this scious/unconscious distinction one by one. view has been used to model controlled First of all, some explanations are based and automatic processing data in the on recognizing that there are two sepa- attention-performance literature (Hunt rate systems in the mind. The difference & Lansman, 1986). Note that this view is between the two systems can be explained the exact opposite of the view advocated in terms of differences in either knowledge by Anderson (1983), in terms of the roles organization, knowledge-processing mech- of the two computational mechanisms anisms, knowledge content, or knowledge involved. Note also that the emphasis in representation: this view is on the processing difference r of the two mechanisms, serial vs. parallel, The SN+PS view: an instance of the expla- and not on knowledge organization. nations based on differences in knowl- r edge organization. As originally proposed The algorithm+instance view: another ins- by Anderson (1983) in his ACT* model, tance of the explanations based on dif- there are two types of knowledge: Declar- ferences in knowledge-processing mech- ative knowledge is represented by seman- anisms. As proposed by Logan (1988) tic networks (SN), and it is consciously and also by Stanley et al. (1989), the accessible, whereas procedural knowl- computation involved in retrieval and edge is represented by rules in a produc- use of instances of past experience is tion system (PS), and it is inaccessible. considered to be unconscious (Stanley P1: JzG 0521857430c07 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 20:23

154 the cambridge handbook of consciousness

et al., 1989) or automatic (Logan 1988), vated views that rely on the interplay whereas the use of “algorithms” involves of various memory systems, such as that conscious awareness. Here the term advocated by Taylor (1997) and McClel- “algorithm” is not clearly defined and land et al. (1995). r apparently refers to computation more The localist+distributed representation complex than instance retrieval/use. view: an instance of the explanations Computationally, it was suggested that based on differences in knowledge rep- the use of an algorithm is under tight resentation. As proposed by Sun (1994, control and carried out in a serial, 2002), different representational forms step-by-step way, whereas instances can used in different components may be be retrieved in parallel and effortlessly used to explain the qualitative difference (Logan, 1988). The emphasis here is between conscious and unconscious again on the differences in process- processes. One type of representation is ing mechanisms. This view is also sim- symbolic or localist, in which one distinct ilar to the view advocated by Neal entity (e.g., a node in a connectionist and Hesketh (1997), which emphasizes model) represents a concept. The other the unconscious influence of what they type of representation is distributed, in called episodic memory. Note that the which a non-exclusive set of entities views by Logan (1988), Stanley et al. (e.g., a set of nodes in a connectionist (1989), and Neal and Hesketh (1997) are model) are used for representing one the exact opposite of the view advo- concept, and the representations of cated by Anderson (1983) and Bower different concepts overlap each other; (1996), in which instances/episodes are in other words, a concept is represented consciously accessed rather than uncon- as a pattern of activations over a set of sciously accessed. entities (e.g., a set of nodes). Conceptual r The episode+activation view: an instance structures (e.g., rules) can be imple- of the explanations based on differences mented in the localist/symbolic system in knowledge content. As proposed by in a straightforward way by connections Bower (1996), unconscious processes are between relevant entities. In distributed based on activation propagation through representations, such structures (includ- strengths or weights (e.g., in a connec- ing rules) are diffusely duplicated in a tionist fashion) between different nodes way consistent with the meanings of the representing perceptual or conceptual structures (Sun, 1994), which captures primitives, whereas conscious processes unconscious performance. There may are based on explicit episodic memory be various connections between corre- of past episodes. What is emphasized sponding representations across the two in this view is the rich spatial-temporal systems. (A system embodying this view, context in episodic memory (i.e., the ad CLARION, is described later.) hoc associations with contextual information, acquired on a one-shot basis), which is termed type-2 associations as In contrast to these two-systems views, opposed to regular type-1 associations there exist some theoretical views that insist (which are based on semantic related- on the unitary nature of the conscious and ness). This emphasis somewhat distin- the unconscious. That is, they hold that con- guishes this view from other views conscious and unconscious processes are differ- cerning instances/episodes (Logan, 1988; ent manifestations of the same underlying Neal & Hesketh, 1997; Stanley et al. system. The difference between conscious 1989).2 The reliance on memory of spe- and unconscious processes lies in the differ- cific events in this view bears some resem- ent processing modes for conscious versus blance to some neurobiologically moti- unconscious information within the same P1: JzG 0521857430c07 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 20:23

computational models of consciousness: a taxonomy and some examples 155

system. There are several possibilities in this role of internal consistency in producing regard: consciousness. There has been support r for this possibility from neuroscience, for The threshold view: As proposed by var- example, in terms of a coherent “thalamo- ious researchers, including Bowers et al. cortical core” (Edelman & Tononi, 2000). (1990), the difference between con- r The access view: As suggested by Baars scious and unconscious processes can be (1988), consciousness is believed to help explained by the difference between acti- mobilize and integrate mental functions vations of mental representations above that are otherwise disparate and inde- a certain threshold and activations of pendent. Thus, consciousness is aimed at such representations below that thresh- solving the relevance problem – ﬁnding old. When activations reach the threshold the exact internal resources needed to level, an individual becomes aware of the deal with the current situation. Some evi- content of the activated representations; dence has been accumulated for this view otherwise, although the activated repre- (Baars, 2002). A computational imple- sentations may inﬂuence behavior, they mentation of Baars’ theory in the form will not be accessible consciously. r of IDA (a running software agent system; The chunking view: As in the mod- Franklin et al., 1998) is described in els described by Servan-Schreiber and detail later. See also Coward and Sun 1987 Anderson ( ) and by Rosenbloom (2004). et al. (1993), a chunk is considered a unitary representation and its internal work- The coexistence of these various views ing is opaque (although its input/output of consciousness seems quite analogous to are accessible). A chunk can be a pro- the parable of the Blind Men and the Ele- duction rule (as in Rosenbloom et al., phant. Each of them captures some aspect of 1993) or a short sequence of perceptual- the truth about consciousness, but the por- motor elements (as in Servan-Schreiber tion of the truth captured is limited by the & Anderson, 1987). Because of the lack view itself. None seems to capture the whole of transparency of the internal working picture. of a chunk, it is equated with implicit In the next two sections, we look into learning (Servan-Schreiber & Anderson, some details of two representative compu- 1987) or automaticity (Rosenbloom et al., tational models, exemplifying either two- 1993). According to this view, the differ- system or one-system views. The models ence between conscious and unconscious illustrate what a plausible computational processes is the difference between using model of consciousness should be like, syn- multiple (simple) chunks (involving some thesizing various psychological notions and consciousness) and using one (complex) relating to various available psychological chunk (involving no consciousness). theories. r The attractor view: As suggested by the model of Mathis and Mozer (1996), being in a stable attractor of a dynamic system A Model Adopting the (a neural network in particular) leads to Representational Difference View consciousness. The distinction between conscious and unconscious processes is Let us look into the representational dif- reduced to the distinction of being in a ference view as embodied in the cogni- stable attractor and being in a transient tive architecture Clarion (which stands for state. O’Brien and Opie (1998) proposed Connectionist Learning with Rule Induction an essentially similar view. This view may ON-line; Sun 1997, 2002, 2003), as an exam- be generalized to a general coherence ple of the two-system views for explaining view – the emphasis may be placed on the consciousness. P1: JzG 0521857430c07 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 20:23

156 the cambridge handbook of consciousness

Top Level interpretive processes), thus exhibiting different psychological properties (see, e.g., explicit representation Berry & Broadbent, 1988; Reber, 1989; more discussions later). explicit knowledge. An example model in this regard is explicit learning process Clarion, which is a two-level model that uses the localist and distributed representations in the two levels, respectively, and learns using two different methods in the two levels, respectively. In developing the model, four criteria were hypothesized (see Sun, 1994), on the basis of the aforemen- implicit representation tioned considerations: (1) direct accessibility implicit knowledge of conscious processes; (2) direct inaccessi- implicit learning process bility of unconscious processes; and furthermore, (3) linkages from localist concepts to distributed features: once a localist concept Bottom Level is activated, its corresponding distributed Figure 7.1. The CLARION model. representations (features) are also activated, as assumed in most cognitive models, ranging from Tversky (1977) to Sun (1995);4 The important premises of subsequent and (4) linkages from distributed features discussions are the direct accessibility of con- to localist concepts: under appropriate cir- scious processes and the direct inaccessibility cumstances, once some or most of the dis- of unconscious processes. Conscious pro- tributed features of a concept are activated, cesses should be directly accessible – that the localist concept itself can be activated to is, directly verbally expressible – without “cover” these features (roughly correspond- involving intermediate interpretive or trans- ing to categorization; Smith & Medin, 1981). formational steps, which is a requirement The direct inaccessibility of unconscious prescribed and/or accepted by many the- knowledge can be best captured by a “sub- oreticians (see, e.g., Clark, 1992; Hadley, symbolic” distributed representation such as 1995).3 Unconscious processes should be, that provided by a backpropagation network in contrast, inaccessible directly (but they (Rumelhart et al., 1986), because represen- might be accessed indirectly through some tational units in a distributed representation

Dimensions bottom top Cognitive phenomena implicit learning explicit learning implicit memory explicit memory automatic processing controlled processing intuition explicit reasoning Source of knowledge trial-and-error external sources assimilation of explicit knowledge extraction from the bottom level Representation distributed (micro) features localist conceptual units Operation similarity-based explicit symbol manipulation Characteristics more context sensitive, fuzzy more crisp, precise less selective more selective more complex simpler

Figure 7.2 . Comparisons of the two levels of the CLARION architecture. P1: JzG 0521857430c07 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 20:23

computational models of consciousness: a taxonomy and some examples 157

ACS NACS

action–centered non–action–centered explicit representation explicit representation

action–centered implicit non–action–centered representation implicit representation

goal structure reinforcement

goal setting

filtering drives selection regulation

MS MCS Figure 7.3. The implementation of CLARION. ACS denotes the action-centered subsystem, NACS the non-action-centered subsystem, MS the motivational subsystem, and MCS the metacognitive subsystem. The top level contains localist encoding of concepts and rules. The bottom level contains multiple (modular) connectionist networks for capturing unconscious processes. The interaction of the two levels and the information ﬂows are indicated with arrows.

are capable of accomplishing tasks but ing, the bottom level uses gradual weight are generally uninterpretable directly (see tuning, whereas the top level uses explicit, Rumelhart et al., 1986; Sun, 1994). In con- one-shot hypothesis testing learning, in cor- trast, conscious knowledge can be cap- respondence with the representational char- tured in computational modeling by a sym- acteristics of the two levels. There are var- bolic or localist representation (Clark & ious connections across the two levels for Karmiloff-Smith, 1993; Sun & Bookman exerting mutual inﬂuences. See Figure 7.1 1994), in which each unit has a clear concep- for an abstract sketch of the model. The dif- tual meaning/interpretation (i.e., a semantic ferent characteristics of the two levels are label). This captures the property of con- summarized in Figure 7.2. scious processes being directly accessible and Let us look into some implementational manipulable (Smolensky, 1988; Sun, 1994). details of Clarion. Note that the details This difference in representation leads to a of the model have been described exten- two-level structure whereby each level uses sively in a series of previous papers, includ- one type of representation (Sun, 1994, 1995, ing Sun (1997, 2002, 2003), Sun and Peter- 1997; Sun et al., 1996, 1998, 2001). The bot- son (1998), and Sun et al. (1998, 2001). tom level is based on distributed represen- It has a dual representational structure – tation, whereas the top level is based on implicit and explicit representations being in localist/symbolic representation. For learn- two separate “levels” (Hadley, 1995; Seger, P1: JzG 0521857430c07 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 20:23

158 the cambridge handbook of consciousness

1994). Essentially it is a dual-process the- ing the action with the highest Q-value). To ory of mind (Chaiken & Trope, 1999). It acquire the Q-values, one may use the Q- also consists of a number of functional sub- learning algorithm (Watkins 1989), a rein- systems, including the action-centered sub- forcement learning algorithm. It basically system, the non-action-centered subsystem, compares the values of successive actions the metacognitive subsystem, and the moti- and adjusts an evaluation function on that vational subsystem (see Figure 7.3). basis. It thereby develops reactive sequential Let us first focus on the action-centered behaviors. subsystem of Clarion. In this subsystem, the The bottom level of the action-centered two levels interact by cooperating in actions, subsystem is modular; that is, a number of through a combination of the action rec- small neural networks coexist, each of which ommendations from the two levels, respec- is adapted to specific modalities, tasks, or tively, as well as cooperating in learning groups of input stimuli. This coincides with through a bottom-up and a top-down pro- the modularity claim (Baars, 1988; Cosmides cess (to be discussed below). Actions and & Tooby, 1994; Edelman, 1987; Fodor, 1983; learning of the action-centered subsystem Hirschfield & Gelman, 1994; Karmiloff- may be described as follows: Smith, 1986) that much processing in the human mind is done by limited, encapsu- 1. Observe the current state x. lated (to some extent), specialized proces- 2. Compute in the bottom level the “values” sors that are highly effcient. Some of these of x associated with each of all the possi- modules are formed evolutionarily; that is, ble actions ai’s: Q(x, a1), Q(x, a2 ),...... , given a priori to agents, reflecting their hard- Q(x, an) (to be explained below). wired instincts and propensities (Hirsch- 1994 3. Find out all the possible actions (b1, field & Gelman, ). Some of them can be learned through interacting with the world b2 ,....,bm) at the top level, based on the input x (sent up from the bottom level) (computationally through various decompo- 1999 and the rules in place. sition methods; e.g., Sun & Peterson, ). In the top level of the action-centered 4. Compare or combine the values of the subsystem, explicit conceptual knowledge a s with those of b s (sent down from i j is captured in the form of rules. Symbolic/ the top level), and choose an appropriate localist representations are used. See Sun action b. (2003) for further details of encoding (they 5 b . Perform the action , and observe the next are not directly relevant here). y r state and (possibly) the reinforcement . Humans are clearly able to learn implicit 6 . Update Q-values at the bottom level knowledge through trial and error, without in accordance with the Q-Learning- necessarily utilizing a priori explicit knowl- Backpropagation algorithm (to be edge (Seger, 1994). On top of that, explicit explained later). knowledge can be acquired, also from ongo- 7. Update the rule network at the top level ing experience in the world, and possibly th- using the Rule-Extraction-Refinement algo- rough the mediation of implicit knowledge rithm (to be explained later). (i.e., bottom-up learning; see Karmilof- 8. Go back to Step 1. Smith, 1986; Stanley et al., 1989; Sun, 1997, 2002; Willingham et al., 1989). The basic In the bottom level of the action-centered process of bottom-up learning is as follows subsystem, implicit reactive routines are (Sun, 2002). If an action decided by the learned: A Q-value is an evaluation of the bottom level is successful, then the agent “quality” of an action in a given state: Q(x, extracts a rule that corresponds to the action a) indicates how desirable action a is in state selected by the bottom level and adds the x (which consists of some sensory input). rule to the top level. Then, in subsequent The agent may choose an action in any state interaction with the world, the agent ver- based on Q-values (for example, by choos- ifies the extracted rule by considering the P1: JzG 0521857430c07 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 20:23

computational models of consciousness: a taxonomy and some examples 159

outcome of applying the rule: If the outcome action-centered knowledge (Sun, 1994). In is not successful, then the rule should be this network, chunks are specified through made more specific and exclusive of the cur- dimensional values. A node is set up at the rent case, and if the outcome is successful, top level to represent a chunk. The chunk the agent may try to generalize the rule to node (a symbolic representation) connects make it more universal (e.g., Michalski, to its corresponding features (dimension- 1983). The details of the bottom-up learning value pairs) represented as nodes in the algorithm (the Rule-Extraction-Refinement bottom level (which form a distributed algorithm) can be found in Sun and Peter- representation). Additionally, links between son (1998). After rules have been learned, a chunks at the top level encode explicit asso- variety of explicit reasoning methods may be ciations between pairs of chunks, known as used. Learning explicit conceptual represen- associative rules. Explicit associative rules tation at the top level can also be useful in may be formed (i.e., learned) in a variety of enhancing learning of implicit reactive rou- ways (Sun, 2003). tines (reinforcement learning) at the bottom On top of associative rules, similarity- level. based reasoning may be employed in Although Clarion can learn even when the non-action-centered subsystem. Dur- no a priori or externally provided knowl- ing reasoning, a known (given or inferred) edge is available, it can make use of it when chunk may be automatically compared with such knowledge is available (cf. Anderson, another chunk. If the similarity between 1983; Schneider & Oliver, 1991). To deal them is sufficiently high, then the latter with instructed learning, externally provided chunk is inferred (see Sun, 2003, for details). knowledge (in the forms of explicit concep- Similarity-based and rule-based reasoning tual structures, such as rules, plans, routines, can be intermixed. As a result of mixing categories, and so on) should (1) be com- similarity-based and rule-based reasoning, bined with autonomously generated concep- complex patterns of reasoning emerge. As tual structures at the top level (i.e., internal- shown by Sun (1994), different sequences ization) and (2) be assimilated into implicit of mixed similarity-based and rule-based reactive routines at the bottom level (i.e., reasoning capture essential patterns of assimilation). This process is known as top- human everyday (mundane, common-sense) down learning. See Sun (2003) for further reasoning. details. As in the action-centered subsystem, The non-action-centered subsystem rep- top-down or bottom-up learning may take resents general knowledge about the world, place in the non-action-centered subsystem, which is equivalent to the notion of seman- either to extract explicit knowledge in the tic memory (as in, e.g., Quillian, 1968). It top level from the implicit knowledge in the may be used for performing various kinds bottom level or to assimilate explicit knowl- of retrievals and inferences. It is under the edge of the top level into implicit knowledge control of the action-centered subsystem in the bottom level. (through the actions of the action-centered The motivational subsystem is concerned subsystem). At the bottom level, associa- with drives and their interactions (Toates, tive memory networks encode non-action- 1986). It is concerned with why an agent centered implicit knowledge. Associations does what it does. Simply saying that an are formed by mapping an input to an out- agent chooses actions to maximizes gains, put. The regular backpropagation learning rewards, or payoffs leaves open the quest- algorithm can be used to establish such asso- ion of what determines these things. The re- ciations between pairs of input and output levance of the motivational subsystem to the (Rumelhart et al., 1986). action-centered subsystem lies primarily in On the other hand, at the top level of the fact that it provides the context in which the non-action-centered subsystem, a gen- the goal and the payoff of the action-cente- eral knowledge store encodes explicit non- red subsystem are set. It thereby influences P1: JzG 0521857430c07 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 20:23

160 the cambridge handbook of consciousness

the working of the action-centered subsys- tures (in the action-centered subsystem), tem and, by extension, the working of the and so on. See Sun (2003) for further non-action-centered subsystem. details of these memories. As touched upon A bipartite system of motivational rep- before, these memories are important for resentation is again in place in Clarion. accounting for various forms of conscious The explicit goals (such as “finding food”) and unconscious processes (also see, e.g., of an agent (which is tied to the working McClelland et al., 1995; Schacter, 1990; of the action-centered subsystem) may be Taylor, 1997). generated based on internal drive states (for Clarion has been successful in account- example, “being hungry”). See Sun (2003) ing for a variety of psychological data. A for details. number of well-known skill learning tasks Beyond low-level drives concerning phys- have been simulated using Clarion; these iological needs, there are also higher-level span the spectrum ranging from simple reac- drives. Some of them are primary, in the tive skills to complex cognitive skills. The sense of being “hardwired.” For example, tasks include serial reaction time (SRT) tasks, Maslow (1987) developed a set of these artificial grammar learning (AGL) tasks, pro- drives in the form of a “need hierarchy.” cess control (PC) tasks, the categorical infer- Whereas primary drives are built-in and rel- ence (CI) task, the alphabetical arithmetic atively unalterable, there are also “derived” (AA) task, and the Tower of Hanoi (TOH) drives, which are secondary, changeable, and task (see Sun, 2002). Among them, SRT, acquired mostly in the process of satisfying AGL, and PC are typical implicit learning primary drives. tasks, very much relevant to the issue of con- The metacognitive subsystem is closely sciousness as they operationalize the notion tied to the motivational subsystem. The of consciousness in the context of psycho- metacognitive subsystem monitors, controls, logical experiments (Coward & Sun, 2004; and regulates cognitive processes for the sake Reber, 1989; Seger, 1994; Sun et al., 2005), of improving cognitive performance (Nel- whereas TOH and AA are typical high-level son, 1993; Sloman & Chrisley, 2003; Smith cognitive skill acquisition tasks. In addition, et al., 2003). Control and regulation may be extensive work have been done on a com- in the forms of setting goals for the action- plex minefield navigation task (see Sun & centered subsystem, setting essential param- Peterson, 1998; Sun et al., 2001). Metacogni- eters of the action-centered and the non- tive and motivational simulations have also action-centered subsystem, interrupting and been undertaken, as have social simulation changing ongoing processes in the action- tasks (e.g., Sun & Naveh, 2004). centered and the non-action-centered sub- In evaluating the contribution of Clarion system, and so on. Control and regulation to our understanding of consciousness, we may also be carried out through setting rein- note that the simulations using Clarion forcement functions for the action-centered provide detailed, process-based interpreta- subsystem on the basis of drive states. The tions of experimental data related to con- metacognitive subsystem is also made up of sciousness, in the context of a broadly two levels: the top level (explicit) and the scoped cognitive architecture and a uni- bottom level (implicit). fied theory of cognition. Such interpreta- Note that in Clarion, there are thus a tions are important for a precise, process- variety of memories: procedural memory based understanding of consciousness and (in the action-centered subsystem) in both other aspects of cognition, leading to bet- implicit and explicit forms, general “seman- ter appreciations of the role of consciousness tic” memory (in the non-action-centered in human cognition (Sun, 1999a). Clarion subsystem) in both implicit and explicit also makes quantitative and qualitative pre- forms, episodic memory (in the non-action- dictions regarding cognition in the areas of centered subsystem), working memory (in memory, learning, motivation, metacogni- the action-centered subsystem), goal struction, and so on. These predictions either P1: JzG 0521857430c07 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 20:23

computational models of consciousness: a taxonomy and some examples 161

have been experimentally tested already or of all other information in the hippocam- are in the process of being tested (see, e.g., pus and with the ongoing events. Weights Sun, 2002; Sun et al., 2001, 2005). Because are adjusted by a small amount after each of the complex structures and their com- experience, so that the overall direction of plex interactions specified within the frame- weight change is governed by the struc- work of Clarion, it has a lot to say about ture present in the ensemble of events and the roles that different types of processes, experiences, using distributed representa- conscious or unconscious, play in human tions (with weights). Therefore, catastrophic cognition, as well as their synergy (Sun et al., interference is avoided in cortical systems. 2005). This model is very similar to the two-level Comparing Clarion with Bower (1996), idea of Clarion, in that it not only adopts the latter may be viewed as a special case of a two-system view but also utilizes repre- Clarion for dealing specifically with implicit sentational differences between the two sys- memory phenomena. The type-1 and type-2 tems. However, in contrast to this model, connections, hypothesized by Bower (1996) which captures only what may be termed as the main explanatory constructs, can be top-down learning (that is, learning that pro- equated roughly to top-level representations ceeds from the conscious to the uncon- and bottom-level representations, respec- scious), Clarion can capture both top-down tively. In addition to making the distinc- learning (from the top level to the bottom tion between type-1 and type-2 connections, level) and bottom-up learning (from the bot- Bower (1996) also endeavored to specify tom level to the top level). See Sun et al. the details of multiple pathways of spread- (2001) and Sun (2002) for details of bottom- ing activation in the bottom level. These up learning. pathways were phonological, orthographi- Turning to the declarative/procedural cal, semantic, and other connections that knowledge models, ACT* (Anderson, 1983) store long-term implicit knowledge. In the is made up of a semantic network (for declar- top level, associated with type-2 connec- ative knowledge) and a production system tions, it was claimed on the other hand (for procedural knowledge). ACT-R is a that rich contextual information was stored. descendant of ACT*, in which procedural These details nicely complement the speci- learning is limited to production formation fication of Clarion and can thus be incorpo- through mimicking, and production firing rated into the model. is based on log odds of success. Clarion The proposal by McClelland et al. (1995) succeeds in explaining two issues that ACT that there are complementary learning sys- did not address. First, whereas ACT takes tems in the hippocampus and neocortex a mostly top-down approach toward learn- is also relevant here. According to their ing (i.e, from given declarative knowledge account, cortical systems learn slowly, and to procedural knowledge), Clarion can pro- the learning of new information destroys ceed bottom-up. Thus, Clarion can account the old, unless the learning of new infor- for implicit learning better than ACT (see mation is interleaved with ongoing expo- Sun, 2002, for details). Second, in ACT sure to the old information. To resolve these both types of knowledge are represented in two problems, new information is initially explicit, symbolic forms (i.e., semantic net- stored in the hippocampus, an explicit mem- works and productions), and thus it does ory system, in which crisp, explicit repre- not explain, from a representational view- sentations are used to minimize interference point, the differences in conscious accessibil- of information (so that catastrophic inter- ity (Sun, 1999b). Clarion accounts for this ference is avoided there). It allows rapid difference based on the use of two different learning of new material. Then, the new forms of representation. Top-level knowl- information stored in the hippocampus is edge is represented explicitly and thus con- assimilated into cortical systems. The assim- sciously accessible, whereas bottom-level ilation is interleaved with the assimilation knowledge is represented implicitly and P1: JzG 0521857430c07 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 20:23

162 the cambridge handbook of consciousness

thus inaccessible. Thus, this distinction in must communicate with sailors via e-mail Clarion is intrinsic, instead of assumed as and in English, understanding the con- in ACT (Sun, 1999b). tent and producing human-like responses. Comparing Clarion with Hunt and Lans- It must access a number of existing Navy man’s (1986) model, there are similari- databases, again understanding the content. ties. The production system in Hunt and It must see that the Navy’s needs are satisfied Lansman’s model clearly resembles the top while adhering to Navy policies. For exam- level in Clarion, in that both use explicit ple, a particular ship may require a certain manipulations in much the same way. Like- number of sonar technicians with the req- wise, the spreading activation in the seman- uisite types of training. It must hold down tic network in Hunt and Lansman’s model moving costs. And it must cater to the needs resembles the connectionist network in the and desires of the sailor as well as possi- bottom level of Clarion, because the same ble. This includes negotiating with the sailor kind of spreading activation was used in via an e-mail correspondence in natural lan- both models, although the representation in guage. Finally, it must authorize the finally Hunt and Lansman’s model was symbolic, selected new billet and start the writing of not distributed. Because of the uniformly the sailor’s orders. symbolic representations used in Hunt and Although the IDA model was not initially Lansman’s model, it does not explain con- developed to reproduce experimental data, vincingly the qualitative difference between it is nonetheless based on psychological conscious and unconscious processes (see and neurobiological theories of conscious- Sun, 1999b). ness and does generate hypotheses and qualitative predictions (Baars & Franklin, 2003: Franklin et al., 2005). IDA successfully An Application of the Access View implements much of the global workspace theory (Baars, 1988), and there is a growing Let us now examine an application of the body of empirical evidence supporting that access view on consciousness in building a theory (Baars, 2002). IDA’s flexible cogni- practically useful system. The access view is tive cycle has also been used to analyze the a rather popular approach in computational relation of consciousness to working mem- accounts of consciousness (Baars, 2002), and ory at a fine level of detail, offering explana- therefore it deserves some attention. It is also tions of such classical working memory tasks presented here as an example of various one- as visual imagery to gain information and the system views. rehearsal of a telephone number (Baars & Most computational models of cognitive Franklin, 2003: Franklin et al., 2005). processes are designed to predict experi- In his global workspace theory (see Fig- mental data. IDA (Intelligent Distribution ure 7.4 and Chapter 8), Baars (1988) postu- Agent), in contrast, models consciousness lates that human cognition is implemented in the form of an autonomous software by a multitude of relatively small, special- agent (Franklin & Graesser, 1997). Specif- purpose processors, which are almost always ically, IDA was developed for Navy appli- unconscious (i.e., the modularity hypoth- cations (Franklin et al., 1998). At the end esis as discussed earlier). Communication of each sailor’s tour of duty, he or she is between them is rare and over a narrow assigned to a new billet in a process called bandwidth. Coalitions of such processes find distribution. The Navy employs almost 300 their way into a global workspace (and people (called detailers) to effect these new thereby into consciousness). This limited assignments. IDA’s task is to play the role of capacity workspace serves to broadcast the a detailer. message of the coalition to all the uncon- Designing IDA presents both communi- scious processors in order to recruit other cation problems and action selection prob- processors to join in handling the current lems involving constraint satisfaction. It novel situation or in solving the current P1: JzG 0521857430c07 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 20:23

computational models of consciousness: a taxonomy and some examples 163

The Dominant Context Hierarchy: Goal Contexts

Conceptual Contexts

Perceptual Contexts Competing Contexts:

Input Processors:

Global Workspace (conscious)

Other Available Contexts:

Figure 7.4. Baars’ global workspace theory.

problem. Thus consciousness, in this theory, dent threads, each of which is specialized for allows us to deal with novel or problem- some relatively simple task. They often play atic situations that cannot be dealt with eff- the role of “demons,”5 waiting for a particu- ciently, or at all, by habituated unconscious lar situation to occur in response to which processes. In particular, it provides access they should act. Codelets also correspond to appropriately useful resources. Global more or less to Edelman’s neuronal groups workspace theory offers an explanation for (Edelman, 1987) or Minsky’s agents (Minsky, the limited capacity of consciousness. Large 1985). Codelets come in a number of vari- messages would be overwhelming to tiny eties, each with different functions to per- processors. In addition, all activities of these form. Most of these codelets subserve some processors take place under the auspices of high-level entity, such as a behavior. How- contexts: goal contexts, perceptual contexts, ever, some codelets work on their own, per- conceptual contexts, and/or cultural con- forming such tasks as watching for incoming texts. Though contexts are typically uncon- e-mail and instantiating goal structures. An scious, they strongly inﬂuence conscious important type of codelet that works on its processes. own is the attention codelets that serve to Let us look into some details of the IDA bring information to “consciousness.” architecture and its main mechanisms. At IDA senses only strings of characters, the higher level, the IDA architecture is which are not imbued with meaning but modular with module names borrowed from which correspond to primitive sensations, psychology (see Figure 7.5). There are mod- like, for example, the patterns of activ- ules for Perception, Working Memory, Auto- ity on the rods and cones of the retina. biographical Memory, Transient Episodic These strings may come from e-mail mes- Memory, Consciousness, Action Selection, sages, an operating system message, or from Constraint Satisfaction, Language Genera- a database record. tion, and Deliberation. The perception module employs analy- In the lower level of IDA, the proces- sis of surface features for natural-language sors postulated by the global workspace the- understanding. It partially implements per- ory are implemented by “codelets.” Codelets ceptual symbol system theory (Barsalou, are small pieces of code running as indepen- 1999); perceptual symbols serve as a uniform P1: JzG 0521857430c07 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 20:23

164 the cambridge handbook of consciousness

Stimulus Sensory Autobiographical from Internal Memory Memory Environment Internal Senses Cue Local Associations

Perception Codelets Working Long-term Stimulus Senses Consolidation Working Memory from External Memory Environment Percept Perceptual Memory Cue Local Associations Action Selected (Slipnet) and Taken (behavior codelets) Transient Episodic Memory Attention Update Codelets Action Selected and Taken Codelets Update (behavior codelets) Conscious Behavior Codelets Broadcast in priming mode instantiate, bind Procedural Update Action activate Winning Competition for Selection Coalition Consciousness (Behavior Net)

Figure 7.5. IDA’s cognitive cycle.

system of representations throughout the The results of this process, information system. Its underlying mechanism consti- created by the agent for its own use, are tutes a portion of the Copycat architecture written to the workspace (working mem- (Hofstadter & Mitchell, 1994). IDA’s per- ory, not to be confused with Baars’ global ceptual memory takes the form of a semantic workspace). (Almost all of IDA’s modules net with activation passing, called the slipnet either write to the workspace, read from it, (see Figure 7.6). The slipnet embodies the or both.) perceptual contexts and some conceptual IDA employs sparse distributed mem- contexts from the global workspace theory. ory (SDM) as its major associative memory Nodes of the slipnet constitute the agent’s (Anwar & Franklin, 2003; Kanerva, 1988). perceptual symbols. Perceptual codelets rec- SDM is a content-addressable memory. ognize various features of the incoming Being content addressable means that items stimulus; that is, various concepts. Percep- in memory can be retrieved by using part of tual codelets descend on an incoming mes- their contents as a cue, rather than having to sage, looking for words or phrases they rec- know the item’s address in memory. ognize. When such are found, appropriate Reads and writes, to and from associative nodes in the slipnet are activated. This acti- memory, are accomplished through a gate- vation passes around the net until it settles. way within the workspace called the focus. A node (or several) is selected by its high When any item is written to the workspace, activation, and the appropriate template(s) another copy is written to the read registers is ﬁlled by codelets with selected items from of the focus. The contents of these read reg- the message. The information thus created isters of the focus are then used as an address from the incoming message is then written to query associative memory. The results of to the workspace (working memory, to be this query – that is, whatever IDA associates described below), making it available to the with this incoming information – are writ- rest of the system. ten into their own registers in the focus. P1: JzG 0521857430c07 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 20:23

computational models of consciousness: a taxonomy and some examples 165

Norfolk norfolk nor NRFK

San Diego Miami Norfolk Jacksonville

location

Information preference acceptance request

Figure 7.6. A portion of the slipnet in IDA.

This may include some emotion and some tion codelets have the task of bringing infor- action previously taken. Thus associations mation to “consciousness.” Each attention with any incoming information, either from codelet keeps a watchful eye out for some the outside world or from some part of IDA particular situation to occur that might call itself, are immediately available. (Writes to for “conscious” intervention. Upon encoun- associative memory are made later and are tering such a situation, the appropriate described below.) attention codelet will be associated with the In addition to long-term memory, IDA small number of information codelets that includes a transient episodic memory carry the information describing the situ- (Ramamurthy, D’Mello, & Franklin, 2004). ation. This association should lead to the Long-term, content-addressable, associative collection of this small number of codelets, memories are not typically capable of together with the attention codelet that col- retrieving details of the latest of a long lected them, becoming a coalition. Codelets sequence of quite similar events (e.g., where also have activations. The attention codelet I parked in the parking garage this morn- increases its activation in proportion to how ing or what I had for lunch yesterday). The well the current situation ﬁts its particular distinguishing details of such events tend to interest, so that the coalition might compete blur due to interference from similar events. for “consciousness,” if one is formed. In IDA, this problem is solved by the addi- In IDA, the coalition manager is respon- tion of a transient episodic memory imple- sible for forming and tracking coalitions of mented with a sparse distributed memory. codelets. Such coalitions are initiated on the This SDM decays so that past sequences of basis of the mutual associations between similar events no longer interfere with the the member codelets. At any given time, latest such events. one of these coalitions ﬁnds it way to “con- The apparatus for producing “conscious- sciousness,” chosen by the spotlight con- ness” consists of a coalition manager, a spot- troller, which picks the coalition with the light controller, a broadcast manager, and a highest average activation among its mem- collection of attention codelets that recog- ber codelets. Baars’ global workspace the- nize novel or problematic situations. Atten- ory calls for the contents of “consciousness” P1: JzG 0521857430c07 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 20:23

166 the cambridge handbook of consciousness

to be broadcast to each of the codelets in of them, the spotlight controller eventually the system, and in particular, to the behav- selects that coalition, and the contents of the ior codelets. The broadcast manager accom- coalition are broadcast to all the codelets. plishes this task. In response to the broadcast, appropri- IDA depends on the idea of a behavior net ate behavior-priming codelets perform three (Maes, 1989; Negatu & Franklin, 2002) for tasks: an appropriate goal structure is instan- high-level action selection in the service of tiated in the behavior net, the codelets bind built-in drives. It has several distinct drives variables in the behaviors of that struc- operating in parallel, and these drives vary ture, and the codelets send activation to the in urgency as time passes and the environ- currently appropriate behavior of the struc- ment changes. A behavior net is composed ture. Eventually that behavior is chosen to be of behaviors and their various links. A behav- acted upon. At this point, information about ior has preconditions as well as additions and the current emotion and the currently exe- deletions. A behavior also has an activation, a cuting behavior is written to the focus by the number intended to measure the behavior’s behavior codelets associated with the cho- relevance to both the current environment sen behavior. The current contents of the (external and internal) and its ability to help write registers in the focus are then writ- satisfy the various drives it serves. ten to associative memory. The rest of the The activation comes from activation behavior codelets associated with the chosen stored in the behaviors themselves, from the behavior then perform their tasks. Thus, an external environment, from drives, and from action has been selected and carried out by internal states. The environment awards means of collaboration between “conscious- activation to a behavior for each of its true ness” and the behavior net. preconditions. The more relevant it is to This background information on the IDA the current situation, the more activation it architecture and mechanisms should enable receives from the environment. (This source the reader to understand IDA’s cognitive of activation tends to make the system cycle (Baars & Franklin, 2003: Franklin opportunistic.) Each drive awards activation et al., 2005). The cognitive cycle speciﬁes to every behavior that, by being active, will the functional roles of memory, emotions, help satisfy that drive. This source of activa- consciousness, and decision making in cognition tends to make the system goal directed. tion, according to the global workspace the- Certain internal states of the agent can also ory. Below, we sketch the steps of the cogni- send activation to the behavior net. This tive cycle; see Figure 7.5 for an overview. activation, for example, might come from a 1. Perception. Sensory stimuli, external or coalition of codelets responding to a “con- internal, are received and interpreted by scious” broadcast. Finally, activation spreads perception. This stage is unconscious. from behavior to behavior along links. 2 IDA’s behavior net acts in consort with . Percept to Preconscious Buffer. The percept its “consciousness” mechanism to select is stored in preconscious buffers of IDA’s actions (Negatu & Franklin, 2002). Sup- working memory. pose some piece of information is written to 3. Local Associations. Using the incoming the workspace by perception or some other percept and the residual contents of the module. Attention codelets watch both it preconscious buffers as cues, local asso- and the resulting associations. One of these ciations are automatically retrieved from attention codelets may decide that this infor- transient episodic memory and from long- mation should be acted upon. This codelet term autobiographical memory. would then attempt to take the informa- 4. Competition for Consciousness. Attention tion to “consciousness,” perhaps along with codelets, whose job is to bring relevant, any discrepancies it may ﬁnd with the help urgent, or insistent events to conscious- of associations. If the attempt is success- ness, gather information, form coalitions, ful, the coalition manager makes a coalition and actively compete against each other. P1: JzG 0521857430c07 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 20:23

computational models of consciousness: a taxonomy and some examples 167

(The competition may also include atten- 9. Action Taken. The execution of a behav- tion codelets from a recent previous ior (goal context) results in the behav- cycle.) ior codelets performing their specialized 5. Conscious Broadcast. A coalition of code- tasks, which may have external or internal lets, typically an attention codelet and its consequences. The acting codelets also covey of related information codelets car- include an expectation codelet (see Step rying content, gains access to the global 6) whose task is to monitor the action and workspace and has its contents broadcast. to try and bring to consciousness any fail- The contents of perceptual memory are ure in the expected results. updated in light of the current contents of consciousness. Transient episodic memory is updated with the current contents IDA’s elementary cognitive activities of consciousness as events. (The contents occur within a single cognitive cycle. More of transient episodic memory are sepa- complex cognitive functions are imple- rately consolidated into long-term mem- mented over multiple cycles. These include ory.) Procedural memory (recent actions) deliberation, metacognition, and voluntary is also updated. action (Franklin, 2000). 6. Recruitment of Resources. Relevant behav- The IDA model employs a methodology ior codelets respond to the conscious that is different from that which is currently broadcast. These are typically codelets typical of computational cognitive models. whose variables can be bound from infor- Although the model is based on experimen- mation in the conscious broadcast. If tal findings in cognitive psychology and brain the successful attention codelet was an science, there is only qualitative consistency expectation codelet calling attention to an with experiments. Rather, there are a num- unexpected result from a previous action, ber of hypotheses derived from IDA as a the responding codelets may be those that unified theory of cognition. The IDA model can help rectify the unexpected situation. generates hypotheses about human cogni- (Thus consciousness solves the relevancy tion and the role of consciousness through problem in recruiting resources.) its design, the mechanisms of its modules, their interaction, and its performance. 7 . Setting Goal Context Hierarchy. The Every agent must sample and act on its recruited processors use the contents of world through a sense-select-act cycle. The consciousness to instantiate new goal frequent sampling allows for a fine-grained context hierarchies, bind their variables, analysis of common cognitive phenomena, and increase their activation. Emotions such as process dissociation, recognition vs. directly affect motivation and determine recall, and the availability heuristic. At a high which terminal goal contexts receive acti- level of abstraction, the analyses support vation and how much. Other (environ- the commonly held explanations of what mental) conditions determine which of occurs in these situations and why. At a finer- the earlier goal contexts receive addi- grained level, the analyses flesh out common tional activation. explanations, adding details and functional 8. Action Chosen. The behavior net chooses mechanisms. Therein lies the value of these a single behavior (goal context). This analyses. selection is heavily influenced by acti- Unfortunately, currently available tech- vation passed to various behaviors influ- niques for studying some phenomena at a enced by the various emotions. The fine-grained level, such as PET, fMRI, EEG, choice is also affected by the current situ- implanted electrodes, etc., are lacking either ation, external and internal conditions, by in scope, in spatial resolution, or in temporal the relation between the behaviors, and resolution. As a result, some of the hypothe- by the residual activation values of vari- ses from the IDA model, although testable ous behaviors. in principle, seem not to be testable at the P1: JzG 0521857430c07 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 20:23

168 the cambridge handbook of consciousness

response system

knowledge modules

declarative conscious awareness system episodic memory

executive system

procedural system

Figure 7.7. Schacter’s model of consciousness.

present time for lack of technologies with so on. In addition to the two views elab- suitable scope and resolution. orated on earlier, let us look into some There is also the issue of the breadth more details of a few other views. Although of the IDA model, which encompasses per- some of the models that are discussed ception, working memory, declarative mem- below are not strictly speaking computa- ory, attention, decision making, procedural tional (because they may not have been learning, and more. How can such a broad fully computationally implemented), they model produce anything useful? The IDA are nevertheless important because they model suggests that these various aspects point to possible ways of constructing com- of human cognition are highly integrated. putational explanations of consciousness. A more global view can be expected to We can examine Schacter’s (1990) model add additional understanding to that produ- as an example. The model is based on neu- ced by more specific models. This assertion ropsychological findings of the dissociation seems to be borne out by the analyses of vari- of different types of knowledge (especially in ous cognitive phenomena (Baars & Franklin, brain-damaged patients). It includes a num- 2003; Franklin et al., 2005). ber of “knowledge modules” that perform specialized and unconscious processing and may send their outcomes to a “conscious Sketches of Some Other Views awareness system,” which gives rise to conscious awareness (see Figure 7.7). Schacter’s As we have seen, there are many attempts explanation of some neuropsychological dis- to explain the difference in conscious acces- orders (e.g., hemisphere neglect, blindsight, sibility. Various explanations have been aphasia, agnosia, and prosopagnosia) is that advanced in terms of the content of knowl- brain damages result in the disconnection edge (e.g., instances vs. rules), the organi- of some of the modules from the conscious zation of knowledge (e.g., declarative vs. awareness system, which causes their inac- procedural), processing mechanisms (e.g., cessibility to consciousness. However, as has spreading activation vs. rule matching and been pointed out by others, this explana- firing), the representation of knowledge tion cannot account for many findings in (e.g., localist/symbolic vs. distributed), and implicit memory research (e.g., Roediger, P1: JzG 0521857430c07 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 20:23

computational models of consciousness: a taxonomy and some examples 169

We can also examine Damasio’s neu- multimodal convergence roanatomically motivated model (Dama- zone sio et al., 1990). The model hypothesizes the existence of many “sensory convergence zones” that integrate information from individual sensory modalities through forward and backward synaptic connections and the visual auditory somatosensory resulting reverberations of activations, with- convergence convergence convergence out a central location for information storage and comparisons; it also hypothesizes the global “multimodal convergence zone,” which integrates information across modal- 7 8 Figure . . Damasio’s model of consciousness. ities also through reverberation (via recurrent connections; see Figure 7.8). Corre- lated with consistency is global informa- 1990). Revonsuo (1993) advocated a similar tion availability; that is, once “broadcast” view, albeit from a philosophical viewpoint, or “reverberation” is achieved, all the infor- largely on the basis of using Schacter’s (1990) mation about an entity stored in difference data as evidence. Johnson-Laird’s (1983) places of the brain becomes available. This model was somewhat similar to Schacter’s was believed to have explained the accessi- model in its overall structure in that there bility of consciousness.6 In terms of Clarion, was a hierarchy of processors and conscious- different sensory convergence zones may be ness resided in the processes at the top of roughly captured by bottom-level modules, the hierarchy. Shallice (1972) put forward each of which takes care of sensory inputs of a model in which a number of “action sys- one modality (at a properly fine level), and tems” could be activated by “selector input” the role of the global multi-modal conver- and the activated action systems correspond gence zone (similar to the global workspace to consciousness. It is not clear, however, in a way) may be played by the top level of what the computational (mechanistic) dif- Clarion, which has the ultimate responsi- ference between conscious and unconscious bility for integrating information (and also processes is in those models, which did not serves as the “conscious awareness system”). offer a mechanistic explanation. The widely recognized role of reverberation We can compare Schacter’s (1990) model (Damasio, 1994; Taylor, 1994) may be cap- with Clarion. It is similar to Clarion in that tured in Clarion through using recurrent it includes a number of “knowledge mod- connections within modules at the bottom ules” that perform specialized and uncon- level and through multiple top-down and scious processing (analogous to bottom-level bottom-up information flows across the two modules in Clarion) and send their out- levels, which leads to the unity of conscious- comes to a “conscious awareness system” ness that is the synthesis of all the informa- (analogous to the top level in Clarion), tion present (Baars, 1988; Marcel, 1983). which gives rise to conscious awareness. Similarly, Crick and Koch (1990) hypoth- Unlike Clarion’s explanation of the con- esize that synchronous firing at 35–75 Hz in scious/unconscious distinction through the the cerebral cortex is the basis for conscious- difference between localist/symbolic versus ness – with such synchronous firing, pieces distributed representations, however, Schac- of information regarding different aspects ter’s model does not elucidate in computa- of an entity are brought together, and thus tional/mechanistic terms the qualitative dis- consciousness emerges. Although conscious- tinction between conscious and unconscious ness has been experimentally observed to processes, in that the “conscious awareness be somewhat correlated with synchronous system” lacks any apparent qualitative dif- firing at 35–75 Hz, there is no explana- ference from the unconscious systems. tion of why this is the case and there is P1: JzG 0521857430c07 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 20:23

170 the cambridge handbook of consciousness

no computational/mechanistic explanation Much more work can be conducted on of any qualitative difference between 35– various issues of consciousness along this 75 Hz synchronous firing and other firing computational line. Such work may include patterns. further specifications of details of compu- Cotterill (1997) offers a “master-module” tational models. It may also include recon- model of consciousness, which asserts that ciliations of existing computational models consciousness arises from movement or the of consciousness. More importantly, it may, planning of movement. The master-module and should, include the validation of compu- refers to the brain region that is responsi- tational models through empirical and the- ble for motor planning. This model sees the oretical means. The last point in particular conscious system as being profligate with its should be emphasized in future work (see resources: Perforce it must plan and organize the earlier discussions concerning Clarion movements, even though it does not always and IDA). In addition, we may also attempt execute them. The model stresses the vital to account for consciousness computation- role that movement plays and is quite com- ally at multiple levels, from phenomenol- patible with the IDA model. This centrality ogy, via various intermediate levels, all the of movement was illustrated by the obser- way down to physiology, which will likely vation that blind people were able to read lead to a much more complete computa- braille when allowed to move their fingers, tional account and a much better picture of but were unable to do so when the dots were consciousness (Coward & Sun, 2004). moved against their still fingers (Cotterill, 1997). Finally, readers interested in the possibil- Acknowledgments ity of computational models of consciousness actually producing “conscious” artifacts Ron Sun acknowledges support in part from 2003 may consult Holland ( ) and other work Office of Naval Research grant N00014-95- along that line. 1-0440 and Army Research Institute grants DASW01-00-K-0012 and W74V8H-04-K- 0002. Stan Franklin acknowledges support Concluding Remarks from the Office of Naval Research and other U.S. Navy sources under grants N00014- This chapter has examined general frame- 01-1-0917,N00014-98-1-0332,N00014-00-1- works of computational accounts of con- 0769, and DAAH04-96-C-0086. sciousness. Various related issues, such as the utility of computational models, explanations of psychological data, and poten- Notes tial applications of machine consciousness, have been touched on in the process. Based 1 on existing psychological and philosophical . There are also various numerical measures evidence, existing models were compared involved, which are not important for the present discussion. and contrasted to some extent. It appears 2 1991 inevitable at this stage that there is the coex- . Cleeremans and McClelland’s ( ) model of artificial grammar learning can be viewed as istence of various computational accounts of instantiating half of the system (the uncon- consciousness. Each of them seems to cap- scious half), in which implicit learning takes ture some aspect of consciousness, but each place based on gradual weight changes in also has severe limitations. To capture the response to practice on a task and the result- whole picture in a unified computational ing changes in activation of various represen- framework, much more work is needed. In tations when performing the task. this regard, Clarion and IDA provide some 3. Note that the accessibility is defined in terms hope. of the surface syntactic structures of the P1: JzG 0521857430c07 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 20:23

computational models of consciousness: a taxonomy and some examples 171

objects being accessed (at the level of out- Bower, G. (1996). Reactivating a reactivation the- comes or processes), not their semantic mean- ory of implicit memory. Consciousness and Cog- ings. Thus, for example, a LISP expression nition, 5(1/2), 27–72. is directly accessible, even though one may Bowers, K., Regehr, G., Balthazard, C., & Parker, not fully understand its meaning. The inter- K. (1990). Intuition in the context of discovery. nal working of a neural network may be inac- Cognitive Psychology, 22, 72–110. cessible even though one may know what the Chaiken, S., & Trope, Y. (Eds.). (1999). Dual network essentially does (through an interpre- process theories in social psychology. New York: tive process). Note also that objects and pro- Guilford Press. cesses that are directly accessible at a certain 1992 level may not be accessible at a ﬁner level of Clark, A. ( ). The presence of a symbol. Con- 193 205 details. nection Science. 4, – . 1993 4. This activation of features is important in sub- Clark, A., & Karmiloff-Smith, A. ( ). The sequent uses of the information associated cognizer’s innards: A psychological and philo- with the concept and in directing behaviors. sophical perspective on the development of thought. Mind and Language, 8(4), 487– 5. This is a term borrowed from computer oper- 519. ating systems that describes a small piece of 1991 code that waits and watches for a particular Cleeremans, A., & McClelland, J. ( ). Learn- event or condition to occur before it acts. ing the structure of event sequences. Journal of Experimental Psychology: General, 120, 235– 6. However, consciousness does not necessar- 253. ily mean accessibility/availability of all the 1975 information about an entity; for otherwise, Collins, A., & Loftus, J. ( ). Spreading activa- conscious inference, deliberate recollection, tion theory of semantic processing. Psychologi- 407 428 and other related processes would be unnec- cal Review, 82, – . essary. Cosmides, L., & Tooby, J. (1994). Beyond intuition and instinct blindness: Toward an evolutionarily rigorous cognitive science. Cognition, References 50, 41–77. Cotterill, R. (1997). On the mechanism of con- Anderson, J. R. (1983). The architecture of cog- sciousness. Journal of Consciousness Studies, 4, nition. Cambridge, MA: Harvard University 231–247. Press. Coward, L. A., & Sun, R. (2004). Criteria for Anderson, J., & Lebiere, C. (1998). The atomic an effective theory of consciousness and some components of thought. Mahwah, NJ: Erlbaum. preliminary attempts. Consciousness and Cog- 268 301 Anwar, A., & Franklin, S. (2003). Sparse dis- nition, 13, – . tributed memory for “conscious” software Crick, F., & Koch, C. (1990). Toward a neurobio- agents. Cognitive Systems Research, 4, 339–354. logical theory of consciousness. Seminars in the 263 275 Baars, B. (1988). A cognitive theory of con- Neuroscience, 2 , – . sciousness. New York: Cambridge University Damasio. A., et al. (1990). Neural regionalization Press. of knowledge access. Cold Spring Harbor Sym- Baars, B. (2002). The conscious access hypothesis: posium on Quantitative Biology, LV. Origins and recent evidence. Trends in Cognitive Damasio, A. (1994). Descartes’ error. New York: Science, 6, 47–52. Grosset/Putnam. Baars, B., & Franklin, S. (2003). How con- Dennett, D. (1991), Consciousness explained. scious experience and working memory inter- Boston: Little Brown. act. Trends in Cognitive Science, 7, 166–172. Edelman, G. (1987). Neural Darwinism. New Barsalou, L. (1999). Perceptual symbol sys- York: Basic Books. tems. Behavioral and Brain Sciences, 22, 577– Edelman, G. (1989). The remembered present: A 609. biological theory of consciousness. New York: Berry, D., & Broadbent, D. (1988). Interac- Basic Books. tive tasks and the implicit-explicit distinction. Edelman, G., & Tononi, G. (2000). A universe of British Journal of Psychology, 79, 251–272. consciousness. New York: Basic Books. P1: JzG 0521857430c07 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 20:23

172 the cambridge handbook of consciousness

Freeman, W. (1995). Societies of brains. Hillsdale, Marcel, A. (1983). Conscious and uncon- NJ: Erlbaum. scious perception: An approach to the rela- Fodor, J. (1983). The modularity of mind. Cam- tions between phenomenal experience and bridge, MA: MIT Press. perceptual processes. Cognitive Psychology, 15, 238 300 Franklin, S. (2000). Deliberation and voluntary – . action in ‘conscious’ software agents. Neural Maslow, A. (1987). Motivation and personality Network World, 10, 505–521. (3d ed.). New York: Harper and Row. Franklin, S., Baars, B. J., Ramamurthy, U., & Ven- Mathis, D., & Mozer, M., (1996). Conscious tura, M. 2005. The Role of consciousness in and unconscious perception: A computational memory. Brains, Minds and Media, 1, 1–38. theory. Proceedings of the 18th Annual Con- 324 Franklin, S., & Graesser, A. C. (1997). Is it an ference of the Cognitive Science Society, – 328 agent, or just a program?: A taxonomy for . autonomous agents. Intelligent agents III, 21–35. McClelland, J., McNaughton, B., & O’Reilly, R. 1995 Franklin, S., Kelemen, A., & McCauley, L. (1998). ( ). Why there are complementary learn- IDA: A cognitive agent architecture. IEEE Con- ing systems in the hippocampus and neocortex: ference on Systems, Man and Cybernetics. Insights from the successes and failures of connectionist models of learning and memory. Psy- Hadley, R. (1995). The explicit-implicit distinc- chological Review, 102(3), 419–457. tion. Minds and Machines, 5, 219–242. Michalski, R. (1983). A theory and methodology Hirschfeld, L., & Gelman, S. (1994). Mapping of inductive learning. Artificial Intelligence, 20, the Mind: Domain Specificity in Cognition and 111–161. Culture. New York: Cambridge University 1985 Press. Minsky, M. ( ). The society of mind. New York: Simon and Schuster. Hofstadter, D., & Mitchell, M. (1994). The copy- 1991 cat project: A model of mental fluidity and Moscovitch, M., & Umilta, C. ( ). Conscious analogy-making. In K. J. Holyoak, & J. A. Barn- and unconscious aspects of memory. In Per- den (Eds.), Advances in connectionist and neural spectives on cognitive neuroscience. New York: computation theory, Vol. 2 : Logical connections. Oxford University Press. Norwood, NJ: Ablex. Neal, A., & Hesketh, B. (1997). Episodic knowl- Holland, O. (2003). Machine consciousness. edge and implicit learning. Psychonomic Bulletin 1 24 37 Exeter, UK: Imprint Academic. and Review, 4( ), – . 2002 Hunt, E., & Lansman, M. (1986). Unified model Negatu, A., & Franklin, S. ( ). An action selec- of attention and problem solving. Psychological tion mechanism for ‘conscious’ software agen- 363 386 Review, 93(4), 446–461. ts. Cognitive Science Quarterly, 2 , – . 1993 Jackendoff, R. (1987). Consciousness and the Nelson, T. (Ed.) ( ). Metacognition: Core computational mind. Cambridge, MA: MIT Readings. Boston, MA: Allyn and Bacon. Press. O’Brien. G., & Opie, J. (1998). A connectionist Johnson-Laird, P. (1983). A computational anal- theory of phenomenal experience. Behavioral 127 148 ysis of consciousness. Cognition and Brain The- and Brain Sciences, 22, – . ory, 6, 499–508. Penrose, R. (1994). Shadows of the mind. Oxford: Kanerva, P. (1988). Sparse distributed memory. Oxford University Press. Cambridge MA: MIT Press. Quillian, M. R. (1968). Semantic memory. In M. Karmiloff-Smith, A. (1986). From meta- Minsky (Ed.), Semantic information processing 227 270 processes to conscious access: Evidence from (pp. – ). Cambridge, MA: MIT Press. children’s metalinguistic and repair data. Ramamurthy, U., D’Mello, S., & Franklin, S. Cognition, 23, 95–147. (2004). Modified sparse distributed memory as LeDoux, J. (1992). Brain mechanisms of emo- transient episodic memory for cognitive soft- tion and emotional learning. Current Opinion ware agents. In Proceedings of the International in Neurobiology, 2 (2), 191–197. Conference on Systems, Man and Cybernetics. Logan, G. (1988). Toward a theory of automati- Piscataway, NJ: IEEE. zation. Psychological Review, 95(4), 492–527. Reber, A. (1989). Implicit learning and tacit Maes, P. (1989). How to do the right thing. Con- knowledge. Journal of Experimental Psychology: 3 219 235 nection Science, 1, 291–323. General, 118( ), – . P1: JzG 0521857430c07 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 20:23

computational models of consciousness: a taxonomy and some examples 173

Revonsuo, A. (1993). Cognitive models of con- Quarterly Journal of Experimental Psychology, sciousness. In M. Kamppinen (Ed.), Conscious- 41A(3), 553–577. ness, cognitive schemata and relativism (pp. 27– Sun, R. (1994). Integrating rules and connectionism 130). Dordrecht, Netherlands: Kluwer. for robust commonsense reasoning. New York: Roediger, H. (1990). Implicit memory: Retention John Wiley and Sons. without remembering. American Psychologist, Sun, R. (1995). Robust reasoning: Integrating 45(9), 1043–1056. rule-based and similarity-based reasoning. Arti- Rosenbloom, P., Laird, J., & Newell, A. (1993). ficial Intelligence, 75(2), 241–296. The SOAR papers: Research on integrated intel- Sun, R. (1997). Learning, action, and cons- ligence. Cambridge, MA: MIT Press. ciousness: A hybrid approach towards mod- Rosenthal, D. (Ed.). (1991). The nature of mind. eling consciousness. Neural Networks, 10(7), Oxford: Oxford University Press. 1317–1331. Rumelhart, D., McClelland, J., & the PDP Sun, R. (1999a). Accounting for the computa- Research Group, (1986). Parallel distributed tional basis of consciousness: A connectionist processing: Explorations in the microstructures of approach. Consciousness and Cognition, 8, 529– cognition. Cambridge, MA: MIT Press. 565. Schacter, D. (1990). Toward a cognitive neu- Sun, R. (1999b). Computational models of ropsychology of awareness: Implicit knowledge consciousness: An evaluation. Journal of Intel- and anosagnosia. Journal of Clinical and Exper- ligent Systems [Special Issue on Consciousness], imental Neuropsychology, 12(1), 155–178. 9(5–6), 507–562. Schneider, W., & Oliver, W. (1991). An Sun, R. (2002). Duality of the mind. Mahwah, NJ: instructable connectionist/control architec- Erlbaum. ture. In K. VanLehn (Ed.), Architectures for Sun, R. (2003). A tutorial on CLARION. intelligence. Hillsdale, NJ: Erlbaum. Retrieved from http://www.cogsci.rpi.edu/ Searle, J. (1980). Minds, brains, and programs. ∼rsun/sun.tutorial.pdf. 417 457 Brain and Behavioral Sciences, 3, – . Sun, R., & Bookman, L. (Eds.). (1994). Computa- Seger, C. (1994). Implicit learning. Psychological tional architectures integrating neural and sym- Bulletin, 115(2), 163–196. bolic processes. Norwell, MA: Kluwer. Servan-Schreiber, E., & Anderson, J. (1987). Sun, R., Merrill, E., & Peterson, T. (2001). From Learning artificial grammars with competitive implicit skills to explicit knowledge: A bottom- chunking. Journal of Experimental Psychology: up model of skill learning. Cognitive Science, Learning, Memory, and Cognition, 16, 592– 25(2), 203–244. 608 . Sun, R., & Naveh, I. (2004, June). Simulating Shallice, T. (1972). Dual functions of conscious- organizational decision making with a cogni- ness. Psychological Review, 79(5), 383–393. tive architecture Clarion. Journal of Artificial Sloman, A., & Chrisley, R. (2003). Virtual Society and Social Simulation, 7(3). Retrieved machines and consciousness. Journal of Con- from http://jasss.soc.surrey.ac.uk/7/3/5.html. sciousness Studies, 10, 133–172. Sun, R., & Peterson, T. (1998). Autonomous Smith, E., & Medin, D. (1981). Categories and learning of sequential tasks: Experiments and concepts. Cambridge, MA: Harvard University analyses. IEEE Transactions on Neural Net- Press. works, 9(6), 1217–1234. Smith, J. D., Shields, W. E., & Washburn, D. A. Sun, R., & Peterson, T. (1999). Multi-agent rein- (2003). The comparative psychology of uncer- forcement learning: Weighting and partition- tainty monitoring and metacognition. Behav- ing. Neural Networks, 12(4–5). 127–153. 317 339 ioral and Brain Sciences. 26, – . Sun, R., Peterson, T., & Merrill, E. (1996). Smolensky, P. (1988). On the proper treatment of Bottom-up skill learning in reactive sequential connectionism. Behavioral and Brain Sciences, decision tasks. Proceedings of 18th Cognitive Sci- 11(1), 1–74. ence Society Conference. Hillsdale, NJ: Lawrence Stanley, W., Mathews, R., Buss, R., & Kotler- Erlbaum Associates. Cope, S. (1989). Insight without awareness: On Sun, R., Merrill, E.. & Peterson, T. (1998). A the interaction of verbalization, instruction and bottom-up model of skill learning. Proceedings practice in a simulated process control task. of the 20th Cognitive Science Society Conference P1: JzG 0521857430c07 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 20:23

174 the cambridge handbook of consciousness

(pp. 1037–1042). Mahwah, NJ: Lawrence Toates, F. (1986). Motivational systems. Erlbaum Associates Cambridge: Cambridge University Press. Sun , R., Slusarz, P., & Terry, C. (2005). The inter- Tversky, A. (1977). Features of similarity. Psycho- action of the explicit and the implicit in skill logical Review, 84(4), 327–352. learning: A dual-process approach. Psychologi- Watkins, C. (1989). Learning with delayed cal Review, 112, 159–192. rewards. PhD Thesis, Cambridge University, Taylor, J. (1994). Goal, drives and consciousness. Cambridge, UK. Neural Networks, 7 (6/7), 1181–1190. Willingham, D., Nissen, M., & Bullemer, P. Taylor, J. (1997). The relational mind. In A. (1989). On the development of procedural Browne (Ed.), Neural network perspectives on knowledge. Journal of Experimental Psychology: cognition and adaptive robotics. Bristol, UK: Learning, Memory, and Cognition, 15, 1047– Institute of Physics. 1060. P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

C. Cognitive Psychology

175 P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

176 P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

CHAPTER 8 Cognitive Theories of Consciousness

Katharine McGovern and Bernard J. Baars

Abstract ing. A specific conscious content, like the sight of a coffee cup, is crucially dependent Current cognitive theories of conscious- on local regions of visual cortex. But, by ness focus on a few common themes, such itself, local cortical activity is not conscious. as the limited capacity of conscious con- Rather, the conscious experience of a cof- tents under input competition; the wide fee cup requires both local and widespread access enabled by conscious events to sen- cortical activity. sation, memory, problem-solving capacities, and action control; the relation between conscious contents and working memory; Introduction and the differences between implicit and explicit cognition in learning, retrieval, and When consciousness became a scientifically other cognitive functions. The evidentiary respectable topic again in the 1980s, it was base is large. A unifying principle in the tackled in a number of different scholarly midst of these diverse empirical findings is to disciplines – psychology, philosophy, neuro- treat consciousness as an experimental vari- science, linguistics, medicine, and others. By able and, then, to look for general capacities the late 1990s, considerable interdisciplinary that distinguish conscious and unconscious cooperation evolved in consciousness stud- mental functioning. In this chapter, we dis- ies, spurred by the biennial Tucson Con- cuss three classes of theories: information- ferences and the birth of two new schol- processing theories that build on modu- arly journals, Consciousness and Cognition lar elements, network theories that focus and the Journal of Consciousness Studies. on the distributed access of conscious pro- The domain of consciousness studies origi- cessing, and globalist theories that combine nated in separate disciplines, but has since aspects of these two. An emerging con- become cross-disciplinary. Thus, a number sensus suggests that conscious cognition is of early theories of consciousness can jus- a global aspect of human brain function- tifiably be called purely cognitive theories 177 P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

178 the cambridge handbook of consciousness

of consciousness, whereas most recent the- able has led to scientific advances before. ories are neurocognitive hybrids – depend- In the late 1600s, contemporaries of Isaac ing on evidence from the brain as well as Newton were frustrated in their attempts behavior. In this chapter, we have, for the to understand gravity. One key to Newton’s most part, restricted discussion to cognitive great achievement was to imagine the pres- or functional models of consciousness with ence and the absence of gravity, thus per- less reference to the burgeoning neuroscien- mitting gravity to be treated as a variable. tific evidence that increasingly supports the In the same way, a breakthrough in the globalist position that we develop here. scientific study of consciousness occurred when psychologists began to understand that consciousness can be treated as a variable. That is, behavioral outcomes can be Operationally Defining Consciousness observed when conscious cognitions are present and when they are absent. The pro- Cognitive Methods That Treat cess of generalizing across these observa- Consciousness as a Variable tions has been called contrastive analysis There is a curious asymmetry between the (explained below). assessment of conscious and unconscious Beginning in the 1980s, a number of processes. Obtaining verifiable experiential experimental methods gained currency as reports works very nicely for specifying con- means of studying comparable conscious scious representations, but unconscious ones and non-conscious processes. In much of are much more slippery. In many cases of cognitive science and neuroscience today, apparently unconscious processes, such as the existence of unconscious cognitive pro- all the things the reader is not paying atten- cesses, often comparable to conscious ones, tion to at this moment, it could be that is taken for granted. Table 8.1 highlights the “unconscious” representations may be methods that have produced behavioral data momentarily conscious, but so quickly or relevant to the study of consciousness. vaguely that we cannot recall them even a fraction of a second later. Or suppose Working Definitions of “Conscious” people cannot report a word shown for and “Unconscious” a few milliseconds: Does this mean that they are truly unconscious of it? Such ques- In the history of science, formal definitions tions continue to lead to controversy today. for concepts like “heat” and “gene” tend to William James understood this problem very come quite late, often centuries after ade- well and suggested, in fact, that there were quate operational definitions are developed. no unconscious psychological processes at The same point may apply to conscious cog- all (1890,p.162ff.). This has been called nition. Although there is ongoing debate the “zero point” problem (Baars, 1988). It about what consciousness “really” is, there should be emphasized, however, that prob- has long been a scientific consensus on its lems with defining a zero point do not pre- observable index of verbal report. This index vent scientists from studying phenomena as can be generalized to any other kind of variables. Even today, the precise nature of voluntary response, such as pressing a but- zero temperature points, such as the freezing ton or even voluntary eye movements in point of water, continues to lead to debate. “locked-in” neurological patients. Experien- But physicists have done extremely produc- tial reports can be analyzed with sophis- tive work on thermodynamics for centuries. ticated methods, such as process dissocia- Zero points are not the sole criterion for use- tion and signal detection. Thus, empirically, ful empirical variables. it is not difficult to assess conscious events The discovery that something we take for in humans with intact brains, given good granted as a constant can be treated as a vari- experimental conditions. We propose the P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

cognitive theories of consciousness 179

Table 8.1. Empirical methods used in the study of conscious and unconscious processes Class of Methods Experimental Paradigm Outcome Divided attention Dichotic listening Two dense streams of speech are offered to the two ears, and only one stream at a time can receive conscious word-level processing. Evidence suggests that the unconscious stream continues to receive some processing. Selective (“double When two overlaid movies are viewed, only exposure”) viewing one is perceived consciously. Inattentional blindness Aspects of visual scenes to which attention is not directed are not consciously perceived; attended aspects of the same scenes are perceived. Binocular rivalry/dichoptic Presenting separate visual scenes to each viewing (including flash eye; only one scene reaches suppression) consciousness, but the unconscious scene receives low-level processing. Dual task paradigms Driving and talking on a cell To the extent that tasks require conscious phone initiation and direction, they compete Rehearsing words and doing and degrade the performance of each word verification other; once automatized, multiple tasks interfere less. Priming Supraliminal and subliminal When a “prime” stimulus is presented prior priming to a “target” stimulus, response to the Priming of one interpretation “target” is influenced by the currently of ambiguous words or unconscious nature and meaning of the pictures “prime.” Supraliminal priming generally results in a more robust effect. Visual backward When supra-threshold visual stimuli are masking followed immediately by visual masking stimuli (visual noise), the original stimuli are not consciously perceived, though they are locally registered in early visual cortex. Implicit learning Miniature grammar learning Consciously perceived stimuli give rise to knowledge structures that are not available to consciousness. Process dissociation Participants are told to exclude certain and ironic effects memorized items from memory reports; if those items nevertheless appear, they are assumed to be products of non-conscious processing. Fixedness, Problem-solving tasks, Set effects in problem solving can exclude decontextualization, functional fixedness tasks otherwise obvious conclusions from and being blind to (Duncker), chess playing, consciousness. “Breaking set” can lead to the obvious (related garden path sentences, recovery of those conclusions in to availability) highly automatized actions consciousness. under novel conditions P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

180 the cambridge handbook of consciousness

Table 8.2 . Contrastive analysis in perception and imagery Conscious Events Comparable Unconscious Events 1. Perceived stimuli 1. Processing of stimuli lacking in intensity or duration, or centrally masked stimuli 2. Preperceptual processing of stimuli 3. Habituated or automatic stimulus processing 4. Unaccessed versions of ambiguous stimuli/words 5. Contexts of interpretation for percepts and concepts 6. Unattended streams of perceptual input (all modalities) 7. Implicit expectations about stimuli 8. Parafoveal guidance of eye movements in reading 9. Stimulus processing under general anesthesia 10. Images in all sense modalities 10. Unretrieved images in memory 11. a. Newly generated visual images 11. Automatized visual images b. Automatic images that encounter some difﬁculty 12. Inner speech: words currently rehearsed in 12. Inner speech, not currently rehearsed in working working memory memory 13. Fleetingly conscious phrases and belief 13. Automatized inner speech; the “jingle channel” statements 14. Visual search based on conjoined features 14. Visual search based on single features 15. Retrieval by recall 15. Retrieval by recognition 16. Explicit knowledge 16. Implicit knowledge

following as de facto operational defini- (d) even under optimal reporting conditions of conscious and unconscious that are tions. already in very wide experimental use in perception, psychophysics, memory, imagery, and the like. The Method of Contrastive Analysis We can say that mental processes are con- Using the logic of experimental research, scious if they consciousness can be treated as a controlled (a) are claimed by people to be conscious; variable; then, measures of cognitive func- and tioning and neural activity can be com- (b) can be reported and acted upon, pared under two levels of the independent variable – consciousness-present and (c) with verifiable accuracy, consciousness-absent. If there is no clearly (d) under optimal reporting conditions (e.g., unconscious comparison condition, a low- with minimum delay between the event level conscious condition may be used, and the report, freedom from distrac- as in drowsiness or stimuli in background tion, and the like). noise. The point, of course, is to have at Conversely, mental events can be defined as least two quantitatively different levels for unconscious for practical purposes if comparison. (a) their presence can be verified (through data from contrastive analysis facilitation of other observable tasks, for Examples of conscious versus non-conscious example); although contrasts from studies of perception, (b) they are not claimed to be conscious; imagery, memory, and attention appear in (c) and they cannot be voluntarily reported, Table 8.2. In the left column, conscious operated upon, or avoided, mental events are listed; on the right are P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

cognitive theories of consciousness 181

Table 8.3. Capability contrasts between comparable conscious and non-conscious processes Conscious Processes Unconscious Processes 1. Are computationally inefﬁcient with 1. Are very efﬁcient in routine tasks with • Many errors • Few errors • Relatively low speed • High speed • Mutual interference between conscious • Little mutual interference. processes. 2. Have a great range of contents 2. Taken individually, unconscious processes have • Great ability to relate different conscious a limited range of contents contents to each other • Each routine process is relatively isolated and • Great ability to relate conscious events to autonomous unconscious contexts • Each routine process is relatively • Flexible context-free, operates in a range of contexts • Fixed pattern 3. 3. The set of routine, unconscious processes, • have high internal consistency at any single taken together, is: moment • diverse, • have seriality over time • can operate concurrently • have limited processing capacity • have great processing capacity 4. The clearest conscious contents are perceptual 4. Unconscious processes are involved in all or quasi-perceptual (e.g., imagery, inner speech, mental tasks, not limited to perception and and internally generated bodily feelings) imagery, but including memory, knowledge representation and access, skill learning, problem-solving, action control, etc. 5. Are associated with voluntary actions 5. Are associated with non-voluntary actions

corresponding non-conscious processes. would need to account for these observed Theoretically, we are interested in finding differences in functioning. Thus, we have out what is common in conscious processing a way of judging the explanatory adequacy across all these cases. of proposals concerning the nature and functioning of consciousness. We can keep these capability contrasts in mind as we Capability Contrasts review contemporary cognitive models of The difference in mental and neural func- consciousness. tioning between consciousness-present and Given the tight constraints that appear consciousness-absent processing – taken repeatedly in studies of conscious process- across many experimental contexts – reveals ing – that is, limited capacity, seriality, stable characteristics attributable to con- and internal consistency requirements – we sciousness. Conscious processes are phenom- might ask, Why? Would it not be adaptive to enally serial, internally consistent, unitary at do several conscious things at the same time? any moment, and limited in capacity. Non- Certainly human ancestors might have ben- conscious mental processes are functionally efited from being able to gather food, be concurrent, often highly differentiated from alert for predators, and keep an eye on their each other, and relatively unlimited in capac- offspring simultaneously; modern humans ity, when taken together. Table 8.3 sum- could benefit from being able to drive their maries these general conclusions. cars, talk on cell phones, and put on lip- These empirical contrasts in the capabil- stick without mutual interference. Yet these ities of conscious and unconscious mental tasks compete when they require conscious- processes can become the criteria against ness, so that only one can be done well at which models of consciousness can be evalu- any given moment. The question then is, ated. Any adequate theory of consciousness Why are conscious functions so limited in P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

182 the cambridge handbook of consciousness

a neuropsychological architecture that is so of the system that are capable of acting large and complex? to recognize and correct it. Consciousness functions as the central distributor of information, which is used by subparts of the Functions of Consciousness in the cognitive system or architecture. Architecture of Cognition Consciousness Creates Access A Note about Architectures A strong case can be made that we can cre- The metaphor of “cognitive architectures” ate access to any part of the brain by way of dates to the 1970s when cognitive psychol- conscious input. For example, to gain volun- ogists created information-processing mod- tary control over alpha waves in the occipi- els of mental processes. In many of these tal cortex we merely sound a tone or turn models, different mental functions, such as on a light when alpha is detected in the memory, language, attention, and sensory EEG, and shortly the subject will be able processes, were represented as modules, or to increase the power of alpha at will. To sets of modules, within a larger information- control a single spinal motor unit we merely processing system. The functional layout pick up its electrical activity and play it back and the interactions of the parts of the sys- over headphones; in a half-hour, subjects tem came to be called the cognitive archi- have been able to play drum rolls on single tecture. We have adopted this terminology motor units. Biofeedback control over sin- here to capture the idea that consciousness gle neurons and whole populations of neu- operates within a larger neuropsychological rons anywhere in the brain is well estab- system that has many constituents interact- lished (Basmajian, 1979). Consciousness of ing in complex ways. the feedback signal seems to be a necessary condition to establish control, though the motor neural activities themselves remain Consciousness Serves Many Functions entirely unconscious. It is as if mere con- William James believed that “[t]he sciousness of results creates access to uncon- study . . . of the distribution of consciousness scious neuronal systems that are normally shows it to be exactly such as we might quite autonomous. expect in an organ added for the sake Psychological evidence leads to similar of steering a nervous system grown too conclusions. The recognition vocabulary of complex to regulate itself (1890,p.141).” educated English speakers contains about More recently, Baars (1988) identified eight 100,000 words. Although we do not use all psychological functions of consciousness, of them in everyday speech, we can under- which are defined in Table 8.4. stand each word as soon as it is presented in Note that each proposed function of con- a sentence that makes sense. Yet each indi- sciousness is served through an interplay of vidual word is already quite complex. The conscious and unconscious processes. It has Oxford English Dictionary devotes 75,000 been argued that consciousness fulfills all words to the many different meanings of eight functions by providing access or pri- the word “set.” Yet all we do as humans ority entrance into various subparts of the to access these complex unconscious bod- cognitive system (Baars, 2002). For example, ies of knowledge is to become conscious of the error-detection function can be accom- a target word. It seems that understanding plished only when information about an language demands the gateway of conscious- impending or actual error, which cannot be ness. This is another case of the general prin- handled by “canned” automatisms, can gain ciple that consciousness of stimuli creates access to consciousness. Subsequently, edit- widespread access to unconscious sources of ing occurs when this conscious information knowledge, such as the mental lexicon, is “broadcast” or distributed to other parts meaning, and grammar. P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

cognitive theories of consciousness 183

Table 8.4. Explaining the psychological functions of consciousness Function of Consciousness Function Explained 1. Definition and context-setting By relating input to its contextual conditions, consciousness defines a stimulus and removes ambiguities in its perception and understanding. 2. Adaptation and learning The more novelty and unpredictability to which the psychological system must adapt, the greater the conscious involvement required for successful problem solving and learning. 3. Prioritizing and access control Attentional mechanisms exercise selective control over what will become conscious by relating input to unconscious goal contexts. By consciously relating an event or circumstance to higher-level goals, we can raise its access priority, making it conscious more often and therefore increasing the chances of successful adaptation to it. 4. Recruitment and control of Conscious goals can recruit subgoals and behavior systems to thought and action organize and carry out flexible, voluntary action. 5. Decision-making and Consciousness creates access to multiple knowledge sources executive function within the psychological system. When automatic systems cannot resolve some choice point in the flow of action, making it conscious helps recruit knowledge sources that are able to help make the decision; in case of indecision, making the goal conscious allows widespread recruitment of conscious and unconscious sources acting for and against the goal. 6. Error detection and editing Conscious goals and plans are monitored by unconscious rule systems that will act to interrupt execution if errors are detected. Though we often become aware of making an error in a general way, the detailed description of what makes an error an error is almost always unconscious. 7. Reflection and self-monitoring Through conscious inner speech and imagery we can reflect upon and to some extent control and plan our conscious and unconscious functioning. 8. Optimizing the tradeoff Automatized, “canned” responses are highly adaptive in between organization and predictable circumstances. However, in unpredictable flexibility environments, the capacity of consciousness to recruit and reconfigure specialized knowledge sources is indispensable in allowing flexible responding.

Or consider autobiographical memory. conscious experience of each picture in its The size of long-term episodic memory is entirety. Here the brain does a marvelous unknown, but we do know that simply job of memory search, with little effort. It by paying attention to as many as 10,000 seems that humans create memories of the distinct pictures over several days with- stream of input merely by paying attention, out attempting to memorize them, we can but because we are always paying attention spontaneously recognize more than 90%a to something, in every waking moment, this week later (Standing, 1973). Remarkable suggests that autobiographical memory may results like this are common when we use be very large indeed. Once again we have a recognition probes, merely asking people vast unconscious domain, and we gain access to choose between known and new pic- to it using consciousness. Mere conscious- tures. Recognition probes apparently work ness of some event helps store a recogniz- so well because they re-present the original able memory of it, and when we experience P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

184 the cambridge handbook of consciousness

it again, we can distinguish it accurately from patterns of system-wide activity: Pribram’s millions of other experiences. Holonomic Theory, Tononi and Edelman’s The ability to access unconscious pro- Dynamic Core Hypothesis, and Walter cesses via consciousness applies also to the Freeman’s Dynamical Systems Approach. vast number of unconscious automatisms The third group includes globalist mod- that can be triggered by conscious events, els that combine aspects of information- including eye movements evoked by con- processing theories and network theories: scious visual motion, the spontaneous inner Baars’ Global Workspace Theory, Franklin’s speech that often accompanies reading, the IDA as an implementation of GW theory, hundreds of muscle groups that control the and Dehaene’s Global Neuronal Network vocal tract, and those that coordinate and Theory. Theories have been selected that control other skeletal muscles. None of these represent the recent history of cognitive automatic neuronal mechanisms are con- modeling of consciousness from the 1970s scious in any detail under normal circum- forward and that account in some way for stances. Yet they are triggered by conscious the evidence described above concerning the events. This triggering function is hampered capability contrasts of conscious and uncon- when the conscious input is degraded by dis- scious processes. traction, fatigue, somnolence, sedation, or low signal fidelity. Information-Processing Theories That Consciousness seems to be needed to Emphasize Modular Processes: access at least four great bodies of uncon- Consciousness Depends on a Kind of scious knowledge: the lexicon of natural lan- Processing guage, autobiographical memory, the automatic routines that control actions, and even Theories in this group emphasize the the detailed firing of neurons and neuronal information-processing and action control populations, as shown in biofeedback train- aspects of the cognitive architecture. They ing. Consciousness seems to create access to tend to explain consciousness in terms of vast unconscious domains of expert knowl- “flow of control” or flow of information edge and skill. among specialist modules. johnson-laird’s operating system model of consciousness Survey of Cognitive Theories Johnson-Laird’s (1988) operating system of Consciousness model of consciousness emphasizes its role in controlling mental events, such as direct- Overview ing attention, planning and triggering action In the survey that follows, cognitive the- and thought, and engaging in purposeful ories of consciousness are organized into self-reflection. Johnson-Laird proposes that three broad categories based on the architec- the cognitive architecture performs paral- tural characteristics of the models. The first lel processing in a system dominated by group consists of examples of information- a control hierarchy. His system involves a processing theories that emphasize modular collection of largely independent processors processes: Johnson-Laird’s Operating System (finite state automata) that cannot modify Model of Consciousness, Schacter’s Model each other but that can receive messages of Dissociable Interactions and Conscious from each other; each initiates computation Experience (DICE), Shallice’s Supervisory when it receives appropriate input from any System, Baddeley’s Early and Later Mod- source. Each passes messages up through the els of Working Memory, and Schneider hierarchy to the operating system that sets and Pimm-Smith’s Message-Aware Control goals for the subsystems. The operating sys- Mechanism. The second group includes net- tem does not have access to the detailed work theories that explain consciousness as operations of the subsystems – it receives P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

cognitive theories of consciousness 185

Response System

Lexical Conceptual Facial Spatial Self

Consciousness operates here Conscious Awareness System Episodic Memory

Executive System

Procedural-Habit System

Figure 8.1. Schacter’s Dissociable Interactions and Conscious Experience (DICE) Model. (Redrawn from Schacter, 1990). Phenomenal awareness depends on intact connections between the conscious awareness system and the individual knowledge modules or episodic memory. The conscious awareness system is the gateway to the executive system, which initiates voluntary action.

only their output. Likewise, the operating cesses that mediate conscious identification system does not need to specify the details and recognition – that is, phenomenal aware- of the actions it transmits to the processors – ness in different domains – should be sharply they take in the overall goal, abstractly spec- distinguished from modular systems that ified, and elaborate it in terms of their own operate on linguistic, perceptual, and other capabilities. kinds of information” (pp. 160–161, 1990). In this model, conscious contents reside Like Johnson-Laird’s model, Schacter’s in the operating system or its working mem- DICE model assumes independent memory ory. Johnson-Laird believes his model can modules and a lack of conscious access to account for aspects of action control, self- details of skilled/procedural knowledge. It reflection, intentional decision making, and is primarily designed to account for mem- other metacognitive abilities. ory dissociations in normally functioning and damaged brains. There are two main obser- schacter’s model of dissociable vations of interest. First, with the excep- interactions and conscious tion of coma and stupor patients, failures experience (dice) of awareness in neuropsychological cases Accumulating evidence regarding the neu- are usually restricted to the domain of the ropsychological disconnections of process- impairment; these patients do not have diffi- ing from consciousness, particularly implicit culty generally in gaining conscious access to memory and anosagnosia, led Schacter other knowledge sources. Amnesic patients, (1990) to propose his Dissociable Interac- for example, do not necessarily have trou- tions and Conscious Experience (DICE) ble reading words, whereas alexic indi- model (see Figure 8.1): “The basic idea moti- viduals do not necessarily have memory vating the DICE model . . . is that the pro- problems. P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

186 the cambridge handbook of consciousness

Second, implicit (non-conscious) mem- control thought and action. Subsequently, ory of unavailable knowledge has been Shallice (1988; Norman & Shallice, 1980) demonstrated in many conditions. For exam- modified and refined the theory to accom- ple, name recognition is facilitated in modate a broader range of conscious func- prosopagnosic patients when the name of tions (depicted in Figure 8.2). the to-be-identified face is accompanied by Shallice describes an information- a matching face – even though the patient processing system as having five charac- does not consciously recognize the face. teristics. First, it consists of a very large Numerous examples of implicit knowledge set of specialized processors, with several in neuropsychological patients who do not qualifications on their “modularity”: have deliberate, conscious access to the r information are known (see Milner & Rugg, There is considerable variety in the way 1992 ). These findings suggest an architec- the subsystems can interact. ture in which various sources of knowl- r The overall functional architecture is seen edge function somewhat separately, because as partly innate and partly acquired, as they can be selectively lost; these knowledge with the ability to read. sources are not accessible to consciousness, r even though they continue to shape volun- The “modules” in the system include not tary action. only input processors but also specialized In offering DICE, Schacter has given information stores, information manage- additional support to the idea of a con- ment specialists, and other processing scious capacity in a system of separable modules. knowledge sources, specifically to explain spared implicit knowledge in patients with Second, a large set of action and brain damage. DICE does not aim to explain thought schemata can “run” on the modules. the limited capacity of consciousness or the These schemata are conceptualized as well- problem of selecting among potential inputs. learned, highly specific programs for rou- In agreement with Shallice (see below) the tine activities, such as eating with a spoon, DICE model suggests that the primary role driving to work, etc. Competition and of consciousness is to mediate voluntary interference between currently activated action under the control of an executive. schemata are resolved by another special- However, the details of these abilities are not ist system, CONTENTION SCHEDUL- spelled out, and other plausible functions are ING, which selects among the schemata not addressed. based on activation and lateral inhibition. Contention scheduling acts during routine operations. shallice’s supervisory system Third, a SUPERVISORY SYSTEM func- Shallice shares an interest in the relation tions to modulate the operation of con- of volition and consciousness with James tention scheduling. It has access to repre- (1890), Baars (1988), and Mandler (1975). sentations of operations, of the individual’s In 1972, Shallice argued that psychologists goals, and of the environment. It comes into needed to have a rationale for using the play when operation of routinely selected data of introspection. To do so, he said, schemata does not meet the system’s goals; the nature of consciousness would need to that is, when a novel or unpredicted sit- be considered. He thought at that time uation is encountered or when an error that the selector of input to the dominant has occurred. action system had properties that corre- Fourth, a LANGUAGE SYSTEM is sponded to those of consciousness. Shallice’s involved that can function either to acti- early theory (1978) focused on conscious vate schemata or to represent the opera- selection for a dominant action system, the tions of the supervisory system or specialist set of current goals that work together to systems. P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

cognitive theories of consciousness 187

Supervisory System Consciousness depends on concurrent and coherent operation of several control systems

Episodic Language Contention Memory System Scheduling

Trigger Database

Special Purpose Processing System

Figure 8.2 . Shallice’s Supervisory System Model of Conscious Processing. Solid arrows represent obligatory communications; dashed arrows represent optional communications. (Drawn from Shallice, 1988.)

Fifth, more recently an EPISODIC tions between willed and ideomotor action” MEMORY component containing event- (p. 319). Shallice identifies consciousness specific traces has been added to the set of with the control of coherent action subsys- control processes. tems and the emphasis on the flow of infor- Thus, the supervisory system, con- mation among the subsystems. tention scheduling, the language system, and episodic memory all serve higher-level or baddeley’s early and later models control functions in the system. As a first of working memory: 1974 to 2000 approximation, one of these controllers or Working memory is a functional account several together might be taken as the “con- of the workings of temporary memory (as scious part” of the system. However, as Shal- distinct from long-term memory). Baddeley lice points out, consciousness cannot reside and Hitch (1974) first proposed their multi- in any of these control systems taken indi- component model of working memory vidually. No single system is either necessary (WM) as an advance over single-store mod- or sufficient to account for conscious events. els, such as Short-Term Memory (STM; Consciousness remains even when one of Atkinson & Shiffrin, 1968). The original WM these control systems is damaged or dis- model was simple, composed of a central abled, and the individual control systems can executive with two subsystems, the phono- all operate autonomously and unconsciously. logical loop and the visuospatial sketch- Instead, Shallice suggests that consciousness pad. WM was designed to account for may arise on those occasions where there is short duration, modality-specific, capacity- concurrent and coherent operation of sev- limited processing of mnemonic informa- eral control systems on representations of a tion. It combined the storage capacity of the single activity. In this event, the contents of older STM model with an executive pro- consciousness would correspond to the flow cess that could “juggle” information between of information between the control systems two slave systems and to and from long-term and the flow of information and control from memory. The evolving model of WM has the control systems to the rest of the cogni- been successful in accounting for behavioral tive system. and neurological findings in normal partic- Shallice’s (1988) model aims primarily ipants and in neuropsychological patients. to “reflect the phenomenological distinc- From the beginning, working memory, P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

188 the cambridge handbook of consciousness

Central Executive Retrieval from the episodic buffer involves conscious processing

Visuospatial Episodic Buffer Phonological Sketchpad Loop

Visual Semantics Episodic LTM Language ...... Long Term Memory......

Figure 8.3. Baddeley’s Model of Working Memory. This model incorporates the episodic buffer. (Adapted from Baddeley, 2000.)

particularly transactions between the central with the visuospatial sketchpad and phono- executive and the subsystems, has been asso- logical loop), in which the central executive ciated with conscious and effortful informa- controls and switches attention while the tion processing. However, these were rarely episodic buffer creates and makes available stated in terms of the question of conscious- multimodal information. ness as such. Baddeley’s episodic buffer resembles Recently, Baddeley (2000, 2001) has pro- other models of consciousness in its abil- posed an additional WM component called ity to brieﬂy hold multimodal information the episodic buffer (see Figure 8.3 for a and to combine many information sources depiction of the most recent model). This into a unitary representation. A major dif- addition to the WM architecture means that ference between WM and other models is the central executive now becomes a purely that WM was not proposed as a model attentional, controlled process while mul- of consciousness in general. It is restricted timodal information storage devolves onto to an accounting of mnemonic processes – the episodic buffer. The episodic buffer both conscious and unconscious. In addition, “comprises a limited capacity system that the WM model does not assume that con- provides temporary storage of information tents of the episodic buffer are “broadcast” held in a multimodal code, which is capa- systemwide as a means of organizing and ble of binding information from the sub- recruiting other non-mnemonic processes. sidiary systems, and from long-term mem- No account is given of the further distribu- ory, into a unitary episodic representation. tion of information from the episodic buffer, Conscious awareness is assumed to be the once it is accessed by the central executive. principal mode of retrieval from the buffer” (Baddeley, 2000,p.417). Baddeley (2001) schneider and pimm-smith’s believes that the binding function served by message-aware control mechanism the episodic buffer is “the principal biologi- Schneider and Pimm-Smith have proposed cal advantage of consciousness” (p. 858). a model of cognition that incorporates Conscious processing in WM appears to a conscious processing component and reside in the transactions of the central exec- allows widespread distribution of informa- utive with the episodic buffer (and perhaps tion from specialist modules (Schneider & P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

cognitive theories of consciousness 189

VISION

Modality specific messages SPEECH Inner Loop TOUCH

Consciousness Activity & consists of this Priority Codes process & its interactions

Message aware controller Gain control Goal systems

Attentional controller

Figure 8.4. A simpliﬁed view of Schneider and Pimm-Smith’s Message-Aware Control Mechanism.

Pimm-Smith, 1997). The model is an to all information flowing within the sys- attempt to capture the adaptive advantage tem, examining data only after lower-level that consciousness adds to cognitive pro- modules have produced invariant codes and cessing. According to Schneider and Pimm- selecting those messages that relate to cur- Smith, “consciousness may be an evolution- rently active goals. Figure 8.4 illustrates the ary extension of the attentional system to functional relations among components in modulate cortical information flow, provide Schneider and Pimm-Smith’s model. awareness, and facilitate learning particu- The message-aware control mechanism larly across modalities. . . . According to [the] model of consciousness depends on local- model, consciousness functions to monitor ized specialist processors – auditory, haptic, and transmit global messages that are gener- visual, speech, motor, semantic, and spatial ally received by the whole system serially to modules – which each have their own inter- avoid the cross-talk problem (pp. 65, 76).” nal levels of processing. These modules feed The conscious component of this model, their output codes serially and separately the conscious controller, stands between to other modules and to the consciousness the serially arriving high-level messages for- controller via an inner loop. According to warded by the specialist modules and the Schneider and Pimm-Smith, “Consciousness attentional controller, which sends scalar is a module on the inner loop that receives messages back to the specialist modules messages that evoke symbolic codes that are according to their value in ongoing activi- utilized to control attention to specific goals ties. The conscious controller is not privy (p. 72).” Furthermore, “consciousness is the P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

190 the cambridge handbook of consciousness

message awareness of the controller. . . . This of neurons, most network theories, including message awareness allows control decisions those described here, are in fact functional based on the specific messages transmitted in descriptions of activities that are presumed the network rather than labeled line activity to occur in the brain. Brain networks them- or priority codes from the various modules selves are not directly assessed. in the network (pp. 72–73).” There is also information sharing through non-conscious pribram’s holonomic theory channels. The attentional controller receives Karl Pribram has developed a holographic activity and priority signals from specialist theory (more recently, holonomic theory) modules marking the availability of input of brain function and conscious processing through non-conscious outer loop channels, (Pribram, 1971). He has built on the mathe- and it sends gain control information back to matical formulations of Gabor (1946), com- the specialists. bining holography and communication the- Schneider and Pimm-Smith’s model has ory. To state things in simple form, the several notable characteristics: (1) informa- holonomic model postulates that the brain tion is widely distributed; (2) particular is holograph machine. That is, the brain content is created by localized specialist handles, represents, stores, and reactivates modules, which themselves have levels of information through the medium of “wet- processing; (3) consciousness is identified as ware” holograms. Holograms, though com- a separable and separate function that mod- plex mathematically, are familiar to us as the ulates the flow of information throughout three-dimensional images found on credit the system: and (4) access to consciousness is cards and driver’s licenses. Such a hologram through reference to goal systems. This ele- is a photograph of an optical interference gant model accounts for the modulation of pattern, which contains information about attention by reference both to current goals intensity and phase of light reflected by an and to the recruitment of specialist modules object. When illuminated with a coherent in the service of goals through the global dis- source of light, it will yield a diffracted tribution of conscious messages. wave that is identical in amplitude and phase distribution with the light reflected from the original object. The resulting three- Network Theories That Explain dimensional image is what we see on our Consciousness as Patterns of credit cards. System-Wide Activity In a holonomic model, Pribram says, Although the information-processing theo- information is encoded by wave interfer- ries in the previous section attempted to ence patterns rather than by the binary units model consciousness in terms of the selec- (BITs) of computer science. Rather than tion and interactions of specialized, semi- encoding sensory experience as a set of fea- autonomous processing modules, the theo- tures that are then stored or used in informa- ries described in the present section build tion processing, sensory input in the holo- on networks and connectionist webs. It nomic view is encoded as the interference may be helpful to view the information- pattern resulting from interacting waves of processing models as macroscopic and the neuronal population activity. Stated in the network theories as microscopic. It is possi- language of visual perception, the retinal ble that the activities attributed to modules image is understood to be coded by a spa- in the information-processing models could, tial frequency distribution over visual cor- in fact, be seen to be carried out by connec- tex, rather than by individual features of the tionist networks when viewed microscopi- visual scene. Because the surface layer of the cally. As we examine network theories, it cortex consists of an entangled “feltwork” should be kept in mind that, although net- of dendrites, Pribram suggests that cortex work theorists often make comparisons with represents a pattern of spatial correlations in and assumptions about networks composed a continuous dendritic medium, rather than P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

cognitive theories of consciousness 191

by discrete, localized features expressed by man in its brain, and so on ad infinitum). the firing of single neurons. The dendritic It is the activity of the dendro-dendritic web of the surface layers of cortex is the web that gives rise to the experience. Cor- medium in which representations are held. respondingly, remembering is a form of re- Functionally, it is composed of oscillating experiencing or re-constructing the initial graded potentials. In Pribram’s view, nerve sensory input, perhaps by cuing a portion impulses have little or no role to play in of the interference pattern. Finally, Pribram the brain web. However, dendritic potentials believes that we “become aware of our con- obviously trigger axonal firing when they rise scious experience due to a delay between an above a neuronal threshold, so that long- incoming pattern of signals before it matches distance axons would also be triggered by a previously established outgoing pattern” the dendro-dendritic feltwork. (Pribram & Meade, 1999,p.207). The holonomic model finds support in Pribram’s holonomic model is attractive the neuropsychological finding that focal in that it can account for the distributed brain damage does not eliminate memory properties of memory and sensory process- content. Further, it is consistent with the fact ing. The model makes use of the dendritic that each sensory neuron tends to prefer one feltwork that is known to exist in the sur- type of sensory input, but often fires to dif- face layer of cortex. However, the model ferent inputs as well. Neurons do not operate fails to help us understand the difference individually but rather participate in differ- between conscious and unconscious process- ent cell assemblies or active populations at ing or the unique functions of consciousness different times; the cell assemblies are them- qua consciousness. Pribram has not treated selves “kaleidoscopic.” consciousness as a variable and cannot tell In a hologram, the information neces- us what it is that consciousness adds to the sary to construct an image is inherently dis- cognitive system. tributed. Pribram explains how the notion of distributed information can be illustrated edelman and tononi’s dynamic with a slide projector (Pribram & Meade, core hypothesis 1999). If one inserts a slide into the projec- Edelman and Tononi’s theory of conscious- tor and shows a “figure,” then removes the ness (2000; see also Tononi & Edelman, lens from the front of the projector, there 1998) combines evidence from large-scale is only a fuzzy bright area. There is noth- connectivities in the thalamus and cortex, ing, “no-thing,” visible on the screen. But that behavioral observation, and mathematical does not mean there is no information in the properties of large-scale brain-like networks. light. The information can be re-visualized Based on neuropsychological and lesion evi- by placing a pair of reading glasses into the dence that consciousness is not abolished light beam. On the screen, one now again by losses of large volumes of brain tissue sees the “figure” in the slide. Putting two (Penfield, 1958), Edelman and Tononi reject lenses of the eyeglasses in the beam, one sees the idea that consciousness depends on par- two “figures” that can be made to appear in ticipation of the whole brain in an undif- any part of the bright area. Thus any part ferentiated fashion. At the same time, they, of the beam of light carries all the informa- along with many others, reject the view that tion needed to reconstruct the picture on the consciousness depends only on local prop- slide. Only resolution is lost. erties of neurons. Tononi and Edelman cite, According to the holonomic theory, no for example, PET evidence suggesting that “receiver” is necessary to “view” the result moment-to-moment awareness is highly cor- of the transformation (from spectral holo- related with increasing functional connec- graphic to “image”), thus avoiding the tivity between diverse cortical regions (see, homunculus problem (the problem of infi- for example, McIntosh, Rajah, & Lobaugh, nite regress, of a little man looking at the 1999). In other words, the same cortical areas visual scene, which in turn needs a little seem to participate in conscious experience P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

192 the cambridge handbook of consciousness

or not at different times, depending on their Before counting occurs, each member of current dynamic connectivity. This idea is parliament is interacting with as many resonant with Pribram’s description of neu- other members as possible not by persua- ral assemblies being “kaleidoscopic.” sive rhetoric . . . but by simply pushing and The fundamental idea in Edelman and pulling. Within 300 msec., a new vote is Tononi’s theory is the dynamic core hypoth- taken. How informed the decision turns out to be will depend on the number of esis, which states that conscious experience diverse interactions within the parliament. arises from the activity of an ever-changing In a totalitarian country, every member will functional cluster of neurons in the thala- vote the same; the information content of mocortical complex of the brain, character- constant unanimity is zero. If there are two ized by high levels of differentiation as well monolithic groups, left and right, such that as strong reciprocal, re-entrant interaction the vote of each half is always the same, the over periods of hundreds of milliseconds. information content is only slightly higher. The particular neurons participating in the If nobody interacts with anyone, the vot- dynamic core are ever changing while inter- ing will be purely random, and no infor- nal integration in the dynamic core is main- mation will be integrated within the sys- tained through re-entrant connections. The tem. Finally, if there are diverse interactions within the parliament, the ﬁnal vote will hypothesis highlights the functional connec- be highly informed (Edelman & Tononi, tions of distributed groups of neurons, rather 2000,pp.245–246). than their local properties; thus, the same group of neurons may at times be part of the A constantly changing array of ever- dynamic core and underlie conscious experi- reorganized mid-sized neuronal groups in ence, whereas at other times, this same neu- a large system of possible groups has high ronal group will not be part of the dynamic levels of complexity and integration – char- core and will thus be part of unconscious acteristics of conscious states. Within this processing. Consciousness in this view is not model, unconscious specialist systems are a thing or a brain location but rather, as local, non-integrated neuronal groups. How William James argued, a process, occurring the unconscious specialists are recruited into largely within the re-entrant meshwork of the dynamic core is not made entirely clear the thalamocortical system. in the theory. Edelman and Tononi say that Edelman and Tononi (2000) take issue consciousness in its simplest form emerges in with Baars’ (1988) concept of global broad- the re-entrant linkage between current per- casting (see below) as a way to explain ceptual categorization and value-category capacity limits and wide access in conscious memory (short-term and long-term mem- processing. In Baars’ view, the information ory). Conscious experience is actually a suc- content of any conscious state is apparently cession of 100-ms snapshots of the current contained in the single message that is being linkages that constitute the “remembered broadcast to specialist systems throughout present.” the brain at any one moment; informa- Perhaps Baars and Tononi and Edelman tion content is thus limited but widely dis- are not so different on closer examination. tributed. Edelman and Tononi (2000) argue Baars’ (1988, 1998) model supposes that for an alternative view: that the informa- there is reciprocal exchange between the tion is not in the message, but rather in the global workspace (GW) and specialist sys- number of system states that can be brought tems in the architecture of consciousness; it about by global interactions within the sys- is difﬁcult to see why this is different from tem itself. In place of Baars’ broadcasting or the re-entrant linkages between neuronal theater metaphor, they offer an alternative: groups in Edelman and Tononi’s theory. Fur- thermore, within any one “snapshot” of the [A] better metaphor would be . . . a riotous system, the pattern of dynamically linked parliament trying to make a decision, sig- elements in Baars’ model – GW and spe- naled by its members raising their hands. cialists that are able to receive the particular P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

cognitive theories of consciousness 193

Figure 8.5. A Hilbert analysis of analytic phase differences in EEG across cortical surface measured over 400 ms in rabbit and human conscious processing. Phase differences are calculated in the beta band (12–30 Hz) for human EEG and in the gamma band (20–50 Hz) for the rabbit EEG. (With permission of the author.) (See color plates.)

message that has been disseminated – looks Freeman & Rogers, 2003). Freeman and his very much like the pattern of momen- colleagues have analyzed EEGs, recorded tarily linked neuronal groups in Tononi from multiple high-density electrode arrays and Edelman’s model that are recruited in (64 electrodes) fixed on the cortex of rab- the moment depending on environmental bits and on the scalp of human volunteers. input and value memories. Two strengths An index of synchronization was obtained of the Tononi and Edelman model are its for pairs of signals located at different cor- acknowledgment of long-distance connec- tical sites to detect and display epochs tivity among specialist brain regions as a of mutual engagement between pairs. The characteristic of conscious processes, as well measure was adapted to derive an index of as the dynamic nature of these connections. global synchronization among all four cortices (frontal, parietal, temporal, and occipital) – global epochs of phase stabilization walter freeman’s dynamical systems (“locking”) involving all cortices under obser- approach: frames in the cinema vation during conscious perceptual activity. Like Pribram, Walter Freeman has worked These epochs of phase locking can be seen in to obtain empirical support for a cortex- the “plateaus” of global coherence in Figure wide dynamic neural system that can 8.5. The peaks in the figure indicate momen- account for behavioral data observed in tary, global decoherence. conscious activities. Freeman’s Dynamical To understand Freeman’s findings, we Systems approach to consciousness is built have to understand the basics of Hilbert on evidence for repetitive global phase analysis as it is shown in Figure 8.5. Hilbert transitions occurring simultaneously over analysis of the EEGs recorded from elec- multiple areas of cortex during normal trode arrays produces a three-dimensional behavior (see Figure 8.5 Freeman, 2004; graphical representation. In it, the phase P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

194 the cambridge handbook of consciousness

difference between pairs of cortical elec- sion that can change in an all-or-none fash- trodes within a particular EEG band is plot- ion with small environmental perturbations. ted against time (in milliseconds) and spatial He says “a large system can hold itself in a location (represented by electrode number). near-unstable state, so that by a multitude The resulting plot is called a Hilbert space. of adjustments it can adapt to environments The Hilbert space can be read like a topo- that change continually and unpredictably” graphical map. In the plot in Figure 8.5, (Freeman & Rogers, 2003,p.2882). we can see many flat plateau areas lying between peaked ridges. The plateaus repre- Globalist Models That Combine Aspects 50 sent time periods (on the order of ms) of Information-Processing Theories in which many pairs of EEG signals from and Network Theories different cortical locations are found to be in phase with each other. The ridges repre- baars’ global workspace theory sent very short intervals when all of these A theater metaphor is the best way to pairs are simultaneously out of phase, before approach Baars’ Global Workspace (GW) returning to phase locking. These out-of- theory (Baars, 1988, 1998, 2001). Con- phase or decoherent epochs appear to be sciousness is associated with a global “broad- non-conscious transitions between moments casting system” that disseminates informa- of consciousness. tion widely throughout the brain. The According to Freeman (2004), metaphor of broadcasting explicitly leaves open the precise nature of such a wide influ- The EEG shows that neocortex processes ence of conscious contents in the brain. It information in frames like a cinema. The could vary in signal fidelity or degree of dis- perceptual content is found in the phase tribution, or it might not involve “labeled plateaus from rabbit EEG; similar content line” transmission at all, but rather activation is predicted to be found in the plateaus of passing, as in a neural network. Metaphors human scalp EEG. The phase jumps show are only a first step toward explicit the- the shutter. The resemblance across a 33- fold difference in width of the zones of coor- ory, and some theoretical decision points are dinated activity reveals the self-similarity explicitly left open. of the global dynamics that may form If consciousness is involved with wide- Gestalts (multisensory percepts). (Caption spread distribution or activation, then con- to cover illustration, p. i) scious capacity limits may be the price paid for the ability to make single momentary Freeman’s data are exciting in their messages act upon the entire system for pur- ability to map the microscopic tempo- poses of coordination and control. Because ral dynamic changes in widespread corti- at any moment there is only one “whole cal activity during conscious perception – system,” a global dissemination capacity something not found in other theories. As a must be limited to one momentary content. theory of consciousness, the dynamical sys- (There is evidence that the duration of each tems approach focuses primarily on describ- conscious “moment” may be on the order of ing conscious perceptual processing at the 100 ms, one-tenth of a second – see Blumen- cortical level. It does not attempt to explain thal, 1977). the conscious/non-conscious difference or Baars develops these ideas through seven the function of consciousness in the neu- increasingly detailed models of a global ropsychological system. With many neu- workspace architecture, in which many par- rocognitive theorists, we share Freeman’s allel unconscious experts interact via a serial, question about how the long-range global conscious, and internally consistent global state changes come about virtually simul- workspace (1983, 1988). Global workspace taneously. Freeman’s hypothesis of Self- architectures or their functional equivalents organized Criticality suggests that the neural have been developed by cognitive scientists system is held in a state of dynamic ten- since the 1970s; the notion of a “blackboard” P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

cognitive theories of consciousness 195

where messages from specialized subsys- of unconscious experts, which report their tems can be “posted” is common to the answer to consciousness. Figure 8.6 shows work of Baars (1988), Reddy and Newell the major constructs in GW theory and the (1974), and Hayes-Roth (1984). The global functional relations among them. workspace framework has a family resem- The second construct is, of course, the blance to the well-known integrative the- global workspace (GW) itself. A global ories of Herbert A. Simon (General Prob- workspace is an architectural capability for lem Solver or EPAM), Allan Newell (SOAR, system-wide integration and dissemination 1992), and John R. Anderson (ACT*, 1983). of information. It is much like the podium Architectures much like this have also seen at a scientific meeting. Groups of experts at some practical applications. GW theory is such a meeting may interact locally around currently a thoroughly developed frame- conference tables, but to influence the meet- work, aiming to explain an large set of evi- ing as a whole any expert must compete with dence. It appears to have fruitful implica- others, perhaps supported by a coalition of tions for a number of related topics, such like-minded experts, to reach the podium, as spontaneous problem solving, voluntary whence global messages can be broadcast. control, and even the Jamesian “self” as agent New links among experts are made possible and observer (Baars, 1988; Baars, Ramsoy, & by global interaction via the podium and can Laureys, 2003). then spin off to become new local proces- GW theory relies on three theoretical sors. The podium allows novel expert coali- constructs: unconscious specialized proces- tions to form that can work on new or dif- sors, a conscious Global Workspace, and ficult problems, which cannot be solved by unconscious contexts. established experts and committees. Tenta- The first construct is the unconscious spe- tive solutions to problems can then be glob- cialized processor, the “expert” of the psy- ally disseminated, scrutinized, and modified. chological system. We know of hundreds The evidence presented in Tables 8.2 of types of “experts” in the brain. They and 8.3 falls into place by assuming that may be single cells, such as cortical fea- information in the global workspace cor- ture detectors for color, line orientation, or responds to conscious contents. Because faces, or entire networks and systems of neu- conscious experience seems to be oriented rons, such as cortical columns, functional primarily toward perception, it is conve- areas like Broca’s or Wernicke’s areas, and nient to imagine that preperceptual pro- basal ganglia. Like human experts, uncon- cessors – visual, auditory, or multimodal – scious specialized processors may sometimes can compete for access to a brain version be quite “narrow-minded.” They are highly of a global workspace. For example, when efficient in limited task domains and able someone speaks to us, the speech stream to act independently or in coalition with receives preperceptual processing through each other. Working as a coalition, they do the speech specialist systems before the mes- not have the narrow capacity limitations of sage in the speech stream is posted in con- consciousness, but can receive global mes- sciousness. This message is then globally sages. By “posting” messages in the global broadcast to the diverse specialist systems workspace (consciousness), they can send and can become the basis for action, for com- messages to other experts and thus recruit posing a verbal reply, or for cuing related a coalition of other experts. For routine mis- memories. In turn, the outcome of actions sions they may work autonomously, with- carried out by expert systems can also be out conscious involvement, or they may dis- monitored and returned to consciousness as play their output in the global workspace, action feedback. thus making their work conscious and avail- Obviously the abstract GW architecture able throughout the system. Answering a can be realized in a number of different question like “What is your mother’s maiden ways in the brain, and we do not know at name?” requires a mission-specific coalition this point which brain structures provide P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

196 the cambridge handbook of consciousness

Context Hier archy Individual Conte xts: Goal, (Unconscious) Perceptual, and Conceptual Contexts

Conscious Global Stimulus Input Contexts W orkspace Vision (Conscious) Input Competition Acoustic

Spresthesia Broadcasting Unconscious Specialized Processors Balance

Language Automatisms Face Recognition Image Generation Skill Components

Speech and Action Output Figure 8.6. Global Workspace Architecture: Basic constructs and their relations.

the best candidates. Although its brain cor- be defined functionally as knowledge struc- relates are not entirely clear at this time, tures that constrain conscious contents with- there are possible neural analogs, includ- out being conscious themselves, just as the ing the reticular and intralaminar nuclei playwright determines the words and actions of the thalamus, one or more layers of of the actors on stage without being visible. cortex, long-range cortico-cortico connec- Conceptually, contexts are defined as pre- tions, and/or active loops between sensory established expert coalitions that can evoke, projection areas of cortex and the corre- shape, and guide global messages without sponding thalamic relay nuclei. Like other themselves entering the global workspace. aspects of GW theory, such neural candi- Contexts may be momentary, as in the dates provide testable hypotheses (Newman way the meaning of the first word in a sen- & Baars, 1993). All of the neurobiologi- tence shapes an interpretation of a later word cal proposals described in this chapter pro- like “set,” or they may be long lasting, as with vide candidates (Freeman, 2004; Dehaene & life-long expectations about love, beauty, Naccache, 2001; Edelman & Tononi, 2000; relationship, social assumptions, professional Tononi & Edelman, 1998), and some have expectations, worldviews, and all the other been influenced by GW theory. things people care about. Although contex- Context, the third construct in GW the- tual influences shape conscious experience ory, refers to the powers behind the scenes without being conscious, contexts can also of the theater of mind. Contexts are coali- be set up by conscious events. The word tions of expert processors that provide the “tennis” before “set” shapes the interpreta- director, playwright, and stagehands behind tion of “set,” even when “tennis” is already the scenes of the theater of mind. They can gone from consciousness. But “tennis” was P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

cognitive theories of consciousness 197

B C B C B C D D D A A A

“Is C relevant?” “Is A relevant?” “Is B relevant?”

Figure 8.7. A naive processor approach to environmental novelty.

initially conscious and needed to be con- out all features of the new environment? scious to create the unconscious context that This task quickly becomes computation- made sense of the word “set.” ally prohibitive. Shanahan and Baars (2005) Thus conscious events can set up uncon- point out that the following: scious contexts. The reader’s ideas about consciousness from years ago may influence What the global workspace architecture his or her current experience of this chapter, has to offer . . . is a model of information flow that explains how an information- even if the memories of the earlier thoughts ally unencapsulated process can draw on do not become conscious again. Earlier expe- just the information that is relevant to the riences typically influence current experi- ongoing situation without being swamped ences as contexts, rather than being brought by irrelevant rubbish. This is achieved to mind. It is believed for example that by distributing the responsibility for decid- a shocking or traumatic event earlier in ing relevance to the parallel specialists life can set up largely unconscious expecta- themselves. The resulting massive paral- tions that may shape subsequent conscious lelism confers great computational advan- experiences. tage without compromising the serial flow of conscious thought, which corresponds to the sequential contents of the limited capac- shanahan: an answer to the ity global workspace.... modularity and frame problems Shanahan and Baars (2005) suggest that Compare the naive processor’s inefficient the global workspace approach may pro- approach (depicted in Figure 8.7) with a vide a principled answer to the widely dis- massively parallel and distributed global cussed “modularity” and “frame” problems. workspace approach (depicted in Figure 8.8) Fodor (1983) developed the view that cog- to dealing with environmental novelty. nitive functions like syntax are performed The key point here is that the GW by “informationally encapsulated” modules, architecture permits widely distributed local an idea that has some empirical plausi- responsibility for processing global signals. bility. However, as stated by Fodor and As was pointed out above, conscious and others, modules are so thoroughly isolated non-conscious process differ in their capabil- from each other that it becomes difficult to ities – they are two different modes of pro- explain how they can be accessed, changed, cessing that, when combined, offer powerful and mobilized on behalf of general goals. adaptive possibilities. A closely related difficulty, called the frame problem, asks how an autonomous agent can franklin’s ida as an implementation deal with novel situations without following of gw theory out all conceivable implications of the novel Stan Franklin and colleagues (Franklin, 2001; event. For example, a mobile robot on a cart Franklin & Graesser, 1999) have developed may roll from one room to another. How a practical implementation of GW theory does it know what is new in the next room in large-scale computational agents to test and what is not, without explicitly testing its functionality in complex practical tasks. P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

198 the cambridge handbook of consciousness

Am I Encapsulated specialist processors Am I relevant? relevant? YES!

Am I Am I relevant? B C B C relevant? D D A A

Broadcasting… Conscious solution

Global Workspace … Figure 8.8. A GW approach to environmental novelty.

IDA, or Intentional Distribution Agent, the (see http://csrg.cs.memphis.edu). Although current implementation of the extended agent simulations do not prove that GW GW architecture directed by Franklin, is architectures exist in the brain, they demon- designed to handle a very complex artifi- strate their functionality. Few if any large- cial intelligence task normally handled by scale cognitive models can be shown to trained human beings (see Chapter 7). The actually perform complex human tasks, particular domain in this case is interac- but somehow the real cognitive architec- tion among U.S. Navy personnel experts and ture of the brain does so. In that respect, sailors who move from job to job. IDA inter- the test of human-level functionality is as acts with sailors via e-mail and is able to important in its way as any other source combine numerous regulations, sailors’ pref- of evidence. erences, and time, location, and travel considerations into human-level performance. dehaene’s global neuronal Although it has components roughly cor- network theory responding to human perception, memory, Stanislas Dehaene and his colleagues and action control, the heart of the sys- (Dehaene & Naccache, 2001; Dehaene, tem is a GW architecture that allows input Kerszberg, & Changeux, 1998) have recently messages to be widely distributed, so that proposed a global neuronal workspace theory specialized programs called “codelets” can of consciousness based on psychological respond with solutions to centrally posed and neuroscientific evidence quite similar problems (see Figure 8.9). to that cited by Baars and others. Dehaene Franklin writes, “The fleshed out global and colleagues identify three empirical workspace theory is yielding hopefully observations that any theory of conscious- testable hypotheses about human cogni- ness must be able to account for: “namely tion. The architectures and mechanisms (1) a considerable amount of processing is that underlie consciousness and intelli- possible without consciousness, (2) atten- gence in humans can be expected to yield tion is a prerequisite of consciousness, and information agents that learn continuously, (3) consciousness is required for some adapt readily to dynamic environments, specific cognitive tasks, including those and behave flexibly and intelligently when that require durable information mainte- faced with novel and unexpected situations” nance, novel combinations of operations, or P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

cognitive theories of consciousness 199

Stimulus Sensory Autobiographical from Internal Memory Memory Environment Internal Senses Cue Local Associations

Perception Codelets Working Long-term Stimulus Senses Consolidation Working Memory from External Memory Percept Environment Perceptual Memory Cue Local Associations Action Selected (Slipnet) and Taken (behavior codelets) Transient Episodic Memory Attention Update Codelets Action Selected and Taken Codelets Update (behavior codelets) Behavior Codelets Conscious in priming mode Broadcast instantiate, bind Procedural Update Action activate Winning Competition for Selection Coalition (Behavior Net) Consciousness

Figure 8.9. Franklin’s IDA Model.

the spontaneous generation of intentional long-distance connectivity is needed that can behavior” (p. 1). The Dehaene and Naccache “potentially interconnect multiple special- model depends on several well-founded ized brain areas in a coordinated, though assumptions about conscious functioning. variable manner” (p. 13). The first assumption is that non-conscious The third assumption concerns the role mental functioning is modular. That is, of attention in gating access to conscious- many dedicated non-conscious modules can ness. Dehaene and Naccache (2001) review operate in parallel. Although arguments evidence in support of the conclusion that remain as to whether psychological mod- considerable processing can occur without ules have immediate correlates in the brain, attention, but that attention is required for Dehaene and Naccache (2001) say that the information to enter consciousness (Mack & “automaticity and information encapsula- Rock, 1998). They acknowledge a similarity tion acknowledged in cognitive theories are between Michael Posner’s hypothesis of an partially reflected in modular brain circuits.” attentional amplification (Posner, 1994) and They tentatively propose that “a given pro- their own proposal. Attentional amplifica- cess, involving several mental operations, can tion explains the phenomena of conscious- proceed unconsciously only if a set of ade- ness as due to the orienting of attention, quately interconnected modular systems is which causes increased cerebral activation available to perform each of the required in attended areas and a transient increase operations” (p. 12; see Figure 8.10). in their efficiency. According to Dehaene & The second assumption, one shared Nacache (2001), by other cognitive theories, is that controlled processing requires an architecture [I]nformation becomes conscious . . . if the in addition to modularity that can estab- neural population that represents it is mobilized by top-down attentional amplification lish links among the encapsulated proces- 1998 into a brain-scale state of coherent activ- sors. Dehaene et al. ( ) argue that a dis- ity that involves many neurons distributed tributed neural system or “workspace” with throughout the brain. The long-distance P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

200 the cambridge handbook of consciousness

connectivity of these ‘workspace neurons’ ity, presumably generating the P1,N1, and can, when they are active for a minimal N400 waveforms, would propagate with- duration, make the information available out necessarily creating a global reverber- to a variety of processes including percep- ant state. However, a characteristic neural tual categorization, long-term memoriza- signature of long-lasting distributed activity tion, evaluation, and intentional action. and g-band emission, presumably generat- (p. 1) ing the P300 waveform, would be associated with global access. (p. 8520) An implication of the Dehaene and Nac- In the simulation, a network modeled the cache model is that consciousness has a gran- cell assemblies evoked by T1 and T2 through ularity, a minimum duration of long-distance four hierarchical stages of processing, two integration, below which broadcast informa- separate perceptual levels and two higher tion will fail to be conscious. association areas. The network was initially It is worth noting a small difference assigned parameters that created sponta- between Baars’ version of global workspace neous thalamocortical oscillations, simulat- and that of Dehaene and colleagues. ing a state of wakefulness. Then, the network (Dehaene & Naccache, 2001; Dehaene et al., was exposed to T1 and T2 stimulation at var- 1998). They believe that a separate atten- ious interstimulus intervals (ISI). T1 exci- tional system intervenes with specialized tation was propagated bottom-up through processors to allow their content to enter all levels of the processing hierarchy, fol- the global workspace and become conscious. lowed by top-down amplification signals Baars (1998), on the other hand, sees atten- that resulted in sustained firing of T1 neu- tion not as a separate system but rather as rons. Dehaene et al. (2003) hypothesized the name for the process of gaining access that this sustained firing and global broad- to global workspace by reference to long- casting may be the neural correlate of con- term or current goals. Clearly, further refine- scious reportability. In contrast, the activa- ment is needed here in thinking through tion evoked by T2 depended closely on its what we mean by attention or an attentional timing relative to T1. For simultaneous and system as separate from the architecture of long ISIs, T2 excitation evoked sustained fir- consciousness, in this case, varieties of GW ing. Importantly, when T2 was presented architecture. during T1-elicited global firing, it evoked Dehaene, Sargent, and Changeux (2003) activation only in the low-level perceptual have used an implementation of the global assemblies and resulted in no global propaga- neuronal workspace model to successfully tion. Dehaene and colleagues conclude that simulate attentional blink. Attentional blink this detailed simulation has provided ten- is a manifestation of the all-or-none char- tative links between subjective reports and acteristic of conscious processing observed “objective physiological correlates of con- when participants are asked to process two sciousness on the basis of a neurally plausible successive targets, T1 and T2. When T2 is architecture” (2003,p.8524). presented between 100 and 500 ms after T1, the ability to report it drops, as if the participants’ attention had “blinked.” Dur- The Globalist Argument: 2 300 ing this blink, T fails to evoke a P An Emerging Consensus potential but still elicits event-related potentials associated with visual and semantic processing (P1,N1, and N400). Dehaene et al. In the last two decades, a degree of (2003) explain, consensus has developed concerning the role of consciousness in the neuropsy- Our simulations aim at clarifying why chological architecture. The general posi- some patterns of brain activity are selection is that consciousness operates as a tively associated with subjective experience. distributed and flexible system offering non- In short, during the blink, bottom-up activ- conscious expert systems global accessibility P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

cognitive theories of consciousness 201

hierarchy of modular high-level processors processors with strong long-distance interconnectivity

Perceptual categorization Long-term memory Evaluation (affect) Intentional action

automatically processors activated mobilized processors into the conscious workspace

Figure 8.10. A global neuronal network account of conscious processes (Dehaene & Naccache, 2001,p.27).

to information that has a high concurrent in some mysterious way, a privileged value to the organism. Although conscious- correlation with consciousness. Instead, ness is not itself an executive system, a global this hypothesis accounts for fundamen- distribution capacity has obvious utility for tal properties of conscious experience by executive control, in much the way that gov- linking them to global properties of par- ernments can control nations by influencing ticular neural processes” (p. 1850). r nation-wide publicity. Llinas et al. (1998): “...the thalamus rep- Excerpted below are the views of promi- resents a hub from which any site in the cor- nent researchers on consciousness revealing tex can communicate with any other such considerable agreement. site or sites. . . . temporal coincidence of specific and non-specific thalamic activity r Baars (1983): “Conscious contents pro- generates the functional states that char- vide the nervous system with coherent, acterize human cognition” (p. 1841). r global information.” Edelman and Tononi (2000): “When r Damasio (1989): “Meaning is reached by we become aware of something . . . it is time-locked multiregional retroactivation as if, suddenly, many different parts of of widespread fragment records. Only the our brain were privy to information latter records can become contents of that was previously confined to some consciousness.” specialized subsystem. . . . the wide dis- r Freeman (1991): “The activity patterns tribution of information is guaranteed that are formed by the (sensory) dynam- mechanistically by thalamocortical and ics are spread out over large areas of cortex, corticocortical reentry, which facilitates the not concentrated at points. Motor out- interactions among distant regions of the flow is likewise globally distributed....In brain” (pp. 148–149). r other words, the pattern categorization Dennett (2001): “Theorists are converg- does not correspond to the selection of ing from quite different quarters on a a key on a computer keyboard but to version of the global neuronal workspace an induction of a global activity pattern.” model of consciousness” (p. 42). r [Italics added] Kanwisher (2001): “ . . . it seems reason- r Tononi and Edelman (1998): “The able to hypothesize that awareness of a dynamic core hypothesis avoids the cat- particular element of perceptual informa- egory error of assuming that certain tion must entail not just a strong enough local, intrinsic properties of neurons have, neural representation of information, but P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

202 the cambridge handbook of consciousness

also access to that information by most of cialist systems, which interact in a the rest of the mind/brain.” dynamic way via a global workspace. r Dehaene and Naccache (2001): “We 2. The function of the workspace is global propose a theoretical framework . . . the distribution of information in order to hypothesis of a global neuronal recruit resources in the service of current workspace. . . . We postulate that this goals. global availability of information through 3. Specialist systems compete for access to the workspace is what we subjectively the global workspace; information that experience as the conscious state.” achieves access to the workspace obtains r Rees (2001): “One possibility is that system-wide dissemination. activity in such a distributed network 4. Access to the global workspace is “gated” might reflect stimulus representations by a set of active contexts and goals. gaining access to a ‘global workspace’ that constitutes consciousness” (p. 679). r dissenting views John et al. (2001): “Evidence has been The globalist position argues that conscious- steadily accumulating that information ness provides a momentary unifying influ- about a stimulus complex is distributed ence for a complex system through global to many neuronal populations dispersed distribution and global access. In this sense, throughout the brain.” consciousness may be said to have unity. r 2001 Varela et al. ( ): “ . . . the brain Alternative views chiefly depart from the . . . transiently settling into a globally con- globalist position on this point. They argue, sistent state . . . [is] the basis for the unity in one way or another, that consciousness is of mind familiar from everyday experi- not fundamentally unified. ence.” r One alternative view is that of Marcel Cooney and Gazzaniga (2003): “Inte- (1993) who argued for “slippage” in the grated awareness emerges from mod- unity of consciousness. In part, he made his ular interactions within a neuronal case based on his observation that differ- workspace. . . . The presence of a large- ent reporting modalities (blink vs. finger tap) scale network, whose long-range con- could produce conflicting reports about connectivity provides a neural workspace scious experience. Marcel took this to indi- through which the outputs of numerous, cate that consciousness itself is not unified specialized, brain regions can be intercon- in any real sense. nected and integrated, provides a promis- Marcel’s argument bears some similar- ing solution . . . In the workspace model, ity to Dennett’s “multiple drafts” argument outputs from an array of parallel pro- (Dennett, 1991). Dennett pointed to the cessors continually compete for influence puzzle posed by the phi phenomenon. In within the network” (p. 162). the phi phenomenon, we observe a green r Block (2005): “Phenomenally conscious light and a red light separated by a few content is what differs between experi- degrees in the field of vision as they are ences as of red and green, whereas access flashed in succession. If the time between conscious content is information which is flashes is about one second or less, the first ‘broadcast’ in the global workspace.” light flashed appears to move to the position of the second light. Further, the color of Although debate continues about the func- the light appears to change midway between tional character of consciousness, the glob- the two lights. The puzzle is explaining how alist position can be summarized in the fol- we could see the color change before we lowing propositions: see the position of the second light. Dennett hypothesizes that the mind creates different 1. The architecture of consciousness com- analyses or narratives (multiple drafts) of the prises numerous, semi-autonomous spe- scene at different moments from different P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

cognitive theories of consciousness 203

sensory inputs. All of the accounts are avail- local specializations as well. It is the integra- able to influence behavior and report. A tion of local and global capacities that marks given scene can give rise to more than these theoretical approaches. Given the fact one interpretation. In contrast, with global that scientists have only “returned to con- broadcasting models, Dennett says there is sciousness” quite recently, this kind of con- no single version of the scene available any- vergence of opinion is both surprising and where in the psychological system. gratifying. Simiarly, Zeki (2001, 2003) has argued Future work should focus on obtain- on neurological grounds that there is “dis- ing neuroscientific evidence and correspond- unity” in the neural correlates of conscious- ing behavioral observations that can address ness. With many others, Zeki notes that the global access as the distinguishing feature of visual brain consists of many separate, func- consciousness. Additional work could con- tionally specialized processing systems that tribute simulations of the kind offered by are autonomous with respect to one another. Dehaene, Sargent, and Changeux (2003), He then supposes that activity at each node supporting the plausibility of all-or-none reaches a perceptual endpoint at a different global propagation of signals as models time, resulting in a perceptual asynchrony in of the neurocognitive architecture of con- vision. From there, Zeki makes the inference sciousness, and of Franklin, documenting that activity at each node generates a micro- the real-world potential of global workspace consciousness. He concludes that visual con- architectures as intentional agents. Further sciousness is therefore distributed in space work is also needed to resolve the issue and time, with an organizing principle of of whether consciousness is all-or-none, as abstraction applied separately within each Baars, Freeman, and Dehaene and his col- processing system. It remains to be seen leagues argue, or whether there are multiple whether Zeki’s microconsciousnesses can be drafts (Dennett, 1991) or microconscious- examined empirically via contrastive anal- nesses (Zeki, 2001, 2003) playing a role in ysis and whether the microconsciousnesses the architecture of consciousness (see also are necessarily conscious or simply poten- Chapter 15). tially conscious. The globalist position would argue that all neural processing is potentially conscious, depending on the needs and goals References of the system. Clearly, this is a point for future discussion. Atkinson, R., & Shiffrin, R. (1968). Human memory: A proposed system and its control processes. In K. W. Spence & J. Spence (Eds.), Conclusion Advances in the psychology of learning and motivation: Research and theory (Vol. 2,pp.89– 195). New York: Academic Press. This chapter suggests that current cognitive 1983 theories have much in common. Almost all Baars, B. J. ( ). Conscious contents provide the nervous system with coherent, global suggest an architectural function for con- information. In R. J. Davidson, G. Schwartz, sciousness. Although the reader’s experience & D. Shapiro (Eds.), Consciousness and self- of these words is no doubt shaped by fea- regulation (Vol. 3,pp.41–79). New York: ture cells in visual cortex, including word Plenum Press. recognition regions, such local activity is not Baars, B. J. (1988). A cognitive theory of conscious- sufficient for consciousness of the words. In ness. New York: Cambridge University Press. addition, some widespread functional brain Baars, B. J. (1998). In the theater of consciousness: capacity is widely postulated. Direct func- The workspace of the mind. New York: Oxford tional imaging evidence for that hypothesis University Press. is now abundant. In that sense, most current Baars, B. J. (2002). The conscious access models are globalist in spirit, which is not to hypothesis: Origins and recent evidence. Trends deny, of course, that they involve multiple in Cognitive Sciences, 6(1), 47–52. P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

204 the cambridge handbook of consciousness

Baars, B. J., Ramsoy, T., & Laureys, S. (2003). Fodor, J. (1983). The modularity of mind: An essay Brain, conscious experience, and the observ- on faculty psychology. Cambridge, MA: MIT ing self. Trends in Neurosciences, 26(12), 671– Press. 675. Franklin, S. (2001). Conscious software: A com- Baddeley, A. D. (2000). The episodic buffer: A putational view of mind. In V. Loia & S. Sessa new component of working memory? Trends in (Eds.), Soft computing agents: New trends for Cognitive Sciences, 4, 417–423. designing autonomous systems (pp. 1–46). Berlin: Baddeley, A. D. (2001). Is working memory still Springer (Physica-Verlag). working? American Psychologist, 56(11), 851– Franklin, S., & Graesser, A. (1999). A software 864. agent model of consciousness. Consciousness 285 301 Baddeley, A. D., & Hitch, G. J. (1974). Working and Cognition, 8, – . memory. In G. A. Bower (Ed.), Recent advances Freeman, W. J. (2004). Origin, structure, and role in learning and motivation (Vol. 8,pp.47–90). of background EEG activity. Part 1. Analytic New York: Academic Press. amplitude. Clinical Neurophysiology, 115, 2077– 2088 (Including issue cover). Basmajian, J. (1979). Biofeedback: Principles and 1991 practice for the clinician. Baltimore: Williams & Freeman, W. J. ( ). The physiology of percep- 78 85 Wilkins. tion. Scientiﬁc American, 264, – . 2003 Block, N. (2005). Two neural correlates of con- Freeman, W. J., & Rogers, L. ( ). A neurobi- sciousness, Trends in Cognitive Sciences, 9, 46– ological theory of meaning in perception. Part 52. V. Multicortical patterns of phase modulation in gamma EEG. International Journal of Bifur- 1977 Blumenthal, A. L. ( ). The process of cognition. cation and Chaos in Applied Sciences and Engi- Englewood Cliffs, NJ: Prentice-Hall. neering, 13(10), 2867–2887. 2003 Cooney, J. W., & Gazzaniga, M. S. ( ). Neu- Gabor, D. (1946, November). Theory of commu- rological disorders and the structure of human nication. Journal of the IEE (London), 93(26), consciousness. Trends in Cognitive Sciences, 7 429–457. (4), 161–165. Hayes-Roth, B. (1984). A blackboard model of Damasio, A. (1989). Time-locked multiregional control. Artiﬁcial Intelligence, 16, 1–84. retroactivation: A systems-level proposal for James, W. (1890). The principles of psychology. the neural substrates of recall and recognition. New York: Holt. Cognition, 33, 25–62. John, E. R., Prichep L. S., Kox, W., Valdes-Sosa, 2001 Dehaene, S., & Naccache, L. ( ). Towards a P., Bosch-Bayard, J., Aubert, E., Tom, M., di cognitive neuroscience of consciousness: Basic Michele, F., & Gugino, L. D. (2001). Invari- evidence and a workspace framework. Cogni- ant reversible qeeg effects of anesthetics. Con- 1 37 tion, 79, – . sciousness and Cognition, 10, 165–183. Dehaene, S., Kerszberg, M., & Changeux, Johnson-Laird, P. N. (1988). A computa- J. P. (1998). A neuronal model of a global tional analysis of consciousness. In A. Mar- workspace in effortful cognitive tasks. Proceed- cel & E. Bisiach (Eds.), Consciousness in ings of the National Academy of Sciences USA, contemporary science (pp. 357–368). Oxford: 95, 14529–14534. Clarendon. Dehaene, S., Sargent, C., & Changeux, J. (2003). Kanwisher, N. (2001). Neural events and A neuronal network model linking subjective perceptual awareness. Cognition, 79, 89– reports and objective physiological data dur- 113. ing conscious perception. Proceedings of the Llinas, R., Ribary, U., Contreras, D., & Pedro- 14 National Academy of Science USA, 100( ), arena, C. (1998). The neuronal basis of con- 8520 8525 – . sciousness. Philosophical Transaction of the Dennett, D. (1991). Consciousness explained. Royal Society, London, 353, 1841–1849. Boston: Back Bay Books. Mack, A., & Rock, I. (1998). Inattentional blind- Dennett, D. E. (2001). Are we explaining con- ness. Cambridge, MA: MIT Press. sciousness yet? Cognition, 79, 221–237. Marcel, A. (1993). Slippage in the unity of con- Edelman, G. M., & Tononi, G. (2000). A universe sciousness. CIBA Foundation Symposium. 174, of consciousness. New York: Basic Books. 168–80; discussion, pp. 180–186. P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

cognitive theories of consciousness 205

Mandler, G. A. (1975). Consciousness: Res- Schacter, D. L. (1990). Toward a cognitive neu- pectable, useful and probably necessary. In ropsychology of awareness: Implicit knowl- R. Solso (Ed.), Information processing and edge and anosognosia. Journal of Clinical cognition: The Loyola Symposium. Hillsdale, NJ: and Experimental Neuropsychology, 12(1), 155– Erlbaum. 178. McIntosh, A. R., Rajah, M. N., & Lobaugh, N. Schneider, W., & Pimm-Smith, M. (1997). Con- J. (1999). Interactions of prefrontal cortex in sciousness as a message-aware control mech- relation to awareness in sensory learning. Sci- anism to modulate cognitive processing. In ence, 284, 1531–1533. J. D. Cohen & J. W. Schooler (Eds.), Scien- Milner, B., & Rugg, M. D. (Eds.), (1992). tific approaches to consciousness (pp. 65–80). The neuropsychology of consciousness. London: Mahwah, NJ: Erlbaum. Academic Press. Shallice, T. (1972). The dual functions of con- Newell, A. (1992). SOAR as a unified theory of sciousness. Psychological Review, 79(5), 383– cognition: Issues and explanations. Behavioral 393. and Brain Sciences, 15(3), 464–492. Shallice, T. (1978). The dominant action system: Newman, J., & Baars, B. J. (1993). A neural atten- An information processing approach to con- tional model for access to consciousness: A sciousness. In K. S. Pope & J. L. Singer (Eds.), global workspace perspective. Concepts in Neu- The stream of consciousness: Scientific investiga- roscience, 4(2), 255–290. tions into the flow of experience (pp. 117–157). Norman, D. A., & Shallice, T. (1980). Attention to New York: Plenum. action: Willed and automatic control of behaviour Shallice, T. (1988). Information-processing mod- (CHIP Report No. 99). San Diego: University els of consciousness: Possibilities and problems. of California. In A. J. Marcel & E. Bisiach (Eds.), Conscious- Penfield, W. (1958). The excitable cortex in con- ness in contemporary science (pp. 305–333). scious man. Springfield, IL: Thomas. Oxford: Clarendon Press. Posner, M. (1994). Attention: The mechanisms Shanahan, M., & Baars, B. J. (2005), Applying of consciousness. Proceedings of the National global workspace theory to the frame problem. Academy of Sciences USA, 91, 7398–7403. Cognition, 98, 157–176. Pribram, K. H. (1971). Languages of the brain: Standing, L. (1973). Learning 10,000 pictures. Experimental paradoxes and principles in neu- Quarterly Journal of Experimental Psychology, ropsychology. Englewood Cliffs, NJ: Prentice- 525, 207–222. Hall. Tononi, G., & Edelman, G. (1998). Consciousness Pribram, K., & Meade, S. D. (1999). Conscious and complexity. Science, 282, 1846–1851. awareness: Processing in the synaptodendritic Varela, F., Lachaux, J., Rodriguez, E., & Mar- web. New Ideas in Psychology, 17(205), 214. tinerie, J. (2001). The brainweb: Phase synchro- Reddy, R., & Newell, A. (1974). Knowledge and nization and large-scale integration. Nature its representations in a speech understand- Neuroscience, 2 , 229–239. ing system. In L. W. Gregg (Ed.), Knowledge Zeki, S. (2001). Localization and globalization and cognition (pp. 256–282). Potomac, MD: in conscious vision. Annual Review of Neuro- Erlbaum. science, 24, 57–86. Rees, G. (2001). Seeing is not perceiving. Nature Zeki, S. (2003). The disunity of consciousness. Neuroscience, 4, 678–680. Trends in Cognitive Science, 7(5), 214–218. P1: KAE 0521857430c08 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:12

206 P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

CHAPTER 9 Behavioral, Neuroimaging, and Neuropsychological Approaches to Implicit Perception

Daniel J. Simons, Deborah E. Hannula, David E. Warren, and Steven W. Day

Abstract differences between perceptual mechanisms are interesting regardless of whether or not For well over a century, the idea that they demonstrate the existence of percep- rich, complex perceptual processes can tion without awareness. occur outside the realm of awareness has either intrigued or exasperated researchers. Although popular notions of implicit pro- Introduction cessing largely focus on the practical consequences of implicit perception, the empir- . . . [T]here is now fairly widespread agree- ical literature has addressed more focused, ment that perception can occur even when basic questions: (a) Does perception occur we are unaware that we are perceiving. in the absence of awareness? (b) what types (Merikle & Joordens, 1997a, p. 219) of information are perceived in the absence Unconscious cognition is now solidly estab- of awareness? and (c) what forms of pro- lished in empirical research (Greenwald, cessing occur outside of awareness? This 1992,p.766). chapter discusses recent advances in the study of implicit perception, considering My contention is that most, if not all, claims the ways in which they do and do not for SA/CI [semantic activation without improve on earlier approaches. We contrast conscious identiﬁcation] in dichotic listening, parafoveal vision, and visual masking the conclusions a skeptic and a believer are in reality based on the failure of these might draw from this literature. Our review experimental methods to reveal whether or considers three distinct but related classes not the meaning of the critical stimulus was of evidence: behavioral studies, neuroimag- available to consciousness at the time of pre- ing, and brain-damaged patient case stud- sentation (Holender, 1986,p.3; brackets ies. We conclude by arguing that qualitative added)

207 P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

208 the cambridge handbook of consciousness

For well over a century, the idea that rich, The more subtle, more interesting ques- complex perceptual processes can occur tion is whether the meaning of a stimulus outside the realm of awareness has either in- is processed without awareness. This prob- trigued or exasperated researchers. The no- lem is of fundamental theoretical impor- tion that many of the cognitive process- tance because any evidence of semantic pro- es that occur with awareness might also cessing in the absence of awareness strongly occur without awareness is both exciting supports late-selection models of attention and frightening; it would not only reveal and awareness (Deutsch & Deutsch, 1963). untapped or unnoticed powers of mind but Presumably, implicit processes occur inde- would also raise the specter of undesirable pendent of explicit attentional selection, so mechanisms of mind. If implicit cognitive if the meaning of a stimulus can be perceived processes are rich and powerful, then given implicitly, selective attention is not neces- the right tools, we might be able to exploit sary for semantic processing. Each of these these resources – we might be capable of questions, at its core, asks how implicit per- using far more information than reaches ception is like explicit perception. awareness. Alternatively, implicit processes For more than a century, strong claims might counteract our explicitly held atti- for the existence of complex perceptual pro- tudes, thereby changing our behavior with- cesses in the absence of awareness have out our knowledge (Greenwald & Banaji, been dismissed on methodological grounds. 1995). In one early study, for example, observers This fear has its roots in psychodynamic viewed a card with a letter or digit on it, views of unconscious processing that attri- but their viewing distance was such that bute many psychological problems to un- the character was hard to see – it was conscious conﬂicts and motivations (Freud, reported to be blurry, dim, or not visible at 1966). It manifests itself in the fear that sub- all. Although subjects could not consciously liminal advertising can affect our beliefs report the nature of the stimulus, they accu- against our will (Pratkanas, 1992). These de- rately guessed whether it was a letter or sires and fears drive a large market in sublim- digit, and they could even guess its identity inal self-help tapes as well as public outcry better than chance (Sidis, 1898). This lack about apparent attempts at implicit inﬂu- of a clear conscious percept combined with ence. Yet, evidence for subliminal persua- better performance on an indirect, guess- sion of this sort is scant at best (Greenwald, ing task might provide evidence for implicit Spangenberg, Pratkanis, & Eskenazi, 1991; perception. However, alternative interpre- Pratkanis, Eskenazi, & Greenwald, 1994). tations that require no implicit perception Although popular notions of implicit pro- are equally plausible. For example, observers cessing focus largely on the practical conse- might simply be more conservative when quences of implicit perception, the empir- asked to produce the name of a digit or letter ical literature has addressed more focused, than they would be when making a forced- basic questions: (a) Does perception occur choice decision (see Azzopardi & Cowey, in the absence of awareness? (b) What types 1998, for a similar argument about blind- of information are perceived in the absence sight). This bias alone could account for bet- of awareness? and (c) What forms of pro- ter performance on a forced-choice task even cessing occur outside of awareness? Few if there were no difference in conscious per- researchers question the idea that some perception. Moreover, the forced-choice task ceptual processing occurs outside of aware- might just be a more sensitive measure of ness. For example, we are not usually aware conscious awareness, raising the possibil- of the luminance changes that lead to the ity that the dissociation between the two perception of motion. Rather, we just per- tasks is a dissociation within conscious perceive the motion itself. Some processing of ception rather than between conscious and the luminance boundaries occurs outside of non-conscious perception. Finally, the mea- awareness even if we are aware of the stim- sure of awareness – the ability to recognize ulus itself. the character from a distance – might be P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

behavioral, neuroimaging, and neuropsychological approaches to implicit perception 209

inadequate as an assessment of awareness, but related classes of evidence: behavioral leaving open the possibility that some con- studies, neuroimaging, and brain-damaged scious perception had occurred. patient case studies. This example illustrates some of the weaknesses inherent in many studies of Limits on the Scope of our Chapter implicit perception. Although the behav- Given the availability of many excellent ioral and methodological tools for studying and comprehensive reviews/critiques of the implicit perception became far more sophis- early literature on implicit perception (e.g., ticated toward the end of the 20th cen- Greenwald, 1992; Holender, 1986; Merikle, tury, and despite some claims to the contrary 1992), our chapter focuses primarily on the (e.g., Greenwald, 1992; Merikle, Smilek, & theoretical and methodological innovations Eastwood, 2001), the controversy over the introduced in recent years. Many disciplines mere existence of implicit perception per- include claims about implicit processing, sists (Dulany, 2004). Often, the same data and incorporating all of them in a single are taken by some as convincing support for overview would be impractical. Instead, the existence of implicit perception and by we highlight claims for implicit perceptual others as unpersuasive (see the critique and or semantic processing of discrete stimuli, responses in Holender, 1986). largely overlooking implicit skill learning, In fact, theoretical reviews of the existing artificial grammar learning, or other forms literature often arrive at strikingly different of procedural knowledge that might well conclusions. Whereas Holender (1986) con- be acquired without awareness. Our neglect cludes that most demonstrations of implicit of these areas does not imply any denigra- semantic processing are unconvincing, oth- tion of the evidence for implicit perception ers consider the converging support for they have produced. Although we limit our implicit effects to be overwhelming (e.g., review to the possibility of semantic process- Greenwald, 1992). In part, these divergent ing without awareness and closely related conclusions simply reflect different default questions, we also consider recent arguments assumptions. “Skeptics” assume the absence about how best to study implicit percep- of implicit perception unless definitive evi- tion. Finally, we discuss how qualitative dif- dence supports its presence. “Believers” ferences in the nature of perceptual process- assume the presence of implicit perception ing may be of theoretical significance even given converging evidence, even if none of without a clear demonstration that process- the evidence is strictly definitive. At its core, ing occurs entirely outside of awareness. the debate often devolves into little more than arguments over parsimony or over the criteria used to infer implicit processing. The goal of this chapter is not to resolve Early Evidence for and against this controversy. Nor is it to provide a Implicit Perception thorough review of this century-old debate. Rather, we discuss recent advances in the Claims for and against implicit perception study of implicit perception, considering received extensive empirical attention start- the ways in which they do and do not ing in the late 1950s, with sentiment in the improve on earlier approaches. We also con- field vacillating between acceptance and ske- trast the conclusions a skeptic and a believer pticism. Many early studies used a dichotic might draw from this literature. Since the listening method in which observers attend mid-1980s, claims about implicit perception to a stream of auditory information in one have become more nuanced, focusing less ear and verbally shadow that content while on the mere existence of the phenomenon simultaneously ignoring another stream in and more on the nature of the informa- their other ear (Cherry, 1953; Moray, 1959; tion that might be implicitly perceived and Treisman, 1960, 1964). If the ignored chan- on the mechanisms underlying implicit per- nel is actually unattended and informa- ception. Our review considers three distinct tion from the ignored channel intrudes into P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

210 the cambridge handbook of consciousness

awareness, then the ignored information der, 1986), few current studies use dichotic must have been processed implicitly. With listening to study implicit perception. The this technique, observers occasionally hear dissociation paradigm, however, remains the their own name in an ignored channel dominant approach to studying implicit (Moray, 1959), and they sometimes momen- perception. tarily shift their shadowing to the ignored The modern use of the dissociation channel when the auditory information pre- paradigm in the study of implicit percep- sented to each ear is swapped (Treisman, tion was triggered by a series of experiments 1960). If ignored information is truly unat- in the 1980s in which masked primes were tended, then these findings support a strong shown to influence subsequent processing of form of late selection in which unattended a target stimulus even though observers did information is processed to a semantic level not notice the primes themselves (Marcel, and sometimes intrudes on awareness. Other 1983a, b). This approach is a classic applica- studies using this dichotic listening tech- tion of the dissociation paradigm: Rule out nique found evidence for skin conductance explicit awareness of the prime stimulus and changes to words in the ignored stream show that it still influences performance in that were semantically related to shock- some other way. Importantly, these studies associated words (Corteen & Dunn, 1974; provided evidence not just that something Corteen & Wood, 1972). was perceived but also that its meaning was Of course, the central assumption under- processed as well; the semantic content of lying these conclusions is that an ignored a masked word served as a prime for a sub- auditory stream is entirely unattended. If sequent response to a semantically related participants periodically shift attention to target word (Marcel, 1983b). Many of the the “ignored” channel, then the influence of recent behavioral studies of implicit per- semantic information in the ignored chan- ception use variants of this masked prime nel might occur only with attention. To con- approach. clude that perception of the semantic content of the ignored stream was caused by The Merits and Assumptions implicit processing, the experimenter must of the Dissociation Paradigm show that it did not result from explicit shifts of attention at the time of presenta- The dissociation paradigm is particularly tion. The difficulty of verifying that atten- appealing because it requires no assumption was never directed to the ignored chan- tions about the nature of or mechanisms nel gave meat to skeptics (Holender, 1986). underlying implicit perception. In its purest In fact, this critique can be applied far form, the dissociation paradigm has a single more generally. The vast majority of stud- constraint: Implicit perception can only be ies of implicit perception, including those demonstrated in the absence of explicit per- in the past 20 years, rely on what is com- ception. Superficially, this constraint seems monly known as the dissociation paradigm straightforward. Yet, it amounts to confirm- (Merikle, 1992). To demonstrate the exis- ing the null hypothesis – demonstrating no tence of implicit perception, experimenters effect of explicit perception – leading some must eliminate explicit perception and show to decry its usefulness for the study of that something remains. Applied to dichotic implicit perception (Merikle, 1994). Given listening, the task for experimenters is to that most claims for implicit perception are rule out attention to the ignored stream based on the dissociation paradigm, most cri- and then show that something remains. The tiques of these claims focus on violations failure of the premise, that ignored means of this assumption, often producing evi- unattended in the case of dichotic listen- dence that some contribution from explicit ing, weakens evidence for implicit percep- perception can explain the residual effects tion. Given the fairly convincing critiques of previously attributed to implicit perception evidence based on dichotic listening (Holen- (see Mitroff, Simons, & Franconeri, 2002 for P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

behavioral, neuroimaging, and neuropsychological approaches to implicit perception 211

a similar approach to critiquing evidence in this chapter, we review new behavioral for implicit change detection). For exam- studies that attempt to meet these assump- ple, critiques of dichotic listening studies tions, but we also note that few of them typically focus on the possibility that sub- systematically demonstrate null explicit jects devoted some attention to the ignored sensitivity to the presence of a stimulus. stream (Holender, 1986). Given that the dichotic listening paradigm does not allow Objective vs. Subjective Thresholds – a direct measure of the absence of atten- What Is the Appropriate Measure tion to the ignored channel, it cannot rule of Awareness? out the possibility that explicit factors contributed to perception of ignored material. One recurring controversy in the study of More subtle critiques raise the possibility implicit perception concerns whether the that observers were momentarily aware of threshold for explicit perception should be ignored or unattended material, but rapidly based on an objective or subjective criterion. forgot that they had been aware. If so, then Although the notion of thresholds has fallen explicit awareness could have contributed to into disfavor with the advent and increased any effects of the “unattended” information. use of signal detection theory in percep- This “amnesia” critique has been applied tion (e.g., Green & Swets, 1966; Macmillan, more recently to such phenomena as inat- 1986), it still has intuitive appeal in the study tentional blindness (Wolfe, 1999). of implicit perception. Later in this chap- To meet the assumptions of the disso- ter, we discuss the importance of using sig- ciation paradigm, the measure of explicit nal detection to measure awareness in the perception must be optimally sensitive – it dissociation paradigm. In the interim, the must exhaustively test for explicit influences distinction between objective and subjective on performance (Merikle, 1992). If a maxi- thresholds may still provide a useful rubric mally sensitive measure reveals no evidence for explaining some of the continuing con- of explicit perception, we can be fairly confi- troversy in the literature. dent that explicit factors did not contribute Most studies of implicit perception rely to performance, and any residual effects on a subjective threshold to determine can be attributed to implicit perception. whether or not a stimulus was explic- This criterion was adopted by some of the itly noticed; this approach assumes that more ardent critics of the early literature on observers will report a stimulus if they are implicit perception (Holender, 1986). The aware of it and will not if they are unaware. explicit measure most typically adopted as For example, blindsight patients typically a sensitive measure of explicit awareness is will report no awareness of a static stim- the simple detection of the presence of a ulus presented to their blind field – the stimulus. If subjects cannot detect the pres- stimulus falls below their subjective thresh- ence of a stimulus, but the stimulus still has old. Use of the subjective threshold to rule an effect on performance, then that effect out explicit perception essentially treats the presumably resulted from implicit percep- observers’ reports of their experiences as tion. In essence, this approach served as the the best indicator of whether or not they basis for early work on priming by masked were aware. More often than not, studies stimuli (Marcel, 1983b). If a masked prime using subjective thresholds are interested in cannot be detected but still influences per- performance on each individual trial, and formance, it must have been implicitly per- claims about implicit perception are derived ceived. Note, however, that even a simple from the consequences of a specific stimu- detection task may not exhaustively mea- lus that was not reported. This approach is sure all explicit influences on performance appealing because it treats observers’ reports and residual effects of a stimulus that can- of their own mental states as more legitimate not be detected might still reflect some than the experimenter’s ability to infer the explicit processing (Merikle, 1992). Later observers’ state of awareness. P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

212 the cambridge handbook of consciousness

Objective thresholds are based on the ing that explicit performance was truly ran- idea that observers might fail to report a dom and not somewhat better than would stimulus even if they did have some explicit be expected by chance alone. awareness of its presence. They might adopt The use of an objective threshold can a conservative response bias, responding only lead to a seeming paradox wherein sub- when certain. Or, they might lack the means jects report no conscious awareness of a to express verbally what they saw. Typically, stimulus (i.e., they report guessing) but still objective thresholds are measured across a show better than chance performance; their large set of trials. The threshold is that level performance exceeds the objective thresh- at which a stimulus is not perceivable rather old even though their subjective impres- than simply not perceived. In using this sion is of guessing. Those adopting a sub- approach, experimenters often adopt the jective threshold approach would conclude standard of null explicit sensitivity required that such a finding reflects implicit process- by the dissociation paradigm, assuming that ing. The appeal of relying on the subjec- if a series of trials show that a stimulus is tive threshold is that it accepts what the not explicitly perceivable, then it could not observer reports at face value. If observers have been perceived on any individual trial. report no awareness, then they had no aware- Consequently, any influence of that stimu- ness. However, it also relies on the observer’s lus must be implicit. Unlike the subjective ability to judge probabilities over a series threshold approach, objective thresholds are of trials. Does the subjective report of no based on the idea that observers might fail awareness really mean that they were guess- to report a stimulus not because they failed ing, or does it mean that they thought that to see it, but because they adopted too con- they were guessing? If observers lack preservative a criterion. This approach does not cise access to their probability of a suc- trust an observer’s subjective experience on cessful response, they might report guess- a given trial to be a true indicator of his or ing when in actuality, they were slightly, her actual awareness of the stimulus. but significantly performing better than In a sense, the terms “objective” and “sub- chance. jective” are misnomers. Both approaches rely The primary difference between the on explicitly reported experiences, so both objective threshold approach and the sub- are subjective. Subjective thresholds are jective threshold approach is that objective based on experiences on each trial, whereas thresholds take the responsibility of estimat- objective thresholds are based on cumula- ing the extent of correct responding out of tive experiences across a larger number of the observer’s hands. Rather than relying trials. Thus, when measuring an objective on the observer to estimate when they felt threshold, responses on individual trials do they were guessing, the objective threshold not necessarily indicate the observer’s aware- technique objectively measures when their ness. Observers might respond that they saw actual performance across a series of trials a stimulus, but that response might simply reflected guessing. In both cases, though, be a guess. Similarly, they might report hav- the subjects’ subjective experience on a ing no conscious experience, even if they given trial contributes to the assessment of had some vague inkling that failed to sur- whether or not they were aware of the crit- pass their criterion for responding. Finding ical stimulus. an objective threshold requires manipulat- Differential reliance on objective and sub- ing the stimulus presentation such that judg- jective thresholds underlies much of the ments of stimulus presence are no better controversy in the field. Most critiques than chance over a reasonably large num- of implicit perception simply show that ber of trials. If responding to this sort of performance actually exceeded an objec- explicit task is at chance over a set of trials, tive threshold for awareness. For exam- then presumably any individual trial is based ple, evidence for implicit priming from on a guess. The challenge is in demonstrat- masked stimuli was premised on the idea P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

behavioral, neuroimaging, and neuropsychological approaches to implicit perception 213

that subjects were no better than chance Momen, Drivdahl, & Simons, 2000 for sim- at determining whether or not the prime ilar examples from the change blindness lit- was present – explicit performance did erature). not exceed the objective threshold (Marcel, Unfortunately, the studies are not ideal 1983b). Yet critiques of those studies sug- for demonstrating implicit perception. gest that the thresholds were not ade- Imagine, for instance, that observers in this quately measured and that explicit perfor- study reported not noticing the gorilla, mance might well have exceeded threshold but then showed priming for the word (Holender, 1986). Even studies that do “monkey.” Would that provide evidence attempt to demonstrate that explicit detec- for implicit perception of the gorilla? The tion was no better than chance rarely meet study uses the dissociation paradigm, and the statistical requirements necessary to subjects subjectively report no awareness infer null explicit sensitivity (Macmillan, of a gorilla. This finding suggests that 1986). Many studies, especially those of any priming effects might be implicit. patients, make no attempt to measure an However, observers might have had some objective threshold, but instead rely entirely awareness of the gorilla, or they might have on the observer’s self-assessment of aware- had momentary awareness of some furry ness, much as early behavioral studies did object, even if they failed to report noticing (e.g., Sidis, 1898). Such studies are open to anything unusual. Given that the method the criticism that explicit perception might only allows one critical trial and the “gorilla” well affect performance even when subjects is demonstrably perceivable (i.e., it is above do not consciously report the presence of a the objective threshold), the possibility of stimulus. some residual explicit awareness cannot be As we discuss later in this chapter, this eliminated. issue is only of importance when question- Arguments for implicit perception on the ing whether or not an example of per- basis of such one-trial studies rest on the ception is entirely implicit. Finding a dis- plausibility of the alternative explanations sociation in the types of processing that for the priming effects. As the measure occur above and below a subjective thresh- of explicit awareness becomes less “objec- old would still be of theoretical (and practi- tive” and more reliant on the observer’s self- cal) import even if explicit perception con- assessment, it is more likely to miss some tributed to both types of processing. For aspect of explicit processing. The sufficiency example, in studies of inattentional blind- of the measure of explicit awareness, regard- ness, observers view a single critical trial and less of whether it is considered objective or quite often fail to notice the presence of subjective, rests on the plausibility of the salient but unexpected objects and events possibility that some explicit awareness was (Mack & Rock, 1998; Most et al., 2001; not tapped by the measure. Of course, even Simons, 2000; Simons & Chabris, 1999). if the gorilla exceeded an objective thresh- When counting the total number of times old for awareness, this hypothetical finding one team of basketball players passes a ball would still be interesting because it would and simultaneously ignoring another team of reveal a discrepancy between what people players passing a ball, approximately 50% see and what they can explicitly report. of observers fail to notice a person in a Moreover, their surprise at having missed gorilla suit who walks through the display the gorilla suggests that their awareness of it (Simons & Chabris, 1999). The interesting likely was limited. Consequently, evidence aspect of these studies is that observers can for inattentional blindness may have impor- fail to notice or consciously detect surpris- tant practical consequences even if some ingly salient unexpected events. Most people residual awareness of the unexpected event expect that they would notice such events, exists. and the fact that they do not report objects Rather than viewing the objective- as unusual as a gorilla is startling (see Levin, subjective difference as a dichotomy, we P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

214 the cambridge handbook of consciousness

prefer to characterize it as a continuum that ciently sensitive to detect any implicit pro- varies along the dimension of the experi- cesses. menter’s confidence in the accuracy of the This exclusivity critique is based on the subjective judgments. With a subjective premise that tasks do not provide a pure judgment on a single trial, the experimenter measure of either implicit or explicit per- should lack confidence in the veracity of the ception. Whether or not this premise is observer’s claim of no explicit awareness. valid, the exclusiveness critique carries less One-trial approaches do not systematically force than the exhaustiveness critique. The eliminate the possibility that the stimulus failure to use an exclusive measure of was perceived and then forgotten, that some explicit awareness is one reason why stud- less-easily-reportable aspect of the stimulus ies using the dissociation paradigm might was consciously perceived, or that the stim- fail to find evidence for implicit perception. ulus was explicitly detected and partially but The lack of exclusivity can only decrease the not completely identified. probability of finding implicit perception, and it should not spuriously produce evidence for implicit perception. Thus, posi- Critiques of the Dissociation Paradigm tive evidence for implicit perception derived Although the dissociation paradigm has from the dissociation paradigm cannot be intuitive appeal, some critics argue that the attributed to the lack of pure measures of exhaustiveness requirement is a fatal short- implicit and explicit processing. If evidence coming – that no task can fully satisfy the for implicit perception using the dissociation exhaustiveness assumption (Merikle, 1992). paradigm is not forthcoming, failed exclu- Even if a task were optimally sensitive to sivity would provide a plausible explana- explicit perception and even if it showed null tion for how implicit perception might occur sensitivity, some other unmeasured aspect but be undetectable via the dissociation of explicit perception could still influence paradigm. performance. Logically, this view is unas- sailable. Even if a task showed null sensitivity for all known explicit influences, it Recent Behavioral Approaches might neglect some as yet unknown and to Studying Implicit Perception unmeasured explicit influence. Practically, however, if a task eliminates all known, Despite concerns about the need for exhaus- plausible explicit influences, then claims of tive measures of awareness, most recent implicit perception might be more parsi- studies of implicit perception have relied monious than defaulting to some unknown heavily on the dissociation logic. The explicit factor. approaches to studying implicit perception A second critique of the dissociation have become somewhat more refined in paradigm rests on the idea that no task mea- their treatment of the problem. In this sec- sures just explicit or just implicit perception tion, we review several relatively new behav- (Reingold & Merikle, 1988). Performance ioral approaches to studying implicit percep- on any task involves a mixture of implicit tion. In some cases, these approaches follow and explicit influences. Consequently, find- the dissociation logic, but with improved ing null sensitivity on an “explicit” task might attempts to exhaustively measure explicit also eliminate implicit perception because influences. Others dismiss the dissociation the task likely measures aspects of both. By paradigm as flawed and propose new app- analogy, a sledgehammer to the head would roaches to measuring implicit perception. eliminate all explicit awareness, but it also For each topic, we consider possible criti- would eliminate most implicit effects on per- cisms of the evidence for implicit percep- formance. Any manipulation that leads to tion, and at the end of the section, we pro- null explicit sensitivity might simply be so vide contrasting conclusions that might be draconian that no measure would be suffi- drawn by a believer and by a skeptic. P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

behavioral, neuroimaging, and neuropsychological approaches to implicit perception 215

Modern Applications of the Dissociation stimulus even when subjects typically were Paradigm unsuccessful at naming the prime. Although this study is consistent with implicit per- Since the mid-1980s, the tools and tech- ception, critics might well raise the objec- niques used to measure implicit perception that the explicit measure (naming) was tion have developed substantially, largely at not an exhaustive test of explicit aware- the goading of skeptics (Holender, 1986). ness. Given that the logic of this task fol- However, straightforward applications of the lows from that of the dissociation paradigm, dissociation logic still dominate studies of unless explicit awareness of the prime is implicit perception, and many (if not most) eliminated, naming improvements could of them neglect to address the standard cri- result from residual explicit awareness. tiques of the dissociation paradigm. This Other studies adopted the repetition shortcoming is particularly true of neu- approach with a more rigorous measure of roimaging work and of studies using patient awareness of the initial stimulus (Watanabe populations, where failures to provide an et al., 2001), although these studies focused adequate exhaustive measure of awareness on perceptual learning rather than priming are commonplace (see Hannula, Simons, & per se. While subjects performed a primary Cohen, 2005 for a detailed discussion of neu- task involving the perception of letters in the roimaging evidence for implicit perception). center of a display, a set of dots behind were In part, the methods in these studies are con- organized into somewhat coherent motion; strained by the need to include imaging mea- most of the dots moved randomly, but a sures or by the nature of the patient’s deficit. subset moved in a coherent direction. Criti- However, behavioral studies of implicit per- cally, a small enough subset of the dots (5%) ception are not limited in these ways, and a was coherent that subjects could not reliably number of new techniques have emerged to discriminate the coherent motion displays provide sensitive and relatively rigorous tests from displays in which all dots moved ran- of the existence of implicit perception. domly. The dots were entirely irrelevant to Some of the simplest approaches are the primary task during the first phase of the based closely on early studies of masked experiment. Then, in a later phase, subjects priming, focusing on the ability to per- attempted to judge the direction of coherent ceive a target as a function of an unseen motion of another set of dot arrays, this time prime (e.g., Bar & Biederman, 1998; Watan- with somewhat more coherence (10%). Sub- abe, Nanez, & Sasaki, 2001). For example, jects were reliably better at determining the one study examined naming accuracy for direction of these dot displays if they moved briefly presented line drawings (Bar & Bie- in the same direction as the previously derman, 1998). For the first time a stimu- viewed displays. Thus, even though subjects lus was presented, subjects were only able were unable to determine that the dots were to name it correctly approximately 15%of moving coherently at all in the first phase of the time. However, when the same stim- the experiment, the frequent repetition of ulus was presented a second time, sub- a particular motion direction led to better jects were far more successful, suggesting performance with a somewhat easier judg- that having seen the stimulus before, even ment task. This indirect test provides evi- without being able to name it, facilitated dence for implicit perception of the coher- subsequent processing. This priming ben- ent motion of dots in the first phase, even efit only occurred when the same object though subjects had no conscious awareness was presented (a different exemplar of the of their motion. This approach is an ele- same category received no priming) and was gant instance of the dissociation paradigm; maximal when the object was presented subjects could not reliably detect the pres- in the same location. These results suggest ence of coherent motion in the prime stim- that implicit processing of the prime stimulus, but the motion coherence still affected ulus led to facilitated naming of the target subsequent judgments. Perceptual learning P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

216 the cambridge handbook of consciousness

approaches like this one have distinct advan- not entirely eliminate explicit perceptibility tages over typical priming experiments in for all subjects, raising some concerns about that awareness of the prime stimulus can the exhaustiveness assumption. be psychophysically eliminated. Other more Although early studies of priming by recent priming studies have attempted to masked stimuli focused on semantic prim- adopt more rigorous measures of awareness ing by words, more recent studies using as well. response compatibility have adopted a host Many of these recent studies exploit of different stimuli and judgment tasks, response compatibility as an indirect, but including left/right discrimination of arrows sensitive measure of perceptual processing. (Eimer, 1999; Eimer & Schlaghecken, 2002; For example, an experiment might measure Klapp & Hinkley, 2002); concrete/abstract response latency to a supraliminal target pre- word discrimination (Damian, 2001); lex- ceded by a supposedly subliminal prime. If ical decision (Brown & Besner, 2002); the target would require a different response words and pictures in animacy judgments than the prime, subjects might be slowed (Dell’Acqua & Grainger, 1999; Klinger, by the presence of the prime. If subjects Burton, & Pitts, 2000); words and non- do not consciously detect the prime, then word stimuli in Stroop interference tasks response compatibility effects likely resulted (Cheesman & Merikle, 1984; Daza, Ortells, from implicit processing of the prime. One & Fox, 2002); words in positive/negative large advantage of this approach over tradi- valence judgments (Abrams & Greenwald, tional semantic priming studies in the disso- 2000; Abrams, Klinger, & Greenwald, 2002); ciation paradigm is that response compat- numerals and number words in relative ibility effects can be positive, negative, or magnitude judgments (Greenwald, Abrams, absent, allowing additional ways to measure Naccache, & Dehaene, 2003; Naccache the effects of an unseen stimulus. & Dehaene, 2001b; Naccache, Blandin, & Given that this approach adopts the dis- Dehaene, 2002;); names in male/female sociation logic, experiments must provide judgment (Greenwald et al., 1996); and direct evidence for the invisibility of the diamonds and rectangles in shape catego- prime. As for studies of masked semantic rization (Jaskowski et al., 2002). Despite priming, most decrease detectability by lim- the varied stimuli and judgment tasks, the iting presentation times and by adding mask- results of these studies are remarkably ing stimuli before and/or after the prime consistent. (e.g., Eimer & Schlaghecken, 2002; Nac- Moreover, all of these approaches to com- cache & Dehaene, 2001b). Others have used patibility effects fall into roughly four types: small differences in contrast to camouflage (1) centrally presented masked primes fol- primes against a background of a similar lowed by a target, (2) centrally presented color (e.g., Jaskowski, van der Lubbe, Schlot- masked primes followed by a target with a terbeck, & Verleger, 2002). Even within limited interval for an allowed response (i.e., the masked presentation approach, how- a “response window”), (3) masked flanker ever, studies vary in terms of how system- tasks, and (4) Stroop tasks. Findings from the atically they manipulate the visibility of the first two approaches are reviewed below. A prime. Some studies use a single stimulus few of these studies were accompanied by duration, contrast level, or type of masking neuroimaging results, some of which are dis- for all subjects (e.g., Naccache & Dehaene, cussed in this section and some of which are 2001b), whereas others adjust the stimulus considered in the section on neuroimaging presentation to account for individual differ- evidence for implicit perception. ences in perceptibility (Greenwald, Draine, & Abrams, 1996). Both approaches can work Masked Priming without provided that neither shows any evidence a Response Window of explicit detection of the prime stimulus. Unfortunately, many of the studies using a The influence of masked primes on response constant prime and mask across subjects do time and accuracy to subsequently presented P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

behavioral, neuroimaging, and neuropsychological approaches to implicit perception 217

target items varies as a function of the ity resulted from delays longer than 350–400 compatibility of the responses mapped to ms (Eimer, 1999; Eimer & Schlaghecken, those items (Dehaene et al., 1998; Eimer, 1998). 1999; Eimer & Schlaghecken, 1998; Koech- The transition from positive to negative lin, Naccache, Block, & Dehaene, 1999; Nac- effects has been characterized more com- cache & Dehaene, 2001b; Neumann & Klotz, pletely using recordings of ERPs. The lateral- 1994). In many cases, target items elicit ized readiness potential (LRP), detected via faster, more accurate responses when the tar- ERP recording, measures the activation from get and prime require the same, compat- motor cortex of the hemisphere opposite the ible response than when they require dif- response hand (Coles, Gratton, & Donchin, ferent or incompatible responses (Dehaene 1988) and provides a direct way to deter- et al., 1998; Koechlin et al., 1999; Nac- mine whether a stimulus leads to activation cache & Dehaene, 2001b; Neumann & Klotz, of motor cortex. On incompatible trials, the 1994). In one task, subjects judged whether prime should elicit transient activation of an Arabic numeral or number word target motor cortex ipsilateral to the responding was greater than or less than 5 (Dehaene hand followed by contralateral motor cor- et al., 1998; Koechlin et al., 1999; Naccache tex activation in response to the target. With & Dehaene, 2001a; Naccache & Dehaene, a compatible prime and target, this ipsi- 2001b). Target numbers were preceded by a lateral activation should be absent. In fact, compatible or incompatible number prime the behavioral compatibility studies often (e.g., if 6 were the target, a prime of 7 incorporated ERP recording and consistently would be compatible and a prime of 4 would found LRPs in response to masked primes be incompatible). In this case, compati- (Dehaene et al., 1998; Eimer, 1999; Eimer ble primes beneﬁted performance regardless & Schlaghecken, 1998). Ipsilateral activa- of whether or not the prime was masked tion was evident shortly after the prime, (Koechlin et al., 1999). Moreover, the com- both for arrow primes and numerical stim- patibility effects persisted even when the uli. Assuming that the masked primes were notation of the target and prime were dif- not consciously perceived, these LRPs pro- ferent (i.e., Arabic numerals primed both vide evidence of processing in the absence of Arabic numerals and number words), sug- awareness. gesting that the priming effect must be The time course of neural activation cor- more abstract than feature-based visual responding to a masked prime might also matching. help explain the paradoxical negative com- Not all studies show a positive effect of patibility effect sometimes observed with compatibility, however. In fact, some studies longer lags between the prime and response show a negative compatibility effect (NCE) (Eimer, 1999; Eimer & Schlaghecken, 1998). in which responses are slower and more The burgeoning neural activity associated error prone for compatible primes (Eimer with a subliminal prime diminishes rapidly & Schlaghecken, 1998)! For example, when when observers do not make an overt a post-masked priming arrow pointed in response. If inhibitory mechanisms, not yet the same direction as a subsequent target fully characterized, are responsible for pre- arrow, subjects were slower and less accurate venting an overt motor response to the than when the prime arrow pointed in the masked prime (Eimer, 1999), then they opposite direction (Eimer & Schlaghecken, might also induce a refractory period dur- 1998). One explanation for these contradic- ing which activation consistent with the tory results appeals to the effects of delays prime is suppressed. Thus, activation in between the prime and the response on com- response to the consistent target would over- patibility effects. In one experiment that lap temporally with this refractory period, systematically manipulated the delay, pos- leading to the paradoxical result of slowed itive compatibility effects were found for responses with compatible primes. Regard- short delays between the prime and the less of whether the prime produces a pos- response, but negative effects of compatibil- itive or negative compatibility effect, these P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

218 the cambridge handbook of consciousness

studies confirm that masked primes activate responses is to maximize any implicit com- corresponding motor cortices. patibility effects based on the premise that Together, the behavioral and ERP evi- such implicit compatibility effects might be dence for compatibility effects suggests that short-lived. As is typical of implicit response unseen primes influence both performance compatibility studies, prime visibility was and neural activity. However, these stud- measured by asking subjects to detect the ies still follow the logic of the dissocia- masked prime stimulus either in a simple tion paradigm, and any claims for implicit detection task or in a discrimination task perception must satisfy the exhaustiveness (e.g., distinguish between a word prime and assumption. Otherwise, differences between a random string of digits). Not surprisingly visible and “subliminal” primes might just given the fixed prime presentation durations, reflect different levels of explicit activation a number of subjects had d levels above 0. rather than a dissociation between explicit However, these studies did not simply look and implicit perception. In most response at performance on the compatibility task compatibility studies, the perceptibility of and then presume that explicit awareness the prime is measured not during the pri- was nil. Rather, a new analytical approach mary task, but in a separate set of tri- was adopted: Regression was used to predict als or separate control experiments (e.g., the level of the compatibility effect when Naccache & Dehaene, 2001b). Although few explicit awareness was absent (d = 0). If the subjects report having seen the primes after intercept of the regression of the compati- the primary task, performance in these sepa- bility effect on explicit sensitivity is greater rate prime perceptibility trials implies some than 0, then the study provides evidence for awareness of the “subliminal” primes. For implicit perception. That is, implicit pro- example, sensitivity using signal detection cessing is revealed when the indirect mea- measure (d ) ranged from 0 for some subjects sure reveals some consequence of the per- to as high as 1.3 for other subjects (Naccache ception of the prime even when explicit & Dehaene, 2001b). Given that d levels of sensitivity is extrapolated to d = 0. This as low as .3 can reflect some reliable sensi- approach revealed significant response com- tivity to the presence of a prime and a d patibility effects for prime durations ranging level of 1 represents fairly good sensitivity, from 17–50 ms when explicit sensitivity was these studies do not adequately eliminate extrapolated to d = 0. explicit awareness of the prime stimuli. Con- This approach was premised on the sequently, claims of compatibility effects assumption that the response window was that are devoid of any explicit awareness are necessary to detect implicit compatibility not entirely supported; the masked primes effects. Another experiment tested the valid- might well have been explicitly detec- ity of this assumption by varying the stim- ted by some of the subjects on some trials. ulus onset asynchrony (SOA) between the prime and the target (Greenwald et al., 1996). If more than 67 ms elapsed between Masked Priming with a Response the prime and target onsets, masking the Window prime eliminated the compatibility effect. In One recent refinement of the masked prim- contrast, unmasked primes produced coming approach involves the use of a speeded patibility effects at a wide range of SOAs. response to maximize the effects of implicit This finding represents an important qual- processing (Draine & Greenwald, 1998; itative difference between visible and sub- Greenwald et al., 1996; Greenwald, Schuh, liminal primes. Moreover, the regression & Klinger, 1995). In this approach, subjects technique and the response-window meth- must make their judgment within a fixed odology are valuable contributions to the temporal window after the presentation of study of implicit perception. the target (e.g., between 383 and 517 ms More importantly, the findings raise some instead of a more typical response latency of important limitations on implicit process- about 600 ms). The goal in forcing speeded ing. Findings from this response-window P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

behavioral, neuroimaging, and neuropsychological approaches to implicit perception 219

technique suggest that implicit effects are be consciously perceived to later serve as an extremely short-lived and are disrupted by effective prime but it must also have been even slight increases to the delay between used as a target such that the word would be the prime and the target. If this form of per- associated with a motor response (Damian, ception proves to be the only reliable way 2001). In the example above, the word to find evidence for implicit perception, it “smile” without prior exposure to “smut” would undermine more radical claims about and “bile” did not prime positive or negative the pervasiveness of implicit processes, espe- words (Abrams & Greenwald, 2000; Green- cially implicit persuasion. Most hypothe- wald et al., 2003). This claim directly con- sized processes of implicit persuasion would tradicts other evidence of implicit semantic require a much longer delay between the processing (Dehaene et al., 1998; Koechlin priming stimulus and the changed belief or et al., 1999; Naccache & Dehaene, 2001b) action. These findings also provide an expla- by suggesting that only fragments of words nation for why studies of implicit percep- are processed implicitly and that the asso- tion often fail to replicate – the effects are ciations they prime are developed through ephemeral. conscious experience as part of the exper- Although the response compatibility iment. Yet, evidence for priming of Arabic effects seem to provide evidence for implicit numerals by number words implies priming semantic processing, many of the findings of more abstract representations, and such could be attributed to motor interference studies also showed priming from stimuli rather than to semantic priming. Subjects that had not previously been the target of learn responses to a stimulus, and it is the a judgment (Naccache & Dehaene, 2001b). responses that conflict, not the abstract or Moreover, switching the required response semantic representations of those stimuli. did not eliminate priming, so the effect In the response-window approach, semantic cannot be entirely due to some automated priming effects are difficult to produce, and form of response priming (Abrams et al., most results can be attributed to response 2002). compatibility rather than any more abstract In a recent intriguing paper, the pri- priming (Klinger et al., 2000). In fact, the mary adversaries in the argument over the effects, at least in some cases of word primes, nature of priming in the response-window seem not due to the semantic content of paradigm combined their efforts to deter- the word, but rather to response associa- mine whether the effects were due to more tions formed earlier in the experiment. In than response compatibility (Greenwald one striking example, subjects were asked to et al., 2003). These experiments adopted make positive/negative valence judgments numerical stimuli (Naccache & Dehaene, about words. In the critical trials, words 2001b) in a response-window task. The stim- that previously had been used as targets uli were all two-digit Arabic numerals, and were recombined in a way that changed the the judgment task required subjects to deter- valence and then were used as a prime word. mine whether the target was greater or less For example, the targets “smut” and “bile” than 55. Unfortunately, the use of two- would become the prime word “smile.” digit numbers precluded the assessment of Although the semantic representation of cross-notation priming, which was one of “smile” should lead to a compatibility benefit the strongest arguments for semantic pro- for a positive target, it instead facilitated processing in earlier experiments (Naccache cessing of negative words (Abrams & Green- & Dehaene, 2001b). The experiment repli- wald, 2000; Greenwald et al., 2003)! More- cated the finding of response compatibil- over, unless a word or part of a word had ity effects with stimuli that had not pre- been consciously perceived as a target dur- viously been used as targets, refuting the ing an earlier phase of the experiment, it argument that subjects must have formed produced no priming at all. In fact, other a response association to a stimulus for it evidence not using a response-window tech- to produce priming (Abrams & Greenwald, nique suggests that not only must a word 2000; Damian, 2001). However, the study P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

220 the cambridge handbook of consciousness

also replicated the counter-intuitive finding the regression approach often depend on that prior judgments affect the directionality extrapolation, with relatively few subjects of priming (Abrams & Greenwald, 2000). If (e.g., 25%) performing at chance on the 73 had served as a target, then subsequently explicit detection task and a sizable minor- using 37 as a prime enhanced response ity of subjects (25%) showing substantial times to numbers greater than 55! Taken explicit sensitivity with d levels greater than together, these findings imply that previ- 1 (Greenwald et al., 1996). If most subjects ously unclassified primes can produce com- show greater than chance explicit sensitivity, patibility effects and that they do so based on the extrapolation to zero sensitivity might long-term semantic representations. How- not be appropriate. A skeptic could eas- ever, such representations are overridden ily imagine a non-linearity in the relation- once a prime has been consciously classified, ship between implicit and explicit measures and then its features lead to priming based when explicit performance is just barely on the response association formed during above d = 0. Perhaps there is a qualitative the experiment. difference between minimal sensitivity and fairly good sensitivity. If so, then extrapolat- ing to no sensitivity from fairly good sensi- Interim Conclusions tivity would not allow a clear conclusion in The response-window and regression favor of implicit effects. Of course, this con- approach lend new credibility to the tradi- cern could be remedied with a more system- tional dissociation technique, and they show atic manipulation of prime visibility within exceptional promise as a way to produce rather than across subjects, thereby obviat- consistent evidence for priming by masked ing the need for any extrapolation. Given stimuli. Although some of the findings the trend toward progressively more sophis- using this method are counter-intuitive and ticated analyses and methodologies in this others are contradictory, the basic approach literature, this new approach shows great represents one of the best existing attempts promise as an effective use of the dissoci- to meet the challenges of critics of implicit ation paradigm. perception (e.g., Holender, 1986). The approach is firmly couched in the disso- Alternatives to Dissociation ciation logic, and most experiments make a laudable attempt to eliminate explicit The concerns about exhaustiveness and the sensitivity. The regression approach, in possible role of failed exclusivity in mini- particular, is a clever way to examine mizing evidence for implicit perception have performance in the absence of awareness. spurred a new approach to studying implicit However, the existing literature does leave perception: Concentrate on qualitative or plenty of wiggle room for skeptics unwilling quantitative differences between tasks that to accept the existence of implicit percep- purportedly measure implicit perception to tion. First, because the approach adopts different degrees. Examining differences in the dissociation paradigm, the measures of performance on these tasks as a function explicit sensitivity might fail to measure of an experimental manipulation can reveal explicit sensitivity exhaustively (Merikle & the operation of distinct implicit and explicit Reingold, 1998). processes. Two types of “relative differences” Perhaps of greater concern to those who methodologies have used this logic: (1) the are otherwise willing to adopt the dissoci- relative sensitivity procedure, which looks for ation logic is the nature of the regression greater sensitivity to stimulus presence with approach itself. The approach has been crit- indirect measures than with direct measures icized for making assumptions about the of awareness, and (2) the process dissociation nature of the relationship between the direct procedure, which looks for qualitatively dif- and indirect tasks and measures (Dosher, ferent performance for implicit and explicit 1998). For example, the conclusions from perception. Neither methodology requires P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

behavioral, neuroimaging, and neuropsychological approaches to implicit perception 221

process-pure measures of implicit or explicit judgment. Therefore, if indirect measures perception. Rather, both assume that all reveal better performance than direct mea- tasks have implicit and explicit components. sures, implicit processes must have influ- Carefully designed experiments can pull enced performance. apart the underlying processes, revealing dif- One critical component of this paradigm ferences between the implicit and explicit is that the two tasks must be equated in processing in a given task. Both approaches most respects. Unless the visual displays are also assume that implicit and explicit pro- equivalent and the task requirements com- cesses underlie functionally different types parable, any performance differences could of behavior; explicit processes underlie be caused by the differences between the intentional actions, whereas implicit pro- displays or the task demands and require- cesses govern automatic (non-intentional) ments. Proponents of this approach rightly behaviors. A behavior may result from a con- take pains to make sure that the only differ- scious, deliberate decision or from an auto- ence between the direct and indirect tasks matic predisposition, or a combination of are in the instructions (a similar approach the two. The relative differences method- has been adopted in the study of implicit ologies attempt to show that such automatic memory; see Schacter, 1987). and deliberate processes can lead to qualita- Note that this criterion – equivalency tively different performance. across direct and indirect measures – is not often met in studies of implicit perception. Many experiments use entirely distinct indi- Relative Sensitivity: An Alternative rect and direct measures, making compara- to Dissociation bility more difficult. When observers report The goal of the relative sensitivity proce- no awareness of a stimulus on a direct mea- dure is to reveal implicit processes by show- sure, indirect measures such as eye move- ing instances in which indirect measures are ments, patterns of neural activation, skin- more sensitive than comparable direct mea- conductance changes, or ERPs might reveal sures in making a given discrimination. This sensitivity to the presence of a stimulus. approach assumes that performance of any Although these sorts of indirect measure- task involves both implicit and explicit con- ments certainly provide important insight tributions, neither of which can be measured into the nature of the processing of the exclusively by any task. Direct tasks measure stimulus, they do not provide conclusive performance when subjects are instructed to evidence for processing in the absence of use their percept of a critical stimulus to awareness. They might only reveal greater make a judgment or discrimination. Indirect sensitivity of the measure itself; using such tasks involve an ostensibly unrelated behav- measures to provide corroborating evidence ior or judgment that nevertheless can be for qualitative differences in implicit and influenced by perception of a critical stim- explicit processing may prove more fruit- ulus. Although the direct task might not ful (see the neuroimaging evidence section exclusively measure explicit contributions, below). For the inference of implicit process- on its face it is demonstrably more explicit ing to follow from the relative sensitivity of than the indirect task. Any decision-making direct and indirect measures, however, the process that relies on conscious awareness measures must be comparable. of the critical stimulus should lead to bet- In one of the first experiments to adopt ter performance on a direct measure than on the relative sensitivity approach for the an indirect measure because subjects should study of implicit perception (Kunst-Wilson optimally rely on their conscious percept. In & Zajonc, 1980), subjects viewed a series contrast, indirect measures do not require of briefly presented pictures of geomet- conscious perception of the critical stimulus, ric shapes. Then, they either performed an so subjects are unlikely to rely on conscious old/new recognition task (the direct task) processing of that stimulus in making their or they picked which of two shapes they P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

222 the cambridge handbook of consciousness

preferred (the indirect task). When perform- words and were asked to read the cued one. ing the direct task, subjects performed no Then, in the test phase, they viewed indi- better than chance at discriminating pre- vidual words against a noise background and viously viewed from novel shapes. In con- either judged whether it was old or new (a trast, when performing the indirect pref- direct recognition task) or judged whether erence judgment task, they preferred the it was presented in high or low contrast (an previously studied shape over a novel shape indirect measure). Performance on the con- at rates significantly above chance levels. In trast judgment task revealed greater sensi- other words, a direct measure of conscious tivity to the presence of a word in the study recognition showed less sensitivity to the phase than did the direct recognition task (at presence of a representation than an indi- least for the first block of trials). Presumably, rect measure of preference. This experiment the prior presentation reduced the process- meets the standards necessary for inferring ing demands, leading to a subjective impres- implicit processing in the relative sensitiv- sion that the words were easier to see against ity approach: (a) the experimental environ- a noisy background even if the words were ment was constant across tasks, with only not recognized. Once again, the study used the task instructions changing across con- comparable stimuli in the direct and indirect ditions, and (b) performance on the indi- tasks and found greater performance for the rect task exceeded that on the direct task. indirect task, suggesting implicit processing. By the logic of the relative sensitivity approach, the direct task represents a puta- Problems with Relative Sensitivity tively better measure of conscious aware- as an Approach ness, so the relatively increased sensitivity of the indirect task must have resulted from Although this approach is touted as an alter- implicit processes. native to the classic dissociation paradigm, One possible concern about this con- any positive evidence for implicit percep- clusion derives from the use of separate tion is subject to many of the same assump- study and test phases rather than testing tions. Positive evidence for implicit percep- performance at the time of presentation; tion requires some task to have a greater subjects may have perceived and forgot- implicit contribution than explicit contribu- ten the consciously experienced shape even tion. Otherwise, performance on the indi- if a vague, explicitly generated preference rect task could not exceed that on any persisted longer and affected performance task with a greater explicit component. If on the indirect task. Furthermore, subjects some task has a greater implicit than explicit might have been less motivated to make the component, then it should also be possi- more difficult intentional recognition judg- ble to make the task sufficiently difficult ments than the simpler preference judg- that the explicit component would be elim- ment, so they were more likely to select inated, leaving only the residual implicit responses randomly. If so, then responding in component. That is, the “indirect > direct” both cases might reflect access to an explicit approach is a superset of the standard dis- representation, with the relative increase in sociation paradigm that does not require sensitivity for the indirect task resulting not the elimination of an explicit component. from implicit processing but from a dif- Yet, any case in which the indirect > direct ferential effect of motivation (cf. Visser & approach reveals implicit perception would Merikle, 1999). also support the possibility that the disso- One representative experiment pitted a ciation paradigm could reveal implicit per- recognition task (direct) against a perceptual ception. In essence, this approach amounts contrast judgment (indirect) in which sub- to a more liberal variant of the dissocia- jects judged the contrast of a word against tion paradigm in which explicit processing the background (Merikle & Reingold, 1991). need not be eliminated. However, as critics In a study phase, subjects viewed pairs of of early work on implicit perception have P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

behavioral, neuroimaging, and neuropsychological approaches to implicit perception 223

noted, whenever a stimulus is consciously pretation of qualitative differences is often perceptible, explicit factors may contami- muddied by the challenge of determin- nate estimates of implicit processing. ing whether differences in performance A more general concern about this are qualitative rather than quantitative. An paradigm is that it assumes a unitary explicit effect that initially appears to reflect a qual- contribution and a unitary implicit contri- itative difference might simply be a differ- bution. In arguing that a direct measure ence along a non-linear dimension. involves a greater explicit contribution than More importantly, though, qualitative an indirect measure, the assumption is that differences in performance can occur even the explicit contributions to each task are when subjects are aware of the stimulus of the same sort. If more than one sort of (Holender, 1986). That is, qualitative dif- explicit contribution exists, then a “direct” ferences are possible within explicit per- task might exceed an “indirect” task on some ception, so the existence of a qualitative forms of explicit contribution but not others. difference in performance alone does not Unless the direct task exceeds the indirect unequivocally demonstrate implicit percep- task on all explicit contributions, the logic tion. Rather, the qualitative difference must underlying the paradigm fails. Just as the dis- be accompanied by an exhaustive measure sociation paradigm suffers from the prob- of explicit awareness. Consequently, qual- lem of exhaustively eliminating all possible itative differences can provide converging explicit contributions, the indirect > direct evidence for the existence of implicit per- approach requires that the two tasks mea- ception, but they are not definitive in and sure the same explicit component and only of themselves (Holender, 1986). Perhaps the that explicit component. Consequently, for best example of the use of qualitative dif- the logic of the paradigm to hold, the exper- ferences in studies of implicit perception imenter must exhaustively eliminate any comes from the use of the process disso- extraneous explicit contributions to the indi- ciation paradigm (otherwise known as the rect task that might explain superior per- “exclusion” paradigm). formance on the indirect task. Given that this exhaustiveness assumption applies to Process Dissociation the relative sensitivity approach, its advantage over the standard dissociation paradigm In the process dissociation technique, is somewhat unclear. implicit and explicit performance are put in opposition (Jacoby, 1991). As in the relative sensitivity approach, direct and indirect Qualitative Differences measures are thought to rely differentially One criterion often used to infer the exis- on explicit and implicit processing. In this tence of implicit perception relies on differ- approach, intentional actions are assumed to ences in the patterns of performance derived be under explicit control, whereas automatic from implicit and explicit processes. When responses are thought to reflect implicit pro- the pattern of performance diverges from cessing. Presumably, people will use con- what would be expected with explicit per- sciously available information to guide their ception, then the processes leading to this intentional actions. In contrast, informa- qualitative difference might well be implicit. tion available only to implicit processes Qualitative differences in performance for will be less subject to intentional con- implicit and explicit tasks or measures often trol. Consequently, when subjects produce provide an intuitive way to infer the exis- responses that differ from those associated tence of implicit perception. The negative with intentional actions, they may have been compatibility effects described earlier pro- influenced by non-conscious processes. The vide one illustration of the importance of critical difference between the process disso- such differences for inferring implicit pro- ciation procedure and the relative sensitivity cessing (Eimer, 1999). However, the inter- procedure is that the task instructions and P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

224 the cambridge handbook of consciousness

goals are constant. Subjects always perform tions, subjects completed the stems with the same task. Rather than manipulating the the “studied” word more often than the task across conditions, the perceptibility of baseline condition. Even when they were the critical stimulus itself is varied so that unable to use their memory for the word to some responses are consistent with aware- guide their intentional actions (i.e., choose ness of the stimulus and others are not. another word), the briefly presented word Subjects are instructed to respond one still received enough processing to increase way if they are aware of a stimulus, but its availability in the stem completion task. implicit or indirect influences lead them to A similar pattern emerges when atten- respond the opposite way by default. For tional focus rather than presentation dura- example, in the original instantiation of this tion is manipulated (see Merikle et al., 2001 approach in the memory literature, subjects for an overview). In one study, subjects studied a list of words and then were asked viewed a briefly presented cross and judged to complete word fragments with words that which of its two lines was longer. During a had not been on the studied list (Jacoby, subset of trials, a word was presented briefly 1991). Presumably, if they remembered the along with the cross (see Mack & Rock, 1998 studied word, they would successfully avoid for the origins of this method). Depend- it in the fragment task. If they did not explic- ing on the condition, subjects were asked to itly remember the word, they might auto- focus attention either on the cross judgment matically or implicitly be more likely to or on the words. In both conditions, subjects complete a fragment with a studied word subsequently attempted to complete a word than a non-studied word. Implicit influences stem with a word that had not been pre- should increase the likelihood of complet- sented (this study was described by Merikle ing fragments with studied words, whereas et al., 2001). Those subjects who focused explicit influences should decrease the like- on the words performed well, rarely using lihood of completing fragments with stud- the presented words to complete the stem. ied words. The same logic can be applied to In contrast, those who focused attention on implicit and explicit perception: If subjects the cross judgment completed the stem with explicitly detect the presence of a word, they presented words more often than would be should avoid using it to complete a word expected based on a previously determined fragment. However, if they do not detect it baseline (see also Mack & Rock, 1998). and it still influences performance implic- When the words were the focus of attention, itly, they should be more likely to complete they presumably were available to aware- a fragment with a studied word than a non- ness, and subjects could use that informa- studied word. tion to exclude them in the stem completion Studies using this procedure have been task. However, when subjects focused atten- taken to support the existence of implicit tion on the cross judgment, they were less perception. For example, one study var- aware of the words, but automatic process- ied the presentation time for words. Imme- ing of the words biased them to use the pre- diately after viewing each word, subjects sented words in the stem completion task. were given a word stem completion task in A variety of exclusion tasks have been which they were asked to complete the stem used to study implicit perception. For exam- with a word other than the one that had ple, subjects show differential effects of been presented (Debner & Jacoby, 1994). interference in a variant of the Stroop task With long presentation durations, subjects when aware and unaware of a stimulus. Typ- were aware of the words and successfully ically, when color patches are incongruent avoided completing stems with the “studied” with a preceding word (e.g., a green patch words relative to the baseline performance preceded by the word “red”), subjects are of subjects who had never been shown the slower to identify the color of the patch than word. In contrast, with shorter presenta- if the two matched. However, if mismatches P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

behavioral, neuroimaging, and neuropsychological approaches to implicit perception 225

occur on a large proportion of trials (80%), Problems with Process Dissociation as a subjects use this information to perform Measure of Implicit Perception faster when the word and patch mismatch This approach has promise as a means of than when they match (see Merikle & Joor- studying implicit perception. One concern dens, 1997b for a discussion of these studies). about this approach, however, is that it When words were presented long enough to might be subject to biases and motivational be consciously detected, subjects used their factors that affect the criterion that sub- explicit knowledge to override Stroop inter- jects adopt. If so, estimates of implicit pro- ference. In contrast, briefly presented words cessing might be inflated (Visser & Merikle, were not consciously detected, and subjects 1999). Any case in which the subject’s cri- were significantly slowed when a word-color terion is differentially affected by exclu- patch mismatch occurred (Merikle & Joor- sion and inclusion instructions can produce dens, 1997a,b). Stroop interference can be a change in the criterion that could then counteracted if subjects are aware of the influence estimates of unconscious process- word and of the predictiveness of the word, ing. For example, increasing incentives to but if the word is not consciously perceived, exclude studied items led to improved per- subjects cannot override these automatic formance, thereby decreasing estimates of interference effects. unconscious processing (Visser & Merikle, Other exclusion studies have used sim- 1999). More broadly, variations in the degree ilar manipulations of target predictability of confidence or certainty in a represen- in response compatibility paradigms (e.g., tation or a percept can lead to different McCormick, 1997). Subjects were asked to degrees of success on the exclusion task. decide whether an X or O was in the dis- Given that the exclusion task provides the play, and this target item was presented basis for inferring implicit representations, either on the right or left side of the fix- such variations are problematic. A word that ation cross. Before the presentation, a cue is explicitly detected, but with low confi- appeared on the left or right side of the dis- dence, might lead to a failure to exclude that play. On approximately 80% of trials, the item on a stem completion task even though cue was on the side opposite where the tar- there was an explicit contribution to percep- get would appear. Thus, the cue predicted tion. In terms of signal detection, if subjects that the target would be on the opposite side were conservatively biased when reporting of the display. When the cue was presented explicit detection, estimates of implicit perfor long enough to be consciously detected, ception would be inflated. Thus, as in the subjects responded more rapidly to targets dissociation paradigm, the explicit task must on the side opposite the cue. In contrast, demonstrably eliminate all explicit detec- when the cue was presented too briefly to tion and must not be subject to conservative be consciously detected, subjects were faster response biases for this paradigm to provide to respond when the cue and target were on a clear estimate of implicit perception. the same side of the display (McCormick, 1997). Presumably, the cue automatically A Believer’s Interpretation attracts attention, and only with awareness can subjects override this automatic shift of The past 15 years have seen tremendous attention. Without awareness, the cue auto- improvements in the behavioral methods matically draws attention, leading to better used to study implicit perception. More performance when the target appears at the importantly, many of the early critiques cued location. Although this finding does not of the implicit perception literature have involve semantic processing without aware- been addressed. Most studies using the dis- ness, it does suggest that attention shifts can sociation paradigm now use signal detec- be induced without awareness of the induction theory to determine the explicit per- ing stimulus. ceptibility of the prime stimulus, thereby P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

226 the cambridge handbook of consciousness

providing a more convincing demonstration dissociation paradigm or other relative sen- that priming results from implicit process- sitivity approaches is a promising avenue for ing rather than from explicit contamina- the exploration of implicit perception. How- tion. The recently introduced technique of ever, none of these approaches or studies regressing performance on an indirect mea- provides airtight evidence for implicit per- sure (e.g., a response compatibility effect) ception, and all are subject to fairly plausible on performance on an explicit detection task alternative explanations that rely solely on provides a more nuanced approach to the explicit mechanisms. For example, in stud- dissociation technique. Even when perfor- ies using the regression approach, the direct mance on the explicit task is extrapolated to measure often reveals sensitivity to the pres- null sensitivity, performance on some indi- ence of the prime stimulus at levels far above rect measures is still better than chance. The d = 0 (Naccache & Dehaene, 2001b); the use of response compatibility allows an indi- prime is readily visible to some subjects. rect measure that can, under the right cir- Consequently, the inference for implicit per- cumstances, reveal implicit semantic pro- ception relies on extrapolation of perfor- cessing. For example, priming persists even mance to d = 0 from a number of sub- when the format of a number (text vs. Ara- jects who show positive sensitivity to the bic numeral) changes from prime to test. stimulus. This extrapolation is potentially The combination of the regression technique hazardous, particularly if the distribution and response compatibility paradigms proof subjects is not centered on d of 0.If vides a powerful new tool to study implicit the relationship between explicit percep- perception, one that has produced consis- tion and the indirect measure is non-linear, tent and replicable evidence for implicit per- the extrapolation may be invalid (Dosher, ception. Finally, work using process disso- 1998). Moreover, the presence of a pos- ciation and relative sensitivity approaches itive indirect effect might require only a reveals evidence for qualitative differences minimal amount of explicit sensitivity. No between implicit and explicit processing. published studies have examined the effect These qualitative differences suggest that of varying explicit sensitivity systematically different mechanisms underlie implicit and (within subjects) on the magnitude of the explicit perception, thereby providing fur- indirect response compatibility effect. Any ther evidence for the existence of implicit application of the dissociation paradigm, perception. In sum, evidence from a wide including the regression approach, depends variety of tasks and measures provides sup- critically on demonstrating null sensitivity port for implicit perception and even for to the presence of the critical stimulus. semantic processing in the absence of aware- None of the studies to date have done so ness. Given the wide variety of tools used adequately. in the study of implicit perception, the con- Evidence from the process dissociation verging evidence for the existence of implicit paradigm suggests a qualitative difference perception is overwhelming. between implicit and explicit perception, something that would be more difﬁcult to explain via explicit contamination. Most A Skeptic’s Interpretation studies of implicit perception simply reveal The tools and techniques used to study “implicit” effects that are weaker versions implicit perception have improved immen- of what would be expected with explicit sely over the past 20 years. Many studies processing. The process dissociation pro- have adopted signal detection theory as a cedure, in contrast, suggests that implicit way to verify the absence of explicit percep- and explicit mechanisms differ. However, as tion, thereby making evidence from the dis- accurately noted in critiques of the implicit sociation paradigm less subject to the stan- perception literature, qualitative differen- dard criticisms. Moreover, seeking evidence ces alone are insufﬁcient to claim eviden- of qualitative differences using the process ce for implicit perception. The qualitative P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

behavioral, neuroimaging, and neuropsychological approaches to implicit perception 227

difference could simply be a dissociation ferences in the pattern of activation for between two forms of explicit perception explicit and implicit perception might reveal rather than between implicit and explicit additional qualitative differences between perception. Moreover, inferences of implicit these forms of processing even if behavioral perception depend on the extent to which responses show no difference; neuroimaging the explicit, intentional response fully mea- might simply provide a more sensitive measures all of the explicit processing. The sure. Finally, the known functions of vari- only studies to address this question suggest ous brain regions can be mapped onto the that performance on the explicit task can pattern of activation produced in response be enhanced via motivation manipulations, to seen and unseen stimuli, allowing yet thereby decreasing the evidence for implicit another way to determine the richness of perception (Visser & Merikle, 1999). implicit percepts. In sum, the new tools introduced to Although such approaches have great study implicit perception may be promising, promise as a new tool for the study of but the evidence for implicit perception is implicit perception, in many respects the not yet convincing. Moreover, the implicit existing research on the neural bases of effects that have been reported are small implicit perception falls prey to the same and tend to vary with the extent of demon- critiques leveled at the behavioral research. strated explicit awareness, hinting that the Perhaps more importantly, as our review sug- “implicit” effects might well be driven by gests, the neural activity elicited by implicit residual explicit processing. For a study using perception often is similar to that corre- the dissociation paradigm to make a strong sponding to overt perception, just dimin- claim for implicit perception, no subject ished in amplitude. In the absence of qualita- should show explicit sensitivity to the vis- tive differences in the pattern of activation, ibility of the critical stimulus; no study to such diminished effects might well result date has met this strict criterion. Converg- from low-level overt perception. In such ing solid evidence from a variety of tech- cases, the same standards and criteria applied niques can provide powerful support for a to the use of the dissociation paradigm claim of implicit perception, but the conver- in behavioral research must be applied to gence of weak and controvertible evidence the neuroimaging methods (see Hannula for implicit perception does not merit strong et al., 2005 for a detailed treatment of support for the claim. If all of the evidence these issues). can be explained by plausible explicit con- A wide array of neuroimaging tools, founds, then there is no need to infer the most notably functional magnetic resonance existence of a separate mechanism or set of imaging (fMRI) and event-related potentials mechanisms. (ERPs), have been adapted to the study of implicit perception. Most often, these investigations draw on existing knowledge of the Evidence for Implicit Perception – functional brain regions likely to be involved Neuroimaging Data in overt perception of a class of stimuli (e.g., emotional faces, words, etc.), and then try Neuroimaging approaches provide sev- to determine whether those same regions are eral distinct advantages over behavioral active even when observers report no aware- approaches in the study of implicit per- ness of the stimuli. Neuroimaging studies of ception. First, the effects of a subliminal implicit perception typically rely on several stimulus can be assessed without an overt different types of processing and stimulus response; neuroimaging techniques provide classes, and for the sake of organizing this an additional dependent measure of the con- rapidly expanding ﬁeld, we consider three sequences of perception, one that may allow types of evidence for implicit perception: dissociations that would be impossible with implicit perception of faces, implicit per- strictly behavioral measures. Moreover, dif- ception of words and numbers, and ERPs P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

228 the cambridge handbook of consciousness

in priming studies. Within the neuroimag- One study required a patient to respond ing literature, most inferences about implicit differently to a stimulus presented solely on perception depend critically on the pattern the left, solely on the right, or simultane- of neural localization or the magnitude of ously on the left and right (Rees et al., 2000). activation resulting from explicitly detected Given that the patient had right inferior and implicitly perceived stimuli. parietal lobe damage, extinction would be revealed by incorrect “right-side” responses when a stimulus was presented on both Implicit Perception of Faces the left and the right simultaneously. By Face processing represents one of the more comparing fMRI data corresponding to promising avenues for the study of implicit correct “right-side” responses and incor- perception because the neural regions acti- rect “right-side” responses on extinction vated in response to faces are fairly well trials, residual neural activity associated with described in the neuroimaging literature. A extinction could be revealed. Extinguished slew of recent neuroimaging studies of face stimuli activated striate and early extras- perception reveal an area in the fusiform triate cortex in the damaged right hemi- gyrus that responds relatively more to faces sphere – a pattern of activation no differ- than to other stimuli (the Fusiform Face ent from that elicited by left-side stimuli Area, or FFA; Kanwisher, McDermott, & that were consciously perceived. This acti- Chun, 1997; McCarthy, Puce, Gore, & Alli- vation of early visual cortex occurred regard- son, 1997). Is this area active even when less of whether the patient was aware of observers are unaware of the presence of a the stimulus, suggesting that these areas face stimulus? Also, fearful faces are associ- are not sufficient for conscious awareness. ated with activation of the amygdala (Breiter More importantly, a region of interest analy- et al., 1996; Morris et al., 1996). Do fear- sis revealed low-threshold, category-specific ful faces lead to amygdala activation even activation in the right FFA in association when they are not consciously perceived? with extinguished face stimuli, suggesting Finally, recent neuroimaging studies using that the extinguished face was processed by the phenomenon of binocular rivalry have the same regions as for consciously perceived explored the areas that are activated by stim- faces (for reviews of evidence for preserved uli when they are consciously perceived and activation in response to unreported stim- when rivalry removes them from awareness. uli, see Driver & Vuilleumier, 2001; Driver, Recent neuroimaging studies of visual Vuilleumier, Eimer, & Rees, 2001). extinction patients have explored whether This basic pattern was replicated in a sim- an extinguished face leads to activation in ilar experiment using both fMRI and ERPs the FFA (Rees et al., 2000; Vuilleumier et (Vuilleumier et al., 2001). An extinction al., 2001). Unilateral brain lesions, particu- patient with right-lateralized posterior infe- larly those located in the right posterior inferior parietal damage indicated on each trial rior parietal lobe, are associated with spa- whether or not a face was presented. Stimuli tial neglect of the contralesional visual field. (i.e., schematic faces and shapes) were pre- Many neglect patients exhibit visual extinc- sented unilaterally in the right or left hemi- tion, accurately detecting isolated stimuli field or bilaterally. Again, extinguished faces presented in either visual field, but failing activated right striate cortex as well as an to identify a contralesional stimulus when area of inferior temporal cortex just lateral items are presented simultaneously in both to the FFA, although the level of activation visual fields. Behavioral research (discussed was much reduced relative to that for visi- in later sections of this chapter) provides evi- ble face stimuli. Furthermore, ERPs revealed dence for residual processing of extinguished a right-lateralized negativity over posterior stimuli, perhaps due to intact striate and temporal regions approximately 170–180 ms extrastriate cortex along with ventral infer- after a face was presented in the left hemi- otemporal areas that process object identity. field. This N170, a component known to be P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

behavioral, neuroimaging, and neuropsychological approaches to implicit perception 229

face-selective, was evident regardless of (although activation was more pronounced whether the face was perceptible or not. for fearful faces). This pattern of results is Interestingly, this experiment varied the consistent with the notion that the amyg- duration of the bilateral presentations in dala is selectively recruited when subliminal order to vary whether or not extinction fear stimuli are presented (for additional evi- occurred. Awareness of the left visual field dence of early affective word processing in stimulus evoked activation of striate cortex the absence of awareness see Bernat, Bunce, and fusiform gyrus coupled with increased & Shevrin, 2001a), but as for similar behav- activation of a network of frontal and pari- ioral results, such dissociations must be inter- etal brain regions, reflecting the sorts of preted with caution because of method- long-range associations or widespread acti- ological shortcomings in the assessment of vation thought to accompany consciousness awareness. For example, awareness was not (Baars, 1988). Thus, differences in activation measured directly on each trial – doing so strength and functional connectivity distin- might change the subject’s strategy from one guish conscious from unconscious percep- of passive viewing to active search (Whalen tion. et al., 1998). Evidence from patients with bilateral In another study, relative to neutral faces, amygdala damage (e.g., Adolphs, Tranel, fearful faces were associated with significant Damasio, & Damasio, 1995) and neuroimag- activation of the left amygdala, left fusiform, ing of intact individuals (Breiter et al., 1996; lateral orbitofrontal, and right intraparietal Morris et al., 1996) support a role for the cortex (Vuilleumier et al., 2002). Activa- amygdala in processing fear-related stimuli, tion of fusiform gyrus in response to extin- such as fearful faces. At least one theory guished faces was much reduced relative to suggests that a direct short-latency path- activation for visible faces though, and the way between the thalamus and the amygdala activation evident in association with extin- might underlie the processing of emotional guished stimuli may be a consequence of stimuli even in the absence of awareness (Le feedback from the amygdala. Together, these Doux, 1996). In one fMRI study (Whalen findings suggest that emotional stimuli can et al., 1998), fearful and happy faces were receive substantial processing even if they presented for 33 ms followed immediately fail to reach awareness. Emotional stimuli by a neutral-face mask. Based on previ- are among the most promising approaches ous behavioral studies, the 33-ms masked to the study of implicit processing, pre- presentation was assumed to be below the cisely because of the hypothesized existence threshold for awareness. Post-study ques- of a direct, perhaps more primitive neu- tioning found that eight of ten subjects ral pathway that bypasses higher cognitive denied having seen emotional faces and did areas. not select these faces as having been in the These studies provide interesting, sug- stimulus set. Under these conditions, the gestive support for the hypothesized short- unnoticed fearful faces did elicit a relatively latency pathway originating in the thalamus circumscribed increase in amygdala activa- (LeDoux, 1996). Such a pathway might rea- tion relative to masked happy faces and a sonably allow for processing even in the fixation baseline. This amygdala activation absence of more complex cognitive pro- was attenuated with repeated exposure to cesses, and by inference without awareness. masked fearful faces, a finding consistently More importantly, amygdala activation was observed with visible faces as well. Fur- not significantly modulated by awareness ther, increased activation in response to both (Vuilleumier et al., 2002), suggesting that masked fearful and happy faces extended processing of extinguished stimuli extends into the adjacent sublenticular substantia beyond early visual processing areas and innominata of the basal forebrain, a region that activation need not be less robust in thought to be involved in more general the absence of conscious detection. Similar processing of emotional stimuli and arousal approaches have been taken in the study P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

230 the cambridge handbook of consciousness

of implicit processing of unnoticed, emo- ceived a face or a house when both were tionally arousing stimuli (see Lane & Nadel, present (in the rivalrous stimulus) was no 2000 for an overview of work on the cogni- different than when the face or house was tive neuroscience of emotion). presented alone, suggesting that the com- Another neuroimaging-based approach petitive neural interactions responsible for to studying processing in the absence of rivalry are largely resolved before conscious awareness relies on the phenomenon of perception occurs. binocular rivalry. When two patterns are This finding might suggest that activation presented simultaneously, one to each eye, in the FFA or the PPA produces visual aware- the contents of conscious awareness spon- ness of the presence of a face or a house. taneously alternate between one monocu- However, the FFA also is active when faces lar percept and the other over time. The are not consciously reported (Rees et al., visual percepts compete for awareness such 2000), suggesting that reliable FFA activa- that only one image is consciously per- tion is not sufficient for conscious percep- ceived, and the other is suppressed (Lev- tion of a face. This discrepancy might result elt, 1965; Wheatstone, 1838). The oscilla- from different degrees of activation, though. tion of perceptual awareness between two If neural activity is graded with respect to simultaneously presented stimuli provides the level of perceptual awareness (i.e., low- a useful tool to identify the neural corre- level activity reflects low-level awareness) lates of conscious awareness (for reviews, see or if activity must surpass some threshold Rees, Kreiman, & Koch, 2002; Tong, 2001, before conscious awareness occurs, then it 2003). is entirely possible that sufficient activation A growing number of investigations have of the FFA or PPA does correspond to con- been conducted using fMRI to address, in scious awareness of a face or house, respec- particular, the contributions of specific brain tively. Stricter criteria for measuring con- regions to perceptual awareness of rival- scious awareness are needed to determine rous stimuli. One recent study using fMRI whether activation in these specialized pro- (Tong, Nakayama, Vaughan, & Kanwisher, cessing regions is sufficient for conscious per- 1998) presented face and house images sep- ception. arately to each eye and measured the neural activity in two predefined regions of Implicit Processing of Words interest: the FFA, which responds prefer- and Numbers entially to faces (Kanwisher et al., 1997; McCarthy et al., 1997), and a parahippocam- Just as consciously perceived emotional pal region that responds most strongly to stimuli activate the amygdala, read words places and less so to faces (the Parahip- tend to activate a prescribed set of brain pocampal Place Area, or PPA; Epstein & regions more than do other stimuli (e.g., Kanwisher, 1998). During imaging, partici- left-lateralized extrastriate cortex, fusiform pants continuously reported whether they gyrus, and precentral sulcus). Therefore, saw a face or a house, and the pattern of neu- studies of implicit word perception can ral activity extracted from a region of inter- use neuroimaging evidence to determine est analysis was time-locked to these con- whether words activate a similar set of scious perceptual experiences. Interestingly, regions without awareness. Such studies first neural activation corresponded to the con- assess the visibility of the critical words using scious perceptual experience, even though behavioral measures. In one study (Dehaene the stimulus pair was invariant within a et al., 2001), masked words were presented trial; FFA activation increased when partic- such that they were detected only 0.7%of ipants reported perception of a face stimu- the time (a rate slightly higher than the false lus, and PPA activation increased when they alarm rate of 0.2% for trials in which no word reported a house. Critically, the pattern of was presented) and almost never named suc- activation when subjects consciously per- cessfully (see also Rees, 2001). Moreover, P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

behavioral, neuroimaging, and neuropsychological approaches to implicit perception 231

recognition tests after the imaging portion a consequence of the naming task. In sum, of the study revealed no memory for the masking resulted in less robust neural acti- masked words. Of course, subjects might vation, but also in reduced correlated neural adopt a conservative criterion for indicating activity that might contribute to conscious whether or not a word was present if they awareness. knew they would then be asked to name Similar patterns have emerged in neu- it. If so, then the task might not exhaus- roimaging studies of the perception of tively measure conscious awareness, raising numerical stimuli (Naccache & Dehaene, the possibility that the masked words were at 2001a). Neuroimaging and lesion data sug- least temporarily available to consciousness. gest a role for the parietal lobe (and partic- Assuming that low detection rates and ularly the intraparietal sulcus) in the men- failed recognition performance imply the tal representation and understanding of the absence of conscious awareness of the pres- quantity meaning of numbers (for a review, ence of the masked words, and if neu- see Dehaene, Piazza, Pinel, & Cohen, 2003). ral activity is consistent with reading, then Can implicit stimuli lead to similar pat- perception presumably occurred implicitly. terns of activation? A recent paper (Nac- Interestingly, when compared to control cache & Dehaene, 2001a) reanalyzed ear- conditions that mimicked the masking con- lier neuroimaging data (Dehaene et al., ditions of the critical trials but without 1998) and addressed this issue by using any masked words, the unseen stimuli acti- the phenomenon of repetition suppression. vated the previously mentioned set of brain A number of imaging studies have shown regions known to be associated with reading that when a stimulus is repeated, local- (Dehaene et al., 2001). This pattern is con- ized neural activity associated with process- sistent with the idea that the unseen stim- ing of that stimulus or its attributes typ- uli were processed similarly to visible words. ically decreases (Schacter, Alpert, Savage, However, the pattern of neural activity Rauch, & Albert, 1996; Schacter & Buckner, evoked by the masked words in the ventral 1998; Squire et al., 1992). Whole brain visual pathway was less widely distributed analysis of fMRI data revealed two iso- and of smaller magnitude than that obtained lated brain regions with reduced activity with consciously perceived words. The dis- when the target repeated the prime rela- crepancy was increasingly evident from pos- tive to an otherwise categorically congru- terior to anterior brain regions, suggesting ent prime: the left and right intraparietal that visual masking begins to suppress neural sulci (Naccache & Dehaene, 2001a). The activity early in the visual processing stream, priming effect was not inﬂuenced by the rendering later stages of visual process- use of different notations for the prime and ing less likely. Furthermore, visible words target (1 vs. one), suggesting that the intra- elicited neural activity in parietal, prefrontal, parietal sulcus encodes numbers in a more and cingulate cortices, but corresponding abstract format. Assuming that the prime activation was not evident when the words stimuli were not consciously perceived, were not available to conscious awareness. these effects indicate that repetition sup- Finally, increased correlated activity among pression can occur even when observers are the ventral visual stream, parietal, and pre- unaware of the repetition. Presumably, this frontal areas was evident only when the effect reﬂects the fairly extensive processing words were visible. Some of these differ- of an implicitly perceived stimulus. ences might well result from the naming Additionally, ERP studies of the response task used in the study rather than from the compatibility effect reveal covert activation perceptibility of the stimuli. Visible words from an incongruent prime – a lateralized could be named, but the masked words were readiness potential (LRP) on the incorrect not. However, it cannot be determined on side of response – presumably because the the basis of these results whether some of incongruent prime activates the incorrect the activity associated with visible words is motor response (Dehaene et al., 1998). fMRI P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

232 the cambridge handbook of consciousness

data revealed greater overall activation in tute an electrophysiological marker of con- right motor cortex when both the prime scious semantic processes. Yet, other exper- and target were consistent with a left hand iments that account for potential method- response (and vice versa), providing addi- ological shortcomings of this experiment tional evidence for processing of the prime induce modulation of the N400 even when stimulus without awareness (Dehaene et the primes were consciously inaccessible al., 1998). In all of these studies, percep- (e.g., Deacon, Hewitt, Yang, & Nagata, tion of the prime in the absence of aware- 2000; Kiefer, 2002). Moreover, the effects of ness was not limited to sensory mecha- a prime on the N400 are qualitatively differ- nisms alone, but also influenced higher-level ent for visible and masked primes. Masked processing. primes modulate the N400 with a short SOA between the prime and target, but not with a longer SOA. In contrast, for vis- ERPs in Priming Experiments ible primes, the modulation of the N400 The influence of an unseen prime stimu- increases as the SOA increases (Kiefer & lus has been explored by examining gen- Spitzer, 2000). This qualitative difference eral changes in ERPs to a target as a result suggests that implicit and explicit percep- of the presence of a prime. These studies tion of prime stimuli might rely on different measure the influence of an implicit prime processing mechanisms. indirectly, looking for changed neural pro- Taken together, these studies provide sup- cessing of the target rather than activation port for N400 activation in response to an directly in response to the prime stimulus. unseen prime. However, they are subject to Studies of priming by masked stimuli rep- many of the critiques leveled at the dis- resent the paradigmatic application of the sociation paradigm (Holender, 1986). For dissociation paradigm, and the use of ERPs example, visibility of the prime on some tri- in conjunction with this approach may well als might well contribute to the observed contribute to a more complete assessment of effects – the measure of awareness might not the processing of an unseen stimulus. To the have been exclusive. One recent ERP study extent that semantic processing of a prime made a valiant effort to address many of the takes place, it should lead to modulation of requirements of the dissociation paradigm the N400 (i.e., a negative-going ERP compo- (Stenberg, Lindgren, Johansson, Olsson, & nent sensitive to manipulations of semantic Rosen, 2000). Most dissociation paradigm relatedness). Experiments with supraliminal studies attempt to render the prime invisible words and sentences consistently find larger using a masking procedure, assuming that deflections in N400 amplitude for incongru- the prime is invisible to all subjects on all ent than for congruent targets (Kutas & Hill- trials. An alternative approach is to vary the yard, 1980). For example, the N400 gener- visibility of the target itself and to measure ated in response to the word “lemon” would the ERP response to an unseen target stim- likely be more negative when preceded by ulus (Stenberg et al., 2000). This approach the unrelated prime “chair” than when pre- has the advantage of allowing the trial-by- ceded by the related prime “citrus.” trial assessment of target visibility. Unfortunately, studies of N400 modula- In these experiments (Stenberg et al., tion by semantically related, unseen primes 2000), a visible category name (the prime) have produced mixed results (for a review, was followed by a word that either was from see Deacon & Shelley-Tremblay, 2000). the primed category or from a different cat- For instance, in one experiment, masked egory. Target perceptibility was varied across primes led to faster responses to semantically blocks so that individual subjects could suc- related targets, but modulation of N400 was cessfully name the target on 50% of trials. evident only when primes were completely Because this subjective naming task leaves visible (Brown & Hagoort, 1993). This find- criterion setting in the hands of the subject ing implied that the N400 might consti- and does not sample conscious awareness P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

behavioral, neuroimaging, and neuropsychological approaches to implicit perception 233

exhaustively, several measures of conscious stringent objective criterion, modulation of awareness were also administered at the end the N400 remained intact (Stenberg et al., of each trial: (a) subjects indicated whether 2000). In fact, when a regression analysis was or not the word had been a member of conducted to determine whether the N400 the prime category, (b) they named the was more sensitive to categorical deviations word (guessing if necessary), and (c) they than the binary-choice discrimination task, attempted to select the target word from the intercept was reliably greater than 0. either a 2-or6-alternative forced-choice This experiment adopts most of the controls test. The 6-alternative test was considered needed to make clear inferences from behav- the most sensitive, hence the most exhaus- ioral studies using the dissociation paradigm, tive measure of awareness. Interestingly, the but also adds a more sensitive neuroimaging semantic priming effect (i.e., N400) dis- measure to provide additional evidence for tinguished between categorically consistent both quantitative and qualitative differences and categorically inconsistent words, irre- in the processing of consciously perceived spective of visibility. Although modulation and implicitly perceived stimuli. of the N400 was less pronounced when the Additional evidence for a change in the words could not be explicitly identified, the ERP pattern in response to an unseen stim- topographical pattern of activation did not ulus comes from studies of the P300, a com- differ across conditions. Qualitative differ- ponent typically occurring 260–500 ms after ences in hemispheric lateralization were evi- exposure to a relatively rare stimulus. In dent in an extended positive-going com- this case, the “rarity” of the target stimulus plex that typically accompanies cognitive depends on its relation to other stimuli pre- tasks like the one employed in these exper- sented in the study. Would the target stim- iments. This ERP component remained ulus reveal this rarity response if the other consistent irrespective of categorical clas- stimuli were not consciously perceived? A sification, but had a different topography number of studies have explored this ques- depending upon whether or not targets were tion (e.g., Brazdil, Rektor, Dufek, Jurak, & explicitly identified. Consciously reported Daniel, 1998; Devrim, Demiralp, & Kurt, targets were associated with left- lateralized 1997), but most are subject to the critique activity, whereas implicitly perceived targets that subjects were aware of the regular or elicited more distributed or right-lateralized frequent stimuli and that they had a strong activity, suggesting that different neural pop- response bias when awareness was assessed ulations were recruited under these circum- in a separate block of trials (Bernat, Shevrin, stances (Stenberg et al., 2000). & Snodgrass, 2001b). Together, the consistency of the N400 One more recent experiment (Bernat et irrespective of visibility and differences in al., 2001b) showed modulation of the P300 lateralization of raw amplitudes for visi- to a rare target word even when more rigor- ble and implicit targets strengthen claims ous criteria for measuring conscious aware- for semantic processing of words that are ness were applied to make sure that the fre- not readily identified. When the crite- quent words were not consciously detected rion for conscious awareness was based on (for a review, see Shevrin, 2001). The words the more conservative 6-alternative forced- LEFT and RIGHT were presented tachisto- choice test, 30% of the words that could scopically in an oddball design with an 80:20 not be named were correctly identified frequent-to-rare ratio. Frequent stimuli were and dropped from subsequent analyses. The made subliminal by presenting them for only binary categorization responses collected on 1 ms, and subjects were given a forced-choice the remaining trials were used to calcu- detection block after the experiment. Col- late d, which was not different from 0 – lapsed across subjects, d did not differ from providing even stronger evidence that the 0, but not all subjects showed a d of 0. Con- remaining target words were not available to sequently, the effect could be driven by a conscious awareness. Despite using a more few subjects who showed awareness on some P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

234 the cambridge handbook of consciousness

trials. However, the correlation between processing was implicit and simultaneously the d score for a given subject and their demonstrate neural consequences of implicit P300 was negative, suggesting that more perception. The strongest evidence comes awareness of the frequent stimuli actually from studies of differences in N400 ampli- diminished the P300 amplitude. Moreover, tude in response to an implicitly perceived a regression of P300 magnitude against d stimulus (Stenberg et al., 2000). Reliable dif- revealed a significant P300 effect even when ferences in N400 amplitude were evident d was extrapolated to 0. even when a fairly conservative 6-alternative forced-choice task was used to rule out Summary explicit awareness on a trial-by-trial basis. By probing for awareness of the critical stimulus One recurrent theme in this overview of immediately after presentation, this study the neuroimaging of implicit perception is reduced concerns about fleeting conscious that, when stimuli are not consciously per- perception of the stimuli (i.e., memory fail- ceptible, activation is often reduced rela- ure following conscious perception). Fur- tive to when they are consciously perceived. ther, the study adopted the regression tech- Importantly, activation in response to an nique (Greenwald et al., 1995) to show that unseen stimulus is not limited to early sen- N400 patterns persisted even when the d sory processing and often activates brain measure was extrapolated to null sensitivity. regions associated with processing that par- Finally, and perhaps most importantly, the ticular type of stimulus. These findings sug- patterns of neural activity elicited by implic- gest that implicit perception might be a itly and explicitly visible stimuli were qual- weaker version of the same processes occur- itatively different, suggesting different neu- ring for explicit perception. As for most ral mechanisms for the processing of implicit studies of implicit perception, neuroimag- and explicit stimuli. Each piece of evidence ing studies rely almost exclusively on the can be criticized if considered in isolation, dissociation paradigm, attempting to elim- but taken together they provide one of the inate explicit awareness and then attribut- most complete and convincing demonstra- ing the residual effects to implicit perceptions of implicit perception. tion. To the extent that these studies fail to Further valuable evidence for implicit meet the exhaustiveness assumption of the perception comes from fMRI studies of dissociation paradigm, they are subject to emotionally valenced faces (Whalen et al., the same critiques often leveled at behav- 1998). Implicitly perceived fearful faces pro- ioral studies (Hannula et al., 2005). The duce amygdala activation, and a subtraction strength of the evidence for implicit per- analysis revealed no additional activation of ception based on neuroimaging approaches visual cortex relative to happy faces, suggest- depends on the extent to which the studies ing the possibility that fearful faces are pro- successfully demonstrate that processing has cessed automatically via a non-cortical route. really occurred in the absence of awareness. Of course, this subtraction does not eliminate the possibility of cortical activation; A Believer’s Interpretation both happy and fearful faces could produce Although some of the experiments fail to visual cortex activation, and the subtraction address the exhaustiveness assumption suffi- just reveals the lack of additional cortical ciently, others provide more convincing tests processing of fearful faces. Even so, the fact of explicit awareness. Few individual studies that amygdala activation was greater for fear- provide unequivocal evidence for the effects ful faces in the absence of greater activation of an unseen stimulus on brain activity; how- of visual cortex is suggestive of an alterna- ever, when considered holistically, the liter- tive, non-cortical source of the activation. ature provides strong converging evidence. Together, for these results provide converg- Some experiments provide evidence that ing support for implicit perception. P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

behavioral, neuroimaging, and neuropsychological approaches to implicit perception 235

A Skeptic’s Interpretation measures and then examined the N400 for both correct and incorrect/absent responses. Although these investigations provide some Although explicit sensitivity (measured of the strongest evidence for implicit per- using d for a binary category decision task) ception, and despite the advantages of using was effectively nil for mistaken responses in sensitive neuroimaging measures, all of them the 6AFC task, it was reliably above chance adopt the dissociation paradigm without for the word identification task (in Experi- fully meeting the exhaustiveness assumption ments 2 and 3). Thus, the identification task for each subject (Hannula et al., 2005). Most clearly was not exhaustive. The 6AFC task of these studies find diminished responses to comes closer, but a skeptic could quibble less visible stimuli, raising the possibility that with several of the procedures in this study. the effect results from residual explicit pro- First, the mean d was often greater than cessing rather than from a different mech- 0, and some subjects had d values greater anism altogether. That is, these findings are than 0.5 on the binary choice. Although the consistent with a failure to meet the exhaus- regression method revealed an intercept sig- tiveness assumption. Moreover, neural acti- nificantly greater than 0, suggesting implicit vation might not be as sensitive a measure processing even when d was extrapolated to as we assume. Perhaps a sizable amount 0, the fact that many observers had greater of conscious processing is necessary to pro- than nil sensitivity raises concerns that a duce robust neural activation and to pro- few of the subjects might partially drive the duce the distributed processing that is typi- results. A better approach would be to set cally attributed to consciousness. If so, then the stimulus characteristics separately for “implicit” stimuli may have been fleetingly each subject such that d is as close as pos- or weakly perceived, and the amount of con- sible to 0 on the explicit task. Another conscious information available might not be cern is that the task used to measure d was a enough to drive robust neural activation. binary category judgment (in the category vs. This distinction might account for qualita- not in the category). This task might not be tive differences in the pattern of activation as sensitive as a presence/absence judgment, for identified and unidentified words (Sten- raising the possibility that a more sensitive berg et al., 2000). The unidentified words measure might reveal some explicit process- might have received an insufficient amount ing even when observers show no sensitiv- of conscious processing to produce the pat- ity in the category judgment. These critiques tern typically associated with full awareness; aside, this study represents one of the great- however, that qualitative difference does not est challenges to a skeptic because it uses imply the absence of explicit processing. multiple explicit measures and a sensitive Implicitly and explicitly perceived stimuli imaging measure to examine implicit pro- may produce qualitatively different patterns cessing. of activation only because the implicit stimuli received less conscious processing (not no conscious processing). Evidence for Implicit Perception – The strongest evidence reviewed here is Patient Data the N400 effect for unseen stimuli. This series of studies represents the most care- Studies of brain-damaged patients provide ful and systematic exploration of implicit some of the most compelling evidence for perception that we are aware of in any implicit perception. In fact, some have noted of the studies discussed in this chapter. the surprising acceptance of evidence for The studies carefully segregated aware and implicit perception in brain-damaged sub- unaware trials on the basis of both subjec- jects even by researchers who reject simi- tive (word identification) and objective (6- lar methods in the study of unimpaired sub- alternative forced-choice [6AFC] decisions) jects (Merikle & Reingold, 1992). In part, this P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

236 the cambridge handbook of consciousness

acceptance of evidence from patient pop- damage to portions of extrastriate visual cor- ulations derives from the belief that brain tex in the ventral visual processing stream. damage can entirely disrupt some aspects Although she can perceive and discriminate of conscious perception or memory. If so, surface features such as color and texture, then the brain damage may provide the most she shows a strikingly impaired ability to effective elimination of explicit perception, visually discriminate figural properties of much more so than simply reducing the vis- objects, such as form, size, and orientation. ibility of a stimulus via masking. In unim- Her preserved haptic and auditory discrimi- paired populations, the mechanisms for con- nation of objects reveals preserved general scious perception are potentially available, knowledge and object recognition abilities; leaving the persistent concern that any evi- her deficit is one of visual object perception. dence for implicit perception might derive Despite her inability to recognize objects from explicit contamination. However, if visually, she can use the visual structure of the mechanisms themselves are eliminated objects to guide her motor responses. For by brain damage, then any residual process- example, she shows normal performance ing must be attributed to implicit processes. when trying to insert a slate into a slot, The challenge for researchers wishing to using the proper orientation and directed provide evidence for implicit perception is movement even though she cannot report different for patient studies. Rather than try- the orientation of the slate in the absence ing to show that a particular task rules out of a motor interaction (Goodale, Milner, the use of explicit perception, researchers Jakobson, & Carey, 1991). Furthermore, she must demonstrate that the patient entirely cannot report the orientations of blocks lacks the capacity for explicit processing placed on tables, but can still reach out and in any task. Given that most such “nat- pick up the blocks with appropriate grip ural experiments” are inherently messy, aperture and limb movements (Jakobson & with some spared abilities intermixed with Goodale, 1991). impairments, conclusions from patient stud- These results countermand the intuition ies depend on a systematic exploration of the that perception produces a unitary represen- nature and extent of the deficit in process- tation of the world, that interactions with ing. In many cases, such studies require the the visual world should rely on the same same level of empirical precision necessary representations and mechanisms as visual in behavioral studies, but they are further interpretation of the world. The dissocia- hampered by the limited subject population. tion in DF’s ability to interpret and act on In this section, we consider three different the world provides evidence for two distinct sorts of evidence for a distinction between mechanisms to process visual information. implicit and explicit processing. In two of One system involves the phenomenal recog- these cases, conclusions rely heavily on the nition of parts of the visual world, and the data of a relatively small number of patients. other, operating without our awareness of First, we consider the implications of stud- the identities of objects, allows us to act on ies of DF who is a visual form agnosic. We the world. In other words, this system seems then consider two different classes of brain- to allow guided motor responses to objects damage phenomena, each of which has led even if we are unaware of what those objects to striking findings of preserved processing might be. in the absence of awareness: blindsight and The case of DF does not provide evidence visual neglect. for implicit perception in the same sense discussed throughout the rest of this chapter; at some level, DF is aware of the existence of DF and the Two Pathways Argument the object even if she cannot name it. How- The visual form agnosic patient DF ever, the case has some interesting parallels, (Goodale & Milner, 1992; Milner & Goodale, and it reveals the importance of looking for 1995) acquired her deficit from bilateral qualitative differences in performance. One P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

behavioral, neuroimaging, and neuropsychological approaches to implicit perception 237

obvious parallel is that a visual stimulus can tion of the light more accurately than would elicit an appropriate action or response even be expected by chance. This finding sug- if some aspects of it are unavailable to con- gests that V1 contributes to visual awareness, sciousness. Visual analysis of an object does because in its absence, patients do not con- not guarantee conscious perception of its sciously experience visual stimuli. Perhaps properties. More importantly, some aspects the most established explanation for blind- of visual processing occur outside of what sight posits two routes to visual perception: can be consciously reported. The case of DF (a) a pathway via V1 that leads to conscious differs from other studies of implicit per- awareness and (b) a more primitive path- ception in that her spared abilities do not way bypassing V1, perhaps via the supe- involve the processing of symbolic repre- rior colliculus. The latter route presumably sentations outside of awareness. Rather, she allows perception in the absence of aware- can engage in actions toward objects without ness. Indeed, in animals, cells in MT special- needing to use a symbolic representation or ized for the detection of motion continue to any recognition-based processes. Most stud- respond normally to moving stimuli in the ies of implicit perception focus on whether scotoma (Rosa, Tweedale, & Elston, 2000). or not implicit symbol manipulation or rep- The two-routes hypothesis provides a resentation is possible (Dulany, 2004). strong claim about the nature of implicit perception, with one route operating outside awareness and the other generating Blindsight awareness. Over the past 20 years, this Neurologists had long speculated that some hypothesis has faced a number of chal- visual functioning might persist even in lenges designed to undermine the claim that patients blinded by cortical damage (see conscious perception is entirely absent in Teuber, Battersby, & Bender, 1960 for a blindsight. In other words, these alterna- review), but the phenomenology of “blind- tive explanations question the exhaustive- sight” was not convincingly demonstrated ness of the measure of conscious aware- until the 1970s(Poppel,¨ Held, & Frost, 1973; ness, which in this case is the subjective Weiskrantz, Warrington, Sanders, & Mar- report of the subject. For example, dam- shal, 1974). Patients suffering damage to pri- age to V1 might be incomplete, with islands mary visual cortex (V1) experience a visual of spared cortex that function normally, scotoma; they fail to consciously perceive thereby allowing degraded visual experience objects that fall into the affected portion in small portions of the scotoma region of their visual field. They do not perceive (Fendrich, Wessinger, & Gazzaniga, 1992, a black hole or an empty space. Rather, the 1993; Gazzaniga, Fendrich, & Wessinger, missing region of the visual field simply does 1994; Wessinger, Fendrich, & Gazzaniga, not reach awareness, much as neurologically 1997). Brain imaging of blindsight patients intact individuals do not normally notice has returned mixed results: at least one their blind spot when one eye is closed. patient (CLT) showed a small region of Blindsight refers to the finding that some cor- metabolically active visual cortex (Fendrich tically blind patients show evidence of per- et al., 1993), whereas other researchers ception in their damaged field in the absence found no evidence for intact visual cor- of awareness. In essence, such patient evi- tex in structural scans of other blindsight dence constitutes an application of the dis- patients (e.g., Trevethan & Sahraie, 2003; sociation logic; the patient reports no aware- Weiskrantz, 2002). Moreover, lesions of V1 ness of the stimulus but still shows some in animals produce blindsight-like behav- effect of it. In a classic study of blind- ior even though these controlled lesions sight (Weiskrantz et al., 1974), lights were likely are complete (e.g., Cowey & Stoerig, flashed in the damaged visual field of patient 1995). Another alternative is that neurolog- DB. Although DB reported no awareness ically spared regions surrounding the sco- of the lights, he could point out the loca- toma receive differential sensory input as a P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

238 the cambridge handbook of consciousness

result of the presence of an item in the blind sured with d (or da where appropriate) region, thereby allowing better-than-chance along with his response criterion for a vari- guessing (Campion, Latto, & Smith, 1983). ety of tasks often used to study blindsight For example, a light source in the blind field (Azzopardi & Cowey, 1998). As expected, might also generate some visual input for responding was unbiased in a forced-choice regions outside the blind field via scatter- task. In contrast, response criterion in a ing of light, thereby indicating the presence presence/absence judgment was fairly con- of something unseen (see the commentary servative (c = 1.867), and interestingly, in Campion et al., 1983 for a discussion). it was substantially reduced by instruct- A final challenge comes from the argument ing GY to guess when unsure (c = .228). that blindsight itself might indicate a change These findings reveal the danger of relying in response criterion rather than a change in on percent correct as a primary measure awareness or sensitivity per se (Azzopardi & of blindsight; with sensitivity set to d = Cowey, 1998). This challenge is based on the 1.5, these levels of bias elicit 75% correct idea that subjective reports on single trials do responding for a forced-choice task, but 55% not fully measure awareness and that a sig- performance for a presence/absence judg- nal detection approach is needed to verify ment. In fact, any d > 1 would lead to an that the response criterion cannot entirely apparent dissociation in percentage correct, account for the results. We address this final but the result could entirely be attributed alternative in more detail here because it is to response criterion rather than differen- the most theoretically relevant to the topic tial sensitivity. Using this signal detection of this chapter. approach, GY showed greater sensitivity for Most evidence for blindsight comes from static displays in forced-choice responses a comparison of performance on two tasks: a than in presence/absence responses, but presence/absence judgment (direct measure the same did not hold for moving dis- of awareness) and a forced- choice task (indi- plays. Thus, evidence for “blindsight” to rect measure of perception), and most data motion stimuli in which patients report no are reported in terms of percentage correct awareness (presence/absence) but still show (Azzopardi & Cowey, 1998). Yet, the use of accurate forced-choice performance might percent correct to compare performance in result entirely from shifts in response cri- these two tasks could well lead to spurious terion. These results underscore the dan- dissociations between implicit and explicit ger of relying on percent correct scores in perception because percent correct mea- investigations employing blindsight patients sures are affected by response biases (Cam- and highlight the benefits of using bias- pion et al., 1983). For example, subjects free tasks. To date, relatively few investi- tend to adopt a fairly conservative response gations of blindsight have adopted these criterion (responding only when certain) important methodological changes, despite when asked to make a presence/absence an active literature that possesses surprising judgment about a near-threshold stimu- scope given the apparent rarity of blindsight lus. Furthermore, subjects may well vary patients. their criterion from trial to trial. In con- Inferences from one recent study are less trast, when subjects are forced to choose subject to the criterion problem (Marcel, between two alternative stimuli or to pick 1998). Two patients (TP and GY) com- which temporal interval contained a stimu- pleted a series of tasks that required forced- lus, response bias is less of a concern; sub- choice judgments, and only some showed jects have to choose one of the two stimuli. evidence of implicit perception. For exam- Direct comparisons of presence/absence and ple, neither showed much priming from sin- forced choice performance are therefore gle letters presented to their blind field when pitting. the task was to pick the matching letter. To examine the possibility of bias, a fre- Also, neither was more likely to select a quently tested blindsight patient’s (GY) sen- synonym of a word presented to the blind sitivity to the presence of stimuli was mea- field. However, when defining a polysemous P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

behavioral, neuroimaging, and neuropsychological approaches to implicit perception 239

word, both showed priming in their choice sphere, secondary to middle cerebral artery of a definition when a word presented to infarction. Although both blindsight and the blind field disambiguated the meaning. neglect are associated with spared process- Given that these tasks all involve forced- ing in the absence of awareness, neglect is choice decisions, differences between them characterized as an attentional (rather than are unlikely to result from response biases. sensory) deficit, and commonly occurs in the Interestingly, the finding that the least direct absence of a visual scotoma (or blind spot). measure shows an effect implies that seman- The damage in neglect occurs later in the tic concepts are activated without activating perceptual processing stream than it does the representation of the word itself. in cases of blindsight, raising the possibil- One intriguing finding is that some blind- ity that neglected stimuli might be processed sight patients apparently consciously expe- semantically to a greater extent as well (see rience afterimages of stimuli presented to Driver & Mattingley, 1998 for a review). their blind field (Marcel, 1998; Weiskrantz, In many patients, the failure to notice or 2002). Such afterimages, if frequently expe- attend to contralesional stimuli is exacer- rienced by blindsight patients, might explain bated when stimuli are presented simulta- some residual perception in the damaged neously to both left and right visual fields, field. Interestingly, the afterimages can arise presumably because these stimuli compete after information from the blind and sighted for attention (visual extinction, as described fields have been combined. When differ- earlier). For example, neglect patients might ent colored filters were used for the blind fail to eat food on the left side of their plate. and sighted field, patient DB experienced an Some patients fail to dress the left side of afterimage that was specific to the combi- their body or to brush the left side of their nation of those two colors, suggesting that hair. Although neglect can affect other pro- information from the blind field was processing modalities (e.g., haptic and auditory cessed beyond the point required to resolve processing), we limit our discussion to visual binocular differences (Weiskrantz, 2002). neglect. The phenomenon of blindsight repre- Evidence for preserved processing of sents one of the most striking demonstra- neglected visual stimuli takes several forms: tions of non-conscious perception. It pro- (a) successful same/different discrimination vides potentially important insights into the of bilaterally presented stimuli despite a fail- need for V1 in order to consciously perceive ure to report the contralesional stimulus, our environment. However, the approaches (b) intact lexical and semantic processing typically used to study blindsight are sub- of extinguished stimuli, and (c) activation ject to methodological critiques because of responses consistent with an extinguished they often do not account for response prime. Here we review evidence from each biases in the measurement of awareness. of these areas, and we also describe exper- Perhaps more importantly, most such stud- iments designed to test the claim that ies are couched in the dissociation frame- extinguished stimuli are not consciously work, inferring implicit perception based on perceived. the absence of direct evidence for conscious Many studies demonstrate the preserved perception. Consequently, blindsight find- ability to discriminate identical pairs of items ings are subject to many of the same objec- from those that are physically or categori- tions raised for behavioral work on implicit cally dissimilar (Berti et al., 1992; Verfaellie, perception. Milberg, McGlinchey-Berroth, & Grande, 1995; Volpe, Ledoux, & Gazzaniga, 1979). Typically, two pictures or words are briefly Parietal Neglect presented to the right and left of fixation, Visual neglect involves deficient awareness and patients judge whether they are the of objects in the contralesional visual field, same (have the same name) or are differ- typically resulting from damage to the pos- ent. Both patients and intact control subjects terior inferior parietal lobe in the right hemi- perform this task better than chance even P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

240 the cambridge handbook of consciousness

when the object orientations differ or when phonological priming by neglected stimuli they are two different exemplars of the same (Schweinberger & Stief, 2001), and seman- category (e.g., two different cameras). Fur- tic priming studies have explored whether ther, patients can reliably indicate that physi- neglected primes receive more extensive cally similar (and semantically related) items cognitive processing (e.g., Ladavas, Paladini, are in fact different from one another. In all & Cubelli, 1993; McGlinchey-Berroth et cases, patients report little or no awareness al., 1993). Primes are presented briefly in of the stimulus in the contralesional visual either the contralesional or ipsilesional visual field. field followed by a visible target stimulus, These findings suggest that neglect and priming is reflected in faster process- patients can process extinguished stimuli ing of the target stimulus. Lexical prim- semantically and that their representation ing apparently survives visual neglect: Prim- of these unseen stimuli is fairly complex ing was evident for both patients and nor- and complete. The dissociation between the mal controls only when word stimuli were naming and matching indicates that visual repeated and not when non-word stimuli processing of extinguished stimuli proceeds were repeated, suggesting that the neglected relatively normally despite the absence of word activated an existing representation awareness (but see Farah, Monheit, & Wal- (Schweinberger & Stief, 2001). Furthermore, lace, 1991). However, more concrete evi- the magnitude of priming was comparable dence for the absence of explicit aware- in the contralesional and ipsilesional visual ness of extinguished stimuli is required fields. In fact, left visual field priming was for a clear conclusion in favor of implicit actually greater than that of normal con- perception. This approach is logically equiv- trols for patients who neglected their left alent to the dissociation paradigm; demon- visual field. This counter-intuitive finding strate that subjects cannot perceive a stim- may result from a center-surround mech- ulus and then look for residual effects on anism that increases activation for weakly performance. Subjects claim no conscious accessible or subconscious visual stimulus experience of extinguished stimuli but still while simultaneously inhibiting activation of are able to perform a fairly complex discrim- other related items (see Carr & Dagenbach, ination on the basis of the stimulus presenta- 1990). tion. In fact, when patients were required to Similar claims have been made with name both stimuli, they frequently named respect to higher-level semantic processing only the ipsilesional item. However, some of of neglected visual stimuli. In one of these them felt that something had appeared on experiments (McGlinchey-Berroth et al., the contralesional side. None of these studies 1993), pictures, used as prime stimuli, were demonstrate that sensitivity to the presence presented peripherally in the left or the right of the extinguished stimulus is objectively visual field, and filler items (a meaningless no better than chance. visual stimulus made up of components of An alternative to examining preserved the target items) were presented on the side judgments about extinguished stimuli is to opposite. After 200 ms, the pictures were explore whether such stimuli produce lexi- replaced by a central target letter string, and cal or semantic priming. To the extent that subjects indicated whether or not the string neglect is only a partial disruption of the was a word. Semantic priming should lead to ability to form representations, the inability faster lexical decisions if the prime pictures to name extinguished stimuli might result were related to the word. Although patients from a failure to access existing represen- responded more slowly than controls, they tations rather than a failure to form a rep- showed semantic priming even though they resentation (McGlinchey-Berroth, Milberg, could not identify the prime pictures in a 2- Verfaellie, Alexander, & Kilduff, 1993). alternative forced-choice task (McGlinchey- Repetition priming tasks have been used Berroth et al., 1993). Other semantic prim- to examine lexical, orthographical, or ing tasks have found faster processing P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

behavioral, neuroimaging, and neuropsychological approaches to implicit perception 241

of a right-lateralized word following a when instructed, direct their attention to the left-lateralized prime word (Ladavas et al., “neglected” contralesional stimulus implies 1993). Given that the patient in this study that they might have had some residual was unable to read a single word presented awareness of flankers presented to the con- in the left visual field, and performed no bet- tralesional visual field. The patients in this ter than chance with lexical decision, seman- experiment also had more diffuse damage tic discrimination, and stimulus detection than is typical in neglect experiments, and tasks even without a bilateral presentation, one had left-lateralized damage. The diffuse the prime presumably was not consciously damage might affect performance on the perceived. flanker task for reasons other than hemispa- However, none of the studies discussed tial neglect. thus far provided an exhaustive test for con- Most studies of implicit perception by scious awareness of left-lateralized stimuli, neglect patients have focused on determin- leaving open the possibility that residual ing the richness of processing of neglected awareness of the “neglected” stimulus might stimuli, but relatively few studies have account for preserved priming effects. More focused on producing convincing demon- generally, the use of different paradigms or strations that neglected stimuli truly escape stimuli in tests of awareness and measures conscious awareness. One recent study of priming does not allow a full assessment adopted the process dissociation procedure of awareness during the priming task; mea- in an attempt to provide a more thorough sures of awareness may not generalize to the demonstration that processing of neglected experiment itself. Many of the priming stud- stimuli is truly implicit (Esterman et al., ies also introduce a delay interval between 2002). A critical picture appeared in one the prime and target, leaving open the pos- visual field and a meaningless filler picture sibility that patients shift their attention to appeared in the other. After a 400-ms delay, the extinguished stimulus in advance of tar- a two-letter word stem appeared in the cen- get presentation (see Cohen, Ivry, Rafal, & ter of the screen, and subjects were either Kohn, 1995). instructed to complete the stem with the Other studies of implicit perception in name of the critical picture (inclusion) or visual neglect have adopted a response com- to complete it with any word other than patibility approach in which a central tar- the picture name (exclusion). Relative to get item is flanked by an irrelevant item normal control subjects, patients were less on either the left or right side of fixation likely to complete word stems with pic- (Cohen et al., 1995). These flanker items ture names in the inclusion task, particu- were either compatible, incompatible, or larly when the picture was presented in the neutral with respect to the response required neglected visual field. In contrast, patients for the target. Interestingly, responses to were more likely than controls to complete the target were slower when the flanker the stems with picture names in the exclu- was incompatible, even when it was pre- sion task when the picture was presented sented to the contralesional visual field. In in the neglected field. Moreover, they com- a control experiment using the same mate- pleted such stems with the picture name rials, the patients were asked to respond more frequently than in a baseline condition to the flankers and were given unlimited with no pictures. If patients had explicitly time to respond. Responses were reliably perceived the stimulus, they would not have slower and more error prone for stimuli pre- used it to complete the word stem. How- sented to the contralesional visual field. This ever, they still processed it enough that it finding confirms the impairment of process- influenced their stem completion. Although ing of stimuli in the contralesional visual this study provides clearer evidence for field, but it also undermines the response implicit perception of neglected stimuli, compatibility results as a demonstration of the methods are subject to the same cri- implicit perception. That subjects could, tiques discussed in our review of behavioral P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

242 the cambridge handbook of consciousness

evidence using the exclusion paradigm. ger in relying solely on subjective reports Clearly, more systematic assessments of of awareness rather than on systematic explicit perception of neglected words are measurement of performance. Most patient needed before unequivocal claims about studies use subjects’ ability to report their implicit perception in neglect are possible. visual experience as the primary measure of explicit awareness, with implicit percep- A Believer’s Interpretation tion inferred from any spared processing in the “blind” field. Such subjective reports in Perhaps more interesting than the evidence the context of the dissociation paradigm do itself is the face validity of evidence for not provide an adequately exhaustive mea- implicit perception in patient populations. sure of explicit perception. Consequently, Neglect and blindsight illustrate the seri- performance on indirect measures might ous behavioral ramifications of the absence reflect residual explicit perception rather of awareness, and their normal behaviors than implicit perception. are in essence a constant, real-world version of the process dissociation paradigm. Such patients’ daily actions reflect their lack of What do Dissociations awareness of some aspects of their visual in Perception Mean? world, and if they had awareness of those aspects, they would perform differently. Evi- Despite protestations to the contrary, the dence for perception despite the absence of century-old debate over the mere exis- awareness in these patients is particularly tence of implicit perception continues to convincing because the absence of explicit this day. The techniques and tools have awareness is their primary deficit. In com- improved, but the theoretical arguments are bination with behavioral and neuroimaging surprisingly consistent. In essence, believers evidence, these data confirm that implicit argue that the converging evidence provides perception is possible in the absence of overwhelming support for the existence of awareness. implicit perception, whereas skeptics argue that almost all findings of implicit percep- A Skeptic’s Interpretation tion fail to provide adequate controls for Studies of patient populations rely exten- explicit contamination. As with most such sively on the logic of the dissociation debates, different conclusions can be drawn paradigm; patients lack awareness of parts from the same data. Believers can point to of their visual world, so any residual improved methodologies that provide more processing of information in those areas sensitive measures of implicit processing or must reflect implicit perception. Unfortu- that more effectively control for explicit pro- nately, few studies exhaustively eliminate cessing. Skeptics can point to the fact that the possibility of partial explicit process- none of these controls are airtight and that ing of visual information in the face of the effects, when present, tend to be small. these deficits. Nobody doubts that explicit Believers point to converging evidence from awareness is affected in both blindsight patients and from imaging studies with neu- and neglect. These deficits of awareness rologically intact individuals, and skeptics have clear behavioral consequences. How- point to the even greater inadequacies of ever, impaired awareness does not mean the controls for explicit processing in those absent awareness. One of the few studies of domains. The conclusions drawn from these blindsight to measure awareness using sig- data are colored by assumptions about the nal detection theory found that many of parsimony of each conclusion. Believers find the most robust findings supporting implicit the conclusion in favor of implicit pro- perception could be attributed to bias cessing more parsimonious because a vari- rather than residual sensitivity (Azzopardi & ety of critiques, some fairly convoluted, are Cowey, 1998). This study reveals the dan- needed to account for all of the converging P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

behavioral, neuroimaging, and neuropsychological approaches to implicit perception 243

support for implicit processing. Skeptics find are aware of those outputs. Someone unfa- conclusions in favor of implicit processing miliar with the workings of a refrigerator unappetizing because they often posit the might assume that the light is on when the existence of additional mechanisms when door is closed. Similarly, given that the only all of the data can potentially be explained information available to consciousness is that using solely explicit processing. information that has reached awareness, an More recent behavioral techniques have intuitive inference would be to assume that made progress toward eliminating the more all processing involves awareness – we never obvious objections of skeptics. Qualitative “see” evidence of processing without aware- differences, signal detection measures of ness. Claims that mental processes happen sensitivity, and regression techniques are outside of consciously mediated operations appropriate first steps toward overcoming run counter to this intuitive belief. That the critiques of the dissociation paradigm, same belief might also underlie the will- although a staunch critic might never be ingness to accept subjective reports as ade- satisfied. From this possibly irresolvable quate measures of conscious processing (see debate, perhaps some additional insights can the subjective vs. objective threshold discus- be gleaned. Regardless of whether or not sion above). The goal of the implicit percep- implicit perception exists, what can we learn tion literature is to determine whether or about perceptual processing from attempts not perceptual processing occurs with the to reveal implicit perception? What do dis- metaphorical refrigerator light out. sociations in perception, whether between A fundamental issue in the implicit per- implicit and explicit or entirely within ception literature concerns the similarities explicit processing, tell us about the nature of the types of processing attributed to of awareness and about the nature of implicit and explicit mechanisms. Is there a perception? What do these dissociations commonality to those operations that apply mean for our understanding of perception? with and without awareness? If implicit per- For the moment, let’s assume that ception exists, do the implicit mechanisms implicit perception exists. If it does exist, apply to everything outside of awareness what does it do? Does it play a functional equally, or is there some selectivity? Does role in the survival of the perceiver? Our the spotlight of consciousness perpetually perceptual systems exist to extract informa- move about, randomly illuminating implic- tion from the world to allow effective behav- itly processed information, thereby bring- ior. The world, then, presents a challenge to ing it to awareness? Or are there fundamen- a perceptual system: The information avail- tal differences between implicit and explicit able far exceeds our ability to consciously processes? This question is, in many respects, encode and retain it. Our perceptual sys- more interesting and important than the tems evolved to extract order and system- question of whether or not implicit percep- aticity from the available data, to encode tion exists at all. Implicit perception would those aspects of the world that are relevant lack its popular appeal and broad implica- for the behavioral demands of our ecolog- tions if it produced nothing more than a ical niche (Gibson, 1966). Our perceptual weak version of the same processing that systems adapted to extract signal from the would occur with awareness. The broad pop- noise, allowing us to survive regardless of ular appeal (or fear) of the notion of implicit whether we are aware of the variables that processing is that it could, under the right influence our behavior. circumstances, lead to behaviors different We are only aware of a subset of our world from those we would choose with awareness. at any time, and we are, of course, unaware One way to conceptualize the implicit/ of those aspects that fail to reach aware- explicit distinction is to map it onto ness. Just as a refrigerator light is always on the intentional/automatic dichotomy. Con- whenever we look inside, any time we exam- sciousness presumably underlies intentional ine the outputs of perceptual processing, we actions (Searle, 1992), those in which P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

244 the cambridge handbook of consciousness

perceivers can perform new operations, itative differences are theoretically signifi- computations, or symbol manipulations on cant regardless of whether or not they reflect the information in the world. Automatic a difference between implicit and explicit behaviors, in contrast, reproduce old oper- processing. Take, for example, the case of ations, computations, or symbol manipula- DF (Goodale & Milner, 1992). She can accu- tions, repeating processes that were effec- rately put an oriented card through a slot, tive in the past in the absence of intentional but lacks conscious access to the orientation control. Automatic computations occur in of the card. Although she cannot subjec- a data-driven, possibly encapsulated fash- tively perceive the orientation, other mech- ion (Fodor, 1986). Much of the evidence anisms allow her to access that information for perception without awareness is based and to use it in behavior. In other words, on this sort of data-driven processing that the dissociation implies the operation of two could potentially affect explicit processing, different processes. Moreover, the finding but that occurs entirely without conscious is significant even if her accurate behavior control. Previous exposure to a stimulus involves some degree of explicit awareness. might lead to more automatic processing The dissociation reveals the operation of two of it the next time. Or, if the implicit and different processes and different uses of the explicit processing mechanisms overlap sub- same visual information. If the behavior hap- stantially, a prior exposure might provide pened to result entirely from implicit per- metaphorical grease for the gears, increas- ception, that would be interesting as well, ing the likelihood that bottom-up processing but it is not as important as the finding will lead to explicit awareness. that two different processes are involved. The vast majority of evidence for Similarly, evidence for amygdala activation implicit perception takes this form, produc- from unreported fearful faces is interesting ing behavioral responses or neural activa- not because the faces cannot be reported tion patterns that mirror those we might but because it suggests a possible alterna- expect from explicit processing. Behavioral tive route from visual information to neu- response compatibility findings are a nice ral activation (LeDoux, 1996). Subsequent example of this form of implicit perception: control experiments might show that the The interference shown in response to an unreported faces were explicitly perceiv- implicitly perceived prime is comparable to able. However, the more interesting ques- that we might expect from a consciously tion is whether a subcortical route exists, perceived prime (except for NCE effects). not whether that subcortical route oper- Similarly, much of the fMRI evidence for ates entirely without awareness. Of course, implicit perception shows that activation to the extent that inferences about alter- from unseen stimuli mirrors the pattern of native processing mechanisms depend on activation that would occur with awareness. the complete absence of awareness, these These findings are not as theoretically findings will always be open to critique. interesting as cases in which the outcome Although evidence from the process disso- of implicit perception differs from what we ciation paradigm is subject to shifts in bias would expect with conscious perception – and motivation (Visser & Merikle, 1999), the cases in which implicit and explicit pro- underlying goal of that paradigm has at its cesses lead to qualitatively different out- base the demonstration of a qualitative dif- comes. When the results are the same for ference in performance. Whether or not this implicit and explicit processing, the standard difference reflects the distinction between skeptical critiques weigh heavily. The fail- implicit and explicit processing or between ure to meet the assumptions of the dissoci- two forms of explicit processing is of sec- ation paradigm adequately leaves open the ondary importance. possibility that “implicit” behaviors result In sum, the evidence for implicit percep- from explicit processing. In contrast, qual- tion continues to be mixed and likely will P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

behavioral, neuroimaging, and neuropsychological approaches to implicit perception 245

remain that way in spite of improved tools Bar, M., & Biederman, I. (1998). Subliminal visual and methods. A diehard skeptic likely will priming. Psychological Science, 9(6), 464–469. be able to generate some alternative, how- Bernat, E., Bunce, S., & Shevrin, H. (2001). Event- ever implausible, in which explicit process- related brain potentials differentiate positive ing alone can explain a dissociation between and negative mood adjectives during both implicit and explicit perception. Similarly, supraliminal and subliminal visual processing. believers are unlikely to accept the skeptic’s International Journal of Psychophysiology, 42, 11 34 discomfort with individual results, relying – . 2001 instead on the convergence of a large body Bernat, E., Shevrin, H., & Snodgrass, M. ( ). 300 of evidence. Methodological improvements Subliminal visual oddball stimuli evoke a P component. Clinical Neurophysiology, (1), might well force the skeptic to adopt more 112 159–171. convoluted explanations for the effects, but are unlikely to eliminate those explanations Berti, A., Allport, A., Driver, J., Dienes, Z., Oxbury, J., & Oxbury, S. (1992). Levels of pro- altogether. cessing for visual stimuli in an “extinguished” Here we propose a somewhat different ﬁeld. Neuropsychologia, 30, 403–415. focus for efforts to explore perception with Brazdil, M., Rektor, I., Dufek, M., Jurak, P., & and without awareness. Rather than trying to Daniel, P. (1998). Effect of subthreshold target eliminate all aspects of explicit perception, stimuli on event-related potentials. Electroen- research should focus instead on demon- cephalography & Clinical Neurophysiology, 107, strating differences in the perceptual mech- 64–68. anisms that vary as a function of manipu- Breiter, H. C., Etcoff, N. L., Whalen, P. J., lating awareness. Qualitative differences in Kennedy, W. A., Rauch, S. L., Buckner, R. L., perception are interesting regardless of et al. (1996). Response and habituation of the whether they reﬂect purely implicit per- human amygdala during visual processing of ception. Differences between performance facial expression. Neuron, 17, 875–887. when most explicit processes are elimi- Brown, C., & Hagoort, P. (1993). The process- nated and performance when all explicit ing nature of the N400: Evidence from masked processes can be brought to bear are interest- priming. Journal of Cognitive Neuroscience, 5, 34 44 ing in their own right and worthy of further – . study. Brown, M., & Besner, D. (2002). Seman- tic priming: On the role of awareness in visual word recognition in the absence of an References expectancy. Consciousness and Cognition, 11(3), 402–422. Abrams, R. L., & Greenwald, A. G. (2000). Parts Campion, J., Latto, R., & Smith, Y. M. (1983). outweigh the whole (word) in unconscious Is blindsight an effect of scattered light, spared analysis of meaning. Psychological Science, 11(2), cortex, and near-threshold vision? Behavioral 118–124. & Brain Sciences, 6(3), 423–486. Abrams, R. L., Klinger, M. R., & Greenwald, A. Carr, T. H., & Dagenbach, D. (1990). Semantic G. (2002). Subliminal words activate semantic priming and repetition priming from masked categories (not automated motor responses). words: Evidence for a center-surround atten- Psychonomic Bulletin & Review, 9(1), 100– tional mechanism in perceptual recognition. 106. Journal of Experimental Psychology: Learning, 341 350 Adolphs, R., Tranel, D., Damasio, H., & Damasio, Memory and Cognition, 16, – . A. R. (1995). Fear and the human amygdala. Cheesman, J., & Merikle, P. M. (1984). Priming Journal of Neuroscience, 15, 5879–5891. with and without awareness. Perception & Psy- 4 387 395 Azzopardi, P., & Cowey, A. (1998). Blindsight chophysics, 36( ), – . and visual awareness. Consciousness & Cogni- Cherry, E. C. (1953). Some experiments upon the tion, 7(3), 292–311. recognition of speech, with one and with two Baars, B. J. (1988). A cognitive theory of conscious- ears. Journal of the Acoustical Society of America, 975 979 ness. New York: Cambridge University Press. 25, – . P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

246 the cambridge handbook of consciousness

Cohen, A., Ivry, R. B., Rafal, R. D., & Kohn, C. processing. Cognitive Neuropsychology, 20(3– (1995). Activating response codes by stimuli in 6), 487–506. the neglected visual ﬁeld. Neuropsychology, 9, Dell’Acqua, R., & Grainger, J. (1999). Uncon- 165–173. scious semantic priming from pictures. Cogni- Coles, M. G., Gratton, G., & Donchin, E. (1988). tion, 73(1), B1–B15. Detecting early communication: Using mea- Deutsch, J. A., & Deutsch, D. (1963). Attention: sures of movement-related potentials to illumi- Some theoretical considerations. Psychological nate human information processing. Biological Review, 70(1), 80–90. 69 89 Psychology, 26, – . Devrim, M., Demiralp, T., & Kurt, A. (1997). The Corteen, R. S., & Dunn, D. (1974). Shock- effects of subthreshold visual stimulation on associated words in a nonattended message: A P300 response. Neuroreport, 8, 3113–3117. test for momentary awareness. Journal of Exper- Dosher, B. A. (1998). The response-window 6 1143 1144 imental Psychology, 102( ), – . regression method – some problematic ass- Corteen, R. S., & Wood, B. (1972). Autonomic umptions – comment on Draine and Green- responses to shock-associated words in an unat- wald (1998). Journal of Experimental Psychology: tended channel. Journal of Experimental Psy- General, 127(3), 311–317. chology, 94(3), 308–313. Draine, S. C., & Greenwald, A. G. (1998). Repli- Cowey, A., & Stoerig, P. (1995). Blindsight in cable unconscious semantic priming. Journal of monkeys. Nature, 373, 247–249. Experimental Psychology: General, 127(3), 286– Damian, M. (2001). Congruity effects evoked 303. by subliminally presented primes: Automatic- Driver, J., & Mattingley, J. B. (1998). Parietal ity rather than semantic processing. Journal of neglect and visual awareness. Nature Neuro- Experimental Psychology: Human Perception and science, 1, 17–22. Performance, 27(1), 154–165. Driver, J., & Vuilleumier, P. (2001). Perceptual Daza, M. T., Ortells, J. J., & Fox, E. (2002). Per- awareness and its loss in unilateral neglect and ception without awareness: Further evidence extinction. Cognition, 79(1–2), 39–88. from a Stroop priming task. Perception & Psy- Driver, J., Vuilleumier, P., Eimer, M., & Rees, G. chophysics, 64(8), 1316–1324. (2001). Functional magnetic resonance imaging Deacon, D., Hewitt, S., Yang, C., & Nagata, and evoked potential correlates of conscious M. (2000). Event-related potential indices of and unconscious vision in parietal extinction semantic priming using masked and unmasked patients. Neuroimage, 14,S68–75. words: Evidence that N400 does not reﬂect a Dulany, D. E. (2004). Higher order representa- post-lexical process. Cognitive Brain Research, tion in a mentalistic metatheory. In R. J. Gen- 9, 137–146. naro (Ed.), Higher-order theories of conscious- Deacon, D., & Shelley-Tremblay, J. (2000). How ness. Amsterdam: John Benjamins. automatically is meaning accessed: A review of Eimer, M. (1999). Facilitatory and inhibitory the effects of attention on semantic processing. effects of masked prime stimuli on motor acti- Frontiers in Bioscience, 5,E82–94. vation and behavioural performance. Acta Psy- Debner, J., & Jacoby, L. (1994). Unconscious chologica, 101(2–3), 293–313. perception: Attention, awareness and control. Eimer, M., & Schlaghecken, F. (1998). Effects Journal of Experimental Psychology: Learning, of masked stimuli on motor activation: Behav- Memory, and Cognition, 20(2), 304–317. ioral and electrophysiological evidence. Journal Dehaene, S., Naccache, L., Cohen, L., Le, B., D., of Experimental Psychology: Human Perception Mangin, J., Poline, J., et al. (2001). Cerebral and Performance, 24, 1737–1747. mechanisms of word masking and unconscious Eimer, M., & Schlaghecken, F. (2002). Links repetition priming. Nature Neuroscience, 4(7), between conscious awareness and response 752–758. inhibition: Evidence from masked priming. Dehaene, S., Naccache, L., Le Clech, G., Koech- Psychonomic Bulletin & Review, 9(3), 514–520. lin, E., Mueller, M., Dehaene-Lambertz, G., et Epstein, R., & Kanwisher, N. (1998). A cortical al. (1998). Imaging unconscious semantic prim- representation of the local visual environment. ing. Nature, 395(6702), 597–600. Nature, 392, 598–601. Dehaene, S., Piazza, M., Pinel, P., & Cohen, Esterman, M., McGlinchey-Berroth, R., Verfael- L. (2003). Three parietal circuits for number lie, M., Grande, L., Kilduff, P., & Milberg, W. P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

behavioral, neuroimaging, and neuropsychological approaches to implicit perception 247

(2002). Aware and unaware perception in Greenwald, A. G., Schuh, E. S., & Klinger, hemispatial neglect: Evidence from a stem M. R. (1995). Activation by marginally per- completion priming task. Cortex, 38(2), 233– ceptible (subliminal) stimuli: Dissociation of 246. unconscious from conscious cognition. Jour- Farah, M. J., Monheit, M. A., & Wallace, M. nal of Experimental Psychology: General, 124(1), A. (1991). Unconscious perception of “extin- 22–42. guished” visual stimuli: Reassessing the evi- Greenwald, A. G., Spangenberg, E. R., Pratkanis, dence. Neuropsychologia, 29, 949–958. A. R., & Eskenazi, J. (1991). Double-blind tests Fendrich, R., Wessinger, C. M., & Gazzaniga, M. of subliminal self-help audiotapes. Psychologi- S. (1992). Residual vision in a scotoma: Impli- cal Science, 2 (119–122). cations for blindsight. Science, 258, 1489–1491. Hannula, D. E., Simons, D. J., & Cohen, N. J. Fendrich, R., Wessinger, C. M., & Gazzaniga, M. (2005). Imaging implicit perception: Promise S. (1993). Sources of blindsight – Reply to Sto- and pitfalls. Nature Reviews Neuroscience, 6, erig and Weiskrantz. Science, 261, 493–495. 247–255. Fodor, J. (1986). Modularity of mind. In Z. W. Holender, D. (1986). Semantic activation with- Pylyshyn & W. Demopoulos (Ed.), Meaning out conscious identification in dichotic listen- and cognitive structure: Issues in the computa- ing, parafoveal vision, and visual masking: A tional theory of mind (pp. 129–167). Norwood, survey and appraisal. Behavioral and Brain Sci- NJ: Ablex. ences, 9, 1–66. Freud, S. (1966). Introductory lectures on psycho- Jacoby, L. L. (1991). A process dissociation analysis (J. Strachey, Trans.). New York: W. W. framework: Separating automatic and inten- Norton. tional uses of memory. Journal of Memory & Gazzaniga, M. S., Fendrich, R., & Wessinger, Language, 30, 513–541. C. M. (1994). Blindsight reconsidered. Current Jakobson, L. S., & Goodale, M. A. (1991). Factors Directions in Psychological Science, 3(3), 93–96. affecting higher-order movement planning: Gibson, J. J. (1966). The senses considered as per- A kinematic analysis of human prehension. ceptual systems. Boston: Houghton Mifflin. Experimental Brain Research, 8, 199–208. Goodale, M. A., & Milner, A. D. (1992). Sepa- Jaskowski, P., van der Lubbe, R. H. J., Schlot- rate visual pathways for perception and action. terbeck, E., & Verleger, R. (2002). Traces left Trends in Neurosciences, 15(1), 20–25. on visual selective attention by stimuli that are Goodale, M. A., Milner, A. D., Jakobson, L. S., & not consciously identified. Psychological Science, Carey, D. P. (1991). A neurological dissociation 13(1), 48–54. between perceiving objects and grasping them. Kanwisher, N., McDermott, J., & Chun, M. M. Nature, 349, 154–156. (1997). The fusiform face area: A module Green, D. M., & Swets, J. A. (1966). Signal detec- in human extrastriate cortex specialized for tion theory and psychophysics. New York: Wiley. face perception. Journal of Neuroscience, 17, 4302 4311 Greenwald, A. G. (1992). New Look 3: Uncon- – . scious cognition reclaimed. American Psycholo- Kiefer, M. (2002). The N400 is modulated by gist, 47, 766–779. unconsciously perceived masked words: Fur- Greenwald, A., Abrams, R., Naccache, L., ther evidence for an automatic spreading acti- 400 & Dehaene, S. (2003). Long-term semantic vation account of N priming effects. Cog- 1 27 39 memory versus contextual memory in uncon- nitive Brain Research, 13( ), – . scious number processing. Journal of Experi- Kiefer, M., & Spitzer, M. (2000). Time course mental Psychology: Learning, Memory, and Cog- of conscious and unconscious semantic brain nition, 29(2), 235–247. activation. Neuroreport, 11(11), 2401–2407. Greenwald, A. G., & Banaji, M. R. (1995). Klapp, S., & Hinkley, L. (2002). The negative Implicit social cognition: Attitudes, self- compatibility effect: Unconscious inhibition esteem, and stereotypes. Psychological Review, influences reaction time and response selec- 102(4–27). tion. Journal of Experimental Psychology: Gen- 2 255 269 Greenwald, A. G., Draine, S. C., & Abrams, R. eral, 131( ), – . L. (1996). Three cognitive markers of uncon- Klinger, M., Burton, P., & Pitts, G. (2000). Mech- scious semantic activation. Science, 273(5282), anisms of unconscious priming: I. Response 1699–1702. competition, not spreading activation. Journal P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

248 the cambridge handbook of consciousness

of Experimental Psychology: Learning, Memory, McCormick, P. (1997). Orienting attention and Cognition, 26(2), 441–455. without awareness. Journal of Experimental Koechlin, E., Naccache, L., Block, E., & Psychology: Human Perception & Performance, Dehaene, S. (1999). Primed numbers: Explor- 23(1), 168–180. ing the modularity of numerical representa- McGlinchey-Berroth, R., Milberg, W. P., Verfael- tions with masked and unmasked semantic lie, M., Alexander, M., & Kilduff, P. T. (1993). priming. Journal of Experimental Psychology: Semantic processing in the neglected visual Human Perception & Performance, 25(6), 1882– field: Evidence from a lexical decision task. 1905. Cognitive Neuropsychology, 10, 79–108. Kunst-Wilson, W. R., & Zajonc, R. B. (1980). Merikle, P. (1992). Perception without awareness. Affective discrimination of stimuli that cannot Critical issues. American Psychologist, 47(6), be recognized. Science, 207, 557–558. 766–779. Kutas, M., & Hillyard, S. A. (1980). Reading Merikle, P. M. (1994). On the futility of attempt- senseless sentences: Brain potentials reflect ing to demonstrate null awareness. Behavioral semantic incongruity. Science, 207, 203–205. & Brain Sciences, 17(3), 412. Ladavas, E., Paladini, R., & Cubelli, R. (1993). Merikle, P., & Joordens, S. (1997a). Parallels Implicit associative priming in a patient with between perception without attention and per- left visual neglect. Neuropsychologia, 31, 1307– ception without awareness. Consciousness & 1320. Cognition, 6(2–3), 219–236. Lane, R. D., & Nadel, L. (2000). Cognitive neuro- Merikle, P. M., & Joordens, S. (1997b). Measuring science of emotion. New York: Oxford Univer- unconscious influences. In J. W. Schooler (Ed.), sity Press. Scientific approaches to consciousness (pp. 109– Le Doux, J. E. (1996). The emotional brain. New 123). Mahwah, NJ: Erlbaum. York: Simon & Shuster. Merikle, P. M., & Reingold, E. M. (1991). Com- Levelt, W. J. M. (1965). On binocular rivalry. paring direct (explicit) and indirect (implicit) Assen, The Netherlands: Royal VanGorcum. measures to study unconscious memory. Jour- Levin, D. T., Momen, N., Drivdahl, S. B., nal of Experimental Psychology: Learning, Mem- 2 224 233 & Simons, D. J. (2000). Change blindness ory, and Cognition, 17( ), – . blindness: The metacognitive error of overes- Merikle, P. M., & Reingold, E. M. (1992). Mea- timating change-detection ability. Visual Cog- suring unconscious perceptual processes. In T. nition, 7, 397–412. S. Pittman (Ed.), Perception without awareness: Mack, A., & Rock, I. (1998). Inattentional blind- Cognitive, clinical, and social perspectives 55 80 ness. Cambridge, MA: MIT Press. (pp. – ). New York: Guilford Press. Macmillan, N. A. (1986). The psychophysics of Merikle, P., & Reingold, E. (1998). On demon- subliminal perception. Behavioral and Brain strating unconscious perception – Comment 1998 Sciences, 9, 38–39. on Draine and Greenwald ( ). Journal of Experimental Psychology: General, (3), Marcel, A. J. (1983a). Conscious and uncon- 127 304–310. scious perception: An approach to the relations between phenomenal experience and Merikle, P., Smilek, D., & Eastwood, J. (2001). perceptual processes. Cognitive Psychology, 15, Perception without awareness: Perspectives 238–300. from cognitive psychology. Cognition, 79(1–2), 115 134 Marcel, A. J. (1983b). Conscious and unconscious – . perception: Experiments on visual masking Milner, A. D., & Goodale, M. A. (1995). The and word recognition. Cognitive Psychology, 15, visual brain in action. New York: Oxford Uni- 197–237. versity Press. Marcel, A. J. (1998). Blindsight and shape Mitroff, S. R., Simons, D. J., & Franconeri, S. L. perception: Deficit of visual conscious- (2002). The siren song of implicit change dete- ness or of visual function? Brain, 121(8), ction. Journal of Experimental Psychology: Hu- 1565–1588. man Perception & Performance, 28(4), 798–815. McCarthy, J. C., Puce, A., Gore, J. C., & Alli- Moray, N. (1959). Attention in dichotic listening: son, T. (1997). Face-specific processing in the Affective cues and the influence of instructions. human fusiform gyrus. Journal of Cognitive Quarterly Journal of Experimental Psychology, 11, Neuroscience, 9, 604–609. 56–60. P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

behavioral, neuroimaging, and neuropsychological approaches to implicit perception 249

Morris, J. S., Frith, C. D., Perrett, D. I., Rowland, without awareness. Perception & Psychophysics, D., Young, A. W., Calder, A. J., et al. (1996). 44(6), 563–757. A differential neural response in the human Rosa, M. G. P., Tweedale, R., & Elston, G. N. amygdala to fearful and happy facial expres- (2000). Visual responses of neurons in the mid- sions. Nature, 383, 812–815. dle temporal area of new world monkeys after Most, S. B., Simons, D. J., Scholl, B. J., Jimenez, lesions of striate cortex. Journal of Neuroscience, R., Clifford, E., & Chabris, C. F. (2001). How 20(14), 5552–5563. not to be seen: The contribution of similarity Schacter, D. L. (1987). Implicit memory: His- and selective ignoring to sustained inattentio- tory and current status. Journal of Experimental nal blindness. Psychological Science, 12(1), 9–17. Psychology: Learning, Memory, and Cognition, Naccache, L., Blandin, E., & Dehaene, S. (2002). 13, 501–518. Unconscious masked priming depends on tem- Schacter, D. L., Alpert, N. M., Savage, C. R., poral attention. Psychological Science, 13(5), Rauch, S. L., & Albert, M. S. (1996). Con- 416–424. scious recollection and the human hippocam- Naccache, L., & Dehaene, S. (2001a). The prim- pal formation: Evidence from positron emis- ing method: Imaging unconscious repetition sion tomography. Proceedings of the National priming reveals an abstract representation of Academy of Sciences, 93(1), 321–325. number in the parietal lobes. Cerebral Cortex, Schacter, D. L., & Buckner, R. L. (1998). On the 11(10), 966–974. relations among priming, conscious recollec- Naccache, L., & Dehaene, S. (2001b). Uncon- tion, and intentional retrieval: Evidence from scious semantic priming extends to novel neuroimaging research. Neurobiology of Learn- unseen stimuli. Cognition, 80, 223–237. ing & Memory, 70(1–2), 284–303. Neumann, O., & Klotz, W. (1994). Motor Schweinberger, S., & Stief, V.(2001). Implicit per- responses to nonreportable, masked stimuli: ception in patients with visual neglect: Lexi- Where is the limit of direct paramater cal specificity in repetition priming. Neuropsy- specification? In M. Moscovitch (Ed.), Atten- chologia, 39(4), 420–429. tion and performance (Vol. XV, pp. 123–150). Searle, J. R. (1992). The rediscovery of the mind. Cambridge, MA: MIT Press. Cambridge, MA: MIT Press. Poppel,¨ E., Held, R., & Frost, D. (1973). Resid- Shevrin, H. (2001). Event-related markers of ual visual function after brain wounds involv- unconscious processes. International Journal of ing the central visual pathways in man. Nature, Psychophysiology, 42, 209–218. 243, 295–296. Sidis, B. (1898). The psychology of suggestion. New Pratkanis, A. R. (1992). The cargo-cult science York: D. Appleton. of subliminal persuasion. Skeptical Inquirer, 16, Simons, D. J. (2000). Attentional capture and 260–272. inattentional blindness. Trends in Cognitive Sci- Pratkanis, A., Eskenazi, J., & Greenwald, A. ences, 4(4), 147–155. (1994). What you expect is what you believe Simons, D. J., & Chabris, C. F. (1999). Gorillas (but not necessarily what you get) – A test of in our midst: Sustained inattentional blindness the effectiveness of subliminal self-help audio- for dynamic events. Perception, 28, 1059–1074. 3 tapes. Basic & Applied Social Psychology, 15( ), Squire, L. R., Ojemann, J. G., Miezin, F. M., 251 276 – . Petersen, S. E., Videen, T. O., & Raichle, M. E. Rees, G. (2001). Seeing is not perceiving. Nature (1992). Activation of the hippocampus in nor- Neuroscience, 4(7), 678–680. mal humans: A functional anatomical study of Rees, G., Kreiman, G., & Koch, C. (2002). Neural memory. Proceedings of the National Academy correlates of consciousness in humans. Nature of Sciences, 89(5), 1837–1841. Reviews Neuroscience, 3, 261–270. Stenberg, G., Lindgren, M., Johansson, M., Ols- Rees, G., Wojciulik, E., Clarke, K., Husain, M., son, A., & Rosen, I. (2000). Semantic process- Frith, C., & Driver, J. (2000). Unconscious acti- ing without conscious identification: Evidence vation of visual cortex in the damaged right from event-related potentials. Journal of Exper- hemisphere of a parietal patient with extinc- imental Psychology: Learning, Memory, & Cog- tion. Brain, 123(8), 1624–1633. nition, 26(4), 973–1004. Reingold, E., & Merikle, P. (1988). Using direct Teuber, H. L., Battersby, W. S., & Bender, M. and indirect measures to study perception B. (1960). Visual field defects after penetrating P1: JzG 0521857430c09 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 16:21

250 the cambridge handbook of consciousness

missile wounds of the brain. Cambridge, MA: etal patient with visual extinction and spatial Harvard University Press. neglect. Neuropsychologia, 40(12), 2156–2166. Tong, F. (2001). Competing theories of binocular Vuilleumier, P., Sagiv, N., Hazeltine, E., Pol- rivalry: A possible resolution. Brain and Mind, drack, R. A., Swick, D., Rafal, R. D., et al. 2 , 55–83. (2001). Neural fate of seen and unseen faces in Tong, F. (2003). Primary visual cortex and visual visuospatial neglect: A combined event-related awareness. Nature Reviews Neuroscience, 4, functional MRI and event-related potential 219–229. study. Proceedings of the National Academy of 6 3495 3500 Tong, F., Nakayama, K., Vaughan, J. T., & Kan- Sciences, 98( ), – . wisher, N. (1998). Binocular rivalry and visual Watanabe, T., Nanez, J., & Sasaki, Y. (2001). Per- awareness in human extrastriate cortex. Neu- ceptual learning without perception. Nature, ron, 21, 753–759. 413(6858), 844–848. Treisman, A. M. (1960). Contextual cues in selec- Weiskrantz, L. (2002). Prime-sight and blind- tive listening. Quarterly Journal of Experimental sight. Consciousness & Cognition, 11(4), 568– Psychology, 12, 242–248. 581. Treisman, A. (1964). Monitoring and storage of Weiskrantz, L., Warrington, E. K., Sanders, M. irrelevant messages in selective attention. Jour- D., & Marshal, J. (1974). Visual capacity in the nal of Verbal Learning and Verbal Behavior, 3, hemianopic field following a restricted occipi- 449–459. tal ablation. Brain, 97, 709–728. Trevethan, C. T., & Sahraie, A. (2003). Spatial Wessinger, C. M., Fendrich, R., & Gazzaniga, M. and temporal processing in a subject with cor- S. (1997). Islands of residual vision in hemi- tical blindness following occipital surgery. Neu- anopic patients. Journal of Cognitive Neuro- ropsychologia, 41(10), 1296–1306. science, 9(2), 203–221. Verfaellie, M., Milberg, W. P., McGlinchey- Whalen, P. J., Rauch, S. L., Etcoff, N. L., McIn- Berroth, R., & Grande, L. (1995). Comparison erney, S. C., Lee, M. B., & Jenike, M. A. of cross-field matching and forced-choice iden- (1998). Masked presentations of emotional tification in hemispatial neglect. Neuropsychol- facial expressions modulate amygdala activity ogy, 9, 427–434. without explicit awareness. Journal of Neuro- Visser, T., & Merikle, P. (1999). Conscious and science, 18, 411–418. unconscious processes: The effects of motiva- Wheatstone, C. (1838). Contributions to the tion. Consciousness & Cognition, 8(1), 94–113. physiology of vision – Part the first. On some Volpe, B. T., Ledoux, J. E., & Gazzaniga, M. S. remarkable and hitherto unobserved phenom- (1979). Information processing of visual stimuli ena of binocular vision. Philosophical Trans- in an “extinguished” field. Nature, 282, 722– actions of the Royal Society of London, 128, 724. 371–394. Vuilleumier, P., Armony, J., Clarke, K., Husain, Wolfe, J. M. (1999). Inattentional amnesia. In V. M., Driver, J., & Dolan, R. (2002). Neural res- Coltheart (Ed.), Fleeting memories: Cognition ponse to emotional faces with and with- of brief visual stimuli (pp. 71–94). Cambridge, out awareness: Event-related fMRI in a pari- MA: MIT Press. P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

CHAPTER 10 Three Forms of Consciousness in Retrieving Memories

Henry L. Roediger III, Suparna Rajaram, and Lisa Geraci

Abstract out intention or awareness. In addition to reviewing the Remember/Know judgment The study of conscious processes during literature and the topic of priming, we con- memory retrieval is a relatively recent sider such related topics as objective mea- endeavor. We consider the issue and review sures of conscious control in Jacoby’s process the literature using Tulving’s distinctions dissociation procedure and the thorny issue among autonoetic (self-knowing), noetic of involuntary conscious memory. (knowing), and anoetic (non-knowing) types of conscious experience during retrieval. One index of autonoetic consciousness is the experience of remembering (mental time Accessing Memories: Three Forms travel to recover past events and the sense of of Consciousness re-experiencing them). We review the literature on judgments of remembering (express- During most of the first hundred years ing autonoetic consciousness) and those of that researchers worked on issues in human knowing (being confident something hap- memory, considerations of conscious expe- pened without remembering it, an expres- rience were rare. Scholars interested in con- sion of noetic consciousness). These intro- scious experience did not consider states of spective judgments during retrieval have consciousness during retrieval from memory, produced a sizable body of coherent liter- and memory researchers rarely considered ature, even though the field remains filled conscious states of awareness of their sub- with interesting puzzles (such as how to jects performing memory tasks. Conscious- account for the experience of remembering ness and memory were considered separate events that never actually occurred). Prim- areas of inquiry, with few or no points of ing on implicit memory tests can be consid- contact. Researchers working on the vex- ered an index of anoetic consciousness, when ing problem of consciousness were inter- past events influence current behavior with- ested in such topics as sleeping and waking, 251 P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

252 the cambridge handbook of consciousness

hypnosis, states of alertness and awareness, lated experiences belongs here...Most of and how skilled tasks become automatic and these experiences remain concealed from appear to drop out of conscious control, consciousness and yet produce an effect among other issues. Researchers working in which is significant and which authenti- traditions of human memory considered the cates their previous existence (Ebbinghaus, products of memory – what people recalled 1885/1964,pp.1–2 ). or recognized when put through their paces in experimental paradigms – but they gen- In today’s terminology, we might say that erally did not concern themselves with the Ebbinghaus was outlining different means state of conscious awareness accompanying of retrieval or accessing information from the memory reports. memory. The first case, voluntary recol- In fact, in the first empirical studies lection, resembles retrieval from episodic of memory, Ebbinghaus (1885/1964) cham- memory (Tulving, 1983) or conscious, con- pioned his savings method of measuring trolled recollection (Jacoby, 1991). The sec- retention because it avoided reliance on ond case, involuntary recollection, has sev- “introspective” methods of assessing mem- eral modern counterparts, but perhaps the ory (recall and recognition). Nonetheless, most direct is the concept of involuntary Ebbinghaus did clearly state his opinion of conscious memory discussed by Richardson- the relation of consciousness and memory, Klavehn, Gardiner, and Java (1996), among and the relevant passage is still worth quot- others. Finally, the idea that aftereffects of ing today: experience may be expressed in behavior and the person may never be conscious of Mental states of every kind – sensations, the fact that current behavior is so guided feelings, ideas, – which were at one time is similar to the contemporary idea of prim- present in consciousness and then have disappeared from it, have not with their disap- ing on implicit or indirect tests of memory 1987 pearance ceased to exist . . . they continue to (Schacter, ). exist, stored up, so to speak, in the memory. Although Ebbinghaus raised the issue of We cannot, of course, directly observe their the relation of consciousness to memory on present existence, but it is revealed by the the first two pages of his great book that effects which come to our knowledge with a began the empirical investigation of mem- certainty like that with which we infer the ory, later generations of researchers gener- existence of stars below the horizon. These ally did not take up the puzzles he posed, effects are of different kinds. at least until recently. Today, research on In a first group of cases we can call back consciousness and memory is proceeding into consciousness by an exertion of the will apace, although the field is still fumbling directed to this purpose the seemingly lost states . . . that is, we can produce them vol- toward a lucid and encompassing theory. untarily . . . One origin of the current interest in con- In a second group of cases this survival is sciousness among memory researchers can even more striking. Often, even after years, be traced, quite after the fact, to the 1980s, mental states once present in consciousness with the rise of the study of priming in what return to it with apparent spontaneity and are now called implicit or indirect memory without any act of the will; that is, they experiments. are produced involuntarily ...inthemajor- Warrington and Weiskrantz (1968, 1970) ity of the cases we recognize the returned first showed that amnesic patients could per- mental state as one that has already been form as well as normal control subjects on experienced; that is, we remember it. indirect tests of memory, such as complet- Finally, there is a third and large group to be reckoned with here. The vanished ing fragmented pictures or words. When mental states give indubitable proof of their the corresponding pictures or words had continuing existence even if they themselves been studied recently, patients could com- do not return to consciousness at all . . . The plete the fragments as well as control sub- boundless domain of the effect of accumu- jects (see too Graf, Squire, & Mandler, 1984). P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

three forms of consciousness in retrieving memories 253

However, these same patients did much other cases the situation is reversed (Church more poorly than controls on free recall & Schacter, 1994). And a variable can even and recognition tests. The critical difference have opposite effects on an explicit and between these types of test was the instruc- implicit task (e.g., Blaxton, 1989; Jacoby, tions given to subjects. In standard memory 1983b; Weldon & Roediger, 1987), even tests like recall and recognition, subjects are under conditions when all variables except explicitly told to think back to the recent instructions to the subjects are held constant experiences to be retrieved. These are called (Java, 1994). There is no doubt that explicit explicit or direct tests of memory. In what and implicit measures are tapping different are now called implicit (or indirect tests), qualities of memory. subjects are presented with material in one Second, one straightforward and appeal- guise or another in an experiment and are ing way to think of the contrast between later told to perform the criterial task (nam- explicit and implicit forms of memory is ing fragmented words or pictures, answering to align them with states of conscious- general knowledge questions, among many ness. Explicit memory tests are thought to others) as well and quickly as possible. Usu- be reflections of memory with awareness, ally no mention is made about the test having or conscious forms of memory, whereas anything to do with the prior study episode, implicit memory tests are thought to reflect and often researchers go to some effort to an unaware, unconscious, or even auto- disguise the relation between the two. The matic form of memory. This appealing argu- finding, as in the work of Warrington and ment, which was put forward in one form Weiskrantz (1968, 1970), is that prior expe- or another by many authors in the 1980s rience with a picture or word facilitates, (e.g., Graf & Schacter, 1985; Jacoby & or primes, naming of the fragmented items. Witherspoon, 1982, among others) seems The phenomenon is called priming and has valid up to a point, but that point is quickly been much studied in the past 25 years. reached, and in 2006, no researcher would Because retention is measured indirectly and agree with this assessment. States of con- the study of memory is implicit in the proce- sciousness (e.g., aware and unaware) can- dure, these tasks are called implicit or indi- not be directly equated with performance rect memory tasks. Graf and Schacter (1985) on explicit and implicit tests. Jacoby (1991) first used the terms “explicit” and “implicit refers to this as a process purity assump- memory” to refer to these different types of tion (that a task and state of consciousness measures. Schacter (1987) provided a fine in performing the task can be equated). He historical review of the concept of implicit argued that it is difficult to provide con- memory, and a huge amount of research has vincing evidence that an implicit test does been conducted on this topic. Here we make not involve some component of conscious only a few points about this research to set awareness and even harder to show that an the stage for the chapter. explicit test does not involve unconscious First, hundreds of experiments have components. His process dissociation proce- shown dissociations between measures of dure was developed in the same paper as a explicit memory and implicit memory with promising way to cut this Gordian knot and both neuropsychological variables (differ- measure conscious and unconscious compo- ent types of patients relative to con- nents underlying task performance. Another trol subjects; see Moscovitch, Vriezen, method of measuring these components that & Goshen-Gottstein, 1993) and variables makes different assumptions is the Remem- under experimental control (e.g., type of ber/Know paradigm originally created by study condition, type of material, and many Tulving (1985) and developed by others others; see Roediger & McDermott, 1993). (Gardiner, 1988; Rajaram, 1993). Still, no Sometimes a variable can have a powerful method is generally agreed upon in the field effect on an explicit task and no effect on to perfectly measure consciousness or its role priming (Jacoby & Dallas, 1981), whereas in (or lack thereof) in various memory tasks. P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

254 the cambridge handbook of consciousness

In this chapter we adopt Tulving’s (1985) is uniquely associated with remembering, tripartite distinction among three states and it is this recognition of oneself in the of consciousness to provide coherence to past that characterizes the kind of mem- our review of the literature. Tulving dis- ory that people with amnesia lack. When tinguished among autonoetic, noetic, and Tulving described K. C., “a man without anoetic forms of consciousness, which refer, autonoetic consciousness” (1985), he illus- respectively, to self-knowing, knowing, and trated the distinction between the rich reliv- non-knowing states of consciousness. (We ing of one’s past, an ability of which most deﬁne each concept more fully below.) Even people are capable, and the cold fact-like though Tulving’s distinctions are not neces- knowledge of one’s life that most memory- sarily used by all psychologists, we see his impaired patients retain. K. C. does not seem theory as a fruitful and useful way of organiz- to have the concept of personal time. Tulving ing our chapter, and we consider it the lead- noted that K. C. can understand the concept ing theory on the relations of conscious states of yesterday, but cannot remember himself of awareness to memory performance. The yesterday. Similarly, he can understand the Remember/Know paradigm just mentioned concept of tomorrow, but cannot imagine was intended to measure autonoetic con- what he might do tomorrow. When asked sciousness (remembering) and noetic con- what he thinks when he hears a question sciousness (knowing). We turn to the ﬁrst about tomorrow, K. C. describes his mind of these concepts in the next section. as “blank” (p. 4).

Autonoetic Consciousness Measurement Issues and Theoretical Accounts Tulving (1985) defined autonoetic con- Tulving (1985) introduced the concepts of sciousness of memory as one’s awareness autonoetic and noetic consciousness and that she has personally experienced an event the Remember/Know paradigm for measur- in her past. This ability to retrieve and, ing these states of consciousness. The basic in a sense, relive events from the past has paradigm for studying these two different been characterized as a kind of mental time forms of subjective experience involves giv- travel that allows one to “become aware ing subjects explicit memory instructions (to of [their] protracted existence across sub- think back to some specific point in time) jective time” (Wheeler, Stuss, & Tulving, and asking them either to recall or recog- 1997,p.334). By this definition, autonoetic nize events from this time. For a recogni- consciousness constitutes what most people tion test, subjects are told that if they recog- think of as memory: thinking back to a par- nize the item from the study episode (if they ticular episode in life and mentally reliving judge it “old” or “studied”), then they should that event. In addition to imagining oneself try to characterize their experience of recog- in the past, autonoetic consciousness allows nition as involving the experience of either one to imagine the future and make long- remembering or knowing. They are told that term plans. In this section, we focus on the they should assign a Remember response to autonoetic consciousness that is associated items when they can vividly remember hav- with mental time travel into the past, or ing encountered the item; that is, they can remembering. remember some specific contextual detail Autonoetic consciousness includes not (e.g., what they were thinking or the item’s only the ability to travel mentally through position in the study list) that would provide time but also the complementary ability to supporting evidence that they are indeed recognize that a particular mental experi- remembering the item’s occurrence (see ence is from one’s past (as opposed to being Rajaram, 1993, for published instructions perceived for the first time). This recogni- given to subjects). Subjects are told that tion of one’s past gives rise to a feeling that they should assign a Know response to a P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

three forms of consciousness in retrieving memories 255

recognized item when they are sure that that, as compared to the sequential method, the item occurred in the study list, but they the simultaneous method leads subjects to cannot recollect its actual occurrence; they respond more liberally and increases the cannot remember any specific details associ- recognition hit rate as well as the false ated with the item’s presentation at study. alarm rate. In short, subjects know it was presented In addition to these differences in proce- in the past, but they cannot remember its dure, several important measurement issues occurrence. have arisen that reflect fundamental points In the example just given, subjects are of disagreement regarding what processes asked first to make a yes/no recognition deci- Remember and Know responses measure. sion and then to assign either a Remember Most of the measurement controversies sur- or Know response to the recognized item. round the theoretical question of what This variety of the task is the most widely states of awareness or processes are reflected used version of the Remember/Know proce- by remembering and by knowing. As we dure. A small number of studies have used describe in the next section, there are sev- a different procedure where the recognition eral accounts of remembering and knowing and Remember/Know judgments are made (see Yonelinas, 2002, for a recent review). simultaneously rather than sequentially; that is, subjects are asked to say for each test item Remembering, Knowing, and Confidence whether they remember or know that the item was on the study list or whether the One proposal is that remembering and item is new. knowing reflect different levels of confi- Research shows that the results using dence that can be explained by appealing the Remember/Know procedure can change to a signal detection model (Donaldson, critically depending on whether the sequen- 1996; Hirshman & Masters, 1997; Inoue & tial or simultaneous method is used (Hicks Bellezza, 1998; Wixted & Stretch, 2004). & R. Marsh, 1999). The standard sequential The idea is that subjects place two thresholds method of measurement implicitly assumes on a continuum of strength; the more strin- that the Remember/Know judgments are gent threshold is used for making Remem- post-recognition judgments, whereas the ber judgments, and the more lenient thresh- other method assumes that remembering old is used for making Know judgments. and knowing drive recognition. As we dis- In other words, when people are very con- cuss later, this difference in assumption par- fident that they recognize an item, they allels the debate as to whether these judg- assign it a Remember response, and when ments should be considered as subjective they are less confident, they assign it a states that a person assesses after recogni- Know response. By this view, certain inde- tion or as subjective states that uniquely pendent variables influence these judgments map onto two processes that drive recog- by affecting the amount of memory infor- nition judgments. Regardless, Hicks and R. mation available at retrieval. The availabil- Marsh argue that the simultaneous decision ity of this information, in turn, determines is more difficult than the sequential method, where people place their criteria (if they do because the judgment for each test item has not have much information, they may be to be weighed against two other possibilities very conservative and set a high threshold for (e.g., “Do I remember or know that the item responding). Thus, these models conceptu- was on the study list, or is it a new item?”). In alize Remember/Know judgments as quan- contrast, in the sequential method, subjects titatively different judgments that vary along only have to weigh the judgment against one a single continuum of degree of confidence. other possibility (e.g., “Was the item pre- It follows from this view that Know judg- sented earlier or not?” and “Do I remem- ments are isomorphic with low-confidence ber it or do I know it?”). In support of judgments and do not capture any other this hypothesis, Hicks and R. Marsh found experiential state. This criterion shift model P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

256 the cambridge handbook of consciousness

can fit several different patterns of Remem- To test for the presumed equivalence ber and Know data (see Dunn, 2004, for a between Remember/Know and high/low recent review from this perspective). confidence judgments, Rajaram, Hamilton Although single-process models do a good and Bolton (2002) adapted Gardiner and job of accounting for various associations and Java’s design and conducted the study with dissociations, they seem to lack explanatory amnesic subjects. Impaired conscious expe- power. It is difficult to know what deter- rience is a hallmark of amnesia, and both mines the placement of criteria at particu- Remember judgments (to a greater extent) lar points on the continuum in these models and Know judgments (to a lesser extent) are and how this placement might vary with dif- impaired in amnesia. Consistent with these ferent experimental conditions. More prob- findings, Rajaram et al. (2002) showed that lematic for this view are the reports that amnesic subjects were severely impaired at both meta-analyses of a large set of stud- making Remember/Know judgments, and ies (Gardiner & Conway, 1999; Gardiner, they did not produce even a hint of the Ramponi, & Richardson-Klavehn, 2002) and double dissociation observed with matched- analyses of sets of individual data (Gardiner control subjects. In contrast, the perfor- & Gregg, 1997) have not supported a key mance of matched-control and amnesic prediction of single-process models. These subjects did not differ for high- and low- models predict that the bias-free estimates confidence judgments. Such findings are for Remember judgments should be com- quite difficult to reconcile with the notion parable to that obtained for overall recog- that Remember and Know judgments are nition (that includes both Remember and redundant with confidence judgments. Know judgments). In other words, Know The findings described here show that judgments are assumed to contribute lit- confidence judgments and experiential judg- tle to the bias-free estimate, but con- ments can be differentiated. Clearly though, trary to this prediction, overall recogni- one is highly confident when reporting tion shows a larger bias-free estimate than that one remembers an event, and so it is do Remember judgments by themselves likely that recollective experience is closely (but see Wixted & Stretch, 2004, for an tied to confidence. Roediger (1999) sug- alternative view). gested that “ . . . theorists may have the rela- Other empirical evidence is also incon- tion backwards. Rather than differing levels sistent with the view that Remember and of confidence explaining Remember/Know Know judgments simply reflect high- and responses, it may well be that the study of low-confidence judgments. Gardiner and retrieval experience through the Remem- Java (1990) had shown in an earlier study ber/Know technique may help explain why that memory-intact subjects give signifi- subjects feel more or less confident” (p. 231). cantly more Remember judgments to words According to this view, confidence and rec- than non-words and significantly more ollective experience can be correlated, but Know judgments to non-words than words. confidence does not explain remembering. In contrast, words and non-words simply Remembering explains confidence. produce a main effect on high- and low- confidence judgments. Single-process mod- Remember/Know Responses and els can account for the double dissociation Dual-Process Models of Recognition observed for Remember/Know judgments and the main effect observed for confidence Dual-process models in general suggest judgments by assuming certain shifts of cri- that remembering and knowing reflect two teria, but it is not clear why the criteria independent processes in memory, termed would shift in different ways for the two “recollection” and “familiarity,” respectively sets of judgments (Remember/Know and (Jacoby, 1991; Jacoby, Yonelinas, & Jen- high/low confidence) if the types of judg- nings, 1997). Applied to the Remember/ ment are isomorphic. Know paradigm, Remember judgments P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

three forms of consciousness in retrieving memories 257

reflect primarily recollection-based memory, differently (e.g., Yonelinas, 1994). Accord- whereas Know judgments reflect primar- ing to this view, only responding based on ily familiarity-based memory performance. familiarity or fluency (as measured by Know According to this model, recollection and responses) can be modeled by a signal detec- familiarity represent independent processes: tion theory that assumes a shifting crite- They can work together or separately to rion. In contrast, responding that is driven affect memory performance. This means by recollection, as measured by Remember that Remember judgments can arise from a responses, does not fit this model. Instead, recollective process alone or from occasions these responses reflect a retrieval process when recollection and familiarity co-occur. that can be characterized as an all-or-none Know responses, on the other hand, arise threshold process. The idea is that partici- from a familiarity process that occurs in the pants can recall various types of information absence of recollection. By this view, Know about an event (e.g., its appearance, sound, responses alone can underestimate the true or associated thoughts) that either exceed or contribution of familiarity-driven processes do not exceed some set retrieval threshold. to recognition performance. If one assumes If any one of these qualities exceeds some that Remember and Know responses reflect threshold, then participants determine that the contribution of recollection and famil- they remember the item. If these qualities iarity processes, then it is important to do not exceed this threshold, then partici- measure Remember and Know responses pants might rely on various levels of familiar- in a slightly different manner than is usu- ity when endorsing the item as recognized. ally reported in the literature. The Inde- Therefore, unlike a signal detection model of pendence Remember Know (IRK) proce- Remember and Know responses, this model dure (Jacoby et al., 1997; Lindsay & Kelley, can be considered a dual-process model (see 1996; Yonelinas & Jacoby, 1995) addresses Jacoby, 1991). Recognition can be character- this issue by measuring familiarity as a ized by two distinct processes that behave proportion of Know judgments divided by differently and give rise to distinct subjective the opportunity to make a Know response states of awareness: Remember responses are (or K/1-Remember). associated with a threshold retrieval process Implicit in this view of remembering and that is driven by the qualitative features of an knowing is the assumption that the pro- event, whereas Know responses are associ- cesses driving Remember responses overlap ated with a familiarity retrieval process that with those that drive Know responses. That is driven by sheer memory strength. In this is, events that are remembered can also be model, remembering is an all-or-none pro- known. However, there are other ways to cess, whereas knowing is based purely on conceive of the relation between remember- familiarity. ing and knowing (see Jacoby et al., 1997). For The varying assumptions of these mod- example, one could assume that everything els are tied to different conceptualizations that is remembered is also known, but events about the ways in which Remember and that are known are not always remembered Know judgments denote states of conscious- (e.g., Joordens & Merikle, 1993). Or, one ness. The assumption that a single pro- could assume that the two responses are cess underlies both states of retrieval puts exclusive: things are either remembered or the emphasis on overall memory perfor- known (Gardiner & Parkin, 1990). How one mance, but at the cost of shifting focus conceives of the relation between these two from capturing (through experimentation) states of awareness and the processes under- the conscious states that accompany per- lying these states will have implications formance. The dual-process assumption rec- for how they should be measured (Rotello, ognizes the role of conscious states more Macmillan, & Reeder, 2004). overtly, although some versions of dual- Another dual-process model proposes process models place greater emphasis on that remembering and knowing are graded the relation of performance to distinct P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

258 the cambridge handbook of consciousness

underlying processes, whereas other dual- Know response is defined as the inability to process models focus directly on the func- recollect any details of the study presentational states of consciousness that accom- tion in combination with a feeling of famil- pany retrieval. iarity or certainty that the word was studied” (p. 134). As we see in the next section on noetic consciousness, these possibilities map Measurement Issues and the Role onto the theoretical debate in the literature of Instructions regarding the definition of knowing. Another measurement issue has to do with The nature of the instructions given to the instructions that one gives to subjects. subjects also has been used to argue for This topic has not been the focus of many exclusivity of the responses. Gardiner and discussions on the Remember/Know pro- Java (1993) have suggested that the instruc- cedure (but see Geraci & McCabe, 2006). tions implicitly suggest that remembering Like the other topics discussed so far, how and knowing are two states of conscious subjects interpret the Remember/Know dis- experience that cannot coexist. They argue, tinction based on the instructions given to “A person cannot at one and the same time them may determine what these responses experience conscious recollection and feel- reflect. As such, the issue of interpretation ings of familiarity in the absence of conscious of instructions also has theoretical impli- recollection” (p. 179). They note that one cations. We note that for Remember judg- state can lead to the other across time and ments, the issue of instructions may be repeated retrievals, but that retrieval on any less of a problem. People tend to be in one occasion will be associated with either agreement over what remembering means. one state of awareness or the other. Although For knowing, the psychological experience this idea is consistent with the instructions that elicits this response may depend on researchers give to subjects, it is at odds the instructions. Some published instruc- with Jacoby’s conception that events that are tions (Rajaram, 1993) tell subjects to give remembered are also familiar (Jacoby et al., a Know judgment when they are certain 1997). Recall that this independence view that the event has occurred, but their assumes that the two underlying processes recognition lacks the recollective detail that (recollection and familiarity) are separate was described to them. By this definition, and can therefore act together or in oppo- Know responses should, and probably do, sition. That is, recognition can be driven by reflect high confidence, similar to Remem- both recollection and familiarity, by just rec- ber responses. With these instructions, the ollection, or by just familiarity. Researchers distinction between the two judgments is disagree on which conception is the right likely to be based on distinct conscious expe- one, but it may be that both are correct riences. However, because Know instruc- but in different situations. This matter awaits tions differ across experimenters and across future research. labs, the definition of knowing may also differ from lab to lab. If the instructions Factors That Increase Autonoetic say something along these lines of “Give a Consciousness Remember judgment if you vividly remember the item from the study list. Otherwise, Another way to understand the distinc- give it a Know judgment,” then knowing tion between remembering and knowing has could reflect high- or low-confident mem- been to examine the various factors that give ory, some sort of feeling or familiarity, or sim- rise to each state of recollective experience. ple guessing (see Java, Gregg, & Gardiner, Several factors increase reports of remem- 1997, for examples of the variety of ways bering and include conceptual or elabora- in which subjects interpret this instruction). tive processing, generation, imagery, and dis- Kelley and Jacoby (1998) capture these var- tinctive processing. Alongside this collection ious interpretations when they suggest, “A of empirical findings, several complementary P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

three forms of consciousness in retrieving memories 259

theories have arisen. In this section, we dis- after generating study targets to semantic cuss these findings and theories together. cues, rather than simply reading them. Lastly, we discuss other factors that selec- A similar hypothesis emphasized the role tively influence remembering by decreasing of processing rather than different memory it. These factors include various forms of systems and proposed that remembering is brain damage associated with amnesia and affected by conceptual processing (Rajaram, the cognitive decline associated with aging. 1993). This work showed that ostensibly Several studies have provided empiri- conceptually driven memory effects, includ- cal evidence for the experiential distinc- ing the levels of processing effect and the tion between remembering and knowing by picture superiority effect, were obtained and showing that these two states are selectively selectively associated with remembering. affected by different independent variables. That is, subjects’ reports of remembering Because the body of evidence is now large, (but not their reports of knowing) increased we classify the effects of various indepen- when they studied words for meaning and dent variables into general categories of con- when they saw pictures at study. This work ceptual processing, imagery and generation, not only showed that remembering is influ- distinctiveness, and emotion. enced by conceptual processing but it also demonstrated that knowing was differently influenced by perceptual processing (we dis- conceptual influences cuss factors that affect knowing in the next on remembering section on noetic consciousness). Several accounts of remembering suggest Subsequently, a number of studies have that this recollective state is driven largely been conducted that support the proposal by prior conceptual processing. This idea fol- that remembering is associated with prior lows from dual-process theories of recog- conceptual processing and knowing is asso- nition (Atkinson & Juola, 1973; 1974; ciated with prior perceptual processing Jacoby, 1983a, b; Jacoby & Dallas, 1981; (see Gardiner & Richardson-Klavehn, 2000; Mandler, 1980) that propose that the rec- Rajaram, 1999; Rajaram & Roediger, 1997; ollective component of recognition mem- Richardson-Klavehn, Gardiner, & Java, 1996; ory is affected by conceptual or elabo- Roediger, Wheeler, & Rajaram, 1993, for rative processing, whereas the familiarity reviews). For example, reports of remem- component is driven by perceptual pro- bering increase after elaborative rehearsal cesses. Based on this processing distinction, (as compared to rote rehearsal) at study Gardiner (1988) proposed that Remember (Gardiner, Gawlick, & Richardson-Klavehn, responses are affected by conceptual pro- 1994). Reports of remembering are also cessing that arises from the episodic mem- affected by attention at study: Remember ory system, and Know responses are affected responses increase after study under full by perceptual processing that arises from attention as compared to divided attention the semantic or procedural memory system. (Gardiner & Parkin, 1990; Mangels, Picton, In support of this idea, Gardiner showed & Craik, 2001; Parkin, Gardiner, & Rosser, that reports of remembering increased when 1995; Yonelinas, 2001). Because dividing subjects performed some meaningful pro- attention at study has been interpreted as cessing at study. Using a level of processing a manipulation that decreases elaborative, manipulation (Craik & Lockhart, 1972), this conceptual processing but not perceptual study found that people’s reports of remem- processing, the findings showing selective bering increased after studying words for effects of dividing attention on Remember their meaning (as opposed to their lexical responses can be taken as support for the or physical properties). Also, using a gener- idea that remembering is associated with ation manipulation (Jacoby, 1978; Slamecka prior conceptual processing. &Graf,1978), this same study showed that Although the evidence just reviewed sup- people’s reports of remembering increased ports the idea that conceptual processes P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

260 the cambridge handbook of consciousness

underlie remembering and perceptual pro- distinctiveness effects cesses underpin knowing, more recent data on remembering have undercut these claims. These data Recent work shows that the distinctiveness are inconsistent with both sides of the of the study episode influences reports of argument: They show that remembering can remembering; importantly, this work shows be influenced by perceptual processes and that both perceptual and conceptual sources that knowing can be influenced by con- of distinctiveness cause increases in remem- ceptual process (Conway, Gardiner, Perfect, bering. First, take for example a manipula- Anderson, & Cohen, 1997; Mantyla, 1997; tion of distinctiveness that arises from per- Rajaram, 1996, 1998; Rajaram & Geraci, ceptual oddities of the word form, such 2000). For now, we focus on research that as orthographic distinctiveness. Words with is inconsistent with the remembering side unusual letter combinations, such as “sub- of the account. Two perceptual manipula- poena,” are remembered better than words tions have been found to affect remember- with more common letter combinations, ing. Changes in both size and orientation such as “sailboat” (Hunt & Elliott, 1980; of objects across study and test influence Hunt & Mitchell, 1978; 1982; Hunt & Toth, Remember responses, but have little effect 1990; Zechmeister, 1972). The effects of such on Know responses (Rajaram, 1996; see also orthographic distinctiveness on Remember Yonelinas & Jacoby, 1995). This work shows and Know judgments were examined in that reports of remembering increase when one study by asking subjects to study a objects are presented in the same size and list of orthographically common and disorientation at study and at test and decrease tinct words. Replicating the standard finding when the size and orientation are differ- in the literature, results showed that peo- ent at test. In retrospect, the possibility ple had superior recognition for the ortho- that both meaning and perceptual features graphically distinct words. Critically, this influence remembering is not altogether sur- manipulation selectively affected Remem- prising because much of autonoetic con- ber responses, and not Know responses sciousness is associated with retrieval of (Rajaram, 1998). In other words, remember- vivid perceptual details. The challenge then ing increased with the perceptual distinc- is to ascertain a priori the nature of vari- tiveness of the items at study. ables – conceptual or perceptual – that Similarly, noting the distinctive features would influence remembering (or auto- of a face affects remembering and not noetic consciousness) and thereby develop knowing (Mantyla, 1997). In this study, a framework that can generate useful subjects studied faces and either exam- predictions. ined the differences among them by not- To this end, Rajaram (1996, 1998; Rajaram ing the facial distinctiveness of various fea- & Roediger, 1997) developed an alternative tures or categorized faces together based theory that involved distinctiveness of pro- on general stereotypes, such as “intellec- cessing. This alternate hypothesis proposes tual” or “party-goer.” Distinctive process- that autonoetic consciousness or remem- ing of individual features increased Remem- bering reflects distinctiveness of the pro- ber responses, whereas categorizing faces cessing at study (Rajaram, 1996, 1998), increased Know responses. These results all whereas knowing, or noetic consciousness, converge on the conclusion that distinctive is influenced by the fluency of processing processing leads to conscious recollection as at study (Rajaram, 1993; Rajaram & Geraci, reflected in Remember responses. 2000). This interpretation that autonoetic consciousness especially reflects distinctive processing during encoding can accommo- effects of emotion date many of the studies mentioned so far, on remembering as well as more recent findings, which are Defining what constitutes a distinctive event discussed next. in memory is difficult (see Schmidt, 1991), P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

three forms of consciousness in retrieving memories 261

although the issue has received a resurgence events that people experience outside the of investigation and theorizing (Geraci & lab. Because the hallmark of a flashbulb Rajaram, 2006; Hunt, 1995, 2006; McDaniel memory is that people report being able & Geraci, 2006). Does distinctiveness refer to remember many contextual details from to information that is simply unusual against having first encoded the emotional event a background context (e.g., von Restorff, (e.g., people often report that they can 1933)? Is something considered distinctive if remember where they were and what they it is surprising or unexpected within a certain were wearing when they first heard the news context? Is particularly salient information that JFK had been shot; Neisser & Harsch, distinctive? Or, does distinctiveness refer to 1992; Rubin & Kozin, 1984), recent investi- a type of processing, as Hunt and McDaniel gation is aimed at examining the quality of (1993) proposed in distinguishing between these emotional memories. distinctive and relational processing? To examine whether emotional memo- Emotionally laden events are often con- ries are qualitatively different from non- sidered distinctive and are well remem- emotional or neutral memories, some studies bered (but see Schmidt, 2006, for conditions have begun examining metamemory judg- under which emotionally arousing events ments for these events, including Remem- affect memory differently from other dis- ber/Know responses and source judgments tinctive events). Evidence that emotional (Kensinger & Corkin, 2003; Ochsner, 2000). information is remembered well and in vivid In the Ochsner study, subjects studied detail comes both from studies investigat- positive pictures (e.g., a flower), negative ing emotional experimental stimuli, such as pictures (e.g., a burned body), and neu- arousing words or pictures, and from stud- tral pictures (e.g., a house) that system- ies investigating powerful emotional occur- atically varied in the amount of arousal rences outside the lab in what are called they produced. Results showed that partic- flashbulb memories (Brown & Kulik, 1977). ipants had best retention of the negative Flashbulb memories are so named because pictures, followed by the positive pictures, dramatic life events seem to be remem- with the worst memory for the neutral pic- bered in striking detail, just as is a picture tures. Importantly for our purposes, people caught in a photographic flash (although were much more likely to indicate that they later research shows that the term may be had a rich and vivid memory for emotional something of a misnomer). Many studies of pictures that were at least mildly arous- flashbulb memory have examined subjects’ ing relative to the neutral pictures. Subjects memory for large-scale naturally occurring assigned a higher proportion of Remember events, such as JFK’s assassination, the Chal- responses to emotional than to neutral pic- lenger explosion of 1996, or more recently, tures, whereas Know responses were associ- the terrorist attacks of September 11, 2001. ated mostly with neutral and positive items. Although there are doubtless important dif- Kensinger and Corkin further demonstrated ferences between laboratory and naturally that not only were people more likely to occurring emotional memories, in general remember the emotional events but they findings from these studies demonstrate that were also more likely to remember accurate emotional events produce more vivid mem- source details from the emotional events. ories for the events in question. We review Similar to the studies just described, peo- evidence from the two types of study in turn. ple often report that their flashbulb mem- Several laboratory experiments demon- ories of real-life events are also extremely strate that retention is superior for emotion- vivid and full (e.g., Christianson & Lof- ally laden items presented as pictures, words, tus, 1990). Of course, the term flashbulb or sentences relative to neutral items (see memory was developed to suggest this very Buchanan & Adolphs, 2002 and Hamann, fact. Although, flashbulb memories may feel 2001, for reviews). This work attempts to vivid, much debate has ensued regarding the provide a laboratory analog to emotional accuracy of these events and the relation P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

262 the cambridge handbook of consciousness

between accuracy and vivid memory reports a computer monitor, and so on. A partic- (see Conway, 1995, for a review). ularly interesting case (called reality moni- A recent study examined the relation toring) asks subjects to determine whether between accuracy and subjective experi- an event actually occurred or was imag- ence using the Remember/Know paradigm ined. In some sense, all recognition judg- to examine flashbulb memories for the ments require source-specifying information events of September 11th (Talarico & Rubin, of some sort. 2003). Participants were asked questions According to the source-monitoring about their flashbulb memories of Septem- framework (e.g., Johnson, 1988), people ber 11th (e.g., when they first heard what often rely on the qualities of their mem- happened, who told them, etc.), and they ories to determine their source by com- were asked questions about everyday sorts of paring the characteristics of the retrieved events from before the attack. People were memory to memories generally associated more likely to assign Remember responses with that source. For example, with a real- to their flashbulb memories as compared to ity monitoring decision, people may rely control (common) events. This pattern held on the knowledge that memories of per- when participants were tested on Septem- ceived or experienced events tend to be ber 12, 2001 and became more pronounced at more vivid and have more associated details longer delays. Talarico and Rubin found that than memories of imagined events. Con- people claimed to remember the emotional versely, they may rely on the knowledge events more vividly than the everyday events that memories from imagined sources tend despite the fact that they were no more accu- to be characterized by more information rate at recalling the flashbulb memories as about cognitive processes associated with compared to the everyday memories. Thus, cognitive effort and elaboration relative to flashbulb memories may be quite suscep- those that are perceived (e.g., Johnson, tible to error despite the great confidence Foley, Suengas, & Raye, 1988; Johnson, with which they are held, especially after Hashtroudi, & Lindsay, 1993; Johnson, Raye, long delays (see also Neisser & Harsch, 1992; Foley, & Foley, 1981). As with the Remem- Schmolck, Buffalo, & Squire, 2000). ber/Know studies, much research on source (or reality) monitoring focuses on defining the information that characterizes memo- Recollective Experience ries from various sources and shows that and Memory for Source sources can be discriminated flexibly among As the preponderance of the evidence many dimensions. The dimensions include reviewed so far indicates, autonoetic con- perceptual features of the target (e.g., sciousness is characterized by vivid, detailed Ferguson, Hashtroudi, & Johnson, 1992; feelings associated with one’s personal past, Henkel, Franklin, & Johnson, 2000; John- at least for distinctive events. Using the son, DeLeonardis, Hashtroudi, & Ferguson, Remember/Know paradigm, “vividness” is 1995; Johnson, Foley, & Leach, 1988), cog- characterized by using a single Remember nitive processes (e.g., Finke, Johnson, & response, whereas the lack of this vivid detail Shyi, 1988), the plausibility of remembered constitutes a Know response. However, this details (e.g., Sherman & Bessenoff, 1999), memory for details has also been exam- as well as related experiences (Geraci & ined by requiring subjects to assess the qual- Franklin, 2004; Henkel & Franklin, 1998). ity of their memories by assigning the cor- Thus, source decisions rely on multiple rect source of these memories (see Johnson, aspects of experience that give rise to auto- Hashtroudi, & Lindsay, 1993). In this line of noetic consciousness. research, source is defined broadly and can Some studies have compared Remem- include, for example, one list of items versus ber/Know responses and source judgments. another, items presented in one voice ver- Conway and Dewhurst (1995) had sub- sus another, in one location or another, on jects watch, perform, or imagine doing a P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

three forms of consciousness in retrieving memories 263

series of tasks. Later, they were presented including perceptual and emotional quali- with the task and asked about its source: ties of the information. Interestingly, auto- Was it watched, performed, or imagined? noetic consciousness as reflected in mem- Accurate source memory for tasks that the ory for details is not always associated with subjects performed was primarily associ- accurate memory, as we also observed in dis- ated with Remember responses, whereas cussing emotional memories. However, the accurate source memory for tasks that qualitatively distinct nature of memory – they only imagined performing was associ- accurate or inaccurate – that is accompa- ated mostly with Know responses. Accurate nied by autonoetic consciousness provides source memory for the observed tasks was greater consistency and confidence in sub- associated more equally with both Remem- jects’ judgments. Finally, these studies also ber and Know responses. These results cor- show that remembering is the central pro- roborate the idea that Remember responses cess in conscious recollections of our past, can reflect detailed perceptual memo- and that Tulving’s (1985) Remember/Know ries associated with personally experienced procedure and Johnson’s Memory Charac- events. teristics Questionnaire are useful method- In conjunction with asking source ques- ological tools for investigating the nature of tions, subjects are often asked to rate the autonoetic consciousness. qualities of their memories using the Mem- ory Characteristics Questionnaire (MCQ; Johnson, Foley, Suengas, & Raye, 1988). The Noetic Consciousness MCQ asks people to assess the qualities of their remembrances; in this context, we may In Tulving’s (1983, 1985) theory, noetic con- think of the MCQ as a further attempt to sciousness is associated with the experience gain introspective knowledge of autonoetic of knowing and with the semantic memory states of consciousness. Remember/Know system (Tulving, 1985). In this sense, Know responses have also been compared to MCQ judgments should be associated with seman- ratings (Mather, Henkel, & Johnson, 1997). tic knowledge. However, in experimental The Mather et al. study was designed to practice, Know judgments seem to capture determine why people falsely remember various types of awareness. In particular, sub- words that they never saw using the Deese- jects generally give Know judgments for two Roediger-McDermott (DRM) false memory types of cognitive experiences – knowledge paradigm (Deese, 1959; Roediger & McDer- and familiarity. In fact, in some theoreti- mott, 1995) in which subjects study lists cal treatments (e.g., Jacoby, Jones & Dolan, of related words (bed, rest, awake, tired, 1998), Know judgments are aligned with the dream . . . ) and often remember a word process of familiarity. that was not presented on the list (sleep, At the experiential level, noetic con- in this example). To examine the basis of sciousness lacks the intensity and immedi- these false Remember responses, this study acy that are associated with autonoetic con- examined their qualities using MCQ rat- sciousness. This is true by definition because ings. (We include more discussion on the noetic consciousness represents the less per- topic of illusory remembering toward the sonal and the more generic sense in which end of this chapter.) Results from the MCQ we retrieve factual events and information. ratings showed that people did report less This lack of personal involvement in the auditory perceptual detail for false Remem- retrieved information applies not only to bered items than for correct Remembered general knowledge about the world (the items (see also Norman & Schacter, 1997), usual definition of semantic memory) but whereas both types of items were associated also can apply to knowledge about ourselves with details of semantic associations. These and our own experiences. We may know we findings show that autonoetic conscious- had a fifth birthday party without it being ness is influenced by a number of variables, remembered. In this case the memory takes P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

264 the cambridge handbook of consciousness

on an impersonal quality. Another example know) when you are certain of recogniz- of the operation of noetic consciousness can ing the words but these words fail to evoke be seen in one’s memory of a trip by air- any specific conscious recollection from the plane taken ten years ago. Although a person study list. may know that he or she traveled from New To further clarify the difference between York to Calcutta, any remembrance of the these two judgments (i.e., R versus K), here are a few examples. If someone asks for events of the trip may have vanished. The your name, you would typically respond argument is that noetic consciousness differs in the “know” sense without becoming con- from autonoetic consciousness not only in sciously aware of anything about a par- terms of the content of retrieval but also in ticular event or experience; however, when the very nature of the retrieval process. We asked the last movie you saw, you would know about the airplane ride in the same typically respond in the “Remember” sense, way that we know that Thomas Jefferson that is, becoming consciously aware again was president of the United States. of some aspects of the experience. It is clear from these instructions that The Influence of Instructions on the Know judgments may be used in the sense Interpretations of Know Judgments of knowledge (as in the semantic sense The interpretation of Know judgments can of knowing one’s own name or knowing be traced back to the specific instructions Thomas Jefferson was president) or as a provided to subjects. An abbreviated version sense of familiarity where no specific details of these instructions (taken from Rajaram, can be evoked from the study phase (e.g., 1996) is provided here to illustrate this recognizing a face as familiar but not being point. We include here the instructions able to recover who the person is or where for Remember judgments as well, because you met her). In other words, Know judg- Know judgments are typically operational- ments are defined in terms of what they are ized in experimental studies in the context not (they are confident memories that lack of Remember judgments: detail), rather than what they are. However, the examples and description provided in Remember judgments: If your recognition the instructions do lead to the two interpre- of the item is accompanied by a conscious tations – knowledge and familiarity – that recollection of its prior occurrence in the study list, then write R. “Remember” is are most commonly associated with Know the ability to become consciously aware judgments. again of some aspect or aspects of what happened or what was experienced at the Knowing and Retrieval time the word was presented (e.g., aspects from Semantic Memory of the physical appearance of the word, or of something that happened in the room, It is common for people to know that Mt. or of what you were thinking and doing Everest is in the Himalayas, that mango is a at the time). In other words, the “remem- tropical fruit, and that Chicago in Novem- bered” word should bring back to mind ber is colder than Houston. However, peo- a particular association, image, or some- ple almost certainly do not know when and thing more personal from the time of study where they learned these bits of knowl- or something about its appearance or posi- edge. The conception of semantic memory tion (i.e., what came before or after that is not without its critics, and the very def- word). inition of semantic memory is sometimes a Know judgments: “Know” responses source of debate. Nevertheless, according to should be made when you recognize that the word was in the study list, but you Tulving’s theory within which the Remem- cannot consciously recollect anything about ber/Know distinction is embedded, seman- its actual occurrence, or what happened, tic memory is defined as a repository of orga- or what was experienced at the time of its nized information about concepts, words, occurrence. In other words, write K (for people, events, and their interrelations in the P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

three forms of consciousness in retrieving memories 265

world. Information from semantic memory lost, information became schematized and is retrieved as facts and without memory conceptually organized. for the details of the learning experience This sense of knowing (or just know- or the time and place where learning took ing) has not been used frequently in the place. literature, even though the distinction led An understanding of the process by which to interesting results in the Conway et al. the sense of knowing may be associated (1997) study. Several reasons may exist for with semantic memory requires a system- the scant use of Know judgments as a atic investigation of the learning and testing measure of semantic knowledge. For exam- conditions. This approach has the potential ple, the Conway et al. study clearly found to lead us to an understanding of the ways numerous repetitions of material to be nec- in which specific experimental conditions essary for schematization to occur and to give rise to noetic consciousness associated lead to just knowing the response. Spac- with memory. ing between repetitions may also be impor- The original sense of knowing as aware- tant for conceptual organization to occur. ness associated with semantic knowledge is The variety of content and learning experi- perhaps best illustrated in a study by Mar- ence accumulated over time may also inter- tin Conway and his colleagues (Conway, act with repetition and spacing and create Gardiner, Perfect, Anderson, & Cohen, memories that can shift from remember- 1997). In this study, subjects gave Remem- ing to just knowing, instead of being sim- ber and Know judgments to different types ply forgotten over time. These are but three of course material learned over an extended notable variables, and there are probably period of time. Conway and colleagues asked more. Classroom education in the Conway subjects to make an important distinction et al. (1997) study brought together these in their study between two different inter- conditions of repetition, spacing, and var- pretations of noetic consciousness by asking ied encoding quite nicely, but it is usu- them to judge between Just Knowing and ally difficult to create such conditions in Familiarity. This distinction was made for the laboratory. the precise reason of separating knowledge In a recent study, Rajaram and Hamil- from a sense of familiarity. The authors also ton (2005) created the following labora- studied two types of courses. One type con- tory conditions as a first step toward test- sisted of lecture materials (Introduction to ing the effects of varied and deep encoding Psychology, Physiological Psychology, Cog- on Remember and Know states of aware- nitive Psychology, and Social and Devel- ness. Subjects studied unrelated word pairs opmental Psychology), and another type either once or twice where the repeated pre- consisted of learning scientific methodol- sentation varied the context and was spaced ogy (research methods courses). The nature apart (e.g., a single presentation might be of learning differs in these two types of penny-cousin, whereas a repeated presen- courses: Lecture courses entail learning mas- tation would have been fence-bread, then sive amounts of content material, whereas guard-bread). At test, subjects gave recogni- methodology courses usually require active tion and Remember/Know judgments to the learning and application in smaller class set- target words (cousin, bread). Among several tings. The nature of awareness systemati- conditions in this study, the most relevant for cally varied with this distinction; subjects our present purposes are those that involved gave more Remember judgments to infor- a deep level of processing at encoding. The mation learned in lecture courses and more results showed that even after 48 hours Just Know judgments to material from the of delay (relative to 30 minutes), subjects methodology courses. Furthermore, there gave significantly more Remember responses was a shift from Remember to Just Know to words repeated under different contexts judgments over time, further supporting than to once-presented target words. Impor- the idea that once episodic details were tantly for the present discussion, subjects P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

266 the cambridge handbook of consciousness

also gave significantly more Know responses rang a bell,” “There was no association, I just after 48 hours than after 30 minutes. Thus, had a feeling that I saw it, I was sure” [p. 7].) Know judgments were responsive to both The fluency-familiarity interpretation of conceptual encoding and varied repetition Know judgments has featured prominently even after considerable delay. in our own conceptualization (see Rajaram, A recent study by E. Marsh, Meade, 1993, 1996, 1999; Rajaram & Geraci, 2000; and Roediger (2003) reported a paradigm Rajaram & Roediger, 1997). This approach that could be very useful for investigating proposes that Remember judgments are knowing as a measure of semantic memory. influenced by the processing of distinctive In this study, E. Marsh et al. investigated attributes of the stimuli, whereas Know whether reading a story before being tested judgments are sensitive to the fluency or on general knowledge questions can influ- ease with which stimuli are processed. Con- ence subjects’ ability to identify correctly siderable evidence supports this interpreta- the story or prior knowledge as the source tion of Know judgments. For example, sub- of their answers. Both immediate source jects give more Know responses to words judgments and retrospective source judg- that are preceded by a masked repetition of ments on the general knowledge answers the same word (hence increasing fluency) showed that subjects attributed many details compared to words that are preceded by a from the story to being part of their prior masked presentation of an unrelated word knowledge. This finding nicely illustrates (Rajaram, 1993). Similarly, words that are that episodic information was converted to preceded by a very brief (250 ms) presen- facts or semantic memory. Other experi- tation of semantically related words elicit ments using this paradigm and ones like more Know judgments than words that it might answer the question about how are preceded by unrelated words (Rajaram information that once held great recollective & Geraci, 2000; see also Mantyla, 1997; detail may be transformed into impersonal Mantyla & Raudsepp, 1996). Along these knowledge over time, or how remembering lines, having modality match across study becomes knowing. and test occasions (e.g., items presented visually in both cases) increases perceptual fluency and selectively increases Know judg- Knowing as Fluency and Familiarity ments relative to when the modalities mis- In contrast to the limited experimental work match between study and test (Gregg & on knowing as a measure of semantic mem- Gardiner, 1994). ory, there has been a flurry of research The idea that fluency or familiarity aimed at characterizing Know judgments as increases Know judgments and the sup- a measure of fluency or familiarity. This porting evidence are also consistent with effort may be attributed, in large part, to Gardiner and colleagues’ original proposal the ways in which Know judgments are about the nature of Know judgments defined through instructions. As described (Gardiner, 1988; Gardiner & Java, 1990; earlier, subjects are asked to give Know Gardiner & Parkin, 1990). According to this judgments when they have a feeling that view, Know judgments are mediated by something has been encountered recently processes of the procedural memory sys- but no details come to mind about that tem. This view ties Know judgments not encounter. In an interesting project on the only to the familiarity process but also to actual reports by subjects, Gardiner, Ram- perceptual priming. In more recent works, poni, and Richardson-Klavehn (1998) exam- Gardiner and his associates have reconsid- ined subjects’ transcripts of reasons they ered Tulving’s original classification system gave Know judgments and found they did of associating Know judgments with seman- so typically based on a feeling of familiarity tic memory that we described in an earlier as expressed in some of the statements made section (see for example, Gardiner & Gregg, by subjects (“It was one of those words that 1997; Gardiner, Java, & Richardson-Klavehn, P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

three forms of consciousness in retrieving memories 267

1996; Gardiner, Kaminska, Dixon, & Java, to mind in response to fragmented word 1996). cues that could be solved with studied words Other recent theories that distinguish (e.g., l ph n for elephant) or non- between a recollective basis and a famil- studied words (e.g., a lb a for iarity basis of recognition memory have sailboat). The advantage in completing frag- also contributed to the interpretation of ments of studied words compared to non- Know judgments as reflecting the fluency studied words is that it gives a measure or familiarity process (Jacoby, et al., 1997; of priming, and priming can be dissociated Yonelinas, 2001). These dual-process models from explicit measures of memory by many are based on the process dissociation pro- variables. Extensive experimental efforts are cedure (Jacoby, 1991) that was developed made in these studies to discourage subjects to measure the independent and oppos- from using explicit retrieval strategies and ing influences of recollective (consciously to exclude subjects who nevertheless use controlled) and familiarity (more auto- explicit or deliberate retrieval to complete matic) processes. According to this view, these tasks. Thus, such tasks are assumed Know judgments in Tulving’s (1985) Re- to measure a relatively automatic process member/Know procedure provide an under- that contributes to memory performance. estimation of the extent to which the Furthermore, a subset of this class of tasks familiarity process contributes to memory is also particularly, though not exclusively, performance. As noted in an earlier sec- sensitive to match or mismatch in the per- tion, Jacoby and colleagues have assumed ceptual attributes of study and test stim- independence between these processes and uli. For example, changes in the presen- have proposed a mathematical correction tation modality (from auditory to visual) (Know/(1-Remember)) to compute the in- reduce the magnitude of priming compared fluence of familiarity. On this point these to matched modality across study and test dual-process models diverge from the frame- (visual to visual; Rajaram & Roediger, 1993; works and models described earlier, but Roediger & Blaxton, 1987; see also Weldon these models are nevertheless in agreement & Roediger, 1987, for similar conclusions with the main point under consideration based on changes in surface format across here – Know judgments measure the effects pictures and words.) Thus, perceptual prim- of fluency or familiarity. ing is a measure of performance on implicit memory tasks where perceptual features exert a strong influence on the magnitude Know Judgments and Perceptual Priming of priming. Some of the theoretical interpretations In contrast to perceptual priming, Know reviewed in the previous section suggest the judgments reflect retention while subjects strong possibility that the same processes are engaged in the explicit retrieval of stud- should mediate Know judgments and per- ied information. Therefore, the possibility ceptual priming. Issues surrounding this pro- that these two measures, one derived from posal are tricky, however, and the evidence implicit memory tasks and the other from is mixed. We review the issues and evidence explicit memory tasks, have the same under- here, albeit briefly. lying basis is intriguing. The earlier proposal Perceptual priming is measured with of Gardiner and colleagues that Know judg- implicit memory tasks, such as word stem ments are influenced by processes of the completion, word fragment completion, and procedural memory system suggests such perceptual identification (see Roediger, 1990 equivalence because perceptual priming is and Roediger & McDermott, 1993, for assumed to be mediated by the procedural reviews.) For example, on a task such as word memory system (Tulving & Schacter, 1990). fragment completion, subjects first study a Furthermore, dual-process models pro- list of words (e.g., elephant) and are later posed by Jacoby, Yonelinas, and colleagues asked to complete the first word that comes also suggest such equivalence. In these P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

268 the cambridge handbook of consciousness

approaches, automatic processes and famil- in ways similar to their effects on prim- iarity processes appear to be interchangeable ing tasks, other studies have not shown concepts; the former is typically associated such parallels. For example, the generation with priming on implicit tests, and the lat- effect (when words generated from seman- ter is associated with the Know component tic cues are better recalled and recognized of explicit memory performance. Whether than words that are simply read) that is reli- or not these measures are isomorphic, and ably observed in explicit memory tasks (e.g., the evidence reviewed below is mixed on Jacoby, 1978; Slamecka & Graf, 1978) is usu- this issue, it seems intuitive to assume that ally reversed in perceptual priming (Blaxton, Know judgments share some of the prop- 1989; Srinivas & Roediger, 1990), but such erties both of perceptual priming and of a reversal is not observed for Know judg- explicit memory. If priming is placed at one ments (Gardiner & Java, 1990) even under end of the continuum of conscious aware- optimally designed conditions (Java, 1994). ness and Remember judgments at the other This result clearly undermines any straight- end, Know judgments by definition fall in forward notion that Know judgments solely between, albeit on the conscious side of this reflect perceptual priming. continuum. By virtue of being at the brink of Word frequency effects across a repetition conscious awareness, some processing com- priming task and a recognition memory task ponent of Know judgments might share its also challenge the notion that Know judg- basis with priming. ments and priming performance are sim- As just mentioned, the evidence seems ilarly responsive to independent variables. mixed on this issue, although only a handful For example, Kinoshita (1995) reported that of studies have addressed it. For instance, on a priming task of making lexical deci- dividing attention during study with an audi- sions (where subjects decide whether a tory tone or digit monitoring task (when letter string is a word or a non-word), the subjects’ main task is to read the words) low-frequency words yielded greater prim- adversely affects explicit memory tasks, but ing than high-frequency words even fol- leaves perceptual priming intact (Jacoby, lowing unattended study conditions. This Woloshyn, & Kelley, 1989; Parkin & Russo, finding suggests that low-frequency words 1990). Parallel effects of tone monitoring should lead to greater Know judgments than during study are observed on Remember high-frequency words. However, Gardiner and Know judgments, respectively (Gar- and Java (1990) had previously reported an diner & Parkin, 1990). Similarly, a study of unambiguous advantage for low-frequency pictures and words dissociates performance words over high-frequency words in Remem- on explicit memory tasks and implicit member judgments, and this variable had little ory tasks such that the picture superiority effect on Know judgments. Other recent evi- effect (better memory for pictures than dence also suggests a distinction between words – Madigan, 1983) is reliably observed types of familiarity processes that mediate on the explicit memory task, but this effect recognition relative to priming. For example, reverses on the perceptual priming task Wagner, Gabrieli, and Verfaellie (1997)have of word fragment completion (Weldon & reported that the familiarity process associ- Roediger, 1987). This dissociative pattern ated with recognition is more conceptually has been reported for Remember and based than processes that underlie percep- Know judgments as well, where the picture tual priming. Wagner and Gabrieli (1998) superiority effect was observed for Remem- have further argued that processes support- ber judgments and its reversal was observed ing perceptual priming and the familiarity on Know judgments (Rajaram, 1993, component of recognition are both function- Experiment 2). ally and anatomically distinct. In contrast to these findings where inde- The evidence on this issue is also mixed pendent variables affected Know judgments in studies of individuals with anterograde P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

three forms of consciousness in retrieving memories 269

amnesia. Whereas explicit memory per- reasons outlined earlier and mixed empiri- formance is severely impaired in amnesia, cal evidence reviewed here, this area is ripe perceptual priming for single words or pic- for extensive investigation. ture is found to be intact (see Moscovitch, Vriezen, & Goshen-Gottstein, 1993; Knowing and Confidence Schacter, Chiu, & Ochsner, 1993; Verfaellie Earlier we considered the contention that & Keane, 2002, for reviews). Preserved judgments of knowing may simply reflect perceptual priming in amnesia suggests low-confidence judgments. That is, perhaps that Know judgments (or the familiarity people give Remember judgments when component of recognition) should also be they are highly confident that an event preserved if perceptual priming and know- occurred previously and Know judgments ing have the same bases. Recent evidence when they believe it occurred but are less has started to delineate conditions in the confident. Proponents of this view (e.g., Remember/Know paradigm that show that Donaldson, 1996) argue that the Remember/ amnesic subjects can indeed utilize the Know distinction is merely a quantitative familiarity component to boost their Know one (how much “memory strength” does the judgments (Verfaellie, Giovanello, & Keane, tested event have?), rather than a qualitative 2001). However, converging evidence difference (reflecting, say, recollection and from different paradigms (see Knowlton & fluency). To review a few points made pre- Squire, 1995; Schacter, Verfaellie, & Pradere, viously, the signal detection models with sev- 1996; Verfaellie, 1994) has also shown a eral criteria can often account for data after deficit in Know judgments in amnesic the fact, but lack true explanatory power patients, although this deficit is far lower in in predicting dissociations and associations magnitude than the deficit in Remember between Remember and Know judgments. judgments (see Yonelinas, Kroll, Dobbins, It is difficult to know what determines the Lazzara, & Knight, 1998). Evidence from placement of criteria at particular points on studies with amnesic patients has further the continuum in these models and how this identified neuroanatomical regions that are placement might vary with different exper- differentially associated with remembering imental conditions. Further, experimental and knowing. Specifically, evidence from evidence cited in the earlier part of the studies of amnesic patients (Moscovitch & chapter (e.g., Rajaram et al., 2002) reveal- McAndrews, 2002, Yonelinas et al., 2002) ing different effects of independent and suggests that Remember judgments are subject variables on Remember/Know judg- mediated by the hippocampus, whereas ments and confidence judgments is incon- familiarity is mediated by parahippocampal sistent with the idea the Know judgments structures in the medial temporal lobe merely reflect confidence. Although con- region. Neuroimaging evidence also sup- fidence and Remember/Know judgments ports this distinction (but see Squire, Stark, are related, remembering may explain con- & Clark, 2004, for a different view on the fidence judgments rather than the other proposed structural dichotomies). way around. Together, a comparison between perceptual priming and Know judgments has Knowing and Guessing revealed at best mixed evidence. As some have noted, the key to the differences A potential problem in interpreting the between processes that affect perceptual nature of Know judgments is the extent priming and Know judgments (or familiar- to which subjects include guesses when ity) may lie in the greater involvement of making Know judgments. In this situation, conceptual processes in explicit recognition Know judgments would reflect not only (Verfaellie & Keane, 2002, Wagner et al., memory processes but also pure guesses, 1997.) It is clear that both for theoretical thereby complicating the inferences we P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

270 the cambridge handbook of consciousness

might draw about the nature of know- of memory that is associated with noetic ing. An even more serious concern might consciousness. be that Know judgments simply reflect guesses and nothing more. However, nei- Knowing and Other Metamemory ther of these concerns seems to compromise Judgments the data typically obtained for Know judg- Remember/Know judgments are by defini- ments. The first concern is circumvented in tion metamemory judgments – subjects indi- most studies by the careful use of instruc- cate their assessment of retrieval experience tions that strongly discourage guessing. In for items that they have recalled or recog- fact, the generally low false alarm rates nized. People are able to make other judg- that are typical in these studies suggest that ments that may seem similar to Know judg- this approach is largely successful. The sec- ments, but are in fact quite distinct. For ond concern about Know judgments sim- example, people can reliably report feelings- ply being equivalent to guesses is addressed of-knowing indicating that they can rec- by Gardiner and colleagues in studies where ognize an item on a multiple-choice test they required subjects to make Remem- even though they are unable to recall it ber, Know, and Guess judgments to items (see Koriat, 1995), and people can also reli- presented in recognition memory tests ably differentiate whether or not they are (see Gardiner & Conway, 1999; Gardiner, in a tip-of-the-tongue state (where they feel Java, & Richardson-Klavehn, 1996; Gardiner, that the answer or the word they are look- Kaminska, Dixon, & Java, 1996; Gardiner, ing for is on the tip of their tongue and Ramponi, & Richardson-Klavehn, 1998; could be retrieved; see Brown, 1991, for a Gardiner, Richardson-Klavehn, & Ramponi, review). Both of these types of experiences 1997). By and large, Know and Guess differ from Know judgments in that the lat- responses turn out to be functionally dis- ter experience is associated with informa- tinct such that Know judgments reflect tion that is already retrieved (as in recall) memory for prior information and Guess or has been presented (as in recognition). judgments do not. As Gardiner and Con- That is, Know judgments characterize a par- way (1999) note, the relevance of taking ticular experiential state that accompanies guesses into account in the Remember/ retrieved information. There has been little Know paradigm seems to be to minimize experimental effort as yet directed toward noise for Know judgments. We note that the possible relation between Know judg- this could be achieved either by including ments and these other states of awareness. a separate response category or by instructing subjects to refrain from guessing. As dis- Concluding Remarks about Interpreting cussed in a previous section, the potential Know Judgments variations across laboratories in communi- cating Remember/Know instructions may The preceding sections bring into focus both account for some differences in the liter- the difficulties in characterizing Know judg- ature. For these reasons, we have empha- ments and the successes that have been sized elsewhere the care and effort that are accomplished so far in doing so. Know judg- needed to administer the Remember/Know ments have been defined both in terms task properly (Rajaram & Roediger, 1997). of what they are (semantic memory, flu- Despite these concerns, the experimental ency, familiarity), what they are not (low- effort of nearly 20years has yielded fairly sys- confidence responses, guesses, and other tematic, informative, and interesting findings metamemorial judgments), and what they about the nature of Know judgments. This might or might not partly reflect (percep- effort has also answered as many questions tual priming.) These efforts show the chal- as it has raised critical issues for future inves- lenges associated with distinguishing dif- tigation. Finally, this empirical effort in the ferent states of consciousness – autonoetic literature has begun to identify the nature and noetic – experimentally and the role P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

three forms of consciousness in retrieving memories 271

research can play in successfully delineat- fine our remarks to responding on implicit ing these mental states. As such, these find- memory tests. ings have refined the questions and sharp- ened the direction for future studies aimed Measurement Issues: Responding on at understanding the relationship between Implicit Tests noetic consciousness and memory. Implicit memory tests differ from explicit ones because they are designed to measure Anoetic Consciousness retention when people are not aware of the influence of prior events on their behavior. As discussed previously, implicit tests mea- So far our discussion has focused on two sure memory indirectly (and therefore are types of conscious experience: autonoetic also called indirect tests) by having subjects consciousness where one feels as though perform tasks that, unbeknownst to them, one is mentally reliving a past experience in can be accomplished using previously stud- the present and noetic consciousness where ied items. These tasks may include filling one simply knows that one has experienced in fragmented words or naming fragmented the event before, but cannot vividly relive pictures, generating items that belong to a it. These two states of conscious aware- category, or simply answering general knowl- ness have in common that they both indi- edge questions. Subjects can perform the cate knowledge of past events. However, task at some level whether or not they have there is a third class of memory phenomena studied relevant material. However, they are that is characterized by the lack of aware- more likely to fill in a fragmented word, for ness that an event occurred in the past example, if they had been exposed to it pre- even though the event changes behavior. viously than if the fragment is filled by a Ebbinghaus (1885/1964) described this class word that was not studied. The study expe- of event and Tulving (1985) referred to this rience is said to prime the correct comple- occurrence as exemplifying anoetic, or non- tion, so the phenomenon is called priming. knowing, consciousness. One could quib- Paul Rozin (1976) was among the first to call ble that the characteristic of “non-knowing” attention to this notion. Importantly, prim- means that subjects are not conscious, and so ing, by definition, occurs without autonoetic this state should not be included in the list. or noetic awareness of the study episode. However, in other realms of inquiry, being In this way, implicit tests measure memory asleep or in a coma is referred to as a state that is associated with anoetic consciousness. of consciousness even though both indicate However, once a person has produced an the absence of awake consciousness. item on the test, he or she may become In anoetic consciousness, a person is fully aware, after the fact, that it was from a awake and alert, but is unaware that some recently experienced episode. In this case, past event is influencing current behavior. retrieval of the item seems to occur auto- Unlike the first two states that accompany matically, but the experience of recognition performance on explicit memory tests like occurs later; this type of experience has been free recall and recognition, anoetic con- referred to as involuntary conscious mem- sciousness is associated with memory perfor- ory by Richardson-Klavehn and Gardiner mance on a separate class of memory tests, (2000) and is discussed in a later section. called implicit memory tests. Of course, many phenomena in other areas of psychol- Perceptual and Conceptual Tests ogy (particularly social psychology) might also be said to refer to anoetic consciousness All implicit tests are designed to measure because all sorts of factors affect behavior anoetic consciousness, but they differ in the without the person becoming aware of the nature of the processes they require. Roedi- critical variables controlling behavior (e.g., ger and Blaxton (1987) first proposed that Wegner, 2002; Wilson, 2002). Here we con- there were (at least) two types of implicit P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

272 the cambridge handbook of consciousness

and explicit tests, perceptual and concep- to complete them with the first words that tual tests. It is probably best to think of per- come to mind (Warrington & Weiskrantz, ceptual and conceptual dimensions as sepa- 1968). There are usually ten or more possible rate continua, so that tests could rely mostly solutions in this kind of test (element, ele- on perceptual processes or conceptual pro- gant, etc.) so priming is measured by the bias cesses or some combination of the two (see to produce elephant after study of that word Roediger, Weldon & Challis, 1989, for a dis- relative to the case in which it has not been cussion of converging operations that can be studied. Similarly, the word fragment com- applied to define tests). Most explicit tests pletion test requires subjects to complete are primarily conceptual in nature. Simi- words with missing letters, such as e e h n , larly, implicit tests have been broadly clas- with the first word that comes to mind (e.g., sified into those that require primarily per- Tulving, Schacter & Stark, 1982). Priming ceptual analysis of the target items and those in both tests is measured by the propor- that require primarily meaningful analysis. tion of fragments completed with studied According to the transfer appropriate pro- solutions minus the proportion completed cessing view (Blaxton, 1989; Kolers & Roedi- with non-studied solutions. The picture frag- ger, 1984; Roediger, 1990; Roediger, Weldon, ment completion test provides fragmented & Challis, 1989), retention on both explicit pictures (following study of intact pictures) and implicit tests benefits to the extent that with instructions for subjects to name the the tests require similar stimulus analysis pictures (Weldon & Roediger, 1987). (i.e., similar conceptual or perceptual pro- Other perceptual tests require naming cessing between study and test), regardless of degraded words or pictures. These are of whether one is consciously aware of the called word identification and picture iden- study episode. A large body of work is con- tification because people are required to sistent with this prediction (although there identify words or pictures presented very are clear exceptions, too). Here, we describe briefly (and sometimes followed by a back- just a few examples of these classes of tests ward mask). A variant on this task requires that have emerged from this line of thinking participants to identify increasingly com- (see Toth, 2000, for a comprehensive list of plete fragments of stimuli. Here, the item is implicit memory tests). revealed slowly, and what is measured is the Perceptual implicit memory tests gener- level of clarity needed to identify the item, ally require participants to identify a phys- either a picture (e.g., Snodgrass & Corwin, ically degraded or rapid presentation of a 1988; Snodgrass, Smith, Feenan, & Corwin, stimulus. Priming on these tests is influenced 1987) or a word (e.g., Hashtroudi, Fergu- by the perceptual format used at encod- son, Rappold, & Cronsniak, 1988; Johnston, ing, such as modality (auditory or visual) Hawley & Elliot, 1991). Again, priming on or form (e.g., picture or word) of presen- these tasks is measured by the percentage of tation. Conversely, these tests are relatively clarity required for identification when the unaffected by meaningful analysis, such as item was studied relative to when it was not the semantic analysis required by levels- studied. of-processing manipulations (e.g., Jacoby & Conceptual tests represent the other class Dallas, 1981). There are several popular per- of implicit memory tests (Roediger & Blax- ceptual implicit memory tests, and here we ton, 1987). These tests are largely unaffected describe only a few. by perceptual manipulations at encoding The word stem completion test is one of (e.g., the modality of presentation does the most popular perceptual implicit tests. not affect priming). Instead, priming on In verbal versions of this task, subjects are these tests is affected by meaningful fac- exposed to a list of words in one phase of tors manipulated during study, such that the experiment (e.g., elephant) and then, in more priming occurs with more meaning- a later phase, presented with the first three ful analyses. One commonly used concep- letters of words (e.g., “ele ”) and are asked tual implicit test is the word association test P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

three forms of consciousness in retrieving memories 273

(see Shimamura & Squire, 1984). In this test, problem for implicit memory research is participants see words (some of which are up for debate (see Roediger & McDermott, associated with the studied words; e.g., ele- 1993), but certainly it is possible that phant) during study and are asked to quickly neurologically intact participants may treat produce all the associated words that come implicit tests like explicit ones (see Geraci to mind in response to the cue word pre- & Rajaram, 2002). Given this possibil- sented during the test (e.g., lion). In a similar ity, several researchers have provided rec- test, the category exemplar production test, ommendations to help limit participants’ participants see category names at test (ani- awareness of the study-test relation and to mals) and are asked to quickly produce as devise procedures for determining when the many examples from the category that come implicit test is compromised by this con- to mind (e.g. Srinivas & Roediger, 1990). scious recollection. Many of these strate- As always, priming in both tests is obtained gies have been described at length elsewhere by comparing performance when a relevant (Roediger & Geraci, 2004; Roediger & Mc- item was studied to when it was not stud- Dermott, 1993) so we discuss them here only ied. The category verification test (e.g., Ten- briefly. penny & Shoben, 1992) is similar to the category production test, except that partici- experimental methods to minimize pants do not have to produce the category contamination exemplar. Instead, they are given the cate- One suggestion has been to give inciden- gory name and a possible exemplar and must tal learning instructions at encoding to try indicate whether or not the item is a member to disguise the fact that participants have of the category (animals: elephant). Priming entered a memory experiment. It may also on this task is measured by examining the help to use several filler tasks between the decrease in reaction time to studied exem- study and test phases of the experiment so plars as compared to non-studied exemplars. that the criterial test seems, to the subjects, Finally, general knowledge tests can func- to be just one more task in a long series. If tion as a conceptual implicit memory test intentional learning instructions are required (e.g. Blaxton, 1989). In this test, participants at encoding, then the implicit test itself can attempt to answer general knowledge ques- be disguised as a filler test before an expected tions (e.g., “What animal did the Carthage- explicit memory test (e.g., Weldon & Roedi- nian general Hannibal use in his attack on ger, 1987). In fact, one can even give an Rome?”). Priming is obtained when partici- example of the expected explicit test (e.g., a pants are more likely to answer the questions recognition test or a cued recall test) before correctly when they have studied the answer encoding, so that participants will be less than when they have not. likely to recognize the implicit test as a memory test and think of it as only another filler task before the explicit test they expect. In The Problem of Contamination addition to including a good cover story, the Implicit memory tests are designed to mea- test list can be constructed such that stud- sure anoetic consciousness, but these tests ied items make up a smaller proportion of can become contaminated by consciously test items than non-studied or filler items, controlled uses of memory. That is, despite and appear later in the list. With this test the test instructions to complete the frag- construction, participants may be less likely ment or answer the question with the first to notice the studied items (but see Challis word that comes to mind, subjects may & Roediger, 1993, for evidence on whether recognize the items they produce as being this factor matters). Finally, there is some from the earlier phase in the experiment, evidence that rapid presentation of the test and they may change their retrieval strategy fragments or stems (for example) helps pro- to attempt explicit recollection. Whether mote performance associated with anoetic contamination should be considered a great consciousness (Weldon, 1993). P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

274 the cambridge handbook of consciousness

methods for detecting autonoetic late the nature of the encoding task (phys- (or noetic) consciousness in ical or semantic processing) to determine implicit tests if the perceptual implicit test is compro- Despite using the recommendations out- mised by explicit processes. If the implicit lined above, it is still possible that subjects test shows a level of processing effect, then will become aware that they encountered one can conclude that the test is contami- the test items recently. There are several nated by explicit recollection; if there is lit- procedures for measuring whether implicit tle or no effect of this powerful variable on tests have been compromised by this level the perceptual implicit test, then it is proba- of awareness. Perhaps the simplest mea- bly a relatively pure measure of priming in an sure is to use a post-test questionnaire to anoetic state (see Roediger, Weldon, Stadler, assess autonoetic and noetic consciousness & Riegler, 1992). Note, however, that this (e.g., Bowers & Schacter, 1990). Many stud- speciﬁc procedure only works for perceptual ies of implicit memory have used this tech- implicit tests, because conceptual implicit nique, and data from these questionnaires tests, by deﬁnition, are sensitive to mean- have permitted the partitioning of subjects ingful processing. Other techniques must be into those who are aware and unaware of used for conceptual tests (e.g., Hashtroudi, the relations between the study and test Ferguson, Rappold, & Cronsniak, 1988). phases; critical dissociations between aware A third procedure for separating con- and unaware participants are sometimes sciously controlled from automatic pro- obtained as a function of independent vari- cesses is the process dissociation procedure ables (e.g., Geraci & Rajaram, 2002). As (Jacoby, 1991). Jacoby argued that attempts an aside, we note that the data from these to isolate pure types of processing (inci- questionnaires may overestimate the level of dental or automatic processing on the one participants’ awareness because (1) partici- hand, and intentional or consciously con- pants may only become aware at the time trolled processing, on the other) are unlikely of the questioning, especially if the ques- to be completely successful even when using tions are leading ones, and (2) participants questionnaires or the retrieval intention- may not have had time or motivation to ality criterion. So, although implicit tests engage in conscious recollection when per- are designed and often assumed to rely on forming the task, even if they did become unconscious automatic processes, they are aware during the test. This latter possibil- not immune to more consciously controlled ity can be thought of as illustrating auto- processes. Similarly, and just as seriously, noetic consciousness occurring after auto- explicit memory tests may be affected by matic retrieval; as noted above, this phe- incidental or automatic retrieval. To address nomenon is called involuntary conscious these issues Jacoby and his colleagues devel- recollection (or memory) and is discussed oped the process dissociation procedure in depth below (Richardson-Klavehn, Gar- (PDP) that incorporates a technique called diner, & Java, 1994). the opposition method (see Jacoby, 1991, A second procedure that has been devel- 1998; Jacoby, Toth, & Yonelinas, 1993). Here oped to assess whether implicit memory we sketch in the logic of the procedure, but tests are compromised by autonoetic con- the method can be a bit tricky to use; per- sciousness is the retrieval intentionality cri- haps the best general “user’s guide” to the terion (Schacter, Bowers, & Booker, 1989). PDP is Jacoby (1998). This procedure is based on the fact that In the PDP technique as applied to explicit tests of memory reliably show cer- implicit memory tests, participants study a tain encoding effects, such as the levels set of material, such as words in a list (often of processing effect (superior memory for under several encoding conditions), and then words processed for meaning as opposed to take one of two types of tests using dif- surface detail). If the criterial test is a percep- ferent retrieval instructions called inclusion tual implicit memory test, one can manipu- and exclusion instructions. The test cues are P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

three forms of consciousness in retrieving memories 275

held constant (e.g., the same word stems tracting performance under the exclusion might be presented on both the inclusion instruction from performance under inclu- and exclusion test). Consider again a word- sion instruction. That is, if Inclusion per- stem completion test (see Jacoby et al., 1993, formance = Probability of retrieval using for an experiment that used this procedure). intentional recollection + Probability of rec- After studying a long list of words such as ollection using automatic retrieval, whereas mercy under various encoding conditions, Exclusion performance = Probability of participants are given the stems of words, recollection using automatic retrieval, then such as “mer ”, that could either be com- the difference between the two reflects the pleted with a studied word on the previ- influence of intentional recollection. ous list (e.g., mercy) or a non-studied (e.g., Probability of recall in the exclusion merit). On a typical cued recall test, peo- condition represents a measure of per- ple would be given a cue and asked to use formance that is driven by incidental or it to remember the studied word. Here, as automatic processes. This automatic use in all explicit memory tests, correct recall of of memory is analogous to implicit mem- the item could be achieved through either ory in that it is the information that intentional recollection of the study episode leaks into memory and affects behavior or by a more automatic process in which the without intention or awareness. However, item pops to mind and is then recognized. several researchers (Richardson-Klavehn & The inclusion test instructions are similar to Gardiner, 1996; Richardson-Klavehn, Gar- those in a typical explicit memory task in diner, & Java, 1994) have suggested that the that participants are asked to respond to the automatic form of memory measured by the cue with an item from the study list; how- PDP may not be completely analogous to ever, if they cannot remember the item, they priming on implicit memory tests because are instructed to guess, so the test includes of involuntary recollection. We turn to that both the product of intentional recollec- issue next. tion and, failing that, incidental or automatic priming due to familiarity. Involuntary Conscious Recollection On an exclusion test, participants are told to respond to the word stem without using a Some researchers have suggested that auto- word from the studied list. So, if mercy comes matic forms of memory measured by the to mind and they recognize it from the list, process dissociation procedure may not be they should not respond with mercy, but completely analogous to priming on implicit they must respond with merit or merchant or memory tests due to the bugaboo of invol- some other word beginning with mer. Now untary conscious recollection (Richardson- participants’ use of conscious recollection Klavehn, et al. (1994); Richardson-Klavehn, opposes their responding with a list word; if et al. 1996). The criticism arises from the they respond with the list word (above the fact that controlled processes and auto- non-studied base rate of producing the list matic processes are often, but not always, word when it has not been recently studied) accompanied by autonoetic consciousness then this effect is due to incidental retrieval and anoetic consciousness, respectively. The that is unopposed by recollection. process dissociation procedure assumes that The logic of the PDP is that inclusion forms of memory that are automatic are performance is driven both by intentional also unconscious. However, it is logically and incidental (or automatic) retrieval, possible that people may vividly remem- whereas exclusion performance is produced ber events after they come to mind spon- only by incidental (automatic) retrieval. If taneously (see the Ebbinghaus quote at the we assume that these processes are inde- beginning of the chapter). A procedure used pendent, then an estimate of intentional to capture this kind of memory experience recollection in a particular condition or a instructs subjects to try not to produce stud- particular participant can be derived by sub- ied items to fit a cue (e.g., a word stem), but P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

276 the cambridge handbook of consciousness

instead to complete the stems with only non- because the PDP overestimates the amount studied words; this is Jacoby’s exclusion test of conscious recollection (by mixing in invol- and embodies the logic of opposition. If the untary conscious recollection). subject produces any studied words under Recently, Kinoshita (2001) has attempted these instructions, the assumption can be to provide a theoretical account of involun- made that the words came to mind automat- tary aware memory. Following Moscovitch’s ically. However, Richardson-Klavehn and his component process model of memory (1992, colleagues altered the test to determine 1994), Kinoshita distinguishes between the whether this spontaneous retrieval is associ- memory systems involved in intentional ated with later awareness. To do this, after retrieval and those involved in awareness the exclusion test they gave subjects an of the past. According to Moscovitch, the opportunity to write the word again next frontal lobes are responsible for our ability to the fragment if they recognize it as hav- to intentionally retrieve the past, whereas ing been in the list earlier. Words that were the medial-temporal lobes are responsible studied and “accidentally” used to complete for binding the features of an event together, the fragments but are then later recognized including time and place information that as having been studied provide a measure of helps define the episode in memory. The involuntary conscious (aware) memory. To idea is that cues at retrieval (either ones the extent that such recognition occurs dur- provided experimentally or those produced ing exclusion tests, the automatic compo- internally) automatically reactivate memo- nent from the PDP may be underestimated. ries, bringing events to mind. (Tulving [1983, Interestingly, this form of memory 1985] referred to this kind of process as appears to be useful in reconciling some con- ecphory). If the subject is in an explicit mem- tradictory results in the literature regard- ory experiment at the time of this ecphoric ing whether cross-modality priming (e.g., process and is by definition required to use the effect of auditory presentation on a the cue to retrieve the past, then volition visual test, Rajaram & Roediger, 1993) results and awareness work together: The subject from explicit memory contamination. Using both intends to retrieve the past and is also the retrieval intentionality criterion, Craik, aware of the past. If, on the other hand, Moscovitch, and McDowd (1994) argued the subject is in an implicit memory exper- that valid cross-modal priming occurs on iment and is not required to intentionally perceptual implicit memory tests and that retrieve the past, then volition and aware- it is not the result of explicit memory pro- ness can either occur together or separately. cesses being used on the implicit test. On Because the medial-temporal lobes bind the other hand, another set of results using episodic information associated with the the process dissociation procedure indicated study context together, this information can that the effect is not associated with auto- become automatically available at retrieval, matic processes (Jacoby et al., 1997). The despite the lack of intention to recall two methods therefore lead to different con- these events. clusions. These paradoxical results can be Our interpretation of this argument is reconciled if it is assumed that the process that if all aspects of the event including the dissociation procedure conflates awareness episodic features of that event (the time and with volition. Using the procedure to study place) are activated, then one may become involuntary conscious recollection outlined aware of the past even in the absence of above, Richardson-Klavehn and Gardiner an intention to retrieve (hence involuntary (1996) showed that cross-modality priming conscious recollection). That is, this process was associated both with awareness and with associated with the medial-temporal lobes automatic retrieval. Cross-modality priming allows for involuntary aware memory. As does occur due to an automatic (priming) Kinoshita suggests, “ . . . this retrieval of a component, and the apparent lack of an trace imbued with consciousness accounts automatic influence using the PDP occurs for the felt experience of remembering, the P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

three forms of consciousness in retrieving memories 277

feeling of reexperiencing the event” (2001, tired, dream, slumber. . . . ) and is called p. 61). the Deese-Roediger-McDermott paradigm Although it is possible for awareness to or DRM (after its originators; Deese, 1959; accompany automatic retrieval, the con- Roediger & McDermott, 1995). The lists are verse is possible as well. Even on a free all associates of one word that is not pre- recall test a person may simply know that sented, sleep in this case, as determined by the retrieved item was presented earlier, word association norms. The finding is that, indicating retrieval accompanied by noetic even on immediate recall tests with warnings consciousness (e.g., Hamilton & Rajaram, against guessing, subjects recall the critical 2003; Tulving, 1985). Similarly, the phe- non-presented words at levels comparable to nomenon of recognition failure of recallable those of words that were presented (Roedi- words (Tulving & Thomson, 1973) indi- ger & McDermott, 1995). The effect also cates that a person can fail to recognize occurs on cued recall tests (e.g., E. Marsh, retrieved items as from the past (an exam- Roediger, & McDermott, 2004). When given ple of anoetic consciousness). The point a recognition test, the subjects falsely rec- is that volition, or intention to retrieve, is ognize the critical item at the same level a separate and orthogonal construct from as the list words. Even more importantly conscious awareness. The thorny thicket of for present purposes, when asked to provide issues surrounding these complex issues of Remember/Know judgments on recognized intention, awareness, and memory perfor- items, subjects judge the critical words like mance are just beginning to be investigated sleep to be remembered just as often as they and understood. do for the list words that were actually studied. The fact that subjects vividly remember these falsely recognized items produces an Illusions of Remembering interesting paradox, because in most recog- and Knowing nition studies (often with unrelated words) false alarms are assigned Know rather than So far we have discussed conscious experi- Remember responses. (After all, if the item ences associated with accurate memories – was never presented, shouldn’t it just be instances when people vividly remember a known because it was familiar? How could past event, when they know that the event subjects remember features associated with has occurred, or when the event (unbe- the moment of occurrence of a word that knownst to them) influences their behavior. was never presented?) In all three cases, we are concerned with the The finding of false recall, false recog- various levels of conscious experience asso- nition, and false remembering using the ciated with memory for events that actually DRM paradigm has been confirmed and occurred. However, one of the most com- studied by many other researchers (e.g., pelling findings from recent studies is that Gallo & Roediger, 2003; Gallo, McDermott, subjects sometimes report vivid conscious Percer, & Roediger, 2001; Neuschatz, Payne, experiences (Remember responses) for Lampinen, & Toglia, 2001). In addition, illu- events that never occurred (e.g., Roediger sory recollection is obtained using other & McDermott, 1995). This phenomenon paradigms that produce high levels of false has been termed false remembering alarms in recognition (Lampinen, Copeland, (Roediger & McDermott, 1995), illusory & Neuschatz, 2001; Miller & Gazzaniga, recollection (Gallo & Roediger, 2003), or 1998) and also in cases of false recall in the phantom recollection (Brainerd, Payne, Loftus misinformation paradigm (Roediger, Wright, & Reyna, 2003). Jacoby, & McDermott, 1996). The paradigm that has been used Remember judgments are sometimes most frequently to study this phenomenon viewed as a purified form of episodic rec- involves having subjects study lists of 15 ollection, so the finding of false Remem- associatively related words (bed, rest, awake ber responses raises the issue of veracity P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

278 the cambridge handbook of consciousness

of Remember judgments. The outcome is Remember/Know judgments. Don’t these perplexing for theories of remembering and results from the DRM paradigm show they knowing (see Rajaram, 1999; Rajaram & are invalid? We believe that such skepticism Roediger, 1997, for further discussion). One is misplaced. The fact that autonoetic con- idea is that people misattribute (Gallo & sciousness is subject to illusions is quite inter- Roediger, 2003) or incorrectly bind (Hicks & esting, but in our opinion does not cast doubt Hancock, 2002) features from studied items on this type of conscious experience. Con- to the related non-studied item (the criti- sider the case of visual perception and our cal lure, in this case). When subjects study conscious experience of seeing the world. lists of words that are associated to the crit- Complex cognitive processes can give rise to ical item, many of the studied words may powerful visual illusions in which our per- spark associative arousal (at either a concepts differ dramatically from the objects in scious or unconscious level, or both). There- the world that give rise to them. Still, no one fore, features from studied events become doubts the conscious experience of seeing bound to the critical item even though it just because what we see sometimes differs is never explicitly presented. Norman and from what we ought to see. Just as errors of Schacter (1997) required subjects to justify perception do not invalidate the notion of their Remember responses in a DRM exper- seeing, so we do not think that errors of rec- iment by providing details they remem- ollection invalidate the concept of remembered. Subjects had no trouble doing so for bering. items such as sleep, and in fact the levels of false remembering were just as high in an instructional condition in which sub- Conclusion jects had to justify responses as in other conditions in which no justification was As noted at the outset of this chapter, required (as is customary in Remember/ the issue of states of consciousness in the Know experiments). study of memory has only recently become A related idea is that reports of illusory an active topic of study. The discussion of recollection are driven in part by accurate mental states associated with various forms episodic memory for the surrounding list of retrieval was avoided through much of context (Geraci & McCabe, 2006). In sup- the history of cognitive psychology, proba- port of this hypothesis, Geraci and McCabe bly because investigators worried about the showed that reports of false Remember legacy of introspection. Introspective studies responses decreased when subjects were of attention and perception conducted early given modified Remember instructions that in the 20th century are today largely consid- did not contain the instruction to use rec- ered blind alleys into which the field was led. ollection of surrounding items. When sub- Yet even Ebbinghaus (1885/1964), the great jects were not instructed to rely on recollec- pioneer who eschewed introspective meth- tion for the surrounding words as a basis for ods in his own work, began his famous book remembering, the magnitude of the illusion with a lucid discussion of mental states dur- decreased (although it did not vanish). These ing retrieval. results suggest that Remember responses to In this chapter, we have described two falsely recognized items are driven partly by empirical movements, with their attendant retrieval of studied items. These findings fur- theoretical frameworks, that have shaped ther highlight the critical role of instructions the recent study of consciousness in rela- in affecting reports of conscious experience, tion to memory. The first breakthrough can which has been a theme of this chapter. be traced to the reports of implicit memory A critic might complain that the fact that in severely amnesic individuals. The disso- subjects can have full-blown recollective ciative phenomena of conscious or explicit experiences of events that never occurred memory and indirect or implicit memory might cast doubt on the utility of studying provided, in retrospect, one important way P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

three forms of consciousness in retrieving memories 279

to characterize two distinct states of con- systematic responsiveness of these measures sciousness associated with retrieval. The to the influence of specific independent and second impetus – this a deliberate effort subject variables. to map the relation between conscious- The study of remembering is in some ness and memory – came from the dis- ways more advanced that the study of tinction Endel Tulving introduced in 1985 knowing, which presents unique challenges. between remembering and knowing. Unlike Knowing is relatively more difficult to com- the explicit/implicit distinction, where the municate in experiments, and the usual tac- former signified conscious memory and the tic of instructions is to define knowing in latter non-conscious memory, the experi- relation to remembering, rather than as an ences of remembering and knowing both experience in its own right. Noetic con- denote conscious memory but of two dif- sciousness can, in the abstract, be defined ferent forms. The two states can be distin- on its own (the conscious state of know- guished by subjects given careful instruc- ing, just as we know the meaning of platy- tions and seem to map onto experiences pus without remembering when and where people have every day. Remembering rep- we first learned about this creature), but in resents the ability to mentally travel across experimental practice has been defined in the temporal continuum of the personal past relation to remembering. This methodolog- with attendant feelings of immediacy and ical challenge of definition through instruc- warmth, whereas knowing represents mem- tions is manifested in our attempts at theory for the past in terms of facts, knowl- oretical interpretations as well. We have edge, or familiarity but without any re- identified these challenges in the section experiencing of the events. on noetic consciousness, and we consider We have used Tulving’s tripartite distinc- these issues to be important topics for future tion among autonoetic, noetic, and anoetic investigation (see also Rajaram, 1999). Bet- states of consciousness to organize some of ter characterization of the nature of knowing the key research findings in memory. This remains an important piece of the puzzle to approach helps us understand the properties solve in our pursuit of relating consciousness of these three states of consciousness in rela- to memory. tion to different forms of memory – remem- The third state of consciousness under bering, knowing, and priming, respectively. consideration here – anoetic consciousness – An interesting observation to emerge from is best described in memory research in this approach is that the experience of terms of priming on implicit memory tests remembering can be documented with con- under conditions when conscious aware- siderable clarity. Even though remember- ness of the study/test relation can be elimi- ing – the mental time travel associated with nated or minimized. Priming has been doc- this form of memory – seems introspective umented most dramatically in individuals and highly personal, it represents a state of with severe anterograde amnesia who show consciousness and a form of memory that intact priming with little or no capacity for neurologically intact subjects can use. The conscious recollection. This non-knowing or experimental work that has been produced non-aware state of consciousness and its using the Remember/Know procedure (and expression in priming on implicit memory related techniques) has resulted in a siz- tests have also been extensively studied in able body of research with consistent and individuals with intact memory. In these lat- replicable effects across laboratories. The ter cases, much effort has been expended unreliability of introspective reports, which to control, minimize, or eliminate the issue undermined certain research programs proof consciously controlled processes affecting mulgated by Wundt and Titchener, does not performance that is supposed to be, in Tul- seem to afflict this modern work, which is ving’s terms, anoetic. The great challenge in robust and replicable. Research on remem- the work reported in this chapter is to sepa- bering, knowing, and priming reveals the rate and study the three states of awareness, P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

280 the cambridge handbook of consciousness

attempts that have been partially successful. Brown, R. (1991). A review of the tip-of-the- Of course, in many situations the subject’s tongue experience. Psychological Bulletin, 109, mind may slip among the various states of 204–223. consciousness in performing tasks, and the Brown, R., & Kulik, J. (1977). Flashbulb memo- challenge would be to chart the ebb and flow ries. Cognition, 5, 73–99. of different states of consciousness during Buchanan, T. W., & Adolphs, R. A. (2002). The memory tasks. role of the human amygdala in emotional mod- In 1885 Ebbinghaus provided examples of ulation of long-term declarative memory. In states of conscious awareness during mem- S. Moore & M. Oaksford (Eds.), Emotional ory retrieval. One hundred years later, Tulv- cognition: From brain to behavior. London: John Benjamins. ing named and delineated a theory of three 1993 states of conscious awareness and, most Challis, B. H., & Roediger, H. L. ( ). The importantly, provided a method by which effect of proportion overlap and repeated testing on primed word fragment completion. they might be studied. We may hope that 2085 Canadian Journal of Experimental Psychology, by great progress will have been made 47, 113–123. in studying consciousness and memory. Our 1990 2006 Christianson, S., & Loftus, E. F. ( ). Some chapter is a progress report (written in ) characteristics of people’s traumatic memories. that marks the steps taken along the first Bulletin of the Psychonomic Society, 28, 195– fifth of the path. Although we as a field have 198. 20 made good progress in the past years, it Church, B. A., & Schacter, D. L. (1994) Percep- is clear to us that great breakthroughs must tual specificity of auditory priming: Implicit lay ahead because we are far from our goal memory for voice intonation and fundamental of understanding the complex relations of frequency. Journal of Experimental Psychology: memory and consciousness. Learning, Memory, and Cognition, 20, 521– 533. Conway, M. A. (1995). Flashbulb memories. Hills- dale, NJ: Erlbaum. References Conway, M. A., & Dewhurst, S. A. (1995). Remembering, familiarity, and source mon- Atkinson, R. C., & Joula, J. F. (1973). Factors influ- itoring. Quarterly Journal of Experimental encing speed and accuracy of word recogni- Psychology: Human Experimental Psychology, tion In S. Kornblum (Ed.), Attention and perfor- 48A, 125–140. mance (Vol. 4,pp.583–612). San Diego: Aca- Conway, M. A., Gardiner, J. M., Perfect, T. J., demic Press. Anderson, S. J., & Cohen, G. M. (1997). Atkinson, R. C., & Joula, J. F. (1974). Search and Changes in memory awareness during learning: decision processes in recognition memory. In The acquisition of knowledge by psychol- D. H. Krantz & R. C. Atkinson (Eds.), Contem- ogy undergraduates. Journal of Experimental porary developments in mathematical psychology: Psychology: General, 126, 393–413. I. Learning, memory, and thinking (p. 299). Craik, F. I. M., & Lockhart, R. S. (1972). Levels of Oxford: W. H. Freeman. processing: A framework for memory research. Blaxton, T. A. (1989). Investing dissociations Journal of Verbal Learning and Verbal Behavior, among memory measures: Support for a 11, 671–684. transfer-appropriate processing framework. Craik, F. I. M., Moscovitch, M., & McDowd, J. M. Journal of Experimental Psychology: Learning, (1994). Contributions of surface and concep- Memory, and Cognition, 15, 657–668. tual information to performance on implicit Bowers, J. S., & Schacter, D. L. (1990). Implicit and explicit memory tasks. Journal of Experi- memory and test awareness. Journal of Experi- mental Psychology: Learning, Memory, and Cog- mental Psychology: Learning, Memory, and Cog- nition, 20, 864–875. nition, 16, 404–416. Deese, J. (1959). On the prediction of occurrence Brainerd, C. J., Payne, D. G., Wright, R., & Reyna, of particular verbal intrusions in immediate V. F. (2003). Phantom recall. Journal of Memory recall. Journal of Experimental Psychology, 58, & Language, 48, 445–467. 17–22. P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

three forms of consciousness in retrieving memories 281

Donaldson, W. (1996). The role of decision pro- memory. Canadian Journal of Experimental Psy- cesses in remembering and knowing. Memory chology, 50, 114–122. & Cognition, 24, 523–533. Gardiner, J. M., Kaminska, Z., Dixon, M., & Java, Dunn, J. C. (2004). Remember-know: A matter R. I. (1996). Repetition of previously novel of confidence. Psychological Review, 111, 524– melodies sometimes increases both remember 542. and know responses in recognition memory. Ebbinghaus, H. (1964). Memory: A contribution Psychonomic Bulletin and Review, 3, 366–371. to experimental psychology (H. A. Ruger & C. E. Gardiner, J. M., & Parkin, A. J. (1990). Atten- Bussenius, Trans.). New York: Dover Publica- tion and recollective experience in recognition tions. (Original work published 1885). memory. Memory & Cognition, 18, 579–583. Ferguson, S. A., Hashtroudi, S., & Johnson, M. K. Gardiner, J. M., Ramponi, C., & Richardson- (1992). Age differences in using source-relevant Klavehn, A. (1998). Experiences of remember- cues. Psychology and Aging, 7, 443–452. ing, knowing, and guessing. Consciousness and Finke, R. A., Johnson, M. K., & Shyi, G. C. (1988). Cognition: An International Journal, 7, 1–26. Memory confusions for real and imagined com- Gardiner, J. M., Ramponi, C., & Richardson- pletions of symmetrical visual patterns. Mem- Klavehn, A. (2002). Recognition memory and ory & Cognition, 16, 133–137. decision processes: A meta-analysis of remem- Gallo, D. A., McDermott, K. B., Percer, J. M., ber, know, and guess responses. Memory, 10, & Roediger, H. L., III. (2001). Modality effects 83–98. in false recall and false recognition. Journal of Gardiner, J. M., & Richardson-Klavehn, A. Experimental Psychology: Learning, Memory, & (2000). Remembering and knowing. In E. Tulv- Cognition, 27, 339–353. ing & F. I. M. Craik (Eds.), Handbook of memory. Gallo, D. A., & Roediger, H. L., III. (2003). The Oxford: Oxford University Press. effects of associations and aging on illusory rec- Gardiner, J. M., Richardson-Klavehn, A., & Ram- ollection. Memory & Cognition, 31, 1036–1044. poni, C., (1997). On reporting recollective Gardiner, J. M. (1988). Functional-aspects of rec- experiences and “direct access to memory sys- ollective experience. Memory & Cognition, 16, tems.” Psychological Science, 8, 391–394, 309–313. Geraci, L., & Franklin, N. (2004). The influence Gardiner, J. M., & Conway, M. A. (1999). Levels of linguistic labels on source monitoring deci- of awareness and varieties of experience. In B. sions. Memory, 12, 571–585. H. Challis & B. M. Velichkovsky (Eds.), Strati- Geraci, L., & McCabe, D. P. (2006). Examin- fication in cognition and consciousness (pp. 237– ing the basis for illusory recollection: The role 254). Amsterdam: John Benjamins. of Remember/Know Instructions. Psychonomic Gardiner, J. M., Gawlick, B., & Richardson- Bulletin & Review, 13, 466–473. Klavehn, A. (1994). Maintenance rehearsal Geraci, L., & Rajaram, S. (2002). The ortho- affects knowing, not remembering; elaborative graphic distinctiveness effect on direct and rehearsal affects remembering, not knowing. indirect tests of memory: Delineating the Psychonomic Bulletin and Review, 1, 107–110. awareness and processing requirements. Jour- Gardiner, J. M., & Gregg, V. H. (1997). Recogni- nal of Memory and Language, 47, 273–291. tion memory with little or no remembering: Geraci, L., & Rajaram, S. (2006). The distinctive- Implications for a detection model. Psycho- ness effect in explicit and implicit memory. In nomic Bulletin and Review, 4, 474–479. R. R. Hunt & J. Worthen, (Eds.), Distinctiveness Gardiner, J. M., & Java, R. I. (1990). Recollective and memory (pp. 211–234). New York: Oxford experience in word and nonword recognition. University Press. Memory & Cognition, 18, 23–30. Graf, P., & Schacter, D. L. (1985). Implicit and Gardiner, J. M., & Java, R. I. (1993). Recognizing explicit memory for new associations in normal and remembering. In A. F. Collins, S. E. Gath- and amnesic subjects. Journal of Experimental ercole, M. A. Conway, & P. E. Morris (Eds.), Psychology: Learning, Memory, and Cognition, Theories of memory (pp. 163–188). Hillsdale, NJ: 11, 501–518. Erlbaum. Graf, P., Squire, L. R., & Mandler, G. (1984). The Gardiner, J. M., Java, R. I., & Richardson- information that amnesic patients do not for- Klavehn, A. (1996). How level of process- get. Journal of Experimental Psychology: Learn- ing really influences awareness in recognition ing, Memory, and Cognition, 10, 164–178. P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

282 the cambridge handbook of consciousness

Gregg, V. H., & Gardiner, J. M. (1994). Recogni- Hunt, R. R. & McDaniel, M. A. (1993). The tion memory and awareness: A large effect of enigma of organization and distinctiveness. the study-test modalities on “know” responses Journal of Memory and Language, 32, 421–445. following a highly perceptual orienting task. Hunt, R. R., & Mitchell, D. B. (1978). Speci- European Journal of Cognitive Psychology, 6, 131– ficity in nonsemantic orienting tasks and dis- 147. tinctive memory traces. Journal of Experimen- Hamilton, M., & Rajaram, S. (2003). States tal Psychology: Human Learning and Memory, of awareness across multiple memory tasks: 4, 121–135. Obtaining a “pure” measure of conscious rec- Hunt, R. R., & Mitchell, D. B. (1982). Indepen- ollection. Acta Psychologica, 112, 43–69. dent effects of semantic and nonsemantic dis- Hamann, S. (2001). Cognitive and neural mecha- tinctiveness. Journal of Experimental Psychology: nisms of emotional memory. Trends in Cognitive Learning, Memory, and Cognition, 8, 81–87. Sciences, 5, 394–400. Hunt, R. R., & Toth, J. P. (1990). Perceptual Hashtroudi, S., Ferguson, S. A., Rappold, V. A., identification, fragment completion, and free & Cronsniak, L. D. (1988). Data-driven and recall: Concepts and data. Journal of Experimen- conceptually driven processes in partial-word tal Psychology: Learning, Memory, and Cogni- identification and recognition. Journal of Exper- tion, 16, 282–290. imental Psychology: Learning, Memory, and Cog- Inoue, C., & Bellezza, F. S. (1998). The detection nition, 14, 749–757. model of recognition using know and remem- Henkel, L. A., & Franklin, N. (1998). Reality mon- ber judgments. Memory & Cognition, 26, 299– itoring of physically similar and conceptually 308. related objects. Memory & Cognition, 26, 659– Jacoby, L. L. (1978). On interpreting the effects 673. of repetition: Solving a problem versus remem- Henkel, L. A., Franklin, N., & Johnson, M. K. bering a solution. Journal of Verbal Learning and (2000). Cross-modal source monitoring confu- Verbal Behavior, 17, 649–667. sions between perceived and imagined events. Jacoby, L. L. (1983a). Perceptual enhancement: Journal of Experimental Psychology: Learning, Persistent effects of an experience. Journal Memory, and Cognition, 26, 321–335. of Experimental Psychology: Learning, Memory, Hicks, J. L., & Hancock, T. (2002). The associ- and Cognition, 9, 21–38. ation between associative strength and source Jacoby, L. L. (1983b). Remembering the data: attributions in false memory. Psychonomic Bul- Analyzing interactive processes in reading. letin & Review, 9, 807–815. Journal of Verbal Learning and Verbal Behavior, Hicks, J. L., & Marsh, R. L. (1999). Remember- 22, 485–508. know judgments can depend on how memory Jacoby, L. L. (1991). A process dissociation is tested. Psychonomic Bulletin & Review, 6, 117– framework: Separating automatic from inten- 122. tional uses of memory. Journal of Memory and Hirshman, E., & Masters, S. (1997). Modeling the Language, 30, 513–541. conscious correlates of recognition memory: Jacoby, L. L. (1998). Invariance in automatic Reflections on the remember-know paradigm. influences of memory: Toward a user’s guide Memory & Cognition, 25, 345–351. for the process-dissociation procedure. Journal Hunt, R. R. (1995). The subtlety of of Experimental Psychology: Learning, Memory, distinctiveness: What von Restorff really & Cognition, 24, 3–26. did. Psychonomic Bulletin & Review, 2 , 105–112. Jacoby, L. L., & Dallas, M. (1981). On the relation- Hunt, R. R. (2006). What is the meaning of ship between autobiographical memory and distinctiveness for memory research? In R. perceptual learning. Journal of Experimental R. Hunt & J. Worthen, (Eds.), Distinctive- Psychology: General, 110, 306–340. ness and memory. Oxford: Oxford University Jacoby, L. L., Jones, T. C. & Dolan, P. O. Press. (1998). Two effects of repetition: Support for Hunt, R. R., & Elliot, J. M. (1980). The role of a dual-process model of knowledge judgments nonsemantic information in memory: Ortho- and exclusion errors. Psychonomic Bulletin & graphic distinctiveness effects on retention. Review, 5, 705–709. Journal of Experimental Psychology: General, Jacoby, L. L., Toth, J. P., & Yonelinas, A. P. 109, 49–74. (1993). Separating conscious and unconscious P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

three forms of consciousness in retrieving memories 283

influences of memory: Measuring recollection. Psychology, Learning, Memory, and Cognition, Journal of Experimental Psychology: General, 17, 210–223. 122, 139–154. Joordens, S., & Merikle, P. M. (1993). Inde- Jacoby, L. L., & Witherspoon, D. (1982). Remem- pendence or redundancy: 2 models of con- bering without awareness. Canadian Journal of scious and unconscious influences. Journal of Psychology, 36, 300–324. Experimental Psychology: General, 122, 462– Jacoby, L. L., Woloshyn, V., & Kelley, C. (1989). 467. Becoming famous without being recognized: Kelley, C. M., & Jacoby, L. L. (1998). Subjec- Unconscious influences of memory produced tive reports and process dissociation: Fluency, by dividing attention. Journal of Experimental knowing, and feeling. Acta Psychologica, 98, Psychology: General, 118, 115–125. 127–140. Jacoby L. L., Yonelinas, A. P., & Jennings, J. Kensinger, E. A., & Corkin, S. (2003). Memory M. (1997). The relation between conscious enhancement for emotional words: Are emo- and unconscious (automatic) influences: A tional words more vividly remembered than declaration of independence. In J. D. Cohen neutral words? Memory & Cognition, 31, 1169– & J. W. Schooler (Eds.), Scientific approaches 1180. to consciousness (pp. 13–47). Hillsdale, NJ: Kinoshita, S. (1995). The word frequency effect in Erlbaum. repetition memory versus repetition priming. Java,R.I.(1994). States of awareness follow- Memory & Cognition, 23, 569–580. ing word stem completion. European Journal Kinoshita, S. (2001). The role of involuntary of Cognitive Psychology, 6, 77–92. aware memory in the implicit stem and frag- Java, R. I., Gregg, V. H., & Gardiner, J. M. ment completion tasks: A selective review. Psy- (1997). What do people actually remember chonomic Bulletin & Review, 8, 58–69. (and know) in “remember/know” experiments? Knowlton, B. J., & Squire, L. R. (1995). Remem- European Journal of Cognitive Psychology, 9, bering and knowing: Two different expressions 187–197. of declarative memory. Journal of Experimental Johnson, M. K. (1988). Discriminating the origin Psychology: Learning, Memory, and Cognition, of information. In T.F. Oltmanns & B. A. Maher 21, 699–710. (Eds.), Delusional beliefs (pp. 34–65). Oxford: Kolers, P. A., & Roediger, H. L., III. (1984). Pro- John Wiley and Sons. cedures of mind. Journal of Verbal Learning and Johnson, M. K., DeLeonardis, D. M., Hashtroudi, Verbal Behavior, 23, 425–449. S., & Ferguson, S. A. (1995). Aging and single Koriat, A. (1995). Dissociating knowing and the versus multiple cues in source monitoring. Psy- feeling of knowing: Further evidence for the chology and Aging, 10, 507–517. accessibility model. Journal of Experimental Johnson, M. K., Foley, M. A., & Leach, K. (1988). Psychology: General, 124, 311–333. The consequences for memory of imagining in Lampinen, J. M., Copeland, S. M., & Neuschatz, another person’s voice. Memory & Cognition, J. S. (2001). Recollections of things schematic: 16, 337–342. Room schemas revisited. Journal of Experimen- Johnson, M. K., Foley, M. A., Suengas, A. G., tal Psychology: Learning, Memory, and Cogni- & Raye, C. L. (1988). Phenomenal character- tion, 27, 1211–1222. istics of memories for perceived and imagined Lindsay, D. S., & Kelley, C. M. (1996). Creating autobiographical events. Journal of Experimen- illusions of familiarity in a cued recall remem- tal Psychology: General, 117, 371–376. ber/know paradigm. Journal of Memory and Johnson, M. K., Hashtroudi, S., & Lindsay, D. S. Language, 35, 197–211. (1993). Source monitoring. Psychological Bul- Madigan, S. (1983). Picture memory. In J. C. letin, 144, 3–28. Yuille (Ed.), Imagery, memory, and cognition: Johnson, M. K., Raye, C. L., Foley, H. J., & Foley, Essays in honor of Allan Paivio (pp. 65–89). M. A. (1981). Cognitive operations and decision Hillsdale, NJ: Erlbaum. bias in reality monitoring. American Journal of Mandler, G. (1980). Recognizing: The judgment Psychology, 94, 37–64. of previous occurrence. Psychological Review, Johnston, W. A., Hawley, K. J., & Elliott, J. M. 87, 252–271. (1991). Contribution of perceptual fluency to Mangels, J. A., Picton, T. W., & Craik, F. I. recognition judgments. Journal of Experimental M. (2001). Attention and successful episodic P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

284 the cambridge handbook of consciousness

encoding: An event-related potential study. news about the Challenger. In E. Winograd & Cognitive Brain Research, 11, 77–95. U. Neisser (Eds.), Affect and accuracy in recall: Mantyla, T. (1997). Recollections of faces: Studies of “flashbulb” memories (pp. 9–31). New Remembering differences and knowing similar- York: Cambridge University Press. ities. Journal of Experimental Psychology: Learn- Neuschatz, J. S., Payne, D. G., Lampinen, J. M., & ing, Memory, and Cognition, 23, 1–14. Toglia, M. P. (2001). Assessing the effectiveness Mantyla, T., & Raudsepp, J. (1996). Recollec- of warnings and the phenomenological charac- tive experience following suppression of focal teristics of false memories. Memory, 9, 53–71. attention. European Journal of Cognitive Psy- Norman, K. A., & Schacter, D. L. (1997). chology, 8, 195–203. False recognition in older and younger adults: Marsh, E. J., Meade, M. L., & Roediger, H. L. Exploring the characteristics of illusory mem- (2003). Learning facts from fiction. Journal of ories. Memory & Cognition, 25, 838–848. Memory and Language, 49, 519–536. Ochsner, K. (2000). Are affective events richly Marsh, E. J., Roediger, H. L., & McDermott, K. B. recollected or simply familiar? The experience (2004). Does test-induced priming play a role and process of recognizing feelings past. Journal in the creation of false memories? Memory, 12, of Experimental Psychology: General, 129, 242– 44–55. 261. Mather, M., Henkel, L. A., & Johnson, M. K. Parkin, A. J., Gardiner, J. M., & Rosser, R. (1995). (1997). Evaluating characteristics of false Functional aspects of recollective experience in memories: Remember/know judgments and face recognition. Consciousness and Cognition: memory characteristics questionnaire com- An International Journal, 4, 387–398. pared. Memory & Cognition, 25, 826–837. Parkin, A. J., & Russo, R. (1990). Implicit and McDaniel, M. A., & Geraci, L. (2006). Encod- explicit memory and the automatic/effortful ing and retrieval processes in distinctiveness distinction. European Journal of Cognitive Psy- effects: Toward an integrative framework. In chology, 2 , 71–80. R. R. Hunt & J. Worthen (Eds.), Distinctiveness Rajaram, S. (1993). Remembering and knowing: and memory (pp. 65–88). Oxford: Oxford Uni- Two means of access to the personal past. Mem- versity Press. ory & Cognition, 21, 89–102. Miller, M. B., & Gazzaniga, M. S. (1998). Creat- Rajaram, S. (1996). Perceptual effects on ing false memories for visual scenes. Neuropsy- remembering: Recollective processes in picture chologia, 36, 513–520. recognition memory. Journal of Experimental Moscovitch, M. (1992). Memory and working- Psychology: Learning, Memory, and Cognition, with-memory: A component process model 22, 365–77. based on modules and central systems. Journal Rajaram, S. (1998). The effects on conceptual of Cognitive Neuroscience, 4, 257–267. salience and conceptual distinctiveness on con- Moscovitch, M. (1994). Memory and working scious recollection. Psychonomic Bulletin & with memory: Evaluation of a component pro- Review, 5, 71–78. cess model and comparisons with other models. Rajaram, S. (1999). Assessing the nature of In D. L. Schacter & E. Tulving (Eds.), Memory retrieval experience: Advances and challenges. systems 1994 (pp. 269–310). Cambridge, MA: In B. H. Challis & B. M. Velichkovsky (Eds.), MIT Press. Stratification in cognition and consciousness (pp. Moscovitch, D. A., & McAndrews, M. P. (2002). 255–275). Amsterdam: John Benjamins. Material-specific deficits in “remembering” in Rajaram, S., & Geraci, L. (2000). Conceptual patients with unilateral temporal lobe epilepsy fluency selectively influences knowing. Journal and excisions. Neuropsychology, 40, 1335–1342. of Experimental Psychology: Learning, Memory, Moscovitch, M., Vriezen, E., & Goshen- and Cognition, 26, 1070–1074. Gottstein, Y. (1993). Implicit tests of mem- Rajaram, S., & Hamilton, M. (2005). Concep- ory in patients with focal lesions or degener- tual processes can enhance both Remember- ative brain disorders. In H. Spinnler & F. Boller ing and Knowing even after delay: Effects (Eds.), Handbook of neuropsychology (Vol. 8,pp. of contextually-varied repetition, meaning- 133–173). Amsterdam: Elsevier. ful encoding, and test delay. Unpublished Neisser, U. & Harsch, N. (1992). Phantom manuscript, Stony Brook University, Stony flashbulbs: False recollections of hearing the Brook, N.Y. P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

three forms of consciousness in retrieving memories 285

Rajaram, S., Hamilton, M., & Bolton, A. (2002). F. Boller & J. Grafman (Eds.), Handbook of neu- Distinguishing states of awareness from confi- ropsychology (Vol. 8,pp.63–131). Amsterdam: dence during retrieval: Evidence from amne- Elsevier. sia. Cognitive, Affective, and Behavioral Neuro- Roediger, H. L., & McDermott, K. B. (1995). science, 2 , 227–235. Creating false memories: Remembering words Rajaram, S. & Roediger, H. L. (1993). Direct com- not presented in lists. Journal of Experimental parison of four implicit memory tests. Journal Psychology: Learning, Memory, & Cognition, 21, of Experimental Psychology: Learning, Memory, 803–814. and Cognition, 19, 765–776. Roediger, H. L., Weldon, M. S., & Challis, B. Rajaram, S. & Roediger, H. L. (1997) Remember- H. (1989). Explaining dissociations between ing and knowing as states of consciousness dur- implicit and explicit measures of retention: ing retrieval. In J. D. Cohen & J. W. Schooler A processing account. In H. L. Roediger & (Eds.), Scientific approaches to consciousness. F. I. M. Craik (Eds.), Varieties of memory and (pp. 213–240). Hillsdale, NJ: Erlbaum. consciousness: Essays in honour of Endel Tulving Richardson-Klavehn, A. & Gardiner, J. M. (2000). (pp. 3–41). Hillsdale, NJ: Erlbaum. Remembering and knowing. In E. Tulving & Roediger, H. L., Weldon, M. S., Stadler, M. L., F. I. M. Craik (Eds.), The Oxford handbook of & Riegler, G. L. (1992). Direct comparison of memory (pp. 229–244). New York, N.Y., U.S.: two implicit memory tests: Word fragment and Oxford University Press. word stem completion. Journal of Experimental Richardson-Klavehn, A., Gardiner, J. M., & Java, Psychology: Learning, Memory, & Cognition, 18, R. I. (1994). Involuntary conscious memory 1251–1269. and the method of opposition. Memory, 2 , 1– Roediger, H. L. III, Wheeler, M. A., & Rajaram, 29. S. (1993). Remembering, knowing, and recon- Richardson-Klavehn, A., Gardiner, J. M, & Java, structing the past. Psychology of Learning and R. I. (1996). Memory: Task dissociations, pro- Motivation: Advances in Research and Theory, cess dissociations and dissociations of con- 30, 97–134. sciousness. In G. D. M. Underwood (Ed.), Rotello, C. M., Macmillan, N. A., & Reeder, J. A. Implicit cognition (pp. 85–158). New York, N.Y., (2004). Sum-difference theory of remember- U.S.: Oxford University Press. ing and knowing: a two-dimensional signal- Roediger, H. L. (1990). Implicit memory: Reten- detection model. Psychological Review, 111, 588– tion without remembering. American Psychol- 616. ogist, 45, 1043–1056. Rozin, P. (1976). The psychobiological approach Roediger, H. L. (1999). Retrieval experience: A to human memory. In M. R. Rosenzweig & E. L. new arena of psychological study. In B. H. Chal- Bennett (Eds.), Neural Mechanisms of Learning lis & B. M. Velichkovsky (Eds.), Stratification and Memory (pp. 3–48). Cambridge, MA: MIT of cognition and consciousness (pp. 229–235). Press. Amsterdam: John Benjamins. Rubin, D. C., & Kozin, M. (1984). Vivid memo- Roediger, H. L., & Blaxton, T. A. (1987). Effects ries. Cognition, 16, 81–95. of varying modality, surface features, and reten- Schacter, D. L. (1987). Implicit memory: His- tion interval on priming in word-fragment tory and current status. Journal of Experimental completion. Memory & Cognition, 15, 379–388. Psychology: Learning, Memory, and Cognition, Roediger, H. L., & Geraci, L. (2004). Conducting 13, 501–518. implicit memory research: A practical guide. In Schacter, D. L., Bowers, J., & Booker, J. (1989). A. Wenzel & D. Rubin (Eds.), A guide to imple- Intention awareness, and implicit memory: menting cognitive methods with clinical popula- The retrieval intentionality criterion. In S. tions. Washington, DC: APA Books. Lewandowsky, J. C. Dunn, et al. (Eds.), Implicit Roediger, H. L., Jacoby, J. D., & McDermott, memory: Theoretical issues (pp. 47–65). Hills- K. B. (1996). Misinformation effects in recall: dale, NJ: Erlbaum. Creating false memories through repeated Schacter, D. L., Chiu, C. Y. P., & Ochsner, K. N. retrieval. Journal of Memory & Language, 35, (1993). Implicit memory: A selective review. 300–318. Annual Review of Neuroscience, 16, 159–182. Roediger, H. L., & McDermott, K. B. (1993). Schacter, D. L., Verfaellie, M., & Pradere, Implicit memory in normal human subjects. In D. (1996). The neuropsychology of memory P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

286 the cambridge handbook of consciousness

illusions: False recall and recognition in tion. Journal of Experimental Psychology: Learn- amnesic patients. Journal of Memory and Lan- ing, Memory, and Cognition, 18, 25–42. guage, 35, 319–334. Toth, J. P. (2000). Nonconscious forms of Schmidt, S. R. (1991). Can we have a distinctive human memory. In E. Tulving & F. I. M. theory of memory? Memory & Cognition, 19, Craik (Eds.), The Oxford handbook of memory 523–542. (pp. 245–261). New York: Oxford University Schmidt, S. (2006). Emotion, significance, dis- Press. tinctiveness, and memory. In R. R. Hunt & Tulving, E. (1983). Elements of episodic memory. J. Worthen (Eds.), Distinctiveness and memory. Oxford: Oxford University Press. Oxford: Oxford University Press. Tulving, E. (1985). Memory and consciousness. Schmolck, H., Buffalo, E. A., & Squire, L. R. Canadian Psychology, 26, 1–12. (2000). Memory distortions develop over time: Tulving, E.. & Schacter, D. L. (1990). Priming Recollections of the O. J. Simpson trial verdict and human memory systems. Science, 247, 301– after 15 and 32 months. Psychological Science, 306. 39 45 11, – . Tulving, E., Schacter, D. L., & Stark, H. A. (1982). Sherman, J. W., & Bessenoff, G. R. (1999). Stereo- Priming effects in word-fragment completion types as source-monitoring cues: On the inter- are independent of recognition memory. Jour- action between episodic and semantic memory. nal of Experimental Psychology: Learning, Mem- Psychological Science, 10, 106–110. ory, and Cognition, 8, 336–342. Shimamura, A. P., & Squire, L. R. (1984). Tulving,E., & Thompson, D. M. (1973). Encoding Paired-associated learning and priming affects specificity and retrieval processes in episodic in amnesia: A neuropsychological study. Jour- memory. Psychological Review, 80, 352–373. nal of Experimental Psychology: General, 113, Verfaellie, M. (1994). A re-examination of 556 570 – . recognition memory in amnesia: Reply to Slamecka, N. J., & Graf, P. (1978). The generation Roediger and McDermott. Neuropsychology, 8, effect: Delineation of a phenomenon. Journal of 289–292. Experimental Psychology: Human Learning and Verfaellie, M., Giovanello, K. S., & Keane, M. 592 604 Memory, 4, – . M. (2001). Recognition memory in amnesia: Snodgrass, J. G., & Corwin, J. (1988). Perceptual Effects of relaxing response criteria. Cogni- identification thresholds for 150 fragmented tive, Affective, and Behavioral Neuroscience, 1, pictures from the Snodgrass and Vanderwart 3–9. 3 picture set. Perceptual and Motor Skills, 67, – Verfaellie, M., & Keane, M. M. (2002). Impaired 36 . and preserved memory processes in amnesia. In Snodgrass, J. G., Smith, B., Feenan, K., & Cor- L. R. Squire & D. L. Schacter (Eds.), Neuropsy- win, J. (1987). Fragmenting pictures on the chology of memory (3d ed., pp. 36–45). New Apple Macintosh computer for experimen- York: Guilford Press. tal and clinical applications. Behavior Research von Restorff, H. (1933). Uber die wirkung von 270 Methods, Instruments and Computers, 19, – Bereichsblidungen im spurenfeld. Psychologishe 274 . Forschung, 18, 299–342. Squire, L. R., Stark, C. E. L., & Clark, R. E. Wagner, A. D., & Gabrieli, J. D. E. (1998). On 2004 ( ). The medial temporal lobe. Annual the relationship between recognition familiar- 27 279 306 Review of Neuroscience, , – . ity and perceptual fluency: Evidence for dis- Srinivas, K., & Roediger, H. L. (1990). Classify- tinct mnemonic processes. Acta Psychologia, ing implicit memory tests: Category association 98, 211–230. and anagram solution. Journal of Memory and Wagner, A. D., Gabrieli, J. D. E., & Verfaellie, Language, 29, 389–412. M. (1997). Dissociations between familiarity Talarico, J. M., & Rubin, D. C. (2003). Con- processes in explicit recognition and implicit fidence, not consistency, characterizes flash- perceptual memory. Journal of Experimental bulb memories. Psychological Science, 14, 455– Psychology: Learning, Memory, and Cognition, 461. 23, 305–323. Tenpenny, P. L., & Shoben, E. J. (1992). Warrington, E. K., & Weiskrantz, L. (1968). A Component processes and the utility of study of learning and retention in amnesic the conceptually-driven/data-driven distinc- patients. Neuropsychologia, 6, 283–291. P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

three forms of consciousness in retrieving memories 287

Warrington, E. K., & Weiskrantz, L. (1970). Yonelinas, A. P. (2001). Consciousness, control, Amnesic syndrome: Consolidation or retrieval? and conﬁdence: The 3 Cs of recognition mem- Nature, 228, 628–630. ory. Journal of Experimental Psychology: Gen- Wegner, D. M. (2002). The illusion of conscious eral, 130, 361–379. will. Cambridge, MA: MIT Press. Yonelinas, A. P. (2002). The nature of recollec- Weldon, M. S. (1993). The time course of per- tion and familiarity: A review of 30 years of ceptual and conceptual contributions to word research. Journal of Memory and Language, 46, fragment completion priming. Journal of Exper- 441–517. imental Psychology: Learning, Memory, and Cog- Yonelinas, A. P., & Jacoby, L. L. (1995). The nition, 19, 1010–1023. relation between remembering and knowing Weldon, M. S., & Roediger, H. L. (1987). Altering as bases for recognition-effects of size congru- retrieval demands reverses the picture superi- ency. Journal of Memory and Language, 34, 622– ority effect. Memory & Cognition, 15, 269–280. 643. Wheeler, M. A., Stuss, D. T., & Tulving,E. (1997). Yonelinas, A. P., Kroll, N. E. A., Dobbins, I., Toward a theory of episodic memory: The Lazzara, M., & Knight, R. T. (1998). Recollec- frontal lobes and autonoetic consciousness. Psy- tion and familiarity deﬁcits in amnesia: Con- chological Bulletin, 121, 331–354. vergence of remember-know, process dissocia- Wilson, T. D. (2002). Strangers to ourselves: Dis- tion, and receiver operating characteristic data. 323 339 covering the adaptive unconscious. Cambridge, Neuropsychology, 12, – . MA: Harvard University Press. Yonelinas, A. P., Kroll, N. E. A., Quamme, J. R., Wixted, J., & Stretch, V. (2004). In defense of Lazzara, M. M., Sauve, M., Widaman, K. F., & 2002 the signal detection interpretation of remem- Knight, R. T. ( ). Effects of extensive tem- ber/know judgments. Psychonomic Bulletin & poral lobe damage or mild hypoxia on recol- Review, 11, 616–641. lection and familiarity. Nature Neuroscience, 5, 1236–1241. Yonelinas, A. P. (1994). Receiver-operating char- 1972 acteristics in recognition memory: evidence for Zechmeister, E. B. ( ). Orthographic dis- a dual-process model. Journal of Experimental tinctiveness as a variable in word recogni- 425 Psychology: Learning, Memory, and Cognition, tion. American Journal of Psychology, 85, – 430 20, 1341–1354. . P1: JzG 0521857430c10 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 24, 2007 9:14

288 P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

CHAPTER 11 Metacognition and Consciousness

Asher Koriat

Abstract reviews research addressing these questions, emphasizing its implications for issues con- The study of metacognition can shed light on cerning consciousness; in particular, the gen- some fundamental issues about conscious- esis of subjective experience, the function of ness and its role in behavior. Metacognition self-reflective consciousness, and the cause- research concerns the processes by which and-effect relation between subjective expe- people self-reflect on their own cognitive rience and behavior. and memory processes (monitoring) and how they put their metaknowledge to use in regulating their information processing and Introduction behavior (control). Experimental research on metacognition has addressed the follow- There has been a surge of interest in ing questions. First, what are the bases of metacognitive processes in recent years, metacognitive judgments that people make with the topic of metacognition pulling in monitoring their learning, remembering, under one roof researchers from tradition- and performance? Second, how valid are ally disparate areas of investigation. These such judgments and what are the factors areas include memory research (Kelley that affect the correspondence between sub- & Jacoby, 1998; Metcalfe & Shimamura, jective and objective indexes of knowing? 1994; Nelson & Narens, 1990; Reder, 1996), Third, what are the processes that underlie developmental psychology (Schneider & the accuracy and inaccuracy of metacogni- Pressley, 1997), social psychology (Bless tive judgments? Fourth, how does the out- & Forgas, 2000; Jost, Kruglanski, & Nel- put of metacognitive monitoring contribute son, 1998; Schwarz, 2004), judgment and to the strategic regulation of learning and decision making (Gilovich, Griffin, & remembering? Finally, how do the metacog- Kahneman, 2002; Winman & Juslin, 2005), nitive processes of monitoring and con- neuropsychology (Shimamura, 2000), trol affect actual performance? This chapter forensic psychology (e.g., Pansky, Koriat, 289 P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

290 the cambridge handbook of consciousness

& Goldsmith, 2005; Perfect, 2002), edu- trol). The object level, in contrast, has no cational psychology (Hacker, Dunlosky, & control over the meta level and no access Graesser, 1998), and problem solving and to it. For example, the study of new mate- creativity (Davidson & Sternberg, 1998; rial involves a variety of basic, object-level Metcalfe, 1998a). The establishment of operations, such as text processing, compre- metacognition as a topic of interest in its hending, rehearsing, and so on. At the same own right is already producing synergy time, metacognitive processes are engaged among different areas of investigation con- in planning how to study, in devising and cerned with monitoring and self-regulation implementing learning strategies, in moni- (e.g. Fernandez-Duque, Baird, & Posner, toring the course and success of object-level 2000). Furthermore, because some of the processes, in modifying them when neces- questions discussed touch upon tradition- sary, and in orchestrating their operation. In ally ostracized issues in psychology, such the course of studying new material, learners as the issues of consciousness and free are assumed to monitor their degree of com- will (see Nelson, 1996), a lively debate prehension online and then decide whether has been going on between metacognitive to go over the studied material once again, researchers and philosophers (see Nelson how to allocate time and effort to different & Rey, 2000). In fact, it appears that the segments, and when to end studying. increased interest in metacognition research We should note, however, that the derives in part from the feeling that perhaps distinction between cognitive and metacog- this research can bring us closer to dealing nitive processes is not sharp because the with (certainly not resolving) some of the same type of cognitive operation may occur metatheoretical issues that have been the at the object level or at the meta level, and province of philosophers of the mind. in some cases it is unclear to which level a particular operation belongs (Brown, 1987). Definition Research Traditions Metacognition concerns the study of what people know about cognition in general, Historically, there have been two main and about their own cognitive and mem- lines of research on metacognition that ory processes, in particular, and how they proceeded almost independently of each put that knowledge to use in regulating their other, one within developmental psychology information processing and behavior. Flavell and the other within experimental memory (1971) introduced the term “metamemory,” research. The work within developmen- which concerns specifically the monitoring tal psychology was spurred by Flavell (see and control of one’s learning and remem- Flavell, 1979; Flavell & Wellman, 1977), who bering. Metamemory is the most researched argued for the critical role that metacog- area in metacognition and is the focus of this nitive processes play in the development chapter. of memory functioning (see Flavell, 1999). Nelson and Narens (1990) proposed Within memory research, the study of a conceptual framework that has been metacognition was pioneered by Hart’s adopted by most researchers. According to (1965) studies on the feeling-of-knowing them, cognitive processes may be divided (FOK), and Brown and McNeill’s (1966) into those that occur at the object level work on the tip-of-the-tongue (TOT). and those that occur at the meta level: The There is a difference in goals and object level includes the basic operations methodological styles between these two traditionally subsumed under the rubric of research traditions. The basic assumption information processing – encoding, rehears- among developmental students of metacog- ing, retrieving, and so on. The meta level is nition is that learning and memory per- assumed to oversee object-level operations formance depend heavily on monitoring (monitoring) and return signals to regulate and regulatory proficiency. This assump- them actively in a top-down fashion (con- tion has resulted in attempts to specify the P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

metacognition and consciousness 291

components of metacognitive abilities, to In addition to the developmental and trace their development with age, and to the experimental-memory lines of research, examine their contribution to memory func- there has been considerable work on tioning. Hence a great deal of the work metacognition in the areas of social psy- is descriptive and correlational (Schneider, chology and judgment and decision mak- 1985). The focus on age differences and indi- ing. Social psychologists have long been vidual differences in metacognitive skills has concerned with questions about metacog- also engendered interest in specifying “definition, although their work has not been ciencies” that are characteristic of children at explicitly defined as metacognitive (see Jost different ages and in devising ways to rem- et al., 1998). In particular, social psychol- edy them. This work has expanded into the ogists share the basic tenets of metacog- educational domain: Because of the increas- nitive research (see below) regarding the ing awareness of the critical contribution of importance of subjective feelings and beliefs, metacognition to successful learning (Paris & as well as the role of top-down regula- Winograd, 1990), educational programs have tion of behavior. In recent years social psy- been developed (see Scheid, 1993) designed chologists have been addressing questions to make the learning process more “metacog- that are at the heart of current research in nitive.” Several authors have stressed specif- metacognition (e.g., Winkielman, Schwarz, ically the importance of metacognition to Fazendeiro, & Reber, 2003; Yzerbyt, Lories, transfer of learning (see De Corte, 2003). & Dardenne, 1998; see Metcalfe, 1998b). The conception of metacognition by Within the area of judgment and decision developmental psychologists is more com- making, a great deal of the work concern- prehensive than that underlying much of ing the calibration of probability judgments the experimental work on metacognition. (Fischhoff, 1975; Lichtenstein, Fischhoff, & It includes a focus on what children know Phillips, 1982; Winman & Juslin, 2005)is about the functioning of memory and par- directly relevant to the issues raised in ticularly about one’s own memory capac- metacognition. ities and limitations. Developmental work has also placed heavy emphasis on strate- Research Questions gies of learning and remembering (Bjorklund & Douglas, 1997; Brown, 1987; Pressley, This chapter emphasizes the work on Borkowski, & Schneider, 1987). In addition, metacognition within the area of adult many of the issues addressed in the area of memory research. It is organized primarily theory of mind (Perner & Lang, 1999) con- around the five main questions that have cern metacognitive processes. These issues been addressed in experimental research on are, perhaps, particularly important for the metamemory. First, what are the bases of understanding of children’s cognition. metacognitive judgments; that is, how do In contrast, the experimental-cognitive we know that we know (e.g., Koriat & Levy- study of metacognition has been driven Sadot, 1999)? Second, how valid are subjec- more by an attempt to clarify basic questions tive intuitions about one’s own knowledge; about the mechanisms underlying monitor- that is, how accurate are metacognitive ing and control processes in adult mem- judgments, and what are the factors that ory (for reviews, see Koriat & Levy-Sadot, affect their accuracy (e.g., Schwartz & 1999; Nelson & Narens, 1990; Schwartz, Metcalfe, 1994)? Third, what are the pro- 1994). This attempt has led to the emergence cesses underlying the accuracy and inaccu- of several theoretical ideas as well as spe- racy of metacognitive judgments? In par- cific experimental paradigms for examining ticular, what are the processes that lead the monitoring and control processes that to illusions of knowing and to dissociations occur during learning, during the attempt between knowing and the feeling of know- to retrieve information from memory, and ing (e.g., Benjamin & Bjork, 1996; Koriat, following the retrieval of candidate answers 1995)? Fourth, what are the processes under- (e.g., Metcalfe, 2000; Schwartz, 2002). lying the strategic regulation of learning P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

292 the cambridge handbook of consciousness

and remembering? In particular, how does the importance of subjective processes and the output of monitoring affect control top-down executive functions (see Koriat, processes (e.g., Barnes, Nelson, Dunlosky, 2000b). The study of metacognition is gen- Mazzoni, & Narens, 1999; Son & Metcalfe, erally predicated on a view of the person as 2000)? Finally, how do the metacognitive an active organism that has at its disposal an processes of monitoring and control affect arsenal of cognitive operations that can be actual memory performance (e.g., Koriat applied at will toward the achievement of & Goldsmith, 1996a; Metcalfe & Kornell, various goals. The strategic choice and reg- 2003)? ulation of these operations are assumed to Although these questions focus on be guided in part by the person’s subjective relatively circumscribed processes of mem- beliefs and subjective feelings. ory and metamemory, they touch upon Embodied in this view are two metathe- some of the issues that are at the heart oretical assumptions (see Koriat, 2002). of the notions of consciousness and self- The first concerns agency – the assumption consciousness. Thus, the study of the sub- that self-controlled processes have measur- jective monitoring of knowledge addresses a able effects on behavior. Although most defining property of consciousness, because researchers would acknowledge that many consciousness implies not only that we cognitive processes, including some that are know something but also that we know subsumed under the rubric of executive that we know it. Thus, consciousness binds function, occur outside of consciousness, together knowledge and metaknowledge there is also a recognition that the per- (Koriat, 2000b). This idea is implied, for son is not a mere medium through which example, in Rosenthal’s (2000) “higher- information flows. Rather, people have some order thought” (HOT) philosophical theory freedom and flexibility in regulating actively of consciousness: A “lower-order” men- their cognitive processes during learning and tal state is conscious by virtue of there remembering. Furthermore, it is assumed being another, higher-order mental state that that such self-regulation processes deserve makes one conscious that one is in the lower- to be studied not only because they can have order state (see Chapter 3). Clearly, the considerable effects on performance but also subjective feelings that accompany cognitive because they are of interest in their own processes constitute an essential ingredient right. of conscious awareness. Rather than taking This assumption presents a dilemma these feelings (and their validity) at their face for experimental researchers because self- value, the study of metacognition attempts controlled processes have been tradition- to uncover the processes that shape subjec- ally assumed to conflict with the desire tive feelings and contribute to their validity of experimenters to exercise strict experi- or to their illusory character. Furthermore, mental control. Of course, there are many the study of monitoring-based control has studies in which learning and remembering implications for the question of the function strategies have been manipulated (through of conscious awareness, and for the benefits instructions) and their effects investigated and perils in using one’s own intuitive feel- (e.g., Craik & Lockhart, 1972). Unlike such ings and subjective experience as a guide to experimenter-induced strategies, however, judgments and behavior. self-initiated strategies generally have been seen as a nuisance factor that should be avoided or neutralized. For example, labo- Basic Assumptions about Agency ratory studies typically use a fixed-rate pre- and Consciousness sentation of items rather than a self-paced The increased interest in metacognition presentation (see Nelson & Leonesio, 1988). seems to reflect a general shift from Also, in measuring memory performance, the stimulus-driven, behavioristic view of sometimes forced-choice tests are preferred the person to a view that acknowledges over free-report tests to avoid having to P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

metacognition and consciousness 293

deal with differences in “guessing,” or else Winkielman & Berridge, 2004). However, some correction for guessing procedure is by and large, much of the experimental used to achieve a pure measure of “true” research in metacognition is predicated on memory (see Koriat & Goldsmith, 1996a; the tacit assumption that the metacognitive Nelson & Narens, 1994). Needless to say, processes studied entail conscious control. people in everyday life have great freedom in Nonetheless, although the term “metacog- regulating their memory processes, and the nition” is generally understood as involving challenge is to find ways to bring these self- conscious awareness, it should be acknowl- controlled metacognitive processes into the edged that monitoring and control pro- laboratory (Koriat, 2000a; Koriat & Gold- cesses can also occur unconsciously (Spehn smith, 1996a). & Reder, 2000). The second assumption concerns the role I now review some of the experimental of self-reflective, subjective experience in work on metamemory, focusing on research guiding controlled processes. This is, of that may have some bearing on general course, a debatable issue. It is one thing to questions about phenomenal experience and equate controlled processes with conscious conscious control. processes (e.g., Posner & Snyder, 1975); it is another to assume that subjective expe- Experimental Paradigms in the Study rience plays a causal role in behavior. Stu- of Online Metamemory dents of metacognition not only place a heavy emphasis on subjective experience A variety of metacognitive judgments have but also assume that subjective feelings, such been studied in recent years that ought to be as the feeling of knowing, are not mere included under the umbrella of metacogni- epiphenomena, but actually exert a causal tion (Metcalfe, 2000). Among these are ease- role on information processing and behavior of-learning judgments (Leonesio & Nelson, (Koriat, 2000b; Nelson, 1996). 1990), judgments of comprehension (Maki A similar growing emphasis on the role of & McGuire, 2002), remember/know judg- subjective feelings in guiding judgments and ments (Gardiner, & Richardson-Klavehn, behavior can be seen in social-psychological 2000), output monitoring (Koriat, Ben-Zur, research (Schwarz & Clore, 2003) and in & Sheffer, 1988), olfactory metacognition decision making (Slovic, Finucane, Peters, & (Jonsson¨ & Olsson, 2003), and source moni- MacGregor, 2002). Also, the work on mem- toring (Johnson, 1997). However, the bulk of ory distortions and false memories brings to the experimental work has concerned three the fore the contribution of phenomenolog- types of judgments. ical aspects of remembering to source mon- First are judgments of learning (JOLs) itoring and reality monitoring (see Kelley & elicited following the study of each item. Jacoby, 1998; Koriat, Goldsmith, & Pansky, For example, after studying each paired- 2000; Mitchell & Johnson, 2000). associate in a list, participants are asked to It should be stressed, however, that not all assess the likelihood that they will be able to students of metacognition subscribe to the recall the target word in response to the cue assumptions discussed above. In particular, in a future test. These item-by-item judg- Reder (1987) has argued that a great deal of ments are then compared to the actual recall strategy selection occurs without conscious performance. deliberation or awareness of the factors that Second are FOK judgments elicited influence one’s choice. Of course, there is following blocked recall. In the Recall- little doubt that many monitoring and con- Judgment-Recognition (RJR) paradigm trol processes occur without consciousness introduced by Hart (1965), participants (Kentridge & Heywood, 2000), so the ques- are required to recall items from memory tion becomes one of terminology, like the (typically, the answers to general knowledge question whether feelings must be conscious questions). When they fail to retrieve or can also be unconscious (Clore, 1994; the answer, they are asked to make FOK P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

294 the cambridge handbook of consciousness

judgments regarding the likelihood that and can even sense its imminent emergence they would be able to select the correct into consciousness. What is peculiar about answer from among several distractors in a this experience is the discrepancy between forced-choice test to be administered later. subjective and objective knowing. So how The validity of FOK judgments is then can people monitor the presence of infor- evaluated by the correspondence between mation in memory despite their failure to these judgments and performance on the retrieve it? In reviewing the verbal learning recognition test. Finally, after retrieving literature more than 30 years ago, Tulving an answer from memory or after selecting and Madigan (1970), in fact, argued that one an answer, the subjective confidence in of the truly unique characteristics of human the correctness of that answer is elicited, memory is its knowledge of its own knowl- typically in the form of a probability judg- edge. They proposed that genuine progress ment reflecting the assessed likelihood that in memory research depends on understand- the answer is correct. Whereas JOLs and ing how the memory system not only can FOK judgments are prospective, involving produce a learned response or retrieve an predictions of future memory performance, image but also can estimate rather accurately confidence judgments are retrospective, the likelihood of its success in doing it. A involving assessments about a memory that great deal of research conducted since 1970 has been produced. has addressed this question. Many different variations of these general paradigms have been explored, including The Direct-Access View variations in the type of memory studied (semantic, episodic, autobiographical, A simple answer to the question about eyewitness-type events, etc.), the format of the basis of feelings of knowing is pro- the memory test (free recall, cued recall, vided by the direct-access view according to forced-choice recognition, etc.), and the which people have direct access to mem- particular judgments elicited (item-by-item ory traces both during learning and during judgments or global judgments, using a prob- remembering and can base their metacog- ability or a rating scale, etc.). nitive judgments on detecting the presence and/or the strength of these traces. For example, in the case of JOLs elicited during study, it may be proposed that learners How Do We Know That We Know? can detect directly the memory trace that is The Bases of Metacognitive Judgments formed following learning and can also monitor online the increase in trace strength that As we see later, metacognitive judgments occurs in the course of study as more time is are accurate by and large. JOLs made for spent studying an item (e.g., Cohen, Sandler, different items during study are generally & Keglevich, 1991). Of course, to the extent predictive of the accuracy of recalling these that learners can do so, they can also decide items at test. FOK judgments elicited fol- to stop studying (under self-paced condi- lowing blocked recall predict the likelihood tions) when trace strength has reached a of recalling or recognizing the elusive target desirable value (Dunlosky & Hertzog, 1998). at some later time, and subjective confidence A direct-access account has also been in the correctness of an answer is typically advanced by Hart (1965) with regard to diagnostic of the accuracy of that answer. FOK. Hart proposed that FOK judgments Thus, the first question that emerges is, How represent the output of an internal moni- do we know that we know? tor that can survey the contents of mem- This question emerges most sharply with ory and can determine whether the trace regard to the tip-of-the-tongue (TOT) state, of a solicited memory target exists in store. in which we fail to recall a word or a name, Thus, the feeling associated with the TOT and yet we are convinced that we know it state may be assumed to stem from direct, P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

metacognition and consciousness 295

privileged access to the memory trace of degree of validity in predicting objective the elusive target (see also Burke, MacKay, memory performance (Benjamin & Bjork, Worthley, & Wade, 1991; Yaniv & Meyer, 1996). To the extent that such indeed is 1987). Hart stressed the functional value of the case, then the accuracy of metacogni- having such a monitor, given the general fal- tive judgments is not guaranteed, but should libility of the memory system: If the mon- depend on the validity of the cues on which itor “signals that an item is not in storage, it rests. then the system will not continue to expend Inferential, cue-utilization accounts gen- useless effort and time at retrieval; instead, erally distinguish between information- input can be sought that will put the item based (or theory-based) and experience- into storage” (Hart, 1965;p.214). based metacognitive judgments (see Kelley Direct-access (or trace-access) accounts, & Jacoby, 1996a; Koriat & Levy-Sadot, which assume that monitoring involves a 1999; Matvey, Dunlosky, & Guttentag, 2001; direct readout of information that appears Strack, 1992). This distinction parallels a dis- in a ready-made format, have two merits. tinction between two modes of thought that The first is that they can explain not only has been proposed in other domains (see the basis of JOLs and FOK judgments but Kahneman, 2003, and see further below). also their accuracy. Clearly, if JOLs are based Thus, it is assumed that metacognitive judg- on accessing the strength of the memory ments may be based either on a deliber- trace that is formed following learning, then ate use of beliefs and memories to reach an they ought to be predictive of future recall, educated guess about one’s competence and which is also assumed to depend on memory cognitions, or on the application of heuristics strength. Similarly, if FOK judgments moni- that result in a sheer subjective feeling. tor the presence of the memory trace of the unrecalled item, they should be expected to Theory-Based Monitoring predict the future recognition or recall of that item. Consider first theory-based metacognitive The second merit is that they would judgments. Developmental students of cog- seem to capture the phenomenal quality of nition placed a great deal of empha- metacognitive feelings: the subjective feel- sis on what Flavell called “metacognitive ing, such as that which accompanies the knowledge;” that is, on children’s beliefs tip-of-the-tongue state, that one monitors and intuitions about their own memory directly the presence of the elusive target in capacities and limitations and about the memory and its emergence into conscious- factors that contribute to memory perfor- ness (James, 1890). In fact, metacognitive mance (Brown, 1987). Such beliefs have feelings are associated with a sense of self- been found to affect the choice of learning evidence, which gives the impression that strategies, as well as people’s predictions of people are in direct contact with the con- their own memory performance (see Flavell, tents of their memories and that their intro- 1999; Schneider & Pressley, 1997). spections are inherently accurate. In contrast, the experimental research on adult metacognition contains only scattered references to the possible contribution of The Cue-Utilization View of theories and beliefs to metacognitive judg- Metacognitive Judgments ments. For example, in discussing the bases Although the direct-access view has not of JOLs, Koriat (1997) proposed to distin- been entirely abandoned (see Burke et al., guish between two classes of cues for theory- 1991; Metcalfe, 2000), an alternative view based online JOLs, intrinsic and extrinsic. has been gaining impetus in recent years. The former includes cues pertaining to the According to this view, metacognitive judg- perceived a priori difficulty of the studied ments are inferential in origin, based on a items (e.g., Rabinowitz, Ackerman, Craik, variety of cues and heuristics that have some & Hinchley, 1982). Such cues seem to affect P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

296 the cambridge handbook of consciousness

JOLs, particularly during the first study trial, the same test; those who believed that the as suggested by the observation that norma- test measured abstract reasoning ability (on tive ratings of ease of learning are predic- which they had rated themselves highly) tive both of JOLs and of recall of different estimated that they had achieved higher items (e.g., Koriat, 1997; Leonesio & Nelson, scores than did those who thought that 1990; Underwood, 1966). The second class they had taken a computer programming includes extrinsic factors that pertain either test. This was so despite the fact that the to the conditions of learning (e.g., number two groups did not differ in their actual of times an item has been presented, pre- performance. sentation time, etc., Mazzoni, Cornoldi, & Another finding that points to the effects Marchitelli, 1990; Zechmeister & Shaugh- of one’s a priori beliefs comes from stud- nessy, 1980) or to the encoding operations ies of the relationship between confidence applied by the learner (e.g., level of pro- and accuracy. People’s confidence in their cessing, interactive imagery, etc.; Begg, Vin- responses is generally predictive of the accu- ski, Frankovich, & Holgate, 1991; Matvey racy of these responses in the case of general et al., 2001; Rabinowitz et al., 1982; Shaw knowledge questions but not in the case of & Craik, 1989). For example, participants’ eyewitness memory (Perfect, 2002). Perfect JOLs seem to draw on the belief that gen- (2004) provided evidence that this occurs erating a word is better for memory than because people’s confidence is based in part reading it (Begg et al., 1991; Matvey et al., on their preconceptions about their abilities. 2001). Koriat (1997) proposed that JOLs are Such preconceptions are generally valid in comparative in nature. Hence, they should the case of general knowledge questions, for be more sensitive to intrinsic cues pertaining which people have had considerable feed- to the relative recallability of different items back and hence know their relative standing. within a list than to factors that affect over- Such is not the case with eyewitness mem- all performance (see Begg, Duft, Lalonde, ory, for which they lack knowledge about Melnick, & Sanvito, 1989; Carroll, Nelson, & how good they are and, by implication, how Kirwan, 1997; Shaw & Craik, 1989). Indeed, confident they ought to be. Thus, people’s he obtained evidence indicating that, in confidence in their performance seems to be making JOLs, the effects of extrinsic factors based in part on their preconceived beliefs are discounted relative to those of intrinsic about their own competence in the domain factors that differentiate between different of knowledge tested. items within a list. Evidence for the effects of beliefs and Another major determinant of people’s theories also comes from studies of correc- metacognitive judgments is their perceived tion processes in judgment. People often self-efficacy (Bandura, 1977). In fact, peo- base their judgments directly on their sub- ple’s preconceived notions about their skills jective feelings (see Schwarz & Clore, 1996; in specific domains predict their assessment Slovic et al., 2002). However, when they of how well they did on a particular task. realize that their subjective experience has For example, when students are asked to been contaminated by irrelevant factors, tell how well they have done on an exam, they may try to correct their judgments they tend to overestimate greatly their per- according to their beliefs about how these formance on the test, and this bias derives judgments had been affected by the irrel- in part from the tendency of people to evant factors (Strack, 1992). For example, base their retrospective assessments on their in the study of Schwarz, Bless, Strack, preconceived, inflated beliefs about their Klumpp, Rittenauer-Schatka, and Simons skills in the domain tested, rather than on (1991), participants who were asked to their specific experience with taking the recall many past episodes demonstrating test (Dunning, Johnson, Ehrlinger, & Kruger, self-assertiveness reported lower self-ratings 2003). In a study by Ehrlinger and Dun- of assertiveness than those who were asked ning (2003), two groups of participants took to recall a few such episodes, presumably P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

metacognition and consciousness 297

because of the greater difficulty experienced tors. Brainerd, Reyna, Wright, and Mojardin in recalling many episodes. However, when (2003) also discussed a process termed “rec- led to believe that the experienced difficulty ollection rejection” in which a distractor that had been caused by background music, paris consistent with the gist of a presented item ticipants relied more heavily on the retrieved may be rejected when the verbatim trace of content, reporting higher ratings under the that item is accessed. However, they argued many-episodes condition than under the that this process can occur automatically, few-episodes condition. These and other outside conscious awareness. findings suggest that the correction pro- The evidence reviewed thus far supports cess is guided by the person’s beliefs about the idea that metacognitive judgments may the factors that make subjective experi- be based on one’s beliefs and theories. For ence an unrepresentative basis for judgment. example, the subjective confidence in the Although most researchers assume that the correctness of one’s memory product (e.g., a correction process requires some degree of selected answer in a quiz) can be based on a awareness (see Gilbert, 2002), others sug- logical, analytic process in which one evalu- gest that it may also occur unconsciously ates and weighs the pros and cons (Gigeren- (Oppenheimer, 2004). zer, Hoffrage & Kleinbolting,¨ 1991; Koriat, More recent work in social cognition (see Lichtenstein, & Fischhoff, 1980). FOK judg- Schwarz, 2004) suggests that the conclu- ments, too, may draw on theories or beliefs sions that people draw from their metacog- resulting in an educated guess about the like- nitive experience, such as the experience lihood of retrieving or recognizing an elusive of fluent processing, depend on the na¨ıve word in the future (Costermans, Lories, & theory that they bring to bear. Further- Ansay, 1992). Such judgments may not be more, people can be induced to adopt qualitatively different from many predic- opposite theories about the implications of tions that people make in everyday life. processing fluency, and these theories modulate experience-based judgments. These sug- Experience-Based Monitoring gestions deserve exploration with regard to judgments of one’s own knowledge. Experience-based metacognitive judgments, Another line of evidence comes from in contrast, are assumed to entail a quali- studies that examined how people deter- tatively different process from that undermine that a certain event did not happen. lying theory-based judgments. Consider, for Strack and Bless (1994) proposed that deci- example, the TOT experience. The strong sions of nonoccurrence may be based on conviction that one knows the elusive target a metacognitive strategy that is used when is based on a sheer subjective feeling. That rememberers fail to retrieve any feature of feeling, however, appears to be the prod- a target event that they have judged to uct of an inferential process that involves be highly memorable. In contrast, in the the application of nonanalytic heuristics (see absence of a clear recollection of a non- Jacoby & Brooks, 1984; Kelley & Jacoby, memorable event, people may infer that 1996a; Koriat & Levy-Sadot, 1999) that oper- the event had actually occurred (but had ate below full consciousness and give rise to a been forgotten). Indeed, non-occurrence sheer subjective experience. Indeed, the idea decisions are made with strong confidence that subjective experience can be influenced for events that would be expected to be and shaped by unconscious inferential pro- remembered (e.g., one’s name, a salient cesses has received support in the work of item, etc.; Brown, Lewis, & Monk, 1977; Jacoby, Kelley, Whittlesea, and their asso- Ghetti, 2003). On the other hand, study- ciates (see Kelley & Jacoby, 1998; Whittle- ing material under conditions unfavorable sea, 2004). Koriat (1993) argued that the for learning (or expecting fast forgetting, nonanalytic, unconscious basis of metacog- Ghetti, 2003) results in a relatively high rate nitive judgments is responsible for the phe- of false alarms for non-memorable distrac- nomenal quality of the feeling of knowing P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

298 the cambridge handbook of consciousness

as representing an immediate, unexplained learning or on the ease with which they intuition, similar to that which is associ- are retrieved. Both of these types of cues ated with the experience of perceiving (see become available in the course of learning Kahneman, 2003). According to this view, and disclose the memorability of the stud- sheer subjective experience, which lies at ied material. Such cues have been assumed the core of conscious awareness, is in fact to give rise to a sheer feeling of knowing. the end product of processes that lie below Indeed, there is evidence suggesting that awareness. JOLs monitor the ease with which stud- Several cues have been proposed as deter- ied items are processed during encoding minants of JOL, FOK, and subjective confi- (Begg, et al., 1989; Koriat, 1997; Matvey dence. These cues have been referred to col- et al., 2001). For example, Begg et al. (1989) lectively as “mnemonic” cues (Koriat, 1997). reported results suggesting that JOLs are With regard to JOLs and FOK, these cues sensitive to several attributes of words (e.g., include the ease or fluency of processing of concreteness-abstractness) that affect ease a presented item (Begg et al., 1989), the of processing. Other findings suggest that familiarity of the cue that serves to probe JOLs are affected by the ease and probabil- memory (Metcalfe, Schwartz, & Joaquim, ity with which the to-be-remembered items 1993; Reder & Ritter, 1992; Reder & Schunn, are retrieved during learning (Benjamin & 1996), the accessibility of pertinent partial Bjork, 1996; Benjamin, Bjork, & Schwartz, information about a solicited memory tar- 1998; Koriat & Ma’ayan, 2005). For example, get (Dunlosky & Nelson, 1992; Koriat, 1993; Hertzog, Dunlosky, Robinson, and Kidder Morris, 1990), and the ease with which (2003) reported that JOLs increased with information comes to mind (Kelley & Lind- the speed with which an interactive image say, 1993; Koriat, 1993; Mazzoni & Nelson, was formed between the cue and the target 1995). Subjective confidence in the correct- in a paired-associates task. Similarly, Matvey ness of retrieved information has also been et al. (2001) found that JOLs increased with claimed to rest on the ease with which infor- increasing speed of generating the targets mation is accessed and on the effort experi- to the cues at study. These results are con- enced in reaching a decision (Kelley & Lind- sistent with the view that JOLs are based say, 1993; Nelson & Narens, 1990; Robinson on mnemonic cues pertaining to the fluency & Johnson, 1998; Zakay & Tuvia, 1998). of encoding or retrieving to-be-remembered These cues differ in quality from items during study. those underlying theory-based judgments. With regard to FOK judgments, sev- Whereas the latter judgments draw upon eral heuristic-based accounts have been pro- the content of domain-specific beliefs and posed. According to the cue familiarity knowledge that are retrieved from memory, account, first advanced by Reder (1987; see the former rely on contentless mnemonic also Metcalfe et al., 1993), FOK is based on cues that pertain to the quality of process- the familiarity of the pointer (e.g., the ques- ing, in particular, the fluency with which tion, the cue term in a paired-associate, etc., information is encoded and retrieved. As see Koriat & Lieblich, 1977) that serves to Koriat and Levy-Sadot (1999) argued, “The probe memory (Reder, 1987). Reder argued cues for feelings of knowing, judgments of that a fast, preretrieval FOK is routinely and learning or subjective confidence lie in struc- automatically made in response to the famil- tural aspects of the information processing iarity of the terms of a memory question system. This system, so to speak, engages in to determine whether the solicited answer a self-reflective inspection of its own oper- exists in memory. This preliminary FOK ation and uses the ensuing information as a can guide the question answering strategy. basis for metacognitive judgments” (p. 496). Indeed, the latency of speeded FOK judg- Consider experience-based JOLs. These ments was found to be shorter than that of have been claimed to rely on the ease providing an answer. Furthermore, in sev- with which the items are encoded during eral studies, the advance priming of the P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

metacognition and consciousness 299

terms of a question was found to enhance attributes, and so on (see Koriat, Levy-Sadot, speeded, preliminary FOK judgments with- Edry, & de Marcas, 2003; Miozzo & Cara- out correspondingly increasing the proba- mazza, 1997). These partial clues may give bility of recall or recognition of the answer rise to a sheer feeling that one knows the (Reder, 1987, 1988). Schwartz and Metcalfe answer. An important assumption of the acc- (1992) extended Reder’s paradigm to show essibility account is that participants have no that cue priming also enhances (unspeeded) direct access to the accuracy of the partial FOK judgments elicited following recall clues that come to mind, and therefore both failure. Additional evidence for the cue- correct and wrong partial clues contribute to familiarity account comes from studies using the FOK. a proactive-interference paradigm (Metcalfe Support for the accessibility account et al., 1993). Remarkable support was also comes from a study on the TOT state (Koriat obtained using arithmetic problems: When & Lieblich, 1977). An analysis of the ques- participants made fast judgments whether tions that tend to induce an overly high they knew the answer to an arithmetic prob- FOK suggested that the critical factor is lem and could retrieve it, or whether they the amount of information they tend to had to compute it, Know judgments were elicit. For example, questions that contain found to increase with increasing frequency redundancies and repetitions tend to proof previous exposures to the same parts of duce inflated feelings of knowing, and so are the problem, not with the availability of the questions that activate many “neighboring” answer in memory (Reder & Ritter, 1992). answers. Thus, accessibility would seem to This was true even when participants did be a global, unrefined heuristic that responds not have enough time to retrieve an answer to the mere amount of information irrespec- (Schunn, Reder, Nhouyvanisvong, Richards, tive of its correctness. Because people can & Stroffolino, 1997; see Nhouyvanisvong & rarely specify the source of partial informa- Reder, 1998, for a review). tion, they can hardly escape the contami- Consistent with the cue-familiarity nating effects of irrelevant clues by attribut- account are also the results of studies of ing them to their source. Such irrelevant the feeling-of-not-knowing. Glucksberg and clues sometimes precipitate a strong illusion McCloskey (1981) and Klin, Guzman, and of knowing (Koriat, 1995, 1998a) or even Levine (1997) reported results suggesting an illusory TOT state – reporting a TOT that lack of familiarity can serve as a basis for state even in response to questions that have determining that something is not known. no real answers (Schwartz, 1998), possibly Increasing the familiarity of questions because of the activations that they evoke. for which participants did not know the Indeed, Schwartz and Smith (1997) answer increased the latency of Don’t Know observed that the probability of reporting responses as well as the tendency to make a a TOT state about the name of a ficti- Know response erroneously. tious animal increased with the amount According to the accessibility account of of information provided about that animal, FOK, in contrast, FOK is based on the overall even when the amount of information did accessibility of pertinent information regard- not contribute to the probability of recalling the solicited target (Koriat, 1993). This ing the name of the animal. In addition, account assumes that monitoring does not FOK judgments following a commission precede retrieval but follows it: It is by try- error (producing a wrong answer) are higher ing to retrieve a target from memory that than following an omission error (Koriat, a person can appreciate whether the target 1995; Krinsky & Nelson, 1985; Nelson & is “there” and worth continuing to search Narens, 1990), suggesting that FOK judg- for. This occurs because, even when retrieval ments are sensitive to the mere accessibility fails, people may still access a variety of of information. partial clues and activations, such as frag- In Koriat’s (1993) study, after participants ments of the target, semantic and episodic studied a nonsense string, they attempted to P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

300 the cambridge handbook of consciousness

recall as many of the letters as they could cate that FOK decreases with the number and then provided FOK judgments regard- of pre-experimental, neighboring concepts ing the probability of recognizing the cor- that are linked to a cue, suggesting that these rect string among lures. The more letters judgments are sensitive to the competition that participants could access, the stronger between the activated elements. was their FOK regardless of the accuracy Subjective confidence in the correctness of their recall. When the number of letters of one’s answers has also been assumed to accessed was held constant, FOK judgments rest sometimes on mnemonic cues deriv- also increased with the ease with which ing from the process of recalling or select- information came to mind, as indexed by ing an answer. Thus, people express stronger recall latency. confidence in the answers that they retrieve If both correct and incorrect partial infor- more quickly, whether those answers are cor- mation contribute equally to the feeling rect or incorrect (Nelson & Narens, 1990). that one knows the elusive memory target, Similarly, in a study by Kelley and Lindsay how is it that people can nevertheless mon- (1993), retrieval fluency was manipulated itor their knowledge accurately? Accord- through priming. Participants were asked to ing to Koriat (1993) this happens because answer general information questions and to much of the information that comes sponta- indicate their confidence in the correctness neously to mind (around 90%; see Koriat & of their answers. Prior to this task, partici- Goldsmith, 1996a) is correct. Therefore, the pants were asked to read a series of words, total amount of partial information accessi- some of which were correct answers and ble is a good cue for recalling or recogniz- some were plausible but incorrect answers ing the correct target. Thus, the accuracy of to the questions. This prior exposure was metamemory is a byproduct of the accuracy found to increase the speed and probabil- of memory: Memory is by and large accu- ity with which those answers were pro- rate in the sense that what comes to mind is vided in the recall test and, in parallel, to much more likely to be correct than wrong. enhance the confidence in the correctness A third account still assumes a combined of those answers. Importantly, these effects operation of the familiarity and accessibility were observed for both correct and incorrect heuristics. According to this account both answers. These results support the view that heuristics contribute to FOK, but whereas retrospective confidence is based in part on a the effects of familiarity occur early in simple heuristic: Answers that come to mind the microgenesis of FOK judgments, those easily are more likely to be correct than those of accessibility occur later, and only when that take longer to retrieve. cue familiarity is sufficiently high to drive The imagination inflation effect also illus- the interrogation of memory for potential trates the heuristic basis of confidence judg- answers (Koriat, & Levy-Sadot, 2001; Ver- ments. Asking participants to imagine some non & Usher, 2003). This account assumes childhood events increased confidence that that familiarity, in addition to affecting FOK these events did indeed happen in the judgments directly, also serves as a gating past (Garry, Manning, Loftus, & Sherman, mechanism: When familiarity is high, par- 1996). Merely asking about the event twice ticipants probe their memory for the answer, also increased subjective confidence. Possi- and then the amount of information acces- bly imagination of an event and attempt- sible affects memory performance. When ing to recall it increase its retrieval fluency, familiarity is low, the effects of potential which in turn contributes to the confidence accessibility on FOK are more limited. that the event has occurred (see also Hastie, It should be noted, however, that results Landsman & Loftus, 1978). obtained by Schreiber and Nelson (1998) In sum, although metacognitive judg- question the idea that FOK judgments are ments may be based on explicit inferences sensitive to the mere accessibility of partial that draw upon a priori beliefs and knowl- clues about the target. These results indi- edge, much of the recent evidence points to P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

metacognition and consciousness 301

the heuristic basis of such judgments, sug- access to memory traces comes from obser- gesting that feelings of knowing are based vations documenting a dissociation between on the application of nonanalytic heuris- subjective and objective indexes of know- tics that operate below conscious awareness. ing. Several such dissociations have been These heuristics rely on mnemonic cues reported. These dissociations also bring to pertaining to the quality of processing and the fore the effects of specific mnemonic result in a sheer noetic experience. Thus, it cues on metacognitive judgments. would seem that sheer subjective feelings, With regard to JOLs, Begg et al. (1989) such as the feeling of knowing, which are found that high-frequency words, presum- at the core of subjective awareness, are the ably fluently processed, yielded higher JOLs product of unconscious processes (Koriat, but poorer recognition memory than low- 2000b). frequency words (see also Benjamin, 2003). The distinction between information- Narens, Jameson and Lee (1994) reported based and experience-based processes has that subthreshold target priming enhanced important implications that extend beyond JOLs, perhaps because it facilitated the pro- metacognition. It shares some features with cessing of the target, although it did not the old distinction between reason and emo- affect eventual recall. tion (see Damasio, 1994), but differs from Bjork (1999) described several conditions it. It implies a separation between two com- of learning that enhance performance dur- ponents or states of consciousness – on the ing learning but impair long-term reten- one hand, sheer subjective feelings and intu- tion and/or transfer. According to Bjork itions that have a perceptual-like quality and Bjork (1992), these manipulations facil- and, on the other hand, reasoned cognitions itate “retrieval strength” but not “storage that are grounded in a network of beliefs and strength.” As a result, the learners, fooled explicit memories. It is a distinction between by their own performance during learning, what one “feels” and “senses” and what one may experience an illusion of competence, “knows” or “thinks.” The extensive research resulting in inflated predictions about their in both cognitive psychology and social psy- future performance. For example, massed chology (e.g., Jacoby & Whitehouse, 1989; practice typically yields better performance Strack, 1992) indicates that these two com- than spaced practice in the short term, ponents of conscious awareness are not only whereas spaced practice yields considerably dissociable, but may actually conflict with better performance than massed practice in each other, pulling judgments and behavior the long term. Massed practice, then, has in opposite directions (Denes-Raj & Epstein, the potential of leading learners to over- 1994). The conflict between these compo- estimate their future performance. Indeed, nents is best illustrated in correction phe- Zechmeister and Shaughnessy (1980) found nomena (e.g., Jacoby & Whitehouse, 1989; that words presented twice produced higher Strack, 1992), which suggest that when JOLs when their presentation was massed people realize that their subjective experi- than when it was distributed, although the ence has been contaminated, they tend to reverse pattern was observed for recall. A change their judgments so as to correct for similar pattern was reported by Simon and the assumed effects of that contamination Bjork (2001) using a motor-learning task: (Strack, 1992). Participants asked to learn each of several movement patterns under blocked conditions predicted better performance than Dissociations between Knowing when those patterns were learned under and the Feeling of Knowing random (interleaved) conditions, whereas actual performance exhibited the opposite The clearest evidence in support of the idea pattern. that metacognitive judgments are based on Benjamin et al. (1998) reported sev- inference from cues rather than on direct eral experiments documenting a negative P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

302 the cambridge handbook of consciousness

relation between recall predictions and reviewed above. These findings indicate that actual recall performance, presumably deriv- manipulations that enhance the familiarity ing from reliance on retrieval fluency when of the terms of a question enhance FOK retrieval fluency was a misleading cue for judgments associated with that question future recall. For example, they had partici- without correspondingly affecting actual pants answer general information questions recall performance. A similar dissociation, and assess the likelihood that they would be inspired by the accessibility account, has able to free recall each answer in a later test. been demonstrated by Koriat (1995): The The more rapidly participants retrieved an results of that study suggest that FOK judg- answer to a question, the higher was their ments for general information questions estimate that they would be able to free tend to be accurate as long as these ques- recall that answer at a later time. In reality, tions bring to mind more correct than incor- however, the opposite was the case. rect partial information. However, deceptive Another type of dissociation was reported questions (Fischhoff, Slovic, & Lichtenstein, by Koriat, Bjork, Sheffer, and Bar (2004). 1977), which bring to mind more incorrect They speculated that, to the extent that than correct information, produce unduly JOLs are based on processing fluency at the high FOK judgments following recall failure time of study, they should be insensitive to and, in fact, yield a dissociation to the extent the expected time of testing. This should that FOK judgments are negatively corre- be the case because the processing fluency lated with subsequent recognition memory of an item at the time of encoding should performance. not be affected by when testing is expected. With regard to confidence judgments, Indeed, when participants made JOLs for Chandler (1994) presented participants with tests that were expected either immediately a series of target and non-target stimuli, each after study, a day after study, or a week after consisting of a scenic nature picture. In a study, JOLs were entirely indifferent to the subsequent recognition memory test, a dis- expected retention interval, although actual sociation was observed such that targets for recall exhibited a typical forgetting func- which there existed a similar stimulus in the tion. This pattern resulted in a dissociation non-target series were recognized less often, such that predicted recall matched actual but were endorsed with stronger confidence recall very closely for immediate testing. For than targets for which no similar non-target a week’s delay, however, participants pre- counterpart was included. Thus, seeing a dicted over 50% recall, whereas actual recall related target seems to impair memory while was less than 20%. enhancing confidence. That study also demonstrated the impor- Busey, Tunnicliff, Loftus, and Loftus tance of distinguishing between experience- (2000) had participants study a series of based and theory-based JOLs: When a new faces appearing at different luminance con- group of participants were presented with ditions. For faces that had been studied in a all three retention intervals and asked to dim condition, testing in a bright condition estimate how many words they would recall reduced recognition accuracy, but increased at each interval, their estimates closely confidence, possibly because it enhanced mimicked the forgetting function exhib- their fluent processing during testing. ited by the first group’s actual recall. Thus, In sum, several researchers, motivated the effects of forgetting on recall per- by the cue-utilization view of metacogni- formance seem to emerge under condi- tive judgments, have deliberately searched tions that activate participants’ beliefs about for conditions that produce a dissociation memory. between memory and metamemory. Inter- Dissociations have also been reported estingly, all of the manipulations explored between FOK judgments and actual mem- act in one direction: inflating metacogni- ory performance. First are the findings tive judgments relative to actual memory in support of the cue-familiarity account performance. Some of the experimental P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

metacognition and consciousness 303

conditions found to engender illusions of tion or bias can also be assessed by elicit- knowing are ecologically unrepresentative, ing global or aggregate predictions (Hertzog, even contrived. However, the demonstrated Kidder, Powell-Moman, & Dunlosky 2002; dissociations clearly speak against the notion Koriat, Sheffer, & Ma’ayan, 2002; Liberman, that metacognitive judgments rest on privi- 2004), for example, by asking participants to leged access to the contents of one’s own estimate how many answers they got right memory. and comparing that estimate to the actual number of correct answers. It should be stressed that calibration can The Validity of Metacognitive be evaluated only when judgments and per- Judgments formance are measured on equivalent scales. Thus, for example, if confidence judgments How valid are subjective feelings of know- are made on a rating scale, calibration can- ing in monitoring actual knowledge? How not be evaluated unless some assumptions accurate are people’s introspections about are made (e.g., Mazzoni & Nelson, 1995). their memory? Earlier research has sought to Such is not the case for the second establish a correspondence between know- aspect of metacognitive accuracy, resolution ing and the feeling of knowing as an attempt (or relative accuracy). Resolution refers to to support the trace-access view of metacog- the extent to which metacognitive judg- nitive judgments. Later studies, in contrast, ments are correlated with memory perfor- inspired by the inferential view, have con- mance across items. This aspect is com- centrated on producing evidence for mis- monly indexed by a within-subject gamma correspondence and dissociation, as just correlation between judgments and perfor- reviewed. Although the conditions used in mance (Nelson, 1984). For example, in the these studies may not be ecologically rep- case of JOLs and FOK judgments, resolu- resentative, the results nevertheless suggest tion reflects the extent to which a partic- that the accuracy of metacognitive judg- ipant can discriminate between items that ments is limited. Furthermore, these results she will recall and those that she will not. In point to the need to clarify the reasons for the case of confidence, it reflects the ability accuracy and inaccuracy and to specify the to discriminate between correct and incor- conditions that affect the degree of corre- rect answers. spondence between subjective and objective The distinction between calibration and measures of knowing. resolution is important. For example, in Two aspects of metacognitive accuracy monitoring one’s own competence dur- must be distinguished. The first is calibra- ing the preparation for an exam, calibration (Lichtenstein et al., 1982) or “bias” or tion is pertinent to the decision when to “absolute accuracy” (see Nelson & Dunlosky, stop studying: Overconfidence may lead to 1991), which refers to the correspondence spending less time and effort than are actu- between mean metacognitive judgments ally needed. Resolution, in turn, is rele- and mean actual memory performance and vant to the decision how to allocate the reflects the extent to which metacogni- time between different parts of the mate- tive judgments are realistic. For example, if rial. Importantly, resolution can be high, confidence judgments are elicited in terms even perfect, when calibration is very poor. of probabilities, then the mean probability Also, calibration and resolution may be assigned to all the answers in a list is com- affected differentially. For example, Koriat pared to the proportion of correct answers. et al. (2002) observed that practice study- This comparison can indicate whether ing the same list of items improves res- probability judgments are well calibrated olution but impairs calibration, instilling or whether they disclose an overconfidence underconfidence. bias (inflated confidence relative to perfor- We should note that much of the exper- mance) or an underconfidence bias. Calibra- imental work on the accuracy of JOLs and P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

304 the cambridge handbook of consciousness

FOK judgments has focused on resolution. the JOL-recall gamma correlation averaged In contrast, research on confidence judg- .54 across several studies that used lists of ments, primarily the work carried out within paired-associates that included related and the judgment and decision tradition, has unrelated pairs (Koriat et al., 2002). In con- concentrated on calibration. trast, in Dunlosky and Nelson’s (1994) study, With regard to JOLs elicited during study, in which all pairs were unrelated, the gamma the results of several investigations indi- correlation averaged .20. cate that by and large item-by-item JOLs Monitoring seems to be particularly poor are well calibrated on the first study-test when it concerns one’s own actions. When trial (e.g., Dunlosky & Nelson, 1994; Maz- participants are asked to perform a series zoni & Nelson, 1995). Judgments of com- of minitasks (so called self-performed tasks) prehension, in contrast, tend to be very and to judge the likelihood of recalling these inflated. One reason for this is that in mon- tasks in the future, the accuracy of their pre- itoring comprehension people assess famil- dictions is poor, and much lower than that iarity with the general domain of the text for the study of a list of words (Cohen et al., instead of assessing knowledge gained from 1991). It has been argued that people some- that text (Glenberg, Sanocki, Epstein, & times have special difficulties in monitoring Morris, 1987). their own actions (e.g., Koriat, Ben-Zur, & Two interesting trends have been Druch, 1991). reported with regard to the calibration of However, two types of procedures have JOLs. First is the aggregate effect. When been found to improve JOL resolution. The learners are asked to provide an aggregate first procedure is repeated practice study- judgment (i.e., predict how many items ing the same list of items. As noted ear- they will recall), their estimates, when lier, although repeated practice impairs cal- transformed into percentages, are substan- ibration, it does improve resolution (King, tially lower than item-by-item judgments. Zechmeister, & Shaughnessy, 1980; Koriat, Whereas the latter judgments tend to be 2002; Mazzoni et al., 1990). Thus, in Koriat relatively well calibrated or even slightly et al.’s (2002) analysis, in which the JOL- inflated, aggregate judgments tend to yield recall gamma correlation averaged .54 for underconfidence (Koriat et al., 2002, 2004; the first study-test cycle, that correlation Mazzoni & Nelson, 1995). A similar effect reached .82 on the third study-test cycle. has been observed for confidence judgments Koriat (1997) produced evidence suggesting (Griffin & Tversky, 1992). that the improved resolution with practice Second is the underconfidence-with- occurs because (a) with increased practice practice (UWP) effect (Koriat et al., 2002): studying a list of items, the basis of JOLs When learners are presented with the same changes from reliance on pre-experimental list of items for several study-test cycles, intrinsic attributes of the items (e.g., per- their JOLs exhibit relatively good calibration ceived difficulty) toward a greater reliance on the first cycle, with a tendency toward on mnemonic cues (e.g., processing fluency) overconfidence. However, a shift toward associated with the study of these items, marked underconfidence occurs from the and (b) mnemonic cues tend to have greater second cycle on. The UWP effect was found validity than intrinsic cues, being sensitive to to be very robust across several experimen- the immediate processing of the items dur- tal manipulations and was obtained even for ing study. Rawson, Dunlosky, and Thiede a task involving the monitoring of memory (2000) also observed an improvement in for self-performed tasks. judgments of comprehension with repeated Turning next to resolution, the within- reading trials. person correlation between JOLs and sub- A second procedure that proved effec- sequent memory performance tends to be tive in improving JOL accuracy is that relatively low, particularly when the stud- of soliciting JOLs not immediately after ied material is homogeneous. For example, studying each item, but a few trials later. P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

metacognition and consciousness 305

In paired-associate learning, delaying JOLs crepancy between the standard conditions has been found to enhance JOL accuracy of learning and the standard conditions of markedly (Dunlosky & Nelson, 1994; Nelson testing. On a typical memory test, people & Dunlosky, 1991). However, the delayed- are presented with a question and are asked JOL effect occurs only when JOLs are cued to produce the answer. In contrast, in the by the stimulus term of a paired-associate, corresponding learning condition, both the not when cued by an intact stimulus- question and the answer generally appear response pair (Dunlosky & Nelson, 1992). in conjunction, meaning that the assessment It would seem that the condition in which of one’s future memory performance occurs JOLs are delayed and cued by the stimu- in the presence of the answer. This differ- lus alone approximates the eventual crite- ence has the potential of creating unduly rion test, which requires access to informa- high feelings of competence that derive from tion in long-term memory in response to the failure to discount what one now knows. a cue. Indeed, Nelson, Narens, and Dun- This situation is similar to what has been losky (2004) reported evidence suggesting referred to as the “curse of knowledge” – that, in making delayed JOLs, learners rely the difficulty in discounting one’s privileged heavily on the accessibility of the target, knowledge in judging what a more igno- which is an effective predictor of subsequent rant other knows (Birch & Bloom, 2003). recall. When JOLs are solicited immediately Koriat and Bjork produced evidence suggest- after study, the target is practically always ing that learners are particularly prone to a retrievable, and hence its accessibility has foresight bias in paired-associate cue-target little diagnostic value. There is still contro- learning when the target (present during versy, however, whether the delayed-JOL study) brings to the fore aspects of the cue effect indeed reflects improved metamem- that are less apparent when the cue is later ory (Dunlosky & Nelson, 1992) or improved presented alone (at test). Subsequent exper- memory (Kimball & Metcalfe, 2003; Spell- iments (Koriat & Bjork, 2006) indicated man & Bjork, 1992). that foresight bias, and associated overconfi- Koriat and Ma’ayan (2005) reported evidence, can be alleviated by conditions that dence suggesting that the basis of JOLs enhance learners’ sensitivity to mnemonic changes with delay: As the solicitation of cues that pertain to the testing situation, JOLs is increasingly delayed, a shift occurs in including study-test experience, particularly the basis of JOLs from reliance on encoding test experience, and delaying JOLs. fluency (the ease with which an item is com- Another way in which JOLs can be made mitted to memory) toward greater reliance more sensitive to the processes that affect on retrieval fluency (the ease with which the performance during testing was explored by target comes to mind in response to the cue). Guttentag and Carroll (1998) and Benjamin In parallel, the validity of retrieval fluency (2003). They obtained the typical result in in predicting recall increases with delay and which learners predict superior recognition becomes much better than that of encoding memory performance for common than for fluency. These results suggest that metacog- uncommon words (although in reality the nitive judgments may be based on the flex- opposite is the case). However, when dur- ible and adaptive utilization of different ing the recognition test learners made post- mnemonic cues according to their relative dictions about the words that they could validity in predicting memory performance. not remember (i.e., judged the likelihood The results of Koriat and Ma’ayan suggest that they would have recognized the word that repeated practice and delay may con- if they had studied it), they actually post- tribute to JOL accuracy by helping learn- dicted superior recognition of the uncom- ers overcome biases that are inherent in mon words. Furthermore, the act of making encoding fluency. Koriat and Bjork (2005) postdictions for one list of items was found described an illusion of competence – fore- to rectify predictions made for a second list sight bias – that arises from an inherent dis- of items studied later. P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

306 the cambridge handbook of consciousness

As far as the accuracy of FOK judgments probability judgments markedly exceeding is concerned, these judgments are relatively the proportion of correct answers (Lichten- well calibrated (Koriat, 1993) and are mod- stein et al., 1982). This overconfidence has erately predictive of future recall and recog- been claimed to derive from a confirmation nition. Thus, participants unable to retrieve bias (see Koriat et al., 1980; Nickerson, 1998; a solicited item from memory can estimate Trope & Liberman, 1996) – the tendency to with above-chance success whether they build toward a conclusion that has already will be able to recall it in the future, produce been reached by selectively gathering or uti- it in response to clues, or identify it among lizing evidence that supports that conclu- distractors (e.g., Gruneberg & Monks, 1974; sion. However, it has also been argued that Hart, 1967). In a meta-analysis, Schwartz part of the observed overconfidence may and Metcalfe (1994) found that the accuracy be due to the biased sampling of items by of FOK judgments in predicting subsequent researchers – the tendency to include too recognition performance increases with the many deceptive items. Indeed, when items number of test alternatives. The highest cor- are drawn randomly, the overconfidence bias relations were found when the criterion test decreases or disappears (Gigerenzer et al., was recall. 1991). Assuming that metacognitive judgments More recently, attempts have been made are based on internal, mnemonic cues, then to show that confidence in a decision is their accuracy should depend on the validity based on the sampling of events from mem- of the cues on which they rest. However, ory, with overconfidence resulting from a only a few studies examined the validity biased sampling (Winman & Juslin, 2005). of the mnemonic cues that are assumed Indeed, Fiedler and his associates (Fiedler, to underlie FOK judgments. Koriat (1993) Brinkmann, Betsch, & Wild, 2000; Freytag showed that the correlation between the & Fiedler, 2006) used a sampling approach amount of partial information retrieved to explain several biases in judgment and about a memory target (regardless of its decision making in terms of the notion accuracy) is a good predictor of eventual of metacognitive myopia. According to this memory performance, and its validity is approach, many environmental entities have equal to that of FOK judgments. Whereas to be inferred from the information given in the overall accessibility of information about a sample of stimulus input. Because sam- a target (inferred from the responses of one ples are rarely representative, an impor- group of participants) predicted the mag- tant metacognitive requirement would be to nitude of FOK judgments following recall monitor sampling biases and control for failure, the output-bound accuracy of that them. People’s responses, however, are finely information was predictive of the accuracy tuned to the information given in the (resolution) of these FOK judgments sample, and biased judgments, including (Koriat, 1995). In a similar manner, cue overconfidence, derive from the failure to familiarity may contribute to the accuracy consider the constraints imposed on the gen- of FOK judgments because in the real world eration of the information sample. cues and targets (or questions and answers) It is important to note that over- typically occur in tight conjunction; there- confidence is not ubiquitous: When it fore familiarity with the clue should predict comes to sensory discriminations, partici- familiarity with the target (Metcalfe, 2000). pants exhibit underconfidence, thinking that Turning finally to retrospective confi- they did worse than they actually did dence judgments, these have received a great (Bjorkman,¨ Juslin, & Winman, 1993). Also, deal of research in the area of judgment whereas item-by-item confidence judg- and decision making. When participants are ments yield overconfidence, aggregate (or presented with general knowledge questions global) judgments (estimating the num- and are asked to assess the probability that ber of correct answers), as noted earlier, the chosen answer is correct, an overconfi- typically yield underconfidence (Gigerenzer dence bias is typically observed, with mean et al., 1991; Griffin & Tversky, 1992). The P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

metacognition and consciousness 307

underconfidence for aggregate judgments level of variability among witnesses in exper- may derive in part from a failure to make an imental laboratory studies. Such studies typ- allowance for correct answers likely to result ically maintain the same conditions across from mere guessing (Liberman, 2004). participants. In contrast, under naturalistic A great deal of research has been car- conditions the correlation is generally much ried out also on the confidence-accuracy higher, and it is that type of correlation that (C-A) relation, with variable results. The would seem to be of relevance in a foren- general pattern that emerges from this sic context (Lindsay, Read & Sharma, 1998). research is that the C-A relation is quite A second reason, mentioned earlier, is that strong when calculated within each partic- retrospective confidence judgments tend to ipant (which is what was referred to as res- be based in part on participants’ precon- olution), but very weak when calculated ceptions about their ability in the domain between participants (see Perfect, 2004). tested, and these preconceptions tend to be Consider the latter situation first. Research of low validity when they concern eyewit- conducted in the domain of eyewitness tes- ness memory (e.g., lineup identification). timony, focusing on the ability of partic- Several studies explored the subjective ipants to recall a particular detail from a mnemonic cues that may mediate the crime or to identify the perpetrator in a within-person C-A correlation. These cues lineup, has yielded low C-A correlations include retrieval latency and the perception (Wells & Murray, 1984). That research has of effortless retrieval. The correlation was typically focused on a between-individual higher for recall than for recognition pre- analysis, which is, perhaps, particularly rele- sumably because recall provides more cues vant in a forensic context: It is important to pertaining to ease of retrieval than recogni- know whether eyewitnesses can be trusted tion (Koriat & Goldsmith, 1996a; Robinson, better when they are confident in the tes- Johnson, & Herndon, 1997). Robinson, timony than when they express low con- Johnson, & Robertson (2000) found that rat- fidence. Similarly, if there are several wit- ings of vividness and detail for a videotaped nesses, it is important to know whether event contributed more strongly to confi- the more confident among them is likely to dence judgments than processing fluency be the more accurate. Thus, in this context and were also more diagnostic of memory the general finding is that a person’s confi- accuracy. Attempts to enhance the C-A rela- dence in his or her memory is a poor prediction in eyewitness identification by inducing tor of the accuracy of that memory. greater awareness of the thoughts and rea- On the other hand, research focusing on soning process involved in the decision pro- within-person variation has typically yielded cess have been largely ineffective or even moderate-to-high C-A correlations. Thus, counterproductive (Robinson & Johnson, when participants answer a number of ques- 1998). tions and for each question report their con- In sum, the accuracy of metacognitive fidence in the correctness of the answer, judgments has attracted a great deal of inter- the cross-item correlation between confi- est because of its theoretical and practical dence and accuracy tends to be relatively implications. The results are quite variable, high (e.g. Koriat & Goldsmith, 1996a). The although by and large JOLs, FOK judg- same is true when the questions concern ments, and confidence ratings are moder- the episodic memory for a previously expe- ately predictive of item differences in actual rienced event (Koriat, Goldsmith, Schnei- memory performance. der, & Nakash-Dura, 2001). Thus, people can generally discriminate between answers (or memory reports) that are likely to be correct The Control Function and those that are likely to be false. of Metacognition Why are the between-participant correlations very low? Several studies suggest that As noted earlier, much of the work in meta- these low correlations stem from the low cognition is predicated on the assumption P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

308 the cambridge handbook of consciousness

that consciousness is not a mere epiphe- First, Thiede and Dunlosky (1999) showed nomenon. Rather, subjective feelings and that when learners are presented with an subjective judgments exert a causal role on easy goal (e.g., to learn a list of 30 items behavior. In metacognition research this idea with the aim of recalling at least 10 of them), has been expressed in terms of the hypoth- they tended to choose the easier rather than esis that monitoring affects control (Nelson, the more difficult items for restudy. Thiede 1996). Indeed, several observations suggest a and Dunlosky took these results to imply a causal link between monitoring and control hierarchy of control levels: At a superordi- so that the output of monitoring serves to nate level, learners may plan to invest more guide the regulation of control processes. effort studying either the easier or the more With regard to the online regulation of difficult items. This strategy is then imple- learning, it has been proposed that JOLs mented at the subordinate level to control affect the choice of which items to relearn the amount of time allocated to each item and how much time to allocate to each item. and to select items for restudy. Indeed, it has been observed that under self- Second, Son and Metcalfe (2000) had par- paced conditions, when learners are given ticipants learn relatively difficult materials the freedom to regulate the amount of time with the option to go back to materials that spent on each item, they tend to allocate had previously been studied. Under high more time to items that are judged to be dif- time pressure, participants allocated more ficult to learn than to those that are judged study time to materials that were judged as to be easier (for a review see Son & Met- easy and interesting. When the time pres- calfe, 2000). It was proposed that the effects sure was not so great, however, they tended of item difficulty on study time allocation to focus on the more difficult items. are mediated by a monitoring process in These results indicate that study time which learners judge the difficulty of each allocation is also affected by factors other item and then invest more effort in studying than the output of online monitoring. the judged-difficult items to compensate for Indeed, other studies indicated, for exam- their difficulty (Nelson & Leonesio, 1988). ple, that learners invest more study time Dunlosky and Hertzog (1998; see also when they expect a recall test than when Thiede & Dunlosky, 1999) proposed a they expect a recognition test (Mazzoni & discrepancy-reduction model to describe the Cornoldi, 1993) and when the instructions relation between JOLs and study time allo- stress memory accuracy than when they cation. Learners are assumed to monitor stress speed of learning (Nelson & Leonesio, online the increase in encoding strength that 1988). Also, the allocation of study time to occurs as more time is spent studying an item a given item varies according to the incen- and to cease study when a desired level of tive for subsequently recalling that item and strength has been reached. This level, which according to the expected likelihood that is referred to as “norm of study” (Le Ny, the item would be later tested (Dunlosky & Denhiere, & Le Taillanter, 1972), is preset Thiede, 1998). on the basis of various motivational factors, Altogether, these results suggest that such as the stress on accurate learning ver- study time allocation is guided by an adap- sus fast learning (Nelson & Leonesio, 1988). tive strategy designed to minimize effort and Thus, in self-paced learning, study continues improve learning. until the perceived degree of learning meets With regard to FOK judgments, several or exceeds the norm of study. studies indicated that they predict how long In their review of the literature, Son and people continue searching for a memory tar- Metcalfe (2000) found that indeed, in 35 get before giving up: When people feel that of 46 published experimental conditions, they know the answer or that the answer is learners exhibited a clear preference for on the tip-of-the-tongue, they search longer studying the more difficult materials. There than when they feel that they do not know are two exceptions to this rule, however. the answer (Barnes et al., 1999; Costermans P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

metacognition and consciousness 309

et al., 1992; Gruneberg, Monks, & Sykes, be expected to rate their performance on a 1977; Schwartz, 2001). FOK judgments are quiz of scientific reasoning lower than men also predictive of the speed of retrieving rate themselves. Such was indeed the case, an answer, so that in the case of commis- although in reality there was no gender dif- sion responses the correlation between FOK ference in actual performance. When asked judgments and retrieval latency is positive, later if they would like to participate in a sci- whereas for omission responses the corre- ence competition, women were more likely lation between FOK and the latency of to decline, and their reluctance correlated the decision to end search is negative (see significantly with their self-rated perfor- Nelson & Narens, 1990). mance on the quiz. Thus, their choices were Search time is also affected by other fac- affected by their confidence even when con- tors in addition to FOK judgments: When fidence was unrelated to actual performance. participants are penalized for slow respond- A systematic examination of the con- ing, they tend to retrieve answers faster trol function of confidence judgments was but produce more incorrect answers (Barnes conducted by Koriat and Goldsmith (1994, et al., 1999). 1996a,b) in their investigation of the strate- As noted earlier, Reder (1987) proposed gic regulation of memory accuracy. Con- that preliminary FOK judgments also guide sider the situation of a person on the witness the selection of strategies for solving prob- stand who is sworn to “tell the whole truth lems and answering questions. In her stud- and nothing but the truth.” To meet this ies, the decision whether to retrieve a solu- requirement, that person should monitor tion to an arithmetic problem (Know) or to the accuracy of every piece of information compute it was affected by manipulations that comes to mind before deciding whether assumed to influence cue familiarity. These to report it or not. Koriat and Goldsmith studies suggest that FOK judgments that are proposed a model that describes the mon- misled by cue familiarity can misguide the itoring and control processes involved. The decision to retrieve or compute the answer. rememberer is assumed to monitor the sub- Retrospective monitoring can also affect jective likelihood that each candidate mem- behavior. When people make an error in per- ory response is correct and then compare forming a task they can detect that without that likelihood to a preset threshold on the an external feedback and can often imme- monitoring output to determine whether diately correct their response. Following the to volunteer that response or not. The set- detection of an error, people tend to adjust ting of the control threshold depends on the their speed of responding to achieve a desir- relative utility of providing as complete a able level of accuracy (Rabbit, 1966). report as possible versus as accurate a report Confidence judgments have also been as possible. Several results provided consis- shown to affect choice and behavior and tent support for this model. First, the ten- do so irrespective of their accuracy. As dency to report an answer was very strongly noted earlier, people are often overconfident correlated with subjective confidence in the in their knowledge. Fischhoff et al. (1977) correctness of the answer (the intra-subject showed that people had sufficient faith in gamma correlations averaged more than .95; their confidence judgments that they were Koriat & Goldsmith, 1996b, Experiment 1; willing to stake money on their validity. see also Kelley & Sahakyan, 2003). This Consider the finding, mentioned earlier, result suggests that people rely completely that when judging how well they have done on their subjective confidence in deciding on a test, people tend to base their judg- whether to volunteer an answer or with- ments on their preconceptions about their hold it. In fact, participants were found to abilities in the domain tested. Ehrlinger rely heavily on their subjective confidence and Dunning (2003) reasoned that because even when answering a set of “deceptive” women tend to perceive themselves as less general knowledge questions, for which sub- scientifically talented than men, they should jective confidence was quite undiagnostic P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

310 the cambridge handbook of consciousness

of accuracy (Koriat & Goldsmith, 1996b, Sahakyan, 2003; Pansky, Koriat, Goldsmith, Experiment 2). Second, participants given a & Pearlman-Avnion, 2002) were found to be high accuracy incentive (e.g., “you win one less effective than young adults (college stu- point for each correct answer but lose all dents) in utilizing the option to withhold of your winnings if even a single answer is answers to enhance their accuracy. These incorrect”) adopted a stricter criterion than results have implications for the dependabil- participants given a more moderate incen- ity of children’s testimony in legal settings. tive (a 1:1 penalty-to-bonus ratio), suggest- Interestingly, results suggest that the rela- ing that the strategic regulation of memory tionship between monitoring and control, reporting is flexibly adapted to the empha- what Koriat and Goldsmith (1996b) termed sis on memory accuracy. Third, the option “control sensitivity,” may be impaired to to volunteer or withhold responses (which some extent in aging (Pansky et al., 2002) is often denied in traditional memory exper- and in certain psychotic disorders, such as iments) allowed participants to boost the schizophrenia (Danion, Gokalsing, Robert, accuracy of what they reported, in compar- Massin-Krauss, & Bacon, 2001; Koren et al., ison with a forced-report test. This increase 2004). In the Koren et al. study, for instance, occurred by sacrificing some of the cor- the correlation between confidence judg- rect answers; that is, at the expense of ments in the correctness of a response and memory quantity performance. This implies the decision to volunteer or withhold that that eyewitnesses generally cannot “tell the response was highly diagnostic of the degree whole truth” and also “tell nothing but the of insight and awareness that schizophrenic truth,” but must find a compromise between patients showed concerning their mental the two requirements. Importantly, how- condition – more so than traditional mea- ever, the extent of the quantity-accuracy sures of executive control, such as the Wis- tradeoff was shown to depend critically on consin Card Sorting Task. Patients exhibit- monitoring effectiveness: In fact, when mon- ing low control sensitivity were also less able itoring resolution is very high (that is, when a to improve the accuracy of their responses person can accurately discriminate between when given the option to choose which correct and wrong answers), the accuracy of answers to volunteer and which to withhold. what is reported may be improved signifi- The research reviewed above has direct cantly under free report conditions at little bearing on the question of how people can or no cost in quantity performance. Thus, avoid false memories and overcome the con- in the extreme case when monitoring is per- taminating effects of undesirable influences. fect, a person should be able to exercise a Using fuzzy-trace theory as a framework, perfect screening process, volunteering all Brainerd et al. (2003) proposed a mechanism correct items of information that come to for false-memory editing that allows chil- mind and withholding all incorrect items. dren and adults to reject false but gist- Koriat and Goldsmith’s model was consistent events. The model also predicts applied to study the strategic regulation of the occurrence of erroneous recollection memory accuracy by school-aged children rejection, in which true events are inappro- (Koriat et al., 2001). Even second-to-third- priately edited out of memory reports. grade children were effective in enhancing Payne, Jacoby, and Lambert (2004) inves- the accuracy of their testimony when given tigated the ability of participants to over- the freedom to volunteer or withhold an come stereotype-based memory distortions answer under a 1:1 penalty-to-bonus ratio, when allowed the option of free report. and they were able to enhance the accu- Reliance on subjective confidence allowed racy of their reports even further when given participants to enhance their overall mem- stronger incentives for accurate reporting. ory accuracy, but not to reduce stereotype However, both the children in this study bias. The results suggested that whereas sub- (see also Roebers, Moga, & Schneider, 2001) jective confidence monitors the accuracy and elderly adults in other studies (Kelley & of one’s report, stereotypes distort memory P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

metacognition and consciousness 311

through an unconscious-accessibility bias to cision or coarseness) of the information which subjective confidence is insensitive. that is reported (Goldsmith & Koriat, 1999; Hence the effects of stereotypes are difficult Goldsmith, Koriat, & Pansky, 2005; Gold- to control. smith, Koriat, & Weinberg-Eliezer, 2002). The work of Johnson and her associates For example, when not completely cer- on source monitoring (see Johnson, 1997; tain about the time of an event, a person Johnson, Hashtroudi, & Lindsay, 1993) also may simply report that it occurred “late in has important implications for the edit- the afternoon” rather than “at four-thirty.” ing of memory reports. According to the Neisser (1988) observed that when answer- source-monitoring framework, there are sev- ing open-ended questions, participants tend eral phenomenal cues that can be used by a to provide answers at a level of general- rememberer to specify the source of a men- ity at which they are not likely to be mis- tal record, including such mnemonic cues as taken. Of course, more coarsely grained vividness, perceptual detail, and spatial and answers, although more likely to be correct, temporal information. Because mental expe- are also less informative. Thus, Goldsmith riences from different sources (e.g., percep- et al. (2002) found that when participants tion versus imagination) differ on average are allowed to control the grain size of their in their phenomenal qualities (e.g., visual report, they do so in a strategic manner, clarity), these diagnostic qualities can sup- sacrificing informativeness (degree of preci- port source monitoring by using either a sion) for the sake of accuracy when their heuristically based process or a more strate- subjective confidence in the more precise gic, systematic process. Both types of pro- informative answer is low, and taking into cesses require setting criteria for making a account the relative payoffs for accuracy and judgment, as well as procedures for compar- informativeness in choosing the grain size ing activated phenomenal information to the of their answers. Moreover, the monitoring criteria. and control processes involved in the reg- The broader implication of the work on ulation of memory grain size appear to be the strategic regulation of memory accu- similar to those underlying the decision to racy (Koriat & Goldsmith, 1996b) is that, to volunteer or withhold specific items of infor- investigate the complex dynamics between mation, implying perhaps the use of com- (a) memory (the quality of the information mon metacognitive mechanisms. A more that is available to the rememberer), (b) recent study by Goldsmith et al. (2005), monitoring, (c) control, and (d) overt accu- which examined changes in the regulation racy and quantity performance, one must of grain size over different retention inter- include a situation in which participants are vals, also yielded results consistent with this free to decide what to report and what not to model: Starting with the well-known finding report. In fact, in everyday life people have that people often remember the gist of an great freedom in reporting an event from event though they have forgotten its details, memory: They can choose what perspective Goldsmith et al. (2005) asked whether to adopt, what to emphasize and what to rememberers might exploit the differential skip, how much detail to provide, and so forgetting rates of coarse and precise infor- forth. Such strategic regulation entails com- mation to strategically regulate the accu- plex monitoring and control processes that racy of the information that they report over go beyond the decision to volunteer or with- time. The results suggested that when given hold specific items of information, and these, control over the grain size of their answers, too, deserve systematic investigation. people tend to provide coarser answers at In fact, the conceptual framework of longer retention intervals, in the attempt to Koriat and Goldsmith was extended to incor- maintain a stable level of report accuracy. porate another means by which people nor- In sum, the few studies concerning the mally regulate the accuracy of what they control function of metacognition suggest report: control over the grain size (pre- that people rely heavily on their subjective, P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

312 the cambridge handbook of consciousness

metacognitive feelings and judgments in ple use under self-paced conditions is largely choosing their course of action. In addition appropriate. These and other results were to the monitoring output, however, they also seen to accord with the region of proxi- take into account a variety of other consid- mal learning framework according to which erations, such as the goals of learning and learning proceeds best by attending to con- remembering, time pressure, emphasis on cepts and events that are nearest to one’s accuracy versus quantity, and the emphasis current understanding and only later going on accuracy versus informativeness. on to integrate items that are more difficult. Thiede, Anderson, and Therriault (2003) used a manipulation that affected the The Effects of Metacognitive learner’s monitoring accuracy in studying Regulation on Memory Performance text. They found that improved accuracy resulted in a more effective regulation of Given the dynamics of monitoring and con- study and, in turn, in overall better test trol processes discussed so far, it is of performance. Thus, learners seem to rely interest to ask, To what extent does the self- on their metacognitive feelings in regulating regulation of one’s processing affect actual their behavior, and to the extent that these memory performance? There are only a few feelings are accurate, such self-regulation studies that have examined this issue sys- helps improve memory performance. tematically. As noted earlier, under self- With regard to confidence judgments, as paced learning conditions, when partici- noted earlier, the work of Koriat and Gold- pants are free to allocate study time to smith (1994, 1996b) indicates that when different items, they tend to divide their given the option of free report, people time unevenly among the items. Does the enhance their memory accuracy consider- self-allocation of study time affect actual ably in comparison to forced-report test- memory performance? Nelson and Leone- ing and do so by relying on the subjective sio (1988) coined the phrase “labor-in-vain confidence associated with each item that effect” to describe the phenomenon that comes to mind. Because confidence is gener- large increases in self-paced study time ally predictive of accuracy, reliance on con- yielded little or no gain in recall. Specif- fidence judgments is effective in enhancing ically, they observed that the amount of accuracy when accuracy is at stake. How- self-paced study time increased substan- ever, the effective regulation of memory tially under conditions that emphasized accuracy comes at the cost of reduced accuracy in comparison with a condition that memory quantity, and both the increase in emphasized speed. However, the increase in memory accuracy achieved under the free- study time resulted in little or no gain in report option and the reduction in mem- recall. ory quantity depend heavily on monitoring Metcalfe and her associates (Metcalfe, effectiveness. 2002; Metcalfe & Kornell, 2003) examined Koriat and Goldsmith (1996b) evaluated systematically the effectiveness of the pol- the effectiveness of the participants’ control icy of study time allocation for enhanc- policies given their actual levels of moni- ing memory performance. They found, for toring effectiveness. The participants were example, that learners allocated most time found to be quite effective in choosing a to medium-difficulty items and studied the control policy that would maximize their easiest items first (in contrast to what would joint levels of free-report accuracy and quan- be expected from the discrepancy-reduction tity performance, compared to an “opti- model, Dunlosky & Hertzog, 1998). When mal” control policy that could be applied study time was experimentally manipulated, directly, based on the confidence judg- the best performance resulted when most ments assigned to the individual answers time was given to the medium-difficulty under forced report. The effectiveness of items, suggesting that the strategy that peo- the participants’ control of grain size in the P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

metacognition and consciousness 313

Goldsmith et al. (2002) study was much less best highlighted by contrasting experience- impressive, however, perhaps because of the based judgment and theory-based judg- greater complexity of the incentive struc- ments. Similar contrasts have been proposed ture (differential payoffs for correct answers by researchers in both cognitive psychology at different grain sizes, a fixed penalty for and social psychology who drew a distinction incorrect answers, regardless of grain size). between two general modes of cognition (see In fact, one of the interesting findings of that Chaiken & Trope, 1999), and each of these study was that participants seemed to adopt contrasts highlights a particular dimension. a simple “satisficing” heuristic based on the Thus, different researchers have conceptual- payoff (whether explicit or implicit) and ized the distinction in terms of such polari- confidence for the more precise-informative ties as Nonanalytic versus Analytic cognition answer alone, rather than to compare the (Jacoby & Brooks, 1984), Associative versus expected subjective utility (confidence mul- Rule-Based Systems (Sloman, 1996), Expe- tiplied by subjective payoff) of potential riential versus Rational Systems (Epstein answers at different grain sizes. Monitor- & Pacini, 1999), Impulsive versus Reflec- ing effectiveness for the correctness of the tive processes (Strack & Deutsch, 2004), answers at different grain sizes was, how- Experience-Based versus Information-Based ever, also relatively poor (see also Yaniv & processes (Kelley & Jacoby, 1996a; Koriat & Foster, 1997). Thus, it may be that there Levy-Sadot, 1999), Heuristic versus Delib- are limits on the complexity and efficiency erate modes of thought (Kahneman, 2003), of both monitoring and control processes and Experiential versus Declarative informa- that in turn place limits on the performance tion (Schwarz, 2004). Stanovich and West benefits that can be achieved through such (2000) used the somewhat more neutral control. terms System 1 versus System 2, which In sum, only a few studies explored the have been adopted by Kahneman (2003)in effectiveness of metacognitive monitoring describing his work on judgmental biases. and control processes in enhancing actual In this chapter I focused on the contrast memory performance. More work in this between theory-based and experience-based vein is needed. judgments, which seems to capture best the findings in metacognition. As far as metacognitive judgments are concerned, the impor- Metacognition and Consciousness: tant assumption is that both experience- Some General Issues based and theory-based judgments are inferential in nature. They differ, however, in two respects. First, theory-based judgments draw In concluding this chapter I would like to upon the content of declarative (semantic comment on how the research on metacog- and/or episodic) information that is typically nition relates to some of the fundamental stored in long-term memory. Experience- issues regarding consciousness and its role based judgments, in contrast, are assumed to in behavior. I discuss three issues: the deter- rely on mnemonic cues stemming from the minants of subjective experience, the concurrent processing of the task at hand. Such trol function of subjective experience, and cues as fluency of processing or ease of access the cause-and-effect relation between con- pertain to the quality and efficacy of object- sciousness and behavior. level processes as revealed online. Hence, as Koriat (1993) argued, experience-based The Genesis of Subjective Experience FOK judgments, for example, monitor the The study of the bases of metacognitive information accessible in short-term mem- judgments and their accuracy brings to the ory rather than the information available in fore an important process that seems to long-term memory. It follows that the accu- underlie the shaping of subjective experi- racy of theory-based judgments depends on ence. The unique qualities of that process are the validity of the theories and knowledge P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

314 the cambridge handbook of consciousness

on which they are based, whereas the accu- by attempting to place the metacognitive racy of experience-based judgments should distinction within a broader framework that depend on the diagnosticity of the effective encompasses other similar distinctions. For mnemonic cues. example, research in social psychology sug- Second, they differ in the nature of gests that the interplay between declara- the underlying process. Theory-based judg- tive and experiential information is greater ments are assumed to rely on an explicitly than has been realized so far (see Schwarz, inferential process: The process is assumed 2004). However, little is known about the to be deliberate, analytic, slow, effortful, and possibility that a similar interplay between largely conscious. In contrast, experience- the effects of theories and knowledge and based judgments involve a two-step process: those of mnemonic cues occurs also with A fast, unconscious, automatic inference regard to metacognitive judgments. Also, lit- results in a sheer subjective experience, and tle research has been carried out that exam- that subjective experience can then serve as ines the possible effects of attribution and the basis for noetic judgments. Therefore, misattribution on metacognitive judgments. as Koriat and Levy-Sadot argued (1999), the Furthermore, processing fluency has been processes that take off from subjective expe- shown to affect a variety of phenomenal rience generally have no access to the pro- experiences, such as liking, truth judgments, cesses that have produced that experience recognition decisions, and so on. Again, in the first place. it is important to examine noetic feelings It is experience-based metacognitive in the context of these other phenomenal judgments that have attracted the atten- experiences. tion of memory researchers who asked the question, How do we know that we know? The Control Function of Subjective (e.g., Hart, 1965; Tulving & Madigan, 1970). Experience Experience-based judgments have the quality of immediate, direct impressions, similar The issue of metacognitive control emerges to what would follow from the trace-access most sharply when we ask, What is the sta- view of metacognitive judgments. However, tus of metacognitive monitoring and con- as argued earlier, this phenomenal quality trol processes within the current distinction could be explained in terms of the idea that between implicit and explicit cognition? In experience-based judgments are based on an light of the extensive research on both of inferential process that is not available to these areas of research, one would expect the consciousness, and hence the outcome of answer to be quite straightforward. How- that process has the phenomenal quality of a ever, such is not the case. In an edited vol- direct, self-evident intuition (see Epstein & ume on Implicit Memory and Metacognition Pacini, 1999). (Reder, 1996), the discussions of the partic- Thus, the work on metacognition nicely ipants revealed a basic ambivalence: Kelley converges on the proposals advanced by and Jacoby (1996b) claimed that “metacog- Jacoby and Kelley (see Kelley & Jacoby, nition and implicit memory are so similar as 1993) and by Whittlesea (2002, 2004)on to not be separate topics” (p. 287). Funnell, the shaping of subjective experience. These Metcalfe, and Tsapkini (1996), on the other proposals also parallel ideas in the area of hand, concluded that “the judgment of what social psychology on the genesis of vari- and how much you know about what you ous subjective feelings (see Bless & Forgas, know or will know is a classic, almost defini- 2000; Strack, 1992). However, although it tional, explicit task” (p. 172). Finally, Reder is theoretically comforting that the distinc- and Schunn (1996) stated, “Given that feel- tion between experience-based and theory- ing of knowing, like strategy selection, tends based metacognitive processes converges on to be thought of as the essence of a metacog- similar distinctions that have emerged in nitive strategy, it is important to defend our other domains, a great deal can be gained claim that this rapid feeling of knowing is P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

metacognition and consciousness 315

actually an implicit process rather than an of immediate feelings, such as experience- explicit process”(p.50). based metacognitive feelings, is to augment Koriat (1998b, 2000b) argued that this self-control; that is, to allow some degree of ambivalence actually discloses the two faces personal control over processes that would of metacognition. He proposed a crossover otherwise influence behavior directly and model that assigns metacognition a piv- automatically, outside the person’s con- otal role in mediating between unconscious sciousness and control. and conscious determinants of information processing. Thus, metacognitive judgments The Cause-and-Effect-Relation between were assumed to lie at the interface between Monitoring and Control implicit and explicit processes. Generally speaking, a rough distinction can be drawn A final metatheoretical issue concerns the between two modes of operation: In the assumption underlying much of the work explicit-controlled mode, which underlies in metacognition (and adopted in the fore- much of our daily activities, behavior is going discussion) – that metacognitive feel- based on a deliberate and conscious eval- ings play a causal role in affecting judgments uation of the available options and on a and behavior. However, the work of Jacoby deliberate and controlled choice of the most and his associates (see Kelley & Jacoby, appropriate course of action. In the implicit- 1998) and of Whittlesea (2004) suggests a automatic mode, in contrast, various fac- process that is more consistent with the tors registered below full consciousness may spirit of the James-Lange view of emotion influence behavior directly and automati- (see James, 1890): Subjective experience is cally, without the mediation of conscious based on an interpretation and attribution control (see Bargh, 1997; Wegner, 2002). of one’s own behavior, so that it follows Metacognitive experiences are assumed rather than precedes controlled processes. to occupy a unique position in this scheme: In fact, the assumption that metacognitive They are implicit as far as their antecedents feelings monitor the dynamics of informa- are concerned, but explicit as far as their tion processing implies that such feelings are consequences are concerned. Although a sometimes based on the feedback from self- strong feeling of knowing or an unmedi- initiated object-level processes. For exam- ated subjective conviction is certainly part ple, the accessibility model of FOK (Koriat, and parcel of conscious awareness, they may 1993) assumes that FOK judgments are themselves be the product of an unconscious based on the feedback from one’s attempt to inference, as reviewed earlier. Once formed, retrieve a target from memory. Hence they however, such subjective experiences can follow, rather than precede, controlled pro- serve as the basis for the conscious control cesses. Thus, whereas discussions of the func- of information processing and action. tion of metacognitive feelings assume that The crossover model may apply to other the subjective experience of knowing drives types of unmediated feelings (Koriat & Levy- controlled action, discussions of the bases Sadot, 1999). Thus, according to this view, of metacognitive feelings imply that such sheer subjective feelings, which lie at the feelings are themselves based on the feed- heart of consciousness, may themselves be back from controlled action, and thus follow the product of unconscious processes. Such rather than precede behavior. feelings represent an encapsulated summary Recent work that addressed the cause- of a variety of unconscious influences, and it and-effect relation between metacogni- is in this sense that they are informative (see tive monitoring and metacognitive con- Schwarz & Clore, 1996): They contain infor- trol (Koriat, in press; Koriat, Ma’ayan, & mation that is relevant to conscious con- Nussinson, 2006; see Koriat, 2000b) sug- trol, unlike the implicit, unconscious pro- gests that the interplay between them is cesses that have given rise to these feelings. bidirectional: Although metacognitive mon- Koriat (2000b) speculated that the function itoring can drive and guide metacognitive P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

316 the cambridge handbook of consciousness

control, it may itself be based on the feed- References back from controlled operations. Thus, when control effort is goal driven, greater effort Bandura, A. (1977). Self-efficacy: Toward a uni- enhances metacognitive feelings, consistent fying theory of behavioral change. Psychological with the “feelings-affect-behavior” hypothe- Review, 84, 191–215. sis. For example, when different incentives Bargh, J. A. (1997). The automaticity of every- are assigned to different items in a study list, day life. In R. S. Wyer, Jr. (Ed.), Advances in learners invest more study time on the high- social cognition (Vol. 10,pp.1–61). Mahwah, NJ: incentive items and, in parallel, make higher Erlbaum. JOLs for these items than for the low- Barnes, A. E., Nelson, T. O., Dunlosky, J., Maz- incentive items. This is similar to the idea zoni, G., & Narens, L. (1999). An integrative that we run away because we are fright- system of metamemory components involved ened, and therefore the faster we run away in retrieval. In D. Gopher & A. Koriat (Eds.), the safer we feel. In contrast, when control Attention and performance XVII: Cognitive regulation of performance: Interaction of theory and effort is data driven, increased effort is corre- application (pp. 287–313). Cambridge, MA: lated with lower metacognitive feelings, con- MIT Press. sistent with the hypothesis that such feelings Begg, I., Duft, S., Lalonde, P., Melnick, R., & are based on the feedback from behavior. For Sanvito, J. (1989). Memory predictions are example, under self-paced learning the more based on ease of processing. Journal of Memory effort learners spend studying an item the and Language, 28, 610–632. lower is their JOL, and also the lower is their Begg, I., Vinski, E., Frankovich, L., & Holgate, subsequent recall of that item. This is simi- B. (1991). Generating makes words memorable, lar to the idea that we are frightened because but so does effective reading. Memory and Cog- we are running away, and therefore the faster nition, 19, 487–497. we run the more fear we should experi- Benjamin, A. S. (2003). Predicting and postdict- ence. Thus, the study of metacognition can ing the effects of word frequency on memory. also shed light on the long-standing issue of Memory and Cognition, 31, 297–305. the cause-and-effect relation between con- Benjamin, A. S., & Bjork, R. A. (1996). Retrieval sciousness and behavior. fluency as a metacognitive index. In L. In sum, some of the current research Reder (Ed.), Implicit memory and metacognition in metacognition scratches the surface (pp. 309–338). Hillsdale, NJ: Erlbaum. of metatheoretical issues concerning con- Benjamin, A. S., Bjork, R. A., & Schwartz, sciousness and its role in behavior and is B. L. (1998). The mismeasure of memory: beginning to attract the attention of philoso- When retrieval fluency is misleading as a phers of mind (see Nelson & Rey, 2000). metamnemonic index. Journal of Experimental Psychology: General, 127, 55–68. Birch, S. A. J., & Bloom, P. (2003). Children Acknowledgments are cursed: An asymmetric bias in mental- state attribution. Psychological Science, 14, 283– 286. The preparation of this chapter was sup- 1999 ported by a grant from the German Fed- Bjork, R. A. ( ). Assessing our own competence: Heuristics and illusions. In D. eral Ministry of Education and Research Gopher & A. Koriat (Eds.), Attention and perfor- (BMBF) within the framework of German- mance XVII: Cognitive regulation of performance: Israeli Project Cooperation (DIP). The chap- Interaction of theory and application (pp. 435– ter was prepared when the author was a 459). Cambridge, MA: MIT Press. fellow at the Centre for Advanced Study, Bjork, R. A., & Bjork, E. L. (1992). A new theory Norwegian Academy of Science, Oslo. I of disuse and an old theory of stimulus fluctu- am grateful to Morris Goldsmith, Sarah ation. In A. F. Healy, S. M. Kosslyn, & R. M. Bar, and Rinat Gil for their help on this Shiffrin (Eds.), Essays in honor of William K. chapter. Estes, Vol. 1: From learning theory to connectionist P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

metacognition and consciousness 317

theory, Vol. 2 : From learning processes to cognitive in a modified recognition test. Memory & Cog- processes (pp. 35–67). Hillsdale, NJ: Erlbaum. nition, 22, 273–280. Bjorklund, D. F., & Douglas, R. N. (1997). Clore, G. L. (1994). Can emotions be noncon- The development of memory strategies. In N. scious. In P. Ekman & R. J. Davidson (Eds.), The Cowan (Ed.), The development of memory in nature of emotion: Fundamental questions. Series childhood (pp. 201–246). Hove, England: Taylor in affective science (pp. 283–299). London: & Francis. Oxford University Press. Bjorkman,¨ M., Juslin, P., & Winman, A. Cohen, R. L., Sandler, S. P., & Keglevich, L. (1993). Realism of confidence in sensory (1991). The failure of memory monitoring in discrimination: The underconfidence phe- a free recall task. Canadian Journal of Psychol- nomenon. Perception and Psychophysics, 54, ogy, 45, 523–538. 75–81. Costermans, J., Lories, G., & Ansay, C. (1992). Bless, H., & Forgas, J. P. (Eds.). (2000). The mes- Confidence level and feeling of knowing in sage within: The role of subjective experience in question answering: The weight of inferential social cognition and behavior. Philadelphia: Psy- processes. Journal of Experimental Psychology: chology Press. Learning, Memory, and Cognition, 18, 142–150. Brainerd, C. J., & Reyna, V. F., Wright, R., & Craik, F. I. M., & Lockhart, R. S. (1972). Levels of Mojardin, A. H. (2003). Recollection rejection: processing: A framework for memory research. False-memory editing in children and adults. Journal of Verbal Learning and Verbal Behavior, Psychological Review, 110, 762–784. 11, 671–684. Brown, A. L. (1987). Metacognition, executive Damasio, A. R. (1994). Descartes’ error: Emo- control, self-regulation, and other more mys- tion, reason, and the human brain. New York: terious mechanisms. In F. E. Weinert & R. Putnam. H. Kluwe (Eds.), Metacognition, motivation, Danion, J. M., Gokalsing, E., Robert, P., Massin- and understanding (pp. 95–116). Hillsdale, NJ: Krauss, M., & Bacon, E. (2001). Defective rela- Erlbaum. tionship between subjective experience and Brown, J., Lewis, V. J., & Monk, A. F. (1977). behavior in schizophrenia. American Journal of Memorability, word frequency and negative Psychiatry, 158, 2064–2066. recognition. Quarterly Journal of Experimental Davidson, J. E., & Sternberg, R. J. (1998). Smart Psychology, 29, 461–473. problem solving: How metacognition helps. In Brown, R., & McNeill, D. (1966). The “tip of the D. J. Hacker, J. Dunlosky, & A. C. Graesser tongue” phenomenon. Journal of Verbal Learn- (Eds.), Metacognition in educational theory and ing and Verbal Behavior, 5, 325–337. practice (pp. 47–68). Mahwah, NJ: Erlbaum. Burke, D. M., MacKay, D. G., Worthley, J. S., De Corte, E. (2003). Transfer as the productive & Wade, E. (1991). On the tip of the tongue: use of acquired knowledge, skills, and moti- What causes word finding failures in young and vations. Current Directions in Psychological Sci- older adults? Journal of Memory and Language, ence, 12, 142–146. 30, 542–579. Denes-Raj, V., & Epstein, S. (1994). Conflict Busey, T.A., Tunnicliff,J., Loftus, G. R., & Loftus, between intuitive and rational processing: E. (2000). Accounts of the confidence accuracy When people behave against their better judg- relation in recognition memory. Psychonomic ment. Journal of Personality and Social Psychol- Bulletin & Review, 7, 26–48. ogy, 66, 819–829. Carroll, M., Nelson, T. O., & Kirwan, A. (1997). Dunlosky, J., & Hertzog, C. (1998). Training pro- Tradeoff of semantic relatedness and degree of grams to improve learning in later adulthood: overlearning: Differential effects on metamem- Helping older adults educate themselves. In D. ory and on long-term retention. Acta Psycholog- J. Hacker (Ed.), Metacognition in educational ica, 95, 239–253. theory and practice (pp. 249–275). Mahwah, Chaiken, S., & Trope, Y. (Eds.). (1999). Dual NJ: Erlbaum. process theories in social psychology. New York: Dunlosky, J., & Nelson, T. O. (1992). Importance Guilford Press. of the kind of cue for judgments of learning Chandler, C. C. (1994). Studying related pictures (JOL) and the delayed-JOL effect. Memory & can reduce accuracy, but increase confidence, Cognition, 20, 374–380. P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

318 the cambridge handbook of consciousness

Dunlosky, J., & Nelson, T. O. (1994). Does the Flavell, J. H., & Wellman, H. M. (1977). sensitivity of judgments of learning (JOLs) to Metamemory. In R. V. Kail & J. W. Hagen the effects of various study activities depend on (Eds.), Perspectives on the development of mem- when the JOLs occur? Journal of Memory and ory and cognition (pp. 3–33). Hillsdale, NJ: Language, 33, 545–565. Erlbaum. Dunlosky, J., & Thiede, K. W. (1998). What Freytag, P. & Fiedler, K. (2006). Subjective valid- makes people study more? An evaluation of ity judgements as an index of sensitivity to factors that affect self-paced study. Acta Psy- sampling bias. In K. Fiedler & P. Juslin (Eds.), chologica, 98, 37–56. Information sampling and adaptive cognition 127 146 Dunning, D., Johnson, K., Ehrlinger, J., & Kruger, (pp. – ). New York: Cambridge Univer- J. (2003). Why people fail to recognize their sity Press. own incompetence. Current Directions in Psy- Funnell, M., Metcalfe, J., & Tsapkini, K. (1996). chological Science, 12, 83–87. In the mind but not on the tongue: Feeling Ehrlinger, J., & Dunning, D. (2003). How chronic of knowing in an anomic patient. In L. M. self-views influence (and potentially mislead) Reder, (Ed.) Implicit memory and metacognition 171 194 estimates of performance. Journal of Personality (pp. – ). Hillsdale, NJ: Erlbaum. and Social Psychology, 84, 5–17. Gardiner, J. M., & Richardson-Klavehn, A. 2000 Epstein, S., & Pacini, R. (1999). Some basic issues ( ). Remembering and knowing. In regarding dual-process theories from the per- E. Tulving & F. I. M. Craik (Eds.), The Oxford 229 244 spective of cognitive-experiential self-theory. handbook of memory (pp. – ). London: In S. Chaiken & Y. Trope (Eds.), Dual process Oxford University Press. theories in social psychology (pp. 462–482). New Garry, M., Manning, C. G., Loftus, E. F., & Sher- 1996 York: Guilford Press. man, S. J. ( ). Imagination inflation: Imag- ining a childhood event inflates confidence that Fernandez-Duque, D., Baird, J., & Posner, M. it occurred. Psychonomic Bulletin and Review, , (2000). Awareness and metacognition. Con- 3 208–214. sciousness and Cognition, 9, 324–326. Ghetti, S. (2003). Memory for nonoccurrences: Fiedler, K., Brinkmann, B., Betsch, T., & Wild, B. The role of metacognition. Journal of Memory (2000). A sampling approach to biases in con- and Language, 48, 722–739. ditional probability judgments: Beyond base rate neglect and statistical format. Journal of Gigerenzer, G., Hoffrage, U., & Kleinbolting,¨ H. (1991). Probabilistic mental models: A Experimental Psychology: General, 129, 399– 418. Brunswikian theory of confidence. Psychologi- cal Review, 98, 506–528. Fischhoff, B. (1975). Hindsight is not equal to Gilbert, D. T. (2002). Inferential correction. In T. foresight: The effect of outcome knowledge on Gilovich, D. Griffin, & D. Kahneman (Eds.), judgment under uncertainty. Journal of Exper- Heuristics and biases: The psychology of intuitive imental Psychology: Human Perception and Per- judgment (pp. 167–184). New York: Cambridge formance, 1, 288–299. University Press. 1977 Fischhoff, B., Slovic, P., & Lichtenstein, S. ( ). Gilovich, T., Griffin, D. & Kahneman, D. (Eds.). Knowing with certainty: The appropriateness (2002). Heuristics and biases: The psychology of of extreme confidence. Journal of Experimental intuitive judgment. New York: Cambridge Uni- Psychology: Human Perception and Performance, versity Press. 3, 552–564. Glenberg, A. M., Sanocki, T., Epstein, W., & Mor- 1971 Flavell, J. H. ( ). First discussant’s comments: ris, C. (1987). Enhancing calibration of com- What is memory development the devel- prehension. Journal of Experimental Psychology: 272 opment of? Human Development, 14, – General, 116, 119–136. 278 . Glucksberg, S., & McCloskey, M. (1981). Deci- Flavell, J. H. (1979). Metacognition and cog- sions about ignorance: Knowing that you nitive monitoring: A new area of cognitive- don’t know. Journal of Experimental Psychology: developmental inquiry. American Psychologist, Human Learning and Memory, 7, 311–325. 906 911 34, – . Goldsmith, M., & Koriat, A. (1999). The strategic Flavell, J. H. (1999). Cognitive development: regulation of memory reporting: Mechanisms Children’s knowledge about the mind. Annual and performance consequences. In D. Gopher Review of Psychology, 50, 21–45. & A. Koriat (Eds.), Attention and performance P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

metacognition and consciousness 319

XVII: Cognitive regulation of performance: Inter- Jacoby, L. L., & Whitehouse, K. (1989). An action of theory and application (pp. 373–400). illusion of memory: False recognition influ- Cambridge, MA: MIT Press. enced by unconscious perception. Journal of Goldsmith, M., Koriat, A., & Pansky, A. (2005) Experimental Psychology: General, 118, 126– Strategic regulation of grain size in memory 135. reporting over time. Journal of Memory and James, W. (1890). The principles of psychology. Language, 52, 505–525. New York: Holt. Goldsmith, M., Koriat, A., & Weinberg-Eliezer, Johnson, M. K. (1997). Identifying the origin A. (2002). Strategic regulation of grain size of mental experience. In M. S. Myslobodsky memory reporting. Journal of Experimental (Ed.), The mythomanias: The nature of deception Psychology: General, 131, 73–95. and self-deception (pp. 133–180). Hillsdale, NJ: Griffin, D., & Tversky, A. (1992). The weighing of Erlbaum. evidence and the determinants of confidence. Johnson, M. K., Hashtroudi, S., & Lindsay, D. S. Cognitive Psychology, 24, 411–435. (1993). Source monitoring. Psychological Bul- Gruneberg, M. M., & Monks, J. (1974). “Feeling letin, 114, 3–28. of knowing” and cued recall. Acta Psychologica, Jonsson,¨ F. U., & Olsson, M. J. (2003). Olfac- 38, 257–265. tory metacognition. Chemical Senses, 28, 651– Gruneberg, M. M., Monks, J., & Sykes, R. N. 658. (1977). Some methodological problems with Jost, J. T., Kruglanski, A. W., & Nelson, T. O. feelings of knowing studies. Acta Psychologica, (1998). Social metacognition: An expansion- 41, 365–371. ist review. Personality and Social Psychology Guttentag, R., & Carroll, D. (1998). Memora- Review, 2 , 137–154. bility judgments for high- and low-frequency Kahneman, D. (2003). A perspective on judg- words. Memory and Cognition, 26, 951–958. ment and choice: Mapping bounded rational- Hacker, D. J., Dunlosky, J., & Graesser, A. C. ity. American Psychologist, 58, 697–720. (Eds.). (1998). Metacognition in educational the- Kelley, C. M., & Jacoby, L. L. (1993). The con- ory and practice. Mahwah, NJ: Erlbaum. struction of subjective experience: Memory Hart, J. T. (1965). Memory and the feeling-of- attributions. In M. Davies & G. W. Humphreys knowing experience. Journal of Educational Psy- (Eds.), Consciousness: Psychological and philo- chology, 56, 208–216. sophical essays. Readings in mind and language 2 74 89 Hart, J. T. (1967). Second-try recall, recognition, (Vol. ,pp. – ). Malden, MA: Blackwell. and the memory-monitoring process. Journal of Kelley, C. M., & Jacoby, L. L. (1996a). Adult Educational Psychology, 58, 193–197. egocentrism: Subjective experience versus ana- Hastie, R., Landsman, R., & Loftus, E. F. (1978). lytic bases for judgment. Journal of Memory and 157 175 The effects of initial questioning on subsequent Language, 35, – . eyewitness testimony. Jurimetrics Journal, 19, Kelley, C. M., & Jacoby, L. L. (1996b). Mem- 1–8. ory attributions: Remembering, knowing, and Hertzog, C., Dunlosky, J., Robinson, A. E., & feeling of knowing. In L. M. Reder (Ed.), 287 Kidder, D. P. (2003). Encoding fluency is a Implicit memory and metacognition (pp. – 308 cue used for judgments about learning. Journal ). Hillsdale, NJ: Erlbaum. of Experimental Psychology: Learning, Memory, Kelley, C. M., & Jacoby, L. L. (1998). Subjec- and Cognition, 29, 22–34. tive reports and process dissociation: Fluency, Hertzog, C., Kidder, D. P., Powell-Moman, A., knowing, and feeling. Acta Psychologica, 98, 127 140 & Dunlosky, J. (2002). Aging and monitor- – . ing associative learning: Is monitoring accuracy Kelley, C. M., & Lindsay, D. S. (1993). Remem- spared or impaired? Psychology and Aging, 17, bering mistaken for knowing: Ease of retrieval 209–225. as a basis for confidence in answers to general Jacoby, L. L., & Brooks, L. R. (1984). Nonana- knowledge questions. Journal of Memory and 1 24 lytic cognition: Memory, perception, and con- Language, 32, – . cept learning. In G. H. Bower (Ed.), The psy- Kelley, C. M., & Sahakyan, L. (2003). Memory, chology of learning and motivation: Advances monitoring, and control in the attainment of in research and theory (pp. 1–47). New York: memory accuracy. Journal of Memory and Lan- Academic Press. guage, 48, 704–721. P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

320 the cambridge handbook of consciousness

Kentridge, R. W., & Heywood, C. A. (2000). (Eds.), Applied metacognition (pp. 261–286). Metacognition and awareness. Consciousness Cambridge: Cambridge University Press. and Cognition, 9, 308–312. Koriat, A. (in press). Are we frightened because Kimball, D. R., & Metcalfe, J. (2003). Delay- we run away? Some evidence from metacogni- ing judgments of learning affects memory, not tive feelings. In B. Uttl, N. Ohta, & A. L. Siegen- metamemory. Memory and Cognition, 31, 918– thaler (Eds.), Memory and emotions: Interdisci- 929. plinary perspectives. Malden, MA: Blackwell. King, J. F., Zechmeister, E. B., & Shaughnessy, J. J. Koriat, A., Ben-Zur, H., & Druch, A. (1991). (1980). Judgments of knowing: The inﬂuence The contextualization of memory for input and of retrieval practice. American Journal of Psy- output events. Psychological Research, 53, 260– chology, 93, 329–343. 270. Klin, C. M., Guzman, A. E., & Levine, Koriat, A., Ben-Zur, H., & Sheffer, D. (1988). W. H. (1997). Knowing that you don’t know: Telling the same story twice: Output monitor- Metamemory and discourse processing. Journal ing and age. Journal of Memory and Language, of Experimental Psychology: Learning, Memory, 27, 23–39. and Cognition, 23, 1378–1393. Koriat, A., & Bjork, R. A. (2005). Illusions of com- Koren, D., Seidman, L. J., Poyurovsky, M., Gold- petence in monitoring one’s knowledge dur- smith, M., Viksman, P., Zichel, S., & Klein, E. ing study. Journal of Experimental Psychology: (2004). The neuropsychological basis of insight Learning, Memory and Cognition, 31, 187– in ﬁrst-episode schizophrenia: A pilot metacog- 194. nitive study. Schizophrenia Research, 70, 195– Koriat, A., & Bjork, R. A. (2006). Illusions of 202. competence during study can be remedied by Koriat, A. (1993). How do we know that we manipulations that enhance learners’ sensitiv- know? The accessibility model of the feeling of ity to retrieval conditions at test. Memory & knowing. Psychological Review, 100, 609–639. Cognition, 34, 959–972. Koriat, A. (1995). Dissociating knowing and the Koriat, A., & Bjork, R. A., Sheffer, L., & Bar, feeling of knowing: Further evidence for the S. K. (2004). Predicting one’s own forgetting: accessibility model. Journal of Experimental The role of experience-based and theory-based Psychology: General, 124, 311–333. processes. Journal of Experimental Psychology: Koriat, A. (1997). Monitoring one’s own knowl- General, 133, 643–656. edge during study: A cue-utilization approach Koriat, A., & Goldsmith, M. (1994). Memory to judgments of learning. Journal of Experimen- in naturalistic and laboratory contexts: Distin- tal Psychology: General, 126, 349–370. guishing the accuracy-oriented and quantity- Koriat, A. (1998a). Illusions of knowing: The link oriented approaches to memory assessment. between knowledge and metaknowledge. In Journal of Experimental Psychology: General, V. Y. Yzerbyt, G. Lories, & B. Dardenne (Eds.), 123, 297–315. Metacognition: Cognitive and social dimensions Koriat, A., & Goldsmith, M. (1996a). Mem- (pp. 16–34). London: Sage. ory metaphors and the real-life/laboratory Koriat, A. (1998b). Metamemory: The feeling of controversy: Correspondence versus store- knowing and its vagaries. In M. Sabourin, F. I. house conceptions of memory. Behavioral and M. Craik, & M. Robert (Eds.), Advances in psy- Brain Sciences, 19, 167–228. chological science (Vol. 2,pp.461–469). Hove, Koriat, A., & Goldsmith, M. (1996b). Monitoring UK: Psychology Press. and control processes in the strategic regulation Koriat, A. (2000a). Control processes in remem- of memory accuracy. Psychological Review, 103, bering. In E. Tulving & F. I. M. Craik (Eds.), 490–517. The Oxford handbook of memory (pp. 333–346). Koriat, A., Goldsmith, M., & Pansky, A. (2000). London: Oxford University Press. Toward a psychology of memory accuracy. Koriat, A. (2000b). The feeling of knowing: Some Annual Review of Psychology, 51, 481–537. metatheoretical implications for consciousness Koriat, A., Goldsmith, M., Schneider, W., & and control. Consciousness and Cognition, 9, Nakash-Dura, M. (2001). The credibility of 149–171. children’s testimony: Can children control the Koriat, A. (2002). Metacognition research: An accuracy of their memory reports? Journal of interim report. In T. J. Perfect & B. L. Schwartz Experimental Child Psychology, 79, 405–437. P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

metacognition and consciousness 321

Koriat, A., & Levy-Sadot, R. (1999). Pro- Liberman, V. (2004). Local and global judg- cesses underlying metacognitive judgments: ments of confidence. Journal of Experimental Information-based and experience-based mon- Psychology: Learning, Memory, and Cognition, itoring of one’s own knowledge. In S. Chaiken 30, 729–732. & Y. Trope (Eds.), Dual process theories in social Lichtenstein, S., Fischhoff, B., & Phillips, L. D. psychology (pp. 483–502). New York: Guilford (1982). Calibration of probabilities: The state Press. of the art to 1980. In D. Kahneman, P. Slovic, & Koriat, A., & Levy-Sadot, R. (2001). The com- A. Tversky (Eds.), Judgment under uncertainty: bined contributions of the cue-familiarity and Heuristics and biases (pp. 306–334). New York: accessibility heuristics to feelings of knowing. Cambridge University Press. Journal of Experimental Psychology: Learning, Lindsay, D. S., Read, D. J., & Sharma, K. 34 53 Memory, and Cognition, 27, – . (1998). Accuracy and confidence in person Koriat, A., Levy-Sadot, R., Edry, E., & de Mar- identification: The relationship is strong when cas,S.(2003). What do we know about what witnessing conditions vary widely. Psychologi- we cannot remember? Accessing the seman- cal Science, 9, 215–218. tic attributes of words that cannot be recalled. Maki, R. H., & McGuire, M. J. (2002). Metacog- Journal of Experimental Psychology: Learning, nition for text: Findings and implications for 1095 1105 Memory, and Cognition, 29, – . education. In T. Perfect & B. Schwartz (Eds.), Koriat, A., Lichtenstein, S., & Fischhoff, B. (1980). Applied metacognition (pp. 39–67). Cambridge: Reasons for confidence. Journal of Experimental Cambridge University Press. Psychology: Human Learning and Memory, 6, Matvey, G., Dunlosky, J., & Guttentag, R. (2001). 107–118. Fluency of retrieval at study affects judgments 1977 Koriat, A., & Lieblich, I. ( ). A study of mem- of learning (JOLs): An analytic or nonanalytical 151 164 ory pointers. Acta Psychologica, 41, – . basis for JOLs? Memory and Cognition, 29, 222– Koriat, A., & Ma’ayan, H. (2005). The effects of 233. encoding fluency and retrieval fluency on judg- Mazzoni, G., & Cornoldi, C. (1993). Strategies ments of learning. Journal of Memory and Lan- in study time allocation: Why is study time 478 492 guage, 52, – . sometimes not effective? Journal of Experimen- Koriat, A., Ma’ayan, H., & Nussinson, R. (2006). tal Psychology: General, 122, 47–60. The intricate relationships between monitor- Mazzoni, G., Cornoldi, C., & Marchitelli, G. ing and control in metacognition: Lessons (1990). Do memorability ratings affect study- for the cause -and-effect relation between time allocation? Memory and Cognition, 18, subjective experience and behavior. Jour- 196–204. nal of Experimental Psychology: General, 135, 1995 36–69. Mazzoni, G., & Nelson, T. O. ( ). Judgments of learning are affected by the kind of encoding Koriat, A., Sheffer, L., & Ma’ayan, H. (2002). in ways that cannot be attributed to the level Comparing objective and subjective learn- of recall. Journal of Experimental Psychology: ing curves: Judgments of learning exhibit Learning, Memory, and Cognition, 21, 1263– increased underconfidence with practice. Jour- 1274. nal of Experimental Psychology: General, 131, 147–162. Metcalfe, J. (1998a). Cognitive optimism: Krinsky, R., & Nelson, T.O. (1985). The feeling of Self-deception or memory-based processing knowing for different types of retrieval failure. heuristics? Personality and Social Psychology 100 110 Acta Psychologica, 58, 141–158. Review, 2 , – . 1998 Le Ny, J. F., Denhiere, G., & Le Taillanter, D. Metcalfe, J. (Ed.). ( b). Metacognition. (1972). Regulation of study-time and interstim- [Special issue]. Personality and Social Psycho- 2 ulus similarity in self-paced learning condi- logical Review, . tions. Acta Psychologica, 36, 280–289. Metcalfe, J. (2000). Metamemory: Theory and Leonesio, R. J., & Nelson, T. O. (1990). Do dif- data. In E. Tulving & F. I. M. Craik (Eds.), ferent metamemory judgments tap the same The Oxford handbook of memory (pp. 197–211). underlying aspects of memory? Journal of London: Oxford University Press. Experimental Psychology: Learning, Memory, Metcalfe, J. (2002). Is study time allocated selec- and Cognition, 16, 464–470. tively to a region of proximal learning? Journal P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

322 the cambridge handbook of consciousness

of Experimental Psychology: General, 131, 349– Nelson, T. O., & Narens, L. (1990). Metamemory: 363. A theoretical framework and new findings. In Metcalfe, J., & Kornell, N. (2003). The dynam- G. Bower (Ed.), The psychology of learning and ics of learning and allocation of study time to motivation: Advances in research and theory (pp. a region of proximal learning. Journal of Exper- 125–173). New York: Academic Press. imental Psychology: General, 132, 530–542. Nelson, T. O., & Narens, L. (1994). Why investi- Metcalfe, J., Schwartz, B. L., & Joaquim, gate metacognition. In J. Metcalfe & A. P. Shi- S. G. (1993). The cue-familiarity heuris- mamura (Eds.), Metacognition: Knowing about tic in metacognition. Journal of Experimental knowing (pp. 1–25). Cambridge, MA: MIT Psychology: Learning, Memory, and Cognition, Press. 19, 851–864. Nelson, T. O., Narens, L., & Dunlosky, J. Metcalfe, J., & Shimamura, A. P. (Eds.). (1994). (2004). A revised methodology for research on Metacognition: Knowing about knowing. Cam- metamemory: Pre-judgment Recall and Moni- bridge, MA: MIT Press. toring (PRAM). Psychological Methods, 9, 53– 69 Miozzo, M., & Caramazza, A. (1997). Retrieval of . lexical-syntactic features in tip-of-the tongue Nelson, T. O., & Rey, G. (Eds.). (2000). Metacog- states. Journal of Experimental Psychology: nition and consciousness: A convergence of Learning, Memory, and Cognition, 23, 1410– psychology and philosophy [Special issue]. 1423. Consciousness and Cognition, 9(2). Mitchell, K. J., & Johnson, M. K. (2000). Source Nickerson, R. S. (1998). Confirmation bias: monitoring: Attributing mental experiences. In A ubiquitous phenomenon in many guises. E. Tulving & F. I. M. Craik (Eds.), The Oxford Review of General Psychology, 2 , 175–220. handbook of memory (pp. 179–195). London: Nhouyvanisvong, A., & Reder, L. M. (1998). Oxford University Press. Rapid feeling-of-knowing: A strategy selection Morris, C. C. (1990). Retrieval processes under- mechanism. In V. Y. Yzerbyt, G. Lories, & B. lying confidence in comprehension judgments. Dardenne (Eds.), Metacognition: Cognitive and Journal of Experimental Psychology: Learning, social dimensions (pp. 35–52). London: Sage. Memory, and Cognition, 16, 223–232. Oppenheimer, D. M. (2004). Spontaneous dis- Narens, L., Jameson, K. A., & Lee, V. A. (1994). counting of availability in frequency judgment Subthreshold priming and memory monitor- tasks. Psychological Science, 15, 100–105. ing. In J. Metcalfe & A. P. Shimamura (Eds.), Pansky, A., Koriat, A., & Goldsmith, M. (2005). Metacognition: Knowing about knowing (pp. 71– Eyewitness recall and testimony. In N. Brewer 92). Cambridge, MA: MIT Press. & K. D. Williams (Eds.), Psychology and law: An Neisser, U. (1988). Time present and time past. In empirical perspective (pp. 93–150). New York: M. M. Gruneberg, P. Morris, & R. Sykes (Eds.), Guilford. Practical aspects of memory: Current research Pansky, A., Koriat, A., Goldsmith, M., & and issues (Vol. 2,pp.545–560). Chichester, Pearlman-Avnion, S. (March, 2002). Memory England: Wiley. accuracy and distortion in old age: Cognitive, Nelson, T. O. (1984). A comparison of current metacognitive, and neurocognitive determinants. measures of the accuracy of feeling-of-knowing Poster presented at the 30th Anniversary Con- predictions. Psychological Bulletin, 95, 109–133. ference of the National Institute for Psychobi- Nelson, T. O. (1996). Consciousness and ology, Jerusalem, Israel. metacognition. American Psychologist, 5, 102– Paris, S. G., & Winograd, P. (1990). How metacog- 116. nition can promote academic learning and Nelson, T. O., & Dunlosky, J. (1991). When instruction. In B. F. Jones & L. Idol (Eds.), people’s judgments of learning (JOLs) are Dimensions of thinking and cognitive instruction 15 51 extremely accurate at predicting subsequent (pp. – ). Hillsdale, NJ: Erlbaum. recall: The “delayed-JOL effect.” Psychological Payne, B. K., Jacoby, L. L., & Lambert, A. J., Science, 2 , 267–270. (2004). Memory monitoring and the control Nelson, T. O., & Leonesio, R. J. (1988). Alloca- of stereotype distortion. Journal of Experimen- 52 64 tion of self-paced study time and the “labor-in- tal Social Psychology, 40, – . vain effect.” Journal of Experimental Psychology: Perfect, T. J. (2002). When does eyewitness con- Learning, Memory, and Cognition, 14, 676–686. fidence predict performance? In T. J. Perfect P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

metacognition and consciousness 323

& B. L. Schwartz (Eds.), Applied metacognition The detrimental effects of attention to the (pp. 95–120). Cambridge: Cambridge Univer- identification process. Law and Human Behav- sity Press. ior, 22, 409–428. Perfect, T. J. (2004). The role of self-rated abil- Robinson, M. D., Johnson, J. T., & Herndon, F. ity in the accuracy of confidence judgments (1997). Reaction time and assessments of cog- in eyewitness memory and general knowledge. nitive effort as predictors of eyewitness mem- Applied Cognitive Psychology, 18, 157–168. ory accuracy and confidence. Journal of Applied Perner, J., & Lang, B. (1999). Development of the- Psychology, 82, 416–425. ory of mind and executive control. Trends in Robinson, M. D., Johnson, J. T., & Robertson, Cognitive Sciences, 3, 337–344. D. A. (2000). Process versus content in eyewit- Posner, M. I., & Snyder, C. R. R. (1975). Atten- ness metamemory monitoring. Journal of Exper- tion and cognitive control. In R. L. Solso imental Psychology: Applied, 6, 207–221. (Ed.), Information processing and cognition: The Roebers, C. M., Moga, N., & Schneider, W. Loyola Symposium (pp. 55–85). Hillsdale, NJ: (2001). The role of accuracy motivation on chil- Erlbaum. dren’s and adults’ event recall. Journal of Exper- Pressley, M., Borkowski, J. G., & Schneider, imental Child Psychology, 78, 313–329. W. (1987). Cognitive strategies: Good strat- Rosenthal, D. M. (2000). Consciousness, con- egy users coordinate metacognition and knowl- tent, and metacognitive judgments. Conscious- edge. Annals of Child Development, 4, 89–129. ness and Cognition, 9, 203–214. Rabinowitz, J. C., Ackerman, B. P., Craik, Scheid, K. (1993). Helping students become strate- F. I. M., & Hinchley, J. L. (1982). Aging and gic learners: Guidelines for teaching. Cambridge, metamemory: The roles of relatedness and MA: Brookline Books. imaginary. Journal of Gerontology, 37, 688–695. Schneider, W. (1985). Developmental trends Rabbitt, P. M. A. (1966). Errors and error correc- in the metamemory-memory behavior tion in choice reaction tasks. Journal of Experi- relationship: An integrative review. In D. mental Psychology, 71, 264–272. L. Forest-Pressley, G. E. MacKinnon, & T. Rawson, K. A., Dunlosky, J., & Theide, K. W. G. Waller (Eds.), Metacognition, cognition, (2000). The rereading effect: Metacomprehen- and human performance (Vol. 1,pp.57–109). sion accuracy improves across reading trials. Orlando, FL: Academic Press. Memory and Cognition, 28, 1004–1010. Schneider, W., & Pressley, M. (1997). Memory Reder, L. M. (1987). Strategy selection in ques- development between two and twenty (2d ed.). tion answering. Cognitive Psychology, 19, 90– Mahwah, NJ: Erlbaum. 138. Schreiber, T. A., & Nelson, D. L. (1998). The Reder, L. M. (1988). Strategic control of retrieval relation between feelings of knowing and the strategies. In G. H. Bower (Ed.), The psychology number of neighboring concepts linked to the of learning and motivation: Advances in research test cue. Memory and Cognition, 26, 869– and theory (Vol. 22,pp.227–259). San Diego: 883. Academic Press. Schunn, C. D., Reder, L. M., Nhouyvanisvong, Reder, L. M. (Ed.). (1996). Implicit memory and A., Richards, D. R., & Stroffolino, P. J. (1997). metacognition. Mahwah, NJ: Erlbaum. To calculate or not to calculate: A source activa- Reder, L. M., & Ritter, F. E. (1992). What deter- tion confusion model of problem familiarity’s mines initial feeling of knowing? Familiarity role in strategy selection. Journal of Experimen- with question terms, not with the answer. Jour- tal Psychology: Learning, Memory, and Cogni- 3 29 nal of Experimental Psychology: Learning, Mem- tion, 23, – . ory, and Cognition, 18, 435–451. Schwartz, B. L. (1994). Sources of information in Reder, L. M. & Schunn, C. D. (1996). Metacogni- metamemory: Judgments of learning and feel- tion does not imply awareness: Strategy choice ing of knowing. Psychonomic Bulletin & Review, 357 375 is governed by implicit learning and mem- 1, – . ory. In L. M. Reder (Ed.), Implicit memory Schwartz, B. L. (1998). Illusory tip-of-the-tongue and metacognition (pp. 45–77). Mahwah, NJ: states. Memory, 6, 623–642. Erlbaum. Schwartz, B. L. (2001). The relation of tip-of-the- Robinson, M. D., & Johnson, J. T.(1998). How not tongue states and retrieval time. Memory and to enhance the confidence-accuracy relation: Cognition, 29, 117–126. P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

324 the cambridge handbook of consciousness

Schwartz, B. L. (2002). Tip-of-the-tongue states: itive judgment (pp. 397–420). New York: Cam- Phenomenology, mechanism, and lexical bridge University Press. retrieval. Mahwah, NJ: Erlbaum. Son, L. K., & Metcalfe, J. (2000). Metacognitive Schwartz, B. L., & Metcalfe, J. (1992). Cue famil- and control strategies in study-time allocation. iarity but not target retrievability enhances Journal of Experimental Psychology: Learning, feeling-of-knowing judgments. Journal of Memory, and Cognition, 26, 204–221. Experimental Psychology: Learning, Memory, Spehn, M. K., & Reder, L. M. (2000). The uncon- and Cognition, 18, 1074–1083. scious feeling of knowing: A commentary on Schwartz, B. L., & Metcalfe, J. (1994). Method- Koriat’s paper. Consciousness and Cognition, 9, ological problems and pitfalls in the study of 187–192. human metacognition. In J. Metcalfe & A. Spellman, B. A., & Bjork, R. A. (1992). When P. Shimamura (Eds.), Metacognition: Knowing predictions create reality: Judgments of learn- about knowing (pp. 93–113). Cambridge, MA: ing may alter what they are intended to assess. MIT Press. Psychological Science, 3, 315–316. Schwartz, B. L., & Smith, S. M. (1997). The Stanovich, K. E., & West, R. F. (2000). Individ- retrieval of related information influences tip- ual differences in reasoning: Implications for of-the-tongue states. Journal of Memory and the rationality debate. Behavioral and Brain Sci- Language, 36, 68–86. ences, 23, 645–665. Schwarz, N. (2004). Meta-cognitive experiences Strack, F. (1992). The different routes to social in consumer judgment and decision making. judgments: Experiential versus informational Journal of Consumer Psychology, 14, 332–348. strategies. In L. L. Martin & A. Tesser (Eds.), Schwarz, N., Bless, H., Strack, F., Klumpp, G., The construction of social judgments (pp. 249– Rittenauer-Schatka, H., & Simons, A. (1991). 275). Hillsdale, NJ: Erlbaum. Ease of retrieval as information: Another Strack, F., & Bless, H. (1994). Memory for look at the availability heuristic. Journal of nonoccurrences: Metacognitive and presuppo- Personality and Social Psychology, 61, 195– sitional strategies. Journal of Memory and Lan- 202. guage, 33, 203–217. Schwarz, N., & Clore, G. L. (1996). Feelings and Strack, F., & Deutsch R. (2004). Reflective and phenomenal experiences. In E. T. Higgins & A. impulsive determinants of social behavior. Per- W. Kruglanski (Eds.), Social psychology: Hand- sonality and Social Psychology Review, 8, 220– book of basic principles (pp. 433–465). New 247. York: Guilford Press. Thiede, K. W., Anderson, M. C. M., & Therriault, Schwarz, N., & Clore, G. L. (2003). Mood D. (2003). Accuracy of metacognitive monitor- as information: 20 years later. Psychological ing affects learning of texts. Journal of Educa- Inquiry, 14, 296–303. tional Psychology, 95, 66–73. Shaw, R. J., & Craik, F. I. M. (1989). Age differ- Thiede, K. W., & Dunlosky, J. (1999). Toward ences in predictions and performance on a cued a general model of self-regulated study: An recall task. Psychology and Aging, 4, 131–135. analysis of selection of items for study and Shimamura, A. P. (2000). Toward a cognitive self-paced study time. Journal of Experimental neuroscience of metacognition. Consciousness Psychology: Learning, Memory, and Cognition, and Cognition, 9, 313–323. 25, 1024–1037. Simon, D. A., & Bjork, R. A. (2001). Metacogni- Trope, Y., & Liberman, A. (1996). Social hypoth- tion in motor learning. Journal of Experimental esis testing: Cognitive and motivational mech- Psychology: Learning, Memory, and Cognition, anisms. In E. T. Higgins & A. W. Kruglan- 27, 907–912. ski (Eds.), Social psychology: Handbook of basic Sloman, S. A. (1996). The empirical case for two principles (pp. 239–270). New York: Guilford systems of reasoning. Psychological Bulletin, 119, Press. 3–22 Tulving,E., & Madigan, S. A. (1970). Memory and Slovic, P., Finucane, M., Peters, E., & MacGre- verbal learning. Annual Review of Psychology, gor,D.G.(2002). The affect heuristic. In T. 21, 437–484. Gilovich, D. Griffin, & Kahneman, D. (Eds.), Underwood, B. J. (1966). Individual and group Heuristics and biases: The psychology of intu- predictions of item difficulty for free learning. P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

metacognition and consciousness 325

Journal of Experimental Psychology, 71, 673– judgment. In J. Musch & K. C. Klauer (Eds.), 679. The psychology of evaluation: Affective processes Vernon, D., & Usher, M. (2003). Dynamics in cognition and emotion (pp. 189–217). Mah- of metacognitive judgments: Pre- and postre- wah, NJ: Erlbaum. trieval mechanisms. Journal of Experimental Winman, A., & Juslin, P. (2005). “I’m m/n confi- Psychology: Learning, Memory, and Cognition, dent that I’m correct”: Confidence in foresight 29, 339–346. and hindsight as a sampling probability. In K. Wegner, D. M. (2002). The illusion of conscious Fiedler & P. Juslin (Eds.), Information sampling will. Cambridge, MA: MIT Press. and adaptive cognition. Cambridge, UK: Cam- Wells, G. L., & Murray, D. M. (1984). Eyewit- bridge University Press. ness confidence. In G. L. Wells & E. F. Loftus, Yaniv, I., & Foster, D. P. (1997). Precision and (Eds.), Eyewitness testimony: Psychological per- accuracy of judgmental estimation. Journal of spectives (pp. 155–170). New York: Cambridge Behavioral Decision Making, 10, 21–32. University Press. Yaniv, I., & Meyer, D. E. (1987). Activa- Whittlesea, B. W. A. (2002). Two routes to tion and metacognition of inaccessible stored remembering (and another to remembering information: Potential bases for incubation not). Journal of Experimental Psychology: Gen- effects in problem solving. Journal of Experi- eral, 131, 325–348. mental Psychology: Learning, Memory, and Cog- 187 205 Whittlesea, B. W. A. (2004). The perception of nition, 13, – . integrality: Remembering through the valida- Yzerbyt, V. Y., Lories, G., & Dardenne, B. tion of expectation. Journal of Experimental (Eds.). (1998). Metacognition: Cognitive and Psychology: Learning, Memory, and Cognition, social dimensions. Thousand Oaks, CA: Sage. 30, 891–908. Zakay, D., & Tuvia, R. (1998). Choice latency Winkielman, P., & Berridge, K. C. (2004). Uncon- times as determinants of post-decisional conscious emotion. Current Directions in Psycholog- fidence. Acta Psychologica, 98, 103–115. ical Science, 13, 120–123. Zechmeister, E. B., & Shaughnessy, J. J. (1980). Winkielman, P., Schwarz, N., Fazendeiro, T. A., When you know that you know and when you & Reber, R. (2003). The hedonic marking of think that you know but you don’t. Bulletin of processing fluency: Implications for evaluative the Psychonomic Society, 15, 41–44. P1: KAE 0521857437c11 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:24

326 P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

CHAPTER 12 Consciousness and Control of Action

Carlo Umilt`a

Abstract scious visual experience. The explanation is that action execution depends on a Any voluntary action involves at least three sensorimotor or “how” system, which con- stages: intention to perform an action, per- trols visually guided behavior without access formance of the intended action, and per- to consciousness. The other is a cogni- ception of the effects of the performed tive or “what” system, which gives rise action. In principle, consciousness may man- to perception, is used consciously in pat- ifest itself at all three stages. Concerning the tern recognition, and produces normal visual ﬁrst stage, research suggests that intentions experience. The processes of this second for carrying out voluntary actions may be stage do not have access to consciousness generated unconsciously and retrospectively either. referred consciously to the action when it In contrast, we are aware of some aspects has been executed. There is a mechanism of the current state of the motor system at that binds together in consciousness the the third stage in the sequence that leads to intention to act and the consequences of the the execution of an action. When perform- intended action, thus producing the experi- ing an action, we are aware of the prediction ence of free will. of its effects, which depend on the motor Human beings consistently show visual commands that were planned in the premo- illusions when they are tested with per- tor and motor cortical areas. ceptual measures, whereas the illusions do not manifest themselves when they are tested with motor measures. These disso- Introduction: The Notion ciations concern the stage of performing of Intentionality the intended action and recall blindsight and visual agnosia, in which patients can In the present chapter, I am concerned perform visually guided tasks without con- exclusively with motor (i.e., bodily) actions.

327 P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

328 the cambridge handbook of consciousness

Although actions usually manifest them- completion, independent of intentional ini- selves as bodily movements, in accord with tiation and online intentional control. That Marcel (2004) an action can be distinguished does not mean, however, that the observer from a mere movement because the former is not conscious of the fact that those cogni- has a goal and an effect, belongs to a seman- tive processes are in progress or, to be more tic category (i.e., it has a content), and has precise, of their intermediate and final out- some degree of voluntariness. puts (i.e., representations; in fact, we are With the term “intentional action” we never conscious of the inner workings of our normally mean an action that one is con- cognitive processes, but only of their out- scious of; that is, an action that is performed puts). Very likely, therefore, the observer consciously (Marcel, 2004). Other terms becomes conscious of the representations that can be used in place of “intentional” that are produced by the cognitive processes action are “deliberate” action or “volitional” that operate automatically. Thus, the ques- action or “willed” action. As Zhu (2004, tion concerning what aspects of action we pp. 2–3) has maintained, are conscious of is relevant for both intentional and automatic actions. A central task for theories of action is In our daily life we quite often experi- to specify the conditions that distinguish ence conscious intentions to perform specific voluntary and involuntary bodily move- actions and have the firm belief that those ments. ...Ageneral way to understand the conscious intentions drive our bodily move- nature of action is to view actions as bod- ments, thus producing the desired changes ily movements preceded by certain forms of in the external world. In this view, which thought, such as appropriate combinations of beliefs, desires, intentions, and reasons. no doubt is shared by the vast majority of It is these particular forms of thought that people, the key components of an inten- characterize the voluntariness of human tional action constitute a causal chain that action. can be described as follows, though in an admittedly oversimplified way. At the begin- The critical question in Zhu’s view is, ning of the chain that leads to an inten- “[H]ow can a certain piece of thought bring tional action there is a goal-directed con- about physical bodily movement?” scious intention. As a direct consequence of It is important to point out, however, the conscious intention, a series of move- that the terms “conscious” and “intentional” ments occurs. Then, effects – that is changes should not be used interchangeably when in the external world – manifest themselves referring to action. In fact, consciously per- and are consciously linked to the inten- formed actions are not necessarily actions tion to act and to the series of performed that are performed intentionally. One can movements. be conscious of performing an action that However, an apparently goal-directed is non-intentional, or automatic, in nature. behavior does not necessarily signal inten- This distinction between conscious and tionality. There can be little doubt that intentional applies to all cognitive processes. an organism may manifest goal-directed Research has found that the vast majority behavior without satisfying the criteria for of human thinking, feeling, and behavior performing an intentional action. Most operates in automatic fashion with little or lower-order organisms do exactly that. The no need for intentional control (see, e.g., characterization of an intentional action that Dehaene & Naccache, 2001; and chapters in I have adopted implies that, to define an Umilta` & Moscovitch, 1994; also, see Proc- action as intentional, one has to consciously hazka, Clarac, Loeb, Rothwell, & Wolpaw, experience a link, through the movements 2000, for a discussion of voluntary vs. reflex- performed to achieve the goal, between the ive behavior). Once certain triggering condi- mental state of intention and the effects of tions occur, cognitive processes can proceed the performed movements. That is, one can automatically and autonomously until their claim that a given goal-directed behavior P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

consciousness and control of action 329

satisfies the criteria for an intentional action reconcile with the reductionism of neuro- only if the organism produces that behavior science and cognitive science. At first sight, along with a conscious mental representa- the toughest problem would be to pro- tion of its own internal state and of the state vide an answer to the question Zhu (2004), of the external world. among many others, asked above; that is, It is generally agreed that intentional how a mental state (i.e., the observer’s con- actions engage processes different from scious intention) interacts with the neural those engaged by automatic actions (e.g., events in the motor and premotor brain areas Prochazka et al., 2000). Here the fundamen- that produce body movements. That would tal principle is that consciousness is necessary no longer be a problem, however, if we for intentional action. In some cases, this prin- accepted that the so-called mental state is ciple is explicitly stated in one of a variety of in fact a specific neural state: It is a neural different forms, whereas in some other cases state representing or mediating intentional- it is simply implied. At any rate, the (explicit ity that acts on another neural state medi- or implicit) accepted view is that conscious ating or representing movements and their awareness of intentions, and of the motor consequences. Therefore, it is only phrasing and environmental consequences (effects) the problem the way Zhu does that creates they cause, is required to construct the sub- a dualistic separation where there may be jective experience of intentional action. none (also, see Wegner, 2005, and commen- In summary, it seems that conscious- taries therein). The question I address in the ness can manifest itself at three stages: first part of this chapter is quite different and intention to perform an action, perfor- is in fact concerned with the relative timing mance of the intended action, and percep- of two sets of neural events, those that rep- tion of the effects of the performed action. resent (accompany?) the subjective experi- It is possible that the closeness in time ence of intentional action and those that rep- of these three stages allows one to unify resent (accompany?) the execution of the the conscious experiences that accompany intended action. them. This unification in turn is how we construct the strong subjective association among intentions, actions, and action effects Intention to Perform an Action (Haggard, Aschersleben, Gehrke, & Prinz, (Unawareness of Intention) 2002a). Note that the second stage, consciousness Regardless of the difficulties of the tradi- of performing the intended action, may be tional view and the danger of dualism it subdivided into two aspects (Marcel, 2004). creates, there can be little doubt that very One aspect is consciousness of the action we often we introspectively feel we can generate are actually performing; that is, the extent to our actions; that is, we are conscious of our which we are aware of what we are doing. actions and of why they are initiated. Our The other aspect is consciousness of some experience of willing our own actions is so events that take place during the action profound that it tempts us to believe that our we are performing; that is, awareness of the actions are caused by consciousness. We first nature of the specific components and of the experience the conscious intention to per- precise details of our current action. In what form a specific action; then, after a variable follows, I am concerned almost exclusively number of intermediate states, the desired with the first aspect because very little exper- action takes place. In spite of its appar- imental evidence is available concerning the ent plausibility, this causal chain is likely second aspect. incorrect. Also, it is worth noting that this tradi- It is important to keep in mind that there tional view of how we perform intentional are two issues here, which should not be con- actions may acquire a dangerously dualist founded. The first is whether the observer’s connotation and becomes thus difficult to conscious intention (i.e., a mental state) P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

330 the cambridge handbook of consciousness

causes the neural events in the motor and the intention and the action, leaving the per- premotor brain areas, which in turn produce son to infer that the intention is causing the body movements. This is a dualistic view- action. point that is at odds with current cognitive Libet et al. (1983; also see McCloskey, neuroscience. The other is whether the neu- Colebatch, Potter, & Burke, 1983) asked ral state representing conscious intention- their participants to watch a clock face with ality precedes or follows the neural states a revolving hand and to report either the that produce movements. I am concerned time at which they “felt the urge” to make almost exclusively with this latter issue, even a freely willed endogenous movement (W if for simplicity I sometimes make recourse judgment) or the time the movement actu- to mentalistic terms. ally commenced (M judgment). The vol- Concerning the dualistic issue, suffice it untary movement consisted in flexing the to say that, contrary to what introspection wrist at a time the participants themselves seems to suggest, conscious intention that chose. Also, they were asked to note and leads to actions in fact arises as a result of to report the position of the hand when brain activity, and not vice versa (Haggard they first became conscious of wanting to & Eimer, 1999; Haggard & Libet, 2001; Hag- move; that is, the moment at which they first gard & Magno, 1999; Haggard et al., 2002a; consciously experienced the will to move. Haggard, Clark, & Kalogeras, 2002b; Libet, The W judgment was considered to be the 1985, 1999; Libet, Gleason, Wright, & Pearl, first moment of conscious intention. The 1983). Or, as Wegner (2003, page 65) puts exact moment at which the action began was it, “You think of doing X and then do X – estimated by measuring the electrical activ- not because conscious thinking causes doing, ity in the muscles involved. The prepara- but because other mental processes (that tory activity in the motor areas of the brain are not consciously perceived) cause both (the readiness potential, RP) was measured the thinking and the doing.” This sentence through the electrical activity recorded by a is phrased in mentalistic terms, but it is scalp electrode placed over the motor cor- easy to rephrase by substituting “conscious tex. The RP is a gradual increase in electrical thinking” with “neural events that represent activity in the motor cortical regions, which conscious thinking” and “other mental pro- typically precedes willed actions by 1 sor cesses” with “neural events that represent longer, and is known to be related closely to other mental processes.” the cognitive processes required to generate The celebrated experiments by Libet and the action. his collaborators (Libet 1985, 1999; Libet Assuming that the cause precedes the et al., 1983) challenged the classical notion effect, the temporal order of the W judg- of conscious intention as action initiator and ment and PR onset allows one to investigate provided evidence that the conscious inten- which event is the cause and which event is tion that is experienced does not correspond the effect. If the moment of the W judgment to causation. In a non-dualistic view, these (i.e., the moment of conscious intention or, experiments suggest that the neural events to be more precise, of the neural event rep- that represent conscious intentions, and that resenting conscious intention) precedes the we think determine voluntary action, actu- onset of RP, then the idea that conscious ally occur after the brain events that under- intention can initiate the subsequent prepa- lie action execution. Precisely, although the ration of movement is tenable. In contrast, experience of conscious intention precedes if the moment of the W judgment follows the movement (flexing a finger, for exam- the onset of RP, then conscious intention ple), it occurs well after the relevant brain (i.e., the neural state representing conscious events. This fact is taken to show that the intention) would be a consequence of activ- experience of consciously willing an action ity in the motor areas, rather than the cause begins after brain events that set the action of it. It should be noted, however, that into motion. Thus, the brain creates both Dennett (1998) has made the interesting P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

consciousness and control of action 331

point that conscious experience does not attention between the external clock and occur instantaneously, but rather develops their own internal states in order to make over time. That means that the conscious the W judgment, the value of 200 ms experience of the onset of the neural event as the interval by which the W judg- that produces a movement may evolve along ment preceded movement onset is uncer- with the movement itself. Therefore, it is tain. However, even the largest estimates of likely wrong to expect a precise point in time the prior entry effect are much smaller than when one becomes conscious of the initia- the gap between RP and W judgment that tion of a movement. Libet et al. found (e.g., Haggard & Libet, The sequence of events observed by Libet 2001; Haggard et al., 2002b). et al. (1983) is as follows: RP began between In conclusion, it seems that Libet et al.’s 1,000 and 500 ms before onset of the actual results, as well as those of Haggard and his body movement, participants only experi- colleagues (Haggard & Eimer, 1999; Haggard enced a conscious intention (i.e., W judg- & Magno, 1999; Haggard et al., 2002a,b), ments) about 200 ms before movement are fully consistent with the view that con- onset, and conscious intention was experi- sciousness of intention to perform an action enced between 500 and 350 ms after RP (i.e., the brain events that represent the con- onset. That indicated that a brain process scious intention to perform an action) is the in the motor areas initiated the intentional consequence, rather than the cause, of activ- process well before participants were aware ity in the cortical motor areas. That does not of the intention to act. The conclusion is mean to deny that there must be an identi- that the brain is preparing the action that is ﬁable link between having an intention and apparently caused by a conscious intention the action that is subsequently performed to before the participant is aware that he or she achieve the goal indicated by the intention. intends to act. All this goes directly against It simply means that the timing between the the traditional view that the conscious inten- brain events involved is not necessarily the tion to perform an action initiates the neural one corresponding to the subjective experi- events in the motor areas of the brain, which ence of the observer. in turn produce the desired action. Also, it Note, however, that the studies by Hag- challenges the folk notion of free will, imply- gard and his colleagues (Haggard & Eimer, ing that the feeling of having made a decision 1999; Haggard & Magno, 1999; Haggard is merely an illusion (e.g., Eagleman & Hol- et al., 2002a,b) suggest that the processes combe, 2002). underlying the so-called lateralized readi- The notion that free will is a mere ness potential (LRP) are more likely than construct of our own minds, and we are those underlying the RP to cause awareness only aware in retrospect of what our brain of movement initiation. The LRP measures had already started some time ago, was the additional negativity contralateral to the not accepted lightly. Several critiques of actual movement that is being performed, Libet et al.’s (1983) study were advanced in over and above that in the ipsilateral motor response to a target article by Libet (1985). cortex, and it can be considered to be an indi- The more damaging one was that partic- cator of action selection. This is based on the ipants are often poor at judging the syn- reasoning that, once the LRP has begun, the chrony of two events, and especially so if selection of which action to make must have the two events occur in different percep- been completed. In the studies by Haggard tual streams. Even more importantly, the so- and his colleagues the onset of the RP and W called prior entry phenomenon may occur, judgment did not covary, whereas the onset according to which events in an attended of the LRP and W judgment did. That is, the stream appear to occur earlier than simul- LRP for early W awareness trials occurred taneous events in an unattended stream. earlier than the LRP for late W awareness tri- Because participants in the Libet et al. als. This ﬁnding would seem to rule out the study very likely attempted to divide their RP as the unconscious cause of the conscious P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

332 the cambridge handbook of consciousness

state upon which W judgment depends, but important consequence of conscious inten- it is consistent with LRP having that role. tion. He claimed that, although actions seem Therefore, there is a clear indication that not to be initiated consciously, they may (a) the initial awareness of a motor action be inhibited or stopped consciously. This is premotor, in the sense of deriving from is because there is sufficient time between the specifications for movement rather than W judgment (i.e., W awareness) and move- the movement itself (i.e., a stage later than ment onset for a conscious veto to operate. intention but earlier than movement itself; Thus, although conscious intention to act Marcel, 2004), and (b) awareness of initi- does not seem to initiate the intended action, ating action relates to preparing a specific consciousness might still allow the uncon- action, rather than a general abstract state of sciously initiated action to go to completion, intending to perform an action of some kind. or veto it and prevent the actual action from In this connection, a study of Fried, Katz, occurring. The view of consciously veto- McCarthy, Sass, Williamson, Spencer, and ing an unconsciously initiated action may Spencer (1991) is of interest. They stimu- be altered to one according to which the lated, through surface electrodes that were action is only consciously modified (Haggard implanted for therapeutic reasons, the pre- & Libet, 2001). That clearly is not a critical motor areas (Broadmann’s area 6,BA6,in alteration. The critical point is that the neu- particular) of epileptic patients. Weak stim- ral onset of a voluntary movement precedes ulation of some of the more anterior elec- the conscious experience of having had the trode sites caused the patients to report a intention to act, and the causal role of con- conscious intention to move or a feeling that scious intention is confined to the possibility they were about to move specific body parts. of suppressing the movement. Stronger stimulation evoked actual move- It must be conceded that one does not eas- ments of the same body parts. Fried et al.’s ily accept the notion that intention of action results too are clearly consistent with a cau- follows its initiation. That is, the activation sation chain that goes from motor areas to of the neural mechanisms that give rise to intention and not vice versa. the intention to act follows the activation of The results of the studies that Haggard the neural mechanisms that cause initiation and his colleagues (Haggard & Eimer, 1999; of the action. What seems to have been asked Haggard & Magno, 1999; Haggard et al., in all studies that to date have addressed 2002a,b) conducted by following Libet’s this issue is whether perception of the time seminal work (also see Marcel, 2004) point of intentionality precedes or follows action. to the stage of motor specifications as the Suppose, however, we have a poor awareness stage at which that initial awareness of action of the time when we believe our intention- arises. By the term “motor specifications” alty occurs: Then, all the evidence collected they refer to the operations of preparing, so far becomes suspect. planning, and organizing forthcoming move- Even if one accepts the notion that the ments (including selection of the effectors), neural state of intention follows the neural which Haggard and Magno (1999) attributed state of action, it is still necessary to indicate to a specific premotor area, the Supplemen- what distinguishes voluntary from involuntary Motor Area (SMA), and which Haggard tary actions. It seems we have not yet uncov- and Eimer (1999) associated with LRPs. The ered the neural correlate of intention forma- stage of motor specifications is downstream tion, which should always be a prerequisite of the stage of intention formation, but prior for voluntary actions, regardless of whether to the stage of activation of the primary such a correlate follows or precedes the neu- motor cortex, which controls execution of ral events that lead to an action. As we see the movements themselves. in one of the following sections, a signal that It is worth pointing out that, having dis- is available to consciousness before a volun- proved the traditional concept of mind-brain tary action is initiated is the prediction of causation, Libet (1985, 1999) salvaged an the sensory consequences of the action; that P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

consciousness and control of action 333

is, the anticipatory representations of their congruence as a result either of experimen- intended and expected effects. Perhaps, this tal intervention in healthy participants or of is the direction in which to look for certain types of brain injury in neurological differences between voluntary and involun- patients. In recent years, many studies (for tary actions. reviews, see Glover, 2002, 2004; Rossetti & Pisella, 2002) have shown that introspection, on the basis of which we are convinced Performing the Intended Action of perceiving a coherent visual world that is identical to the visual world that is the (Unawareness of Action) object of our actions, is in error: At least two visual representations operate simulta- Studies on Neurologically Intact neously and in parallel when we interact Participants with the visual world. Several of those stud- In what follows I summarize evidence that ies have exploited perceptual illusions to an unconscious visual system can accurately explore double dissociations between per- control visually guided behavior even when ceptual and sensorimotor visual systems in the unconscious visual representation on non-brain-damaged participants. which it depends conflicts with the con- One of these illusions is a variation of scious visual representation on which per- the Roelofs effect (Roelofs, 1935), which ception depends. has been studied extensively by Bridgeman In the human brain, visual information and his colleagues (e.g., Bridgeman, 2002; follows many distinct pathways that give Bridgeman, Peery, & Anand, 1997). The rise to many distinct representations of the location of a rectangular frame shown off- visual world (e.g., Milner & Goodale, 1995; center in the visual field is misperceived so Rossetti & Pisella, 2002). Of these, two that a frame presented on the left side of are more important than the others. One the visual field, for example, will appear produces conscious perceptual representa- less eccentric than it is, and the right edge tions (i.e., perception) and governs object will appear somewhat to the right of the recognition, whereas the other produces observer’s center. In a typical experiment, non-conscious representations and controls the target is presented within the asym- visually guided behavior. Because of this metrically located frame, and its location organization, human beings can simultane- is misperceived in the direction opposite ously hold two representations of the same the offset of the frame. That is, mispercep- visual display without becoming aware of tion of frame position induces misperception the conflict when they are incongruent. One of target position (i.e., an induced Roelofs representation is perceptual and is perceived effect). In their studies, Bridgeman and his consciously and directly. The other represen- colleagues asked participants to perform tation is not consciously perceived, is uncon- two tasks when faced with a display that scious, and manifests itself only through induced the Roelofs effect. One task was to its behavioral effects. In spite of that, a describe the target’s position verbally, and healthy observer experiences one coher- based on the verbal response, the Roelofs ent visual world and performs fully appro- effect was reliably observed. In contrast, a priate motor actions to interact with that jab at the target, performed just after it had world, a conscious representation of which disappeared from view, was not affected by he or she is at the same time perceiving. the frame’s position: The direction of the This occurs because the outputs of the two jab proved accurate despite the consciously visual pathways normally do not conflict, perceived (and consciously reported) per- but rather lead to perceptual experiences ceptual mislocation. The result is differ- and to actions that are consistent with one ent, however, if a delay is imposed between another. Therefore, demonstrating their dis- disappearance of the target and execution sociability requires studies that disrupt this of the motor response. After a delay of P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

334 the cambridge handbook of consciousness

just 4 s, participants have the tendency observed in two tasks in which participants to jab in the direction of the perceptual were required to reproduce the length of the mislocation. line with the same two fingers. As is the case Aglioti, DeSouza, and Goodale (1995) for the Roelofs effect (see above), the dif- made use, for a similar purpose, of ference between grasp and perception dis- the Ebbinghaus illusion (also called the appeared when participants were asked to Titchener circle illusion). In it, a circle is pre- delay their response. With the delay, the illu- sented in the center of a circular array, com- sion in the motor condition became as large posed of circles of either smaller or larger as in the perceptual condition. than the central one. The circle in the cen- Additional evidence of a dissociation ter appears to be larger if it is surrounded between perception and action derives from by smaller than by larger circles. One can other studies that made use of paradigms dif- build displays with central circles of physi- ferent from those based on visual illusions. cally different sizes that appear perceptually The perturbation paradigm is one of these equivalent in size. In a 3-D version of this (see review in Desmurget, Pelisson, Rossetti, illusion, Aglioti et al. required participants & Prablanc, 1998). It involves a task in which to grasp the central circle between thumb the participant is asked to reach and grasp and index finger and measured the maximal a target. Then, often coincident with the grip aperture during the reaching phase of onset of the movement by the participant, a the movement. They found that grip size characteristic of the target, typically its loca- was largely determined by the true, physi- tion and/or size, is changed suddenly. Many cal size of the circle to be grasped and not studies have demonstrated the ability of the by its illusory size. In a subsequent study, sensorimotor system to adjust to change in Haffenden and Goodale (1998) measured the characteristics of the target well before the circle illusion either by asking partic- the perceptual system can even detect the ipants to indicate the apparent size of a change. For example, Paulignan, MacKen- circle or to pick it up, without vision of zie, Marteniuk, and Jeannerod (1991) placed hand and target. In both tasks the dependent three dowels on a table and, by manipu- variable was the distance between thumb lating the lighting of the dowels, were able and forefinger, so that the output mode was to create the impression that the target had controlled, and only the source of informa- changed location on some trials. They found tion varied. The illusion appeared in both that the acceleration profile of the grasping tasks, but was much smaller for the grasp movement changed only 100 ms after the response. perturbation. Castiello, Bennett, and Stel- Similarly, Daprati and Gentilucci (1997; mach (1993) studied the effect of size per- also see Gentilucci, Chieffi, Daprati, Saetti, turbation on hand shaping in a thumb and & Toni, 1996) used the Mueller-Lyer illusion finger grasp of a target object and found that for contrasting grasp and perception. This hand shaping could respond to a size pertur- illusion induces the perception of longer or bation in as little as 170 ms. shorter length of a line ended by outward- In a series of experiments (Castiello or inward-pointing arrows. In the study by & Jeannerod, 1991; Castiello, Paulignan, & Daprati and Gentilucci, participants were Jeannerod, 1991; also, see Jeannerod, 1999), required to reach and grasp a wooden bar participants made a simple vocal utterance that was superimposed over the line. Results to signal their awareness of the object per- showed that the illusion was smaller when turbation in a version of the perturbation measured with grasp than with perception, paradigm. Comparison of the hand motor even though there was some illusion under reaction time and the vocal reaction time both conditions. That is, hand shaping while showed that the vocal response consistently grasping the bar with thumb and index took place after the motor correction had finger was influenced by the illusion con- started. As in the Paulignan et al. (1991) figurations on which it was superimposed. study, the change in the hand trajectory This effect, however, was smaller than that occurred as early as 100 ms following the P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

consciousness and control of action 335

object’s perturbation, whereas the vocal tions. Not even forced-choice guesses about response by which they reported awareness the direction of the jump could discrimi- of the perturbation was not observed until nate between forward and backward target more than 300 ms later. The conclusion perturbations. was that awareness of the perturbation Visual masking too has been used to study lagged behind the motor action performed the dissociation between motor control and in response to this perturbation. conscious perception (e.g., Kunde, Kiesel, & Because our perceptual system is able to Hoffman, 2003; Neumann & Klotz, 1994; disregard motion of images on the retina dur- Taylor & McCloskey, 1990; Vorberg, Mat- ing eye movements, it is very difficult to tler, Heinecke, Schmidt, & Schwarzbach, detect small displacements (perturbations) 2003; see Bar, 2000, and Price, 2001, for that occur in the visual field during a saccade; reviews). It seems that masking (and meta- that is, a rapid eye movement from one point contrast) eliminates conscious perception to another. Often objects can be moved sev- of the stimulus, whereas the ability of the eral degrees of visual angle during a saccade (non-perceived) stimulus to trigger an action without the displacement being perceived. (a motor response) remains largely intact. This phenomenon is known as saccadic sup- In the study of Neumann and Klotz, for pression (see, e.g., Bridgeman, Hendry, & example, the observer was unable to dis- Stark, 1975, and Chekaluk & Llewelynn, criminate reliably the presence from the 1992, for a review). A seminal study on this absence of the masked stimulus, but the subject is the one by Bridgeman, Lewis, Heit, masked (and undetected) stimulus affected and Nagle (1979). It was instrumental in the speed of voluntary responses, even in a starting the whole line of research on the two-choice situation that required integrat- distinction between processing for action, ing form information with position infor- which produces non-conscious representa- mation. Thus, this study clearly confirmed tions, and processing for perception, which that motor action in response to a visual produces conscious representations. Partic- stimulus can be dissociated from the verbal ipants were asked to point at a target that report about detection of that same stimu- had been displaced during the saccade and lus. Similarly, Vorberg et al. (2003) showed then extinguished. It was found that the that experimental manipulations that mod- displacement often went undetected (sac- ify the subjective visual experience of invis- cadic suppression), but was not accompa- ible masked stimuli do not affect the speed nied by corresponding visuomotor errors. of motor responses to those same stimuli. In That is, the pointing movement after a tar- addition, they found that, over a wide range get jump remained accurate, irrespective of of time courses, perception and unconscious whether the displacement could be verbally behavioral effects of masked visual stimuli reported or not. Not only did participants obey different temporal laws. fail to detect the target displacement but The studies I have summarized above they also failed to detect their own move- used different procedures, but the results ment corrections. converge toward supporting a rather coun- The paradigm just described is also called ter-intuitive conclusion: An unconscious the double-step paradigm, in which the first visual system controls visually guided actions step is target presentation and the second and operates more or less simultaneously step is target displacement. It was further with the conscious visual system. When exploited by Goodale, Pelisson,´ and Prablanc the representations produced by the uncon- (1986) and Pelisson,´ Prablanc, Goodale, and scious visual system conflict with the repre- Jeannerod (1986). They confirmed that par- sentations produced by the conscious visual ticipants altered the amplitude of their system, the former prevail and guide action. movements to compensate for (most) of Here, as well as in the studies that are sum- the target displacement, even though they marized in the following section, two disso- were not able to detect either the tar- ciations emerge. One is between (conscious) get jump or their own movement correc- perceptual and (unconscious) sensorimotor P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

336 the cambridge handbook of consciousness

visual systems. The other, more interesting ticipants, grip size still correlates well with dissociation is between awareness of the rep- object width even when a temporal delay of resentations, produced by the perceptual up to 30 s is interposed between disappear- system, which one believes guide action, ance of the target object and execution of and the lack of awareness of the representa- the motor response. In DF, instead, evidence tions, produced by the sensorimotor system, of grip scaling was no longer present after a which actually guide action. delay of just 2 s. If DF’s performance is compared with the performance of patients with posterior Studies on Brain-Damaged Patients parietal lesions, impairments in perceptual The deficits observed in some neuropsycho- recognition of objects and in object-directed logical patients lend further (and perhaps action seem to be clearly dissociated. more convincing) support to the notion of Jeannerod (1986) and Perenin and Vighetto dissociability between perceptual and sen- (1988) reported patients with lesions to the sorimotor visual systems. In particular, optic posterior parietal lobe, who showed a deficit ataxia and visual agnosia patients not only that is termed optic ataxia. Patients with strongly support the case for a double dis- optic ataxia have difficulties in directing sociation between perceptual recognition of actions to objects presented in their periph- objects and reaching and grasping of the eral visual field. Their visually directed same objects but also suggest that the neuro- reaching movements are inaccurate, often logical substrates for these two systems are systematically in one direction. In addi- located selectively in the ventral (object per- tion, these movements are altered kinemat- ception) and dorsal (object-directed action) ically, especially concerning duration, peak streams of the visual pathways (e.g., Milner velocity, and deceleration phase. However, & Goodale, 1995). patients with optic ataxia, in contrast to A patient, DF, first reported by Goodale, visual agnosic patients, are not impaired in Milner, Jakobson, and Carey (1991; also see the recognition of the same objects that they Milner & Goodale, 1995), perhaps provides are unable to reach correctly. the clearest evidence to date for dissocia- In addition to optic ataxia and visual tion between perception and action, show- agnosia, blindsight is another neurological ing a reciprocal pattern to optic ataxia (see deficit that provides support to the disso- below). DF developed a profound visual- ciation between conscious perception and form agnosia following a bilateral lesion non-conscious motor control. Patients with of the occipito-temporal cortex. She was extensive damage to the primary visual area unable to recognize object size, shape, and in the occipital lobe (area V1 or BA17) orientation. She failed even when she was are regarded as cortically blind because asked purposively to match her forefinger- they do not acknowledge seeing stimuli in thumb grip aperture to the size of visually the affected parts of their visual fields. It presented target objects. In sharp contrast, is possible, however, to demonstrate that when instructed to pick up objects by per- their behavior can still be controlled by forming prehension movements, the patient visual information (e.g., Farah, 1994; Mil- was quite accurate, and the maximum size of ner & Goodale, 1995; Weiskrantz, 1986). her grip correlated normally with the size of The paradoxical term “blindsight” was ini- the target object. Apparently, DF possessed tially coined by Sanders, Warrington, Mar- the ability to reach out and grasp objects shall, and Weiskrantz (1974) to refer to all with remarkable accuracy and thus could such non-conscious visual capacities that process visual information about object fea- are spared in cortically blind parts of the tures that she could not perceive accurately, visual field. The first of the many reports if at all. However, although her ability to that patients with cortical blindness can grasp target objects was truly remarkable, use visuospatial information to guide their it had certain limitations. In normal par- actions within the blind field showed this by P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

consciousness and control of action 337

asking patients to move their eyes toward a The evidence summarized above strongly light that they insisted they could not see. supports the notion that the human brain Their movements were statistically corre- produces at least two functionally distinct lated with the location of the light (Poeppel, representations of a given visual display, Held, & Frost, 1973). Further, Zihl (1980; which, under some conditions, can be dif- Zihl & von Cramon, 1985) found that accu- ferent. One originates from a conscious sys- racy of their saccadic responses can be tem that performs visual object (pattern) improved markedly as a consequence of recognition. The other is a motor-oriented training. Even more striking evidence was representation that is unconscious, can be provided by Perenin and Jeannerod (1978), in conflict with the conscious perceptual who showed accurate pointing within the representation, and can accurately control blind fields in several patients. It is clear visually guided behavior in spite of the that considerable visual control of the direc- potential conflict with the conscious rep- tion and amplitude of both eye and arm resentation. The distinction between these movements is present in cortically blind two pathways originated nearly 40 years patients. ago (Schneider, 1967; Trevarthen, 1968). Tre- In addition, it turns out that some patients varthen named the two systems “focal” and with blindsight seem to have a residual “ambient”: The focal system was postulated ability to use shape and size information. to be based on the geniculostriate path- Perenin and Rossetti (1996) asked a com- way and to be devoted to pattern recogni- pletely hemianoptic patient to “post” a card tion; the ambient system was thought to be into a slot placed at different angles within based on the superior colliculus and related his blind field. His performance was statis- brainstem structures and to be devoted to tically well above chance. Yet, when asked visually guided behavior. This anatomical to make perceptual judgments of the slot’s and functional distinction became known orientation, either verbally or by manual as the now classical distinction between a matching, the patient was at chance level system specialized for answering the ques- in his affected field. Perenin and Rossetti tion “what is it” and a system specialized also demonstrated that the patient, when for answering the question “where is it” tested with the same tasks as those used (Schneider, 1969). Later on, both systems to test agnosic patient DF, and the task of were shown also to have a cortical repre- “posting” the card was one of them, could sentation: The successor to the focal sys- reach out and grasp rectangular objects with tem (i.e., the “what” system) comprises a certain accuracy in his affected field: As an occipito-temporal pathway, whereas the in normal participants, the wider the object, ambient system (i.e., the “where” system) the greater the anticipatory hand-grip size includes an occipito-parietal pathway as well during reaching. Yet again the patient failed as the superior colliculus (Ungerleider & when asked to make perceptual judgments: Mishkin, 1982). With either verbal or manual response his More precisely, the cortical pathway for attempts were uncorrelated with object size. the “what” system is the ventral occipito- As was the case with agnosic patient DF, ori- temporal route that links striate cortex (V1 entation and size can be processed in corti- or BA17) to prestriate areas and from there cally blind visual fields, but only when used reaches the inferotemporal cortex on both to guide a motor action and not for percep- sides via callosal connections. Lesions to this tual tasks. Remember, however, that patients pathway abolish object discrimination with- with optic ataxia show the converse dissoci- out damaging perception of spatial relations ation. They are unable to use visual infor- among objects (visual-form agnosia). The mation to guide motor acts, such as reaching other, dorsal pathway diverges from the ven- and grasping, but they still retain the ability tral one after the striate cortex and links to perceive consciously the objects on which the prestriate areas to the posterior part of they are unable to act. the parietal lobe. Lesions to this pathway P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

338 the cambridge handbook of consciousness

produce visuospatial deﬁcits characterized ular, the delay seems to bring about a shift by errors in establishing the relative posi- from reliance on a (non-conscious) represen- tions of spatial landmarks and by localiza- tation formed online, and used for immedi- tion deﬁcits during object-directed actions ate action control, to a (conscious) represen- (optic ataxia). tation retrieved from memory and used for As described above, cases of optic ataxia, verbal report. The former would be short- visual-form agnosia, and blindsight, as well lived and would compute stimulus location as a number of studies on healthy par- in egocentric coordinates, whereas the latter ticipants, have shown that the anatomical would last much longer and would compute dorsal-ventral distinction relates more prestimulus location in allocentric coordinates. cisely to a distinction between the process- Egocentric coordinates have their origin in ing of what an object is and of how to direct the observer, often in its body midline. Allo- action, rather than of where an object is centric coordinates have their origin outside located (Milner & Goodale, 1995). the observer, often in an external object. The phenomenon of blindsight merits a While an action is being performed, the loca- few additional words (Milner & Goodale, tion of its target seems to be computed in 1995; Stoerig & Cowey, 1997). The dorsal terms of egocentric coordinates. In contrast, stream also has substantial inputs from sev- if a delay is introduced between target pre- eral subcortical visual structures in addition sentation and verbal report, target location to input from V1. In contrast, the ventral seems to be computed in terms of allocentric stream depends on V1 almost entirely for coordinates. its visual input. As a consequence, damage Evidence is available that supports this to V1, even though it affects either stream, distinction. In a study reported by Rossetti deprives the ventral stream of all its input, and Pisella (2002), for example, immedi- leaving instead the dorsal stream with much ate and delayed pointing toward propriocep- of its input from its associated subcortical tive targets was tested in blindfolded par- structures (the superior colliculus, in partic- ticipants. On each trial a target was pre- ular). It is possible, thus, to demonstrate that sented on one out of six possible locations their behavior can be controlled by visual lying on a circle centered in the starting information provided by intact subcortical point. In a preliminary session participants visual structures via the dorsal stream. What had been trained to point to these positions these patients are unable to do is to pro- and to associate a number (from 1 to 6) cess visual information through the ventral with each target. When the task was to stream, which depends primarily on V1 for point to the target immediately, the partici- its visual input. In addition, it is impossible pants’ pointing distribution was unaffected for these patients to give any but the most by the target array and was elongated in rudimentary kind of perceptual report about the movement direction. That showed the their visual experiences (regions of bright- use of an egocentric reference frame. In ness, motion, or change). contrast, when pointing was delayed and/or As already noted, differences between was accompanied by a simultaneous verbal immediate and delayed actions have been response of the target number, the partici- reported in neurologically intact observers, pants’ pointing distribution tended to align as well as in neurological patients (see Ros- with the target array, perpendicular to move- setti & Pisella, 2002, for an extensive discus- ment direction. That showed the use of an sion of this issue). In general, what emerges is allocentric frame of reference. In the words that the insertion of a temporal gap between of Rossetti and Pisella (2002,p.86), “when target presentation and response execution action is delayed and the object has dis- renders the features of the representation appeared, the parameters of object posi- on which the motor action depends more tion and characteristics that are used by the similar to the perceptual representation on action system can only be accessed from a which the verbal report depends. In partic- cognitive sustained representation. This type P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

consciousness and control of action 339

of representation . . . relies on different refer- of executive functions (see below), we are ence frames with respect to the immediate not aware of the predicted consequences of action system.” our actions, as well as of the intended actions In conclusion, it is clear that the sen- and of the motor programs that are executed sorimotor representations that support to achieve the actions’ goals. immediate action are not conscious and Representations of action effects have are short-lived. If a delay is introduced, at least two functions in performing an long-term and conscious perceptual repre- action (a special issue of Psychological sentations take over. The frames of reference Research/Psychologische Forschung, edited by on which immediate and delayed actions Nattkemper and Ziessler, 2004, was devoted depend also differ, consistent with the to the role of action effects in the cog- needs of an action system and a perceptual nitive control of action). First, after hav- representational system. ing executed a particular action, one needs to compare the obtained effects with the effects that the action intended to accom- Perception of the Effects of the plish. Hence, anticipatory effect represen- Performed Action tations are involved in the evaluation of action results. Second, one plans and exe- So far I have reviewed evidence that sug- cutes actions with the aim of producing gests that many aspects of action, from ini- some desired effects. Hence, anticipations of tiation to appreciation of the percepts that action goals are involved in action control. guide them, occur without awareness (i.e., Both functions require representations of an unconsciously). Now I argue that one aspect action goal and of action effects. of an action that is normally available to The notion that intentional actions are awareness is the sensory consequence(s) of controlled by anticipatory representations that action, or, more precisely, the prediction of their intended and expected effects goes of the sensory consequences of that action back at least to James (1890). It is termed (Blakemore & Frith, 2003; also see Frith, the “ideomotor theory,” which basically pro- Blakemore, & Wolpert, 2000; Jeannerod, poses that actions are represented in mem- Farrer, Franck, Fourneret, Posada, Daprati, & ory by their effects and in turn these effects Georgieff, 2003). However, although there are used to control actions. In the last is only limited awareness of the actual sen- 15 years or so, the ideomotor approach has sory consequences of an action (i.e., of action been reformulated within the framework of effects) when they are successfully predicted cognitive psychology by Prinz and his colin advance, we are very often aware of the leagues (e.g., Prinz, 1997; Hommel, Mues- actual action effects when they deviate from seler, Aschersleben, & Prinz, 2001), who what we expect. used the term “common coding theory of As was pointed out by Blakemore and perception and action.” Frith (2003), here the important point is The simplest idea would, of course, be to distinguish between the predicted action that our awareness of initiating an action effects and the actual action effects. Nor- originated from sensory signals arising in the mally, we are aware of the former but not moving limbs. This seems unlikely, though, of the latter, unless the latter do not con- because such signals are not available form to our expectations, in which case until after the limbs have started to move. they too become conscious. In some circum- Instead, our awareness seems to depend on a stances, however, not even quite large devia- signal that precedes the action. A signal that tions from the expected action effects reach is available before an action is initiated is consciousness (see, e.g., Fourneret & Jean- the prediction of the sensory consequences nerod, 1998). When a task is overlearned and of the action; that is, the anticipatory rep- becomes automatic with practice, and thus resentations of their intended and expected can be carried out without the intervention effects. P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

340 the cambridge handbook of consciousness

Haggard and Magno (1999), by using ally were. Awareness of the voluntary key- Libet’s paradigm (see above), found a disso- press was shifted later in time, toward the ciation between perceived onset and action consequent tone, whereas awareness of the initiation. They showed that the perceived tone was shifted earlier in time, toward time of action onset is slightly delayed (by the action. The involuntary, TMS-induced about 75 ms) if the motor cortex is stim- movement produced perceptual shifts in the ulated with transcranial magnetic stimula- opposite direction. The sham TMS showed tion (TMS), whereas this stimulation causes minimal perceptual shifts. Based on these a greater delay (of about 200 ms) in the initi- findings Haggard et al. (2002b) suggested ation of the actual movement. This finding is the existence of a mechanism that associates compatible with the notion that it is the pre- (or binds) together awareness of a voluntary diction of the effects of the action that corre- action and awareness of its sensory conse- sponds to the onset of the action. However, quences, bringing them closer in perceived it is not evidence of such a relation. In a sub- time. In other words, “the brain would pos- sequent study, which too exploited Libet’s sess a specific mechanism that binds inten- paradigm, Haggard et al. (2002b; also, see tional actions to their effects to construct a Eagleman & Holcombe, 2002; Frith, 2002) coherent conscious experience of our own explored the time course of the binding of agency” (p. 385). Note that several other actions and their effects in consciousness by behavioral studies confirmed that represen- investigating the sensory consequences of an tations of actions and their effects tend to event being causally linked to an observer’s be integrated (see chapters by Hazeltine, by action. Their aim was to clarify what hap- Stoet & Hommel, and by Ziesser & Nattkem- pens to our subjective judgment of timing of per in Prinz & Hommel, 2002) and the obser- events when an event is causally linked to an vation of the existence of attraction effects observer’s intentional action. They showed between percepts of stimuli and percepts that the perceived time of intentional actions of movements that might have caused the and the perceived time of their sensory con- occurrence of those same actions. sequences were attracted together in con- Even if ample empirical evidence sug- sciousness, so that participants perceived gests that motor actions are cognitively intentional movements as occurring later coded based on their sensory effects, it does and their sensory consequences as occur- not imply that they are initiated by con- ring earlier than they actually did. In the sciously accessing their sensory effects. It voluntary condition, participants noted, by is entirely possible that binding between watching a revolving hand on a clock face, actions and their consequences and/oraction the time of onset of their intention to per- retrieval via action consequences occur form an action (i.e., a key-press). In the TMS unconsciously (Haggard et al., 2002a). For condition, they noted the time of a muscle example, Kunde (2004) reasoned that, based twitch produced by magnetic stimulation of on the ideomotor theory, initiating a cer- the motor cortex. In the sham TMS condi- tain action is mediated by retrieving its tion, they noted the time of an audible click perceptual effects, which then in turn acti- made by TMS in the absence of motor stim- vate the particular motor pattern that nor- ulation. In the auditory condition, they just mally brings about these anticipated per- noted the time of a tone. ceptual effects. Also, the ideomotor theory When the first three conditions were fol- implies that actions (as motor patterns) lowed, after 250 ms, by a tone, large percep- should become activated by presenting the tual shifts occurred in the time of awareness effect codes that represent their conse- of conscious intention, the TMS-induced quences. In other words, actions should be twitch, and click produced by the sham induced by perceiving their effects. Kunde TMS. Only awareness of the voluntary key- found that responding to a visual target was press and awareness of the tone were per- faster and more accurate when the target ceived as being closer in time than they actu- was briefly preceded by the visual effect P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

consciousness and control of action 341

of the required response. Interestingly, this Umilta,` Fogassi, Gallese, & Rizzolatti, 2002; response priming that was induced by the Umilta,` Kohler, Gallese, Fogassi, Fadiga, & action effects was independent of prime per- Rizzolatti, 2001). Visual clues of the action ceptibility and occurred when the prime was can be critical to trigger mirror neurons only not consciously perceived. That indicates if they allow the observer to understand that consciousness is not a necessary condi- the action. Auditory clues originating from tion for action effects to evoke their associ- an action performed behind a screen can ated actions (motor patterns). replace visual clues if auditory clues convey Regardless of whether binding among the crucial information about the meaning of intention, action, and consequence occurs the action. However, an fMRI study (Buc- consciously or unconsciously, or, more likely, cino, Lui, Canessa, Patteri, Lagravinese, & in part consciously and in part uncon- Rizzolatti, 2004) showed that only actions sciously, it has both unconscious and con- belonging to the motor repertoire of the scious elements (Haggard et al., 2002a); observer excite the human mirror system, inference based on timing plays a critical whereas actions that do not belong to the role in producing the illusion of consciously motor repertoire of the observer do not. willed actions. The conscious experiences of They are recognized through a different, intention, action, and action effects are com- purely visual mechanism. pressed in time. This unification is instru- Especially interesting in the present con- mental in producing the experience of vol- text are studies (e.g., Schubotz & von Cra- untary action. Even though he overlooked mon, 2002) that showed that the human the role of action effects, Wegner (2003, frontal mirror region is important not p. 67) has clearly expressed this notion, by only for the understanding of goal-directed saying, “When a thought appears in con- actions but also for recognizing predictable sciousness just before an action (priority), is patterns of visual change. In conclusion, it consistent with the action (consistency), and seems clear that mirror neurons play a role is not accompanied by conspicuous alter- in coding intended actions and their conse- native causes of the action (exclusivity), quences. we experience conscious will and ascribe authorship to ourselves for the action.” Before leaving this section it is perhaps Executive Functions useful briefly to touch on the work on mirror neurons. Mirror neurons (see Rizzolatti Another popular view is that investigation of & Craighero, 2004, for a recent and exten- the processes underlying intentional actions sive review) are found in the premotor and are related closely to executive functions parietal cortex of the monkey. They selec- (also termed “control processes”), such as tively discharge both when the monkey per- those involved in planning, problem solv- forms a given action and when the mon- ing, inhibition of prepotent response, and key observers another living being perform response to novelty. If the notion is to be a similar action. Thus, mirror neurons code preserved that consciousness is necessary for specific actions performed by the agent or intentional actions, and intentional actions by others. Evidence exists that a mirror- belong to the realm of executive functions, neuron system, similar to that of the mon- then drawing on the Supervisory Attentional key, is present in humans. The notion that System model (SAS; Norman & Shallice, mirror neurons are involved in action under- 1986; Shallice, 1988, 1994) can be useful in standing is corroborated by the observation this context. that they discharge also in conditions in It is widely accepted that the vast major- which the monkey does not see the action ity of our cognitive processes operate in but nonetheless is provided sufficient clues automatic fashion, with little or no need to create a mental representation of what for conscious, intentional control: When action is being performed (Kohler, Keysers, specific conditions occur, automatic mental P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

342 the cambridge handbook of consciousness

processes are triggered and run autonomo- contain a variety of special-purpose subsys- usly to completion, independent of inten- tems localized in different parts of the pre- tional initiation and conscious guidance frontal cortex and in the anterior cingulate. (e.g., Dehane & Naccache, 2001, and chap- This model relates to consciousness ters in Umilta` & Moscovitch, 1994). In of action if one assumes that conscious the information-processing accounts of con- processes require the mediation of the SAS sciousness developed in the 1970s, the uni- and lead directly to the selection in con- tary nature and control functions of contention scheduling of a schema for thought sciousness were explained in terms of the or action. On this view, consciousness of a involvement of a limited capacity higher- particular content causes the selection of a level processing system (Mandler, 1975; Pos- schema. Once a schema is selected, and pro- ner & Klein, 1973; Umilta,` 1988). However, vided that it does not conflict with a strongly because of the diversification of processing established schema for a different action (as systems that originated from the research in cases requiring inhibition of prepotent in cognitive psychology, cognitive neuropsy- response), then action may proceed without chology, and cognitive neuroscience, and any transfer of information from the SAS. because of the realization that processing Therefore, the principle that consciousness systems are often informationally encapsu- is necessary for intentional action can be lated (Fodor, 1983; Moscovitch & Umilta,` stated more precisely as the hypothesis that 1990), it became less plausible to associate tasks involving intentional action recruit the unitary characteristics of consciousness conscious processes, whereas automatic with the operations of any single process- actions do not. Keep in mind, however, that, ing system. Shallice (Norman & Shallice, as Block (1995) argued, such attempts to 1986; Shallice, 1988, 1994) put forward an account for the functional role of conscious alternative approach by proposing that a information (“access-consciousness”) do not number of high-level systems have a set of address the phenomenological properties of special characteristics that distinguish them conscious experience (“phenomenological- from the cognitive systems devoted to rou- consciousness”). tine informationally encapsulated processes. Some results are especially important He maintained that the contrast between for a precise specification of the rela- the operations of these special systems and tion between consciousness and intentional those realizing informationally encapsulated action. Studies on blindsight (see above) processes corresponded in phenomenologi- show that awareness of the location of the cal terms to that between conscious and non- stimulus is not necessary for accurate per- conscious processes. formance on a simple pointing task when His model is clearly concerned with participants are asked to guess. As was main- action selection and thus is very relevant tained by Marcel (1988; also, see Natsoulas, to the present discussion. It has three main 1992), blindsight patients can learn to ini- processing levels. The lowest is that of tiate goal-directed actions (e.g., reaching), special-purpose processing subsystems, each which means that their actions can be visu- specialized for particular types of opera- ally guided in the absence of conscious visual tions. There are a large number of actions representations on which the actions are and thought schemas, one for each well- based. Very strikingly, patients protest that learned routine task or subtask. Schemas they are not performing the visually guided are selected for operation through a process behavior that they are in fact performing. involving mutually inhibitory competition However, awareness of the presence of the (contention scheduling). To cope with non- visual stimulus has to be provided by an routine situations, an additional system – auditory cue for the initiation of the point- SAS – provides modulating activating input ing action to occur (Weiskrantz, 1997). That to schemas in contention scheduling. In later seems to indicate that the blindsight patient versions of the SAS, the SAS is held to requires input via the SAS to initiate a P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

consciousness and control of action 343

pre-existing schema for pointing. Once that ing while driving (Norman, 1981; Norman & schema is initiated, non-conscious informa- Shallice, 1986; Reason, 1984; Shallice, 1988). tion held in special-purpose processing sys- It is not clear whether it would be cor- tems can serve to guide action. rect to speak of these actions as completely Lesion studies in monkeys and humans unconscious. Probably not. These anoma- indicate that prefrontal lesions (i.e., SAS lous cases are explained by distinguishing lesions) have little effect on performance in between the influence of the stimulus on automatic tasks, but instead have a strong contention scheduling, and the influence on effect on tasks that seem to be crucially the SAS. The selection of well-learned and dependent on conscious processes, among relatively undemanding schemas need not which is the spontaneous generation of require the SAS involvement. As already intentional actions. A striking consequence mentioned in the context of action effects, of a deficit of the SAS is the so-called overlearned tasks are common examples of utilization behavior sign. It manifests itself actions (schemas) that can be carried out as a component of a very grave dysexecu- with very limited awareness. With sufficient tive syndrome and is typically observed in practice many tasks can become automatic patients with a bilateral focal frontal lesion. and can be carried out without the interven- If there is some object that can be used tion of the SAS; that is, without any need to or manipulated within the patient’s field consciously control the actions required by of view and within reach, the patient will the task. use it to perform actions appropriate to the object, even though they have been explicitly and insistently asked not to do Deficits in the Control of Action That so. It is clear that in utilization behavior Co-Occur with Abnormalities in the non-intended actions are environment- Consciousness driven with the mediation of contention scheduling. The SAS plays very little role or At first sight one of the most striking abnor- no role at all in them. In effect, most current malities in the control of action would not explanations of this bizarre behavior sug- seem to be accompanied by abnormality of gest a weakening of whatever mechanism is consciousness, but rather by disownership of responsible for ensuring the implementation action. This abnormality is termed the “anar- of intended actions, such that it overrides, chic hand” sign, which was first described under normal circumstances, automatic or by Della Sala, Marchetti, and Spinnler (1991, environmentally driven actions. Apparently, 1994; also see Frith, Blakemore, & Wolpert, utilization behavior is attributable to the 2000; Marcel, 2004, for extensive discus- failure to inhibit inappropriate actions, sions). It is often confused with the “alien rather than a failure to select appropriate hand” sign (Marchetti & Della Sala, 1998; actions. Prochazka et al., 2000) or with the utiliza- Environmentally driven actions are quite tion behavior sign (e.g., Lhermitte, 1983; common in normal people under circum- also, see above), which too are abnormalities stances that suggest a diminished influence in the control of action. The anarchic hand of the SAS, such as when we are in a sign is unimanual, and patients describe distracted state or when we are engaged the anarchic hand as having a “will of its in another task; this is especially so when own,” which, of course, terrifies them. The the level of arousal is low. In these cases, affected hand performs unintended (even an apparently intentional action is initiated socially unacceptable) but complex, well- in the absence of full awareness. Examples executed, goal-directed actions that com- include reaching for and drinking from a pete with those performed by the non- glass while talking, slips of highly routine affected hand. Sometimes the patient talks actions that involve action lapses of the “cap- to his or her anarchic hand, asking it to desist, ture” error type, and changing gears or brak- and often the patient succeeds in stopping P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

344 the cambridge handbook of consciousness

it only by holding it down with the other vincingly argued that that is not the case. His hand. The patient seems to lack any sense main point is that the patient who shows uti- of intention for what concerns the anarchic lization behavior does not disown the hand action performed by the affected hand. that performs it. If the SMA is the source of It is important to keep in mind that the LRP (see above) and the LRP is corre- the patient is aware that the anarchic hand lated with voluntary action, it makes sense belongs to him or her. What is alien, what that damage to the SMA would be associ- the patient disowns, is the action the hand ated with lack of awareness, or distortions in performs, not the hand itself. In contrast, in awareness, associated with movement initi- the case of the alien hand sign, the affected ation and execution. hand does not feel to be one’s own. This A possibility, with which Marcel (2004) seems to be a sensory phenomenon and has would not agree, though, is that disowner- little to do with motor action. As I have ship of action in the anarchic hand sign is already noted when I discussed the dysex- attributable to lack of awareness of the inten- ecutive syndrome, patients with damage to tion that guides the affected hand’s behavior. the frontal lobe may show utilization behav- In other words, the patient would disown ior that is characterized by the fact that the the action (not the hand that performs it, sight of an object elicits a stereotyped action as happens in the alien hand sign) because that is inappropriate in the current context. the intention that has initiated that action The abnormal aspect of consciousness that is not available to consciousness. If that is accompanies utilization behavior is not so so, then the anarchic hand sign would after much lack of awareness of the action or all be a deﬁcit in the control of action that lack of awareness of the intention to act, co-occurs with an abnormality of conscious- but rather the erroneous experience that ness. Perhaps, it is caused by an abnormal- those inappropriate and unwanted actions ity of consciousness of the predicted action are experienced as intended (also, see Frith effects. et al., 2000). By following Frith et al. (2000), I have The anarchic hand sign is often associ- maintained (see above) that what normally ated with unilateral damage to the SMA con- reaches consciousness is the prediction of tralateral to the hand affected by the patho- the action effects, rather than the actual logical sign. Considering that the function action effects. Frith et al. suggest that the of the anterior part of the SMA is likely to same is true of the state of the limbs after a be essentially inhibitory, and a movement movement; that is, the conscious experience can only be initiated by the primary motor of a limb would normally be based on its cortex (M1 or BA4) when activity in the predicted state, rather than on its actual anterior part of the SMA declines, Frith et state. In effect, the predicted state would al. (2000) have suggested that the anarchic play a greater role than sensory feedback. hand sign manifests itself when the anterior That, after all, is not surprising, considering part of the SMA is damaged. In the absence that one of the effects of an action is to of the inhibitory inﬂuence by the SMA on bring about a new state of the limbs that one side, appropriate stimuli would trigger were involved in performing the action. the automatic action of the corresponding As argued by Frith et al. (2000), the hand. In support of the inhibitory role of notion that the conscious experience of a the SMA is the observation of its preferen- limb originates from its predicted state is tial activation when movements are imag- supported by the phenomenon of the “phan- ined but their execution must be inhibited. tom limb.” After amputation of all or part It is clear that this hypothesis renders the of a limb, patients may report that, in spite anarchic hand sign very similar, if not iden- of the fact they know very well that there tical, to the utilization behavior sign. Actu- is no longer a limb, they still feel the pres- ally, the former would be a special case of ence of it. Although the limb is missing the latter. Marcel (2004), however, has con- because of the amputation, the premotor P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

consciousness and control of action 345

and the motor cortex can still program a ing a movement is based on a representa- movement of it, which causes computation tion of the predicted action effects, rather of the predicted state. Because the predicted than of the actual action effects. A repre- state becomes available, the phantom limb sentation of the predicted action effects can will be experienced as moving. It is interest- be formed as long as the motor commands ing to note that Ramachandran and Hirstein can be issued. Thus, a patient with a par- (1998) have proposed that the experience of alyzed limb would have the normal expe- the phantom limb depends on mechanisms rience of initiating a movement with that located in the parietal lobes, and Frith et al. limb as long as the motor commands can (2000) have independently suggested that be issued. The belief that the movement parietal lobes are involved in the represen- was performed is not contradicted by the tation of predicted limb positions. discrepancy between the predicted action The phantom limb may also manifest effects and the actual action effects because, itself after deafferentation of a limb that in as I have already noted, even healthy individ- fact is still present. Patients may or may not uals may have a remarkably limited aware- be aware of the existing but deafferented ness of this discrepancy (Fourneret & Jean- limb. If they are, then the phantom limb is nerod, 1998). In addition, when patients experienced as a supernumerary limb. One have suffered a parietal lesion, damage to the or more supernumerary limbs can be experi- parietal cortex is likely to impair awareness enced even if the real limbs are not deaffer- of the state of the motor system and cause a ented. Frith et al. (2000) have proposed that failure to detect the discrepancies between these phenomena are all attributable to the the actual and the predicted action effects. failure to integrate two independent sources Finally, according to Frith et al. (2000; of information concerning the position of also, see Kircher & Leube, 2003) the same the limbs. One derives from the motor com- explanation is applicable to those patients mands, which are issued from the cortical with schizophrenia who describe experi- premotor and motor areas independent of ences of alien control, in which actions (as whether the limb is still present. The other well as thoughts or emotions) are performed source derives from sensory feedback and of by external agent, rather than by their own course is available only if the limb is still will. In healthy individuals, self-monitoring present. systems enable one to distinguish the prod- After right-hemisphere damage, leading ucts of self-generated actions (or thoughts) to paralysis of the left side (usually associated from those of other-generated actions (or with anesthesia of that same side), patients thoughts). It has been postulated that may show anosognosia for hemiplegia. By self-monitoring is normally based on a com- that term it is meant that they are unaware parison between the intention underlying an of the impairment that concerns the motor action and its observed effects (Jeannerod, control of their left limb(s) (see, e.g., Pia, 1999; Jeannerod et al., 2003). The proposal Neppi-Modona, Ricci, & Berti, 2004, for a of Frith et al. is that the experience of alien review). Anosognosia for hemiplegia is often control arises from a lack of awareness of associated with unilateral neglect for the left the predicted limb position (action effect). side of space and the location of the lesion In particular, they suggest “that, in the is in the right parietal lobe. The interest- presence of delusions of control, the patient ing question here is why patients with this is not aware of the predicted consequences condition develop the false belief that there of a movement and is therefore not aware is nothing wrong with the paralyzed limb, of initiating a movement” (2000,p.1784). even to the point of claiming to have moved In conclusion, it would seem that the pro- it to command when in fact no movement posal according to which it is the prediction has occurred. The explanation provided by of the action effects, rather than the actual Frith et al. (2000) once more makes recourse action effects, that reaches consciousness can to the hypothesis that awareness of initiat- explain some odd and apparently unrelated P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

346 the cambridge handbook of consciousness

phenomena. In the case of the phantom action, performance of the intended action, limb, for example, the limb is of course miss- and perception of the effects of the per- ing because of the amputation. However, formed action. In principle, therefore, con- the premotor and motor cortical areas are sciousness may manifest itself at all three intact and can still program a movement of stages. At the first stage, conscious intention the missing limb. The action effects are com- to perform an action may arise. At the sec- puted and reach consciousness, thus produc- ond stage, the intended action may be con- ing the conscious experience of a limb. Sim- sciously performed. At the third stage, the ilarly, anosagnosic patients may be unaware effects of the performed action may be con- of their paralyzed limb because they can still sciously perceived. compute the predicted action effects of that For the first stage, research suggests that limb. Conversely, in the case of the anar- intentions for carrying out voluntary actions chic hand and of schizophrenic patients, the are generated unconsciously and retrospec- action effects would not be computed in tively referred consciously to the action advance, causing the feeling of lack of inten- when it has been executed. That is, the evi- tion, which in turn would induce the patient dence is that consciousness of intention to to disown the action and attribute it to an perform an action is the consequence, rather external agent. than the cause, of activity in the brain. There is a mechanism that binds together in consciousness the intention to act and the conse- Conclusion quences of the intended action, thus producing the illusion of free will. That represents Although being aware of initiating and con- a paradox: An individual may accurately trolling actions is a major component of con- attribute the origin of an action to him- or scious experience, empirical evidence shows herself and yet lack online consciousness of that much of the functioning of our motor the events that have led to that action. system occurs without awareness. That is The interpretation outlined above is the to say, it appears that, in spite of the con- most obvious one of the available evidence. tents of our subjective experience, we have However, there is still the possibility that only limited conscious access to the system there is something wrong with it. In partic- by which we control our actions. This lim- ular, what seems to be missing in that inter- ited access is certainly not concerned with pretation is a clear indication of what dif- the detailed mechanisms of the motor sys- ferentiates willed (voluntary) actions from tem. Even higher-order processes seem to be involuntary actions. Perhaps what is needed denied access to consciousness. Awareness is an intention whose brain correlates begin of initiating an action occurs after the move- to initiate an action before we can sig- ment has begun in the brain area devoted nal exactly when that intention began. The to motor processes. Similarly, awareness of advanced representations of the effects of an choosing one action rather than another action might prove instrumental in contrast- occurs after brain correlates indicate that ing a voluntary action with one that is truly the choice already has been made. There- automatic. fore, it is important to provide an answer Human beings consistently show visual to the question of what are the (few) lev- illusions when they are tested with per- els, in the process of action generation, and ceptual measures, whereas the illusions do execution, that can be accessed consciously not manifest themselves when they are while many more levels occur without tested with motor measures. They can point awareness. accurately to targets that have been illu- With much simplification, it can be said sorily displaced by induced motion. A tar- that voluntary actions, from the simplest to get can be moved substantially during a the most complex ones, involve the follow- saccadic eye movement, and the observer ing three stages: intention to perform an will deny perceiving the displacement even P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

consciousness and control of action 347

while pointing correctly to the new, dis- References placed position. These dissociations recall the situation in blindsight and visual agnosia, Aglioti, S., DeSouza, J. F. X., & Goodale, M. where patients can perform visually guided A. (1995). Size-contrast illusions deceive the tasks without visual experience and without eye but not the hand. Current Biology, 5, 679– awareness of the accuracy of their behav- 685. ior. The explanation is that action execution Bar, M. (2000). Conscious and nonconscious pro- depends on one of two visual systems. There cessing of visual identity. In Y. Rossetti & A. is a sensorimotor or “how” system, which Revonsuo (Eds.), Beyond dissociation: Interac- controls visually guided behavior without tion between dissociated implicit and explicit pro- 153 174 access to consciousness. Its memory is very cessing (pp. – ). Amsterdam: John Ben- brief, only long enough to execute an act, jamins. 2003 but it possesses an egocentrically calibrated Blakemore, S. J.-, & Frith, C. D. ( ). Self- metric visual space that the other system awareness and action. Current Opinions in Neu- robiology, 13, 219–224. lacks. The other is a cognitive or “what” sys- 1995 tem, which gives rise to perception and is Block, N. ( ). On a confusion about a function of consciousness. Behavioral Brain Sciences, 18, used consciously in pattern recognition and 227–287. normal visual experience. The processes that Bridgeman, B. (2002). Attention and visually compose the second stage do not have access guided behavior in distinct systems. In W. Prinz to consciousness either. & B. Hommel (Eds.), Attention and performance In contrast, we certainly are aware of XIX: Common mechanisms in perception and some aspects of the current state of the action (pp. 120–135). Oxford: Oxford Univer- motor system at the third stage in the sity Press. sequence that leads to execution of an Bridgeman, B., Hendry, D., & Stark, L. (1975). action. These, however, do not seem to be Failure to detect displacement of the visual concerned with the perception of the action world during saccadic eye movements. Vision effects. Normally, we are not aware of the Research, 15, 719–722. action effects if they match our prediction of Bridgeman, B., Lewis, S., Heit, F., & Nagle, M. the expected effects; that is, of the effect we (1979). Relation between cognitive and motor- predict the action should produce. Rather, oriented systems of visual perception. Journal we are aware of the stream of motor com- of Experimental Psychology: Human Performance 692 700 mands that have been issued to the system. and Attention, 5, – . Or, to be more precise, when performing Bridgeman, B., Peery, S., & Anand, S. (1997). an action, we are aware of the prediction of Interaction of cognitive and sensorimotor maps its effects, which, of course, depend on the of visual space. Perception and Psychophysics, 59, 456–469. motor commands that were planned in the premotor and motor cortical areas. Buccino, G., Lui, F., Canessa, N., Patteri, I., Lagravinese, G., & Rizzolatti, G. (2004). Neu- The conclusion that we are not aware of ral circuits involved in the recognition of most of our own behavior is disturbing, but actions performed by non-conspeciﬁcs: An the evidence to date clearly indicates that fMRI study. Journal of Cognitive Neuroscience, very few aspects of action generation and 16, 1–14. execution are accessible to consciousness. Castiello, U., Bennett, K., & Stelmach, G. (1993). Reach to grasp: The natural response to perturbation of object size. Experimental Brain Acknowledgments Research, 94, 163–178. Castiello, U., & Jeannerod, M. (1991). Measuring Preparation of this chapter was supported time of awareness. Neuroreport, 2 , 797–800. by grants from MIUR and the University of Castiello, U., Paulignan, Y., & Jeannerod, Padua. The author thanks Morris Moscov- M. (1991). Temporal dissociation of motor itch for very helpful suggestions on a previ- responses and subjective awareness. Brain, 114, ous version of the chapter. 2639–2655. P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

348 the cambridge handbook of consciousness

Chekaluk, E., & Llewelynn, K. (1992). Saccadic Frith, C. D. (2002). Attention to action and suppression: A functional viewpoint. In E. awareness of other minds. Consciousness and Chekaluk & K. Llewelynn (Eds.), Advances Cognition, 11, 481–487. in psychology, 88: The role of eye movements Frith, C. D., Blakemore, S.-J., & Wolpert, D. M. in perceptual processes (pp. 3–36). Amsterdam: (2000). Abnormalities in the awareness and Elsevier. control of action. Philosophical Transactions of Daprati, E., & Gentilucci, M. (1997). Grasping an the Royal Society, London B, 355, 1771–1788. illusion. Neuropsychologia, 35, 1577–1582. Gentilucci, M., Chiefﬁ, S., Daprati, E., Saetti, M. Dehaene, S., & Naccache, L. (2001). Towards a C., & ToniI. (1996). Visual illusions and action. cognitive neuroscience of consciousness: Basic Neuropsychologia, 34, 369–376. evidence and a workspace framework. Cogni- Glover, S. (2002). Visual illusions affect planning tion, 79, 1–37. but not control. Trends in Cognitive Sciences, 6, Della Sala, S., Marchetti, C., & Spinnler, H. 288–292. (1991). Right-sided anarchic (alien) hand: A Glover, S. (2004). Separate visual representations longitudinal study. Neuropsychologia, 29, 1113– in the planning and control of action. Behav- 1127. ioral and Brain Sciences, 27, 3–78. Della Sala, S., Marchetti, C., & Spinnler, H. Goodale, M. A., Milner, D. A., Jakobson, L. S., & (1994). The anarchic hand: A fronto-mesial Carey, D. P. (1991). A neurological dissociation sign. In F. Boller & J. Grafman (Eds.), Hand- between perceiving objects and grasping them. book of neuropsychology (Vol. 9,pp.204–248). Nature, 349, 154–156. Amsterdam: Elsevier. Goodale, M. A., Pelisson,´ D., & Prablanc, C. Dennett, D.C. (1998). The myth of double trans- (1986). Large adjustments in visually guided duction. In S. R. Hameroff, A. W. Kaszniak, reaching do not depend on vision of the hand & A. C. Scott (Eds), Towards a science of con- or perception of target displacement. Nature, sciousness. II. The second Tucson discussions and 320, 748–750. debates (pp. 97–107). Cambridge, MA: MIT Haffenden, A., & Goodale, M. A. (1998). The Press. effect of pictorial illusion on prehension and Desmurget, M., Pelisson, D., Rossetti, Y., & perception. Journal of Cognitive Neurosciences, Prablanc, C. (1998). From eye to hand: 10, 122–136. Planning goal-directed movements. Neuro- Haggard, P., Aschersleben, G., Gehrke, J., & science and Biobehavioral Review, 22, 761– Prinz, W. (2002a). Action, binding, and aware- 788. ness. In W. Prinz & B. Hommel (Eds.), Attention Eagleman, D. M., & Holcombe, A. O. (2002). and performance XIX: Common mechanisms in Causality and perception of time. Trends in perception and action (pp. 266–285). Oxford: Cognitive Sciences, 6, 323–325. Oxford University Press. Farah, M. J. (1994). Visual perception and vis- Haggard, P., Clark, S., & Kalogeras, J. (2002b). ual awareness after brain damage: A tutorial Voluntary actions and conscious awareness. review. In C. Umilta` & M. Moscovitch (Eds.), Nature Neuroscience, 5, 382–385. Attention and performance XV: Conscious and Haggard, P., & Eimer M. (1999). On the rela- nonconscious information processing (pp. 37– tion between brain potentials and awareness 76). Cambridge, MA: MIT Press. of voluntary movements. Experimental Brain Fodor, J. (1983). The modularity of mind. Cam- Research, 126, 128–133. bridge, MA: MIT Press. Haggard, P., & Libet, B. (2001). Conscious inten- Fourneret, P., & Jeannerod, M. (1998). Limited tion and brain activity. Journal of Consciousness conscious monitoring of motor performance in Studies, 11, 47–63. normal subjects. Neuropsychologia, 36. Haggard, P., & Magno, E. (1999). Localising Fried, I., Katz, A., McCarthy, G., Sass, K. J., awareness of action with transcranial magnetic Williamson, P., Spencer, S. S., & Spencer, D. D. stimulation. Experimental Brain Research, 17, (1991). Functional organization of human sup- 102–107. plementary motor cortex studied by electrical Hazeltine, E. (2002). The representational nat- stimulation. Journal of Neuroscience, 11, 3656– ure of sequence learning: evidence for goal- 3666. based codes. In W. Prinz & B. Hommel (Eds.), P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

consciousness and control of action 349

Attention and performance XIX: Common mech- (readiness-potential): The unconscious initia- anisms in perception and action (pp. 673–689). tion of freely voluntary act. Brain, 106, 623– Oxford: Oxford University Press. 642. Hommel, B., Muesseler, J., Aschersleben, G., & Mandler, G. (1975). Mind and emotion. New York: Prinz, W. (2001). The theory of event cod- Wiley. ing (TEC): A framework for perception and Marcel, A. J. (1988). Phenomenal experience action planning. Behavioral and Brain Sciences, and functionalism. In A. J. Marcel & E. Bisi- 24, 849–937. ach (Eds.), Consciousness in contemporary sci- James, W. (1890). Principles of psychology. New ence (pp. 121–158). Oxford: Oxford University York: Holt. Press. Jeannerod, M. (1986). The formation of finger Marcel, A. J. (2004). The sense of agency: Aware- grip during prehension: A cortically mediated ness and ownership of action. In J. Roessler visuomotor pattern. Behavioral Brain Research, & N. Eilan (Eds.), Agency and self-awareness 19, 99–116. (pp. 48–93). Oxford: Clarendon. Jeannerod, M. (1999). To act or not to act: Marchetti, C., & Della Sala, S. (1998). Disen- Perspectives on the representation of actions. tangling the alien and anarchic hand. Cognitive Quarterly Journal of Experimental Psychology, Neuropsychiatry, 3, 191–207. 52A, 1–29. McCloskey, D. I., Colebatch, J. G., Potter, E. K., Jeannerod, M., Farrer, M., Franck, N., Fourneret, & Burke, D. (1983). Judgements about onset of P., Posada, A., Daprati, E., & Georgieff, N. rapid voluntary movements in man. Journal of (2003). Action recognition in normal and Neurophysiology, 49, 851–863. schizophrenic subjects. In T. Kircher & A. Milner, D. A., & Goodale, M. A. (1995). The David (Eds.), The self in neuroscience and psy- visual brain in action (Oxford Psychology Series chiatry (pp. 119–151). Cambridge: Cambridge 27). Oxford: Oxford University Press. University Press. Moscovitch, M., & Umilta,` C. (1990). Modu- Kircher, T. T. J., & Leube, D. T. (2003). Self- larity and neuropsychology: Implications for consciousness, self-agency, and schizophrenia. the organization of attention and memory in Consciousness and Cognition, 12, 656–669. normal and brain-damaged people. In M. F. Kohler, E., Keysers, C., Umilta,` M. A., Fogassi, L., Schwartz (Ed.), Modular deficits in Alzheimer- Gallese, V., & Rizzolatti, G. (2002). Hearing type dementia (pp. 1–59). Cambridge, MA: MIT sounds, understanding actions: Action repre- Press. sentation in mirror neurons. Science, 297, 846– Natsoulas, T. (1992). Is consciousness what psy- 848. chologists actually examine?American Journal Kunde, W. (2004). Response priming by supral- of Psychology, 105, 363–384. iminar and subliminar action effects. Psycho- Nattkemper, D., & Ziessler, M. (Eds.). (2004). logical Research / Psychologische Forschung, 68, Cognitive control of action: The role of action 91–96. effects. Psychological Research / Psychologische Kunde, W., Kiesel, A., & Hoffman, J. (2003). Forschung, 68(2/3). Conscious control over the content of uncon- Neumann, O., & Klotz, W. (1994). Motor scious cognition. Cognition, 88, 223–242. responses to nonreportable, masked stimuli: Lhermitte, F. (1983). ‘Utilization behavior’ and Where is the limit of direct parameter speci- its relation to lesions of the frontal lobes. Brain, fication? In C. Umilta` & M. Moscovitch (Eds.), 106, 237–255. Attention and performance XV: Conscious and Libet, B. (1985). Unconscious cerebral initiative nonconscious information processing (pp. 123– and the role of conscious will in voluntary 150). Cambridge, MA: MIT Press. action. Behavioral and Brain Sciences, 8, 529– Norman, D. A. (1981). Categorisation of action 566. slips. Psychological Review, 88, 1–15. Libet, B. (1999). Do we have free will? Journal of Norman, D. A., & Shallice, T. (1986). Attention Consciousness Studies, 6, 47–57. to action: Willed and automatic control of Libet, B., Gleason, C. A., Wright, E. W., & Pearl, behaviour. In R. J. Davidson, G. E. Schwartz, & D. K. (1983). Time of conscious intention to D. Shapiro (Eds.), Consciousness and self regula- act in relation to onset of cerebral activity tion (Vol. 4,pp.189–234). New York: Plenum. P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

350 the cambridge handbook of consciousness

Paulignan, Y., MacKenzie, C. L., Marteniuk, Ramachandran, V. S., & Hirstein, W. (1998). R. G., & JeannerodM. (1991). Selective pertur- The perception of phantom limbs. Brain, 121, bation of visual input during prehension move- 1603–1630. ments. I. The effect of changing object position. Reason, J. T. (1984). Lapses of attention. In R. Experimental Brain Research, 83, 502–512. Parasuraman, R. Davies, & J. Beatty (Eds.), Pelisson,´ D., Prablanc, C., Goodale, M. A., Varieties of attention (pp. 151–183). Orlando, & Jeannerod, M. (1986). Visual control of FL: Academic Press. reaching movements without vision of the Rizzolatti, G. & Craighero, L. (2004). The limb. II. Evidence of fast unconscious pro- mirror-neuron system. Annual review of cesses correcting the trajectory of the hand neuroscience 27, 169–192. to the final position of double-step stimulus. Roelofs, C. (1935). Optische Localisation. Archiv 303 311 Experimental Brain Research, 62, – . fuer Augenheilkunde, 109, 395–415. 1978 Perenin, M.-T., & Jeannerod, M. ( ). Visual Rossetti, Y., & Pisella, L. (2002). Several ‘vision function within the hemianoptic field follow- for action’ systems: A guide to dissociating ing early cerebral hemidecortication in man. and integrating dorsal and ventral functions 1 1 . Spatial localization. Neuropsychologia, 16, – (Tutorial). Attention and visually guided 13 . behavior in distinct systems. In W. Prinz & Perenin, M.-T., & Rossetti, Y. (1996). Grasping B. Hommel (Eds.), Attention and performance without form discrimination in a hemianopic XIX: Common mechanisms in perception and field. Neuroreport, 7, 793–797. action (pp. 62–119). Oxford: Oxford University Perenin, M.-T., & Vighetto, A. (1988). Optic Press. ataxia. Brain, 111, 643–674. Sanders, M. D., Warrington, E. K., Marshall, J., Pia, L., Neppi-Modona, M., Ricci, R., & Berti, A. & Weiskrantz, L. (1974). ‘Blindsight’: Vision (2004). The anatomy of anosognosia for hemi- in a field defect. Lancet, 20, 707–708. plegia: A meta-analysis. Cortex, 40, 367–377. Schneider, G. E. (1967). Contrasting visuomotor Poeppel, E., Held, R., & Frost, D. (1973). functions of tectum and cortex in the golden Residual visual function after brain wounds hamster. Psychologische Forschung, 31, 52–62. involving the central visual pathways in man. Schneider, G. E. (1969). Two visual systems. Nature, 243, 295–296. Science, 163, 895–902. Posner, M. I., & Klein, R. M. (1973). On the func- Schubotz, R. I., & von Cramon, D. Y. (2002). tion of consciousness. In S. Kornblum (Ed.), Predicting perceptual events activates corre- Attention and Performance IV (pp. 21–35). New sponding motor schemes in lateral premotor York: Academic Press. cortex: An fMRI study. Neuroimage, 15, 787– 796 Price, M.C. (2001). Now you see it, now you . don’t. Preventing consciousness with visual Shallice, T. (1988). From neuropsychology to masking. In P.G. Grossenbacher (Ed.), Finding mental structure. Cambridge: Cambridge consciousness in the brain: A neurocognitive University Press. approach (Advances in consciousness research Shallice, T. (1994). Multiple levels of control (Vol. 8,pp.25–60). Amsterdam: John Ben- processes. In C. Umilta` & M. Moscovitch jamins. (Eds.), Attention and performance XV: Con- Prinz, W. (1997). Perception and action planning. scious and nonconscious information processing European Journal of Cognitive Psychology, 9, (pp. 395–420). Cambridge, MA: MIT Press. 129–154. Stoerig, P., & Cowey, A. (1997). Blindsight in Prinz, W., & Hommel, B. (Eds.). (2002) Attention man and monkey. Brain, 120, 535–559. and Performance XIX: Common mechanisms Stoet, G., & Hommel, B. (2002). Interaction in perception and action. Oxford: Oxford between feature binding in perception and University Press. action. In W. Prinz & B. Hommel (Eds.), Prochazka, A., Clarac, F., Loeb, G. E., Rothwell, Attention and performance XIX: Common mech- J. C., & Wolpaw, J. R. (2000). What do reflex anisms in perception and action (pp. 538–552). and voluntary mean? Modern views on an Oxford: Oxford University Press. ancient debate. Experimental Brain Research, Taylor, T. L., & McCloskey, D. (1990). Triggering 130, 417–432. of preprogrammed movements as reactions P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

consciousness and control of action 351

to masked stimuli. Journal of Neurophysiology, Wegner, D. M. (2003). The mind’s best trick: 63, 439–446. How we experience conscious will. Trends in Trevarthen, C. B. (1968). Two mechanisms of Cognitive Sciences, 7, 65–69. vision in primates. Psychologische Forschung, Wegner, D. M. (2005). Precis´ of the illusion of 31, 299–337. conscious will. Behavioral and Brain Sciences, Umilta,` C. (1988). The control operations of con- 27, 649–659. sciousness. In A. J. Marcel & E. Bisiach (Eds.), Weiskrantz, L. (1986). Blindsight. A case study and Consciousness in contemporary science (pp. 334– implications. Oxford: Oxford University Press. 356). Oxford: Oxford University Press. Weiskrantz, L. (1997). Consciousness lost and Umilta,` C., & Moscovitch, M. (1994) (Eds.). Att- found. A neuropsychological exploration. Ox- ention and performance XV: Conscious and non- ford: Oxford University Press. conscious information processing. Cambridge, Zhu, J. (2004). Locating volition. Consciousness MA: MIT Press. and Cognition, 13, 1–21. Umilta,` M. A., Kohler, E., Gallese, V., Fogassi, Ziesser, M., & Nattkemper, D. (2002). Effect L., Fadiga, L., & Rizzolatti, G. (2001). “I know anticipation in action planning. In W. Prinz what you are doing”: A neurophysiological & B. Hommel (Eds.), Attention and perfor- study. Neuron, 32, 91–101. mance XIX: Common mechanisms in perception Ungerleider, L. G., & Mishkin, M. (1982). Two and action (pp. 645–672). Oxford: Oxford cortical visual systems. In D. Ingle, M. A. University Press. Goodale, & R. J. W. Mansﬁeld (Eds.), Analysis Zihl, J. (1980). Blindsight: Improvement of of visual behavior (pp. 549–586). Cambridge, visually guided eye movements by systematic MA: MIT Press. practice in patients with cerebral blindness. Vorberg, D., Mattler, U., Heinecke, A., Schmidt, Neuropsychologia, 18, 71–77. T., & Schwarzbach, J. (2003). Different time Zihl, J., & von Cramon D. (1985). Visual ﬁeld courses for visual perception and action recovery from scotoma in patients with priming. Proceedings of the National Academy postgeniculate damage: A review of 55 cases. of Science USA, 100, 6275–6280. Brain, 108, 335–365. P1: KAE 0521857430c12 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw January 16, 2007 19:25

352 P1: KAE 0521857430c13 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:57

D. Linguistic Considerations

353 P1: KAE 0521857430c13 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:57

354 P1: KAE 0521857430c13 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:57

CHAPTER 13 Language and Consciousness

Wallace Chafe

Abstract imagistic and the ideational components of consciousness are held to be central to This chapter focuses on two distinct, lin- thought as thought is usually understood. guistically oriented approaches to language Both approaches are supported with linguis- and consciousness taken by Ray Jackendoff tic data, but data that are different in kind. and Wallace Chafe. Jackendoff limits consciousness to uninterpreted imagery, and he presents evidence that such imagery is exter- Introduction nal to thought because it is too particular, does not allow the identiﬁcation of individ- How are language and consciousness rela- uals, and fails to support reasoning. If all we ted? Can the study of language shed light are conscious of is imagery and imagery does on the nature of consciousness? Can an not belong to thought, it follows that we improved understanding of consciousness are not conscious of thought. Chafe distin- contribute to an understanding of language? guishes between immediate consciousness, Some scholars in the past have gone so far involved in direct perception, and displaced as to equate conscious experience with lin- consciousness, involved in experiences that guistic expression. Within philosophy, as one are recalled or imagined. He sees the former example, we ﬁnd Bertrand Russell (1921, as including not only the sensory experiences p. 31) writing, “A desire is ‘conscious’ when discussed by Jackendoff but also their inter- we have told ourselves that we have it.” In pretation in terms of ideas. Displaced con- psychology we have Jean Piaget’s statement sciousness includes sensory imagery that is (1964/1967,p.19) that “thought becomes qualitatively different from immediate sen- conscious to the degree to which the child is sory experience, but it too is accompanied able to communicate it.” Given our current by ideational interpretations that resemble knowledge, however, we might want to go those of immediate experience. Both the beyond simply equating consciousness with

355 P1: KAE 0521857430c13 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:57

356 the cambridge handbook of consciousness

language. This chapter focuses on some pos- includes various authors who show that at sible insights originating in linguistics, the least some aspects of language structure lie discipline for which language is the primary outside of consciousness. Langacker (1997, focus, including the relation of language to p. 73), for example, concludes his chapter by broader aspects of humanness. It may be saying that “it should be evident that gram- that a language-centered approach can pro- mar is shaped as much by what we are not vide some answers that have eluded those in consciously aware of as by what we are.” other disciplines who have approached these Although that may be true, the discussion to questions from their own perspectives. follow explores the question of whether and Consciousness-related questions were to what extent consciousness and the struc- rarely asked within linguistics in the 20th ture of language are mutually interactive. century, a period during which most scholars This exploration can hardly proceed within that field would have found them irrel- out taking into account ways in which both evant or pointless. Stamenov (1997,p.6) language and consciousness are related to mentions correctly that “there is at present other mental phenomena, and especially to little research in linguistics, psycholinguis- those captured by such words as thought and tics, neurolinguistics and the adjacent disci- imagery. These are, of course, words in the plines which explicitly addresses the prob- English language, and they can mean dif- lem of the relationships between language ferent things to different people. Natsoulas and consciousness.” This neglect can be (1987) discusses six different ways of under- traced historically to the association of con- standing consciousness, and exactly what is sciousness with “mentalism,” a supposedly meant by thought and imagery is hardly unscientific approach to language that was uncontroversial. The underlying problem, of forcefully rejected by Leonard Bloomfield course, is that consciousness, thought, and and his followers, who dominated linguistics imagery all refer to private phenomena that during the second quarter of the 20th cen- are inaccessible to direct public observation. tury. Bloomfield often repeated statements Although their referents may seem obvious such as the following, taken from his obitu- to introspection and may be pervasive ingre- ary of Albert Paul Weiss, a behaviorist psy- dients of people’s mental lives, it can be frus- chologist who strongly influenced him: “Our tratingly difficult to achieve agreement on animistic terms (mind, consciousness, sensa- what they include or where their bounda- tion, perception, emotion, volition, and so ries lie. on) were produced, among the folk or by It may seem presumptuous to suggest philosophers, in pre-scientific times, when new solutions to problems that have occu- language was taken for granted” (Bloom- pied scholars for millennia, but continu- field, 1931,p.219). Bloomfield’s enormous ing advances in both scholarship and sup- influence led his followers to regard such porting technology have encouraged new notions as pointless appeals to something approaches to old puzzles in many areas, akin to magic. Noam Chomsky, whose influ- including the study of language and its ence eclipsed that of Bloomfield in the lat- relation to the mind. This chapter focuses ter half of the 20th century, departed from on two partially different answers to the Bloomfield by expressing an interest in the questions raised at the beginning of this mind, but his interests were restricted to introduction. One has been proposed by an abstract variety of syntax that in basic Ray Jackendoff, the other by Wallace Chafe. respects remained loyal to Bloomfield’s con- Both agree that people are conscious of straints and that, by its very nature, pre- imagery and affect. Beyond that, how- cluded a significant role for consciousness. ever, their conclusions differ. Jackendoff lim- Language Structure, Discourse and the its consciousness to imagery and excludes Access to Consciousness, a book character- imagery from thought, above all because ized by its editor, M. I. Stamenov, as “the it is unable to account for inferences and first one dedicated to a discussion of some other logical processes. It follows that peo- of the aspects of this topic” (1997,p.6), ple are not conscious of thought. Chafe P1: KAE 0521857430c13 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:57

language and consciousness 357

suggests that imagery is inseparable from its conscious of what they are. We may hear the interpretation in terms of ideas and orienta- sounds of language overtly or in our heads, tions of ideas and that thought exhibits both but the thoughts expressed by those sounds imagistic and ideational components that are lie outside of consciousness. “We become simultaneously available to consciousness. aware of thought taking place – we catch Both scholars support their suggestions with ourselves in the act of thinking – only when evidence from language, but the evidence is it manifests itself in linguistic form, in fact of different kinds. phonetic form” (1997,p.188). Nonverbal imagery provides other conscious manifestations of thought, but it too is distinct from Is Thought Unconscious? thought itself. Jackendoff discusses several kinds of evi- We can consider first the view that peo- dence that thought is unconscious. For one ple are conscious only of imagery and that thing, our minds sometimes seem to solve imagery is not an element of thought, a view problems without our being conscious of that was forcefully presented by Jackendoff them doing it: “We have to assume that in his book, Consciousness and the Computa- the brain is going about its business of tional Mind (1987) and subsequently refined solving problems, but not making a lot of in Chapter 8 of The Architecture of the Lan- conscious noise about it; reasoning is tak- guage Faculty (1997), which was in turn a ing place without being expressed as lan- somewhat modified version of Jackendoff guage” (1997,p.188). He emphasizes the (1996). He identifies three basic levels of role of reasoning as a fundamental aspect of information processing. At the outermost thought, suggesting that if he were to say level, closest to “the outside world,” are brain to you Bill killed Harry, you would know processes of which we are totally uncon- from that statement that Harry died. You scious; among these processes are such visual might at first think you knew it because phenomena as fusing the retinal images from you had an image of Bill stabbing Harry the two eyes into a perception of depth or and Harry falling dead, but this image is stabilizing the visual field when the eyes are too specific because “the thoughts expressed moving. We are conscious of the results of by the words kill and die, not to mention these processes – depth perception and a sta- the connections between them, are too gen- bilized visual field – but not of how they hap- eral, too abstract to be conveyed by a visual pen. When it comes to auditory phenomena, image.” Thus, the knowledge that killing we may be conscious of speech sounds, such entails dying must belong to unconscious as vowels, consonants, syllables, and pitch thought. He also mentions the problem of contours, but not of how raw sound comes to identification. “How do you know that those be interpreted in those ways. Mechanisms at people in your visual image are Bill and the outer layer of perception do their work Harry respectively?” Beyond that, there are outside of consciousness. many concepts like virtue or social justice or What we are conscious of are the visual seven hundred thirty-two for which no useful forms, speech sounds, and so on that belong images may be available. Such words express to what Jackendoff calls the intermediate elements of thought of which there is no way level of information processing. It is inter- to be directly conscious. mediate between unconscious perceptual Jackendoff takes pains to separate processes and a deeper level inhabited by thought from language. A major ingredient thought, of which we are also not con- of consciousness is inner speech, and we may scious. In short, we are conscious of imagery for that reason be tempted to equate inner associated with the various sense modali- speech with thought. But, for one thing, ties, and we are also conscious of affect, “thinking is largely independent of what but that is all. We certainly have thoughts, language one happens to think in” (1997, but it is only through their manifestations in p. 183). Whether one is speaking English imagery, nonverbal or verbal, that we can be or French or Turkish, one can, he says, be P1: KAE 0521857430c13 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:57

358 the cambridge handbook of consciousness

having essentially the same thoughts. The attention as a process for zeroing in on what form of a particular language is irrelevant consciousness makes available: to the thoughts it conveys, as shown by I am claiming that consciousness happens the fact that an English speaker can be to provide the basis for attention to pick having the same thoughts as a Japanese out what might be interesting and thereby speaker, even though the English speaker put high-power processing to work on it. puts the direct object after the verb and the In turn, the high-power processing resulting Japanese speaker before it. Furthermore, from attention is what does the intelligent one can be conscious of linguistic form work; and at the same time, as a byproduct, that is dissociated from any thought at it enhances the resolution and vividness all, as with the rote learning of a ritual in of the attended part of the conscious field an unfamiliar language. Conversely, in the (p. 200). tip-of-the-tongue phenomenon “one feels Furthermore, “without the phonetic form as as though one is desperately trying to fill a conscious manifestation of the thought, a void . . . One is aware of having requisite attention could not be applied, since atten- conceptual structure” but the phonological tion requires some conscious manifestation structure is missing (1987,p.291). He cites as a ‘handle’” (p. 201). Language is particu- various other examples to show that thought larly useful because it “is the only modality is unconscious and that all we are conscious of consciousness in which the abstract and of is the phonetic imagery that expresses relational elements of thought correspond thought, often accompanied by nonverbal even remotely to separable units” (p. 202). imagery. “The picture that emerges from Units of that kind are not available to other these examples is that although language components of consciousness, such as visual expresses thought, thought itself is a separate imagery. brain phenomenon” (1997,p.185). Third, language gives access to what Jack- endoff calls valuations of percepts. Valua- The Relation of Language to Thought tions include judgments that something is novel or familiar, real or imagined, volun- In spite of this disconnect between con- tary or involuntary, and the like. Language scious phonetic imagery and unconscious provides “words like familiar, novel, real, thought, Jackendoff discusses three ways in imaginary, self-controlled, hallucination that which “language helps us think.” First of all, express valuations and therefore give us a language makes it possible to communicate conscious link to them. This conscious link thoughts: permits us to attend to valuations and sub- Without language, one may have abstract ject them to scrutiny” (p. 204). On awaken- thoughts, but one has no way to commu- ing from a dream, for example, we can say, nicate them (beyond a few stereotyped ges- “It was just a dream,” but a dog cannot do tures such as head shaking for affirmation that – it cannot bring the valuation into con- and the like). . . . Language permits us to sciousness in that way. have history and law and science and gossip....Asa result of our having language, vastly more of our knowledge is collective Summary and cumulative than that of nonlinguistic Jackendoff finds thought to be totally uncon- organisms . . . Good ideas can be passed on scious, but he suggests that it is mani- much more efficiently (1997,p.194). fested in consciousness by way of language, So, even though thought itself is uncon- which itself enters consciousness only by scious, language provides an important way way of phonetic imagery. Nevertheless, lan- of sharing thoughts. guage enhances the power of thought in Second, “having linguistic expressions in three ways: by allowing thought to be com- consciousness allows us to pay attention municated, by making it possible to focus to them” (1997,p.200). Jackendoff views attention on selected aspects of thought P1: KAE 0521857430c13 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:57

language and consciousness 359

(particularly on its relational and abstract in ways that have not been sufficiently stud- elements), and by providing access to val- ied (but see Chafe, 2002; Wierzbicka, 1999), uations of thought. and it is mentioned only in passing below. Jackendoff has presented a serious and It should also be apparent that conscious- responsible challenge to those who would ness may be focused on the environment like to believe that we are conscious of immediately surrounding the conscious self, thought. If his suggestion that we are not in which case we can speak of an immedi- conscious of thought seems on the face of ate consciousness. However, it may also be it to conflict with ordinary experience, it focused on experiences remembered or ima- is a conclusion that follows inevitably from gined or learned from others, in which case these two propositions: we can characterize it as a displaced consciousness. Immediate consciousness invol- 1 ( ) Consciousness is limited to phonetic and ves direct perception, whereas displaced nonverbal imagery. consciousness takes the form of imagery – an (2) Thought is independent of those two attenuated experiencing of visual, auditory, kinds of imagery. and/or other sensory information. Immedi- ate and displaced consciousness are qualita- Behind these propositions lie particular ways tively different (Chafe, 1994). of understanding consciousness, thought, Consciousness is dynamic, constantly and imagery, and the subjective nature of changing through time, and it resembles these phenomena leaves room for other vision in possessing both a focus and a interpretations. The remainder of this chap- periphery – a fully active and a semiactive ter explores other possibilities, based in large range – analogous to foveal and peripheral part on Chafe (1994, 1996a,b), with occa- vision. Its focus is severely limited in con- sional references to William James (1890), tent and duration, typically being replaced whose views on “the stream of thought” are every second or two, whereas its periphery still worth taking seriously (Chafe, 2000). changes at longer, more variable intervals. In that respect consciousness shares the properties of eye movements as they are monitored, Preliminaries to an Alternative View for example, while people look at pictures (Buswell, 1935) and also while they talk Because consciousness, thought, and imagery about them (Holsˇanov´ a,´ 2001). Language do refer to subjective experiences, further reflects the foci of active consciousness in discussion can benefit from specifying how prosodic phrases, whereas larger coherences these words are used here. Consciousness is of semiactive consciousness appear in hier- notoriously difficult to define in an objective archically organized topics and subtopics way, but some of its properties will emerge (Chafe, 1994). This view sees the process as we proceed. termed attention as simply the process by which consciousness is deployed. Some Properties of Consciousness The Nature of Imagery There is agreement by Jackendoff, Chafe, and others that conscious experiences have It is well known that direct perception of two obvious components that may be the environment is not a matter of simply present simultaneously. One of them is registering what is “out there.” Information related to sensory experience, and the other that enters the eyes and ears is always inter- is affect. A case could be made for the preted in ways that are shaped in part by affective component being in some ways the genetic endowments, in part by cultural and more basic. It is certainly the oldest in terms individual histories. Imagery is no different. of evolution, and it still underlies much of It is not an attenuated replaying or imagin- human behavior. It is reflected in language ing of raw visual, auditory, or other sensory P1: KAE 0521857430c13 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:57

360 the cambridge handbook of consciousness

input. Sensory experiences do not appear ity that his friend was not restricted to the in that form even in immediate conscious- sounds of language, but had conscious access ness, and images are subject to considerably to ideational interpretations of those sounds more interpretive processing. Experiencing devoid of sensory accompaniments. That is imagery is not like looking at a picture. But the more interesting interpretation because even pictures are interpreted as containing it suggests that one can indeed experience particular people, structures, trees, or what- ideational consciousness while at the same ever, which are oriented and related in par- time imagery is restricted (in some people) ticular ways. to the sounds of language. In other words, Although images always involve inter- an absence of non-verbal imagery need not pretations, there may be some circum- deprive individuals like James’s friend of stances under which it is possible to be a consciousness of ideas. This conscious- conscious of interpretive phenomena in the ness of ideas, with or without accompanying absence of accompanying sensory experi- imagery, can plausibly be considered a major ence. It is thus useful to distinguish these two component of thought. aspects of consciousness. The word imagery is restricted here to sense-related phenom- Components of Language ena, real or imagined, whereas the interpretive experiences that are often (but need Figure 13.1 lays out some basic stages that not be) accompanied by imagery are termed intervene between a person’s interaction ideational. with the outside world and the utterance of The imagistic component of conscious- linguistic sounds. These phenomena interact ness may or may not include the sounds of in ways that are obscured by their assign- language. On that basis one can distinguish ment to separate boxes. In the ﬁnal analysis verbal from non-verbal consciousness, the they are realized in structures and processes latter being focused on non-verbal aspects of within the brain, distributed in ways that are experience. One may experience non-verbal surely not separated so neatly. Nevertheless, imagery without language, one may experi- there are certain distinguishable principles ence language as well as non-verbal imagery, of organization that do lend themselves to or one may experience language alone, either discussion in these terms. overtly or covertly. There are evidently indi- The boxes on the far left and right, vidual differences in this regard (Poltrock & labeled reality and sound, represent phenom- Brown, 1984). William James, having asked ena external to the human mind. We need people about the images they had of their not concern ourselves with problems that breakfast table, reported that might be associated with the word reality; here it is shorthand for whatever people an exceptionally intelligent friend informs think and talk about. It includes events and me that he can frame no image whatever states and people and objects encountered of the appearance of his breakfast-table. in the course of living that may in some way When asked how he then remembers it at affect a person’s thoughts, but that would all, he says he simply ‘knows’ that it seated exist whether they were processed by a mind four people, and was covered with a white or not. Sound on the right represents exter- cloth on which were a butter-dish, a coffee- pot, radishes, and so forth. The mind-stuff nal, physical manifestations of the sounds of which this ‘knowing’ is made seems to of language: their articulation by the vocal be verbal images exclusively (James, 1890, organs, their acoustic properties outside the p. 265). human body, and their reception in the ears of a listener. There are two possible interpretations of The boxes labeled thought on the left and the phrase verbal images in this quote. James phonology on the right represent immediate probably concluded that his friend expe- results of the interpretive processes that the rienced the sounds of language and noth- mind applies to those external phenomena. ing more. More interesting is the possibil- Each has its own patterns of organization. P1: KAE 0521857430c13 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:57

language and consciousness 361

mental processing

reality → thought → semantics → syntax → phonology → sound

language Figure 13.1. Stages in the production of language.

Thought need not be as directly related sciousness of what he calls either phonologi- to external reality as Figure 13.1 suggests, cal structure or phonetic form, pointing out because a great deal of it involves remember- that “we experience the inner speech stream ing one’s own earlier experiences or expe- as segmented into words and possibly fur- riences learned from others, or imagining ther into syllables or individual segments. things for which no direct contact with real- In addition, the rhythm, stress pattern, and ity ever existed. Linguists have devoted a intonation of inner speech must come from great deal of time and effort to understand- phonological units as well” (1987,p.288). ing phonological organization, but have con- In educational psychology there have been cerned themselves much less with the orga- numerous studies of children’s awareness of nization of thought, for reasons that have segments (“phonemes”) and syllables and the more to do with the history of the discipline relevance of such awareness to the acquisi- than with the ultimate relevance of thought tion of reading (e.g., Anthony & Lonigan, to language. 2004; Castles & Coltheart, 2004). It is clear The two remaining boxes, labeled seman- that we are conscious of interpreted sounds, tics and syntax, represent ways in which ele- not physical sounds alone. Sounds are orga- ments of thought are adjusted to fit the nized by the mind into linguistically rele- requirements of language so that they can be vant elements of which we can be conscious, associated with sounds, manipulated sym- although their physical manifestations can bolically, and communicated to other minds. have a place in consciousness too. We orga- If languages never changed, semantic struc- nize sound into syllables, words, intonational tures could be submitted directly to phonol- patterns, and the rest, in association with ogy. But languages do change, and seman- their sensory manifestations. The point is tic structures are reshaped through processes that phonological consciousness has both of lexicalization and grammaticalization to imagistic and interpretive components. Can form syntactic structures that are no longer the same be said of thought, our focus in this related to thoughts fully or directly. It is chapter? these syntactic structures that then proceed to be associated with sounds (Chafe, 2005). Missing from Figure 13.1 is consciousness, Consciousness of Thought and its place in this picture is our chief concern. It is helpful at this point to consider an example of actual speech. The three lines below were taken from a conversation among Consciousness of Phonology three women, in the course of which they discussed some local forest fires. As they It can be instructive to consider briefly what developed this topic, one of them said, is involved in consciousness of the sounds of language, because what happens on the (a) You’d look at the sun, right side of Figure 13.1 may be simpler (b) it just looked red; and easier to comprehend than what hap- (c) I mean you couldn’t see the sun. pens on the left. The important lesson is that the mind does not operate exclusively She was talking about something she had with raw sound. Jackendoff describes con- experienced 2 days earlier, when she had P1: KAE 0521857430c13 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:57

362 the cambridge handbook of consciousness

direct access to the unusual appearance of so common as to seem trivial, but they shed the sun. At that time her perception of it light on the basic nature of consciousness, was in her immediate consciousness. It must thought, imagery, and communication by have been primarily a visual experience, but providing evidence for properties of imme- it was probably accompanied by a feeling of diate and displaced consciousness, for the awe or wonder at this rather extreme depar- mutually supportive roles of ideational and ture from normality. It was not just a regis- imagistic consciousness, and for the relation tering of sensory information, but an inter- of consciousness to thought. pretation that included a selection of certain salient ideas from those that must have been The Organization of Thought potentially available. Among them was the idea of an event she verbalized in line (a) as The dynamic quality of consciousness is evi- look and an object she verbalized as the sun. dent in the constantly changing content of The visual input to her eyes had been inter- its focus. The prosody of language divides preted by her mind, and she was conscious of speech into a sequence of prosodic phrases, both the interpretation and its sensory cor- each expressing a focus of the speaker’s con- relates. If the word thought means anything sciousness. Each focus is replaced by another at all, this immediate experience, with its at intervals in a typical range of 1 or 2 s ideational and imagistic components, was at (cf. Poppel,¨ 1994). This progression through that moment present in her thought. It was time was described by James in his often- an immediate perceptual experience, it had quoted statement regarding the stream of both ideational and imagistic components, thought, where, it may be noted, he equated she was conscious of it, and she was think- thought and consciousness: ing about it. As we take, in fact, a general view of When, 2 days later, she uttered the lan- the wonderful stream of our consciousness, guage above, the earlier experience was once what strikes us first is this different pace more active in her consciousness. This time, of its parts. Like a bird’s life, it seems to be however, it was no longer immediate but dis- made of an alternation of flights and perch- placed. It was probably experienced in part ings. The rhythm of language expresses this, as imagery, but her language expressed an where every thought is expressed in a sen- interpretation in terms of ideas. By utter- tence, and every sentence closed by a period. ing this language she must have intended The resting-places are usually occupied by to activate a partially similar experience in sensorial imaginations of some sort, whose the consciousness of her listeners. They, for peculiarity is that they can be held before their part, must have experienced their own the mind for an indefinite time, and contemplated without changing; the places of displaced consciousness, but for them it was flight are filled with thoughts of relations, twice displaced. We can note a difference static or dynamic, that for the most part between the relative success of communi- obtain between the matters contemplated cating the ideational elements and commu- in the periods of comparative rest (James, nicating their sensory accompaniments. The 1890,p.243). speaker’s ideas as such could be communicated more or less intact, but their sensory The example cited above shows three foci accompaniments necessarily differed for her of consciousness expressed by the prosodic listeners from what she was experiencing. phrases transcribed in the three lines: The ideas expressed as the sun and looked red (a) You’d look at the sun, could survive the communication process, (b) it just looked red; but their sensory manifestations could only (c) I mean you couldn’t see the sun. be reshaped by the listeners’ imaginations. Experiential and communicative events Within these foci, language shows the mind of this kind are ubiquitous in everyday organizing experience into elements that speech, and indeed in everyday life. They are can be called ideas, among which are ideas P1: KAE 0521857430c13 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:57

language and consciousness 363

of events and states, typically expressed as In line (a), you’d look at the sun,itwas clauses. This woman expressed first the idea expressed with the words the sun, which of a remembered event, you’d look at the were given prosodic prominence. They were sun, including the idea of the looking event spoken slowly, relatively loudly, and with a itself along with ideas of its participants – high falling pitch. In line (b), it just looked unspecified individuals expressed by the red, the same idea was expressed with the word you and the object expressed as the sun. pronoun it, which was spoken rapidly with She then proceeded to verbalize the idea of a low volume and low pitch. The prosodi- state with it just looked red, including the idea cally most prominent word in (b) was red. of looking red and once more the sun, this In line (c), I mean you couldn’t see the sun, time expressed as it. Finally, in I mean you the same idea was expressed once more as couldn’t see the sun she rephrased her expe- the sun, but this time these words were spo- rience. Language provides copious evidence ken softly with the pitch deteriorating into that the processing of experience into ideas creaky voice. The most prominent words in of events and states and their participants is (c) were couldn’t see. fundamental to the mind’s organization of The function of pronouns like the it in thought. Outside linguistics the segmenta- line (b) is well explained in terms of con- tion of experience into ideas of events has sciousness. Pronouns are used when an idea been investigated experimentally by Newt- (like the idea of the sun) is in a speaker’s son, Engquist, and Boyd (1977) and more fully active consciousness and is assumed to recently with brain imaging (Zacks et al., be in the listeners’ fully active consciousness 2001). as well. This idea was activated for the listen- The thoughts of this woman may have ers during the utterance of line (a). Because focused on these ideas, but there was clearly in (b) it was assumed to be already active an affective accompaniment that was in in their consciousness, it could be expressed fact foreshadowed by an earlier statement with a minimum of phonological material. describing the most weird day I’ve ever seen There was no need to assign it the promi- in my entire adult life. In the above, affect nence it had in (a). English and many other is most noticeable in line (b) in a rise-fall languages typically express such already con- pitch contour on the word red, expressive of scious ideas with unaccented pronouns. In a feeling engendered by the striking quality Asian languages there is a tendency to give of the sun’s appearance, and in line (c) in a such ideas no phonological representation rise-fall contour on see, conveying an emo- whatsoever. tionally laden contrast with what one would In line (c) there was a reversion to the normally see. words the sun, but this time they were Language, in summary, shows thought spoken without the prominence they had being organized into successive foci of con- in (a). Because the idea of the sun was pre- sciousness, each typically activating ideas sumably still active in the consciousness of of events and states and their participants, the speaker and listeners, why was it not often with accompanying affect. It is at least again expressed with it? Line (c) is discussed plausible to suppose that people are con- further below, but here it can be noted that scious of these ideational elements together it involved a second attempt at verbalizing with their imagistic correlates and that both the speaker’s consciousness, a kind of start- the ideas and the images contribute to the ing over. If she had said I mean you couldn’t flow of thought. see it, she would have tied this act of verbalization too closely to what preceded. With its lack of prominence the idea of the sun Pronouns was presented as still in active consciousness, The idea of the sun was verbalized in three but with the words the sun, rather than it,it different ways in this example, although was presented as part of a new attempt at only two are visible in the transcription. verbalizing this experience. P1: KAE 0521857430c13 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:57

364 the cambridge handbook of consciousness

of the event could be appropriately verbal- Harry’s Death Revisited ized as an instance of the kill category. He We can now return to reasons cited by Jack- might have decided to categorize it differ- endoff for excluding imagery and conscious- ently – as an instance of the murder, stab, or ness from thought. As noted, he imagined poison categories, for example – each with a situation in which he said to someone, its own entailments. Choosing the kill cat- Bill killed Harry. (Linguists often use violent egory accomplished two things: It allowed examples like this because they highlight in the speaker to use the word kill, but at the starkest terms the properties of transi- the same time it associated the idea of the tive verbs.) It is an example that conﬂicts event with expectations that would apply to with observations of actual speech, where instances of the category, and among them it is only under special circumstances that was the expectation that the victim would nouns like Bill and Harry are spelled out be dead. Categories give speakers words for in this way. In a more realistic context one their ideas, but they also carry expecta- might have said, for example, he killed him, tions that are thereby associated with those with pronouns conveying an assumption ideas. Categorization is an essential step that the ideas of Bill and Harry were already in the conversion of thoughts into sounds in the consciousness of the speaker and because of the obvious impossibility of asso- listener. ciating every unique thought with a unique But let us suppose that this sentence sound, but categories are also the locus of was actually uttered and that the speaker entailments. and listener both experienced their own Jackendoff asks further how a particular images of the event in question. The ques- image could allow one to know the identi- tion is whether those images belonged to ties of Bill and Harry. That question, how- the speaker’s and listener’s thoughts. Jack- ever, seems to be based on the view that an endoff says they did not, because an image image resembles an uninterpreted picture. would have to be particular. It would have to If images are always accompanied by inter- “depict Bill stabbing or strangling or shoot- pretations, the identities of Bill and Harry ing or poisoning or hanging or electrocut- would necessarily accompany the sense- ing Harry. . . . And Harry had to fall down, related experience. or expire sitting down, or die hanging from a rope” (1997,p.188). How could any of those particular images be the general con- The Priority of Thought cept of killing? But of course the thought over Phonology expressed by this sentence was not a thought of killing in general; it was a thought of a par- Let us suppose that one can be conscious ticular event for which a particular image of both thoughts and sounds. Language was entirely appropriate. The thought and provides evidence of several kinds that the image were thus not at odds. consciousness of thoughts has priority over Jackendoff’s question, however, goes consciousness of sounds in ordinary mental beyond the relation of a particular image to life. a particular thought to ask how a listener would know Harry was dead, knowledge Familiar and Unfamiliar Languages that would depend on a general knowledge of killing and its results. It is important at It is instructive to compare the experience this point to distinguish between the idea of listening to one’s own language with that of an event, an element of thought, and of listening to an unfamiliar language. In the the way an idea is categorized for expres- latter case it is only the sounds of which one sion in language. That distinction is captured can be conscious. They may be recognized in Figure 13.1 by the placement of thought as the sounds of a language, but that is all. and semantics in separate boxes. Whoever Listening to one’s own language is a very said Bill killed Harry decided that the idea different experience. Normally one is hardly P1: KAE 0521857430c13 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:57

language and consciousness 365

conscious of the sounds at all, but rather of gorically. Availability to consciousness can the thoughts the language conveys, thoughts be a graded matter, as people are conscious that may or may not appear in consciousness more easily, immediately, or readily of some as imagery but that always have an ideational things than others, but are able under the component. right circumstances to shift consciousness to Spoken language provides the best exam- less accustomed phenomena. ples of this distinction, but it can be mimicked with written symbols. To any- Rote Learning one who is not familiar with the Japanese As mentioned by Jackendoff (1996,pp.6– language and writing system the symbols 7), one may have learned “by heart” the below may at best be recognized as exam- sounds of a poem, ritual, or song with little ples of such writing, but otherwise they or no consciousness of thoughts associated are incomprehensible marks. To a literate with those sounds. In such rote learning con- Japanese they elicit consciousness of the sciousness of thought is excluded, and con- sounds ame ga furu, but consciousness is sciousness of sound is all that is available. primarily focused on a thought, a thought This experience has something in common that might be expressed in English as it’s with listening to an unfamiliar language, but raining. in this case one is, oneself, actually speaking. A basic element of linguistic processing is missing, however, and one recognizes that it is unusual to produce sounds with no Linguists who conduct fieldwork with accompanying thoughts. The fact that we little-studied languages sometimes ask a con- have such experiences is indirect evidence sultant to repeat a certain word so they can that we are, under normal circumstances, focus on its phonetic properties. It is not conscious of thoughts and know when they unusual for the consultant to fail to coop- are absent. erate, preferring to discuss at length what the word means. Uppermost in the consul- Ambiguity tant’s consciousness is the thought behind the word, and questions regarding its sound One other linguistic phenomenon that sug- are regarded as irrelevant and intrusive. This gests a consciousness of thoughts is ambi- same priority in consciousness of thoughts guity: the fact that certain words can be over sounds may explain why it is difficult phonologically identical but associated with to teach writing to speakers of previously different thoughts. Chafe (1996b, p. 184) unwritten languages. To write accurately it mentions the word discipline, which may is necessary to shift one’s consciousness from categorize an idea similar to that catego- thoughts to sounds, and that is not always an rized as academic field or may alternatively easy thing to do. involve harsh training to obey a set of rules. The same preference for thoughts can Because the sound and spelling are the same, be instructive in another way. Although there must be consciousness of a difference thoughts may be uppermost in conscious- in thought. Beyond that, one can con- ness, they can with prodding and effort be sciously compare one of the ideas expressed replaced by consciousness of sounds. Con- as discipline with the idea expressed as aca- sciousness thus has priorities that are pre- demic field and judge the closeness of fit, a sumably established by their salience within judgment that can only be accomplished the manifold varieties of human experience. in the realm of thought. Ambiguity can Some experiences enter consciousness more extend beyond lexical expressions like these readily or more naturally than others. We are to include aspectual distinctions, as when not usually conscious of breathing or blink- I’m flying to Washington may mean that the ing our eyes, but we can become so. It fol- speaker is in the middle of the flight, is plan- lows that questions regarding the content of ning to undertake it, or is doing it gener- consciousness should not be phrased cate- ically these days. Consciousness is capable P1: KAE 0521857430c13 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:57

366 the cambridge handbook of consciousness

of focusing on such distinctions. Without about this difficulty: I don’t know quite how conscious access to thought, the experiences to say it,orthat’s not exactly what I meant. mentioned in this paragraph would be diffi- Noam Chomsky wrote, cult to explain. Now what seems to me obvious by introspection is that I can think without language. In fact, very often, I seem to be think- Distinguishing Thought Organization ing and finding it hard to articulate what I From Semantic Organization am thinking. It is a very common experience at least for me and I suppose for everybody In Figure 13.1 thought and semantics were to try to express something, to say it and to assigned to separate boxes, with the seman- realize that is not what I meant and then to try to say it some other way and maybe tics box representing ways in which thoughts come closer to what you meant; then some- are organized to fit the requirements of lan- body helps you out and you say it in yet guage. Language often shows a conscious- another way. That is a fairly common expe- ness of this distinction, a fact which suggests rience and it is pretty hard to make sense of that thought and its semantic organization that experience without assuming that you are both consciously available. think without language. You think and then you try to find a way to articulate what you think and sometimes you can’t do it at all; Inadequacies of Language as an you just can’t explain to somebody what Expression of Thought you think (Chomsky, 2000,p.76). We can return to our example: This kind of experience implies an ability to (a) You’d look at the sun, compare what one is thinking with possible (b) it just looked red; ways of organizing it linguistically, ways that (c) I mean you couldn’t see the sun. depend in the first instance on the semantic resources of one’s language. Again, it seems In line (c) the phrase I mean is a device that people are conscious of both thoughts made available in the English language to and semantic options because they are able show that one is having difficulty expressing to evaluate differences between them. thoughts with available semantic resources. All three lines of this example recorded the The Tip-of-the-Tongue Experience speaker’s attempt to verbalize her memory of what she had experienced 2 days earlier, In the familiar tip-of-the-tongue phenome- and evidently by the end of line (b) she was non (A. S. Brown, 1991; Brown & McNeill, not fully satisfied with what she had said. 1966; Caramazza & Miozzo, 1997), one Line (c) was a further attempt to put what experiences a thought and its categoriza- she was thinking into words. In some logical tion, but has difficulty accessing a word or sense, line (c) contradicted what she said in name. What is missing is the connection to (a) and (b), but the total effect of all three phonology. Suppose one is thinking of a rel- lines was to convey in different ways the atively unfamiliar object like an astragal but manner in which the sun was obscured by is unable to retrieve that sound. One can smoke, the larger idea on which her thoughts reflect on many aspects of the idea, cate- were focused. Line (c) provides one kind gorize it, and thus attribute to it a variety of evidence of a mismatch between thought of traits, but the phonological representation and possible linguistic choices. that is usually provided by categorization is When people talk, their speech often absent. Jackendoff (1987,p.291) mentions exhibits disfluencies: hesitations, false starts, this experience as evidence for the inabil- and rewordings that are evidence that people ity to be conscious of thought, but it can experience difficulty “turning thoughts into easily be seen as evidence for just the oppo- words.” People are sometimes quite explicit site: consciousness of a thought but only a P1: KAE 0521857430c13 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:57

language and consciousness 367

glimmering of phonology at best. The expe- some of the viewers returned on one or more rience may or may not include imagery, occasions to talk about it again (Chafe, 1991). but it always has an ideational component. The film depicted a theft of pears, and the Chafe (1996b, p. 184) mentions temporary following excerpts show how one person consciousness of a thought expressed by the talked about certain events in three succes- word inconclusiveness without access to that sive versions: sound, and probably in that case without Version 1 (after approximately 15 min- imagery. This experience, when imagery is utes) lacking, can be the purest kind of evidence for “imageless thought,” a major issue for the (a) And he ended up swiping one of his bas- Wurzburg¨ school a century ago (Humphrey, kets of pears, 1951). (b) and putting it on his bicycle. (c) On the front of his bicycle, Categories Lacking a Sound Altogether (d) and riding off. There are certain categories that play a Version 2 (after 6 weeks) useful role in the organization of thought (e) And so, but lack any phonological representation. In some cases a word may be available to spe- (f) he finally gets it, cialists, but it is a word of which most people (g) on his bike, are ignorant. An example might be the small (h) and he starts riding away. sheath on the end of a shoelace that allows Version 3 (after a year) the lace to be passed through a small hole. One might or might not learn the word aglet (i) And just put it right on his . . . on the sooner or later, but probably most who are front of his bicycle, totally familiar with both the thought and its (j) and rode off with it. categorization have never done so. There is a resemblance here to the tip-of-the-tongue One cannot expect thoughts to remain experience, but in this case the phonologi- constant over time, of course, but these three cal representation is wholly unavailable, not versions have enough in common to suggest simply difficult to retrieve. the retention of certain ideas in contrast to the rather different language that was chosen to express those ideas each time. All Repeated Verbalizations three versions conveyed ideas of two events: One promising way to compare thoughts the boy’s placing a basket of pears on with their semantic organization is to exam- his bicycle followed by his departure from ine repeated verbalizations of what can the scene. The participants in those events be assumed to be more or less the same included the boy, verbalized as he or simply thoughts experienced by the same person omitted; the basket of pears, verbalized as it; on different occasions (Chafe, 1998). Data and the bicycle, verbalized as his bicycle or of this kind arise only fortuitously in the his bike. Probably the speaker experienced recording of actual speech, but instances can imagery each time, most vividly in version 1 be elicited by giving people something to and least vividly in version 3. A listener (or talk about and recording what they say at dif- reader of the above) might experience ima- ferent times. Bartlett (1932) pioneered this gery too, but it would necessarily be differ- method with a written stimulus and written ent from that of someone who saw the film. responses. A similar procedure was followed Data of this kind highlight the distinction by Chafe and his associates using a short film between relatively constant thoughts and and spoken reponses (Chafe, 1980). Descrip- the less constant language used to express tions of what happened in the film were them. The idea of positioning the basket of recorded shortly after it was viewed, and pears was expressed as put it on his bicycle or P1: KAE 0521857430c13 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:57

368 the cambridge handbook of consciousness

put it on the front of his bicycle in versions 1 and it onto his bicycle. There is no correspond- 3, but get it on his bike in version 2. The idea ing category in English. One could say he of the boy’s departure was expressed as ride lifted it with difficulty but the effect is not off or start riding away or ride off with it. These the same, lacking the association with weight are not just partially different words, but par- and with a broader range of circumstances tially different ways of organizing thoughts under which the German category might semantically. be employed. Line (a), furthermore, com- As with the pronoun it that expressed the bines its elements in ways that are foreign to idea of the sun in the earlier example, the English. The entire sequence might be trans- pronouns he and it in this example show lated he has trouble lifting the basket, which is again that speakers are conscious not only almost as big as he is, onto his big bicycle, and of imagery but also of ideas. One may expe- rides off, but neither this nor any other trans- rience imagery of a boy placing a basket of lation can capture the effect of the original pears on his bicycle, but it is interpreted in with any precision. terms of ideas of the boy, the basket, and Translation is an attempt to join two lan- the bicycle, and consciousness of those ideas guages in the area of thought, and a com- is essential for the production of language. pletely successful translation would be one In this case the ideas themselves were rel- in which exactly the same thoughts were atively constant across time and across the conveyed by both the source and target lan- communicative act. Their imagistic accom- guages. Because the semantic resources of paniments may have had some constancy for different languages are never identical, it the speaker, but they were left to the listen- is a goal that can never be fully achieved. ers’ imaginations. Because the goal is to connect the languages in the realm of thought, translating cannot Language Differences be accomplished at all without a consciousness of thoughts. One does the best one Different languages provide their speakers can to express the thoughts that were origi- with different semantic resources, including nally expressed with the semantic resources different inventories of categories, different of one language while being unavoidably orientations, and different combinatory pat- constrained by the semantic resources of terns. It is possible for thoughts to be more or another. less the same regardless of the language used This German example provides a fur- to verbalize them, but they can be molded ther illustration of the way repeated tellings differently by different semantic resources. can triangulate on constancies of thought. Semantic choices may then feed back on the As this speaker watched the film she was thoughts themselves. evidently impressed with how difficult it The film just mentioned was shown to was for the boy to transfer the basket from a number of German speakers, one of the ground to his bicycle. With the lan- whom expressed her thoughts concerning guage quoted above she chose the wuchten the events described above as follows: category to express that difficulty. Later she (a) er wuchtet den fast so grossen Korb wie used the following language: er selbst es auch ist, (d) (er) nahm einen der Korbe,¨ (b) auf das grosse Fahrrad, (e) und hiefte ihn mit großer Anstrengung (c) und fahrt¨ dann davon. auf sein Fahrrad, German provides a semantic category that (f) und fuhr davon. allows the idea of an event to be verbalized with the verb wuchten. That choice asso- A possible translation is he took one of the ciates the idea of lifting with weight, entail- baskets, and heaved it with great effort onto ing in this case that the weight of the bas- his bicycle, and rode off. In line (e) the ket caused the boy to have difficulty lifting idea of lifting was expressed with the verb P1: KAE 0521857430c13 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:57

language and consciousness 369

hiefte, similar to English heaved, capturing tints of the gardens. Beyond the lawn, with again an impression of great effort that was its pyramidal pale-gold maples and velvety then made explicit with the phrase mit firs, sloped pastures dotted with cattle; and großer Anstrengung. But this time her choice through a long glade the river widened like focused on the difficulty that was suggested a lake under the silver light of September by the boy’s way of moving, not by the (p. 79). basket’s weight. Every choice of a seman- Language like this allows readers to experi- tic category is a way of molding thought ence vicariously what was passing through with a unique complex of associations, using the consciousness of the protagonist at the resources that often differ from language to very time she was sitting and leaning her language. head, with a succession of olfactory and visual experiences. The impression conveyed by such language is that these experi- Manipulations of Consciousness ences were not being reported from the per- in Literature spective of a later displaced consciousness, but that they were immediate. They were, nevertheless, presented in the Examining how people talk is not the only past tense, and in ordinary speech the past linguistic avenue to an understanding of con- tense is compatible with a displaced consciousness. Authors of fiction have discov- sciousness, with experiences recalled from ered various ways to involve their readers in a time preceding the time the language a fictional consciousness, and studying such was produced. Its use here, as well as the devices can lead to understandings of con- third-person references to herself and her, sciousness that might otherwise be difficult imply the existence of a narrating con- to achieve (Chafe, 1994; Cohn, 1978). The sciousness that is separate from the protag- examples below are taken from Edith Whar- onist’s. With relation to that consciousness ton’s novel The House of Mirth, first pub- the time of Lily’s experiences and her iden- lished in 1905 (Wharton, 1987). tity were displaced – thus the past tense There are several ways in which literature and third person. Chafe (1994) termed this highlights qualitative differences between an artifice displaced immediacy. The experi- immediate and a displaced consciousness. ences of the protagonist, although displaced The most obvious difference may be in the with tense and person, achieve immediacy richness of detail that is available to immedi- through Lily’s actions and the sensory detail. ate consciousness. When something is avail- Her consciousness is available to the reader able to direct perception, even though con- as an immediate consciousness while the sciousness can focus on only a small portion narrating consciousness responsible for the at a time, a wealth of information is poten- tense and person remains unacknowledged, tially available. With a displaced conscious- providing little more than a tense and person ness, as with imagery, detail is necessarily baseline. impoverished. The act of writing gives writ- There was also a consciousness of affect- ers a freedom to verbalize details of a kind laden judgments and comparisons: appropriate to an immediate consciousness, allowing readers to share in a fictional imme- Lily smiled at her classification of her diacy: friends. How different they had seemed to her a few hours ago! Then they had sym- Seating herself on the upper step of the ter- bolized what she was gaining, now they race, Lily leaned her head against the hon- stood for what she was giving up. That eysuckles wreathing the balustrade. The very afternoon they had seemed full of bril- fragrance of the late blossoms seemed an liant qualities; now she saw that they were emanation of the tranquil scene, a land- merely dull in a loud way. Under the glitter scape tutored to the last degree of rural ele- of their opportunities she saw the poverty gance. In the foreground glowed the warm of their achievement. It was not that she P1: KAE 0521857430c13 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:57

370 the cambridge handbook of consciousness

wanted them to be more disinterested; but as generally conceived. At the other extreme she would have liked them to be more pic- the word could be restricted arbitrarily to turesque. And she had a shamed recollec- conscious experiences alone, so that thought tion of the way in which, a few hours since, would be conscious by definition. That alter- she had felt the centripetal force of their native need not be rejected out of hand, but standards (pp. 88–89). it excludes various phenomena that many This passage, expressing thoughts that would prefer to include within the realm passed through Lily’s consciousness, retains of thought. We can briefly examine three of the quality of an immediate consciousness them. combined with the displacement expressed by the past tense and the third-person she. Reasoning But the passage is noteworthy, in addition, One such phenomenon is reasoning. Jack- for the fact that its adverbial expressions – endoff (1997), in fact, presents reasoning a few hours ago, then, now, that very after- as a core ingredient of thought and uses noon – have their baseline in the time of its supposed unavailability to consciousness the immediate consciousness, not of the dis- as crucial evidence that thought is uncon- placed consciousness responsible for tense scious. The form that reasoning takes in ordi- and person. The now, for example, was nary experience, however, is far from well the now of Lily’s immediate experiences, understood, and traditional logic is of lit- not the now of the language production. tle help. People do very often take advan- Although tense and person are anchored in tage of things they have experienced in order what Chafe (1994) has called the representing to make inferences about things they have consciousness – the consciousness producing not experienced, such as distal events that the language – temporal adverbs like now are have occurred or might hypothetically occur. anchored in the represented immediate con- All languages have devices of some sort for sciousness. In their treatments of conscious- expressing the inferential nature of ideas, ness novelists thus manage to throw light on although different languages do it in differ- the separate functioning of tense and person ent ways (Aikhenvald & Dixon, 2003; Chafe as opposed to the functioning of adverbs, & Nichols, 1986; Palmer, 2001). A simple a distinction that might not otherwise be example from English is the use of must as apparent. in you must have had a haircut, where direct experience of your appearance has led to the idea of a displaced event. Whether inferen- Is There Thought Outside tial reasoning of this sort operates outside of Consciousness? of consciousness is an open question, but it does seem that people use words like must without any primary awareness of what they If we are conscious of what we are thinking are doing. in ways suggested above, that need not imply that all of thought is conscious. Whether Orientations there is unconscious thought is a question that calls first for agreement on the mean- Within the realm of thought, ideas are ing of thought. To consider and reject one positioned in a multidimensional matrix of extreme possibility, it would not be use- orientations. They may be positioned, for ful for this word to encompass everything example, in time, space, epistemology (as the brain does, including totally unconscious above with must), affect, social interaction, processes like those involved in the regula- and with relation to the context provided by tion of body chemistry, or semi-involuntary neighboring ideas. These orientations affect processes like breathing, none of which the shape of languages in many ways, and would belong within the domain of thought the semantic resources of different languages P1: KAE 0521857430c13 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:57

language and consciousness 371

give different prominence to different ones. of ideas, but also of the ways in which those English, for example, pays a great deal of ideas are oriented. On the one hand, if spe- attention to tense. In the English retelling cific orientations express variable semantic quoted earlier, the first version was in the choices and not elements of thought per se, past tense: they may not bear directly on consciousness of thought but only on the semantic, linguis- (a) And he ended up swiping one of his bas- tically imposed organization of thought. On kets of pears, the other hand, semantic choices may often (b) and putting it on his bicycle. feed back into thought itself, so that rigidly (c) On the front of his bicycle, separating semantics from thought may in (d) and riding off. the end be misleading. Whether people are or can be conscious of orientations may turn whereas the second was in the present: out to depend on the particular orientations (e) And so, involved – some being more available to con- (f) he finally gets it, sciousness than others – and even perhaps on (g) on his bike, the varying semantic sensitivities of different individuals. (h) and he starts riding away. Although English is semantically con- Relations strained to orient events in this way and so to anchor ideas with relation to the time the The elements of consciousness are not inde- language is produced, these examples show pendent of each other, but are interrelated that the choice of a particular tense varies in a variety of ways, and there is a question easily from one telling to the next. It thus as to whether at least some of these relations seems to be a choice that is made during lie outside of consciousness. A simple exam- the act of producing language, not a prop- ple is provided by the first two lines of the erty of underlying thought. The thoughts in example quoted above: these two excerpts may have been more or (a) You’d look at the sun, less the same, but the speaker chose to ori- (b) it just looked red. ent them with two different tenses as she adjusted them to the semantic constraints of The foci of consciousness expressed in lines English. Other languages may pay less atten- (a) and (b) bear a relation that might tion to tense and be more preoccupied, say, be described as conditionality. There is an with aspectual or epistemological distinc- understanding that the event that was ver- tions. But whatever a language does, such balized in (a) took place one or more times choices are not likely to be in the forefront and that each time it occurred it was a nec- of consciousness. essary condition for the state that was ver- As another example, the ubiquitous balized in (b). For this person to experi- English word the, which orients an idea ence the appearance of the sun, she first had as identifiable to the listener, has been to look at it. Relations like this are often responsible for numerous articles and books verbalized overtly with little words like if and is still an object of controversy (e.g., and when, but in this case the conditional Lyons, 1999). If its semantic contribution has relation was only implied (Chafe, 1989). As proved so difficult to specify and if, further- with orientations, the question of whether more, it is so hard for Japanese, Korean, and speakers of a language are conscious of such other learners of English as a second lan- relations, marked or not, remains open. Are guage to assimilate, its availability to con- they so integral to the succession of consciousness may be questioned. scious ideas that they are conscious too, or do The general question is whether or to speakers and thinkers employ them without what extent people are conscious, not just being conscious they are doing so? Language P1: KAE 0521857430c13 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:57

372 the cambridge handbook of consciousness

itself does not provide clear answers to such helpful comments on this chapter. If Jack- questions. endoff and I do not always agree, his work has always provided a stimulating balance to tendencies that might otherwise have Summary been weighted too strongly in a different direction. This chapter has focused on two different approaches to the relation between language References and consciousness, both linguistically oriented. Both agree that this relation cannot be explored without taking other mental phe- Aikhenvald, A. Y., & Dixon, R. M. W. (Eds.). 2003 nomena into account, in particular thought ( ). Studies in evidentiality. Amsterdam: John Benjamins. and imagery. They agree, furthermore, that 2004 conscious experience includes both imagery Anthony, J. L., & Lonigan, C. J. ( ). The nature of phonological awareness: Converging and affect, whether or not those experiences evidence from four studies of preschool and can be considered elements of thought. early grade school children. Journal of Educa- Jackendoff sees consciousness as limited tional Psychology, 96, 43–55. to uninterpreted imagery, whose qualities Bartlett, F. C. (1932). Remembering: A study in mirror those of uninterpreted visual, audi- experimental and social psychology. Cambridge: tory, or other raw sensory information. He Cambridge University Press. presents evidence that such imagery is exter- Bloomﬁeld, L. (1931). Obituary of Albert Paul nal to thought, in part because it is too par- Weiss. Language, 7, 219–221. ticular, in part because it does not allow Brown, A. S. (1991). A review of the tip-of-the- the identiﬁcation of individuals, and in part tongue experience. Psychological Bulletin, 109, because it fails to support reasoning. If all we 204–223. are conscious of is imagery and imagery does Brown, R., & McNeill, D. (1966). The “tip of the not belong to thought, it follows that we are tongue” phenomenon. Journal of Verbal Learn- not conscious of thought. ing and Verbal Behavior, 5, 325–337. Chafe distinguishes between immediate Buswell, G. T. (1935). How people look at pictures: and displaced consciousness, the former A study of the psychology of perception in art. engaged in direct perception and the latter Chicago: University of Chicago Press. in experiences that are recalled or imagined. Caramazza, A., & Miozzo, M. (1997). The Immediate consciousness includes not only relation between syntactic and phonological sensory experiences but also their interpre- knowledge in lexical access: Evidence from tation in terms of ideas, which are positioned the “tip-of-the-tongue” phenomenon. Cogni- 309 343 within a complex web of orientations and tion, 64, – . relations. Displaced consciousness includes Castles, A., & Coltheart, M. (2004). Is there a sensory imagery that is different in quality causal link from phonological awareness to suc- 77 111 from immediate sensory experience, but it cess in learning to read?Cognition, 91, – . 1980 is always accompanied by ideational inter- Chafe, W. (Ed.). ( ). The pear stories: Cogni- pretations that resemble those of immedi- tive, cultural, and linguistic aspects of narrative production. Norwood, NJ: Ablex. ate experience. Both the imagistic and the 1989 ideational components of consciousness are Chafe, W.( ). Linking intonation units in spoken English. In J. Haiman & S. A. Thompson held to be central components of thought, (Eds.), Clause combining in grammar and dis- as thought is ordinarily understood. course (pp. 1–27). Amsterdam: John Benjamins. Chafe, W. (1991). Repeated verbalizations as evi- Acknowledgments dence for the organization of knowledge. In W. Bahner, J. Schildt, & D. Viehweger (Eds.), Pro- ceedings of the Fourteenth International Congress I am especially grateful to Ray Jackendoff of Linguists, Berlin 1987 (pp. 57–68). Berlin: and the editors of this volume for their Akademie-Verlag. P1: KAE 0521857430c13 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:57

language and consciousness 373

Chafe, W. (1994). Discourse, consciousness, and James, W. (1890). The principles of psychology. time: The flow and displacement of conscious New York: Henry Holt. experience in speaking and writing. Chicago: Langacker, R. W. (1997). Consciousness, con- University of Chicago Press. strual, and subjectivity. In M. I. Stamenov Chafe, W. (1996a). How consciousness shapes (Ed.), Language structure, discourse and the language. Pragmatics and Cognition, 4, 35–54. access to consciousness (pp. 49–75). Amster- Chafe, W. (1996b). Comments on Jackendoff, dam: John Benjamins. Nuyts, and Allwood. Pragmatics and Cognition, Lyons, C. (1999). Definiteness. Cambridge: Cam- 4, 181–196. bridge University Press. Chafe, W. (1998). Things we can learn from Natsoulas, T. (1987). The six basic concepts of repeated tellings of the same experience. Nar- consciousness and William James’s stream of rative Inquiry, 8, 269–285. thought. Imagination, Cognition and Personal- Chafe, W. (2000). A linguist’s perspective on ity, 6, 289–319. William James and the stream of thought. Con- Newtson, D., Engquist, G., & Bois, J. (1977). The sciousness and Cognition, 9, 618–628. objective basis of behavior units. Personality Chafe, W. (2002). Prosody and emotion in a sam- and Social Psychology, 35, 847–862. ple of real speech. In P. Fries, M. Cummings, D. Palmer, F. R. (2001). Mood and modality (2nd ed.). Lockwood, & W. Sprueill (Eds.), Relations and Cambridge: Cambridge University Press. 277 functions within and around language (pp. – Piaget, J. (1967). Six psychological studies.(D. 315 ). London: Continuum. Elkind, Ed. & A. Tenzer, Trans.). New Chafe, W. (2005). The relation of grammar York: Vintage. (Original work published to thought. In C. S. Butler, M. de losdr 1964) ´ Angeles Gomez-Gonz´ alez,´ & S. M. Doval- Poltrock, S. E., & Brown, P. (1984). Individual dif- Suarez´ (Eds.), The dynamics of language use: ferences in visual imagery and spatial ability. Functional and contrastive perspectives. Amster- Intelligence, 8, 93–138. dam: John Benjamins. Poppel,¨ E. (1994). Temporal mechanisms in per- 1986 Chafe, W., & Nichols, J. (Eds.). ( ). Evi- ception. In O. Sporns & G. Tononi (Eds.), Selec- dentiality: The linguistic coding of epistemology. tionism and the brain: International review of Norwood, NJ: Ablex. neurobiology (Vol. 37,pp.185–201). San Diego: Chomsky, N. (2000). The architecture of language. Academic Press. New York: Oxford University Press. Russell, B. (1921). Analysis of mind. London: Allen Cohn, D. (1978). Transparent minds: Narrative & Unwin. modes for presenting consciousness in fiction. Stamenov, M. I. (Ed.). (1997). Language structure, Princeton, NJ: Princeton University Press. discourse and the access to consciousness. Ams- Holsˇanov´ a,´ J. (2001). Picture viewing and picture terdam/Philadelphia: John Benjamins. description: Two windows on the mind. Lund Wharton, E. (1987). The house of mirth. New 83 University Cognitive Studies . Lund: Lund York: Macmillan. (Original work published University Cognitive Science. 1905) 1951 Humphrey, G. ( ). Thinking. New York: Wierzbicka, A. (1999). Emotions across language Wiley. and cultures: Diversity and universals. Cam- Jackendoff, R. (1987). Consciousness and the com- bridge: Cambridge University Press. putational mind. Cambridge, MA: MIT Press. Zacks, J. M., Braver, T. S., Sheriden, M. A., Don- Jackendoff, R. (1996). How language helps us aldson, D. I., Snyder, A. Z., Ollinger, J. M., think. Pragmatics and Cognition, 4, 1–34. Buckner, R. L., & Raichle, M. E. (2001). Human Jackendoff, R. (1997). The architecture of the lan- brain activity time-locked to perceptual event guage faculty. Cambridge, MA: MIT Press. boundaries. Nature Neuroscience, 4, 651–655. P1: KAE 0521857430c13 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:57

374 P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

CHAPTER 14 Narrative Modes of Consciousness and Selfhood

Keith Oatley

Abstract centre of agency in relation to others, has emerged gradually during evolution, during Beyond mere awareness, human conscious- cultural development, and during individ- ness includes the reflexive idea of conscious- ual development. It functions importantly ness of self as a centre of agency and expe- in social interaction and allows integrative rience. This consciousness of self has been understanding both of oneself and of others. thought to involve narrative: a distinct mode of thinking about the plans and actions of agents (self and others), about vicissitudes Introduction encountered, about attempts to solve problems posed by these vicissitudes, and about Consciousness of the kind we value – of our the emotions that arise in these attempts. surroundings, of our thoughts and emotions, Philosophical discussion of this idea has of our selves, and of other people – often included the question of to what extent this takes narrative forms. However, there is de- narrative-of-consciousness is epiphenome- bate about how to define narrative. A recent nal and to what extent it may have causal discussion of the question is given by Wilson effects on action. To the extent that the self (2003). A minimalist definition is that nar- takes narrative forms and is constructive, it rative must include at least two events with will tend to assimilate narratives encoun- some indication of their temporal order- tered in the public space: stories that occur ing. The difficulty with such definitions is in conversation and elsewhere. Both the style that on the one hand they are unenlight- and content of stories that circulate in a cul- ening, and on the other that even the see- ture will potentially contribute to the extent mingly unobjectionable ones seem to in- and contents of consciousness, and therefore vite contention and counter-example. I do to the development of selfhood. A narratiz- not base this chapter, therefore, on a defi- ing consciousness, in which self is a unifying nition of narrative. Instead, I adopt the more

375 P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

376 the cambridge handbook of consciousness

r psychological stance of Bruner (1986) who Characters have conscious awareness of writes that narrative “deals in human or vicissitudes – that is to say, problems – human-like intention and action and the as well as of their own problem-solving vicissitudes and consequences that mark attempts and outcomes. r their course” (p. 13). Readers and listeners are affected by the The theory of narrative is known as nar- characters’ problem-solving attempts and ratology (Bal, 1997; Groden & Kreiswirth, by the inferred experiences of the charac- 1994). Literary theorists such as Booth ters (emotional and otherwise). 1988 r ( ), as well as cognitive scientists such as A story offers implications of meanings, 1975 1990 Rumelhart ( ) and Schank ( ), have often moral meanings, in relation to the contributed to it, as have theorists of identity culture in which the story arose. and biography, such as Brockmeier & Car- baugh (2001). Following Bruner’s proposal, Narratives include utterances of one person here is a sketch of prototypical narrative telling another what happened at the office and its elements. I label some as interaction- that day. Stories are structured more con- type elements. They are typical of narrated sciously. They include anecdotes that may sequences of intended social interaction. I be told to friends, newspaper articles, films, label a further set as story-type elements that novels, and so forth. When I use the term are typical of those additional features that “narrative framework,” I mean the kind of make narratives stories. prototype indicated above, with or without characteristics of a story. One may notice Interaction-Type Elements of Narrative that such frameworks are not purely syntactic. They include both structure and content. Interaction-type elements include the fol- The analyses I offer here are Western in lowing: their provenance. Although I believe that r There are agents who may be called char- many of the themes I discuss are universal acters. (see, e.g., Hogan, 2003), there would be dif- r There is a focus on intentions of the char- ferences in this kind of argument if it were acters. elaborated within, say, Chinese or Indian r Events occur in a causal, time-ordered culture. sequence. These events include mental, One must of course distinguish between interpersonal, and physical actions, some narratives as cultural objects and narrative 2003 of which flow from the characters’ emo- thinking (see, e.g., Goldie, ). One must tions and intentions. also distinguish between a life lived and a r life thought about in narrative terms as elab- Vicissitudes of intentions and actions orated, for instance, by a biographer. Given occur and affect characters emotionally. r these distinctions, however, an advantage of Outcomes include further physical, men- adopting Bruner’s (1986) view of narrative as tal, and social events, which affect char- a specific mode of thought about agents and acters emotionally. their intentions is that it is accompanied by his parallel proposal of the mode of thought Story-Type Elements of Narrative that he calls paradigmatic, which “attempts to fulfill the ideal of a formal, mathemat- Story-type elements of narrative include the ical system of description and explanation. following: It employs categorization or conceptualiza- r A story is a narrative account that uses tion and the operations by which categories interaction-type elements (as indicated are established, instantiated, idealized, and above), and that is related or depicted by related one to another to form a system” an explicit or implicit narrator in some (p. 12). Bruner’s starting point is helpful for medium (talk, text, film, etc.). this chapter because narrative seems to be P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

narrative modes of consciousness and selfhood 377

the mode in which important aspects of categorized section. Those readers who wish human consciousness occur and because it to see what work has been done in a par- offers a basis for making meaningful sense of ticular area may turn to the corresponding ourselves and of our interactions with others. section for discussion of representative work By contrast the paradigmatic mode is that by and references. which science, including a science of con- r The question of whether consciousness sciousness, is expressed most typically. has causal properties: Does conscious- Although consciousness used to be regar- ness affect action, or does it occur only ded almost as a non-topic because it was as retrospective commentary? unanalyzable, it has recently become the r object of considerable interest; witness this Four aspects of consciousness: (i) Simple handbook. Likewise, the topic of narrative awareness, (ii) the stream of inner con- is drawing increasing scholarly attention not sciousness, (iii) conscious thought as it just within narratology but also, for instance, may affect decisions and action, (iv) consciousness of self-with-other. in narrative analyses of people’s lives and r identities, with the idea that thinking within Narrative and mutuality: Personal and narrative frameworks in the social sciences interpersonal functions of consciousness amounts to a paradigm shift (Brockmeier & in humans as highly social beings who Harre,´ 2001). depend on mutuality with others. r The idea of speciﬁcally narrative con- The evolution of narrative: Thinking and sciousness is also growing (see, e.g., Fire- consciousness in hominid species and man, McVay, & Flanagan, 2003) and can humans among whom both self and oth- be thought of as the intersection between ers are known as individuals with inten- interest in consciousness and interest in nar- tions; questions of mental models, emo- rative. In itself this intersection is by no tions, and imagination. r means new; in the theory of literary nar- The developmental psychology of narratives, for instance, consciousness and its rative consciousness: Individual develop- transformations have been topics of schol- ment of narratizing consciousness in chil- arly attention for many years (see Auerbach, dren, the role of language, and the idea of 1953; Watt, 1957). In addition, for a hun- narrative consciousness in psychotherapy. r dred years, in areas touched by psychoanal- The rise of consciousness in Western ysis, this intersection has been central as imaginative literature: Changes that have the formative idea of psychoanalysis is that occurred in the depiction of conscious- consciousness is reclaimed from the uncon- ness from the earliest narrative writings scious precisely by turning it into language to the present. within narrative frameworks (see, e.g., Edel- r Coda: The relation of conscious to son, 1992; Spence, 1982). What is relatively unconscious actions and thoughts. new is that this intersection has become of interest for cognitive psychology and for the For those who want something more philosophical theory of mind. It is within like a story, please read on. I have argued these frameworks that this chapter is writ- elsewhere (Oatley, 1992a) that psychology ten, although with links to other disciplinary is a subject in which we expect both an areas. approach directed to the attainment of Given that this is a handbook, a reader’s insight (characteristic of the narrative mode) expectation will properly be of a system of and information of the technical kind (in the categories and explanations in what Bruner paradigmatic mode). Styles of writing in psy- has called the paradigmatic mode. A list of chology need typically to address both these section headings is therefore given below. expectations. In something more like a story The clauses that succeed the main headings framework, therefore, I have conceived this in this list are pointers to the contents of each chapter partly as a debate of a protagonist P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

378 the cambridge handbook of consciousness

(me) and an antagonist. The protagonist puts ticism about the functionality of conscious- forth the idea that consciousness has causal ness remains strong. properties. The antagonist is sceptical of this Thus, although emotions are salient in claim. At the same time, I tell a story (a true consciousness (as discussed below), LeDoux story I hope) about the evolution and devel- (1996), prominent for his work on emotions opment of consciousness. and the amygdala, has written, “The brain states and bodily responses are the fundamental facts of an emotion, and the con- The Question of Whether scious feelings are the frills that have added Consciousness Has Causal Properties icing to the emotional cake” (p. 302). This is the kind of view that is represented by I propose that narrative consciousness has Dennett, not just for emotional conscious- functions. Each of us is, at least in part, a ness, but for consciousness generally. In this centre of consciousness: a unified narrative view the brain computes what it needs to agent. Many though not all of our decisions, compute. It recruits the neurons necessary thoughts, and actions flow from this centre. to produce behavior. What we experience By means of a narrative consciousness, we as consciousness is an extra, with no causal can begin to understand ourselves and oth- effects. Human behavior could occur with- ers in relation to the societies in which we out it. Consciousness is a rather narrow sum- live and to make sense of what would other- mary in narrative form of what has already wise be disconnected. My proposal derives happened. The real workings of mind occur from the idea that, to understand anything elsewhere. satisfactorily in psychology, one must understand its functions. In this chapter I begin to All varieties of perception – indeed all vari- answer the question of what the functions eties of thought or mental activity – are accomplished in the brain by parallel, mul- of narrative consciousness might be. titrack processes of interpretation of sensory Intuitively, we think that we act on our inputs (Dennett, 1991,p.111). conscious perceptions, beliefs, and desires. But this intuition that consciousness is func- Humphrey and Dennett (2002) offer an tional is an expression of folk theory. The analogy with a termite colony, which “builds opposition in the debate is provided by a elaborate mounds, gets to know its territory, sceptic, such as Daniel Dennett (e.g., 1991). organizes foraging expeditions” (p. 28). But Dennett is representative of many cogni- for all this apparent purposefulness, it is just tive scientists who believe that scientific a collection of termites with various roles, understandings of the brain have replaced, influenced by each other but not by any mas- or will replace, explanations based in folk ter plan. theory. Neurons cause behavior, they say. Dennett calls the idea of a unifying spec- Conscious thoughts do not. Folk theory – tacle that constitutes our point of view the which includes ideas of beliefs, desires, and Cartesian Theater. The epithet comes from emotions; the idea that we act for reasons; Descartes’ (1649/1911) idea of a soul housed and the idea of a conscious self that func- in the brain. The idea of a soul witnessing tions as a unified agent – is false, say such and forming meanings is too thoroughly old scientists and philosophers (e.g., Church- fashioned. There is no Cartesian Theater, land, 1986). It is false not just in its details says Dennett, no “Oval Office in the brain, but altogether, just as were notions that housing a Highest Authority” (1991,p.428). stars and planets revolve in perfect circles The conscious narratizing process is mis- round the earth because they were stuck taken about how the mind works. Instead, to crystalline spheres. Although Church- Dennett says, think of the conscious self as land’s brand of radical eliminativism (elim- “a center of narrative gravity” (1991,p.418). ination of all folk-theoretical terminology He wants us to realize that a centre of gravity from cognitive science) is on the wane, scep- is an abstraction, not real. He writes, P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

narrative modes of consciousness and selfhood 379

Our tales are spun, but for the most part we ken, and you need a shovel to clean out the don’t spin them; they spin us. Our human chicken shed” (p. 90). The left (language- consciousness, and our narrative selfhood, equipped) side of the brain had seen the claw is their product, not their source (p. 418). and chose the chicken picture. But this same I suppose Dennett wants us to understand left side had no consciousness of the snow the terms “spun” and “spin” in the foregoing scene; instead it saw the left hand choosing quotation as the output of a spin doctor. In the shovel and it produced a reason without another paper he has extended this idea: Not knowing it was – as one might say – a piece only does consciousness have no causal sta- of fiction to fit what it knew. Also there are tus but it is also mischievously misleading: people who, because of brain damage, say they see nothing on one side of the visual [W]e are virtuoso novelists, who find our- field and yet react perfectly well to objects selves engaged in all sorts of behavior, more there. The phenomenon is called blindsight. or less unified, but sometimes disunified, Humphrey (2002) writes, “I have met such and we always put the best “faces” on it a case: a young man who maintained that he we can. We try to make all of our material could see nothing at all to the left of his nose, cohere into a single good story. And that and yet could drive a car through busy traf- story is our autobiography (Dennett, 1986, fic without knowing how he did it” (p. 69). p. 114). And, in multiple personality disorders there The chief fictional character at the cen- is not one self but several (Humphrey & tre of that autobiography is one’s self. And if Dennett, 2002). you still want to know what the self really is, Further along the spectrum of doubters you are making a category mistake (Dennett, of the value of narrativity is Galen Strawson 1992,p.114). (2004), who argues that the idea of a nar- So a narratizing consciousness does not rative self is regrettable. There are plenty of just happen after the fact; it’s a phoney. people, he says, whose selves are what he Compelling, for Dennett, are the data of calls “episodic,” with episodes being sepa- simultaneous but disparate brain processes. rate rather than bound together in any nar- For instance, in a group of people who have rative way. He remembers, for instance, hav- had operations to sever their left and right ing fallen out of a boat into the water when hemispheres to relieve epilepsy (split-brain he was younger and can recall details of this patients), the hemispheres can know contra- incident, but claims that this did not hap- dictory things (Gazzaniga, 1998). Only the pen to the self who was writing his article. left side has language, and this side can also Strawson’s intuition of disconnectedness in recount narratives. The right side is silent. life’s episodes is accompanied by the claim Gazzaniga (1992) says his favorite split-brain that to argue for narrative selves is to endorse phenomenon was illustrated by flashing to falsehood: to tell stories is (as Dennett and the left hemisphere of a split-brain patient, Gazzaniga also argue) invariably to select, to PS, a picture of a chicken claw and to the revise, to confabulate, and thereby to falsify. right hemisphere a picture of a snow scene. Because of this falsification, the idea that we Then PS was asked to choose by hand which live narrative lives is therefore not just falla- of a set of pictures in full and continuous cious but ethically flawed. view was associated with the pictures he Should the idea of a self with causal pow- had seen flashed previously. His left hand ers that involve a narratizing consciousness (activated by the right side of his brain and therefore be replaced? I argue that it should prompted by the snow scene) chose a pic- not. If we may take Dennett, again, as the ture of a shovel, and the right hand (acti- central antagonist to the idea of a functional vated by the left side of the brain) chose a narrative consciousness, we can say that picture of a chicken. When asked why he although he is thoroughly immersed in cog- made these choices, he said. “Oh, that’s sim- nitive science, the adjacent area of human- ple. The chicken claw goes with the chic- computer interaction is also important. In P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

380 the cambridge handbook of consciousness

that area there occurs the idea of the inter- I keep track of my memories, my plans, my face. In personal computers nowadays we are commitments, and my interactions with oth- offered the interface of a simulated desk- ers. With it, I know about those activities top, which is supported by layers of com- that make me member of the society in puter programs. Although computers work which I live. by means of electron flows and semiconduc- In this chapter I propose that we accept tors or at a another level by binary opera- not only Dennett’s metaphor of self-as- tions, users would not find any inspection of novelist but also that, as proposed by them useful. They want to do such things Richard Velleman (2002), different conclu- as create documents and spreadsheets, make sions may be drawn than those offered by mathematical calculations, compose musi- Dennett. Following Velleman’s argument, I cal collections, or look at their photographs. explore the idea of a conscious unitary self, Modern interfaces offer interfaces in such based on functional properties of narrative. forms as enhanced virtual papers and other objects on a virtual desk, which enable them to do such things easily. Similarly, as users of Four Aspects of Consciousness our bodies, we do not want to know what our neurons are up to, nor yet the state Consciousness has many meanings. To bring of our hippocampus. We live in a world of out some of the properties of consciousness agents somewhat like ourselves with whom that we value, I propose four aspects (Oat- we interact. We want to know what and who ley, 1988), each named after an originator: is in the world in terms of possibilities of Helmholtz, Woolf, Vygotsky, and Mead. interactions with them. We need an interface on our brain states in terms of inten- Helmholtzian Consciousness tions, beliefs, and emotions. That is what consciousness offers us: a functional con- The first and most basic conception of conception of ourselves as agents, with certain sciousness in modern cognitive science was memories, plans, and commitments. Narra- formulated by Hermann von Helmholtz tive is, as Bruner has proposed, that way of (1866). He proposed that perception was thinking that connects actions to agents and the drawing of unconscious conclusions by their plans. analogy (see, e.g., Oatley, Sullivan, & Hogg, Confabulations certainly occur, and not 1988). He pointed out that we are not con- just by split-brain patients. But one can also scious of the means by which we reach demonstrate the experience of light when cognitive conclusions (neural and cognitive one presses the side of the closed eye in the computations). We are only conscious of the dark. The pressure triggers receptors in the conclusions themselves. retina. This does not mean that a world seen Dennett’s (1991) theory is a version of via light rays detected by the light-sensitive Helmholtz’s idea of perception as uncon- retina is an illlusion. Certainly human conscious inference to reach certain conscious sciousness has some properties of post-hoc conclusions. Dennett says distributed pro- rationalization, as Dennett claims. Certainly cesses of neural computation deliver such its properties are generated by brain pro- perceptual and other mental conclusions, cesses of which, in the ordinary course of which are not necessarily veridical nor events, we know nothing, just as the inter- authoritative. Instead, as he argues, it is as face on the computer on which I am writing if iterative drafts are produced of such con- this chapter is produced by semiconductors clusions, which change in certain respects as whose workings I do not know in detail. But a function of other input. the conclusion is not that consciousness is a So far so good. I agree. Helmholtz’s pro- fraud, any more than the icons on my com- posal is, I believe, a deep truth about how puter screen are hallucinations. A conscious the brain works. Information flows from sen- understanding of self is the means by which sory systems to perceptual and other kinds P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

narrative modes of consciousness and selfhood 381

of interpretation, and from the operations William used to say a lady is known by her of neural systems to experience. Although I shoes and her gloves. He had turned on his did not include this aspect in my 1988 arti- bed one morning in the middle of the War. cle, I think one must include in this proposal He had said, “I have had enough.” Gloves the consciousness of emotions (Panksepp, and shoes; she had a passion for gloves, but 2005), which I discuss further below. Much her own daughter, Elizabeth, cared not a straw for either of them (p. 12). of Helmholtzian awareness is a one-way flow. For example, when one is looking at a No longer based in perception of the visual illusion, even when one knows that it purely Helmholtzian kind, Mrs Dalloway is an illusion, one cannot, by taking thought, does not just see a tailor’s shop and fish- alter the conclusion our visual processes mongers. Her consciousness is an admixture reach. of perceptions, memories (where her father bought his suits), and judgements (“That is Woolfian Consciousness all”). Associations occur. Gloves remind her Helmholtz’s theory is applied principally to of her uncle, and they prompt memories. perception, in which we receive input from And “That is all,” has a mental association excitations of the sense organs that provide with her uncle saying, “I have had enough.” cues to objects and events in the world that The thought of gloves also stirs her preoc- produce these excitations. We are, however, cupation with her daughter Elizabeth, from also conscious of images that occur verbally whom she is on the verge of painful estrange- and visually, with no perceptual input. These ment. Underlying this are other emotional images can be thought of as described by themes: her relationship with her father, her William James (1890) in his metaphor, the status as a lady, repercussions of the war, the “stream of consciousness,” or as the states fact of death. depicted by Virginia Woolf in Mrs Dalloway In Woolfian consciousness, thought (1925): a changing kaleidoscope of thoughts, breaks free of the immediacy of perception. ideas, and images. The verbal and visual No doubt Dennett would be pleased to images depicted by Woolf have as much acknowledge the hints in Woolf’s depiction to do with preoccupations, memories, emo- that consciousness is not entirely unified. tions, and trains of inward thought as they do with perceptual input. Vygotskyan Consciousness As with Helmholtzian consciousness, in A third aspect is the influence of conscious- Woolfian consciousness there is a principal ness on thoughts, decisions, and actions. Mrs direction of flow: from neural process to Dalloway is about a single day on which mental image. What Woolf did, starting with Clarissa Dalloway gives a party. The novel’s her novel Mrs Dalloway, was to make inner opening line is, “Mrs Dalloway said she consciousness recognizable in the form of a would buy the flowers herself.” If Clarissa narrative, a novel. Here we are, for instance, Dalloway were a real person, Dennett would inside the mind of Clarissa Dalloway as she say a set of neurocomputational processes walks up Bond Street in London on a June generated her utterance: “I’ll buy the flow- morning, a year or two after the end of World ers myself.” She gives a reason: “For Lucy War I. [her servant] had her work cut out for her,” . . . a roll of tweed in the shop where her preparing for the party. Part of that work father had bought his suits for fifty years; a included arrangements with Rumpelmayer’s few pearls; salmon on an iceblock. men who were coming to take the doors “That is all,” she said, looking at the fish- off their hinges to allow the guests at the mongers. “That is all,” she repeated, paus- party to move about more freely. According ing for a moment at the window of a glove to Dennett, all such thoughts of reasons for shop where, before the War, you could buy going out to buy the flowers would be post- almost perfect gloves. And her old Uncle hoc elaborations, perhaps confabulations. P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

382 the cambridge handbook of consciousness

Lev Vygotsky (a literary theorist before program learns from them and rewrites parts he became a psychologist) proposed that of itself. For instance it might want to put a thoughts could affect actions. His basic con- block on top of one that is on the floor. But ception was of children developing until say the block on the floor already has some- about the age of 2 years much as apes deve- thing on it. The program cannot complete lop, immersed in their relationships with pa- its plan. Prompted by the discrepancy (the rents and siblings and solving as well as they problem) between its current state and its might the problems that confront them. But goal, it constructs a new piece of plan (pro- then, with the entry of each child into lan- gram). It draws from a library of mistakes guage, individual resources are augmented (bugs) and tries to generalize: For instance, by the resources of culture. Vygotsky says it might conclude, “If I want to put a block the social world becomes internalized as on another block which is on the floor, I must mind. So it is not that the mind is a con- always clear off anything that is on the top tainer that holds thoughts or that conscious of the block on the floor.” thoughts are solely the output of neural pro- Sussman’s program has two kinds of code. cesses. Rather, the mind is an internal social The first is code comparable to that in world, which one can consult and in which any program: detailed plans made up of one can take part. sequences of actions that will achieve goals. A typical example of Vygotsky’s think- The second kind is not represented explicitly ing is from a study by his colleague Levina in most programs. It is an explicit account of (Vygotsky, 1930). She was studying a little the goals of each procedure (e.g., to put one girl who had been given the task of trying to block on top of another). Only with such retrieve some candy, with potential tools of a representation could a program analyze a a stool and a stick available. The child talks mismatch between an intended plan and an to herself. But she does not just say aloud outcome that was not intended. With such a what she has done, as one might think if one representation the program can reason back- had been reading Dennett. She makes sug- wards about the effects of each action in rela- gestions to herself: “‘No that doesn’t get it,’” tion to a goal. With such a representation it she says. “‘I could use the stick.’ (Takes stick, can write new pieces of program – patches – knocks at the candy.) ‘It will move now.’ to solve problems that are encountered and (Knocks candy.)” Moreover the child reflects thereby achieve the goal. on the problem and analyses its solution: “‘It Here is the analogy with consciousness. moved. I couldn’t get it with the stool, but When introducing a patch to the program the, but the stick worked’” (p. 25). (itself), Hacker runs in what Sussman calls Consciousness, here, is not an equivocal “careful mode” in which the patch is com- post-hoc account, but a mobilization of the pared line by line with the range of possi- resources of human culture, which become bly interacting subgoals and goals. Here is a potentially available to each of us. In this case unified agency, trying out imagined actions the child instructs herself as she has been in a model of the world before commit- instructed by adults. According to Vygotsky, ting to action. Hacker itself is not conscious. this is how mind becomes mind. Conscious- But Sussman’s account contains the idea ness is not just a result; it can be a cause. that consciousness is a unifying process that Among the most interesting proposals includes a model of goals and the possible about the functions of a unifying conscious- results of our actions in relation to them. ness is one from the early days of artifi- A unifying consciousness is needed to learn cial intelligence: a program called Hacker anything new that will change the substance (Sussman, 1975), which would learn skills of of self. building structures with (virtual) children’s Dennett (1992) compares a storytelling building blocks. (The name “Hacker” comes self with a character whose first words, in from a time that seems now almost prehis- Melville’s Moby Dick, are “Call me Ishmael.” toric when it meant something like “com- But says, Dennett, Ishmael is no more the puter nerd.”) When it makes mistakes, the author of that novel than we are the authors P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

narrative modes of consciousness and selfhood 383

of stories we tell about ourselves. Dennett program had. A narrative processor – based develops the idea further by imagining a on goals that can be represented explicitly robot, Gilbert: and that can generate plans of action – is exactly the kind of representation needed by “Call me Gilbert,” it says. What follows is an autonomous self, able to explain action in the apparent autobiography of this fictional terms of goals and other reasons. Gilbert. Now Gilbert is a fictional, created 1934 self, but its creator is no self. Of course there Claparede’s` ( ) law of awareness is were human designers who designed the that people become conscious of an action machine, but they did not design Gilbert. when it is disrupted. Then consciousness is Gilbert is a product of a process in which flooded with emotion. Why should this be? there are no selves at all . . . the robot’s It is because emotions occur with the unex- brain, the robot’s computer, really knows pected, when what is familiar, what is habit- nothing about the world; it is not a self. ual, what is well practiced, no longer works. It’s just a clanky computer. It doesn’t know Negative emotions occur, as Peterson (1999) what it’s doing. It doesn’t even know that has proposed, when an anomaly occurs it’s creating this fictional character. (The or, in narrative terms, when a vicissitude same is just as true of your brain: it doesn’t arises. know what it’s doing either.)(pp. 107–108). Emotions are central to narrative because Velleman (2002) has contested these points. they are the principal processes in which Imagine Gilbert getting locked in a closet by selfhood is constructed. One falls in love, mistake. It has to call for help to be released. and the self expands to include the new per- The narratizing robot could then give an son. One is thwarted, and one forms venge- account of these matters. Velleman takes ful plans to overcome the purposes of the two important steps. In the first, he argues antagonist. One suffers a severe loss, and that such a robot as Gilbert would have one’s consciousness searches to find what subroutines that enabled him to avoid and in one’s theory of self was mistaken, per- escape from danger. Being locked in a closet haps perseverating in denial and blaming would be such a danger, and if it occurred, others or perhaps recognizing what should subroutines would be activated. Let us sup- be changed in oneself. Consciousness of this pose these subroutines were labeled “fear.” urgent problem-solving kind functions in the Balleine and Dickinson (1998) argue that kind of way that Sussman has proposed. If access to such affective states is a core pro- we are to learn from mistakes and change cess in identity. implicit theories in which we are lodged, a Velleman says that if Gilbert were to unifying consciousness is necessary. make an attribution to such a state, his narra- The second step in Velleman’s argument tive autobiography might include the state- is that Gilbert’s very abilities at telling sto- ment: “I’m locked in the closet and I’m ries also allow him to plan actions. Suppose starting to get frightened” (Velleman, p. 7). Gilbert works in a university: He does use- If the fear module recommended breaking ful things to achieve goals that he is given by down the door, then in his autobiograph- members of the faculty. He might have been ical narrative he might say, “I broke down given the goal of going to library to fetch the door, because I was frightened of being Dennett’s book, Consciousness Explained.If locked in for the whole weekend.” Gilbert at the same time his batteries had started thereby would have taken a step to being an to run down, Gilbert might notice this autonomous agent. He would be acting for a internal state and plan to go to the closet reason (his fear) and in a way that was intel- to get some new batteries. He might say, ligible to other people (us). Notice too, that, “I’m going into the closet.” Then, argues to act for a reason, the robot must be not Velleman, balanced between the possibili- just responsive to stimuli (internal or exter- ties of the library and the closet, “the robot nal) that prompt it to do this or that. It must now goes into the closet partly because of have a representation of goal states (such as having said so” (2002,p.6; see also Velle- avoiding danger) of the kind that Sussman’s man, 2000). P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

384 the cambridge handbook of consciousness

Narrative is based on plans that flow that.” This form of consciousness rests on from intentions. A computer sophisticated an awareness of self and other. It is in this enough to construct its autobiography could form that the possibility arises of being able run the narrative processor either forward to know what other people are thinking. in planning mode or backward in autobio- Mead describes how children at about the graphical mode. It would, in other words, age of 4 years play games that require them achieve what I have called the Vygotskyan to take roles and to experiment with chang- state of consciousness, of being able to direct ing these roles. In hide-and-seek, you can’t its own actions. Gilbert is able to act not only have much fun hiding unless you imagine in terms of reasons such as internal states but seekers looking for you. Developmental psy- also – running the narrative/planning proces- chologists have discovered that, about the sor forward – to plan to act in coherence with age of 4, children develop a theory of mind the story he is telling about himself. (Astington, Harris, & Olson, 1988). They Velleman’s idea, here, is that with begin to understand that other people might coherence-making processes a central orga- know quite different things than they do. nizing agency comes into being. Oatley The hider has a representation of herself; she (1992b) has argued that narrative is a mode knows who and where she is. She knows, too, by which we humans give meaning to life that the seeker does know who she is but events beyond our control, to our human does not know where she is. limitations, to mortality. It would be an illu- Once a person has reached this stage of sion to think that all such matters can be consciousness, the idea of self as narrator made meaningful, but we become human in comes fully into its own. Not many sto- the ways that we value by making some such ries are autobiographies. The fully devel- matters so, and these we share with others oped narrator does not simply offer output in a meaning-drenched community. We are of some autobiographical module. Nor does meaning-making beings who, by means of he or she only use the narrating module to narrative, make some aspects of disorderly instruct him- or herself. Such a narrator has reality comprehensible, and in some ways a theory of other minds and a theory of his tractable. Coherence in the computer gener- or her own mind persisting through time. ation of stories has been achieved by Turner Such a narrator is able thereby to act in a (1994), and Oatley (1999) has argued that world with other beings constituted in a sim- coherence is a principal criterion of success- ilar way. Saying “I will do x,” counts as a ful narrative. By adhering to this criterion, commitment to other people to do x, and then, the narratizing agency becomes capa- other people organize their actions around ble of influencing action and creating a self this commitment. So if I were to extend the who, in retrospect, could achieve an auto- story of Gilbert begun by Dennett, and con- biography that has some meaningful coher- tinued by Velleman, to include interaction ence. Art stops merely imitating life. Life with others, it would stop being autobiogra- starts to imitate art. The Woolfian storyteller, phy. It would become more like an episode aware of inner processes, becomes Vygot- in a story, something like this: skyan. Coherent with the story in which she is the principal character, Mrs Dalloway Gilbert was nearing the end of the day, but instructs herself, “I’ll buy the flowers myself.” like all good robots he was conscientious. Having said so, she does. He was about to set off to the library to fetch Consciousness Explained, when he noticed his batteries were low. He felt anx- Meadean Consciousness ious. If there were too long a wait at the The fourth aspect of consciousness is social. library elevator, he might not make it back As described by George Herbert Mead to the Department. (1913), it is a consciousness of voices in “I’ll go to the library to get the Dennett book debate: “If I were to say this, she might say in a few minutes,” he said to Keith, “First P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

narrative modes of consciousness and selfhood 385

I have to go to the closet for some new bat- Helmholtzian, Woolfian, Vygotskyan, and teries.” Meadean) and their role in the construction of narratives, let me turn to the crux Keith was staring at his computer screen. “OK,” he said, without paying much atten- of why consciousness is a narratizing pro- tion. cess that must be based in folk theory. Selves are social. In our species we as individuals The phone rang. It was a teacher at can’t manage much on our own. Together his daughter’s school. His daughter had we have built the communities in which we had an accident while playing basketball. live and all the accoutrements of the mod- Could he come immediately? Keith had a ern world. Our accomplishments depend sudden image of his daughter lying on the floor, immobile. He quickly put his stuff on culture, on being able to propose and in his book bag, shut down his computer, carry out mutual plans in which people locked the door to the closet, and ran down share objectives and do their parts to ensure the hallway. they are accomplished, and then converse in a narrative way about outcomes. Oatley Gilbert heard the closet door being locked. and Larocque (1995) have found that peo- In the middle of changing his batteries, ple typically initiate something of the order he couldn’t move and didn’t have enough power to call out. of ten explicitly negotiated new joint plans each day. We (Oatley & Larocque, 1995; “Carbon-based life forms,” he thought. Grazzani-Gavazzi & Oatley, 1999) had people keep diaries of joint plans that went In his writings on consciousness, Dennett wrong. Errors occurred in only about 1 in has worried about whether a conscious self 20 new joint plans. People succeed in meet- might really be autonomous. Perhaps he ing friends for coffee, succeed in working should set aside his worries. The self is much together on projects, succeed in fulfilling more. It is a model of self-with-others: first contractual arrangements. If it were not for the self-with-parent or other caregiver, then an interface of folk theory that enables each self-with-friends and suchlike others, then of us to discuss and adopt goals, exchange self with the generalized other of a certain relevant beliefs about the world, and elab- society, then perhaps self with significant orate parts of plans to ensure mutual goals other, and so forth, each with a basis of emo- are reached (Power, 1979), joint plans could tions. Self is not just a kernel of autonomy. not be constructed. And if we were not Self is self-in-relation-to-other, an amalgam able to narrate them to each other after- of the implicit theories we inhabit, suffused ward, we could scarcely understand each with the emotions that prompt our lives and other. If those psychologists, neuroscientists, are generative of our actions in the social and philosophers were correct who main- world. The very processes that allow us to tain that beliefs, desires, and emotions are make inferences about others’ minds, about figments of a radically false folk theory, we what another person may be thinking and would not find that 19 out of 20 joint plans feeling, that allow us to give a coherent were accomplished. The number would be account of that person as an autonomous zero. And when things went wrong, con- agent, are the same as those that enable us to fabulating narrators could never know why. form models of our own selfhood, goals, and Humans would be able to do certain things identity and to project ideas of our ourselves together, but we would perhaps live in much into the future. the same way as do the chimpanzees. When joint plans go wrong, errors occur for such reasons as memory lapses or because Narrative and Mutuality a role in a joint plan has been specified inexactly. When such a thing occurs the Having reviewed the aspects of con- participants experience a strong conscious- sciousness that I proposed (Oatley, 1988; ness of emotion, most frequently anger but P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

386 the cambridge handbook of consciousness

sometimes anxiety or sadness. People typ- injunction written long ago in the temple ically give narrative accounts of their own at Delphi: “Know thyself.” In doing so, we intentions and often of the failure of the each form a mental model of our goals, other. When a person who has not showed inner resources, limitations, emotions, com- up for a meeting is a loved one, vivid fears of mitments, and values that is no doubt inac- accidental death may occur; but if there is a curate, but that is somewhat serviceable. In failure of a plan made with someone who is the course of life, we come to know how far not an intimate, a mental model is elaborated and in what ways we can rely on ourselves. of that person as unreliable and untrustwor- With such knowledge we can then also be thy. Here is a piece of narrative recorded in dependable for others. a joint error diary (from Oatley & Larocque, Humphrey (1976, 2002) has argued that 1995): a principal function of introspective consciousness of ourselves is the understanding My co-worker was measuring some circum- of others as having selfhood and attributes ferences of pipes, converting them to diam- eters and reporting them to me. I recorded that are similar to own. By using our the ﬁgures and used them to drill holes later. model of ourselves to imagine what oth- The drilled holes were incorrect for diame- ers are thinking and feeling, we become ters. It could have been the conversion or what Humphrey calls “natural psycholo- measurement. I had to modify the holes. gists.” Although this method is not very accurate, it equips us far better to under- Continuing the story, our participant take social interaction than if we were to rely elaborated a mental model of his co-worker solely on behavioral methods. By elaborating and formulated a plan that was coherent mental models of others based on our own with what had happened, in a way that introspection, on our experience with them depended on his analysis: “My co-worker is in joint plans, and on what they and oth- not as careful about numbers as I am – maybe ers say in conscious narrative accounts about I should do this kind of task with someone them, we come to know how far and in what else.” Further plans were elaborated – con- ways we can rely on them in the joint plans tinuing the narrative in the forward direc- of our extended relationships. tion – concerning the relationship with the Here, then, I believe is the principal rea- co-worker: “I need to and want to do some- son for thinking narrative consciousness to thing about this kind of thing with him.” be functional. It is a reason that Dennett neglects because he, like most of us West- Personal and Interpersonal Functions erners who work in the brain and behavioral of Consciousness sciences, tends to think in terms of individ- The novelist constructs characters as vir- ual selves, individual minds, and individual tual people who typically have an ensemble brains. But if our species is predominantly of emotion-based goals (intentions), from social and depends for its being on mutual- which ﬂow plans of interaction (Oatley, ity and joint planning, we need to consider 2002). In daily life, we each construct our also such interfaces as the interface of lan- actions (actions, not just behavior) in a sim- guage along with its conscious access to what ilar way, in the light of what is possible, by we take to be our goals and plans, by which coherence with a mental model of our goals, we arrange our lives with others. The lan- of our resources, and of our commitments. guage is not just of describing, in the fash- The person (character) whom each of us ion of post-hoc autobiography, as Dennett constructs, improvising as we go along, is not suggests. It is based in what one might call virtual, but embodied. This person accom- a language of action (mental models, goals, plishes things in the world and interacts with plans, outcomes), of speech acts to others others whom we assume are constituted in (see, e.g., Searle, 1969) as we establish mutu- a way that is much like our self. ality, create joint plans, and subsequently The functions of consciousness include analyze outcomes in shared narrative terms. an ability to follow, to some extent, the The language is also one of explanation in P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

narrative modes of consciousness and selfhood 387

terms of agents’ intentions and the vicissi- in eastern Africa to use sticks as tools to tudes they meet. Narrative is an expression poke into termite mounds to fish out ter- of this language. mites, which they eat (McGrew, 1992). A stick left near a termite mound by a mother chimpanzee may suggest to her daughter The Evolution of Narrative that she pick it up and poke it into the mound. What chimpanzees don’t do – can’t do according to Tomasello (1999) – is for For the period up to about 30,000 years ago, one (say the mother) to show another (say the palaeontological record of hominid evo- her daughter) how by using a stick she can lution is largely of fossilized bone fragments fish for termites. The daughter, although and stone tools, with nothing in writing until she may see from the mother’s activity that 5,000 years ago. Judicious comparisons with there is something interesting to look at, and evidence of our living primate relatives have although she may work out herself how to allowed a number of important inferences. fish for termites with a stick, does not understand that her mother is intentionally using Intentions and Consciousness a tool. She does not thereby see quickly of Intentions how to use it herself. Chimpanzees thus lack Narrative is about intentions. Its sequences an essential step for forming true cultures, involve chains of both human and physical in which useful technical innovations are causation (Trabasso & van den Broek, 1985). preserved and passed on, in a ratchet-like A surprising finding from primatology is that process. our closest relatives, chimpanzees and bono- In an experiment by Povinelli and O’Neill bos, have only a very limited understand- (2000), two of a group of seven chimpanzees ing of causality in both the physical and the were each trained separately by the usual social world. Although they certainly have techniques of reinforcement-based learning intentions, they do not seem conscious that to pull on a rope to bring toward them a they have them. Although they are good at weighted box with fruit on it. When they interacting with others, they do not seem had become proficient at this task, their conscious, in the way that we are, that other training changed. The box was made heav- primates have intentions. These conclusions ier so that one chimpanzee could not move have been reached independently by two it. Now, both chimpanzees were trained by research groups who have conducted exten- reinforcement techniques to work together sive series of experiments with chimpanzees: to pull the box and retrieve the food. All Tomasello (1999) and Povinelli (2000). seven of the chimpanzees knew each other Chimpanzees are successful instrumen- well, and all had previously taken part in tally in the wild, and they are very social. other experiments that involved lone pulling Their brains generate intentions, which they on ropes to retrieve food. Here is the ques- carry out. If they were equipped with post- tion. If, now, one who had not been taught to hoc autobiography constructors they would pull the rope in this apparatus were paired be the creatures whom Dennett describes. with a chimpanzee who was experienced They seem to have no autonomous selves. in cooperative pulling, would the experi- Although they solve problems and even use enced one show the na¨ıve one what to do? primitive tools such as sticks and leaves, their The answer was no. The experienced chim- lack of any sense of their own intentions, or panzee would pick up its rope, perhaps pull, of plans mediated by tools, makes them inca- and wait for a bit, looking over toward the pable of instructing other animals in a new other. One of the five na¨ıve chimpanzees, technique or, in the wild at least, of receiv- Megan, discovered how to pull the rope ing technical instruction. The occurrence of independently of her experienced partner. sticks near termite mounds, together with She thereby managed to work with both chimpanzees’ interest in food and in manip- of the experienced animals to retrieve the ulating objects, had led some groups of them heavy box with the food on it. But for the P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

388 the cambridge handbook of consciousness

most part the experienced and the other stand a bit about intention (in terms of na¨ıve partners failed in the joint task. On no mirror neurons) and that chimpanzees half- occasion did either of the experienced chim- understand it in a more explicit way. They panzees make a gesture or attempt to direct still, however, lack of grasp of full mutual the na¨ıve one’s attention to relevant features intention, which is the centre of narrative of the task. On no occasion did the expe- consciousness. rienced animal pick up and offer the rope Although they intend and feel, monkeys to the na¨ıve one. The experienced chim- and apes do not fully know that they and panzee did not seem to infer that the na¨ıve their fellows are beings who can intend chimpanzee lacked the proper intentions or and feel. As John Donne put it: “The beast knowledge in the task. does but know, but the man knows that he Another striking ﬁnding is that neurons knows” (1615–1631/1960,p.225). This seem- have been discovered in the premotor area ingly small extra step has made a great dif- of monkeys’ brains that respond both when ference. It is an essential step to culture, to a human hand is seen picking up a raisin and narrative consciousness, and to selfhood. when the monkey itself reaches intentionally to pick up a raisin. Rizzolatti, Fogasse, and Mental Models of Others Gallese (2001) call them “mirror neurons.” They do not respond to a human hand when In her ethological work with chimpanzees it is moving without any intention to pick in the wild, Jane Goodall (1986) learned to up the raisin. They do not respond when a recognize each one as an individual. Chim- human hand picks up the raisin with a pair panzees recognize each other as individuals, of pliers. Rizzolatti et al. argue that the brain and this is the basis for their elaborate social recognizes action not in terms of a purely life. They know who the alpha animal is, who visual analysis, but in a process of analysis everyone’s friends and allies are. Because by synthesis, by means of its own motor chimpanzees are promiscuous in their mat- programs of carrying out the action. Riz- ing, no one knows who the father is of any zolatti and Arbib (1998) have further sug- youngster, but everyone knows other aspects gested that this discovery is a preadapta- of kinship. Only when Goodall had started tion for learning language based on a case to recognize individuals did the social lives grammar (Fillmore, 1968) around verbs of and actions of the chimpanzees start to make intention. Gallese, Keysers, and Rizzolatti sense. (2004) have proposed that mirror neurons Chimpanzees know everyone in their enable simulation of intentions and emo- social group, which can reach a maximum tions of other individuals in the social world, size of about 50 individuals. Each one forms and hence afford a neural basis of social a mental model of each other one, which cognition. includes something of that one’s history, The issue remains mysterious, as indi- habits, and allies. Dunbar (1993, 2003) and cated by recent experiments by Tomasello Aiello and Dunbar (1993) have found that and his group. For instance Behne et al. the size of the brain in primate species has a (2005) found that 18-month-old human close correlation with the maximum size of infants, but not 6-month-olds, could tell the its social group. difference between an experimenter who Each chimpanzee spends some 20%ofits was able but unwilling to give them a toy time sitting with others, one at a time, and (teasing) and one who was willing but unable grooming: taking turns in sorting through (e.g. because the toy dropped out of reach). their fur, removing twigs and insects. It is Call et al. (2004) found that chimpanzees a relaxed activity. It is the way primates could also make this same distinction, thus sustain affectionate friendships with each demonstrating that they know more about other. Dunbar (1996) shows a graph in which others’ intentions than had previously been the data points are separate hominid species thought. We might say that monkeys under- (australopithecus, homo erectus, etc.) with an P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

narrative modes of consciousness and selfhood 389

x-axis of time over the past 3 million years to Panksepp’s conjecture, the animal’s dis- and a y-axis of amount of grooming time tress is much the same as a human infant required according to the species’ brain size feels when separated from its mother. It is and the inferred size of its social group. also the core of the distress that we might Humans maintain individual mental mod- feel as adults if we had arranged to meet our els of about 150 others. If we used the same partner, and after a half-hour he or she did procedures as chimpanzees, we would have not show up, and we start to worry that he to spend about 40% of our time groom- or she had suffered an accident. Notice that ing to maintain our friendships. As group the consciousness involved is vividly salient. size increased, a threshold was reached, Its concern is interpersonal. according to Dunbar, between 250,000 and Consider another emotion: interpersonal 500,000 years ago, of about 30% of time fear. De Waal (1982) describes how, in the spent grooming. This is the maximum any group of 25 or so chimpanzees who lived in a primate could afford and still have time for park-like enclosure at Arnheim Zoo, Yeroen foraging and sleeping. It was at this point, was an alpha male until he was deposed by argues Dunbar, that language emerged as the then-beta male, Luit. As alpha, Yeroen conversation: a kind of verbal grooming. received between 75% and 90% of the rit- Conversation is something one can do ual submissive greetings made by individ- while gathering food or performing other uals in the troop. This submissive greeting tasks, and one can do it with several oth- is well marked in chimpanzees. Typically ers. What do we human beings talk about it includes bowing, making short panting in friendly conversation? Dunbar (1996) has grunts, sometimes making offerings such as found that, in a university refectory, about a leaf or stick, or sometimes giving a kiss on 70% of talk is about the doings of ourselves the feet or neck. In the early summer of 1976 and others: conversation, including gossip, Luit stopped making this kind of greeting to most typically the elaboration of mental Yeroen. On 12 June, he mated with a female models of self and others and of people’s just 10 metres from Yeroen, something of goals in the social group. Dunbar might have which Yeroen was normally extremely intol- added that the incidents about which people erant. On this occasion Yeroen averted his talk are recounted in narrative form. eyes. Later that afternoon, Luit made angry aggressive displays toward Yeroen. De Waal said he thought at first that Yeroen was ill. Consciousness and Emotions But this was not so. Only later did he realize Before conversation and narrative conscious- that these were the first moves of a take- ness could emerge, preadapations were nec- over that took 21/2 months to accomplish essary. I have mentioned one suggested by and required much interindividual manoeu- the findings of Rizzolatti and his colleagues. vering. It involved Luit inducing the adult In this subsection and in the next I sketch females in the group to abandon their alle- two more preadaptations, the first of which giance to Yeroen and enter an alliance with concerns emotions. Panksepp (1998, 2001, him. The take-over was completed on day 72 2005) has argued that the most parsimo- of the sequence when Yeroen made his first nious explanation of a range of data is that submissive greeting to Luit. When de Waal emotion is the basic form of consciousness. describes the events of this dominance take- He calls it primary process affective con- over, he recounts a narrative that imposes sciousness. We share it with other mam- meaning on the events for us humans. It is mals, and it is subserved by a homologous not, of course, a narrative that the chim- region of the brain, the limbic system, in panzees could construct. De Waal says that different mammalian species. Thus when a 12 June was the first time he ever saw Yeroen baby mammal is separated from its mother, scream and yelp and the first time he saw it utters distress calls. This behavior is gener- him seek support and reassurance. We may ated in a specific limbic area, and according imagine that on that day, Luit felt angry P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

390 the cambridge handbook of consciousness

toward Yeroen – his hair stood on end, and by disruptions of action or expectancy (vicis- though he had normally looked small, he situdes), and they can completely fill con- now looked the same size as the alpha male sciousness. Fourth, only with the coming Yeroen. We may imagine too, that this was of language (or perhaps the concepts of the first time Yeroen felt afraid of him. The prelanguage) do the interindividual scripts Panksepp conjecture is that Luit did indeed or frames of emotions become the bases feel angry and Yeroen did indeed feel afraid of episodes in stories. At this point, a new in the way that we humans would when our kind of consciousness can emerge that sees position was threatened. itself and others as instigators of intended What we may say about this in terms of actions and as experiencers of the emotions consciousness is that, for the chimpanzees that result from these actions. Oatley and in this group, events and emotions would Mar (2005) have argued that social cogni- have unfolded in a sequence of present- tion is based on narrative-like simulations, tense happenings. Among them only what the conclusions of which can (as Helmholtz I have called the interaction-type elements insisted) become conscious. The elaborated of narrative would be present. My conjec- story-type consciousness, in which agents ture is that immediate emotions – anger, fear, act for reasons and are responsible for their friendly alliance, deference – conferred the actions, is built on preadapted bases (includ- structure on each episode of interaction, but ing those of mirror neurons and emotion- the whole sequence would not have, could based relating) that were in place in hominid not have, a plot of the kind one expects in ancestors. a story (either prospectively or retrospectively) as far as the animals were concerned. Mimesis, Metaphor, and Imagination Among humans, however, sequences of such events not only lend themselves to Homo erectus emerged about 1.9 million narratization but we are also unable to years ago, and was our first ancestor to look avoid turning them into story form (Oat- more humanlike than apelike. Although ley & Yuill, 1985). At some time since the simple stone tools pre-existed this species, line that would lead to humans split from it was with these beings that elaboration of that which would lead to the chimpanzees such tools began. With them, also, came the some 6 million years ago, the perception of first strong evidence for the importance of interaction-type elements of narrative added meat in the diet, and perhaps even according the elements of story-type narratization. to Wrangham (2001), the control of fire and Thus when de Waal (1982) describes the hence cooking. With them came a second events of Luit taking over the alpha position important preadaptation to language (fol- from Yeroen in 1976, he can only do so in lowing that of consciousness of emotions). terms of a humanly recognizable story told, Donald (1991) calls it mimesis: a non-verbal inevitably, by a narrator, a kind of novelist representation of action and an ability to not of the self as Dennett has postulated, reproduce actions. It enabled fundamental but of others. cognitive developments of cultural forms of My hypothesis, then, is that first, as group enactment, some of which are still Panksepp has proposed, emotion is a primary with us, such as dance and ritual. form of Helmholtzian consciousness (with Donald says that the next important evo- sensory awareness being another). Second, lutionary transformation did involve lan- emotions are frames or scripts for interindi- guage. It was to myth, which I take to derive vidual relationships. Anger (such as Luit’s from an early form of narrative conscious- when he displayed aggressively to Yeroen) ness. He dates myth as far back as the time sets up a script for conflict. Emotions struc- that Dunbar has postulated for the emer- ture relationships so that sequences of inter- gence of conversation, but his emphasis is action are prompted. Third, as Paulhan on narrative aspects. Myth, says Donald, is (1887/1930) proposed, emotions are caused the preliterate verbal way of understanding P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

narrative modes of consciousness and selfhood 391

how the world works. It can pervade every ments were related. Rather than simply hav- aspect of people’s lives. For instance among ing separate domain-specific areas of knowl- the !Kung of the Kalahari (Lee, 1984), ill- edge – social knowledge of the group, tech- nesses are typically seen as caused by peo- nical knowledge of how to make stone tools, ple who died a short time before (parents natural historical knowledge of plants and and grandparents). While they were alive animals – the people of those times began these people were good, but many say that to relate one kind of knowledge to another. once dead they generally became malevo- Here began imagination and metaphor: A lent (notice the narrative structure). Others this (in one domain) is a that (in another say they are harmful only when the living domain). This charcoal mark on the wall of don’t behave properly. Living people are by a cave is a rhinoceros, this person who is no means powerless, however. The means dead lives on in another kind of existence, of maintaining health, and healing the sick, this animal bone can be shaped to become have to do with interpreting the interven- a harpoon tip. During this period, argues tions of these spirits and sometimes combat- Mithen, human culture began to accelerate. ing them. So, says Donald, Imagination is the type of consciousness that makes narrative possible. It offers us stories Myth is the authoritative version, the of possibility and of people not currently debated, disputed, filtered product of gener- present in situations that are not directly ations of narrative interchange about real- visible. ity . . . the inevitable outcome of narrative skill and the supreme organizing force in Upper Paleolithic society (p. 258). The Developmental Psychology Myth is a narrative form of social exchange, of Consciousness used for thinking and arguing about how the world works. It is active today, and not Although ontogeny may not exactly recapit- just in the Kalahari. “We are on the side ulate phylogeny, there are parallels between of good, dedicated to the fight against evil,” the rise of narrative consciousness during for instance, and “You can do whatever you hominid evolution and the development of really want to do,” are mythic sentiments narrative skills and consciousness in chil- of some vitality in North America. (Notice, dren. Zelazo and Sommerville (2001) and again, the narrative elements of goals and Zelazo (2004) have proposed a set of levels actions.) Myths have pragmatically impor- of consciousness, reached progressively dur- tant properties. By casting a matter into a ing the preschool years (see Chapter 15). The symbolic form they make it potentially con- earliest, which they call minimal conscious- scious and an object of cultural considera- ness, corresponds to Helmholtzian con- tion. It becomes a potent means of organiz- sciousness (described above). Subsequent ing both individual and societal behavior. levels include recursive representations. The It was between 50,000 and 30,000 years second level includes minimal consciousness ago that the first art began to appear of a thing plus the name of that thing. Higher in the human record. Caves in south- levels include further recursions. So reflec- east France contain paintings of animals on tive consciousness of the kinds I have called their walls from 31,000 years ago (Chauvet, Woolfian and Vygotskyan (discussed above) Deschamps, & Hillaire, 1996). Several cul- include not just actions but also the con- tural developments, in addition to paintings sciousness of self-in-action. Finally a level and the production of ornamental artifacts, is reached at about the age of 4 years that occurred around the same time. One was includes the social attributes of theory of treating the dead in a special way, burying mind (Astington, Harris, & Olson, 1988)in them with ceremony. Another was a sudden which one can know what another knows, proliferation in the types of tools. Mithen even when the other person’s knowledge is (1996) has proposed that these develop- different from what one knows oneself. It P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

392 the cambridge handbook of consciousness

corresponds to Meadean consciousness (dis- grammar, as postulated by Fillmore (1968), cussed above), in which the social self can for which the preadaptation suggested by be explicitly represented and thought about. Rizzolatti and Arbib (1998) is the basis. Tomasello and Rakoczy (2003), however, As Tomasello points out, the child sym- argue that at a much earlier stage, occur- bolizes exactly the kind of intention that ring about the age of 1 year, children come to chimpanzees enact, but do not know that understand themselves and others as inten- they know. The symbolization can then be tional agents. They call this the ”real thing”: used as a tool to affect the attention of It is what separates us from other animals. the person to whom the communication They argue that it enables skills of cultural is made. Infant communications are some- learning and shared intentions to occur and times requests, like “more,” or even “more that this stage must precede any acquisition juice,” but conversational priorities early of theory of mind that would need to be become evident: Parents and children draw based recursively on this ability. each others’ attention, in conversation, to Recursion also occurs in a way that things in the world. extends the scheme proposed by Zelazo. Then in development comes a phase that, Dunbar (2004) points out that from school on palaeontological grounds, would be unex- age onward a person who takes part in a con- pected: the appearance of monologue. Chil- versation or who tells a story must know dren start talking out loud to themselves. (level 1) that a listener can know (level 2) Nelson (1989) arranged for the parents of a what a person depicted in the conversa- small child, Emily, to place a tape recorder tion or story knows (level 3). Dunbar argues by her bed before she went to sleep. After that skilled storytellers can work with about some bedtime conversation with a parent, ﬁve recursive levels of consciousness. Thus the parent would leave and Emily would in Othello, Shakespeare writes (level 1)so often enter into monologues that, with the that audience members know (level 2) that help of her mother, have been transcribed. Iago contrives (3) that Othello believes (4) Nelson (1996) has argued that narrative that Desdemona is in love with (5) Cassio. binds memories of autobiographical events All this, argues Dunbar, depends on neural together in the meaningful form that we machinery present in human brains that is think of as selfhood. Here is an example not present in chimpanzee brains. from Emily aged 21 months: “The broke. Car Recursion is an idea that became impor- broke, the . . . Emmy can’t go in the car. Go in tant in cognitive science in the 1960s. It is green car” (Nelson, 1989,p.64). This mono- that representations may include representa- logue illustrates verb island constructions of tions of themselves. Productively, then, the the kind identiﬁed by Tomasello around the idea of successive steps of recursion has been verbs “break” and “go.” It also offers an agent, proposed by Zelazo as successive levels of “Emmy,” and the narrative structuring of an consciousness, achieved in successive stages autobiographical event. The car was broken, of development. Although each level of con- and therefore Emmy had to go in a different sciousness depends on a previous one, and car. Here, already, we see some of the ele- each emerges at a certain stage during devel- ments of story-type narration: the telling of a opment, the earlier level it is not superceded. story by a narrator, a self (character) persist- It continues to be available. ing through time, a vicissitude, and the over- Tomasello’s (1999) studies of the devel- all possibility of making meaningful sense of opment of language indicate that soon after events (although not yet with a theory of single words (at Zelazo’s second level) come other minds). verb islands, in which different words can From the age of 2 or so, children are able be put into slots for agent of the action, to run the narrative process not just back- for object, and for outcome. Thus in the ward to link memories with an agent but verb island of “throw” we get “Sam throws also forward. Here is Emily in another mono- the ball.” This is close to the idea of case logue at the age of 2 years and 8 months. P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

narrative modes of consciousness and selfhood 393

Tomorrow when we wake up from bed, first self is experienced as unified, not merely me and Daddy and Mommy, you . . . eat as a disparate bundle of reflexes. From psy- breakfast eat breakfast, like we usually choanalytic thought, the somewhat ironic do and then we’re going to p-l-a-y, and but nonetheless suggestive idea comes that then soon as Daddy comes, Carl’s going the moment of this realization occurs as a to come over, and then we’re going to play child can recognize him- or herself in a mir- a little while. And then Carl and Emily ror (Lacan, 1949/1977). Other, more empir- are both going down the car with somebody, and we’re going to ride to nursery ically minded psychologists have taken the school . . . (Nelson, 1989,pp.68–69). infant’s ability to touch a patch of rouge on his or her forehead, when the image of the Here, again, is a recognizable self, “Emily,” infant bearing the patch of rouge is seen by who is a protagonist in a plan-like account him- or herself in the mirror, as an indica- with many of the attributes of a narrative tion of the dawning of selfhood (Lewis et al., structure in story form. 1989). Nelson’s proposal is that, with the Fivush (2001), in pursuit of similar ques- dawning of narrative consciousness, the self tions, has studied children discussing pieces is experienced in a world of widening possi- of autobiographical narrative with their par- bilities, of other people and of other minds. ents and other adults. Children as young as Zelazo’s (2004) proposal is that this ability 3 years make such discussion an important to represent the self explicitly is fundamen- part of their social activity. They use it to tal to the development of consciousness and evaluate experience. For instance, a child of a unified as opposed to a fragmented con- of 3 years and 4 months said to an inter- trol of action. viewer, “There was too much music, and Harris (2000) has argued that a major they play lots of music, and when the circus developmental accomplishment for children is over we went to get some food at the food occurs when they start to construct, in their place” (p. 40). Part of the autobiographical imagination, things that are not immedi- point is to assert selfhood and subjectivity via ately present. They make such constructions things that the narrator liked or didn’t like not just in stories but also in pretend role (“too much music”). By the age of 4 children play (discussed above). They start imagin- beome aware, in a further step to individual- ing what others might know (other minds). ity, that their experience might be different The idea of specifically narrative conscious- from that of others. ness has been extended in the recent work Nelson’s narratives from the crib suggest of developmental psychologists on the ques- how the inner chattering that is a feature tion of how children conceptualize a self of everyday Woolfian consciousness devel- as persisting through time. Barresi (2001) ops from spoken monologue. This idea is and Moore and Macgillivray (2004)have strengthened by Baddeley’s (1993) account shown that the imagination by which one of PET scanning of his own brain when he can anticipate future experiences of the was engaged in quiet inner speech. It showed self (which they call prudence) is likely to involvement of those brain areas that are involve the same processes as those in which typically involved both in speech production one can become interested in others (proso- and speech understanding. cial behavior). This same ability of the imag- Nelson (2003) concludes that conscious- ination is the central core of understand- ness is one of the functions that develops ing narrative. Arguably stories nurture it. as an emergent property with the devel- As Vygotsky (1962) has proposed, language- opment of language, and most specifically based culture offers children resources that with the development of narrative language. they do not innately possess, but also do It is with this ability that a move occurs not have to invent for themselves. Chil- from simple Helmholtzian awareness of the dren in modern times aged 4 and 5 years here and now to the idea of a self. Several have reached the point of the imagination things are accomplished in this move. The that can take wing in the way that Woolf P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

394 the cambridge handbook of consciousness

depicted and that Mithen (2001) argued had visits to doctors’ offices subsequently. They been reached in evolution by our ancestors, also underwent improvements in immune 30,000 years ago, when one may suppose function. The argument is that traumatic with Donald (1991) that story-type accounts and stressful events that are not integrated of meaning-making had become routine. by the meaning-making processes of narra- Conscious selfhood, in these kinds of tive can impair health. Making such events accounts, is a meaning-making function that conscious and integrating them in story form depends on, but has properties beyond, has beneficial effects. those of the separate neural processes and modalities of the kind on which Dennett concentrates. Although in this new field, The Rise of Consciousness in Western there is debate about how best to charac- Imaginative Literature terize phases of child development, it seems that the systems pass through emergent Writers have been fascinated by conscious- stages of consciousness similar to those I ness. The great book on the subject in West- have postulated – Helmholzian, Woolfian, ern literature is by Auerbach (1953). Its Vygotskyan, and Meadean – or through 20 chapters span 3,000 years, from Genesis those that Zelazo and Sommerville write to Virginia Woolf. On the one hand West- of in terms of progressively recursive repre- ern literature involves the “representation sentations. By adulthood, consciousness can of reality.” On the other it offers chronicles often be of the minimal Helmholtzian per- at successive moments in a history of mind ceptual kind, or consciousness of an emo- turning round, recursively, to reflect upon tion, but it can also easily switch to narrative itself. Each chapter in Auerbach’s book starts consciousness of self and of self in relation with a quotation, ranging in length from a to others. paragraph to several pages, from a particular What developmental psychologists have writer in a particular time and society. For shown is that, with each level of emer- each, Auerbach analyzes the subject mat- gence of new abilities, there arrives a set ter, the words, and the syntax. Immediately of functions that these abilities subserve the reader is in the middle of a scene, with that are unavailable to creatures without knights in medieval times, or with Dante and these abilities. So, from the first movements Virgil descending into the Inferno, or in La of naming and drawing attention of others Mancha with Don Quixote, or in Paris with to things, infants enter into a world of a Proust’s narrator Marcel. In each there is a shared folk-theoretical understanding. They society with certain understandings. In each, become creatures of different kinds than any the characters inhabit a certain implicit the- wild-living chimpanzee. They have taken ory of consciousness. the first step toward constructing a narra- The opening sequences of the Bible, writ- tive consciousness of themselves and others ten by the Hebrews about 3,000 years ago, as actors in the world, who cause intended narrate the story of God’s creation of the effects and struggle with the vicissitudes world. When the first human beings, Adam of life. and Eve, enter the scene (attributed to the A kind of culmination to the develop- writer J., Genesis, Chapter 3), they eat of the ment of narrative by Nelson, Fivush, and fruit of the tree of knowledge of good and others has been offered by Pennebaker et al. evil, and they become ashamed. It is hard to (Pennebaker, Kiecolt-Glaser, & Glaser, 1988; avoid the interpretation that in the moment Pennebaker & Seagal, 1999). The basic find- of becoming conscious of good and evil they ing is that adults who wrote narratives for as became self-conscious. little as 20 minutes a day for 3 days on top- In the Greek tradition, The Iliad (Homer, ics that were emotionally important to them 850 bce/1987) was written from previously (vicissitudes), as compared with people who oral versions about the same time as Genesis. wrote about neutral subjects, made fewer Jaynes (1976) has made several provocative P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

narrative modes of consciousness and selfhood 395

proposals that further Auerbach’s idea that emnon. At that moment a goddess appears a different kind of consciousness from our to Achilles. No one else sees her. It is the god- own is found in Genesis and The Iliad. His dess Athene, who utters to Achilles the tribal theory is that consciousness is not only a injunction: “Do not kill the commander-in- narrative and unifying agency, the basis for chief.” The two chambers of mental gov- modern selfhood, but that in early writings ernment are here in conflict. The outcome: such as The Iliad one sees the fading of an Achilles obeyed Athene, but went into a older mentality. He argues that before about sulk. His refusal to fight resulted in the 3,000 years ago, human beings had what Greeks almost losing the Trojan War. In The he called bicameral minds. Bicameral means Iliad, Jaynes and a number of other scholars having two chambers, like two chambers of have noted that there is no word for mind. government, a senate and a house of rep- Emotions occur not as conscious preoccupa- resentatives. One chamber was of species- tions prompting actions, but as physiologi- typical behavior: mating, attacking if one is cal perturbations: thumos (livingness, which attacked, and so on. The second was of obe- may include agitation), phrenes (breathing, dience to the injunctions of a ruler. These which can become urgent), kradie (the injunctions tended to be heard in acoustic heart, which may pound), and etor (guts, form – do this, don’t do that – and they which may churn). In The Iliad the word allowed hierarchical societies to live in cities psuche (from which later derived such con- beyond the immediacy of face-to face concepts as soul and mind) was an insubstantial tact. One feature of this second chamber of presence that may persist after death. The government, argued Jaynes, was that when word noos did not mean mind, but was an a ruler died, his injunctions could still be organ of visual images. heard. The phenomena were referred to in For the preclassical Greeks, there were terms of gods. Each new ruler would take bodily agitations of thumos, phrenes, and so on the mantle of command and be trans- on, and there were voices of gods offering lated into the role of god. So it was not tribal injunctions but – according to Jaynes – that the figurines that archaeologists have no narrative self-consciousness. Plans were unearthed from Mesopotamia were statues not decided by mortals but by immortals. of gods. They were gods. When one visited In the opening sequence the question is their shrines in a house or public building asked by the narrator: Whose plan (boule) set and looked at a god, one could hear his or Achilles and Agamemnon to contend? The her words in one’s mind’s ear. answer: the plan of a god, Zeus. The read- Modern narratizing consciousness arose, ers of this chapter might sense something argued Jaynes, as people started to travel familiar. Here were people whose important beyond the rather simply governed city- actions were determined not by themselves states and began to encounter others, to but in some other way, by what they called take part in trade, and in wars. The sim- gods. The familiarity is that Dennett offers ple two-chamber mental government broke an echo of this same idea. For him the actions down. It was inadequate for dealing with of all of us are determined not by human people from different cultures. At first a few agents qua agents, but by something imper- and then more individuals started to enter- sonal: brain processes. tain thoughts and reasons that were nei- By the time of The Odyssey, something ther instinctive responses nor obedience to else was beginning to enter mental life. authority. They began to tell more elabo- It was human cunning. But still the con- rate stories about themselves and to reason scious, reflective mind had not emerged or in inner debate. at least had not emerged in its modern form. In the opening sequences of The Iliad It had, as Snell (1953/1982) put it, to be Achilles draws his sword in anger. He is res- invented. Its invention perhaps began 200 ponding instinctively to an insult from the years after Homer with the poet Sappho, Greek army’s commander-in-chief, Agam- as she began to reflect on the idea that her P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

396 the cambridge handbook of consciousness

continual falling in love with young women metaphor has not been the mirror held up had something repetitive about it. It was to nature, but the lamp that illuminates. Just brought more fully into being by Aeschy- as the young child by pointing and saying lus and Sophocles in their narrative plays “fire truck” is directing the attention of her about how we human beings are responsible or his companion, so dramatists and novel- for our actions, although we cannot foresee ists direct the attention of their readers. and do not necessarily intend all their conse- It is appropriate, perhaps, that the grow- quences. It reached fully modern form with ing elaboration of consciousness in the great Socrates in his teachings about how we can bourgeois European novels of the 19th cen- choose to think and decide how to act for tury should reach a kind of culmination with the good. Freud’s cases, in which the gaps in the sub- Not all scholars agree that large-scale ject’s consciousness of intentions are filled changes of the kind described by Jaynes and precisely by elaborating a story (see Freud, Snell did occur (see, e.g., Williams, 1993)in 1905/1979; Marcus, 1974; Oatley, 1990). a progression of consciousness from Home- In recent times, the most influential liter- ric to classical times. If they did, however, it ary theorist has been Bakhtin (1963/1984), seems likely that they were cultural changes, who has moved beyond analyses of sin- rather than the neurological ones postulated gle narrators, such as Robinson Crusoe. by Jaynes. With the fully developed novel, for which The beginnings of the modern world are Bakhtin takes Dostoyevsky as the paradig- set by many scholars (following Burckhardt, matic author, we do not so much listen to the 1860/2002) with Dante and the Renaissance. monological voice of an author or narrator. At this time the literary idea came into its We take part in something more like a con- own of character that included the concept versation. In such a novel, there are several that some of it is hidden: People not only act centres of consciousness, and if the author as they did in Homer, but reflect consciously does it right, one of these is the reader’s. on what kind of person they would be to act Bakhtin’s idea of the dialogical basis of in such and such a way. In the 20th century, social life as depicted in the novel is present literary consciousness took an inward turn. in practical affairs in the West in the elabo- Now we read to recognize the images and rate procedures by which justice is admin- thought-sequences of our own minds, of the istered in courts of law (Bruner, 2002). In kind that Woolf portrayed. a criminal trial, two narratives are related, Following Auerbach, Watt (1957) traced one by the prosecution and one by the some of the movements of consciousness defense. The critical narrative, however, is in English novels of the 18th century in constructed by others (judge and jury) who response to social and economic changes. have listened and supplied a narrative end- So Defoe’s (1719) Robinson Crusoe empha- ing in terms of one of a small number of sizes the individualism of the times, particu- prescribed outcomes of morality stories that larly in economic matters, as well as the trait have been told and retold in our society. that would be characteristic of the novel, Such completions are known as verdicts the reflective examination of the self and its (Pennington & Hastie, 1991). doings, which Watt traces to Puritanism. There are universals in the telling of Romanticism, the literary era that we stories. All cultures use narrative for pur- still inhabit, with its emphases on emotions poses that are somewhat didactic (Schank and on style, started around 1750. Although & Berman, 2002). That is to say, all sto- there has been postmodern debate about ries have attributes that Donald attributed whether language can represent anything to myths. They explain human relationships at all outside itself, Abrams (1953) makes to a problematic world, to the gods, to soci- clear that throughout the Romantic period ety, as well as to individual emotions and such representation has not really been the the self. Some kinds of stories are universal intention of literary writers. The dominant (Hogan, 2003), the love story for instance. P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

narrative modes of consciousness and selfhood 397

Its typical plot is of two lovers who long to ism that is thereby promoted. In Lodge’s be united. Their union is prevented, typi- (2001) novel, Thinks, the literary idea is rep- cally by a father. In the comic version, after resented in a talk given by the female pro- vicissitudes the lovers are united. They live tagonist, novelist Helen Reed, at a cognitive happily together, and the previously oppo- science conference on consciousness. She sitional father rejoices in the union. In the presents Marvell’s (1637–1678/1968) poem, tragic version, the union is prevented and “The Garden,” and ends her talk with the the lovers die, perhaps to unite on some non- stanza in which occur these lines: material plane. In Western literature, something changes My soul into the boughs does glide: There like a Bird, it sits, and sings . . . as we move from the mentality of Achilles (p. 50) and Abraham to that of Mrs Dalloway just after World War I or more recently to that We can smile at Descartes’ or Marvell’s pre- of Jacques Austerlitz, whose life was frac- scientific idea of the soul. Or we can see it tured by World War II (Sebald, 2001). Per- as a metaphor of what is most human in us: haps literature has changed in response Perhaps like a bird, we are able to take to the to changes in society. In the West, these wings of imagination toward what may yet changes have included a growing individu- be possible for our species. alism and a growing faculty of inwardness No matter how we conceptualize these and reflection. Perhaps, too, literary narrative issues, we might agree that stories depend has itself been partly responsible for some of on mental simulation, which in Victorian these changes: a workshop of the mind. We times was spoken of in terms of imagination can link Dennett to this series. To be thor- and of dreams. To simulate, make models, oughly Western, thoroughly modern (per- derive analogies, project metaphors, form haps postmodern) he says, we must give up theories . . . this is what mind does. The con- the idea of mind and accept that there is only cept of simulation is used in several senses in the brain. Meanwhile, however, others, such psychology. In one sense, the idea of simula- as Flanagan (2002), reject the dichotomy tion is used by Barsalou (2003) to concep- of mind and brain: Perhaps we can accept tualize how multimodal and multisensory brain science and still retain the idea of a mappings between conceptualization and soul. situated action are unified. A second sense A recent addition to the series of liter- is that of Harris (2000), who argues princi- ary analyses begun by Auerbach is by novel- pally in the context of child development ist and literary theorist David Lodge, who for simulation as the means by which we has become interested in the comparison may read another mind and more generally of literary and scientific approaches to con- project ourselves into situations not imme- sciousness. In the title essay of Conscious- diately present. I describe a third, related ness and the Novel (2002), Lodge describes sense (Oatley, 1999): that stories are sim- how, while writing Consciousness Explained, ulations that run not on computers but on Dennett came across one of Lodge’s novels, minds. Although our human minds are good Nice Work (1988), and found in it a satirical at understanding single causes and single portrait of literary postmodernism very like intentions, we are not good at understand- the idea of the absent self that Dennett was ing several of them interacting. A simula- proposing. tion (in Oatley’s sense) is a means we use Lodge concludes that, although cogni- to understand such complex interactions: It tive scientists have started to devote them- can be run forward in time for planning selves to the question of consciousness, they and backward for understanding. In their may neglect both the substance of what has constructions of characters’ intentions and been discovered by writers of literature – actions, novels and dramas explore the what- which allows us to explore and under- if of problems we humans face (vicissitudes) stand our consciousness – and the human- when repercussions of our actions occur P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

398 the cambridge handbook of consciousness

beyond the horizon of habitual understand- certain emotional processes, however, we ing. Multiple centres of consciousness can be may be closer to the unconscious end of the set up in literary simulations and enable us spectrum. Certainly Shakespeare, that most to enter the dialogues they afford. accomplished of all narrators and one who has perhaps prompted more insightful consciousness than any other, put it to us that a fair amount of our judgement may not be as Coda conscious as we might like to believe, especially when we are in the grip of an emo- If selfhood and the conscious understand- tion. In A Midsummer Night’s Dream, love ing of ourselves and others are the work induced by the administration of juice of “a of the novelist in each of us, we may have little western flower “ (a mere neurochem- learned some of this work from writers. If, ical substance as one might nowadays say) furthermore, we take Vygotsky’s view, sto- is followed not so much by conscious ratio- ries are conduits by which thinking about nality as by rationalization. It is the emotion the self, about human plans with others and rather than any consciously narrated choice their vicissitudes, and about human emo- that sets the frame of relationship, as it did tions has become explicitly available for our I think among our hominid ancestors. The use via culture. Imaginative literature is a one so dosed with the juice gazes upon the means by which a range of human situa- new loved one, and language comes from the tions and predicaments are explored so that love, not love from the language-based deci- they can be made part of ourselves. It is sions. Here in such a condition is Titania as a range we could not hope to encounter she speaks to Bottom (the weaver) who has directly. been changed into an ass. He is the first indi- I have argued that some of who we are vidual she sees when she opens her eyes after flows from a unifying sense of ourselves sleeping. as a character whom we improvise. Within narrative-like constraints we come to accept I pray thee gentle mortal sing again. who we are and consciously to create our- Mine ear is much enamoured of thy note; selves to be coherent with it. We commit So is mine eye enthrallèd to thy shape; ourselves to certain others, and to a cer- And thy fair virtue’s force perforce doth tain kind of life. But Dennett is also right. move me Plenty of our brain processes – those that On the first view, to say, to swear, I love thee (3, 1, 124–128). produce an emotion or mood here, a lack of attention there, a piece of selfishness The question for us humans is how to inte- when we ought to be thinking of someone grate the various phenomena, caught as else – can proceed not from any conscious we are between emotional urgencies and a or narratively coherent self, but from ill- meaning-making consciousness. Such inte- understood brain processes and unintegrated grative capacities are a recent emergence impulses. At present we have few estimates in human beings. Perhaps they are not yet of what proportion of our acts and thoughts working as well as they might. is, in the terms of this chapter, consciously decided within a unifying narrative frame and what proportion derives from unintegrated elements of what one might call the Acknowledgments Dennettian unconscious. In terms of making arrangements with other people, it looks I am very grateful to Robyn Fivush, Keith as if as many as 95% of actions based on Stanovich, Richard West, Philip Zelazo, and explicitly mutual plans can be made as con- two anonymous referees whose suggestions sciousness decides, based on folk-theoretical much helped my thoughts and understand- categories of goals and beliefs. In terms of ing in revising this chapter. P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

narrative modes of consciousness and selfhood 399

and identity: Studies in autobiography, self, and References culture (pp. 39–58). Amsterdam: Benjamins. Bruner, J. (1986). Actual minds, possible worlds. Abrams, M. H. (1953). The mirror and the Cambridge, MA: Harvard University Press. lamp: Romantic theory and the critical tradition. Bruner, J., S. (2002). Making stories: Law, litera- Oxford: Oxford University Press. ture, life. New York: Farrar, Straus and Giroux. Aiello, L. C., & Dunbar, R. I. M. (1993). Neo- Burckhardt, J. (2002). The civilization of the cortex size, group size, and the evolution of Renaissance in Italy (S. G. C Middlemore, language. Current Anthropology, 34, 184–193. Trans.). New York: Random House. Astington, J. W., Harris, P. L., & Olson, D. R. Call, J., Hare, B., Carpenter, M., & Tomasello, (Eds.). (1988). Developing theories of mind. New M. (2004). ‘Unwilling’ versus ‘unable’: Chim- York: Cambridge University Press. panzees’ understanding of human intentional Auerbach, E. (1953). Mimesis: The representation action. Developmental Science, 7, 488–498. of reality in Western literature (W. R. Trask, Chauvet, J.-M., Deschamps, E. , & Hillaire, C. Trans.). Princeton, NJ: Princeton University (1996). Dawn of art: The Chauvet cave. New Press. York: Abrams. 1993 Baddeley, A. D. ( ). Verbal and visual subsys- Churchland, P. S. (1986). Neurophilosophy. tems of working memory. Current Biology, 3, Cambridge, MA: MIT Press. 563–565. Claparede,` E. (1934). La genèse de la hypothèse. 1984 Bakhtin, M. ( ). Problems of Dostoevsky’s poet- Geneva: Kundig. ics (C. Emerson, Trans.). Minneapolis: Univer- Defoe, D. (1719). Robinson Crusoe. London: Pen- sity of Minneapolis Press (Original work pub- guin. lished 1963). Dennett, D. C. (1991). Consciousness explained. Bal, M. (1997). Narratology: Introduction to the Boston: Little, Brown & Co. theory of narrative (2nd ed). Toronto: Univer- 1992 sity of Toronto Press. Dennett, D. C. ( ). The self as a center of narrative gravity. In F. S. Kessel, P. M. Cole, & D. Balleine, B., & Dickinson, A. (1998). Conscious- L. Johnson (Eds.), Self and consciousness: Mul- ness – the interface between affect and cogni- tiple perspectives (pp. 103–115). Hillsdale, NJ: tion. In J. Cornwell (Ed.), Consciousness and Erlbaum. human identity (pp. 57–85). Oxford: Oxford 1911 University Press. Descartes, R. ( ). Passions of the soul.InE.L. Haldane & G. R. Ross (Eds.), The philosophical Barresi, J. (2001). Extending self-consciousness works of Descartes. New York: Dover (Original into the future. In C. Moore & K. Lemmon work published 1649). (Eds.). The self in time: Developmental perspec- 1982 tives (pp. 141–161). Mahwah, NJ: Erlbaum. de Waal, F. ( ). Chimpanzee politics. New York: Harper & Row. Barsalou, L. W.(2003). Situated simulation in the 1991 human conceptual system. Language and Cog- Donald, M. ( ). Origins of the modern mind. nitive Processes, 18, 513–562. Cambridge, MA: Harvard University Press. 1960 Behne, T., Carpenter, M., Call, J., & Tomasello, Donne, J. ( ). The sermons of John Donne, M. (2005). Unwilling versus unable: Infants’ Vol. 8. (G. R. Potter & E. M. Simpson, Eds.). understanding of intentional action. Develop- Berkeley: University of California Press (Orig- 1615 1631 mental Psychology, 41, 328–337. inal work published - ). 1993 Booth, W. C. (1988). The company we keep: An Dunbar, R. I. M. ( ). Coevolution of neocor- ethics of fiction. Berkeley, CA: University of tical size, group size, and language in humans. 681 735 California Press. Behavioral and Brain Sciences, 16, – . 1996 Brockmeier, J., & Carbaugh, D. (2001). Introduc- Dunbar, R. I. M. ( ). Grooming, gossip and the tion. In J. Brockmeier & D. Carbaugh (Eds.), evolution of language. London: Faber & Faber. Narrative and identity: Studies in autobiogra- Dunbar, R. I. M. (2003). The social brain: Mind, phy, self, and culture (pp. 1–22). Amsterdam: language, and society in evolutionary perspec- Benjamins. tive. Annual Review of Anthropology, 32, 163– Brockmeier, J., & Harre,´ R. (2001). Problems 181. and promises of an alternative paradigm. In J. Dunbar, R. I. M. (2004). The human story: A new Brockmeier & D. Carbaugh (Eds.), Narrative history of mankind’s evolution. London: Faber. P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

400 the cambridge handbook of consciousness

Edelson, M. (1992). Telling and enacting stories in Helmholtz, H. (1962). Treatise on physiological psychoanalysis. In J. W. Barron, M. N. Eagle, & optics, Vol 3. New York: Dover (Original work D. L. Wolitzsky (Eds.), Interface of psychoanal- published 1866). ysis and psychology (pp. 99–123). Washington, Hogan, P. C. (2003). The mind and its stories. DC: American Psychological Association. Cambridge: Cambridge University Press. Fillmore, C. J. (1968). The case for case. In E. Bach Homer (1987). The Iliad (M. Hammond, Ed. & R. T. Harms (Eds.), Universals in linguistic & Trans.). Harmondsworth: Penguin (Original theory (pp. 1–88). New York: Holt, Rinehart & work in 850 bce). Winston. Humphrey, N. (1976). The social function of Fireman, G. D., McVay, T. E., & Flanagan, O. J. the intellect. In P. Bateson & R. Hinde 2003 (Eds.). ( ). Narrative and consciousness: Lit- (Eds.), Growing points in ethology. Cambridge: erature, psychology, and the brain. New York: Cambridge University Press. Oxford University Press. Humphrey, N. (2002). The mind made flesh: 2001 Fivush, R. ( ). Owning experience: Develop- Frontiers of psychology and evolution. Oxford: ing subjective perspective in autobiographical Oxford University Press. narratives. In C. Moore & K. Lemmon (Eds.), Humphrey, N., & Dennett, D. (2002). Speaking The self in time: Developmental perspectives for ourselves: An assessment of multiple per- (pp. 35–52). Mahwah, NJ: Erlbaum. sonality disorder. In N. Humphrey (Ed.), The 2002 Flanagan, O. ( ). The problem of the soul: Two mind made flesh: Frontiers of psychology and evo- visions of mind and how to reconcile them. New lution (pp. 19–48). Oxford: Oxford University York: Basic Books. Press. 1979 Freud, S. ( ). Fragment of an analysis of a James, W. (1890). The principles of psychology. case of hysteria (Dora). In J. Strachey & A. New York: Holt. Richards (Eds.), The Pelican Freud library, Vol. Jaynes, J. (1976). The origin of consciousness in 9: Case histories, II (Vol. 7,pp.7–122). London: the breakdown of the bicameral mind. London: Penguin. Allen Lane. Gallese, V., Keysers, C., & Rizzolatti, G. (2004). Lacan, J. (1977). The mirror stage as formative of A unifying view of the basis of social cognition. the function of the I as revealed in psychoana- Trends in Cognitive Sciences, 8, 396–403. lytic experience. In A. Sheridan (Ed.), Jacques 1992 Gazzaniga, M. S. ( ). Brain modules and belief Lacan: Ecrits, a selection (pp. 1–7). London: formation. In F. S. Kessel, P. M. Cole, & D. Tavistock. (Original work published 1949.) L. Johnson (Eds.), Self and consciousness: Mul- 1996 The emotional brain: The mys- tiple perspectives (pp. 88–102). Hillsdale, NJ: LeDoux, J. ( ). terious underpinnings of emotional life Erlbaum. . New York: Simon & Schuster. Gazzaniga, M. S. (1998). The split brain revisited. Lee, R. B. (1984). The Dobe !Kung. New York: Scientific American, 279(1), 35–39. Holt, Rinehart & Winston. Goldie, P. (2003). One’s remembered past: Lewis, M., Sullivan, M. W., Stanger, C., & Weiss, Narrative thinking, emotion, and the exter- M. (1989). Self development and self-conscious nal perspective. Philosophical Papers, 32, 301– emotions. Child Development, , 146–156. 309. 60 1988 Goodall, J. (1986). The chimpanzees of Gombe: Lodge, D. ( ). Nice work. London: Secker & Patterns of behavior. Cambridge, MA: Harvard Warburg. University Press. Lodge, D. (2001). Thinks. London: Secker & Grazzani-Gavazzi, I., & Oatley, K. (1999). Warburg. The experience of emotions of interdepen- Lodge, D. (2002). Consciousness and the novel. dence and independence following interper- Cambridge, MA: Harvard University Press. sonal errors in Italy and Anglophone Canada. Marcus, S. (1984). Freud and Dora: Story, his- Cognition and Emotion, 13, 49–63. tory, case history. In S. Marcus (Ed.), Freud and Groden, M., & Kreisworth, M. (1994). The Johns the culture of psychoanalysis (pp. 42–86). New Hopkins guide to literary theory and criticism. York: Norton. Baltimore: Johns Hopkins University Press. Marvell, A. (1968). Complete poetry (George deF. Harris, P. L. (2000). The work of the imagination. Lord, Ed.). New York: Modern Library. (Orig- Oxford: Blackwell. inal work 1637–1678). P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

narrative modes of consciousness and selfhood 401

McGrew, W.(1992). Chimpanzee material culture. J. Wellenkamp (Eds.), Everyday conceptions Cambridge: Cambridge University Press. of emotions: An introduction to the psychology, Mead, G. H. (1913). The social self. Journal of Phi- anthropology, and linguistics of emotion. NATO losophy, Psychology and Scientific Methods, 10, ASI Series D 81 (pp. 145–165). Dordrecht: 374–380. Kluwer. Mithen, S. (1996). The prehistory of the mind: Oatley, K., & Mar, R. A. (2005). Evolutionary The cognitive origins of art and science. London: pre-adaptation and the idea of character in fic- Thames and Hudson. tion. Journal of Cultural and Evolutionary Psy- 181 196 Mithen, S. (2001). The evolution of imagina- chology, 3, – . tion: An archaeological perspective. SubStance, Oatley, K., Sullivan, G. D., & Hogg, D. (1988). #94–95, 28–54. Drawing visual conclusions from analogy: A Moore, C., & Macgillivray, S. (2004). Social theory of preprocessing, cues and schemata in understanding and the development of pru- the perception of three dimensional objects. dence and prosocial behavior. In J. Baird & B. Journal of Intelligent Systems 1, 97–133. Sokol, (Eds.). New directions for child and ado- Oatley, K., & Yuill, N. (1985). Perception of per- lescent development. New York: Jossey Bass. sonal and interpersonal actions in a cartoon Nelson, K. (1989). Narratives from the crib. New film. British Journal of Social Psychology, 24, York: Cambridge University Press. 115–124. Nelson, K. (1996). Language in cognitive devel- Panksepp, J. (1998). Affective neuroscience: The opment: Emergence of the mediated mind. New foundations of human and animal emotions. York: Cambridge University Press. Oxford: Oxford University Press. Nelson, K. (2003). Narrative and the emergence Panksepp, J. (2001). The neuro-evolutionary cusp of consciousness of self. In G. D. Fireman, between emotions and cognitions: Implications T. E. McVay, & O. Flanagan (Eds.), Narrative for understanding consciousness and the emer- and consciousness: Literature, psychology, and the gence of a unified mind science. Evolution and brain (pp. 17–36). New York: Oxford Univer- Cognition, 7, 141–163. sity Press. Panksepp, J. (2005). Affective consciousness: Oatley, K. (1988). On changing one’s mind: A Core emotional feelings in animals and hu- possible function of consciousness. In A. J. mans. Consciousness and Cognition, 14, 30–80. Marcel & E. Bisiach (Ed.), Consciousness in Paulhan, F. (1930). The laws of feeling. London: contemporary science (pp. 369–389). Oxford: Kegan-Paul, French, Trubner & Co. (Original Oxford University Press. work published 1887). Oatley, K. (1990). Freud’s psychology of inten- Pennebaker, J. W., Kiecolt-Glaser, J. K., & Glaser, tion: The case of Dora. Mind and Language, 5, R. (1988). Disclosure of traumas and immune 69–86. function: Health implications of psychother- Oatley, K. (1992a). Best laid schemes: The psychol- apy. Journal of Consulting and Clinical Psychol- ogy of emotions. New York: Cambridge Univer- ogy, 56, 239–245. sity Press. Pennebaker, J. W., & Seagal, J. D. (1999). Forming Oatley, K. (1992b). Integrative action of narrative. a story: The health benefits of narrative. Journal In D. J. Stein & J. E. Young (Eds.), Cognitive of Clinical Psychology, 55, 1243–1254. science and clinical disorders (pp. 151–170.). San Pennington, N., & Hastie, R. (1991). A cogni- Diego: Academic Press. tive theory of juror decision making: The story Oatley, K. (1999). Why fiction may be twice as model. Cardozo Law Review, 13, 519–557. true as fact: Fiction as cognitive and emotional Peterson, J. (1999). Maps of meaning. New York: simulation. Review of General Psychology, 3, Routledge. 101–117. Povinelli, D. J. (2000). Folk physics for apes: The Oatley, K. (2002). Character. Paper presented at chimpanzee’s theory of how the world works. New the Society for Personality and Social Psychol- York: Oxford University Press. ogy Annual Convention, Savannah, GA. Povinelli, D. J., & O’Neill, D. K. (2000). Do Oatley, K., & Larocque, L. (1995). Everyday chimpanzees use their gestures to instruct concepts of emotions following every-other- each other? In S. Baron-Cohen, H. Tager- day errors in joint plans. In J. Russell, J.- Flusberg, & D. Cohen (Eds.), Understanding M. Fernandez-Dols, A. S. R. Manstead, & other minds: Perspectives from developmental P1: KAE 0521857430c14 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 22:58

402 the cambridge handbook of consciousness

cognitive neuroscience (pp. 459–487). Oxford: Trabasso, T., & van den Broek, P. (1985). Causal Oxford University Press. thinking and the representation of narrative Power, R. (1979). The organization of purposeful events. Journal of Memory and Language, 24, dialogues. Linguistics, 17, 107–152. 612–630. Rizzolatti, G., & Arbib, M. A. (1998). Language Turner, S. R. (1994). The creative process: A within our grasp. Trends in Neuroscience, 21, computer model of storytelling and creativity. 188–194. Hillsdale, NJ: Erlbaum. 2000 Rizzolatti, G., Fogassi, L., & Gallese, V. (2001). Velleman, J. D. ( ). The possibility of practical Neurophysiological mechanisms underlying reason. Oxford: Oxford University Press. the understanding and imitation of action. Velleman, J. D. (2002). The self as narrator. First Nature Reviews: Neuroscience, 2 , 661–670. of the Jerome Simon Lectures in Philosophy, 2002 Rumelhart, D. E. (1975). Notes on a schema October University of Toronto. Retrieved ∼ for stories. In D. G. Bobrow & A. M. Collins from www-personal.umich.edu/ velleman/ (Eds.), Representation and understanding: Stud- Work/Narrator.html. ies in cognitive science. New York: Academic Vygotsky, L. (1962). Thought and language (E. H. Press. G. Vakar, Trans.). Cambridge, MA: MIT Press. Schank, R. C. (1990). Tell me a story: A new look at Vygotsky, L. (1978). Tool and symbol in child real and artiﬁcial memory. New York: Scribner. development. In M. Cole, V. John-Steiner, S. Scribner, & E. Souberman (Eds.), Mind in Schank, R. C., & Berman, T.R. (2002). The perva- society: The development of higher mental pro- sive role of stories. In M. C. Green, J. J. Strange, cesses (pp. 19–30). Cambridge, MA: Harvard & T. C. Brock (Eds.), Narrative impact: Social University Press. (Original work published and cognitive foundations (pp. 287–313). Mah- 1930). wah, NJ: Erlbaum. Watt, I. (1957). The rise of the novel: Studies Searle, J. R. (1969). Speech acts: An essay in the in Defoe, Richardson, and Fielding. London: philosophy of language. Cambridge: Cambridge Chatto & Windus. University Press. Willams, B. (1993). Shame and necessity. Berkeley, Sebald, W. G. (2001). Austerlitz. Toronto: Knopf CA: University of California Press. Canada. Wilson, G. M. (2003). Narrative. In J. Levin- Shakespeare, W. (1997). The Norton Shakespeare son (Ed.), The Oxford handbook of aesthetics (S. Greenblatt, Ed.). New York: Norton. (Orig- (pp. 392–407). New York: Oxford University inal work published 1623). Press. Snell, B. (1982). The discovery of the mind in Greek Woolf, V. (1925). Mrs. Dalloway. London: Ho- philosophy and literature. New York: Dover. garth Press. (Original work published 1953). Wrangham, R. (2001). Out of the Pan, into the Spence, D. P.(1982). Historical truth and narrative ﬁre: How our ancestors’ evolution depended truth. New York: Dover. on what they ate. In F. B. M. de Waal (Ed.), 2004 Strawson, G. ( ). Against narrativity. Ratio, Tree of origin: What primate behavior can tell 428 452 17, – . us about human social evolution (pp. 121–143). Sussman, G. J. (1975). A computer model of skill Cambridge, MA: Harvard University Press. acquisition. New York: American Elsevier. Zelazo, P. D. (2004). The development of con- Tomasello, M. (1999). The cultural origins of scious control in childhood. Trends in Cognitive human cognition. Cambridge, MA: Harvard Sciences, 8, 12–17. University Press. Zelazo, P. D. , & Sommerville, J. A. (2001). Levels Tomasello, M., & Rakoczy, H. (2003). What of consciousness of the self in time. In C. Moore makes human cognition unique? From individ- & K. Lemmon (Eds.). The self in time: Devel- ual to shared to collective intentionality. Mind opmental perspectives (pp. 229–252). Mahwah, and Language, 18, 121–147. NJ: Erlbaum. P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

E. Developmental Psychology

403 P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

404 P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

CHAPTER 15 The Development of Consciousness

Philip David Zelazo, Helena Hong Gao, and Rebecca Todd

Abstract there is a relative dearth of research on the way in which consciousness develops This chapter examines the extent to during ontogeny. This may be due in part which consciousness might develop during to a widespread belief that it goes without ontogeny. Research on this topic is con- saying that children are conscious in same verging on the suggestion that conscious- way as adults. Indeed, most people probably ness develops through a series of levels, each believe that newborn infants – whether of which has distinct consequences for the protesting their arrival with a vigorous cry or quality of subjective experience, the poten- staring wide-eyed and alert at their mother – tial for episodic recollection, the complexity are conscious in an essentially adult-like of children’s explicit knowledge structures, fashion. So, although there are dramatic and the possibility of the conscious control differences between infants and toddlers of thought, emotion, and action. The dis- and between preschoolers and adolescents, crete levels of consciousness identified by these differences are often assumed to developmental research are useful for under- reflect differences in the contents of chil- standing the complex, graded structure of dren’s consciousness, but not in the nature conscious experience in adults, and they of consciousness itself. reveal a fundamental dimension along which There is currently considerable debate, consciousness varies: the number of itera- however, concerning when a fetus first tions of recursive reprocessing of the con- becomes capable of conscious experience tents of consciousness. (including pain). This debate has been instigated by proposed legislation requiring physicians in the United States to Introduction inform women seeking abortions after 22 weeks gestational age (i.e., developmental Despite the explosion of scientific interest in age plus 2 weeks) that fetuses are able consciousness during the past two decades, to experience pain (Arkansas, Georgia, and 405 P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

406 the cambridge handbook of consciousness

Minnesota have already passed similar laws). does not feel like anything to be an infant. Professor Sunny Anand, a paediatrician and Carruthers (1996) further tests our credulity an expert on neonatal pain, recently testified when he argues that children are not actu- before the U.S. Congress (in relation to ally conscious until 4 years of age (because the Unborn Child Pain Awareness Act of it is not until then that children can for- 2005; Anand, 2005) that the “substrate and mulate beliefs about psychological states). mechanisms for conscious pain perception” Kagan and Carruthers characterize infants develop during the second trimester. Others and/or young children essentially as uncon- (e.g., Burgess & Tawia, 1996; Lee et al., 2005) scious automata or zombies (cf. Chalmers, have suggested that consciousness develops 1996; see Chapter 3) – capable of cognitive later, during the third trimester (at around function but lacking sentience. 30 weeks gestational age) – because that is Although many theorists treat consci- when there is first evidence of functional ousness as a single, all-or-nothing pheno- neural pathways connecting the thalamus to menon, others distinguish between first- sensory cortex (see Chapter 27). order consciousness and a meta-level of Inherent in these claims is the assump- consciousness. For example, they may dis- tion that consciousness is an “all-or-nothing” tinguish between consciousness and meta- phenomenon – one is either conscious or consciousness (Schooler, 2002), primary not, and capable of consciousness or not consciousness and higher-order conscious- (cf. Dehaene & Changeux, 2004). This ness (Edelman & Tononi, 2000), or core assumption is also reflected, for exam- consciousness and extended consciousness ple, in information-processing models (e.g., (Damasio, 1999). In the developmental lit- Moscovitch, 1989; Schacter, 1989) in which erature, this dichotomous distinction is usu- consciousness corresponds to a single sys- ally described as the difference between tem and information is either available to consciousness and self-consciousness (e.g., this system or not. From this perspective, Kagan, 1981; Lewis, 2003). In most cases, it is natural to think of consciousness as the first-order consciousness refers to aware- something that emerges full-blown at a par- ness of present sensations (e.g., an integrated ticular time in development, rather than multimodal perceptual scene; Edelman & something that itself undergoes transforma- Tononi, 2000), whereas the meta-level con- tion – something that develops, perhaps grad- sciousness is generally intended to cap- ually. Although the current debate concern- ture the full complexity of consciousness ing pain and abortion has centered on the as it is typically experienced by healthy prenatal period, there are those who believe human adults. Edelman and Tononi (2000), that consciousness emerges relatively late. for example, suggest that higher-order con- Jerome Kagan (1998), for example, writes sciousness is, according to their model, that “sensory awareness is absent at birth “accompanied by a sense of self and the but clearly present before the second birth- ability . . . explicitly to construct past and fu- day” (p. 48). It should be noted that Kagan ture scenes” (p. 102). Developmentally, the believes that consciousness does develop implication is that infants are limited to rel- beyond the emergence of sensory aware- atively simple sensory consciousness until an ness, but if you’ve ever met a toddler (say, enormous transformation (usually presumed a 14-month-old), it may be difficult to imag- to be neurocognitive in nature and involving ine that children at this age lack sensory the acquisition of language or some degree awareness – which Kagan (1998) defines as of conceptual understanding) occurs that “awareness of present sensations” (p. 46). simultaneously adds multiple dimensions The implications of Kagan’s claim are (e.g., self-other, past-future) to the qualita- profound. For example, if we follow Nagel tive character of experience. This profound (1974), who asserts that the essence of sub- metamorphosis has typically been hypothe- jective experience is that it is “like some- sized to occur relatively early in infancy (e.g., thing” to have it (see Chapter 3), we might Stern, 1990; Trevarthen & Aitken, 2001)or conclude, as Carruthers (1989) does, that it some time during the second year (e.g., P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

the development of consciousness 407

Kagan, 1981; Lewis, 2003; Wheeler, 2000), tributions by showing how children’s con- depending on the criteria used for inferring sciousness, including the way in which chil- higher-order self-consciousness. dren experience reality, is changed during In this chapter, we propose that dis- particular developmental transformations. cussions of the development of conscious- That is, they all understood that the struc- ness have been hampered by a reliance on ture of consciousness itself – and not just relatively undifferentiated notions of con- the contents of consciousness – develops sciousness. Indeed, we argue that develop- over the course of childhood. For Baldwin, mental data suggest the need for not just two, infants’ experience can initially be charac- but many dissociable levels of conscious- terized as a state of adualism – meaning ness; information may be available at one that they are unaware of any distinctions level but not at others (see Morin, 2004, that might be implicit in the structure of 2006, for a review of recent models of con- experience (e.g., subject vs. object, ego vs. sciousness that rely on the notion of lev- alter; e.g., Baldwin, 1906). During the course els; see also Cleeremans & Jimenez,´ 2002, of development, however, children proceed for a related perspective on consciousness through a series of “progressive differentia- as a graded phenomenon). Consideration of tions between the knower and the known” these levels and of their utility in explain- (Cahan, 1984,p.131) that culminates in ing age-related changes in children’s behav- transcending these dualisms and recogniz- ior has implications for our understanding of ing their origin in what Baldwin calls the consciousness in general, including individ- dialectic of personal growth (by personal, ual differences in reﬂectivity and mindful- Baldwin refers both to oneself and to per- ness (see Chapter 19), but the focus here is sons). Baldwin, therefore, suggests that con- on development during childhood. Many of sciousness develops through a circular pro- the arguments regarding when consciousness cess of differentiation and then integration emerges – for example, at 30 months ges- (cf. Eliot’s poem, “Little Gidding”: “And the tational age (e.g., Burgess & Tawia, 1996), end of all our exploring/ Will be to arrive at 12 to 15 months after birth (Perner & where we started/ And know the place for Dienes, 2003), at the end of the second year the ﬁrst time.”). (Lewis, 2003), or around the fourth birthday Imitation plays a key role in this dialectic, (Carruthers, 2000) – have merit, and we which starts when an infant observes behav- propose that some of the most salient dis- ior that is (at least partially) outside of his crepancies among these accounts can be rec- or her behavioral repertoire. At this point, onciled from the perspective that conscious- the infant cannot identify with the behavior ness has several levels. Different theorists or the agent of the behavior, so the behavior have directed their attention to the emer- is viewed solely in terms of its outward or gence of different levels of consciousness; a projective aspects. By imitating this behav- developmental perspective allows us to inte- ior, however, the infant discovers the sub- grate these levels into a more comprehen- jective side of it, including, for example, sive model of consciousness as a complex, the feeling that accompanies it. Once this dynamic phenomenon. happens, the infant automatically ejects this newly discovered subjectivity back into his or her understanding of the original behavior. Baldwin (1894) gives the example of a Early Accounts of the Development girl who watches her father prick himself of Consciousness with a pin. Initially, she has no appreciation of its painful consequence. When she For early theorists such as Baldwin (e.g., imitates the behavior, however, she will feel 1892), Piaget (1936/1952), and Vygotsky the pain and then immediately understand (1934/1986), consciousness was the problem that her father felt it too. Subsequently, she to be addressed by the new science of psy- will view the behavior of pin pricking in chology, and these theorists made major con- a different light; her understanding of the P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

408 the cambridge handbook of consciousness

behavior will have been transformed from ume on the topic, The Grasp of Conscious- projective to subjective to ejective. In effect, ness, where he wrote, “The study of cog- she will have brought the behavior into the nizance [i.e., consciousness] has led us to scope of her self- and social-understanding, place it in the general perspective of the expanding the range of human behavior with circular relationship between subject and which she can identify. Baldwin (1897,p.36) object. The subject only learns to know him- writes, “It is not I, but I am to become it,” a self when acting on the object, and the lat- formulation that seems to capture the same ter can become known only as a result of fundamental insight about the development progress of the actions carried out on it” of consciousness as Freud’s (1933/1940, (p. 353). p. 86) famous “Wo Es war, soll Ich warden.” Vygotsky (1934/1986), in contrast to both (“Where it was, there I shall be”). Baldwin and Piaget, noted that children’s Piaget similarly saw “increasing self- consciousness was transformed mainly via awareness of oneself as acting subject” the appropriation of cultural tools, chiefly (Ferrari, Pinard, & Runions, 2001,p.207)–or language. Vygotsky, and then Luria (e.g., decreasing egocentrism – as one of the major 1959, 1961), proposed that thought and dimensions of developmental change, and speech first develop independently but then he tied this development to the emergence become tightly intertwined as a result of of new cognitive structures that allowed internalization – a process whereby the for new ways of knowing or experienc- formal structure inherent in a cultural ing reality. Indeed, for Piaget, conscious- practice, such as speaking, is first acquired ness (the experience of reality) is depen- in overt behavior and then reflected in one’s dent on one’s cognitive structures, which private thinking. Initially, speech serves a are believed to develop through a series of communicative purpose, but later it also stages primarily as a result of a process of acquires semantic, syntactic, and regulatory equilibration, whereby they become increas- functions. The emergent regulatory function ingly abstract (from practical to concep- of speech is inherently self-conscious, and tual) and reflect more accurately the logic it allows children to organize and plan their of the universe. Consciousness also devel- behavior, essentially rendering them capable ops in a characteristic way regardless of chil- of consciously controlled behavior (Luria, dren’s developmental stage; at all stages, 1961; Vygotsky, 1934/1986). Vygotsky from practical to conceptual, consciousness (1978) wrote, “With the help of speech “proceeds from the periphery to the cen- children, unlike apes, acquire the capacity ter” (Piaget, 1974/1977,p.334), by which to be both the subjects and objects of their Piaget meant that one first becomes aware own behavior” (p. 26). For Vygotsky, then, of goals and results and then later comes consciousness was transformed by language, to understand the means or intentions by with important consequences for action. which these results are accomplished. For Contemporary theorists have elaborated older (formal operational) children, Piaget on some of these seminal ideas about the (1974/1977) noted, this development from development of consciousness – although periphery to center can occur quite quickly they have not always explicitly addressed via the reflexive abstraction of practical senso- the implications for the character of chil- rimotor knowledge. This process, which cor- dren’s subjective experience. Barresi and responds to conceptualization or reflective Moore (1996), for example, offered a model redescription, allows children more rapidly of the development of perspective taking to transform knowledge-in-action into artic- that builds on Baldwin’s (1897) dialectic ulate conceptual understanding. In all cases, of personal growth. According to Barresi however, we see consciousness developing and Moore, young children initially take from action to conceptualization. Piaget’s a first-person, present-oriented perspective (1974/1977) emphasis on the role of action in on their own behavior (e.g., “I want candy the development of consciousness was sum- now”) and a third-person perspective on the marized concisely at the end of his key vol- behavior of others – seeing that behavior P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

the development of consciousness 409

from the outside, as it were. Because simul- sciousness develops during the first 5 years of taneous consideration of first- and third- life (and potentially beyond). Empirical and person perspectives is required for a repre- theoretical contributions to our understand- sentational understanding of mental states ing of the development of consciousness are (e.g., “I know there are sticks in the box, but reviewed in the context of the LOC model. he thinks there is candy”), young children have difficulty understanding false beliefs – both their own and those of others (see Overview of the Levels of 2001 Wellman, Cross, & Watson, , for a re- Consciousness (LOC) Model view). With development, however, children are better able to adopt a third-person The Levels of Consciousness (LOC) model perspective on their own behavior, imagine describes the structure of consciousness and a first-person perspective on the behavior attempts to show the consequences that of others, and coordinate these perspectives reflection has on the structure and functions into a single schema. of consciousness, including the key role that As another example, Karmiloff-Smith reflection plays in the conscious control of (1992) builds on Piaget’s (1974/1977) idea of thought, action, and emotion via explicit reflexive abstraction with her model of Rep- rules. In what follows, we consider the impli- resentational Redescription, in which con- cations of the LOC model for (1) the struc- sciousness develops as a function of domain- ture of consciousness, (2) cognitive control specific experience. According to this model, via the use of rules at different levels of knowledge is originally represented in an complexity, (3) the functions of prefrontal implicit, procedural format (Level I), but, cortex, and (4) the development of con- with sufficient practice, behavioral mastery sciousness in childhood. of these procedures is achieved and the knowledge is automatically redescribed into The Structure of Consciousness a more abstract, explicit format (Level E1). This representational format reveals the According to the LOC model, conscious- structure of the procedures, but is still not ness can operate at multiple discrete levels, conscious: Consciousness comes with yet and these levels have a hierarchical struc- additional levels of redescription or ‘explic- ture – they vary from a first-order level itation,’ which occur “spontaneously as part of consciousness (minimal consciousness) to of an internal drive toward the creation higher-order reflective levels that subsume of intra-domain and inter-domain relation- lower levels. Higher levels of conscious- ships” (1992,p.18). Level E2 is conscious but ness are brought about through an itera- not verbalizable, whereas Level E3 is both tive process of reflection, or the recursive conscious and verbalizable. reprocessing of the contents of conscious- Finally, Zelazo and his colleagues have ness via thalamocortical circuits involving expounded a model of consciousness, the regions of prefrontal cortex. Each degree Levels of Consciousness (LOC) model (e.g., of reprocessing results in a higher level of Zelazo, 1999, 2004; Zelazo & Jacques, 1996; consciousness, and this in turn allows for Zelazo & Zelazo, 1998) that builds on the the integration of more information into an work of Baldwin, Piaget, and Vygotsky and experience of a stimulus before a new stim- Luria – but especially Vygotsky and Luria. ulus is experienced; it allows the stimulus to Because this model is relatively compre- be considered relative to a larger interpretive hensive and addresses explicitly the poten- context. In this way, each additional level of tial implications of neurocognitive develop- consciousness changes the structure of expe- ment for children’s subjective experience, rience, and the addition of each level has we describe it in some detail. In what fol- unique consequences for the quality of sub- lows, we first provide an overview of the jective experience: The addition of higher model and then show how it aims to pro- levels results in a richer, more detailed expe- vide an account of the way in which con- rience and generates more “psychological P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

410 the cambridge handbook of consciousness

distance” from stimuli (e.g., Carlson, Davis, tion and diminution of attention to poten- & Leach, 2005; Dewey, 1931/1985; Sigel, tial influences on thought (inferences) and 1993). But the addition of new levels also has action when multiple possible influences are implications for the potential for episodic present. recollection (because information is pro- According to the LOC model, increases cessed at a deeper level; Craik & Lockhart, in rule complexity – whether age-related 1972), the complexity of children’s explicit (see below) or in response to situational knowledge structures, and the possibility of demands – are made possible by corre- the conscious control of thought, emotion, sponding increases in the extent to which and action. one reflects on one’s representations: They are made possible by increases in level of consciousness. Rather than taking rules for granted and simply assessing whether their Control by Rules at Various Levels antecedent conditions are satisfied, reflec- of Complexity tion involves making those rules themselves an object of consideration and considering According to the LOC model, conscious them in contradistinction to other rules at control is accomplished, in large part, by that same level of complexity. the ability to formulate, maintain in work- Figure 15.1 contrasts relatively automatic ing memory, and then act on the basis of action at a lower level of consciousness (a) explicit rule systems at different levels of with relatively deliberate action at a higher complexity – from a single rule relating a level of consciousness (b). The former type stimulus to a response, to a pair of rules, to a of action (a) is performed in response to hierarchical system of rules that allows one the most salient, low-resolution aspects of to select among incompatible pairs of rules, a situation, and it is based on the formu- as explained by the Cognitive Complex- lation of a relatively simple rule system – ity and Control (CCC) theory (e.g., Frye, in this case, nothing more than an explicit Zelazo, & Palfai, 1995; Zelazo, Muller,¨ Frye, representation of a goal maintained in work- & Marcovitch, 2003). On this account, rules ing memory. The more deliberate action are formulated in an ad hoc fashion in poten- (b) occurs in response to a more carefully tially silent self-directed speech. These rules considered construal of the same situation, link antecedent conditions to consequences, brought about by several degrees of repro- as when we tell ourselves, “If I see a mailbox, cessing the situation. The higher level of con- then I need to mail this letter.” When peo- sciousness depicted in Figure 15.1b allows for ple reflect on the rules they represent, they the formulation (and maintenance in work- are able to consider them in contradistinc- ing memory) of a more complex and more tion to other rules and embed them under flexible system of rules or inferences (in this higher-order rules, in the same way that we case, a system of embedded rules considered might say, “If it’s before 5 p.m., then if I see against the backdrop of the goal that occa- a mailbox, then I need to mail this letter, sions them). otherwise, I’ll have to go directly to the post The tree diagram in Figure 15.2 illus- office.” In this example, the selection of a trates the way in which hierarchies of rules simple conditional statement regarding the can be formed through reflection – the mailbox is made dependent on the satisfac- way in which one rule can first become an tion of another condition (namely, the time). object of explicit consideration at a higher More complex rule systems, like the system level of consciousness and then be embed- of embedded if-if-then rules in this exam- ded under another higher-order rule and ple, permit the more flexible selection of controlled by it. Rule A, which indicates certain rules for acting when multiple con- that response 1 (r 1) should follow stimulus 1 flicting rules are possible. The selection of (s 1), is incompatible with rule C, which con- certain rules then results in the amplifica- nects s1 to r2 . Rule A is embedded under, and P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

the development of consciousness 411

semantic LTM levels of consciousness working memory

G1 descs descA rel items (e.g., goals) recC

Iobjs IobjA cc action programs minC procedural LTM

objA response

levels of consciousness working memory

ER1=PR1+PR2 rel SdescC goal; embedded rules refC2 G1; PR1=R1+R2 semantic LTM rel SdescB goal; pair of rules refC1 G1; R1=C1 A1 rel Sdescs SdescA goal; rule=if c then act selfC rel G1 descs descA Items (e.g., goals) recC cc Iobjs IobjA action programs minC procedural LTM

objA response Figure 15.1. The implications of reflection (levels of consciousness) for rule use. (a, top): Relatively automatic action on the basis of a lower level of consciousness. An object in the environment (objA) triggers an intentional representation of that object (IobjA) in semantic long-term memory (LTM); this IobjA, which is causally connected (cc) to a bracketed objA, becomes the content of consciousness (referred to at this level as minimal consciousness or minC). The contents of minC are then fed back into minC via a re-entrant feedback process, producing a new, more reflective level of consciousness referred to as recursive consciousness or recC. The contents of recC can be related (rel) in consciousness to a corresponding description (descA) or label, which can then be deposited into working memory (WM) where it can serve as a goal (G1) to trigger an action program from procedural LTM in a top-down fashion. (b, bottom): Subsequent (higher) levels of consciousness, including self-consciousness (selfC), reflective consciousness 1 (refC1), and reflective consciousness 2 (refC2). Each level of consciousness allows for the formulation and maintenance in WM of more complex systems of rules. (Reprinted with permission from Zelazo, P. D. (2004). The development of conscious control in childhood. Trends in Cognitive Sciences, 8, 12–17.) (See color plates.) P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

412 the cambridge handbook of consciousness

(F) and maintenance in working memory of

c1 c2 a more complex rule system. A particular level of consciousness (SelfC) is required (E) to use a single explicit rule such as (A); a higher level of consciousness (refC1)is (A) (B) (C) (D) required to select between two univalent rules using a rule such as (E); a still higher s1 s2 s1 s2 level (refC2) is required to switch between two bivalent rules using a rule such as (F).

r1 r2 r2 r1 Figure 15.2 . Hierarchical tree structure depicting formal relations among rules. Note: c 1 The Role of Prefrontal Cortex in and c 2 = contexts; s 1 and s 2 = stimuli; r 1 and Higher Levels of Consciousness r 2 = responses. (Adapted from Frye, D., 1995 Zelazo, P. D., & Palfai, T. ( ). Theory of mind The potential role of prefrontal cortex in and rule-based reasoning. Cognitive Development, reflection is arguably revealed by work on 10, 483–527). the neural correlates of rule use (see Bunge, 2004). Bunge and Zelazo (2006) summa- controlled by, a higher- order rule (rule E) rized a growing body of evidence that pre- that can be used to select rule A or rule B, frontal cortex plays a key role in rule use and rule E, in turn, is embedded under an and that different regions of prefrontal cor- even higher-order rule (rule F) that can be tex are involved in representing rules at dif- used to select the discrimination between ferent levels of complexity – from learned rules A and B as opposed to the discrimi- stimulus-reward associations (orbitofrontal nation between rules C and D. This higher- cortex; Brodmann’s area [BA] 11), to sets order rule makes reference to setting condi- of conditional rules (ventrolateral prefrontal tions or contexts (c 1 and c 2 ) that condition cortex [BA 44, 45, 47] and dorsolateral pre- the selection of lower-order rules, and that frontal cortex [BA 9, 46]), to an explicit would be taken for granted in the absence of consideration of task sets (rostrolateral pre- reflection. Higher-order rules of this type (F) frontal cortex [or frontopolar cortex; BA 10]; are required in order to use bivalent rules in see Figure 15.3). which the same stimulus is linked to differ- Figure 15.3 illustrates the way in which ent responses (e.g., rules A and C). Simpler regions of prefrontal cortex correspond to rules like E suffice to select between uni- rule use at different levels of complexity. valent stimulus-response associations – rules Notice that the function of prefrontal cor- in which each stimulus is uniquely associ- tex is proposed to be hierarchical in a way ated with a different response, as when mak- that corresponds, roughly, to the hierarchi- ing discriminations within a single stimulus cal complexity of the rule use underlying dimension. conscious control. As individuals engage in To formulate a higher-order rule such reflective processing, ascend through levels as F and deliberate between rules C and of consciousness, and formulate more com- D, on the one hand, and rules A and B, on plex rule systems, regions of lateral pre- the other, one has to be aware of the fact frontal cortex are recruited and integrated that one knows both pairs of lower-order into an increasingly elaborate hierarchy of rules. Figuratively speaking, one has to view prefrontal cortical function via thalamo- the two rule pairs from the perspective of cortical circuits. As the hierarchy unfolds, (F). This illustrates how increases in reflec- information is first processed via circuits tion on lower-order rules are required for connecting the thalamus and orbitofrontal increases in embedding to occur. Each level cortex. This information is then reprocessed of consciousness allows for the formulation and fed forward to ventrolateral prefrontal P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

the development of consciousness 413

44 46

45 10

47 11

increasing rule complexity

stimulus-reward univalent rules bivalent rules higher-order rules associations

C1 C2 C1 C2 S S S1 S2 1 2

R R 1 2 S1 S2 S1 S2 S1 S2 S1 S2

R1 R2 R2 R1 R1 R2 R2 R1

Figure 15.3. A hierarchical model of rule representation in lateral prefrontal cortex. A lateral view of the human brain is depicted at the top of the ﬁgure, with regions of prefrontal cortex identiﬁed by the Brodmann areas (BA) that comprise them: Orbitofrontal cortex (BA 11), ventrolateral prefrontal cortex (BA 44, 45, 47), dorsolateral prefrontal cortex (BA 9, 46), and rostrolateral prefrontal cortex (BA 10). The prefrontal cortex regions are shown in various colors, indicating which types of rules they represent. Rule structures are depicted below, with darker shades of blue indicating increasing levels of rule complexity. The formulation and maintenance in working memory of more complex rules depend on the reprocessing of information through a series of levels of consciousness, which in turn depends on the recruitment of additional regions of prefrontal cortex into an increasingly complex hierarchy of prefrontal cortex activation. Note:S = stimulus; check = reward; cross = nonreward; R = response; C = context, or task set. Brackets indicate a bivalent rule that is currently being ignored. (Reprinted with permission from Bunge, S., & Zelazo, P. D. (2006). A brain-based account of the development of rule use in childhood. Current Directions in Psychological Science 15, 118–121.) (See color plates.)

cortex via circuits connecting the thalamus Developmental Increases in Children’s and ventrolateral prefrontal cortex. Further Highest Level of Consciousness processing occurs via circuits connecting the thalamus to dorsolateral prefrontal cortex. According to the LOC model, there are Thalamocortical circuits involving rostrolat- four major age-related increases in the high- eral prefrontal cortex play a transient role in est level of consciousness that children the explicit consideration of task sets at each are able to muster (although children may level in the hierarchy. operate at different levels of consciousness P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

414 the cambridge handbook of consciousness

in different situations). These age-related or painful (avoid), but one is not conscious of increases in children’s highest level of con- seeing what one sees or that one (as an agent) sciousness correspond to the growth of is seeing what one sees. And because minC prefrontal cortex, which follows a pro- is tied to ongoing stimulation, one cannot tracted developmental course that mirrors recall seeing what one saw. MinC is hypoth- the development of the ability to use rules at esized to characterize infant consciousness higher levels of complexity. In particular, prior to the end of the first year of life. developmental research suggests that the In adults, this level of consciousness corre- order of acquisition of rule types shown sponds to so-called implicit information pro- in Figure 15.3 corresponds to the order in cessing, as when we drive a car without full which corresponding regions of prefrontal awareness, perhaps because we are conduct- cortex mature. Gray matter volume reaches ing a conversation at a higher level of con- adult levels earliest in orbitofrontal cortex, sciousness (in this example, we are operat- followed by ventrolateral prefrontal cortex, ing at two different levels of consciousness and then by dorsolateral prefrontal cortex simultaneously). Our behavioral routines (Giedd et al., 1999; Gogtay et al., 2004). are indeed elicited directly and automati- Measures of cortical thickness suggest that cally, but they are elicited as a function of dorsolateral prefrontal cortex and rostrolat- consciousness of immediate environmental eral prefrontal cortex (or frontopolar cor- stimuli (cf. Perruchet & Vinter, 2002). It fol- tex) exhibit similar, slow rates of structural lows that implicit processing does not occur change (O’Donnell, Noseworthy, Levine, & in a zombie-like fashion; it is simply unre- Dennis, 2005). With development, children flective (because the contents of minC are are able to engage neural systems involv- continually replaced by new stimulation) ing the hierarchical coordination of more and, as a result, unavailable for subsequent regions of prefrontal cortex – a hierarchical recollection. coordination that develops in a bottom-up Consider how minC figures in the pro- fashion, with higher levels in the hierarchy duction of behavior according to the LOC operating on the products of lower levels model (Fig. 15.1a). An object in the envi- through thalamocortical circuits. ronment (objA) triggers a “description” from Minimal consciousness. The LOC model semantic long-term memory. This particu- starts with the assumption that simple sen- lar description (or IobjA, for “intentional tience is mediated by minimal consciousness object”) then becomes an intentional object (minC; cf. Armstrong, 1980), the first-order of minC and automatically triggers an asso- consciousness on the basis of which more ciated action program that is coded in pro- complex hierarchical forms of consciousness cedural long-term memory. A telephone, for are constructed (through degrees of repro- example, might be experienced by a minC cessing). MinC is intentional in Brentano’s baby as ‘suckable thing,’ and this descrip- (1874/1973) sense – any experience, no mat- tion might trigger the stereotypical motor ter how attenuated, is experience of some- schema of sucking. Sensorimotor schemata thing (see Brentano’s description of presen- are modified through practice and accom- tations, p. 78 ff.), and it motivates approach modation (i.e., learning can occur; e.g., and avoidance behavior, a feature that is DeCasper et al., 1994; Kisilevsky et al., 2003; essential to the evolutionary emergence of Siqueland & Lipsitt, 1966; Swain, Zelazo, minC (e.g., Baldwin, 1894; Dewey, 1896; & Clifton, 1993), and they can be coordi- Edelman, 1989). However, minC is unre- nated into higher-order units (e.g., Cohen, flective and present-oriented and makes no 1998; Piaget, 1936/1952), but a minC infant reference to an explicit sense of self; these cannot represent these schemata in minC features develop during the course of child- (the infant is only aware of the stimuli hood. While minimally conscious, one is that trigger them). In the absence of reflec- conscious of what one sees (the object of tion and a higher level of consciousness, the one’s experience) as pleasurable (approach) contents of minC are continually replaced P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

the development of consciousness 415

by new intero- and exteroceptor stimula- and (b) a sense of self-agency that arises from tion and cannot be deposited into working interaction with the world. memory. Rochat has emphasized the role of pro- Thus, minC infants exhibit learning and prioception in the experience of self, and memory and may well perceive aspects of several studies have explored the role of themselves and their current state implic- contingency between visual and propriocep- itly, but they have no means by which they tive information in the process of distin- can consciously represent past experiences guishing self from the world between the or states or entertain future-oriented repre- ages of 2 and 5 months (Rochat & Morgan, sentations. That is, they cannot engage in 1995). For example, in one study, infants conscious recollection, although they pro- were shown split-screen images of their legs vide clear behavioral evidence of memory, that were either congruent or incongruent and they cannot entertain conscious expec- with the view they would normally have of tations or plans, although their behavior is their own legs (i.e., the incongruent images often future-oriented (e.g., Haith, Hazan, were shown from a different angle). They & Goodman, 1988; see Reznick, 1994, for looked significantly longer at the unfamiliar, a discussion of different interpretations of incongruent view of their own legs, espe- future-oriented behavior). At present, there cially if the general direction of depicted is no behavioral evidence that young infants movement conflicted with the direction of are capable of conscious recollection (as the actual, felt movement. The authors con- opposed to semantic memory; Tulving 1985) cluded that the infants have expectancies or explicit self awareness; their experience of about what constitutes self-movement, that events seems to be restricted to the present self is specified by the temporal and dynamic (see Figure 15.4a) – including objects in the contingency of sensory information in differ- immediate environment and current physi- ent modalities, and that by 3 months of age, cal states. infants have an intermodal body schema that Within the constraints of minC, however, constitutes an implicit bodily self. infants may come to learn quite a bit about Based on a series of empirical studies their bodies in relation to the world (e.g., examining infants’ actions, then, Rochat has Gallagher, 2005; Meltzoff, 2002; Rochat, described the self-differentiation process as 2001). Rochat (2003), for example, pro- the systematic exploration of perceptual poses five levels of self-understanding that, experience, scaffolded by dyadic interaction, from the perspective of the LOC model, that allows the emergence of an implicit can be seen to unfold within particular sense of self and other (Rochat & Striano, levels of consciousness. The levels of self- 2000). For Rochat, such an implicit, interac- understanding that develop in early infancy tive, and somatically based sense of self pro- are characterized by somatic sensation and vides the foundation for the more explicit expectancies about the world, but they are integration of first- and third-person infor- not accompanied by explicit, higher-order mation that will come later in childhood (see representations of self and other. For exam- below). And in terms of the LOC model, all ple, the emergence of level 1, or the differ- of this implicit learning about the self takes entiated self, begins at birth, as infants learn place at the level of minC. to distinguish their own touch from that of Given this characterization of minC – another. Level 2 , which refers to what Rochat as the simplest, first-order consciousness on calls the situated self, emerges at around 2 the basis of which more complex conscious- months and involves implicit awareness of ness is constructed, but one that allows an the self as an agent situated in space. In this implicit understanding of self, we might model, the differentiated and situated selves return to the question of first emergence: emerge from the development of (a) expec- When does minC emerge? According to this tations of contingency between different account, the onset of minC may be tied to a sensory modalities (intermodal contingency) series of anatomical and behavioral changes P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

416 the cambridge handbook of consciousness

(a) (d) temporally decentered self

now

1 2 3 now 4 5 6

history of self and world

(b) (e) temporally decentered self

familiar now desired

1 2 3 now 4 5 6

(c)

history of self

past now future

1 2 3 now 4 5 6

history of world Figure 15.4. Levels of consciousness and their implications for the experience of events in time. (a) MinC. The contents of minimal consciousness are restricted to present intero- and exteroreceptor stimulation (Now). (b) RecC. Past and future events can now be considered but toddlers cannot simultaneously represent the present when representing past or future events. When descriptions of past experiences become the contents of recursive consciousness, they will feel familiar. Future-oriented states (goals) may be accompanied by a feeling of desire. (c) SelfC. Children can consider descriptions of past or future-oriented events in relation to a present experience. For example, while conscious of their current state (Now), 2-year-olds can appreciate that Yesterday they went to the zoo. This creates the conditions for a subjective experience of self-continuity in time, but it does not allow simultaneous consideration of events occurring at two different times. (d) RefC1. From this higher level of consciousness, which allows for a temporally decentered perspective, children can consider two events occurring at two different times, including an event occurring in the present. For example, they can consider that Now, EventA is occurring, but Yesterday, EventB occurred. This is an important advance in the development of episodic memory, but at this point, the history of own’s own subjective experiences (history of self) and the history of the world are confounded – there is no means of conceptualizing the history of the world as independent of one’s own experiences of the world. (e) RefC2. From a temporally decentered perspective, children can coordinate two series, the history of the self and the history of the world. (Reprinted from Zelazo, P.D., & Sommerville, J. (2001). Levels of consciousness of the self in time. In C. Moore & K. Lemmon (Eds.), Self in time: Developmental issues (pp.229–252). Mahwah, NJ: Erlbaum.) P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

the development of consciousness 417

that occur during the third trimester of pre- explained by the emergence of the first new natal development – between about 24 and level of consciousness – recursive conscious- 30 weeks gestational age. First, and perhaps ness (recC). This new level of consciousness foremost, is the development of thalamo- allows for recollection and the maintenance cortical fibres that show synaptogenesis in of a goal in working memory, key functional sensory cortex (and not in prefrontal cortex – consequences. the provenance of higher levels of conscious- Recursive consciousness. The term ‘recur- ness). These thalamocortical connections are sive’ is used here in the sense of a com- established as early as 24 weeks gestational puter program that calls itself (see Chap- age, but the first evidence of functionality ter 21). In recC (see Fig. 15.1a), the contents does not occur until about 30 weeks, as indi- of minC at one moment are combined with cated by sensory-evoked potentials recor- the contents of minC at another moment via ded in preterm infants (Hrbek, Karlberg, & an identity relation (rel), allowing the tod- Olsson, 1973; Klimach & Cooke, 1988). A dler to label the initial object of minC. The number of other neural events also occur at 1-year-old toddler who sees a dog and says, this time, including the emergence of bilat- “dog,” for example, combines a perceptual erally synchronous electroencephalographic experience of a dog with a label from seman- (EEG) patterns of activation (bursts) and the tic long-term memory, effectively indicat- emergence of EEG patterns that distinguish ing, “That [i.e., the object of minC] is between sleep and wakefulness (Torres & a dog.” Similarly, pointing effectively indi- Anderson, 1985). Fetal behavior also changes cates, “That is that.” There must be two at this age in ways that may indicate the things, the experience and the label, for one onset of minC. For example, fetuses begin of them, the experience interpreted in terms to show clear heart rate increases to vibroa- of the label, to become an object of recC. coustic stimuli (Kisilevsky, Muir, & Low, Whereas the contents of minC are con- 1992), evidence of habituation to vibroa- tinually replaced by new perceptual stimu- coustic stimuli (Groome, Gotlieb, Neely, & lation, recC allows for conscious experience Waters, 1993), and sharp increases in cou- in the absence of perceptual stimulation. pling between their movement and their Because a label can be decoupled from the heart rate (DiPietro, Caulfield, Costigan, et experience labelled, the label provides an al., 2004). There are also good reasons to enduring trace of that experience that can believe that fetuses at this age are capable be deposited into both long-term memory of pleasure and pain (e.g., Anand & Hickey, and working memory. The contents of work- 1987; Lipsitt, 1986). So, on this account, we ing memory (e.g., representations of hidden agree with those who hold that conscious- objects) can then serve as explicit goals to ness first emerges during the third trimester trigger action programs indirectly so that the of fetal development, but we emphasize the toddler is no longer restricted to responses relatively simple nature of this initial level of triggered directly by minC of an immedi- consciousness. ately present stimulus. Now when objA trig- According to this account, attribution of gers IobjA (see Figure 15.1a) and becomes minC manages to explain infant behavior the content of minC, IobjA does not trigger until the end of the first year, when numer- an associated action program directly, but ous new abilities appear within just a few rather IobjA is fed back into minC (called months, from about 9 to 13 months of age. recC after one degree of reflection) where For example, during this period, most infants it can be related to a label (descA) from speak their first words, begin to use objects in semantic long-term memory. This descA a functional way, point proto-declaratively, can then be decoupled and deposited in and start searching in a more flexible way for working memory where it can serve as a hidden objects (e.g., passing Piaget’s A-not- goal (G1) that triggers an action program B task), among other milestones. According even in the absence of objA and even if to the LOC model, these changes can all be IobjA would otherwise trigger a different P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

418 the cambridge handbook of consciousness

action program. For example, when pre- Patients with lesions to striate cortex may sented with a telephone, the recC toddler deny that they can see anything in a partic- may activate a specific semantic association ular part of their visual field. Nonetheless, and put the telephone to her ear (functional if they are asked to guess, they are often play) instead of putting the telephone in her quite good at locating objects in that field mouth (a generic, stereotypical response). or even describing features of the objects. The toddler responds mediately to the label Perner and Dienes suggest that the normal in working memory rather than immediately healthy visual perception of objects involves to an initial, minC gloss of the situation. conscious awareness, whereas the impaired Despite these advances, recursively con- perception displayed by blindsight patients scious toddlers still cannot explicitly con- involves unconscious awareness. In terms of sider the relation between a means and an the LOC model, this distinction would seem end (e.g., Frye, 1991) and hence cannot fol- to map onto the distinction between minC low arbitrary rules (i.e., rules linking means and recC. and ends or conditions and actions). More- Given this distinction, Perner and Dienes over, although they are no longer exclusively (2003) then consider three behaviors for present-oriented, their experience of events which consciousness seems necessary in in time is limited because they have no way adults: verbal communication, executive to consider relations among two or more function (or voluntary control over action), explicit representations. As a result, they and explicit memory (i.e., conscious recol- cannot consider past- or future-oriented rep- lection). They note that most babies say resentations from the perspective of the their first words at about 12 to 13 months present (i.e., from the perspective of an of age and that the earliest signs of exec- explicit representation of the present, or utive function also appear at about this Now), because this would require an addi- age. They also argue, on the basis of work tional element to be represented (namely, with amnesic patients (McDonough, Man- a description of Now). Therefore, it should dler, McKee, & Squire, 1995), that delayed be impossible for these toddlers to appre- imitation requires explicit memory. Melt- ciate past or future representations as such zoff’s work (1985, 1988) suggests that infants because the concepts of both past and future first exhibit delayed imitation sometime are only meaningful when considered in rela- between 9 and 14 months (although see tion to a perception of the present circum- Meltzoff & Moore, 1994). stances. This situation is depicted in Figure In addition to these potential behavioral 15.4b. As shown in the figure, recursively indices of consciousness, Perner and Dienes conscious infants are no longer restricted to (2003) consider when children might be Now, but they cannot explicitly consider said to possess the cognitive capabilities events as occurring in the future from the that would allow them to entertain higher- perspective of the present. Similarly, they order thoughts about their experiences – cannot explicitly consider past events as consistent with higher-order thought theo- occurring in the past from the perspective ries of consciousness (e.g., Armstrong, 1968; of the present. Rosenthal, 1986; see Chapter 3). Higher- Perner and Dienes (2003) present an order thought theories claim that conscious- account of the emergence of consciousness ness consists in a belief about one’s psycho- (i.e., when children “become consciously logical states (i.e., a psychological state is aware of events in the world,” p. 64) conscious when one believes that one is in that seems to be congruent with this that state), which would seem to require account of recC. These authors first distin- a fairly sophisticated conceptual under- guish between “unconsciousness awareness” standing of one’s own mind. According to and “conscious awareness” and illustrate one version of higher-order thought the- the distinction in terms of blindsight (e.g., ory (Rosenthal, 2005), however, the relevant Weiskrantz, Sanders, & Marshall, 1974). higher-order thoughts may be relatively P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

the development of consciousness 419

simple thoughts that just happen to be about in front of me is a circle.” Rather, to be con- one’s psychological state. Perner and Dienes sciously aware of the circle, on this view, one observe that children start referring to their must represent one’s psychological attitude own mental states between about 15 and toward the factuality of the representation: 24 months of age, but they caution that “‘I see that [it’s a fact that (the object in front reliance on a verbal measure may lead us of me is a circle)]’” (p. 78). to underestimate the abilities of younger This version of a higher-order thought children. theory is hardly compelling from a devel- Another version of these theories (Car- opmental perspective, however. For exam- ruthers, 2000) holds fast to the suggestion ple, it is by no means inconceivable that a that children will not be conscious until they 3-year-old could be conscious of a fact are capable of meta-representation – in par- (“There are pencils in the Smarties box”) ticular appreciating the distinction between without being conscious of her attitude appearance and reality, or the notion of sub- (belief) or being conscious that she herself jective perspective on independent reality, as is entertaining the attitude toward the fact. assessed by measures of false belief under- Indeed, this is exactly what the LOC model standing. It is fairly well established that maintains: RecC allows for conscious expe- children do not understand these concepts riences that can persist in the absence of explicitly until about 4 years of age (e.g., perceptual stimulation, but this conscious Flavell, Flavell, & Green, 1983; Wellman, experience is still simpler than the com- Cross, & Watson, 2001), and it is on these plex conscious state described by Perner and grounds that Carruthers (2000) suggests that Dienes. From the perspective of the LOC children do not have consciousness until this model, a relatively high level of conscious- age. Perner and Dienes, however, raise the ness (see below), effected by several degrees intriguing possibility that perhaps higher- of reﬂection, is required to represent one’s order thoughts do not require an explicit psychological attitude toward a fact about understanding of subjective perspective, but an object. rather simply a procedural grasp of the As Perner and Dienes (2003) imply, the notion – as might be manifested in children’s developmental conclusions to be drawn on pretend play during the second year of life the basis of higher-order thought theories (e.g., Harris & Kavanaugh, 1993). are not entirely clear. These authors sug- Evidently, it remains unclear exactly what gest, however, that the balance of the evi- kinds of higher-order thoughts might be dence suggests that children become con- required for conscious experience. Perner sciously aware between 12 and 15 months and Dienes attempt to clarify this issue (plus or minus 3 months). In terms of the in terms of Dienes and Perner’s (1999) LOC model, the changes occurring in this framework of explicit knowledge, and they age range (or slightly earlier) do not corre- end up adopting a stance that resembles spond to the emergence of consciousness per Carruthers’ (2000) view more than Rosen- se, but they do correspond to an important thal’s (2000) view. According to Perner and developmental change in the character of Dienes (2003), “If one saw [an] object as experience – one that allows the (singular) a circle, but only unconsciously, one might contents of consciousness to be made avail- minimally represent explicitly only a feature, able to the child in more explicit fashion. e.g., ‘circle.’ A minimal representation, ‘cir- Consider again the example of long-distance cle’ would not provide conscious awareness driving. The difference between MinC and or conscious seeing since it does not qual- RecC is the difference between the ﬂeeting, ify as a full constituent of a higher order unrecoverable awareness of a stop sign that thought” (p. 77). Other explicit representa- is responded to in passing (without any tions that capture some fact about the object elaborative processing) and the recoverable also fail to qualify: “The object in front of awareness that occurs when one not only me is a circle” and “It is a fact that the object sees the stop sign but also labels it as such. P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

420 the cambridge handbook of consciousness

Although the neural correlates of the rior regions are coming to be controlled by behavioral changes at the end of the first frontal function. year are still relatively unknown, there are More recently, Baird and colleagues several reasons to believe that these changes (2002) used near infrared spectroscopy correspond to important developments in (NIRS) to compare blood flow in prefrontal prefrontal cortical function. For example, cortex in infants who reliably searched for in a pioneering study using positron emis- hidden objects (i.e., keeping a goal in mind) sion tomography (PET), Chugani and Phelps and those who did not. These authors found (1986) assessed resting glucose metabolism that infants who reliably searched for hid- in the brains of nine infants. Although there den objects showed an increase in frontal was activity in primary sensorimotor cortex blood flow after the hiding of the object, in infants as young as 5 weeks of age, and whereas those who failed to search showed there were increases in glucose metabolism a decrease. in other areas of cortex at about 3 months of Self-consciousness. Although a 12-month- age, it was not until about 8 months of age old behaves in a way that is considerably that increases were observed in prefrontal more controlled than, say, a 6-month-old, cortex. there is currently no convincing evidence As another example, Bell and Fox (1992) that children are explicitly self-conscious measured EEG activity longitudinally in (e.g., at Rochat’s third level of self- infants between 7 and 12 months of age and awareness) until midway through the sec- found a correlation between putative mea- ond year of life, at which point they begin sures of frontal function and performance use personal pronouns, first appear to recon Piaget’s A-not-B task. In the A-not-B ognize themselves in mirrors, and first dis- task, infants watch as an object is hidden play self-conscious emotions like shame and at one location (location A), and then they embarrassment (see Kagan, 1981, and Lewis are allowed to search for it. Then infants & Brooks-Gunn, 1979, for reviews). In the watch as the object is hidden at a new loca- famous mirror self-recognition paradigm, an tion (location B). When allowed to search, experimenter may surreptitiously put rouge many 9-month-old infants proceed to search on a toddler’s nose and then expose chil- at A, but performance on this task devel- dren to a mirror. It is well established that ops rapidly at the end of the first year of most children first exhibit mark-directed life and seems to provide a good measure of behavior in this situation between about 18 keeping a goal in mind and using it to con- and 24 months of age (e.g., Amsterdam, trol behavior despite interference from pre- 1972; Lewis & Brooks-Gunn, 1979), and this potent response tendencies (see Marcovitch has been taken to reflect the development & Zelazo, 1999, for a review). The putative of an objective self-concept (e.g., Lewis & measures of frontal function were frontal Brooks-Gunn, 1979) or the sense of a ‘me’ EEG power (in the infant “alpha” range; 6– as opposed to the sense of an ‘I’ (James, 9 Hz) and frontal/parietal EEG coherence. 1890/1950). Frontal EEG power reflects the amount of Kagan (1981) also noted the way in which cortical neuronal activity as measured at 2-year-olds respond when shown a complex frontal sites on the scalp, whereas EEG series of steps in the context of an imita- coherence reflects the correlation between tive routine. Kagan found that infants at this signals within a particular frequency band age (but not before) sometimes exhibited but measured at difference scalp sites. Bell signs of distress, as if they knew that the and Fox (1992, 1997) have suggested that series of steps was beyond their ken and was changes in power may be associated with not among the means that they had at their increased organization and excitability in disposal. According to Kagan, this implies frontal brain regions and that increases in that children are now able to consider coherence may indicate that more poste- their own capabilities (i.e., in the context P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

the development of consciousness 421

of a goal of imitating the experimenter). C1 → A1, from Figure 15.1b, is instantiated Consideration of a means relative to the as follows: Now; Tomorrow → EventA. goal that occasions it is a major advance When children can consider past or future that allows children consciously to follow events as such, they will have a subjective rules linking means to ends. According to experience of self-continuity in time. As a the LOC model, the further development of result, they should now be able to engage prefrontal cortex during the second year of in episodic recollection, which, accord- life allows children to engage in a higher level ing to Tulving (e.g., 1985), involves con- of consciousness – referred to as SelfC. This sciously recalling having experienced some- new level of consciousness is what allows thing in the past and so depends, by defi- children to use a single rule to guide their nition, on self-consciousness (or autonoetic behavior. [self-knowing] consciousness, in Tulving’s As shown in Figure 15.1b, a self-conscious terms). The close relation between the toddler can take as an object of con- changes in children’s self-consciousness that sciousness a conditionally specified self- are indexed, for example, by mirror self- description (SdescA) of their behavioral recognition, and the onset of episodic rec- potential – they can consider conditionally ollection has been noted by several authors specified means to an end. This SdescA can (e.g., Howe & Courage, 1997; Wheeler, then be maintained in working memory as 2000), although others, such as Perner a single rule (R1, including a condition, C, and Ruffman (1995), believe that genuine and an action, A), considered against the episodic recollection does not emerge until background of a goal (G1). Keeping a rule in later, coincident with changes in children’s working memory allows the rule to govern theory of mind (see below). responding regardless of the current environ- Although the changes occurring during mental stimulation, which may pull for inap- the second half of the second year are propriate responses. remarkable – so remarkable that Piaget Among the many changes in children’s (1936/1952) imagined they reflected the qualitative experience will be changes in emergence of symbolic thought – there con- their experience of themselves, and of them- tinues to be considerable room for develop- selves in time – changes in conscious recol- ment. For example, an additional degree of lection. Unlike recursively conscious infants, recursion is required for children to consider self-conscious children can now consider simultaneously two different events occur- descriptions of events as past- or future- ring at two different times (e.g., EventA, oriented, relative to a present experience further described as occurring Now, consid- (see Fig. 15.4c). For example, while con- ered in contradistinction to EventB, further scious of their current state (Now), 2-year- described as occurring Tomorrow). olds can appreciate that yesterday they went This characterization of the limitations to the zoo (Friedman, 1993). The concepts on 2-year-old children’s sense of them- yesterday and tomorrow are intrinsically rela- selves in time is similar in some respects tional because they are indexed with respect to that offered by Povinelli (1995, 2001) to today. Thus, for children to comprehend and McCormack and Hoerl (1999, 2001), that an event occurred yesterday (or will although it differs in others. Povinelli (1995) occur tomorrow), children must be con- suggests that, although children between scious of Now and consider two linked 18 and 24 months of age can pass mir- descriptions: a description of the event and ror self-recognition tasks (e.g., Amsterdam, a further description of the event as occur- 1972; Lewis & Brooks-Gunn, 1979), they ring yesterday (or tomorrow). Doing so cor- do not yet possess an objective and responds in complexity to the use of a sin- enduring self-concept. Instead, Povinelli gle rule considered against the backdrop of (1995) suggests that children at this age a goal that occasions its use. That is, G1; maintain a succession of present-oriented P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

422 the cambridge handbook of consciousness

representations of self (termed present eral, poor performance on tests of delayed selves; Povinelli, 1995,p.165), and they self-recognition does not necessarily indicate cannot compare these representations or an immature self-concept, although it may “integrate previous mental or physical states well reflect more general limitations on the with current ones” (p. 166). Consequently, highest level of consciousness that children their sense of self-continuity in time is tem- are able to adopt. porally restricted, and they might still be Reflective consciousness 1. The LOC model said to live in the present. In support of holds that, in contrast to 2-year-olds, 3-year- these claims, Povinelli and colleagues (1996) olds exhibit behavior that suggests an even found that even 3-year-olds perform poorly higher LOC, reflective consciousness 1 (refC1). on measures of delayed self-recognition. In For example, they can systematically employ their studies, children played a game dur- a pair of arbitrary rules (e.g., things that ing which an experimenter surreptitiously make noise vs. are quiet) to sort pictures – placed a sticker on their heads. About 3 min- behavior hypothesized to rely on lateral pre- utes later, children were presented with a frontal cortex. According to the model, 3- video image of the marking event. Whereas year-olds can now reflect on a SdescA of the majority of older children (4-year-olds) a rule (R1) and consider it in relation to reached up to touch the sticker, few of another Sdesc (SdescB) of another rule (R2). the younger children (2- and 3-year-olds) Both of these rules can then be deposited did so. into working memory where they can be Zelazo, Sommerville, and Nichols (1999) used contrastively (via a rule like E in Fig- argued that children perform poorly on mea- ure 15.2) to control the elicitation of action sures of delayed self-recognition not because programs. As a result, unlike 2-year-olds, they lack a subjective experience of self- 3-year-olds do not perseverate on a single continuity in time, but rather because they rule when provided with a pair of rules to have difficulty adjudicating between con- use (Zelazo & Reznick, 1991). flicting influences on their behavior (which Of course, there are still limitations on requires the use of higher-order rules and a 3-year-olds’ executive function, as seen higher level of consciousness). More specif- in their perseveration in the Dimensional ically, children in Povinelli et al.’s (1996) Change Card Sort (DCCS). In this task, chil- experiment have a strong expectation that dren are shown two bivalent, bidimensional they do not have a sticker on their head target cards (e.g., depicting a blue rabbit and (because they do not see it placed there a red boat), and they are told to match a and cannot see it directly at the time series of test cards (e.g., red rabbits and blue of testing). When they are provided with boats) to these target cards first according to conflicting information via a dimly under- one dimension (e.g., color) and then accord- stood representational medium (e.g., video; ing to the other (e.g., shape). Regardless of Flavell, Flavell, Green, & Korfmacher, 1990), which dimension is presented first, 3-year- the new, conflicting information may be olds typically perseverate by continuing to ignored or treated as somehow irrelevant sort cards by the first dimension after the rule to the situation. Empirical support for this is changed. In contrast, 4-year-olds seem to suggestion comes from a study showing know immediately that they know two dif- that although 3-year-olds can use delayed- ferent ways of sorting the test cards. Zelazo video (and delayed-verbal) representations et al. (2003) have argued that successful per- to guide their search for a hidden object formance on this task requires the formu- in the absence of a conflicting expectation lation of a higher-order rule like F in Fig- about the object’s location, they have diffi- ure 15.2 that integrates two incompatible culty doing so in the presence of a conflict- pairs of rules into a single structure. ing expectation (Zelazo et al., 1999, Exp. 3). Performance on measures such as the Because children have difficulty managing DCCS is closely related to a wide range conflicting delayed representations in gen- of metacognitive skills studied under the P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

the development of consciousness 423

rubric of theory of mind (e.g., Carlson & was possible at SelfC. For example, children Moses, 2001; Frye et al., 1995). In one can now conceptualize Now from a tempo- standard task, called the representational rally decentered perspective (McCormack & change task (Gopnik & Astington, 1988), Hoerl, 1999, 2001; see Figure 15.4d). From children are shown a familiar container (e.g., this perspective, children are now able to a Smarties box) and asked what it contains. consider two events occurring at two differ- Subsequently, the container is opened to ent times. For example, they can consider reveal something unexpected (e.g., string), that Now, EventA is occurring, but Yester- and children are asked to recall their ini- day, EventB occurred. tial incorrect expectation about its contents: This important developmental advance “What did you think was in the box before allows children to make judgments about I opened it?” To answer the representa- history (e.g., now vs. before). For example, in tional change question correctly, children a control task used by Gopnik and Astington must be able to recollect (or reconstruct) (1988, Exp. 1), most 3-year-olds were able their initial false belief. Most 3-year-olds to judge that Now there is a doll in a closed respond incorrectly, stating (for example) toy house but Before there was an apple. At that they initially thought that the box this level of consciousness, however, chil- contained string. dren cannot differentiate between the his- Three-year-old children’s difficulty on tory of the world and the history of the self. this type of task has proven remarkably That is, the objective series and the subjec- robust. Zelazo and Boseovksi (2001), for tive series remain undifferentiated; the two example, investigated the effect of video series are conflated in a single dimension. As reminders on 3-year-olds’ recollection of a result, 3-year-olds typically fail Gopnik and their initial belief in a representational Astington’s (1988) representational change change task. Children in a video support task, where they must appreciate that they condition viewed videotapes of their initial, themselves changed from thinking Smarties incorrect statements about the misleading to thinking string, even while the contents container immediately prior to being asked of the box did not change. According to to report their initial belief. For example, the LOC model, this failure to differentiate they watched a videotape in which they between the history of the world and the saw a Smarties box for the first time and history of the self occurs because children said, “Smarties.” They were then asked about who are limited to refC1 are only able to use the videotape and about their initial belief. a single pair of rules, which allows them to Despite correctly acknowledging what they make a discrimination within a single dimen- had said on the videotape, children typi- sion, but prevents them from making com- cally failed the representational change task. parisons between dimensions (e.g., between When asked what they initially thought was shape and color in the DCCS). in the box (or even what they initially said), Reflective consciousness 2 . Research has they answered, “String.” revealed that between 3 and 5 years of age, At 3 years of age, then, children are there are important changes in children’s able to consider two rules in contradistinc- executive function and theory of mind, tion (i.e., they can consider a single pair assessed by a wide variety of measures, and of rules) from a relatively distanced per- these changes tend to co-occur in individ- spective – even if they still cannot adopt ual children (e.g., Carlson & Moses, 2001; the level of consciousness required for such Frye et al., 1995; Perner & Lang, 1999). For measures as the DCCS and the representa- the measures of executive function and the- tional change task. The relatively psycholog- ory of mind that show changes in this age ically distanced perspective made possible range, children need to understand how two by RefC1 and the consequent increase in the incompatible perspectives are related (e.g., complexity of children’s rule representations how it is possible to sort the same cards first allow for a richer qualitative experience than by shape and then by color; how it is possible P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

424 the cambridge handbook of consciousness

for someone to believe one thing when I Behavioral evidence of children’s ability know something else to be true). According to differentiate and yet coordinate the his- to the LOC model, this understanding is tory of the self and the history of the world made possible by the further growth of pre- can be seen in 4- and 5-year-olds’ success frontal cortex and the development of the on Gopnik and Astington’s (1988) repre- ability to muster a further level of conscious- sentational change task. In this task, chil- ness – reflective consciousness 2 (refC2). At dren now appreciate that they themselves refC2, the entire contents of refC1 can be changed from thinking Smarties to think- considered in relation to a Sdesc of compara- ing string, but that the contents of the box ble complexity. This perspective allows chil- did not change. Thus, against the backdrop dren to formulate a higher-order rule that of Now, children appreciate the history of integrates the two incompatible perspectives the world, on the one hand; that is, they (e.g., past and present self-perspectives in appreciate that in the past, EventA (string the representational change task or color in the box) occurred and Now, EventA is vs. shape rules in the DCCS) into a sin- still occurring. However, they also appreci- gle coherent system and makes it possible ate the history of the self, on the other hand: to select the perspective from which to rea- In the past, EventA (believed Smarties in the son in response to a given question. (In the box) occurred, and Now, EventB is occurring absence of the higher-order rule, children (believe string in the box). will respond from the prepotent perspec- Because refC2 allows children to inte- tive.) In terms of Figure 15.3, RefC2 allows grate two incompatible pairs of rules within children to formulate and use a rule like F. a single system of rules, it allows them to Being able to reflect on their discrimina- understand that they can conceptualize a tions within a dimension (e.g., shape) and single thing in two distinct ways. For exam- considering two (or more) dimensions in ple, they understand that they can concep- contradistinction, allows children to conceptualize a red rabbit as a red thing and as a tualize dimensions qua dimensions (see also rabbit in the DCCS, and they understand Smith, 1989). In terms of their understand- that they can acknowledge that a sponge ing of the self in time, this ability to con- rock looks like a rock even though it is really sider dimensions qua dimensions (or series a sponge (Flavell, Flavell, & Green, 1983). qua series) allows children to differentiate When applied to time, this understanding and coordinate two series, the history of the permits children potentially to appreciate self and the history of the world, from a multiple temporal perspectives on the same temporally decentered perspective (see Fig- event (e.g., that time present is time past ure 15.4e). As Bieri (1986) notes, to have a in time future). This acquisition, at about genuine temporal awareness, one must diffe- 4 or 5 years of age, corresponds to the major rentiate the progression of the self from the developmental change identified in McCor- progression of events in the world and then mack and Hoerl’s (1999) account of tempo- understand the former relative to the lat- ral understanding: At a higher level of tem- ter. (The latter corresponds to the objective poral decentering, children appreciate that series, which ultimately serves as the unify- multiple temporal perspectives are perspec- ing temporal framework.) In Bieri’s (1986, tives onto the same temporal reality, and p. 266, italics in the original) words: “In they acquire the concept of particular times order to have a genuine temporal awareness, (i.e., that events occur at unique, particular a being must be able to distinguish between times). the history of the world and the history of its Children’s ability to conceptualize a sin- encounters with this world. And the contin- gle thing in multiple ways can also be applied uously changing temporal perspective . . . is to their understanding of themselves in time, nothing but the continuous process of con- where it allows children potentially to connecting these two series of events within a ceptualize themselves from multiple tempo- representation of one unified time.” ral perspectives – to understand themselves P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

the development of consciousness 425

as exhibiting both continuity and change spontaneously, they should be capable of in time. Muller¨ and Overton (1998) dis- comprehending two perspectives on a sin- cuss this understanding in terms of Stern’s gle item (as indicated, e.g., by successful per- (1934/1938) notion of mnemic continuity:“I formance on the Dimensional Change Card am the same one who now remembers what Sort and a variety of measures of perspec- I then experienced” (p. 250; italics in the tive taking (Carlson & Moses, 2001; Frye original), and they note that Stern described et al., 1995). Therefore, the model predicts this understanding as emerging around the that asking 4-year-old children to label their fourth year of life. perspective on Selection 1 (e.g., “Why do Work with children during this period those two pictures go together?”) should of development – the transition to refC2 – cause them to make that subjective per- has been useful in revealing one of the key spective an object of consciousness, neces- roles that language can play in fostering the sarily positioning them at a higher level of adoption of higher levels of consciousness; consciousness from which it is possible to namely, that it can promote reflection within reflect on their initial perspective. From this developmental constraints on the highest higher level of consciousness (i.e., the per- level of consciousness that children are able spective of Rule F in Figure 15.2), it should to obtain. In particular, labeling one’s sub- be easier to adopt a different perspective on jective experiences helps make those expe- Selection 2, which is exactly what Jacques riences an object of consideration at a higher et al. (2006) found. This was true whether level of consciousness. Increases in level of children provided the label themselves or consciousness, in turn, allow for the flexible whether the experimenter generated it selection of perspectives from which to rea- for them. son. Therefore, for children who are capable In general, the adoption of a higher-order in principle of adopting a particular higher perspective allows for both greater influ- level of consciousness, labeling perspectives ence of conscious thought on language and at the next lower level will increase the like- greater influence of language on conscious lihood that they will in fact adopt this higher thought. On the one hand, it allows for more level of consciousness, facilitating cognitive effective selection and manipulation of rules flexibility. (i.e., it permits the control of language in the The effect of labeling on levels of con- service of thought). On the other hand, it sciousness and flexibility can be illustrated permits children to respond more appropri- by work by Jacques, Zelazo, Lourenco, and ately to linguistic meaning despite a mislead- Sutherland (2007), using the Flexible Item ing context – allowing language to influence Selection Task (see also Jacques & Zelazo, thought. An example comes from a recent 2005). On each trial of the task, children study of 3-to5-year-olds’ flexible under- are shown sets of three items designed so standing of the adjectives “big” and “little” one pair matches on one dimension, and a (Gao, Zelazo, & DeBarbara, 2005). When different pair matches on a different dimen- shown a medium-sized square together with sion (e.g., a small yellow teapot, a large yel- a larger one, 3-year-olds had little diffi- low teapot, and a large yellow shoe). Chil- culty answering the question, “Which one dren are first told to select one pair (i.e., of these two squares is a big one?” How- Selection 1), and then asked to select a dif- ever, when the medium square was then ferent pair (i.e., Selection 2). To respond paired with a smaller one, and children were correctly, children must represent the pivot asked the same question, only 5-year-olds item (i.e., the large yellow teapot) according reliably indicated that the medium square to both dimensions. Four-year-olds generally was now the big one. This example shows perform well on Selection 1 but poorly on an age-related increase in children’s sensi- Selection 2, indicating inflexibility (Jacques tivity to linguistic meaning when it con- & Zelazo, 2001). According to the LOC flicts with children’s immediate experience, model, although 4-year-olds may not do so and it reveals that interpretation becomes P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

426 the cambridge handbook of consciousness

decoupled, to some degree, from stimulus muster, and each level has distinct conse- properties. quences for the quality of subjective expe- Another example of the same phe- rience, the potential for episodic recollec- nomenon comes from a study by Deak´ tion, the complexity of children’s explicit (2000), who examined 3-to6-year-olds’ knowledge structures, and the possibility of use of a series of different predicates (“looks the conscious control of thought, emotion, like a . . . ,” “is made of . . . ,” or “has a . . . ”) and action. Higher levels of consciousness to infer the meanings of novel words. He in this hierarchical model are brought about found that 3-year-olds typically used the first by the iterative reprocessing of the contents predicate appropriately to infer the mean- of consciousness via thalamocortical cir- ing of the first novel word in a series, but cuits involving regions of prefrontal cortex. then proceeded to use that same predicate Each degree of reprocessing recruits another to infer the meanings of subsequent words region of prefrontal cortex and results in a despite what the experimenter said. In con- higher level of consciousness, and this in turn trast, older children used the most recent allows for a stimulus to be considered rela- predicate cues. Again, children are increas- tive to a larger interpretive context. ingly likely to use language to restrict their This model aims to provide a compre- attention to the appropriate aspects of a sit- hensive account of the development of con- uation (or referent). sciousness in childhood that addresses extant Notice that language and conscious data on the topic and establishes a frame- thought become increasingly intertwined in work from which testable predictions can be a complex, reciprocal relation, as Vygotsky derived. Naturally, however, there is consid- (1934/1986) observed. Thus, language (e.g., erable work to be done. Among the many labeling) influences thought (e.g., by pro- questions for future research, a few seem moting a temporary ascent to a higher level particularly pressing. First, future research of consciousness), which in turn influences will need to explore the possibility that there language, and so on. This reciprocal rela- are further increases in children’s highest tion can be seen in the growing richness level of consciousness beyond the refC2 level of children’s semantic understanding and in identified in the LOC model. Compared to the increasing subtlety of their word usage. early childhood, relatively little is known Consider, for instance, children’s developing about the development of consciousness in understanding of the semantics of the verb adolescence, although it is clear that the hit. Children first understand hit from its conscious control of thought, action, and use to depict simple accidental actions (e.g., emotion shows considerable development in an utterance by a child at 2;4.0: Table hit adolescence. Indeed, to the extent that these head; Gao, 2001,pp.220). Usage is initially functions are dependent on prefrontal cor- restricted to particular contexts. Eventually, tex, which continues to develop into adult- however, reflection on this usage allows chil- hood (e.g., Giedd et al., 1999), further age- dren to employ the word in flexible and cre- related increases in the highest level of conative ways (e.g., I should hit her with a pencil sciousness seem likely. and a stick uttered metaphorically by the Second, future research should continue same child at 3;8.6; Gao, 2001,pp.219). to search for more precise neural markers of the development of consciousness. Among the possible indices are increases in neural coherence, dimensional complexity, Summary and Topics for and/or the amount or dominant frequency Future Research of gamma EEG power. Such increases could be associated with the binding together of According to the LOC model, there are at the increasingly complex hierarchical net- least four age-related increases in the high- works of prefrontal cortical regions that we est level of consciousness that children can have proposed are associated with higher P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

the development of consciousness 427

levels of consciousness. Dimensional com- enter a more coherent (coordinated) hierar- plexity (DCx), for example, is a non- chy of levels of consciousness. linear measure of global dynamical complexity (for review see Anokhin et al., 2000 ) that can be derived from EEG data. Conclusion In a cross-sectional study of children and adults (ages 7 to 61 years), Anokhin et al. Discussions regarding the development of (1996) found that whereas raw alpha and consciousness have focused on the question theta power only changed until early adult- of when consciousness emerges, with differ- hood, structural DCx continued to increase ent authors relying on different notions of across the life span. Other research indicates consciousness and different criteria for deter- that DCx may show particularly promi- mining whether consciousness is present. In nent increases during adolescence (Anokhin this chapter, we presented a comprehensive et al., 2000; Farber & Dubrovinskaya, 1991; model of consciousness and its development Meyer-Lindenberg, 1996). In a study com- that we believe helps clarify the way dif- paring children and adolescents (mean ages, ferent aspects of consciousness do indeed 7.5, 13.8, 16.4 years), Anokhin et al. (2000) emerge at different ages. Our hope is that found that both resting and task-related this model provides a useful framework for complexity (in visual and spatial cognitive thinking about consciousness as a complex, tasks) increased with age, as did the differ- dynamic phenomenon that is closely tied to ence between resting and task-related DCx. neural function, on the one hand, and cog- Finally, although formulated to explain nitive control, on the other. developmental data, the LOC model suggests a framework for understanding the vagaries of human consciousness across the life span, and future research should explore Acknowledgments the extent to which this framework is useful for understanding the role of conscious- Preparation of this article was supported ness in adult behavior. One application is by grants from the Natural Sciences and to research on mindfulness (e.g., Brown & Engineering Research Council (NSERC) of Ryan, 2003; see Chapter 19). Acting mind- Canada and the Canada Research Chairs fully (and “super-intending” one’s behavior) Program. may involve adopting higher levels of consciousness and coordinating these levels so that they are all focused on a single thing – References a single object of consciousness. This coordination of levels of consciousness on a sin- 1972 gle object would result in an experience that Amsterdam, B. ( ). Mirror self-image reac- differs dramatically from the kind of multi- tions before age two. Developmental Psychobi- ology, 5, 297–305. tasking observed, for example, when driving 2005 a car at the level of MinC but carrying on a Anand, K. J. S. ( ). A scientiﬁc appraisal of fetal pain and conscious sensory perception. Report conversation at a higher level of conscious- to the Constitution Subcommittee of the U.S. ness. Conceptualising mindfulness in terms House of Representatives (109th United States of the LOC model yields predictions regard- Congress). ing the effects of mindfulness meditation Anand, K. J. S., & Hickey, P. R. (1987). Pain and on behavior (e.g., attentional control) and its effects in the human neonate and fetus. New neural function (e.g., increasingly elaborated England Journal of Medicine, 317, 1321–1329. hierarchies of prefrontal cortical regions). Anokhin, A. P., Birnbaumer, N., Lutzenberger, From this perspective, mindfulness medita- W., Nikolaev, A., & Vogel, F. (1996). Age tion practice can be seen as a type of training increases brain complexity. Electroencephalog- that may increase an individual’s ability to raphy and Clinical Neurophysiology, 99, 63–68. P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

428 the cambridge handbook of consciousness

Anokhin, A. P., Vedeniapin, A. B., Sirevaag, E. J., neuroscience. Cognitive, Affective, and Behav- Bauer, L. O., O’Conner, S. J., Kuperman, S., ioral Neuroscience, 4, 564–579. Porjesz, B., Reich, T., Begleiter, H., Polich, J., Bunge, S. A., & Zelazo, P. D. (2006). A brain- & Rohrbaugh, J. W. (2000). The P300 brain based account of the development of rule use in potential is reduced in smokers. Psychopharma- childhood. Current Directions in Psychological cology, 149, 409–413. Science, 15, 118–121. Armstrong, D. M. (1968). A materialist theory of Burgess, J. A., & Tawia, S. A. (1996). When did the mind. London: Routledge. you ﬁrst begin to feel it? Locating the beginning Armstrong, D. M. (1980). The nature of mind of consciousness. Bioethics, 10, 1–26. and other essays. Ithaca, NY: Cornell Univer- Cahan, E. (1984). The genetic psychologies of sity Press. James Mark Baldwin and Jean Piaget. Devel- Baird, A. A., Kagan, J., Gaudette, T., Walz, opmental Psychology, 20, 128–135. K. A., Hershlag, N., & Baos, D. A. (2002). Carlson, S. M., Davis, A. C., & Leach, J. G. Frontal lobe activation during object perma- (2005). Less is more: Executive function and nence: Data from near-infrared spectroscopy. symbolic representation in preschool children. NeuroImage, 16, 1120–1126. Psychological Science, 16, 609–616. Baldwin, J. M. (1892). Origin of volition in child- Carlson, S. M., & Moses, L. J. (2001). Individual hood. Science 20, 286–288. differences in inhibitory control and theory of Baldwin, J. M. (1894). Imitation: A chapter in the mind. Child Development 72, 1032–1053. natural history of consciousness. Mind, 3, 25– Carruthers, P. K. (1989). Brute experience. Jour- 55. nal of Philosophy, 86, 258–269. Baldwin, J. M. (1897). Social and ethical interpre- Carruthers, P. K. (1996) Language, thought, and tations in mental development: A study in social consciousness: An essay in philosophical psychol- psychology. New York: Macmillan. ogy.New York: Cambridge University Press. Baldwin, J. M. (1906). Thought and things: A study Carruthers, P. K. (2000). Phenomenal conscious- of the development and meaning of thought, or ness: A naturalistic theory. New York: Cam- genetic logic, Vol. 1. London: Swan Sonnenschein bridge University Press. & Co. Chalmers, D. J. (1996). The conscious mind. Barresi, J., & Moore, C. (1996). Intentional rela- Oxford: Oxford University Press. tions and social understanding. Behavioral and Chugani, H. T., & Phelps, M. E. (1986). Mat- 104 154 Brain Sciences, 19, – . urational changes in cerebral function in Bell, M. A., & Fox, N. A. (1992). The relations infants determined by 18 FDG positron emission between frontal brain electrical activity and tomography. Science, 231, 840–843. cognitive development during infancy. Child Cleeremans, A., & Jimenez,´ L. (2002). Implicit 1142 1163 Development, 63, – . learning and consciousness: A graded, dynamic Bell, M. A., & Fox, N. A. (1997). Individual differ- perspective. In R. M. French & A. Cleeremans ences in object permanence performance at 8 (Eds.), Implicit learning and consciousness: An months: Locomotor experience and brain elec- empirical, philosophical and computational con- trical activity. Developmental Psychology, 31, sensus in the making (pp. 1–40). Hove, England: 287–297. Psychology Press. Bieri, P. (1986). Zeiterfahrung und Personalitat.¨ Cohen, L. B. (1998). An information-processing In H. Burger (Ed.), Zeit, natur und mensch approach to infant perception and cognition. In (pp. 261–281). Berlin: Arno Spitz Verlag. F. Simion & G. Butterworth (Eds.), The devel- Brentano, F. (1973). Psychology from an empirical opment of sensory, motor, and cognitive capaci- standpoint. London: Routledge & Kegan Paul. ties in early infancy (pp. 277–300). East Sussex: (Original work published 1874) Psychology Press. Brown, K. W., & Ryan, R. M. (2003). The beneﬁts Craik, F., & Lockhart, R. (1972). Levels of proof being present: The role of mindfulness in cessing: A framework for memory research. psychological well-being. Journal of Personality Journal of Verbal Learning and Verbal Behavior, and Social Psychology, 84, 822–848. 11, 671–684. Bunge, S. A. (2004). How we use rules to select Damasio, A. R. (1999). The feeling of what hap- actions: A review of evidence from cognitive pens. New York: Harcourt Press. P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

the development of consciousness 429

Deak,´ G. O. (2000). The growth of flexi- Freud, S. (1940). Neue folge der Vorlesungen ble problem-solving: Preschool children use zur Einfuhrung¨ in die Psychoanalyse [New changing verbal cues to infer multiple word introductory lectures on psychoanalysis]. In A. meanings. Journal of Cognition and Develop- Freud, E. Bibring, & E. Kris (Eds.), Gesammelte ment, 1, 157–192. werke: XV (Whole volume). London: Imago DeCasper, A., Lecanuet, J-P., Busnel, M-C., Publishing. (Original work published 1933) Granier-Deferre, C., & Mangeais, R. (1994). Friedman, W. J. (1993). Memory for the time of Fetal reactions to recurrent maternal speech. past events. Psychological Bulletin, 113, 44–66. Infant Behavior and Development, 17, 159–164. Frye, D. (1991). The origins of intention in infancy. Dehaene, S., & Changeux, J.-P. (2004). Neu- In D. Frye & C. Moore (Eds.), Children’s theories ral mechanisms for access to consciousness. In of mind: Mental states and social understanding M. Gazzaniga (Ed.), The cognitive neurosciences (pp. 15–38). Hillsdale, NJ: Erlbaum. (3rd ed., pp. 1145–1157). Cambridge, MA: MIT Frye, D., Zelazo, P. D., & Palfai, T. (1995). The- Press. ory of mind and rule-based reasoning. Cognitive Dewey, J. (1896). Review of studies in the evo- Development, 10, 483–527. lutionary psychology of feeling. Philosophical Gallagher, S. (2005). How the body shapes 292 299 Review, 5, – . the mind. Oxford: Oxford University Press/ Dewey, J. (1985). Context and thought. In J. A. Clarendon Press. Boydston (Ed.) & A. Sharpe (Textual Ed.), John Gao, H. (2001). The physical foundation of the pat- 61931 Dewey: The later works, 1925–1953 (Vol. – terning of physical action verbs. Lund, Sweden: 1932 3 21 ,pp. – ). Carbondale, IL: Southern Illi- Lund University Press. nois University Press. (Original work published Gao, H. H., Zelazo, P. D., & DeBarbara, K. 1931) (2005, April). Beyond early linguistic compe- 1999 Dienes, Z., & Perner, J. ( ). A theory of tence: Development of children’s ability to inter- implicit and explicit knowledge (target arti- pret adjectives flexibly. Paper presented at the 735 cle). Behavioral and Brain Sciences, 22, – 2005 Biennial Meeting of Society for Research 755 . in Child Development, Atlanta, GA. DiPietro, J. A., Caulfield, L. E., Costigan, K. A., Giedd, J. N., Blumenthal, J., Jeffries, N. O., Merialdi, M., Nguyen, R. H., Zavaleta, N., & Castellanos, F. X., Liu. H, Zijdenbos, A., Paus, 2004 Gurewitsch, E. D. ( ). Fetal neurobehav- T., Evans, A. C., & Rapoport, J. L. (1999). Brain ioral development: A tale of two cities. Devel- development during childhood and adoles- 445 456 opmental Psychology, 40, – . cence: adolescence: A longitudinal MRI study. Edelman, G. M. (1989). Neural Darwinism: The Nature Neuroscience, 2 , 861–863. theory of group neuronal selection. Oxford Uni- Gogtay, N., Giedd, J. N., Lusk, L., Hayashi, K. M., versity Press, Oxford. Greenstein, D., Vaituzis, A. C., et al. (2004). Edelman, G. M., & Tononi, G. (2000). A universe Dynamic mapping of human cortical develop- of consciousness. New York: Basic Books. ment during childhood through early adult- Farber, D. A., & Dubrovinskaya, N. V. (1991). hood. Proceedings of the National Academy of Organization of developing brain functions: SciencesUSA, 101(21), 8174–8179. Age-related differences and some general prin- Gopnik, A., & Astington, J. W. (1988). Children’s ciples. Human Physiology, 19, 326–335. understanding of representational change and Ferrari, M., Pinard, A., & Runions, K. (2001). its relation to the understanding of false belief Piaget’s framework for a scientific study of con- and the appearance-reality distinction. Child sciousness. Human Development, 44, 195–213. Development, 59, 26–37. Flavell, J. H., Flavell, E. R., & Green, F. L. (1983). Groome, L. J., Gotlieb, S. J., Neely, C. L., Development of the appearance-reality dis- Waters, M. D. (1993). Developmental trends in tinction. Cognitive Psychology, 17, 99–103. fetal habituation to vibroacoustic stimulation. 46 49 Flavell, J. H., Flavell, E. R., Green, F. L., & American Journal of Prerinatology, 10, – . Korfmacher, J. E. (1990). Do young children Haith, M. M., Hazan, C., & Goodman, G. S. think of television images as pictures or real (1988). Expectation and anticipation of dyna- objects?Journal of Broadcasting and Electronic mic visual events by 3.5-month-old babies. Media, 34, 399–419. Child Development, 59, 467–479. P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

430 the cambridge handbook of consciousness

Harris, P. L., & Kavanaugh, R. D. (1993). Young systematic multidisciplinary review of the evi- children’s understanding of pretence. Mono- dence. Journal of the American Medical Associ- graphs of the Society for Research in Child Devel- ation, 294, 947–954. opment (Serial No. 237). Lewis, M. (2003). The development of self- Howe, M., & Courage, M. (1997). The emer- consciousness. In J. Roessler & N. Eilan gence and early development of autobiograph- (Eds.), Agency and self-awareness (pp. 275– ical memory. Psychological Review 104, 499– 295). Oxford: Oxford University Press. 523. Lewis, M., & Brooks-Gunn, J. (1979). Social cog- Hrbek, A., Karlberg, P., & Olsson, T. (1973). nition and the acquisition of self. New York: Development of visual and somatosensory Plenum. responses in pre-term newborn infants. Elec- Lipsitt, L. P. (1986). Toward understanding the troencephalography and Clinical Neurophysiol- hedonic nature of infancy. In L. P.Lipsitt & J. H. ogy, 34, 225–232. Cantor (Eds.), Experimental child psychologist: Jacques, S., & Zelazo, P. D. (2001). The flexible Essays and experiments in honor of Charles C. item selection task (FIST): A measure of exec- Spiker (pp. 97–109). Hillsdale, NJ: Erlbaum. utive function in preschoolers. Developmental Luria, A. R. (1959). The directive function of Neuropsychology, 20, 573–591. speech in development and dissolution. Part I. Jacques, S., & Zelazo, P. D. (2005). Language and Development of the directive function of the development of cognitive flexibility: Impli- speech in early childhood. Word, 15, 341– cations for theory of mind. In J. W. Astington & 352 J. A. Baird (Eds.), Why language matters for the- Luria, A. R. (1961). Speech and the regulation of ory of mind (pp. 144–162). New York: Oxford behaviour. London: Pergamon Press. University Press. Marcovitch, S., & Zelazo, P. D. (1999) The A-not- Jacques, S., Zelazo, P. D., Lourenco, S. F., & B error: Results from a logistic meta-analysis. Sutherland, A. E. (2007). The roles of label- Child Development, 70, 1297–1313. ing and abstraction in the development of McCormack, T., & Hoerl, C. (1999). Memory cognitive flexibility. Manuscript submitted for and temporal perspective: The role of temporal publication. frameworks in memory development. Develop- James, W. (1950). The principles of psychology mental Review, 19, 154–182. 1 (Vol. ). New York: Dover. (Original work pub- McCormack, T., & Hoerl, C. (2001). The child 1890 lished ) in time: Episodic memory and the concept of Kagan, J. (1981). The second year: The emergence of the past. In C. Moore & K. Lemmon (Eds.), self-awareness. Cambridge, MA: Harvard Uni- Self in time: Developmental issues (pp. 203–227). versity Press. Mahwah, NJ: Erlbaum. Kagan, J. (1998). Three seductive ideas. Cam- McDonough, L., Mandler, J. M., McKee, R. D., bridge, MA: Harvard University Press. & Squire, L. R. (1995). The deferred imita- Karmiloff-Smith, A. (1992). Beyond modularity: tion task as a nonverbal measure of declarative A developmental perspective on cognitive science. memory. Proceedings of the National Academy Cambridge, MA: MIT Press. of Sciences USA, 92, 7580–7584. Kisilevsky, B. S., Hains, S. M. J., Lee, K., Xie, Meltzoff, A. N. (1985). Immediate and deferred X., Huang, H., Ye, H. H., Zhang K., & Wang, imitation in 14- and 24-month-old infants. Z. (2003). Effects of experience on fetal voice Child Development, 56, 62–72. recognition. Psychological Science, 14, 220–224. Meltzoff, A. N. (1988). Infant imitation and Kisilevsky, B. S., Muir, D. W., & Low, J. A. memory: Nine month olds in immediate and (1992). Maturation of human fetal responses to deferred tests. Child Development, 59, 217–225. vibroacoustic stimulation. Child Development, Meltzoff, A. (2002). Imitation as a mechanism 63, 1497–1508. of social cognition: Origins of empathy, the- Klimach, V. J., & Cooke, R. W. I. (1988). Matu- ory of mind, and the representation of action. ration of the neonatal somatosensory evoked In U. Goswami (Ed.), Handbook of child- response in preterm infants. Developmental hood cognitive development (pp. 6–25). London: Medicine & Child Neurology, 30, 208–214. Blackwell. Lee, S. J., Ralston, H. J. P., Drey, E. A., Partridge, Meltzoff, A. N., & Moore, M. K. (1994). Imi- J. C., & Rosen, M. A. (2005). Fetal pain: A tation, memory, and the representation of P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

the development of consciousness 431

persons. Infant Behavior and Development, 17, Povinelli, D. J. (1995). The unduplicated self. 83–99. In P. Rochat (Ed.), The self in infancy: The- Meyer-Lindenberg, A. (1996). The evolution of ory and research (pp. 161–192). New York: complexity in human brain development: An Elsevier. EEG study. Electroencephalography and Clinical Povinelli, D. J. (2001). The Self: Elevated in con- Neurophysiology, 99, 405–411. sciousness and extended in time. In C. Moore Morin, A. (2004, August). Levels of conscious- & K. Lemmon (Eds.), The self in time: Devel- ness. Science & Consciousness Review, 2. opmental perspectives (pp. 73–94). New York: Morin, A. (2006). Levels of consciousness and Cambridge University Press. self-awareness: A comparison and integration Povinelli, D. J., Landau, K. R., & Perilloux, H. K. of various neurocognitive views. Consciousness (1996). Self-recognition in young children and Cognition, 15, 358–371. using delayed versus live feedback: Evidence Moscovitch, M. M. (1989) Confabulation and of a developmental asynchrony. Child Develop- 1540 1554 the frontal systems: Strategic versus associa- ment, 67, – . tive retrieval in neuropsychological theories of Reznick, J. S. (1994). In search of infant expec- memory. In H. L. Roediger & F. I. M. Craik tation. In M. Haith, J. Benson, B. Pennington, (Eds.), Varieties of memory and consciousness: & R. Roberts (Eds.), The development of future- Essays in honour of Endel Tulving (pp. 133–160). oriented processes (pp. 39–59). Chicago: Uni- Mahwah, NJ: Erlbaum. versity of Chicago Press. Muller,¨ U., & Overton, W.F. (1998). How to grow Rochat, P. (2001). The infant’s world. Cambridge, a baby: A reevaluation of image-schema and MA: Harvard University Press. Piagetian action approaches to representation. Rochat, P. (2003). Five levels of self-awareness Human Development, 41, 71–111. as they unfold early in life. Consciousness and Nagel, T. (1974). What is it like to be a bat? Cognition, 12, 717–731. The Philosophical Review, 83, 435–450. Rochat, P., & Morgan, R. (1995). Spatial O’Donnell, S., Noseworthy, M. D., Levine, B., & determinants in the perception of self- Dennis, M. (2005). Cortical thickness of the produced leg movements in 3-to5-month- frontopolar area in typically developing chil- old infants. Developmental Psychology, 31, 626– dren and adolescents. NeuroImage, 24, 948– 636. 954. Rochat, P., & Striano, T. (2000). Perceived self in Perner, J., & Dienes, Z. (2003). Developmental infancy. Infant Behavior and Development, 23, aspects of consciousness: How much of a the- 513–530. ory of mind do you need to be consciously Rosenthal, D. M. (1986). Two concepts of con- aware?Consciousness and Cognition, 12, 63– sciousness. Philosophical Studies, 49, 329–359. 82 . Rosenthal, D. (2000). Consciousness, content, Perner, J., & Lang, B. (1999). Development of the- and metacognitive judgements. Consciousness ory of mind and executive control. Trends in and Cognition, 9, 203–214. 337 344 Cognitive Sciences, 3, – . Rosenthal, D. M. (2005). Consciousness, inter- Perner, J., & Ruffman, T. (1995). Episodic mem- pretation, and higher-order thought. In P. ory and autonoetic consciousness: Develop- Giamperie-Deutsch (Ed.), Psychoanalysis mental evidence and a theory of childhood as an empirical, interdisciplinary science: amnesia. Journal of Experimental Child Psychol- Collected papers on contemporary psychoan- ogy, 59, 516–548. alytic research (pp. 119-142). Vienna: Verlag Perruchet, P., & Vinter, A. (2002). The self- der Osterreichischen¨ Akademie der Wis- organizing consciousness. Behavioral and Brain senschaften (Austrian Academy of Sciences Sciences, 25, 297–388. Press). Piaget, J. (1952). The origins of intelligence in chil- Schacter, D. L. (1989). On the relation between dren. (M. Cook, Trans.). New York: Vintage. memory and consciousness: Dissociable inter- (Original work published 1936) actions and conscious experience. In H. L. Piaget, J. (1977). The grasp of consciousness (S. Roediger & F. I. M. Craik (Eds.), Varieties of Wedgewood, Trans.). Cambridge, MA: Har- memory and consciousness: Essays in honour vard University Press. (Original work pub- of Endel Tulving (pp. 355–389). Mahwah, NJ: lished 1974.) Erlbaum. P1: KAE 0521857430c15 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:39

432 the cambridge handbook of consciousness

Schooler, J. W. (2002). Re-presenting conscious- Wellman, H. M., Cross, D., & Watson, J. (2001). ness: dissociations between experience and Meta-analysis of theory-of-mind development: meta-consciousness. Trends in Cognitive Sci- The truth about false belief. Child Development ences, 6, 339–344. 72, 655–684. Sigel, I. (1993). The centrality of a distancing Wheeler, M. (2000). Varieties of consciousness model for the development of representational and memory in the developing child. In E. Tulv- competence. In R. R. Cocking & K. A. Ren- ing (Ed.), Memory, consciousness, and the brain: ninger (Eds.), The development and meaning of The Tallinn conference (pp. 188–199). London: psychological distance (pp. 91–107). Hillsdale, Psychology Press. NJ: Erlbaum. Zelazo, P. D. (1999). Language, levels of con- Siqueland, E. R., & Lipsitt, L. P. (1966). Condi- sciousness, and the development of intentional tioned head-turning in human newborns. Jour- action. In P. D. Zelazo, J. W. Astington, & D. R. nal of Experimental Child Psychology, 3, 356– Olson (Eds.), Developing theories of intention: 376. Social understanding and self-control (pp. 95– Smith, L. B. (1989). From global similarities 117). Mahwah, NJ: Erlbaum. to kinds of similarities: The construction of Zelazo, P. D. (2004). The development of con- dimensions in development. In S. Vosriadou & scious control in childhood. Trends in Cognitive A. Ortony (Eds.), Similarity and analogical rea- Sciences, 8, 12–17. 146 178 soning (pp. – ). Cambridge: Cambridge Zelazo, P. D., & Boseovski, J. (2001). Video University Press. reminders in a representational change task: Stern, D. (1990). Diary of a baby. New York: Basic Memory for cues but not beliefs or statements. Books. Journal of Experimental Child Psychology, 78, Stern, W. (1938). General psychology from the 107–129. personalistic standpoint. New York: Macmillan. Zelazo, P. D., & Jacques, S. (1996). Children’s (Original work published 1934.) rule use: Representation, reﬂection and cogni- Swain, I. U., Zelazo, P. R., & Clifton, R. K. (1993). tive control. Annals of Child Development, 12, Newborn infants’ memory for speech sounds 119–176. 24 retained over hours. Developmental Psychol- Zelazo, P. D., Muller,¨ U., Frye, D., & Marcovitch, ogy, 29, 312–323. S. (2003). The development of executive func- Torres, F., & Anderson, C. ( 1985). The normal tion in early childhood. Monographs of the Soci- EEG of the human newborn. Journal of Clinical ety for Research in Child Development, 68(3), Neurophysiology, 2 , 89–103. Serial No. 274. Trevarthen, C. , & Aitken, K. J. (2001). Infant Zelazo, P. D., & Reznick, J. S. (1991). Age-related intersubjectivity: Research, theory, and clini- asynchrony of knowledge and action. Child cal applications. Journal of Child Psychology & Development, 62, 719–735. 3 48 Psychiatry 42, – . Zelazo, P. D., & Sommerville, J. (2001). Lev- Tulving, E. (1985). Memory and consciousness. els of consciousness of the self in time. In C. Canadian Psychology, 25, 1–12. Moore & K. Lemmon (Eds.), Self in time: Devel- Vygotsky, L. S. (1978). Mind in society. Cam- opmental issues (pp. 229–252). Mahwah, NJ: bridge, MA: Harvard University Press. Erlbaum. Vygotsky, L. S. (1986). Thought and language Zelazo, P. D., Sommerville, J. A., & Nichols, S. (A. Kozulin, Ed.). Cambridge, MA: MIT Press. (1999). Age-related changes in children’s use of (Original work published 1934.) representations. Child Development, 35, 1059– Weiskrantz, L., Sanders, M. D., & Marshall, J. 1071. (1974). Visual capacity in the hemianopic ﬁeld Zelazo, P. R., & Zelazo, P. D. (1998). The emer- following a restricted occipital ablation. Brain, gence of consciousness. Advances in Neurology, 97, 709–728. 77, 149–165. P1: KAE 0521857430c16 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:43

F. Alternative States of Consciousness

433 P1: KAE 0521857430c16 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:43

434 P1: KAE 0521857430c16 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:43

CHAPTER 16 States of Consciousness: Normal and Abnormal Variation

J. Allan Hobson

Abstract principally located in the brainstem, upon the thalamus and cortex located in the The goal of this chapter is to give an upper brain. This means that consciousness account of the phenomenology of the vari- is state dependent and that understanding ations in conscious state, and to show how the mechanisms of brain state control con- three mediating brain processes – activation, tributes indirectly to a solution of the mind- input-output gating, and modulation – brain problem. interact over time so as to account for those The normal and abnormal variations variations in a unified way. in conscious state operate through three The chapter focuses on variations in fairly well-understood physiological pro- consciousness during the sleep-wake cycle cesses: activation (A), input-output gating across species and draws on evidence from (I), and modulation (M) (see Figure 16.1). lesion, electrophysiological, and functional neuroimaging studies. A four-dimensional model called AIM pictorializes both normal Definition and Components and abnormal changes in brain state and provides a unified view of the genesis of a wide of Consciousness variety of normal and abnormal changes in conscious experience. Consciousness may be defined as our awareness of our environment, our bodies, and ourselves. Awareness of ourselves implies an awareness of awareness; that is, the con- Introduction scious recognition that we are conscious beings. Awareness of oneself implies meta- The changes in brain state that result in nor- awareness. mal and abnormal changes in the state of To develop an experimental, scientific the mind all share a common process: an approach to the study of consciousness, alteration in the influence of lower centers, it is convenient to subdivide the mental 435 P1: KAE 0521857430c16 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:43

436 the cambridge handbook of consciousness

What differentiates humans from their fellow mammals and gives humans what Edelman calls secondary consciousness Mode depends upon language and the associated enrichment of cognition that allow humans to develop and to use verbal and numeric abstractions. These mental capacities contribute to our sense of self as agents and as Input creative beings. It also determines the aware- Activation ness of awareness that we assume our animal collaborators do not possess. 16 1 Figure . . The AIM model. Because the most uniquely human cognitive faculties are likely to be functions of our massive cerebral cortex, it is unlikely that the elements that constitute consciousness. We study of animal brains will ever tell us what may discern at least the ten distinct capac- we would like to know about these aspects ities of mind defined in Table 16.1. These of consciousness. Nonetheless, animals can are the faculties of the mind that have been and do tell us a great deal about how other investigated by scientific psychologists since components of consciousness change with their formulation by William James in 1890. changes in brain physiology. The reader who From an examination of this table, it can wishes to learn more about the brain basis of be appreciated that consciousness is compo- consciousness may wish to consult Hobson nential. That is to say, consciousness is made (1998). up of the many faculties of mind, which are It is obvious that when we go to sleep seamlessly integrated in our conscious expe- we lose sensation and the ability to act rience. It should be noted that all of these upon the world. In varying degrees, all functions are also mediated unconsciously or the components of consciousness listed in implicitly. Table 16.1 are changed as the brain changes Only human beings fulfill all of the state. According to the conscious state demands of the definition of consciousness paradigm, consciousness changes state in given above and the components listed in a repetitive and stereotyped way over the Table 16.1. And humans are only fully con- sleep-wake cycle. These changes are so dra- scious when they are awake. It is evident that matic that we can expect to make strong higher mammals have many of the components of consciousness and may thus be considered to be partially conscious. Conscious- Table 16.1. Definition of components of ness is thus graded in both the presence and consciousness intensity of its components. Attention Selection of input data In Edelman’s terms (1992), animals pos- Perception Rpresentation of input data sess primary consciousness, which comprises Memory Retrieval of stored sensory awareness, attention, perception, representations memory (or learning), emotion, and action. Orientation Representation of time, place, This point is of more than theoretical inter- and person est because so much that we know about the Thought Reflection upon brain physiology upon which consciousness representations depends comes from experimental work in Narrative Linguistic symbolization of animals. In making inferences about how our representations own conscious experience is mediated by the Emotion Feelings about representations brain, the attribution of primary conscious- Instinct Innate propensities to act Intention Representations of goals ness to animals is not only naturalistic but Volition Decisions to act also strategic. P1: KAE 0521857430c16 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:43

states of consciousness: normal and abnormal variation 437

inferences about the major physiological ated as the EEG spindles of Stage II NREM underpinnings of consciousness. sleep block the thalamocortical transmission Two conclusions stem from this recogni- of both external and internal signals within tion: The first is that consciousness is graded the brain. When the spindles of Stage II are within and across individuals and species. joined by high-voltage slow waves in over The second is that consciousness is altered half the record, the sleep is called NREM more radically by diurnal changes in brain Stage III; it is called NREM Stage IV when state than it has been by millions of years of the whole record comes to be dominated by evolution. We take advantage of these two the slow waves. facts by studying normal sleep in humans Arousal from Stage NREM IV is difficult, and in those subhuman species with primary often requiring strong and repeated stimula- consciousness. tion. On arousal, subjects evince confusion and disorientation that may take minutes to subside. The tendency to return to sleep is The Sleep-Waking Cycle strong. This process, which is called sleep inertia, is enhanced in recovery sleep follow- 1997 When humans go to sleep they rapidly ing sleep deprivation (Dinges et al., ). become less conscious. The initial loss of As the activation level is falling, result- awareness of the external world that may ing in the sequence of sleep Stages I to occur when we are reading in bed is asso- IV, muscle tone continues to abate pas- ciated with the slowing of the EEG that is sively and the rolling eye movements cease. called Stage 1 (see Fig. 16.2). Frank sleep In Stage IV, the brain is maximally deacti- onset is defined by the appearance of a vated, and responsiveness to external stim- characteristic EEG wave, the sleep spindle, uli is at its lowest point. Consciousness, if which reflects independent oscillation of the it is present at all, is limited to low-level, thalamocortical system. non-progressive thought. It is important to Consciousness is altered in a regular way note three points about these facts. The first at sleep onset. Although awareness of the is that, because consciousness rides on the outside world is lost, subjects may con- crest of the brain activation process, even tinue to have visual imagery and associated slight dips in activation level lead to lapses reflective consciousness. Sleep-onset dreams in waking vigilance. The second is that even are short-lived, and their content departs in the depths of Stage IV NREM sleep when progressively from the contents of previous consciousness appears to be largely obliter- waking consciousness. They are associated ated, the brain remains highly active and is with Stage I EEG, rapidly decreasing muscle still capable of processing its own informa- tone, and slow rolling eye movements. As the tion. From PET and single neurone studies, brain activation level falls further, conscious- it can safely be concluded that the brain 80 ness is further altered and may be obliter- remains about % active in the depths of sleep. These conclusions not only emphasize the Waking graded and state-dependent nature of consciousness. They also indicate how small a fraction of brain activation is devoted to consciousness and that most brain activity is NREM not associated with consciousness. From this it follows that consciousness, being evanescent, is not only a very poor judge of its own causation and of information processing by the brain. It is evident that consciousness REM requires a very specific set of neurophysio- Figure 16.2 . The sleep-wake cycle. logical conditions for its occurrence. P1: KAE 0521857430c16 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:43

438 the cambridge handbook of consciousness

REM Sleep emotion centers of the brain are activated (or even hyperactivated) in REM sleep, and In 1953, Aserinsky and Kleitman reported indeed, we have found that this is the case. that the sleep EEG was periodically acti- At the same time that the perceptual and vated to near waking levels and that rapid emotional components of consciousness are eye movements (REMs) could then be enhanced in dreams, such cognitive func- recorded. When aroused from this REM tions as memory, orientation, and insight are sleep state, subjects frequently reported hal- impaired. It is difficult upon awakening to lucinoid dreaming (Dement & Kleitman, remember one’s dreams but also previous 1957). It was later discovered by Jouvet and scenes may be lost even as the dream unfolds Michel (1959) that the EMG of the cat was (M. Fosse et al., 2002). It has recently been actively inhibited as the brain was sleep acti- shown that even well-remembered dreams vated and that the same inhibition of motor do not faithfully reproduce waking experi- output occurs in humans during REM sleep ence (M. Fosse et al., 2002). Perhaps related (Hodes & Dement, 1964). to the memory defect is the microscopic dis- The overnight tendency is for the periods orientation called dream bizarreness, which of Stage I–IV brain deactivation to become results in extreme inconstancy of the unities shorter and less deep while the REM peri- of time, place, person, and action (R. Fosse ods become longer and more intense. As the et al., 2001). It is these unities that constitute brain is activated more and more, the dif- the anchors of waking consciousness. ferentiation in consciousness is correspond- Reports of thinking are rare on arousal ingly less marked, with reports from early from REM sleep, and the thinking that is morning Stage II coming more and more to reported, although logical within the fan- resemble those of Stage I. Dreaming, it can ciful assumptions of the dream, is almost thus be reasonably concluded, is our con- wholly lacking in insight as to the true state scious experience of brain activation in sleep. of the mind (R. Fosse et al., 2001). Thus, in Because brain activation is most intense in dreams, we typically assume we are awake REM sleep, dreaming is most highly corre- when we are, in fact, asleep. The converse lated with that brain state. almost never occurs, weakening the thesis Waking and dreaming consciousness con- of such skeptical philosophers as Malcolm trast along many of the dimensions shown (1956), who hold that we never know cer- in Table 16.2. It can be seen that, although tainly what state we are in and that reports dreaming constitutes a remarkable percep- of dreaming are fabricated upon awakening. tual and emotional simulacrum of waking, it has equally remarkable cognitive deficiencies. The internally generated visual percepts The Neurophysiology of Sleep with of dreaming are so rich and vivid that they Special Reference to Consciousness regularly lead to the delusion that we are awake. When they are associated with strong The deactivation of the brain at sleep onset is emotions (principally joy-elation, fear- seen as the characteristic EEG change and is anxiety, and anger), they can even be surreal. experienced as an impairment of consciousness. It is related to decreases in activity of the neurones that constitute the brainstem “Why Does the Eye See a Thing More reticular formation. This finding is in con- Clearly in Dreaming Than When We cordance with the classical experiments of Are Awake?” Moruzzi and Magoun (1949) who showed that arousal and EEG activation were a func- As Leonardo da Vinci pointed out, dream tion of the electrical impulse traffic in the consciousness may be even more intense brainstem core. than that of normal waking. Such phe- Since 1949, the reticular activating sys- nomenology suggests that perception and tem has been shown to be anything but P1: KAE 0521857430c16 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:43

states of consciousness: normal and abnormal variation 439

Table 16.2 . Contrasts in the phenomenology of waking and dreaming consciousness Function Nature of Difference Causal Hypothesis Sensory input Blocked Presynaptic inhibition Perception (external) Diminished Blockade of sensory input Perception (internal) Enhanced Disinhibition of networks storing sensory representations Attention Lost Decreased aminergic modulation causes a decrease in signal to noise ratio Memory (recent) Diminished Because of aminergic demodulation, activated representations are not restored in memory Memory (remote) Enhanced Distinhibition of networks storing mnemonic representations increases access to consciousness Orientation Unstable Internally inconsistent orienting signals are generated by cholinergic system Thought Reasoning ad hoc, logical Loss of attention memory and volition leads rigor weak, processing, to failure of sequencing and rule hyperassociative inconstancy; analogy replaces analysis Insight Self-reflection lost (failure Failure of attention, logic, and memory to recognize state as weaken second (and third) order dreaming) representations Language (internal) Confabulatory Aminergic demodulation frees narrative synthesis from logical restraints Emotion Episodically strong Cholinergic hyperstimulation of amygdala and related temporal lobe structures triggers emotional storms, which are unmodulated by aminergic restraint Instinct Episodically strong Cholinergic hyperstimulation of hypothalamus and limbic forebrain triggers fixed action motor programs, which are experienced fictively but not enacted Volition Weak Top-down motor control and frontal executive power cannot compete with disinhibited subcortical network activation Output Blocked Postsynaptic inhibition

non-specific (Hobson & Brazier, 1980). choline (on the cholinergic side). The state Instead, it consists of highly specific of the brain and consciousness is thus deter- interneurones that project mainly locally mined not only by its activation level but also but also reach upward to the thalamus and by its mix of neuromodulators. downward to the spinal cord. By means of Sngle-cell recording studies in cats have these connections, reticular formation neu- revealed that in REM sleep, when global rones regulate muscle tone, eye movements, brain activation levels are as high as in wak- and other sensorimotor functions necessary ing, the firing of two aminergic groups is to waking consciousness. shut off (Hobson, McCarley, & Wyzinski, The reticular formation also contains 1975; McCarley & Hobson, 1975). Thus the chemically specific neuronal systems whose activated brain of REM sleep is aminer- axons project widely throughout the brain gically demodulated with respect to nore- where they secrete the so-called neuro- pinephrine and serotonin. Because nore- modulators: dopamine, norepinephrine, and pinephrine is known to be necessary for serotonin (on the aminergic side) and acetyl- attention (Foote, Bloom, & Aston-Jones, P1: KAE 0521857430c16 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:43

440 the cambridge handbook of consciousness

1983) and serotonin is necessary for memory off-line to sensory inputs and motor out- (Martin et al., 1997), we can begin to under- puts. That is to say, we are anesthetized stand the cognitive deficiencies of dreaming and paralyzed in addition to experiencing consciousness in physiological terms. hallucinated emotion and being disoriented What about the enhancement of internal and amnesic. This is the activation-synthesis perception and emotion that characterizes theory of dreaming (Hobson & McCarley, dream consciousness? Could it be related to 1977). What other evidence can be brought the persistence of the secretion of dopamine to test these hypotheses? and the increase in output of the cholinergic neurones of the brainstem? It turns out that the cholinergic neurones of the retic- A Four-Dimensional Model ular formation are indeed hyperexciteable of Conscious State in REM; in fact, they fire in bursts that are tightly linked in a directionally specific way Three factors – activation level (A), input- to the eye movements that give REM sleep output gating (I), and neuromodulation ratio its name. The result is that such forebrain (M) – determine the normal changes in structures as the amygdala (in the limbic, the state of the brain that give rise to emotion-mediating brain) and the postero- changes in the state of consciousness that dif- lateral cortex (in the multimodal sensory ferentiate waking, sleeping, and dreaming. brain) are bombarded with cholinergically Because these three variables can be mea- mediated internal activation waves during sured in animals, it is appropriate and heuris- REM. tically valuable to model them. In so doing, In the transition from waking to REM, we replace the traditional two-dimensional consciousness has shifted from exterocep- model with the four-dimensional model in tive perception to interoceptive and from Figure 16.1. moderated to unmoderated emotion. To In the AIM model, time is the fourth explain this shift, cholinergic hypermodu- dimension because the instantaneous values lation together with persistent dopaminer- of A, I, and M are points that move in the gic modulation is a candidate mechanism. three-dimensional state space. They form an The mind has simultaneously shifted from elliptical trajectory that represents the sleep- oriented to disoriented and from mnemonic wake sequence as a cyclical function, rather to amnesic cognition. To explain this shift, than as the stairway that is represented in the aminergic demodulation is the best current traditional two-dimensional model in which candidate mechanism. activation is plotted against time (look again at Figures 16.1 and 16.2). To understand the AIM model, it is help- Input-Ouput Gating ful to grasp the fact that the waking domain If the brain is activated in sleep, why don’t is in the back upper right corner of the we wake up? One reason is the aminer- state space. It is there, and only there, that gic demodulation. Another powerful rea- activation (A) level is high, input-output son is that in REM sleep sensory input gates (I) are open, and the modulatory mix and motor output are actively blocked. This (M) measured as the aminergic/cholinergic closing of the input and output gates is ratio is also high. Because all three measures an active inhibitory process in the spinal change from moment to moment, the AIM and the motor neurones that convey move- points form a cloud in the waking domain of ment commands to the muscles. Sensorimo- the state space. tor reticular formation neurones inhibit the When sleep supervenes, all three AIM sensory afferent sensory fibers coming from variables fall. The net result is that the the periphery. NREM (N) sleep domain is the center of The net result is that in dreams we are not the state space. With the advent of REM, only perceptually and emotionally hyper- the activation level rises again to waking lev- conscious but also cognitively deficient and els, but the input-output gates are actively P1: KAE 0521857430c16 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:43

states of consciousness: normal and abnormal variation 441

closed and aminergic neurones are shut off. insights to the brain basis of conscious expe- Factors I and M therefore fall to their lowest rience via the conscious state paradigm. possible levels. The REM sleep domain (R) is thus in the right anterior lower corner of Brain Imaging the state space. The AIM model clearly differentiates REM sleep from waking. It also Taking advantage of PET technology, three affords a valuable picture of how and why separate independent groups have imaged the conscious states of waking and dreaming the human brain in normal waking and sleep 1997 2000 differ in the way that they do. (Braun et al., ; Maquet, ; Nofzinger 1997 As shown by the dashed line forming et al., ). At sleep onset, the blood flow an elliptical trajectory through the state to all regions of the brain declines. When space, the sleep-wake cycle is represented REM sleep supervenes most brain regions as a recurrent cycle. Actually the sequential resume the wake state brain perfusion lev- cycles of sleep move to the right (as the acti- els (from which we infer a restored activa- vation level increases overnight) and down- tion level compared to waking). But several ward as the brain comes to occupy the REM brain regions are selectively hyperactivated domain for longer and longer periods of in REM. They include the pontine reticu- time. lar formation (which previous animal stud- Lucid dreaming is a normal variation in ies have shown to regulate REM sleep), conscious state that serves to illustrate and the amygdala and the deep basal forebrain emphasize the value of the AIM model. (which are thought to mediate emotion), When subjects learn to recognize that they the parietal operculum (which is known are dreaming while they are dreaming, they to be involved in visuospatial integration), obviously have elements of both REM and and the paralimbic cortices (which integrate waking consciousness. They can continue to emotion with other modalities of conscious hallucinate, but they are no longer deluded experience). about the provenance of the imagery. Lucid dreamers typically report that, Spontaneous Brain Damage although they may learn to watch and con- Patients who have suffered brain damage sciously influence the course of their dreams due to stroke report a complete cessation of and even to voluntarily awaken to enhance dreaming when their lesion impairs either recall, lucidity is difficult to maintain. Often, the parietal operculum or the deep frontal they are either pulled back down into non- white matter (Solms, 1997). This finding lucid dreaming or wake up involuntarily. suggests that those structures mediate con- The lucid dreaming domain lies between nections that are essential to dream con- REM and wake in the middle of the state sciousness. When damage is restricted to space near the right side wall. Subjects nor- the visual brain, subjects continue to dream mally cross the REM-wake transition zone vividly, but they lack visual imagery. rapidly, suggesting that lucid dreaming is a forbidden zone of the state space. Such Intentional Lobotomy unwelcome processes as sleep paralysis and hypnopompic hallucinations occur when The clinical histories of patients with mental subjects wake up but one or another REM illness who had undergone frontal lobotomy process persists. in the 1950s revealed an effect on dreaming. This surgical procedure was designed to cut the fibers connecting the frontal lobes to other parts of the limbic lobe on the Brain Imaging and Lesion Studies assumption that the emotion that was in Humans thought to be driving the patient’s psychosis was mediated by these fibers. Some patients Over the past decade, two parallel lines of did indeed benefit from the surgery, but scientific inquiry have contributed striking many reported a loss of dreaming, again P1: KAE 0521857430c16 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:43

442 the cambridge handbook of consciousness

suggesting that fronto-limbic connections dimension of AIM in the direction of REM. were as essential to that normal hallucina- Such a formulation is compatible with the tory process as they were to psychosis. PET finding of selective temporal lobe activation in normal REM sleep. It is reasonable Other Abnormal Conditions to propose that the kinship of temporal lobe epilepsy “dreamy” states and normal dream- When traumatic brain damage or stroke ing is due to a shared selective activation of affects the brainstem, the resulting injury to limbic structures. neurones mediating activation, input-output However, this local excitability change gating, and modulation can render subjects cannot be easily modeled by AIM because comatose for long periods of time. Such sub- the activation measure is global, and as jects may be unable to wake or to sleep nor- PET studies indicate, the activation of REM mally, in which case they are said to be in (and TLE) is regionally selective, there being a chronic vegetative state. They have been some brain areas (like the limbic lobe) that permanently moved to the left half of the are turned on and others (like the dorsolat- AIM state space. As they move further and eral prefrontal cortex) that are turned off. further to the left, they may lose the capac- The only way to deal with this reality is to ity to activate their thalamocortical system add brain regions as a fifth dimension to the even to the NREM sleep level. A flat EEG AIM model. Because it is impossible to rep- indicates a complete absence of activation resent brain regions within the state space of and intrinsic oscillation. AIM, the easiest way to represent and visualize this modification is to see the brain as a locked-in syndrome regionally diverse set of AIM models. Thus Patients with amytrophic lateral sclerosis the value of the AIM may be locally altered (popularly known as Lou Gehrig’s Disease) with profound effects upon consciousness. remain conscious during waking, but are unable to signal out because of motor neuronal death. Recent research suggests that Conclusions they can be taught to signal out and say “yes” or “no” by raising or lowering their cortical By studying the way that consciousness is DC potentials (Hinterberger et al., 2004). normally altered when we fall asleep and It is not known whether these subjects have when we dream, it is possible to obtain normal sleep cycles, but the assumptions of insights about how the brain mediates con- the AIM model predict that they should. sciousness. So stereotyped and so robust are the corresponding changes in brain and temporal lobe epilepsy and conscious state as to assure the following “dreamy states” conclusions: When neuronal excitability is locally altered 1 (as in temporal lobe epilepsy), patients . Consciousness is componential. It com- sometimes experience the intrusion of prises many diverse mental functions dream-like states into waking consciousness. that, in waking, operate in a remarkably This phenomenon serves to illustrate both unified fashion to mediate our experience the value and the limitations of the AIM of the world, our bodies, and ourselves. model. 2. Consciousness is graded. Within and If the abnormal discharge of the epileptic across species, animals are continually focus in the temporal lobe is strong enough, more or less conscious depending upon it can come to dominate the rest of the the componential complexity and the brain and cause it to enter an altered state of state of their brains. waking consciousness akin to dreaming. This 3. Consciousness is state dependent. During shift, which is caused by an increase in inter- normal sleep, consciousness undergoes nal stimulus strength, causes a change in the I both global and selective componential P1: KAE 0521857430c16 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:43

states of consciousness: normal and abnormal variation 443

differentiation as the brain regions medi- cycle. An H2(15)O PET study. Brain, 120(7), ating the components of consciousness 1173–1197. are globally or selectively activated and Dement, W. C., & Kleitman, N. (1957). The rela- deactivated. tion of eye movements during sleep to dream 4. Conscious state is a function of brain activity: An objective method for the study of dreaming. Journal of Experimental Psychology, state. Experimental studies of sleep have 53(3), 339–346. identified three factors that determine Dinges, D. F., Pack, F., Williams, K., Gillen, brain state: activation level (A), input- K. A., Powell, J. W., Ott, G. E., Aptowicz, output gating (I), and modulation (M). C., & Pack, A. I. (1997). Cumulative sleepi- With time as a fourth dimension, the ness, mood disturbance, and psychomotor vig- resulting AIM model represents the sleep ilance performance decrements during a week cycle as an ellipse and more clearly dif- of sleep restricted to 4–5 hours per night. Sleep, ferentiates waking and REM as the sub- 20(4), 267–277. strate of the conscious states of waking Edelman, G. M. (1992). Bright air,brilliant fire: On and dreaming. the matter of the mind. New York: Basic Books. 5. Recent brain imaging and brain lesion Foote, S. L., Bloom, F. E., & Aston-Jones, G. studies in humans indicate that activation (1983). Nucleus locus coeruleus: New evidence (A) is not only global but also regional of anatomical physiological specificity. Physio- 844 914 and that selective activations and inacti- logical Review, 63, – . vations of specific brain subregions con- Fosse, M. J., Fosse, R., Hobson, J. A., & Stick- 2002 tribute to differences in conscious expe- gold, R. ( ). Dreaming and episodic memory: A functional dissociation? Journal of Cog- rience. A fifth dimension may therefore nitive Neuroscience, 15(1), 1–9. have to be added to the AIM model. Fosse, R., Stickgold, R., & Hobson, J. (2001). 6. Armed with the AIM model, it is possible A. Brain-mind states: Reciprocal variation to obtain a unified view of the genesis of in thoughts and hallucinations. Psychological a wide variety of normal and abnormal Science, 12(1), 30–36. changes in conscious experience. Hinterberger, T., Neumann, N., Pham, M., Kubler, A., Grether, A., Hofmayer, N., Wilhelm, B., Flor, H., & Birbaumer, N. (2004). Acknowledgements A multimodal brain-based feedback and communication system. Experimental Brain The author gratefully acknowledges the fol- Research, 154(4), 521–526. lowing sponsors of his research: the National Hobson, J. A. (1998). Consciousness. New York: Institutes of Health; the National Science W. H. Freeman. 1980 Foundation; the John D. and Catherine T. Hobson, J. A., & Brazier, M. A. B. (Eds.). ( ). MacArthur Foundation; and the Mind Sci- The reticular formation revisited: Specifying function for a nonspecific system. New York: Raven ence Foundation. Technical assistance was Press. provided by Katerina di Perri and Nicholas Hobson, J. A., & McCarley, R. W. (1977). The Tranguillo. brain as a dream state generator: An activation synthesis hypothesis of the dream process. American Journal of Psychiatry, 134(12), 1335– References 1348. Hobson, J. A., McCarley, R. W., & Wyzinski, Aserinsky, E., & Kleitman, N. (1953). Regularly P. W. (1975). Sleep cycle oscillation: Reciprocal occurring periods of eye motility and con- discharge by two brain stem neuronal groups. comitant phenomena during sleep. Science, 118, Science, 189, 55–58. 273–274. Hodes, R., & Dement, W. C. (1964). Depression Braun, A. R., Balkin, T. J., Wesenten, N. J., Car- of electrically induced reflexes (“H-reflexes”) in son, R. E., Varga, M., Baldwin, P., Selbie, S., man during low voltage EEG sleep. Electroen- Belenky, G, & Herscovitch, P. (1997). Regional cephalography and Clinical Neurophysiology, 17, cerebral blood flow throughout the sleep-wake 617–629. P1: KAE 0521857430c16 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:43

444 the cambridge handbook of consciousness

Jouvet, M. (1962). Recherches sur les structures Kandel, E. R. (1997). Synapse-speciﬁc, long- nerveuses et les mecanismes responsables des term facilitation of aplysia sensory to motor differentes phases du sommeil physiologique. synapses: A function for local protein syn- Archives Italiennes de Biologie, 100, 125– thesis in memory storage. Cell, 91(7), 927– 206. 938. Jouvet, M. (1969). Biogenic amines and the states McCarley, R. W., & Hobson, J. A. (1975). Neu- of sleep. Science, 163(862), 32–41. ronal excitability modulation over the sleep Jouvet, M., & Michel, F. (1959). Correlation cycle: A structural and mathematical model. electromyographiques du sommeil chez le Science, 189, 58–60. chat decortique mesencephalique chronique. Moruzzi, G., & Magoun, H W. (1949). Brainstem Comptes Rendues des Seances de la Societe de reticular formation and activation of the EEG. Biologie et de Ses Filiales, 153, 422–425. Electroencephalography and Clinical Neurophys- Malcolm, N. (1956). Dreaming and skepticism. iology, 1, 455–473. Philosophical Review, 65, 14–37. Nofzinger, E. A., Mintun, M. A., Wiseman, M., Maquet, P. (2000). Functional neuroimaging of Kupfer, D. J., & Moore, R. Y. (1997). Forebrain, sleep by positron emission tomography. Journal activation in REM sleep: An FDG PET study. of Sleep Research, 9, 207–231. Brain Research, 770(1–2), 192–201. Martin, K. C., Casadio, A., Zhu, H., Yaping, Solms, M. (1997). The neuropsychology of dreams. E., Rose, J. C., Chen, M., Bailey, C. H., & Hillsdale, NJ: Erlbaum. P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

CHAPTER 17 Consciousness in Hypnosis

John F. Kihlstrom

Abstract correlates of hypnotic suggestion revealed by brain imaging. In hypnosis, subjects respond to suggestions for imaginative experiences that can involve alterations in conscious perception, memory, and action. However, these phenom- Consciousness in Hypnosis ena occur most profoundly in those subjects who are highly hypnotizable. The chap- Hypnosis is a process in which one per- ter reviews a number of these phenomena, son (commonly designated the subject) including posthypnotic amnesia; hypnotic responds to suggestions given by another analgesia; hypnotic deafness, blindness, and person (designated the hypnotist) for imag- agnosia; and emotional numbing, with inative experiences involving alterations in an eye toward uncovering dissociations perception, memory, and the voluntary con- between explicit and implicit memory, per- trol of action. Hypnotized subjects can be ception, and emotion. These dissociative oblivious to pain; they hear voices that aren’t phenomena of hypnosis bear a phenotypic there and fail to see objects that are clearly similarity to the “hysterical” symptoms char- in their ﬁeld of vision; they are unable to acteristic of the dissociative and conversion remember the things that happened to them disorders. The experience of involuntariness while they were hypnotized; and they carry in hypnotic response is considered in light of out suggestions after hypnosis has been ter- the concept of automatic processing. Hyp- minated, without being aware of what they nosis may be described as an altered state are doing or why. In the classic case, these of consciousness based on the convergence experiences are associated with a degree of of four variables: induction procedure, sub- subjective conviction bordering on delusion jective experience, overt behavior, and psy- and an experience of involuntariness border- chophysiological indices – including neural ing on compulsion.

445 P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

446 the cambridge handbook of consciousness

The Importance of Individual tizable subjects are compared to those who Differences are insusceptible to hypnosis. In any case, measurement of hypnotizability is crucial to hypnosis research: There is no point in The phenomena of hypnosis can be quite studying hypnosis in individuals who cannot dramatic, but they do not occur in every- experience it. one. Individual differences in hypnotizabil- Some clinical practitioners believe that ity are measured by standardized psycho- virtually everyone can be hypnotized, if logical tests, such as the Harvard Group only the hypnotist takes the right approach, Scale of Hypnotic Susceptibility, Form A but there is little evidence favoring this (HGSHS:A) or the Stanford Hypnotic Sus- point of view. Similarly, some researchers ceptibility Scale, Form C (SHSS:C). These believe that hypnotizability can be enhanced psychometric instruments are essentially by developing positive attitudes, motiva- work samples of hypnotic performance, con- tions, and expectancies concerning hypnosis sisting of a standardized induction of hypno- (Gorassini & Spanos, 1987), but there is also sis accompanied by a set of 12 representative evidence that such interventions are heavily hypnotic suggestions. For example, on both laced with compliance (Bates & Kraft, 1991). HGSHS:A and SHSS:C, subjects are asked As with any other skilled performance, hyp- to hold out their left arm and hand, and then notic response is probably a matter of both it is suggested that there is a heavy object in aptitude and attitude: Negative attitudes, the hand, growing heavier and heavier, and motivations, and expectancies can interfere pushing the hand and arm down. The sub- with performance, but positive ones are not ject’s response to each suggestion is scored by themselves sufficient to create hypnotic according to objective behavioral criteria virtuosity. (for example, if the hand and arm lower Hypnotizability is not substantially corre- at least 6 inches over a specified interval lated with most other individual differences of time), yielding a single score represent- in ability or personality, such as intelligence ing his or her hypnotizability, or responsive- or adjustment (Hilgard, 1965). However, in ness to hypnotic suggestions. Hypnotizabil- the early 1960s, Ronald Shor (Shor, Orne, ity, so measured, yields a quasi-normal distri- & O’Connell, 1962), Arvid As (As, 1962), bution of scores in which most people are at and others found that hypnotizability was least moderately responsive to hypnotic sug- correlated with subjects’ tendency to have gestions, relatively few people are refractory hypnosis-like experiences outside of formal to hypnosis, and relatively few fall within hypnotic settings, and an extensive interview the highest level of responsiveness (Hilgard, study by Josephine Hilgard (1970) showed 1965). that hypnotizable subjects displayed a high Although most people can experience level of imaginative involvement in such hypnosis to at least some degree, the most domains as reading and drama. In 1974, Tel- dramatic phenomena of hypnosis – the ones legen and Atkinson developed a scale of that really count as reflecting alterations in absorption to measure the disposition to have consciousness – are generally observed in subjective experiences characterized by the those “hypnotic virtuosos” who comprise the full engagement of attention (narrowed or upper 10 to 15% of the distribution of hyp- expanded), and blurred boundaries between notizability. Accordingly, a great deal of hyp- self and object (Tellegen & Atkinson, 1974). nosis research involves a priori selection of Episodes of absorption and related phenom- highly hypnotizable subjects, to the exclu- ena such as “flow” (Csikszentmihalyi, 1990; sion of those of low and moderate hypno- Csikszentmihalyi & Csikszentmihalyi, 1988) tizability. An alternative is a mixed design are properly regarded as altered states of in which subjects stratified for hypnotizabil- consciousness in their own right, but they ity are all exposed to the same experimental are not the same as hypnosis and so are not manipulations, and the responses of hypno- considered further in this chapter. P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

consciousness in hypnosis 447

Conventional personality inventories, ations in consciousness. The sensory alter- such as the Minnesota Multiphasic Person- ations exemplified by hypnotic analgesia ality Inventory and California Psychological or deafness, as well as posthypnotic amne- Inventory, do not contain items related to sia, are disruptions in conscious awareness: absorption, which may explain their failure The subject seems to be unaware of per- to correlate with hypnotizability (Hilgard, cepts and memories that ought to be acces- 1965). However, absorption is not wholly sible to phenomenal awareness. Similarly, unrelated to other individual differences in posthypnotic suggestion, as well as the expe- personality. Recent multivariate research rience of involuntariness that frequently has revealed five major dimensions – the accompanies suggested hypnotic experi- “Big Five” – which provide a convenient ences, reflects a loss of control over cognition summary of personality structure: neuroti- and behavior. cism (emotional stability), extraversion, Despite these considerations, the status of agreeableness, conscientiousness, and open- hypnosis as an altered state of consciousness ness to experience (John, 1990; Wiggins & has been controversial (e.g., Gauld, 1992; Trapnell, 1997). Absorption and hypnotiz- Hilgard, 1971; Kallio & Revensuo, 2003; ability are correlated with those aspects of Kirsch & Lynn, 1995; Shor, 1979a).1 For openness that relate to richness of fantasy example, psychoanalytically inclined theo- life, aesthetic sensitivity, and awareness of rists classified hypnosis as an instance of inner feelings, but not those that relate to adaptive regression, or regression in the intellectance or sociopolitical liberalism service of the ego (Fromm, 1979; Gill & (Glisky & Kihlstrom, 1993; Glisky, Tataryn, Brenman, 1959). Orne believed that the Tobias, & Kihlstrom, 1991). essence of hypnosis was to be found in Absorption is the most reliable corre- “trance logic” (Orne, 1959), whereas Hil- late of hypnotizability; by contrast, vivid- gard argued that the phenomena of hyp- ness of mental imagery is essentially uncor- nosis were essentially dissociative in nature related with hypnosis (Glisky, Tataryn, & (Hilgard, 1973b, 1977). By contrast, Sarbin Kihlstrom, 1995). However, the statistical and Coe described hypnosis as a form of relations between hypnotizability and either role-enactment (Sarbin & Coe, 1972); Bar- absorption or openness are simply too weak ber asserted that the phenomena of hyp- to permit confident prediction of an indi- nosis could be produced by anyone who vidual’s actual response to hypnotic sug- held appropriate attitudes, motivations, and gestion (Roche & McConkey, 1990). So expectancies (Barber, 1969). far as the measurement of hypnotizabil- More recently, both Woody and Bowers ity is concerned, there is no substitute for (Woody & Bowers, 1994; Woody & Sadler, performance-based measures such as the 1998) and Kihlstrom (Kihlstrom, 1984, Stanford and Harvard scales. 1992a, 1998) embraced some version of Hilgard’s neodissociation theory of divided consciousness. By contrast, the “sociocogni- The Controversy over State tive” approach offered by Spanos (1986a, 1991) emphasized the motivated subject’s Consciousness has two principal aspects: attempt to display behavior regarded as monitoring ourselves and our environment, characteristic of a hypnotized person and so that objects and events are accurately the features of the social context that represented in phenomenal awareness, and shaped these displays. Kirsch and Lynn controlling ourselves and the environment (Kirsch, 2001a,b; Kirsch & Lynn, 1998a,b) through the voluntary initiation and termi- offered a “social cognitive” theory of hyp- nation of thought and action (Kihlstrom, nosis that attributed hypnotic phenom- 1984). From this point of view, the phe- ena to the automatic effect of subjects’ nomena that mark the domain of hypno- response expectancies. Following Kuhn sis (Hilgard, 1973a) seem to reflect alter- (1962), the “state” and “nonstate” views of P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

448 the cambridge handbook of consciousness

hypnosis have sometimes been construed wider sociocultural context. A truly ade- as competing paradigms (e.g., Spanos & quate, comprehensive theory of hypnosis Chaves, 1970, 1991). will seek understanding in both cognitive and interpersonal terms. We do not yet have such a theory, but even if we did individ- Consciousness and Social Influence ual investigators would naturally emphasize Part of the problem is the multifaceted one aspect, whether altered consciousness or nature of hypnosis itself. Hypnosis entails social context, over the other in their work. changes in conscious perception, memory, The interindividual competition that is part and behavior, to be sure, but these changes and parcel of science as a social enterprise also occur following specific suggestions often leads investigators to write as if alter- made by the hypnotist to the subject. As ations in consciousness and social influence White (1941) noted at the dawn of the mod- were mutually exclusive processes – which ern era of hypnosis research, hypnosis is they simply are not. a state of altered consciousness that takes Taken together with the null-hypothesis place in a particular motivational context – statistical tests that remain part and parcel the motivation being to behave like a hypno- of the experimental method, and a propen- tized subject. Orne (1959), who was White’s sity for making strong rather than weak infer- protege as both an undergraduate and a grad- ences from experimental data, investigators uate student at Harvard, famously tried to will often present evidence for one process distinguish between artifact and essence of as evidence against the other. But if there hypnosis, but a careful reading of his work is one reason why hypnosis has fascinated makes it clear that the demand characteris- successive generations of investigators, since tics that surround hypnosis are as important the very dawn of psychology as a science, it as any “trance logic” that arises in hypnosis. is that hypnosis exemplifies the marvelous Similarly, at the dawn of what might complexity of human experience, thought, be called the “golden age” of hypnosis and action. In hypnosis and elsewhere, com- research, Sutcliffe published a pair of semi- prehensive understanding will require a cre- nal papers that contrasted a credulous view ative synthesis in the spirit of discovery, of hypnosis, which holds that the mental rather than the spirit of proof – a creative states instigated by suggestion are identical synthesis of both-and, as opposed to a stance to those that would be produced by the of either-or. actual stimulus state of affairs implied in the suggestions, with a skeptical view that Defining an Altered State holds that the hypnotic subject is acting as if the world were as suggested (Sutcliffe, Part of the problem as well are the diffi- 1960, 1961). This is, of course, a version culties of defining precisely what we mean of the familiar state-nonstate dichotomy, by an altered state of consciousness (Lud- but Sutcliffe also offered a third view: that wig, 1966). Some theorists have argued that hypnosis involves a quasi-delusional alter- every altered state should be associated with ation in self-awareness – an altered state a unique physiological signature, much as of consciousness that is constructed out of dreaming is associated with the absence of the interaction between the hypnotist’s sug- alpha activity in the EEG and the occur- gestions and the subject’s interpretation of rence of rapid eye movements (REM). The those suggestions. lack of a physiological indicator for hypno- Thus, hypnosis is simultaneously a state sis, then, is taken as evidence that hypnosis of (sometimes) profound cognitive change, is not a special state of consciousness after involving basic mechanisms of perception, all. But of course, this puts the cart before memory, and thought, and a social interac- the horse. Physiological indices are validated tion, in which hypnotist and subject come against self-reports: Aserinsky and Kleitman together for a specific purpose within a (1953) had to wake their subjects up during P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

consciousness in hypnosis 449

periods of REM and ask them if they were hypnotizable subjects, even the induction dreaming. As such, physiological correlates procedure may be unnecessary. have no privileged status over introspective 2. Subjective Experience: Introspective self- self-reports: Aserinsky and Kleitman were reports of changes in subjective experi- in no position to contradict subjects who ence would seem to be central to any said “no.” It is nice when our altered states altered state of consciousness. As noted have distinct physiological correlates, but earlier, the domain of hypnosis is defined our present knowledge of mind-body rela- by changes in perception, memory, and tions is simply not sufficient to make such the voluntary control of behavior – anal- correlates a necessary part of the definition. gesia, amnesia, the experience of involun- After all, cognitive neuroscience has made tariness, and the like. If the hypnotist gives very little progress in the search for the a suggestion – for example, that there is an neural correlates of ordinary waking con- object in the subject’s outstretched hand, sciousness (Metzinger, 2000). How far in getting heavier and heavier – and the sub- the future do the neural correlates of altered ject experiences nothing of the sort, it states of consciousness, like hypnosis, await? is hard to say that he or she has been In the final analysis, it may be best to treat hypnotized. hypnosis and other altered states of con- 3. Overt Behavior: Of course, a reliance on sciousness as natural concepts, represented self-reports has always made psycholo- by a prototype or one or more exem- gists nervous, so another residue of radi- plars, each consisting of features that are cal behaviorism (the first was the reliance only probabilistically associated with cate- on operational definitions) is a focus on gory membership, with no clear boundaries overt behavior. If a subject hallucinates between one altered state and another, or an object in his outstretched hand, and between altered and normal consciousness feels it grow heavier and heavier, eventu- 1984 (Kihlstrom, ). And because we can- ally his arm ought to drop down to his not have direct knowledge of other minds, side. As noted earlier, individual differ- altered states of consciousness must also ences in hypnotizability are measured in remain hypothetical constructs, inferred from terms of the subject’s publicly observable, a network of relations among variables that overt, behavioral response to suggestions. are directly observable (Campbell & Fiske, But in this instance, the overt behavior is, 1959 1956 ; Garner, Hake, & Eriksen, ; Stoyva to borrow a phrase from the Book of Com- 1968 & Kamiya, ), much in the manner of mon Prayer, an outward and visible sign of a psychiatric diagnosis. From this point of an inward and spiritual grace: It is a con- view the diagnosis of an altered state of consequence of the subject’s altered subjec- sciousness can be made with confidence to tive experience. Behavioral response is of the extent that there is convergence among no interest in the absence of correspond- four kinds of variables: ing subjective experience. For this rea- 1. Induction Procedure: Operationally, a spe- son, requests for “honesty reports” (Bow- cial state of consciousness can be defined, ers, 1967; Spanos & Barber, 1968) or other in part, by the means employed to induce appropriate postexperimental interviews it – or, alternatively, as the output result- (Orne, 1971; Sheehan & McConkey, 1982) ing from a particular input. Barber (1969) can help clarify subjects’ overt behavior employed such an input-output defini- and serve as correctives for simple behav- tion as the sole index of hypnosis, largely ioral compliance. ignoring individual differences in hyp- 4. Psychophysiological Indices: Because both notizability. At the very least, hypnosis self-reports and overt behaviors are under would seem to require both a hypnotic voluntary control, and thus subject to dis- induction and a hypnotizable individual tortion by social-influence processes, hyp- to receive it. But in the case of very highly nosis researchers have been interested in P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

450 the cambridge handbook of consciousness

psychophysiological indices of response – affecting tactile sensitivity, “special senses” including, of course, various brain imag- such as vision and audition, and motor func- ing techniques. Over the years, a number tion. Charcot held that these symptoms, of such indices have been offered, includ- in turn, were the products of “functional” ing skin conductance and alpha activ- lesions in the nervous system produced by ity, but these have usually proved to emotional arousal and suggestion. be artifacts of relaxation and not intrin- Charcot’s interest in hysteria passed to sic to hypnosis. In retrospect, it was his proteg´ e´ Pierre Janet, who held that probably a mistake to expect that there the fundamental difficulty in hysteria was would be any physiological correlates of a restriction in awareness – such that, for hypnosis in general, following an induc- example, hysterically deaf patients were not tion procedure but in the absence of aware of their ability to hear and hysteri- any specific suggestions (Maquet et al., cally paralyzed patients were not aware of 1999), because subjects can have a wide their ability to move (Janet, 1907). Like variety of experiences while they were Charcot, Janet was particularly impressed by hypnotized. Progress on this issue is the apparently paradoxical behavior of hys- more likely to occur when investiga- terical patients, as exemplified by ostensibly tors focus on the physiological correlates blind individuals who nevertheless displayed of specific hypnotic suggestions – as in visually guided behavior. Janet argued that brain imaging work that shows specific these behaviors were mediated by mental changes in brain activity corresponding to structures called psychological automatisms. hypnotic visual hallucinations (Kosslyn, In his view, these complex responses to Thompson, Costantini-Ferrando, Alpert, environmental events were normally acces- & Spiegel, 2000) or analgesia (Rainville, sible to conscious awareness and control, Hofbauer, Bushnell, Duncan, & Price, but had been “split off” from the normal 2002). stream of conscious mental activity by traumatic stress – a situation that Janet called desaggregation, or, in English translation, Hypnosis and Hysteria “dissociation.” At least since the late 19th century, interest Although the hegemony of Freudian psy- in hypnosis has had its roots in the med- choanalysis in psychiatry during the first half ical and psychiatric phenomenon known of the 20th century led to a decline of inter- as hysteria (for historical overviews and est in the classical syndromes of hysteria, detailed references, see Kihlstrom, 1994a; the syndrome as such was listed in the early Veith, 1965). This term originated some (1952 and 1968) editions of the Diagnos- 4,000 years ago in ancient Egyptian (and tic and Statistical Manual for Mental Dis- later Greek) medicine to refer to a variety of orders (DSM) published by the American diseases thought to be caused by the migra- Psychiatric Association. Beginning in 1980, tion of the uterus to various parts of the more recent versions of DSM dropped the body. In the 17th century, the English physi- category “hysteria” in favor of separate list- cian Thomas Sydenham reformulated the ings of dissociative disorders – including psy- diagnosis so that hysteria referred to physi- chogenic amnesia and multiple personal- cal symptoms produced by non-organic fac- ity disorder – and conversion disorder, listed tors. In the 19th century, the concept of hys- under the broader rubric of the somato- teria was refined still further, by Briquet, a form disorders (Kihlstrom, 1992b, 1994a). French neurologist, to include patients with As the official psychiatric nosology is cur- multiple, chronic physical complaints with rently constituted, only the functional dis- no obvious organic basis (Briquet, 1859). orders of memory (Kihlstrom & Schac- Sometime later, Charcot noticed that the ter, 2000; Schacter & Kihlstrom, 1989) are symptoms of hysteria mimicked those of cer- explicitly labeled as dissociative in nature. tain neurological illnesses, especially those However, it is clear that the conversion P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

consciousness in hypnosis 451

disorders also involve disruptions in con- Dissociative Phenomena in Hypnosis scious awareness and control (Kihlstrom, 1992b, 1994a; 2001a; Kihlstrom & Schac- As intriguing and historically important as ter, 2000; Kihlstrom, Tataryn, & Hoyt, 1993; the syndromes of hysteria and dissociation Schacter & Kihlstrom, 1989). Renewed are, it is also true that they are very rare interest in the syndromes of hysteria, recon- and for that reason (among others) have strued in terms of dissociations affecting rarely been subject to controlled experi- conscious awareness, was foreshadowed by mental investigation. However, beginning Hilgard’s “neodissociative” theory of divided with Charcot’s observation that hysterical consciousness, which re-established the link patients are highly suggestible, a number of between hypnosis and hysteria (Hilgard, theorists have been impressed by the phe- 1973b, 1977; see also Kihlstrom, 1979, 1992a; notypic similarities between the symptoms Kihlstrom & McGlynn, 1991). of hysteria and the phenomena of hypno- Viewed from a theoretical perspective sis. Accordingly, it has been suggested that centered on consciousness, the dissociative hypnosis might serve as a laboratory model disorders include a number of different syn- for hysteria (Kihlstrom, 1979; Kihlstrom & dromes all involving disruptions in the mon- McGlynn, 1991; see also Oakley, 1999). In itoring and/or controlling functions of con- this way, study of alterations in conscious- sciousness that are not attributable to brain ness in hypnosis might not just help us insult, injury, or disease (Kihlstrom, 1994a, understand hypnosis, but also hysteria and 2001a). These syndromes are reversible, the dissociative and conversion disorders as in the sense that it is possible for the well. In this regard, it is interesting to note patient to recover the lost functions. But that hypnotically suggested limb paralysis even during the symptomatic phase of seems to share neural correlates, as well as the illness, the patient will show evi- surface features, with conversion hysteria dence of intact functioning in the affected (Halligan, Athwal, Oakley, & Frackowiak, system, outside awareness. Thus, patients 2000; Halligan, Oakley, Athwal, & Frack- with psychogenic (dissociative) amnesia, owiak, 2000; Terao & Collinson, 2000). fugue, and multiple personality disorder may show impaired explicit memory but spared Implicit Memory in Posthypnotic Amnesia implicit memory (Kihlstrom, 2001a; Schac- ter & Kihlstrom, 1989). In the same way, Perhaps the most salient alteration in con- patients with conversion disorders affect- sciousness observed in hypnosis is the one ing vision and hearing may show impaired that gave hypnosis its name: posthypnotic explicit perception but spared implicit amnesia. Upon termination of hypnosis, perception (Kihlstrom, 1992b; Kihlstrom, some subjects ﬁnd themselves unable to Barnhardt, & Tataryn, 1992). In light of remember the events and experiences that these considerations, a more accurate tax- transpired while they were hypnotized – an onomy of dissociative disorders (Kihlstrom, amnesia that is roughly analogous to that 1994a) would include three subcategories of experienced after awakening from sleep- syndromes: ing. Posthypnotic amnesia does not occur in the absence of direct or implied suggestions 1 . those affecting memory and identity (e.g., (Hilgard & Cooper, 1965), and the forgot- functional amnesia, fugue, and multiple ten memories are not restored when hyp- personality disorder); nosis is reinduced (Kihlstrom, Brenneman, 2. those affecting sensation and perception Pistole, & Shor, 1985). Posthypnotic amnesia (e.g., functional blindness and deafness, is so named because the subject’s memory analgesia, and tactile anesthesia); is tested in hypnosis, but hypnotic amnesia, 3. those affecting voluntary action (e.g., fun- in which both the suggestion and the test ctional weakness or paralysis of the limbs, occur while the subject is hypnotized, has aphonia, and difﬁculty swallowing). the same properties. Although posthypnotic P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

452 the cambridge handbook of consciousness

amnesia typically covers events and experi- various sorts of priming effects observed in ences that transpired during hypnosis, it is amnesic patients, is for all intents and pur- also possible to suggest amnesia for events poses unconscious memory. that occurred while the subject was not Early evidence that posthypnotic amne- hypnotized (Barnier, 1997; Bryant, Barnier, sia impaired explicit memory but spared Mallard, & Tibbits, 1999). Both features fur- implicit memory came from a pair of exper- ther distinguish posthypnotic amnesia from iments by Kihlstrom (1980), which were state-dependent memory (Eich, 1988). in turn inspired by an earlier investiga- In contrast to the amnesic syndrome asso- tion by Williamsen and his colleagues (see ciated with hippocampal damage, posthyp- also Barber & Calverley, 1966; Williamsen, notic amnesia is temporary: On administra- Johnson, & Eriksen, 1965). Kihlstrom found tion of a prearranged cue, the amnesia is that hypnotizable subjects, given an amne- reversed and the formerly amnesic subject is sia suggestion, were unable to recall the now able to remember the previously forgot- items in a word list that they had mem- ten events (Kihlstrom & Evans, 1976; Nace, orized during hypnosis. However, they Orne, & Hammer, 1974) – although there is remained able to use these same items some evidence that a small residual amnesia as responses on free-association and cat- may persist even after the reversibility cue egory instance-generation tasks. Kihlstrom has been given (Kihlstrom & Evans, 1977). originally interpreted this as reflecting a Reversibility marks posthypnotic amnesia as dissociation between episodic and seman- a disruption of memory retrieval, as opposed tic memory – as did Tulving (1983), who to encoding or storage, somewhat like the cited the experiment as one of four contemporary retrograde amnesias observed vincing demonstrations of the episodic- in individuals who have suffered concus- semantic distinction. However, Kihlstrom sive blows to the head (Kihlstrom, 1985; also noted a priming effect on the pro- Kihlstrom & Evans, 1979). The difference, duction of list items as free associations of course, is that posthypnotic amnesia is a and category instances, compared to con- functional amnesia – an abnormal amount trol items that had not been learned; fur- of forgetting that is attributable to psycho- thermore, the level of priming observed logical factors, rather than to brain insult, was the same as that shown by insuscep- injury, or disease (Kihlstrom & Schacter, tible subjects who were not amnesic for 2000). In fact, as noted earlier, posthypnotic the word list.2 amnesia has long been considered to be a Spared priming during posthypnotic laboratory model of the functional amne- amnesia was subsequently confirmed by sias associated with hysteria and dissociation Spanos and his associates (Bertrand, Spanos, (Barnier, 2002; Kihlstrom, 1979; Kihlstrom & Radtke, 1990; Spanos, Radtke, & Dubreuil, & McGlynn, 1991). 1982), although they preferred to interpret Probably the most interesting psychologi- the results in terms of the demands con- cal research concerning posthypnotic amne- veyed by test instructions rather than disso- sia concerns dissociations between explicit ciations between explicit and implicit mem- and implicit memory (Schacter, 1987), and ory. Later, Dorfman and Kihlstrom (1994) posthypnotic amnesia is no exception. Fol- bolstered the case for spared priming by cor- lowing Schacter (1987), we can identify recting a methodological oversight in the explicit memory with conscious recollec- earlier studies: The comparison of prim- tion, as exemplified by performance on tra- ing with free recall confounded explicit ditional tests of recall and recognition. By and implicit memory with the cue environ- contrast, implicit memory refers to the influ- ment of the memory test. The dissociation ence of some past event on current expe- between explicit and implicit memory was rience, thought, and action in the absence confirmed when a free-association test of of (or independent of) conscious recollec- priming was compared to a cued-recall test tion. Implicit memory, as exemplified by of explicit memory. Similarly, Barnier and P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

consciousness in hypnosis 453

her colleagues extended the dissociation to 1987). Interestingly, source amnesia was explicit and implicit memory for material first identified in the context of hypnosis learned outside as well as within hypnosis (Cooper, 1966; Evans, 1979a,b, 1988; Evans (Barnier, Bryant, & Briscoe, 2001). & Thorne, 1966). Evans and Thorne (1966) Whereas most studies of implicit mem- found that some amnesic subjects retained ory in the amnesic syndrome employ tests world-knowledge that had been taught to of repetition priming, such as stem and frag- them during hypnosis, (e.g., the color an ment completion, the studies just described amethyst turns when exposed to heat or employed tests of semantic priming, which the difference between the antennae of cannot be mediated by a perceptual repre- moths and butterflies), although they did sentation of the stimulus materials. How- not remember the circumstances in which ever, David and his colleagues (David, they acquired this information. In a later Brown, Pojoga, & David, 2000) found study, Evans (1979a) showed that source that posthypnotic amnesia spared repetition amnesia did not occur in insusceptible sub- priming on a stem-completion task. Simi- jects who simulated hypnosis and posthyp- lar results were obtained by Barnier et al. notic amnesia. Although the methodology of (2001). In an especially important twist, Evans’ study has been criticized (Coe, 1978; David et al. employed Jacoby’s process dis- Spanos, Gwynn, Della Malva, & Bertrand, sociation paradigm (Jacoby, 1991) to con- 1988), most of these criticisms pertain to firm that the priming spared in posthypnotic the real-simulating comparison and do not amnesia is a reflection of involuntary uncon- undermine the phenomenon itself. Along scious memory, rather than either involun- with the notion of demand characteris- tary or voluntary conscious memory.3 That is tics (Kihlstrom, 2002a; Orne, 1962, 1973), to say, the spared priming is a genuine reflec- source amnesia is one of the most salient tion of implicit, or unconscious, memory. examples of a concept developed in hyp- With the benefit of hindsight, we can nosis research that has become part of the trace studies of implicit memory in posthyp- common parlance of psychological theory.4 notic amnesia at least as far as the clas- Source amnesia might be interpreted as sic work of Hull (Hull, 1933; Kihlstrom, a form of implicit learning (Berry & Dienes, 2004a), who demonstrated that posthyp- 1993; Reber, 1967, 1993; Seger, 1994). In line notic amnesia impaired recall but had no with the traditional definition of learning, effect on practice effects, savings in relearn- as a relatively permanent change in behaving, or retroactive interference (see further ior that occurs as a result of experience, we discussion below). Hull concluded merely may define implicit learning as the acquisi- that the forgetting observed in posthypnotic tion of new knowledge in the absence either amnesia was “by no means complete” (p. of conscious awareness of the learning expe- 138) – much as Gregg (1979, 1982) later rience or conscious awareness of what has interpreted the evidence as reflecting the dis- been learned, or both. Although evidence tinction dissociation between optional and for implicit learning can be construed as evi- obligatory aspects of memory performance. dence for implicit memory as well (Schac- But we can now interpret the same evidence ter, 1987), we may distinguish between the as illustrating a strong dissociation between two phenomena with respect to the sort explicit and implicit memory. of knowledge affected. In implicit mem- In addition to priming, the dissocia- ory, the memories in question are episodic tion between explicit and implicit mem- in nature, representing more or less discrete ory is revealed by the phenomenon of episodes in the life of the learner. Mem- source amnesia, in which the subject retains ories are acquired in implicit learning as knowledge acquired through some learn- well, of course, but in this case we are con- ing experience while forgetting the learn- cerned with new semantic and procedural ing experience itself (Schacter, Harbluk, & knowledge acquired by the subject. When McClachlan, 1984; Shimamura & Squire, implicit and explicit learning are dissoci- P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

454 the cambridge handbook of consciousness

ated, subjects have no conscious access to the Implicit Perception in Hypnotic Analgesia knowledge – in which case implicit learning counts as a failure of metacognition (Flavell, In addition to their effects on memory, 1979; Metcalfe & Shimamura, 1994; Nelson, hypnotic suggestions can have very dra- 1992,1996; Nelson & Narens, 1990; Reder, matic effects on the experience of pain 1996; Yzerbyt, Lories, & Dardenne, 1998). (Hilgard & Hilgard, 1975; Montgomery, Because the subjects in Evans’ experiments DuHamel, & Redd, 2000). Although hyp- were aware of what they had learned, though notic analgesia was supplanted by more reli- they were amnesic for the learning experi- able chemical analgesia almost as soon as ence, source amnesia is better construed as its efficacy was documented in the mid- an example of implicit memory. 19th century, modern psychophysical stud- Preserved priming on free-association and ies confirm that hypnotizable subjects given category-generation tasks, in the face of suggestions for analgesia can experience impaired recall, is a form of dissociation considerable relief from laboratory pain between explicit and implicit memory. Pre- (Faymonville et al., 2000; Hilgard, 1969; served learning, in the face of amnesia for Knox, Morgan, & Hilgard, 1974). In fact, a the learning experience, is also a form of comparative study found that, among hyp- dissociation between explicit and implicit notizable subjects, hypnotic analgesia was memory. But the case of posthypnotic amne- superior not just to placebo but also to mor- sia is different, in at least three respects, phine, diazepam, aspirin, acupuncture, and from other amnesias in which these disso- biofeedback (Stern, Brown, Ulett, & Sletten, ciations are observed. First, in contrast to 1977). Although hypnosis can serve as the the typical explicit-implicit dissociation, the sole analgesic agent in surgery, it is proba- items in question have been deeply pro- bly used more appropriately as an adjunct cessed at the time of encoding. In the prim- to chemical analgesics, where it has been ing studies, for example, the critical targets shown to be both effective and cost effec- were not just presented for a single trial, tive in reducing actual clinical pain (Lang, but rather were deliberately memorized over Benotsch et al., 2000; Lang, Joyce, Spiegel, the course of several study-test cycles to a Hamilton, & Lee, 1996).5 strict criterion of learning (Dorfman & Hypnotic analgesia is not mediated by Kihlstrom, 1994; Kihlstrom, 1980). Second, relaxation, and the fact that it is not the priming that is preserved is semantic reversed by narcotic antagonists would seem priming, which relies on the formation dur- to rule out a role for endogenous opiates ing encoding and preservation at retrieval (Barber & Mayer, 1977; Goldstein & Hil- of a semantic link between cue and target. gard, 1975; Moret et al., 1991; Spiegel & This priming reflects deep, semantic process- Albert, 1983). There is a placebo coming of a sort that cannot be mediated by a ponent to all active analgesic agents, and perceptual representation system. Third, the hypnosis is no exception; however, hypno- impairment in explicit memory is reversible: tizable subjects receive benefits from hyp- Posthypnotic amnesia is the only case I know notic suggestion that outweigh what they where implicit memories can be restored to or their insusceptible counterparts achieve explicit recollection. from plausible placebos (McGlashan, Evans, Taken together, then, these priming & Orne, 1969; Stern et al., 1977). It has results reflect the unconscious influence of also been argued that hypnotized subjects semantic representations formed as a result employ such techniques as self-distraction, of extensive attentional activity at the time stress inoculation, cognitive reinterpreta- of encoding. The priming itself may be an tion, and tension management to reduce automatic influence, but again it is not the pain (Nolan & Spanos, 1987; Spanos, 1986b). sort that is produced by automatic pro- Although there is no doubt that cognitive cesses mediated by a perceptual represen- strategies can reduce pain, their success, tation system. unlike the success of hypnotic suggestions, P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

consciousness in hypnosis 455

is not correlated with hypnotizability and istration that can be tapped by the hidden thus is unlikely to be responsible for the observer method. effects observed in hypnotizable subjects The paradox of hypnotic analgesia can (Hargadon, Bowers, & Woody, 1995; Miller also be viewed through an extension of & Bowers, 1986, 1993). the explicit-implicit distinction from learn- Rather, Hilgard suggested that hypnotic ing and memory to perception (Kihlstrom, analgesia entails a division of consciousness 1996; Kihlstrom et al., 1992). Explicit per- that prevents the perception of pain from ception refers to the conscious perception of being represented in conscious awareness a stimulus event, whereas implicit percep- (Hilgard, 1973b, 1977). In other words, ver- tion refers to the effect of such an event on bal reports of pain and suffering reflect the the subject’s ongoing experience, thought, conscious perception of pain, whereas phys- and action in the absence of, or indepen- iological responses reflect the processing of dent of, conscious awareness. Just as explicit pain processed outside of conscious aware- and implicit memory can be dissociated in ness. Hilgard’s “hidden observer” is both a the amnesic syndrome and in posthypnotic metaphor for the subconscious perception amnesia, so explicit and implicit perception of pain and a label for a method by which can be dissociated in “subliminal” percep- this subconscious pain can be accessed tion (Marcel, 1983) or prosopagnosia (Bauer, (Hilgard, Morgan, & Macdonald, 1975; Knox 1984). In the case of hypnotic analgesia, et al., 1974). Although it has been suggested explicit perception of the pain stimulus is that hidden observer reports are artifacts reflected in subjects’ self-reports of pain, of experimental demands (Spanos, 1983; whereas implicit perception is reflected in Spanos, Gwynn, & Stam, 1983; Spanos & their physiological responses to the pain Hewitt, 1980), Hilgard showed that both the stimulus. overt and covert pain reports of hypnotized subjects differed from those given by sub- Implicit Perception in Hypnotic Deafness jects who are simulating hypnosis (Hilgard, Hilgard, Macdonald, Morgan, & Johnson, Dissociations between explicit and implicit 1978; Hilgard, Macdonald, Morgan, & John- perception can also be observed in two son, 1978; see also Laurence, Perry, & other classes of hypnotic phenomena. In Kihlstrom, 1983). hypnotic esthesia, the subject experiences a The division in consciousness in hypnotic marked reduction in sensory acuity: Exam- analgesia, as proposed by Hilgard, would ples include hypnotic deafness, blindness, help explain one of the paradoxes of hyp- and tactile anesthesia. In hypnotic negative notic analgesia, which is that it alters sub- hallucinations, the subject fails to perceive jects’ self-reports of pain but has little or no a particular object (or class of objects) in the effect on reflexive, physiological responses environment, but otherwise retains normal to the pain stimulus (e.g., Hilgard & Morgan, levels of sensory function (hypnotized sub- 1975; Hilgard et al., 1974). One interpreta- jects can experience positive hallucinations tion of this difference is that hypnotized sub- as well, perceiving objects that are not actu- jects consciously feel the pain after all. How- ally present in their sensory fields). Although ever, we know on independent grounds that the hypnotic esthesias mimic sensory disor- physiological measures are relatively unsatis- ders, the content-specificity of the negative factory indices of the subjective experience hallucinations marks them as more percep- of pain (Hilgard, 1969). From the perspec- tual in nature. tive of neodissociation theory, the dimin- Careful psychophysical studies, employ- ished self-ratings are accurate reflections of ing both magnitude-estimation (Crawford, the subjects’ conscious experience of pain, Macdonald, & Hilgard, 1979) and signal- whereas the physiological measures show detection (Graham & Schwarz, 1973) that the pain stimulus has been registered paradigms, have documented the loss and processed outside of awareness – a reg- of auditory acuity in hypnotic deafness. P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

456 the cambridge handbook of consciousness

Nevertheless, as is the case in posthyp- subjects would select, from a visually pre- notic amnesia and hypnotic analgesia, sub- sented array, words that were phonetically jects experiencing these phenomena show (but not semantically) similar to words that through their behavior that stimuli in the tar- had been spoken to them (Nash, Lynn, Stan- geted modality continue to be processed, if ley, & Carlson, 1987). Because the hypnotic outside of awareness. For example, Hilgard’s subjects selected fewer such words compared hidden observer has also been observed in to baseline control subjects, this counts as an hypnotic deafness (Crawford et al., 1979). instance of negative phonological priming, Hypnotically deaf subjects continue to man- and thus of implicit perception as well. ifest speech dysfluencies when subjected to delayed auditory feedback (Scheibe, Gray, & Implicit Perception in Hypnotic Blindness Keim, 1968; Sutcliffe, 1961) and in the case of unilateral deafness show substantial num- Similar paradoxes are observed in the visual bers of intrusions from material presented domain. Inspired by an earlier experimen- to their deaf ear (Spanos, Jones, & Malfara, tal case study of hypnotic blindness (Brady, 1982). Nor does hypnotic deafness abol- 1966; Brady & Lind, 1961; see also Bryant ish the “beats” produced by dissonant tones & McConkey, 1989d; Grosz & Zimmerman, (Pattie, 1937) or cardiovascular responses to 1965), Sackeim and his colleagues (Sack- an auditory conditioned stimulus (Sabourin, eim, Nordlie, & Gur, 1979) asked a hyp- Brisson, & Deschamb, 1980). notically blind subject to solve a puzzle in Spanos and Jones (Spanos et al., 1982) which the correct response was indicated preferred to interpret their findings as reveal- by the illumination of a lamp. Performance ing that hypnotically deaf subjects heard per- was significantly below chance. Bryant and fectly well, but Sutcliffe (1960, 1961) offered McConkey (1989a,b) conducted a similar a more subtle interpretation. In his view, the experiment, with a larger group of subjects, persisting effects of delayed auditory feed- generally finding above-chance performance. back certainly contradicted the “credulous” The difference in outcomes may reflect a view that hypnotic deafness was identical number of factors, including the subjects’ to the actual stimulus state of affairs that motivation for the experiment and individ- might arise from damage to the auditory ual differences in cognitive style (Bryant & nerve or lesions in the auditory projec- McConkey, 1990a,b,c), but either outcome tion area – or, for that matter, the simple shows that the visual stimulus was processed absence of an auditory stimulus (Erickson, by the hypnotically blind subjects. 1938a,b). But instead of drawing the “skep- Dissociations between explicit and tical” conclusion that hypnotized subjects implicit perception are also suggested by were engaged in mere role-playing activity, a series of studies by Leibowitz and his Sutcliffe suggested that they were deluded colleagues, who found that ablation of about their experiences – that is, that they the background did not affect perception believed that they heard nothing, when in of the Ponzo illusion (Miller, Hennessy, fact they did. Sutcliffe’s emphasis on delu- & Leibowitz, 1973) and that suggestions sion can be viewed as an anticipation of for tubular (tunnel) vision had no effect Hilgard’s (1977) neodissociation theory of on the size-distance relation (Leibowitz, divided consciousness, where the subjects’ Lundy, & Guez, 1980) or on illusory feelings delusional beliefs reflect their actual phe- of egomotion (roll vection) induced by nomenal experience, and the evidence of viewing a rotating object (Leibowitz, Post, preserved hearing reflects something like Rodemer, Wadlington, & Lundy, 1980). implicit perception. These experiments are particularly inter- Only one study has used priming to exam- esting because they make use of a class of ine implicit perception in hypnotic deafness. perceptual phenomena known as perceptual Nash and his colleagues found that hyp- couplings, which are apparently inviolable notic deafness reduced the likelihood that links between one perceptual organization P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

consciousness in hypnosis 457

and another (Epstein, 1982; Hochberg, hypnotic blindness compared to trials where 1974; Hochberg & Peterson, 1987; Peterson the subjects saw the primes clearly, but any & Hochberg, 1983). If an observer sees evidence of priming counts as evidence of two lines converging in a distance, he or implicit perception – and the magnitude of she must see two identical horizontal bars priming in both studies was substantial by arranged vertically along these lines as any standards. differing in length. In the Miller et al. study, ablation of the converging lines is a failure Color, Meaning, and the Stroop Effect of explicit perception, but the persistence of the perceptually coupled Ponzo illusion In addition to total (binocular or uniocu- indicates that they have been perceived lar) or tubular blindness, hypnotic subjects implicitly. can also be given suggestions for color blind- Perceptual couplings also seem to be ness. Although some early research indi- involved in the ﬁnding of Blum and his col- cated that hypnotic colorblindness affected leagues that hypnotic ablation of surround- performance on the Ishihara test and other ing stimuli did not alter either the magni- laboratory-based tests of color perception tude of the Titchener-Ebbinghaus illusion (Erickson, 1939), the claim has long been (Blum, Nash, Jansen, & Barbour, 1981)orthe controversial (Grether, 1940; Harriman, perception of slant in a target line (Jansen, 1942a,b), and the most rigorous study of Blum, & Loomis, 1982). They are also impli- this type found no effects (Cunningham & cated in the observation that hypnotic anes- Blum, 1982). Certainly, hypnotically color- thesia of the forearm does not affect percep- blind subjects do not show patterns of test tual adaptation of the pointing response to performance that mimic those of the con- displacing prisms (Spanos, Dubreuil, Saad, & genitally colorblind. Nor do hypnotic sug- Gorassini, 1983; Spanos, Gorassini, & Petru- gestions for colorblindness abolish Stroop sic, 1981; Spanos & Saad, 1984). The subjects interference effects (Harvey & Sipprelle, may not feel their arms moving during the 1978; Mallard & Bryant, 2001). All of these pointing trials, but the fact that adaptation results are consistent with the hypothesis occurs indicates that the kinesthetic infor- that color is processed implicitly in hypnot- mation has been processed anyway.6 ically induced colorblindness, even if it is Although the evidence from perceptual not represented in the subjects’ phenome- couplings is consistent with the notion of nal awareness. spared implicit perception, only two stud- However, hypnotic suggestions of a dif- ies have used priming methodologies to seek ferent sort may indeed abolish Stroop inter- evidence of unconscious vision in hypnotic ference. Instead of suggesting that subjects blindness. Bryant and McConkey (1989a) were colorblind, Raz and his colleagues sug- showed subjects pairs of words consisting of gested that the color words were “meaning- a homophone and a disambiguating context less symbols ...like characters of a foreign word (e.g., window-pane), half under con- language that you do not know ...gibberish” ditions of ordinary vision and half during (Raz, Shapiro, Fan, & Posner, 2002,p.1157). hypnotically suggested blindness. On a later The focus on meaning, rather than color, memory test, the subjects generally failed makes this suggestion more akin to the to recall words they had been shown while hypnotic agnosia (or, perhaps, alexia) stud- blind. On a subsequent test, however, when ied by Spanos and his colleagues in rela- the words were presented auditorially, they tion to hypnotic amnesia (Spanos, Radtke tended to spell them in line with their ear- et al., 1982). In contrast to the effects of lier visual presentation (e.g., pane rather than suggested colorblindness, suggested agnosia pain). A subsequent study found a similar completely abolished the Stroop interfer- priming effect on word-fragment comple- ence effect. Subsequent research, employ- tion (Bryant & McConkey, 1994). In both ing a drug to induce cycloplegia and thus cases, priming was diminished somewhat by eliminate accommodation effects, ruled out P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

458 the cambridge handbook of consciousness

peripheral mechanisms, such as visual blur- and implicit emotion (Kihlstrom, Mulvaney, ring or looking away from the stimulus (Raz Tobias, & Tobis, 2000).7 Kihlstrom et al. et al., 2003). However, preliminary fMRI have proposed that, in the absence of self- research suggests that the reduced Stroop reported emotion, behavioral and physio- interference reflects a nonspecific dampen- logical indices of emotional response, such ing of visual information processing (Raz, as facial expressions and heart rate changes, Fan, Shapiro, & Posner, 2002) – a general- might serve as evidence of implicit, uncon- ized effect of visual information processing, scious emotional responding. In fact, a study rather than an effect mediated at linguistic by Bryant and Kourch found that hypnotic or semantic levels. This generalized effect on suggestions for emotional numbing dimin- visual information processing may explain ished self-reported emotional responses, but why Stroop interference did not persist as an had no effect on facial expressions of emo- implicit expression of semantic processing, tion (Bryant & Kourch, 2001). Although despite the conscious experience of agnosia. this finding is suggestive of a dissociation between explicit and implicit expressions of emotion, two other studies found that Implicit Emotion emotional numbing diminished both subjec- Hypnotic suggestions can alter conscious tive reports and facial expressions (Bryant emotion as well as perception and memory. & Mallard, 2002; Weiss et al., 1987). With In fact, the suggested alteration of emotion respect to the dissociation between explicit has been a technique for psychotherapy at and implicit emotion, then, the effects of least since the time of Janet (Ellenberger, hypnotically induced emotional numbing 1970), and has played a role in hypnotic are currently uncertain. studies of psychodynamic processes (Blum, 1961, 1967, 1979; Reyher, 1967). Aside from Anomalies of Dissociation in Hypnosis its inclusion in an advanced scale of hypnotic susceptibility (Hilgard, 1965), the phe- Most of the classic phenomena of hypnosis – nomenon and its underlying mechanisms amnesia, analgesia, and the like – appear to have not been subject to much empirical be dissociative in two related but different study. However, more recent studies leave senses. In the first place, hypnotized sub- little doubt that hypnotic suggestions can jects lack awareness of percepts and mem- alter subjects’ conscious feeling states, just ories that would ordinarily be accessible to as they can alter their conscious percepts consciousness. This disruption in conscious and memories (Bryant & McConkey, 1989c; awareness is the hallmark of the dissociative Weiss, Blum, & Gleberman, 1987). disorders encountered clinically, including As with perception and memory, how- “functional” amnesia and “hysterical” deaf- ever, special interest attaches to the ques- ness. In the second place, these percepts tion of whether the “blocked” emotional and memories continue to influence the responses can nevertheless influence the subject’s ongoing experience, thought, and person’s ongoing experience, thought, and action outside awareness – creating dissocia- action outside of conscious awareness. Until tions between explicit and implicit memory, recently, the idea of unconscious emo- or explicit and implicit perception, similar tion has generally been seen as a holdover to those that have now become quite famil- from an earlier, more psychodynamically iar in the laboratory or neurological clinic. As oriented period in the history of psychol- Hilgard (1977) noted, it is as if consciousness ogy. However, in an era where dissociations has been divided, with one stream of mental between explicit and implicit perception life (e.g., a failure of conscious recollection) and memory are widely accepted as evidence proceeding in phenomenal awareness while of unconscious cognitive processing, there another stream (e.g., the implicit expression seems little reason to reject out of hand the of memory encoding, storage, and retrieval) prospect of dissociations between explicit proceeds outside of awareness. P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

consciousness in hypnosis 459

Co-Consciousness and Trance Logic Bryant, Bibb, & Kihlstrom, 1991). Spanos (Spanos, DeGroot, & Gwynn, 1987) sug- Sometimes, however, the suggested and gested that the occurrence of trance logic actual state of affairs are both represented was an artifact of incomplete response to the in conscious awareness, leading to a set of suggestion on the part of the subject, but this inconsistencies and paradoxes that Orne, in proposal seems to be based on the assump- a classic paper, labeled “trance logic” (Orne, tion that a “complete” image or hallucination 1959). Orne defined trance logic as the would be tantamount to “the real thing” – “apparently simultaneous perception and the actual perceptual state of affairs pro- response to both hallucinations and reality duced by an adequate environmental stimu- without any apparent attempts to satisfy a lus. On the other hand, it may well be that need for logical consistency” (p. 295) – or, as the hallucination is quite complete, in the he often put it in informal conversation, “the sense of being subjectively compelling to the peaceful coexistence of illusion and reality.” person who experiences it – but the accom- For example, in the double hallucination, it panying division of consciousness might be is suggested that the subject will see and incomplete. In this case, trance logic reflects interact with a confederate sitting in a chair a kind of co-consciousness in which two dif- that is actually empty. When the subject’s ferent and mutually contradictory streams of attention is drawn to the real confederate, mental activity – one perceptual, one imag- who has been quietly sitting outside his or inary – are represented simultaneously in her field of vision, Orne reported that hyp- phenomenal awareness. notized subjects typically maintained both the perception of the real confederate and Making the Unconscious Conscious the hallucinations, exhibiting confusion as to which was the real confederate. Similarly, In the case of posthypnotic amnesia and hyp- many subjects reported that they could see notic analgesia, as well as the hypnotic esthe- through the hallucinated confederate to the sias and negative hallucinations, it seems back of the armchair. Thus, the subjects were that hypnotized subjects are able to become simultaneously aware of two mutually con- unaware of percepts and memories that tradictory states of affairs, apparently with- would ordinarily be represented in phenom- out feeling the need to resolve the contra- enal awareness. In contrast, it has some- dictions inherent in the experience. times been suggested that hypnosis also has Orne’s initial report of trance logic was the opposite capacity – to enable subjects somewhat impressionistic in nature, but to become aware of percepts and memo- later investigators have attempted to study ries that would not ordinarily be accessible the phenomenon more quantitatively – to conscious introspection. For example, in with somewhat mixed results (Hilgard, hypnotic hypermnesia subjects receive sug- 1972; R.F.Q. Johnson, 1972; Johnson, Maher, gestions that they will be able to remember & Barber, 1972; McConkey, Bryant, Bibb, events that they have forgotten. In hypnotic Kihlstrom, & Tataryn, 1990; McConkey & age regression, it is suggested that they will Sheehan, 1980; Obstoj & Sheehan, 1977; relive a previous period in their lives – an Sheehan, Obstoj, & McConkey, 1976). On experience that is often accompanied by the the other hand, everyone who has ever apparent recovery of long-forgotten child- worked with hypnotized subjects has seen hood memories. the phenomenon. Although Orne (Orne, Hypermnesia suggestions are sometimes 1959) held the view that trance logic was employed in forensic situations, with for- a defining characteristic of hypnosis, this getful witnesses and victims, or in thera- does not seem to be the case – not least peutic situations, to help patients remember because similar inconsistencies and anoma- traumatic personal experiences. Although lies of response can occur in ordinary imag- field studies have sometimes claimed that ination as well as in hypnosis (McConkey, hypnosis can powerfully enhance memory, P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

460 the cambridge handbook of consciousness

these reports are mostly anecdotal in nature sis enhanced the recovery of valid mem- and generally fail to seek independent cor- ory of actual personal experiences (Hofling, roboration of the memories produced dur- Heyl, & Wright, 1971). This study has not ing hypnosis. Moreover, they have not been been replicated, however, and another study, supported by studies run under labora- also employing lifelike stimulus materials – tory conditions. A report by the Com- a gangland assassination staged before an mittee on Techniques for the Enhance- audience of law enforcement officers – ment of Human Performance, a unit of the found no advantage for hypnosis whatso- U.S. National Research Council, concluded ever (Timm, 1981). Perhaps not surpris- that gains in recall produced by hypnotic ingly, many legal jurisdictions severely limit suggestion were rarely dramatic and were the introduction of memories recovered matched by gains observed when subjects through hypnosis, out of a concern that such are not hypnotized (Kihlstrom & Eich, 1994; memories may be unreliable and tainted Nogrady, McConkey, & Perry, 1985). In fact, by suggestion and inappropriately high lev- there is some evidence from the labora- els of confidence. An abundance of caution tory that hypnotic suggestion can interfere seems to be appropriate in this instance, but with normal hypermnesic processes (Reg- in the present context it seems that hyp- ister & Kihlstrom, 1987). To make things notic suggestion is better at making percepts worse, any increases obtained in valid recol- and memories inaccessible to consciousness lection can be met or exceeded by increases than it is at making unconscious percepts in false recollections (Dywan & Bowers, and memories accessible to phenomenal 1983). Moreover, hypnotized subjects (espe- awareness. cially those who are highly hypnotizable) may be vulnerable to distortions in memory produced by leading questions and Automaticity in Hypnosis other subtle, suggestive influences (Sheehan, 1988). Even before the discovery of implicit mem- Similar conclusions apply to hypnotic ory and the rediscovery of “subliminal” per- age regression (Nash, 1987). Although age- ception, psychology’s renewed interest in regressed subjects may experience them- unconscious mental life was signalled by the selves as children and may behave in a child- general acceptance of a distinction between like manner, there is no evidence that they automatic and controlled mental processes. actually undergo either abolition of char- As a first approximation, automatic pro- acteristically adult modes of mental func- cesses are executed unconsciously in a tioning or reinstatement of childlike modes reflex-like fashion, whereas controlled proof mental functioning. Nor do age-regressed cesses are executed consciously and delib- subjects experience the revivification of for- erately (Kihlstrom, 1987, 1994b). A popu- gotten memories of childhood. Hypnotic age lar example of automaticity is the Stroop regression can be a subjectively compelling color-word effect, in which subjects have experience for subjects, but it is first and difficulty naming the colors in which words foremost an imaginative experience. As with are printed when the words themselves hypnotic hypermnesia, any memories recov- name a different color (MacLeod, 1991, 1992; ered during hypnotic age regression cannot Stroop, 1935). Despite the subjects’ con- be accepted at face value, in the absence of scious intention to name the ink colors independent corroboration. and to ignore the words, they automatically Some clinical practitioners have objected process the words anyway, and this pro- to these conclusions, on the ground that cessing activity interferes with the naming laboratory studies of memory generally task. lack ecological validity (Brown, Scheflin, & According to traditional formulations Hammond, 1998). In fact, one diary-based (LaBerge & Samuels, 1974; Posner & Snyder, study did find some evidence that hypno- 1975; Schneider & Shiffrin, 1977; Shiffrin & P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

consciousness in hypnosis 461

Schneider, 1977, 1984), automatic processes as voluntary doings (Sarbin & Coe, 1972). share five properties in common: Not all responses to hypnotic suggestion are experienced as completely involuntary, but 1. Inevitable Evocation: Automatic pro- the experience is strongest among those who cesses are necessarily engaged by the are most highly hypnotizable (Bowers, 1982; appearance of specific cues in the stimu- Bowers, Laurence, & Hart, 1988). lus environment, independent of the per- Automaticity lies at the heart of the son’s conscious intentions. “social cognitive” theory of hypnosis pro- 2. Incorrigible Execution: Once invoked, posed by Kirsch and Lynn (Kirsch, 2000; automatic processes proceed unalterably Kirsch & Lynn, 1997, 1998b), which asserts to their conclusion and cannot be modi- that hypnotic behaviors are generated auto- fied by conscious activity. matically by subjects’ expectancies that they 3. Effortlessness: The execution of an auto- will occur – much in the manner of a matic process consumes little or no self-fulfilling prophecy (Rosenthal & Rubin, attentional resources and therefore does 1978; Snyder & Swann, 1978). This view, in not interfere with other ongoing mental turn, is rooted in James’s (1890) theory of processes. ideomotor action (see also Arnold, 1946), 4. Speed: Automatic processes are executed which held that motor behavior was gener- rapidly, on the order of seconds or even ated automatically by the person’s idea of fractions of a second – too quickly to be it. Conscious control over behavior, then, is vulnerable to conscious control. accomplished by exerting conscious control over one’s cognitive and other mental states; 5. Unavailability: Perhaps because they con- but once a subject attends to a particular sume no attentional resources, perhaps idea, the resulting behavior occurs naturally. because they are fast, or perhaps because Kirsch and Lynn’s social cognitive, ideo- they are represented as procedural rather motor theory of hypnosis is distinct from than declarative knowledge (Anderson, Spanos’s “sociocognitive” approach (Spanos, 1992), automatic processes are uncon- 1986b), which holds either that subjects fab- scious in the strict sense of being unavail- ricate reports of involuntariness to convince able to conscious introspection in princi- the hypnotist that they are, in fact, deeply ple, and they can be known only by infer- hypnotized (Spanos, Cobb, & Gorassini, ence from performance data. 1985) or that certain features of the hypnotic context lead subjects to misattribute their responses to the hypnotist’s sugges- The Experience of Involuntariness tions, instead of to their own voluntary in Hypnosis actions (Spanos, 1986a). Spanos’s latter view, As indicated at the outset of this chapter, that the hypnotic experience of involuntari- there is much about hypnosis that appears ness is illusory, was also embraced by Weg- to be automatic. Indeed, the experience of ner (2002; but see Kihlstrom, 2004b). Work- involuntariness – sometimes called the clas- ing from a neuropsychological perspective, sic suggestion effect (Weitzenhoffer, 1974)– Woody and Bowers have suggested that is part and parcel of the experience of hyp- the experience of involuntariness is a gen- nosis. Hypnotic subjects don’t simply imag- uine reflection of the effects of hypnosis on ine heavy objects in their hands and allow frontal-lobe structures involved in executive their arms to lower accordingly. They out- functioning (Woody & Bowers, 1994; Woody stretch their hands voluntarily, as an act of & Sadler, 1998). ordinary compliance with the hypnotist’s On the other hand, it is possible that instruction or request to do so, but when the hypnotic experience of involuntariness he or she starts giving the suggestion they is illusory after all – though not for the rea- feel the heaviness in their hands, their arms sons suggested by Spanos and Wegner. After drop, as involuntary happenings rather than all, as Shor noted, “A hypnotized subject P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

462 the cambridge handbook of consciousness

is not a will-less automaton. The hypnotist some views of hypnosis as an altered state does not crawl inside a subject’s body and of consciousness, because the phenomenon take control of his brain and muscles” occurs after the hypnotic state has been (Shor, 1979b, p. 127). From the frame- ostensibly terminated. So far as we can work of Hilgard’s neodissociation theory of tell, subjects do not re-enter hypnosis while divided consciousness (Hilgard, 1977; see they are responding to the posthypnotic also Kihlstrom, 1992a), the experience of suggestion. At least, they are not particu- involuntariness reﬂects an amnesia-like bar- larly responsive to other hypnotic sugges- rier that impairs subjects’ conscious aware- tions during this time (Reyher & Smyth, ness of their own role in producing hypnotic 1971). We cannot say that hypnosis caused responses. In this view, the hypnotic subject the behavior to occur, because the sub- actively imagines a heavy object in his out- jects are not hypnotized when they make stretched hand, and actively lowers his hand their response. Nevertheless, some alteration and arm as if it were heavy, but is not aware of consciousness has occurred, because at of doing so. Thus, the subject’s behavior is the very least they are not aware of what technically voluntary in nature, but is experi- they are doing or why (Sheehan & Orne, enced as involuntary – as occurring automat- 1968). ically – because the subject is unaware of his In the present context, posthypnotic sug- or her own role as the agent of the behavior. gestion is of interest because it seems to In other words, the apparent disruption of occur automatically in response to the prear- conscious control actually occurs by virtue of ranged cue (Erickson & Erickson, 1941). Cer- a disruption of conscious awareness – a pro- tainly posthypnotic suggestion differs from posal that (perhaps) gains credence from the ordinary behavioral compliance. Damaser dissociations between explicit and implicit (Damaser, 1964; see also Orne, 1969)gave memory and perception discussed earlier. subjects a posthypnotic suggestion to mail the experimenter one postcard per day, a control group received an ordinary social Automaticity in Posthypnotic Suggestion request to perform the same behavior, and a Perhaps the most dramatic demonstration third group received both the posthypnotic of apparent automaticity in hypnosis is suggestion and the social request. Surpris- posthypnotic suggestion, in which the sub- ingly, the subjects who received the social ject responds after the termination of hyp- request mailed more postcards than did nosis to a suggestion administered while he those who received only the posthypnotic or she was still hypnotized. On the group- suggestion (see also Barnier & McConkey, administered HGSHS:A, for example, it 1999b). Apparently, those who agreed to is suggested that when the subjects hear the social request felt that they were under two taps, they will reach down and touch some obligation to carry it out, but those their left ankles, but forget that they were who received the posthypnotic suggestion instructed to do so. After the termination of carried it out only so long as they felt the hypnosis, many highly hypnotizable subjects urge to do so. This urge can be powerful: will respond quickly to such a prearranged Subjects who fail to respond to a posthyp- cue – without knowing why they are doing so notic suggestion on an initial test appear to or confabulating a reason, such as that they show a persisting tendency to perform the feel an itch. They may even be unaware that suggested behavior at a later time (Nace & they are doing anything unusual at all. Orne, 1970). Posthypnotic behavior can per- Any suggested experience that can occur sist for long periods of time (Edwards, 1963), during hypnosis can also occur posthypnoti- even after the posthypnotic suggestion has cally, provided that the subject is sufﬁciently been formally canceled (Bowers, 1975). hypnotizable. For this reason, posthypnotic Nevertheless, close examination shows suggestion has always been problematic for that posthypnotic behavior does not meet P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

consciousness in hypnosis 463

the technical definition of automaticity, and influences subsequent behavior in the as it has evolved within cognitive psy- absence of conscious recollection. chology (Barnier, 1999). In the first place, posthypnotic suggestion fails the test of inevitable evocation. Except under special Hypnosis in Mind and Body circumstances (Orne, Sheehan, & Evans, 1968), response to a posthypnotic sugges- Researchers have long been interested in bio- tion declines markedly outside the experi- logical correlates of hypnosis. In the 19th mental context in which the suggestion is century, Braid likened hypnosis to sleep, originally given (Barnier & McConkey, 1998; whereas Pavlov considered it to be a state of Fisher, 1954; Spanos, Menary, Brett, Cross, cortical inhibition (Gauld, 1992). In the mid- & Ahmed, 1987). Moreover, like all other 20th century revival of interest in conscious- aspects of hypnosis, posthypnotic behavior ness, some theorists speculated that hyp- depends intimately on the both the subject’s nosis entailed an increase in high-voltage, interpretation of the hypnotist’s suggestion low-frequency alpha activity in the EEG, and the context in which the cue appears though this proved to be an artifact of relax- (Barnier & McConkey, 1999a, 2001). It is ation and eye closure (Dumas, 1977; Evans, in no sense reflexive in nature. Moreover, 1979b). The discovery of hemispheric spe- posthypnotic suggestion is not effortless. Sub- cialization, with the left hemisphere geared jects respond to simple posthypnotic sug- to analytic and the right hemisphere to non- gestions more frequently than to complex analytic tasks, coupled with the notion that ones (Barnier & McConkey, 1999c), sug- the right hemisphere is “silent” or uncon- gesting that the activity makes demands on scious,” led to the speculation that hypnotic the subject’s information-processing capac- response is somehow mediated by right- ity. Responding to a posthypnotic sugges- hemisphere activity (Bakan, 1969). Studies tion interferes with responding to a waking employing both behavioral and electrophys- instruction, even when the response require- iological paradigms (e.g., MacLeod-Morgan ments of the two tasks do not conflict (Hoyt & Lack, 1982; Sackeim, 1982) have been & Kihlstrom, 1986). Thus, responding to a interpreted as indicating increased activa- posthypnotic suggestion seems to consume tion of the right hemisphere among highly more information-processing capacity than hypnotizable individuals, but positive results would be expected of a truly automatic have proved difficult to replicate (e.g., process. Graffin, Ray, & Lundy, 1995; Otto-Salaj, Posthypnotic suggestion does not appear Nadon, Hoyt, Register, & Kihlstrom, 1992), to be an instance of automaticity, but it and interpretation of these findings remains does appear to be an instance of prospec- controversial. tive memory (Einstein & McDaniel, 1990), It should be understood that hypnosis in which subjects must remember to per- is mediated by verbal suggestions, which form a specified activity at some time must be interpreted by the subject in the in the future. Awareness of the posthyp- course of responding. Thus, the role of notic suggestion does not seem to inter- the left hemisphere should not be min- fere with posthypnotic behavior (Barnier imized (Jasiukaitis, Nouriani, Hugdahl, & & McConkey, 1999c; Edwards, 1956; Gan- Spiegel, 1995; Rainville, Hofbauer, Paus, dolfo, 1971). But when accompanied by Bushnell, & Price, 1999). One interesting posthypnotic amnesia, posthypnotic behav- proposal is that hypnotizable individuals ior takes on some of the qualities of implicit show greater flexibility in deploying the left memory. Even though subjects may for- and right hemispheres in a task-appropriate get the suggestion, the fact that they carry manner, especially when they are actually out the suggestion on cue shows clearly that hypnotized (Crawford, 2001; Crawford & the prospective memory has been encoded Gruzelier, 1992). Because involuntariness P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

464 the cambridge handbook of consciousness

is so central to the experience of hyp- Maquet et al. asked their subjects to review nosis, it has also been suggested that the a pleasant life experience (Maquet et al., frontal lobes (which organize intentional 1999). action) may play a special role in hyp- Although the concept of “neutral” hypnosis, and especially in the experience of nosis has had its proponents (Kihlstrom involuntariness (Woody & Bowers, 1994; & Edmonston, 1971), in subjective terms Woody & Sadler, 1998). Along these lines, the state, such as it is, differs little from Farvolden and Woody have found that eyes-closed relaxation (Edmonston, 1977, highly hypnotizable individuals perform 1981) and bears little resemblance to the relatively poorly on neuropsychological dissociative and hallucinatory experiences tasks that assess frontal-lobe functioning associated with specific hypnotic sugges- (Farvolden & Woody, 2004). tions. Moreover, it is unlikely that imaging subjects who are merely in neutral hypnosis and not responding to particular hyp- “Neutral” Hypnosis notic suggestions will tell us much about the Although most work on the neural correlates neural correlates of hypnosis, because the of hypnosis has employed psychophysiolog- experiences of hypnotic subjects are so var- ical measures such as the EEG and event- ied, depending on the suggestion to which related potentials, it seems likely that a bet- they are responding. A more fruitful tack ter understanding of the neural substrates will likely involve imaging subjects while of hypnosis may come from the applica- they are responding to particular hypnotic tion of brain imaging technologies (Barnier & suggestions. Just as the neural correlates of McConkey, 2003; Killeen & Nash, 2003; Ray NREM sleep differ from those of REM sleep & Tucker, 2003; Woody & McConkey, 2003; (Hobson, Pace-Schott, & Stickgold, 2000), Woody & Szechtman, 2003). One approach so the neural correlates of neutral hypnosis has been to scan subjects after they have will differ from those of specific, suggested received a hypnotic induction but before hypnotic phenomena. they have received any specific suggestions, on the assumption that such a procedure will Hypnotic Analgesia reveal the neural correlates (if indeed any exist) of hypnosis as a generalized altered Perhaps because of the added interest value state of consciousness. For example, one PET that comes with clinical application, most study found that the induction of hypnosis brain imaging studies of hypnotic sugges- generated widespread activation of occipital, tions have focused on analgesia. A pio- parietal, precentral, premotor, and ventro- neering study using the 133 Xe technique lateral prefrontal cortex in the left hemi- found bilateral increases in the activation sphere, and the occipital and anterior cin- of the orbitofrontal region, as well as in gulate cortex of the right hemisphere – in somatosensory cortex, during analgesia com- other words, pretty much the entire brain pared to resting baseline and a control (Maquet et al., 1999). At the same time, condition in which subjects attended to another PET study found that the induction the pain (Crawford, Gur, Skolnick, Gur, & of hypnosis was accompanied by increased Benson, 1993). They suggested that these activation of occipital cortex and decreases changes reflected the increased mental effort in the right inferior parietal lobule, left pre- needed to actively inhibit the process- cuneus, and posterior cingulate (Rainville, ing of somatosensory information. A more Hofbauer et al., 1999). As is so often the recent PET study implicated quite different case in brain imaging experiments, the dif- regions, particularly the anterior cingulate ference in results may be due to differences cortex (ACC). However, this later study also in control conditions. Whereas Rainville employed quite a different procedure, mod- et al. asked their hypnotized subjects simply ulating pain perception through a pleasant to relax (Rainville, Hofbauer et al., 1999), autobiographical reverie instead of a specific P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

consciousness in hypnosis 465

suggestion for analgesia (Faymonville et al., ing from the external stimulus environment 2000). (Woody & Szechtman, 2000a,b). Another Because the specific wording of sugges- interpretation, based on the role of ACC tions is so important in hypnosis, perhaps in emotion, is that the activity in this the most interesting brain imaging studies region reflects affective arousal to experi- of analgesia compared suggestions targeting ences, whether perceptual or hallucinatory, sensory pain, which relates to the location which surprise the subject; mental images, and physical intensity of the pain stimu- being deliberately constructed by the sub- lus, with suggestions targeting suffering, or ject, would not have this surprise value. the meaning of the pain (Melzack, 1975). In another study, Kosslyn and his col- Standard hypnotic suggestions for analge- leagues studied the modulation of color sia affect both sensory pain and suffering perception through hypnotic suggestion (Hilgard & Hilgard, 1975), but these two (Kosslyn et al., 2000). After PET imaging dimensions can also be dissociated by alter- identified a region (in the fusiform area) that ing the specific wording of the suggestion was differentially activated by the presen- (Rainville, Carrier, Hofbauer, Bushnell, & tation of chromatic and gray-scale stimuli, Duncan, 1999). Using hypnotic suggestions, these investigators gave suggestions to highly Rainville and his colleagues have found that hypnotizable subjects that they would per- suggestions that alter the unpleasantness of a ceive the colored stimulus in gray scale, pain stimulus, without altering its intensity, and the gray-scale stimulus as colored. The are associated with changes in ACC but not result was that the fusiform region was acti- in somatosensory cortex (Rainville, Duncan, vated in line with subjects’ perceptions – Price, Carrier, & Bushnell, 1997; Rainville actual and hallucinated color or actual and et al., 2002). hallucinated gray scale, independent of the stimulus. In contrast to nonhypnotic color imagery, which appears to activate only Hallucinations and Imagery the right fusiform region (Howard et al., Brain imaging studies also bear on the rela- 1998), hypnotically hallucinated color acti- tion between hypnotic hallucinations and vated both the left and right hemispheres. normal imagery. On the surface, at least, Taken together with the Szechtman et al. imagery would seem to be a cognitive study (1998), these results suggest that hyp- skill relevant to hypnosis, and some theo- notic hallucinations are in at least some sense rists sometimes write as if hypnosis were distinct from mental images. only a special case of a larger domain of mental imagery (for reviews, see Bow- ers, 1992; Glisky et al., 1995; Kunzen- Brain States and States dorf, Spanos, & Wallace, 1996; Sheehan, of Consciousness 1982). On the contrary, Szechtman and his colleagues found that hypnotized subjects The controversy over the very nature of hyp- experiencing suggested auditory hallucina- nosis has often led investigators to seek evi- tions showed activation of the right ACC; dence of neural and other biological changes this area was also activated during normal to demonstrate that hypnosis is “real” – or, hearing, but not during auditory imagery alternatively, to debunk the phenomenon as (Szechtman, Woody, Bowers, & Nahmias, illusion and fakery. For example, the lack of 1998). Interestingly, a parallel study found reliable physiological correlates of hypnotic that schizophrenic patients also showed response has been interpreted by Sarbin right ACC activation during their audi- as supporting his role-enactment interpre- tory hallucinations (Cleghorn et al., 1992). tation of hypnosis (Sarbin, 1973; Sarbin & They suggested that activation of this region Slagle, 1979). On the other hand, Kosslyn might cause internally generated thoughts and his colleagues argued that the activity and images to be confused with those aris- of the fusiform color area in response to P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

466 the cambridge handbook of consciousness

suggestions for altered color vision “support that makes for consciousness. And the neu- the claim that hypnosis is a psychological ral correlates of that difference are the neural state with distinct neural correlates and is correlates of consciousness. not just the result of adopting a role” (Koss- lyn et al., 2000,p.1279). Neither position is quite correct. Physio- Acknowledgements logical correlates are nice when they exist, and they may enable otherwise skeptical The point of view represented in this paper is observers to accept the phenomena of hyp- based on research supported by Grant #MH- nosis as real. But such correlates are neither 35856 from the National Institute of Mental necessary nor sufficient to define an altered Health. state of consciousness. In the final analysis, consciousness is a psychological construct, not a biological one, and can only be defined Notes at a psychological level of analysis. The phenomena of hypnosis – amnesia, analgesia, 1. This was true even before hypnosis received positive and negative hallucinations, and the its name (Braid, 1843; Gravitz & Gerton, like – obviously represent alterations in con- 1984; Kihlstrom, 1992c) – and for that mat- scious perception and memory. The neural ter even before that, the status of hypnosis correlates of these phenomena are a matter as an altered organismal state was controversial. In the 18th century, Mesmer thought his of considerable interest, but they are another “crises” were induced by animal magnetism, matter entirely. but the Franklin Commission chalked them At the same time, the phenomena of hyp- up to mere imagination (Kihlstrom, 2002b). In nosis seem to offer a unique vantage point the 19th century, Charcot thought that hypno- from which consciousness and its neural cor- sis was closely related to hysteria and to neu- relates can be studied, because they remind rological disease, whereas Liebeault and Bern- us that consciousness is not just a matter of heim attributed its effects to simple sugges- attention and alertness. Mental states are also tion. Perhaps because he was writing in the a matter of aboutness: They have intentional- heyday of functional behaviorism, Hull (1933) ity, in that they refer to objects that exist and did not confront the “state-nonstate” issue: For events that occur in the world outside the him, hypnosis was an intrinsically interesting phenomenon that psychology ought to be able mind. Hypnotized subjects are conscious, in to explain (Kihlstrom, 2004a). the sense of being alert and attentive, but 2. Lacking the explicit-implicit distinction sub- when certain suggestions are in effect they sequently introduced by Schacter (see also are not conscious of some things – of some Graf & Schacter, 1985; Schacter, 1987; Schac- event in the past or some object in their cur- ter&Graf,1986), Kihlstrom noted simply that rent environment. The fact that percepts and the priming represented “a residual effect of memories can be explicit or implicit means the original learning episode on a subsequent that mental states themselves can be con- task involving retrieval from ‘semantic’ mem- scious or unconscious. ory” (p. 246), that it “took place outside of phe- The phenomena of hypnosis remind us nomenal awareness,” and that it was “similar to that there is a difference between being one which occurs in patients diagnosed with 246 aware of something explicitly and being the amnesic syndrome” (p. ). A similar 1985 unaware of something that nonetheless, interpretation appeared in (Kihlstrom, 1985), in a paper that had been written in implicitly influences our ongoing experi- 1984, and the relevance of the explicit-implicit ence, thought, and action. Almost uniquely, distinction was made explicit (sorry) in 1987 hypnosis allows us to create, and reverse, (Kihlstrom, 1987). dissociations between the explicit and 3. Interestingly, David et al. obtained a similar the implicit – between the conscious and the pattern of results for directed forgetting in unconscious – at will in the laboratory. The the normal waking state. Posthypnotic amne- difference between implicit and explicit per- sia and directed forgetting are both exam- cepts and memories, then, is the difference ples of retrieval inhibition (Anderson & Green, P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

consciousness in hypnosis 467

2001; Anderson et al., 2004; Geiselman, Bjork, Glover, G. H., & Gabrieli, J. D. E. (2004). & Fishman, 1983; Levy & Anderson, 2002), Neural systems underlying the suppression of but the two paradigms generally differ greatly unwanted memories. Science, 303, 232–235. in other respects (Kihlstrom, 1983); for exam- Arnold, M. B. (1946). On the mechanism of sug- ple, the role of incidental or intentional learn- gestion and hypnosis. Journal of Abnormal and ing, the amount of study devoted to the items, Social Psychology, 41, 107–128. the temporal location of the cue to forget. As, A. (1962). Non-hypnotic experiences related the retention interval involved, and the means to hypnotizability in male and female college by which memory is measured – as well as students. Scandinavian Journal of Psychology, 3, the degree to which the to-be-forgotten items 112–121. are actually inaccessible, whether the forget- Aserinsky, E., & Kleitman, N. (1953). Regularly ting is reversible, and the extent of inter- occurring periods of eye motility, and concomi- ference between to-be-forgotten and to-be- tant phenomena, during sleep. Science, 118, remembered items. 273–274. 4. Source amnesia is a failure of source monitor- Bakan, P. (1969). Hypnotizability, laterality of eye ing (Johnson, Hashtroudi, & Lindsay, 1993), a movements and functional brain asymmetry. process that in turn is closely related to reality Perceptual and Motor Skills, 28, 927–932. monitoring (Johnson & Raye, 1981). It proba- 1941 bly lies at the heart of the experience of déjàvu Banister, H., & Zangwill, O. L. ( a). Experi- (Brown, 2003). As noted by Evans and Thorne mentally induced olfactory paramnesia. British 155 175 (1966), their work had been anticipated by Journal of Psychology, 32, – . Banister and Zangwill (1941a,b) who used hyp- Banister, H., & Zangwill, O. L. (1941b). Exper- notic suggestion to produce visual and olfac- imentally induced visual paramnesias. British tory “paramnesias” in which subjects recognize Journal of Psychology, 32, 30–51. a previously studied item but confabulate the Barber, J., & Mayer, D. (1977). Evaluation of effi- context in which it has been studied. cacy and neural mechanism of a hypnotic anal- 5. A thorough discussion of experimental and gesia procedure in experimental and clinical clinical research on hypnotic analgesia is dental pain. Pain, 4, 41–48. beyond the scope of this chapter. Interested Barber, T. X. (1969). Hypnosis: A scientific readers may wish to consult Kihlstrom (2000, approach. New York: Van Nostrand Reinhold. 2001 b). Barber, T. X., & Calverley, D. S. (1966). Toward 6. Note, however, Wallace and his colleagues a theory of “hypnotic” behavior: Experimental have found that hypnotic anesthesia actu- analyses of suggested amnesia. Journal of Abnor- ally abolishes prism adaptation, so this find- mal Psychology, 71, 95–107. 1980 ing remains in some dispute (Wallace, ; Barnier, A. J. (1997). Autobiographical amnesia: 1982 1984 Wallace & Fisher, , a,b; Wallace & An investigation of hypnotically created per- 1973 1975 Garrett, , ). sonal forgetting. Proposal to Australian Research 7. McClelland and his colleagues have made a Council. distinction between explicit (conscious) and Barnier, A. J. (1999). Posthypnotic suggestion: implicit (unconscious) motivation, as well Attention, awareness, and automaticity. Sleep 1989 (McClelland, Koestner, & Weinberger, ), & Hypnosis, 1, 57–63. but to date there have been no studies of hyp- Barnier, A. J. (2002). Posthypnotic amnesia for nosis along these lines. autobiographical episodes: A laboratory model of functional amnesia? Psychological Science, 13(3), 232–237. References Barnier, A. J., Bryant, R. A., & Briscoe, S. (2001). Posthypnotic amnesia for material learned Anderson, J. R. (1992). Automaticity and the before or during hypnosis: Explicit and implicit ACT* theory. American Journal of Psychology, memory effects. International Journal of Clini- 105(2), 165–180. cal and Experimental Hypnosis, 49(4), 286–304. Anderson, M. C., & Green, C. (2001, March). Barnier, A. J., & McConkey, K. M. (1998). Suppressing unwanted memories by executive Posthypnotic responding: Knowing when to control. Nature, 410(15), 366–369. stop helps to keep it going. International Journal Anderson, M. C., Ochsner, K. N., Kuhl, B., of Clinical & Experimental Hypnosis, 46, 204– Cooper, J., Robertson, E., Gabrieli, S. W., 219. P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

468 the cambridge handbook of consciousness

Barnier, A. J., & McConkey, K. M. (1999a). visual illusion as reflected in perceptual reports Hypnotic and posthypnotic suggestion: Find- and cortical event related potentials. Academic ing meaning in the message of the hypnotist. Psychology Bulletin, 3, 251–271. International Journal of Clinical & Experimental Bowers, K. S. (1967). The effect for demands of Hypnosis, 47, 192–208. honesty upon reports of visual and auditory Barnier, A. J., & McConkey, K. M. (1999b). hallucinations. International Journal of Clinical Posthypnotic responding away from the hyp- and Experimental Hypnosis, 15, 31–36. notic setting. Psychological Science, 9, 256–262. Bowers, K. S. (1975). The psychology of Barnier, A. J., & McConkey, K. M. (1999c). subtle control: An attributional analysis of Posthypnotic suggestion, response complexity, behavioural persistence. Canadian Journal and amnesia. Australian Journal of Psychology, of Behavioral Science, 7, 78–95. 51(1), 1–5. Bowers, K. S. (1992). Imagination and dissocia- Barnier, A. J., & McConkey, K. M. (2001). Post- tion in hypnotic responding. International Jour- hypnotic responding: The relevance of nal of Clinical & Experimental Hypnosis, 40, suggestion and test congruence. International 253–275. Journal of Clinical and Experimental Hypnosis, Bowers, P. (1982). The classic suggestion effect: 49, 207–219. Relationships with scales of hypnotizability, Barnier, A. J., & McConkey, K. M. (2003). Hyp- effortless experiencing, and imagery vividness. nosis, human nature, and complexity: Inte- International Journal of Clinical and Experimen- grating neuroscience approaches into hypnosis tal Hypnosis, 30, 270–279. research. International Journal of Clinical and Bowers, P., Laurence, J. R., & Hart, D. (1988). Experimental Hypnosis, 51(3), 282–308. The experience of hypnotic suggestions. Inter- Bates, B. L., & Kraft, P. M. (1991). The nature national Journal of Clinical and Experimental of hypnotic performance following administra- Hypnosis, 36, 336–349. tion of the Carleton Skills Training Program. Brady, J. P. (1966). Hysteria versus malingering: A International Journal of Clinical & Experimental response to Grosz & Zimmerman Bound with Hypnosis, 39, 227–242. Brady & Lind. Behavior Research & Therapy, 4, 321 322 Bauer, R. M. (1984). Autonomic recognition of – . names and faces in prosopagnosia: A neuropsy- Brady, J. P., & Lind, D. I. (1961). Experimental chological application of the guilty knowledge analysis of hysterical blindness. Archives of Gen- test. Neuropsychologia, 22, 457–469. eral Psychiatry, 4, 331–339. Berry, D. C., & Dienes, Z. (1993). Implicit learn- Braid, J. (1843). Neurypnology: or the rationale of ing: Theoretical and empirical issues. Hove, UK: nervous sleep considered in relation to animal Erlbaum. magnetism. London: Churchill. Bertrand, L. D., Spanos, N. P., & Radtke, H. L. Briquet, P. (1859). Traite clinque et therapeutique (1990). Contextual effects on priming during a` l’hysterie. Paris: Balliere et Fils. hypnotic amnesia. Journal of Research in Per- Brown, A. S. (2003). A review of the déjàvu sonality, 24, 271–290. experience. Psychological Bulletin, 129, 394– Blum, G. S. (1961). A model of the mind: Explored 413. by hypnotically controlled experiments and exam- Brown, D., Scheflin, A. W., & Hammond, D. C. ined for its psychodynamic implications. New (1998). Memory, trauma treatment, and the law. York: Wiley. New York: W.W. Norton. Blum, G. S. (1967). Hypnosis in psychodynamic Bryant, R. A., Barnier, A. J., Mallard, D., & research. In J. E. Gordon (Ed.), Handbook of Tibbits, R. (1999). Posthypnotic amnesia for clinical and experimental hypnosis (pp. 83–109). material learned before hypnosis. International New York: Macmillan. Journal of Clinical & Experimental Hypnosis, 47, Blum, G. S. (1979). Hypnotic programming 46–64. techniques in psychological experiments. In Bryant, R. A., & Kourch, M. (2001). Hypnoti- E. Fromm & R. E. Shor (Eds.), Hypnosis: cally induced emotional numbing. International Developments in Research and new perspectives Journal of Clinical and Experimental Hypnosis, (pp. 457–481). New York: Aldine. 49(3), 220–230. Blum, G. S., Nash, J. K., Jansen, R. D., & Bar- Bryant, R. A., & Mallard, D. (2002). Hypnotically bour, J. S. (1981). Posthypnotic attenuation of a induced emotional numbing: A real-simulating P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

consciousness in hypnosis 469

analysis. Journal of Abnormal Psychology, 111, R. O. Stanley, & P. B. Bloom (Eds.), Advances in 203–207. clinical hypnosis (pp. 61–84). New York: Wiley. Bryant, R. A., & McConkey, K. M. (1989a). Crawford, H. J., & Gruzelier, J. H. (1992). A Hypnotic blindness, awareness, and attribu- midstream view of the neuropsychophysiol- tion. Journal of Abnormal Psychology, 98, 443– ogy of hypnosis: Recent research and future 447. directions. In E. Fromm & M. R. Nash (Eds.), Bryant, R. A., & McConkey, K. M. (1989b). Hyp- Contemporary hypnosis research (pp. 227–266). notic blindness: A behavioral and experiential New York: Guilford. analysis. Journal of Abnormal Psychology, 98, Crawford, H. J., Gur, R. C., Skolnick, B., Gur, R. 71–77. E., & Benson, D. M. (1993). Effects of hypnosis Bryant, R. A., & McConkey, K. M. (1989c). on regional cerebral blood flow during ischemic Hypnotic emotions and physical sensations: A pain with and without suggested hypnotic anal- real-simulating analysis. International Journal of gesia. International Journal of Psychophysiology, 181 195 Clinical and Experimental Hypnosis, 37, 305– 15, – . 319. Crawford, J. H., Macdonald, H., & Hilgard, E. Bryant, R. A., & McConkey, K. M. (1989d). R. (1979). Hypnotic deafness – psychophysical Visual conversion disorder: A case analysis of study of responses to tone intensity as modi- the influence of visual information. Journal of fied by hypnosis. American Journal of Psychol- 193 214 Abnormal Psychology, 98, 326–329. ogy, 92, – . Bryant, R. A., & McConkey, K. M. (1990a). Hyp- Csikszentmihalyi, M. (1990). Flow: The psychol- notic blindness and the relevance of attention. ogy of optimal experience. New York: Harper & Australian Journal of Psychology, 42, 287–296. Row. Bryant, R. A., & McConkey, K. M. (1990b). Hyp- Csikszentmihalyi, M., & Csikszentmihalyi, E. S. notic blindness and the relevance of cognitive (Eds.). (1988). Optimal experience: Psychologi- style. Journal of Personality & Social Psychology, cal studies of flow in consciousness. New York: 59, 756–761. Cambridge University Press. Bryant, R. A., & McConkey, K. M. (1990c). Hyp- Cunningham, P. V., & Blum, G. S. (1982). Fur- notic blindness: Testing the influence of moti- ther evidence that hypnotically induced color vation instructions. Australian Journal of Clini- blindness does not mimic congenital defects. cal & Experimental Hypnosis, 18, 91–96. Journal of Abnormal Psychology, 91, 139–143. Bryant, R. A., & McConkey, K. M. (1994). Hyp- Damaser, E. (1964). An experimental study of long- notic blindness and the priming effect of visual term post-hypnotic suggestion. Unpublished doc- material. Contemporary Hypnosis, 12, 157–164. toral dissertation, Harvard University, Cam- Campbell, D. T., & Fiske, D. W. (1959). Con- bridge, MA. vergent and discriminant validation by the David, D., Brown, R., Pojoga, C., & David, A. multitrait-multimethod matrix. Psychological (2000). The impact of posthypnotic amnesia Bulletin, 56, 82–105. and directed forgetting on implicit and explicit Cleghorn, J. M., Franco, S., Szechtman, H., memory: New insights from a modified process Brown, G. M., Nahmias, C., & Garnett, E. S. dissociation procedure. International Journal 3 (1992). Toward a brain map of auditory hallu- of Clinical and Experimental Hypnosis, 48( ), 267 289 cinations. American Journal of Psychiatry, 149, – . 1062–1069. Dorfman, J., & Kihlstrom, J. F. (1994, Novem- Coe, W. C. (1978). Credibility of post-hypnotic ber). Semantic priming in posthypnotic amnesia. amnesia – a contextualist’s view. International Paper presented at the Psychonomic Society, Journal of Clinical and Experimental Hypnosis, St. Louis, MO. 26, 218–245. Dumas, R. A. (1977). EEG alpha-hypnotizability Cooper, L. M. (1966). Spontaneous and sug- correlations: A review. Psychophysiology, 14, gested posthypnotic source amnesia. Interna- 431–438. tional Journal of Clinical & Experimental Hyp- Dywan, J., & Bowers, K. (1983). The use of hypnosis, 2 , 180–193. nosis to enhance recall. Science, 222, 184–185. Crawford, H. J. (2001). Neuropsychophysiology Edmonston, W. E.,(1977). Neutral hypnosis as of hypnosis: Towards an understanding of how relaxation. American Journal of Clinical Hyp- hypnotic interventions work. In G. D. Burrows, nosis, 20, 69–75. P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

470 the cambridge handbook of consciousness

Edmonston, W. E. (1981). Hypnosis and relax- source amnesia. International Journal of Clinical ation: Modern verification of an old equation. and Experimental Hypnosis, 14(2), 162–179. New York: Wiley. Farvolden, P., & Woody, E. Z. (2004). Hypno- Edwards, G. (1956). Post-hypnotic amnesia and sis, memory, and frontal executive functioning. post-hypnotic effect. British Journal of Psychia- International Journal of Clinical & Experimental try, 11, 316–325. Hypnosis, 52, 3–26. Edwards, G. (1963). Duration of post-hypnotic Faymonville, M. E., Laureys, S., Degueldre, C., effect. British Journal of Psychiatry, 109, 259– Del Fiore, G., Luxen, A., Franck, G., Lamy, 266. M., & Maquet, P. (2000). Neural mechanisms Eich, E. (1988). Theoretical issues in state depen- of antinociceptive effects of hypnosis. Anesthe- dent memory. In H. L. Roediger & F. I. M. Craik siology, 92(5), 1257–1267. (Eds.), Varieties of memory and consciousness: Fisher, S. (1954). The role of expectancy in the Essays in honor of Endel Tulving (pp. 331-354). performance of posthypnotic behavior. Journal Hillsdale, NJ: Erlbaum. of Abnormal & Social Psychology, 49, 503–507. Einstein, G. O., & McDaniel, M. A. (1990). Nor- Flavell, J. H. (1979). Metacognition and cog- mal aging and prospective memory. Journal of nitive monitoring: A new area of cognitive- Experimental Psychology: Learning, Memory, & developmental inquiry. American Psychologist, Cognition, 16, 717–726. 34(10), 906–911. Ellenberger, H. F. (1970). The discovery of Fromm, E. (1979). The nature of hypnosis and the unconscious: The history and evolution of other altered states of consciousness: An ego dynamic psychiatry. New York: Basic Books. psychological theory. In E. Fromm & R. E. Epstein, W. (1982). Percept-percept couplings. Shor (Eds.), Hypnosis: Developments in research Perception, 11, 75–83. and new perspectives (pp. 81–103). New York: Erickson, M. H. (1938a). A study of clinical and Aldine. experimental findings on hypnotic deafness: I. Gandolfo, R. L. (1971). Role of expectancy, amne- Clinical experimentation and findings. Journal sia, and hypnotic induction in the performance of General Psychology, 19, 127–150. of posthypnotic behavior. Journal of Abnormal Erickson, M. H. (1938b). A study of clinical and Psychology, 77, 324–328. experimental findings on hypnotic deafness: Garner, W. R., Hake, H. W., & Eriksen, C. W. II. Experimental findings with a conditioned (1956). Operationism and the concept of per- response technique. Journal of General Psychol- ception. Psychological Review, 63, 149–159. ogy, 19, 151–167. Gauld, A. (1992). A history of hypnotism. New Erickson, M. H. (1939). The induction of color York : Cambridge University Press. blindness by a technique of hypnotic sugges- Geiselman, R. E., Bjork, R. A., & Fishman, D. L. tion. Journal of General Psychology, 20, 61–89. (1983). Disrupted retrieval in directed forget- Erickson, M. H., & Erickson, E. M. (1941). Con- ting: A link with posthypnotic amnesia. Journal cerning the nature and character of posthyp- of Experimental Psychology: General, 112, 58–72. notic suggestion. Journal of General Psychology, Gill, M. M., & Brenman, M. (1959). Hypno- 24, 95–133. sis and related states: Psychoanalytic studies Evans, F. J. (1979a). Contextual forgetting: (Vol. 2). New York: International Universities Posthypnotic source amnesia. Journal of Abnor- Press. mal Psychology, 88, 556–563. Glisky, M. L., & Kihlstrom, J. F. (1993). Hypno- Evans, F. J. (1979b). Hypnosis and sleep: Tech- tizability and facets of openness. International niques for exploring cognitive activity during Journal of Clinical & Experimental Hypnosis, sleep. In E. Fromm & R. E. Shor (Eds.), Hypno- 41(2), 112–123. sis: Developments in research and new perspec- Glisky, M. L., Tataryn, D. J., & Kihlstrom, J. F. tives (pp. 139–183). New York: Aldine. (1995). Hypnotizability and mental imagery. Evans, F. J. (1988). Posthypnotic amnesia: Disso- International Journal of Clinical & Experimen- ciation of context and context. In H. M. Petti- tal Hypnosis, 43(1), 34–54. nati (Ed.), Hypnosis and memory (pp. 157–192). Glisky, M. L., Tataryn, D. J., Tobias, B. A., & New York: Guilford. Kihlstrom, J. F. (1991). Absorption, openness to Evans, F. J., & Thorne, W. A. F. (1966). Two types experience, and hypnotizability. Journal of Per- of posthypnotic amnesia: Recall amnesia and sonality & Social Psychology, 60(2), 263–272. P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

consciousness in hypnosis 471

Goldstein, A. P., & Hilgard, E. R. (1975). Lack hypnotic analgesia? Journal of Abnormal Psy- of influence of the morphine antagonist nalox- chology, 104(3), 508–516. one on hypnotic analgesia. Proceedings of the Harriman, P. L. (1942a). Hypnotic induction of National Academy of Sciences USA, 72, 2041– color vision anomalies: I. The use of the Ishi- 2043. hara and the Jensen tests to verify the accep- Gorassini, D. R., & Spanos, N. P. (1987). A social tance of suggested color blindness. Journal of cognitive skills approach to the successful mod- General Psychology, 26, 289–298. ification of hypnotic susceptibility. Journal of Harriman, P. L. (1942b). Hypnotic induction of Personality and Social Psychology, 50, 1004–1012. color vision anomalies: II. Results on two other Graf, P., & Schacter, D. L. (1985). Implicit and tests of color blindness. Journal of General Psy- explicit memory for new associations in normal chology, 27, 81–92. and amnesic subjects. Journal of Experimental Harvey, M. A., & Sipprelle, C. N. (1978). Color Psychology: Learning, Memory, and Cognition, blindness, perceptual interference, and hypno- 11, 501–518. sis. American Journal of Clinical Hypnosis, 20, Graffin, N. F., Ray, W. J., & Lundy, R. (1995). 189–193. EEG concomitants of hypnosis and hypnotic Hilgard, E. R. (1965). Hypnotic susceptibility. New susceptibility. Journal of Abnormal Psychology, York: Harcourt, Brace, & World. 123 131 104, – . Hilgard, E. R. (1969). Pain as a puzzle for psy- Graham, K. R., & Schwarz, L. M. (1973, August). chology and physiology. American Psychologist, Suggested deafness and auditory signal detectabil- 24, 103–113. ity. Paper presented at the annual meeting of Hilgard, E. R. (1971, September). Is hypnosis the American Psychological Association, Mon- a state, trait, neither? Paper presented at the treal. American Psychological Association, Washing- Gravitz, M. A., & Gerton, M. I. (1984). Origins ton, DC. of the term hypnotism prior to Braid. American Hilgard, E. R. (1972). A critique of Johnson, Journal of Clinical Hypnosis, 27, 107–110. Maher, and Barber’s “Artifact in the ‘essence of Gregg, V. H. (1979). Posthypnotic amnesia and hypnosis: An Evaluation of trance logic”, with general memory theory. Bulletin of the British a recomputation of their findings. Journal of Society for Experimental and Clinical Hypnosis, Abnormal Psychology, 79, 221–233. 1979(2), 11–14. Hilgard, E. R. (1973a). The domain of hyp- Gregg, V. H. (1982). Posthypnotic amnesia for nosis, with some comments on alternative recently learned material: A comment on the paradigms. American Psychologist, 28, 972– paper by J. F. Kihlstrom (1980). Bulletin of 982. the British Society of Experimental & Clinical Hilgard, E. R. (1973b). A neodissociation inter- Hypnosis, 5, 27–30. pretation of pain reduction in hypnosis. Psy- Grether, W.F. (1940). A comment on “The induc- chological Review, 80, 396–411. tion of color blindness by a technique of hyp- Hilgard, E. R. (1977). Divided consciousness: Mul- notic suggestion”. Journal of General Psychol- tiple controls in human thought and action. New ogy, 23, 207–210. York: Wiley-Interscience. Grosz, H. J., & Zimmerman, J. (1965). Exper- Hilgard, E. R., & Cooper, L. M. (1965). Spon- imental analysis of hysterical blindness: A taneous and suggested posthypnotic amnesia. follow-up report and new experimental data. International Journal of Clinical & Experimental Archives of General Psychiatry, 13, 255– Hypnosis, 13, 261–273. 260. Hilgard, E. R., & Hilgard, J. R. (1975). Hypnosis Halligan, P. W., Athwal, B. S., Oakley, D. A., & in the relief of pain. Los Altos, CA: Kaufman. Frackowiak, R. S. J. (2000, March 18). Imaging Hilgard, E. R., Hilgard, J. R., Macdonald, H., hypnotic paralysis: Implications for conversion Morgan, A. H., & Johnson, L. S. (1978). hysteria. Lancet, 355, 986–987. Covert pain in hypnotic analgesia: Its reality as Halligan, P. W., Oakley, D. A., Athwal, B. S., & tested by the real-simulator paradigm. Journal Frackowiak, R. S. J. (2000). Imaging hypnotic of Abnormal Psychology, 87, 655–663. paralysis – Reply. Lancet, 356(9224), 163. Hilgard, E. R., Macdonald, H., Morgan, A. Hargadon, R., Bowers, K. S., & Woody, E. Z. H., & Johnson, L. S. (1978). The reality of (1995). Does counterpain imagery mediate hypnotic analgesia: A comparison of highly P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

472 the cambridge handbook of consciousness

hypnotizables with simulators. Journal of Jacoby, L. L. (1991). A process dissociation frame- Abnormal Psychology, 87, 239–246. work: Separating automatic from intentional Hilgard, E. R., & Morgan, A. H. (1975). Heart uses of memory. Journal of Memory & Lan- rate and blood pressure in the study of labora- guage, 30, 513–541. tory pain in man under normal conditions and James, W. (1890). Principles of psychology. New as influenced by hypnosis. Acta Neurobiologica York: Holt. Experimentalis, 35, 741–759. Janet, P. (1907). The major symptoms of hysteria. Hilgard, E. R., Morgan, A. H., Lange, A. F., New York: Macmillan. Lenox, J. R., Macdonald, H., Marshall, G. D., & Jansen, R. D., Blum, G. S., & Loomis, J. M. (1982). Sachs, L. B. (1974). Heart rate changes in pain Attentional alterations of slant specific interfer- and hypnosis. Psychophysiology, 11, 692–702. ence between line segments in eccentric vision. Hilgard, E. R., Morgan, A. H., & Macdonald, H. Perception, 11, 535–540. (1975). Pain and dissociation in the cold pressor Jasiukaitis, P., Nouriani, B., Hugdahl, K., & test: A study of hypnotic analgesia with “hid- Spiegel, D. (1995). Relateralizing hypnosis; or den reports” through automatic key pressing have we been barking up the wrong hemi- and automatic talking. Journal of Abnormal Psy- sphere? International Journal of Clinical & chology, 84, 280–289. Experiemental Hypnosis, 45, 158–177. Hilgard, J. R. (1970). Personality and hypnosis: A John, O. P. (1990). The “big five” factor taxon- study in imaginative involvement. Chicago: Uni- omy: Dimensions of personality in the natural versity of Chicago Press. language and in questionnaires. In L. A. Per- Hobson, J. A., Pace-Schott, E., & Stickgold, vin (Ed.), Handbook of personality: Theory and R. (2000). Dreaming and the brain: Towards research (pp. 66–100). New York: Guilford. a cognitive neuroscience of conscious states. Johnson, M. K., Hashtroudi, S., & Lindsay, D. S. Behavioral & Brain Sciences, 23(6). (1993). Source monitoring. Psychological Bul- Hochberg, J. (1974). Higher-order stimuli and letin, 114(1), 3–28. interresponse coupling in the perception of Johnson, M. K., & Raye, C. L. (1981). Reality the visual world. In R. B. MacLeod & H. L. monitoring. Psychological Review, 88, 67–85. Pick (Eds.), Perception: Essays in honor of James Johnson, R. F. Q. (1972). Trance logic revisited: A J. Gibson (pp. 17–39). Ithaca, NY: Cornell reply to Hilgard’s critique. Journal of Abnormal University Press. Psychology, 79, 234–238. Hochberg, J., & Peterson, M. A. (1987). Johnson, R. F. Q., Maher, B. A., & Barber, T. X. Piecemeal organization and cognitive compo- (1972). Artifact in the “essence of hypnosis”: An nents in object perception: Perceptually cou- evaluation of trance logic. Journal of Abnormal pled responses to moving objects. Journal of Psychology, 79, 212–220. Experimental Psychology: General, 116, 370– Kallio, S., & Revensuo, A. (2003). Hypnotic phe- 380. nomena and altered states of consciousness: A Hofling, C. K., Heyl, B., & Wright, D. (1971). multilevel framework of description and expla- The ratio of total recoverable memories to con- nation. Contemporary Hypnosis, 20, 111–164. scious memories in normal subjects. Compre- Kihlstrom, J. F. (1979). Hypnosis and psy- hensive Psychiatry, 12, 371–379. chopathology: Retrospect and prospect. Jour- Howard, R. J., Ffytche, D. H., Barnes, J., nal of Abnormal Psychology, 88(5), 459–473. McKeefry, D., Ha, Y., Woodruff, P. W., Kihlstrom, J. F. (1980). Posthypnotic amnesia Bullmore, E. T., Simmons, A., Williams, S. C. for recently learned material: Interactions with R., David, A. S., & Brammer, M. (1998). The “episodic” and “semantic” memory. Cognitive functional anatomy of imagining and perceiv- Psychology, 12, 227–251. 1019 1023 ing colour. Neuroreport, 9, – . Kihlstrom, J. F. (1983). Instructed forgetting: Hoyt, I. P., & Kihlstrom, J. F. (1986, August). Hypnotic and nonhypnotic. Journal of Experi- Posthypnotic suggestion and waking instruction. mental Psychology: General, 112(1), 73–79. 94 Paper presented at the th annual meeting Kihlstrom, J. F. (1984). Conscious, subconscious, of the American Psychological Association, unconscious: A cognitive perspective. In K. S. Washington, DC. Bowers & D. Meichenbaum (Eds.), The uncon- Hull, C. L. (1933). Hypnosis and suggestibility: An scious reconsidered (pp. 149–211). New York: experimental approach. New York: Appleton. Wiley. P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

consciousness in hypnosis 473

Kihlstrom, J. F. (1985). Posthypnotic amnesia Kihlstrom, J. F. (2002b). Mesmer, the Franklin and the dissociation of memory. Psychology of Commission, and hypnosis: A counterfactual Learning and Motivation, 19, 131–178. essay. International Journal of Clinical & Exper- Kihlstrom, J. F. (1987). The cognitive uncon- imental Hypnosis, 50, 408–419. scious. Science, 237(4821), 1445–1452. Kihlstrom, J. F. (2004a). Clark L. Hull, hypno- Kihlstrom, J. F. (1992a). Dissociation and dissoci- tist [Review of Hypnosis and Suggestibility: An ations: A comment on consciousness and cog- Experimental Approach by C. L. Hull]. Con- nition. Consciousness & Cognition: An Interna- temporary Psychology, 49, 141–144. tional Journal, 1(1), 47–53. Kihlstrom, J. F. (2004b). “An unwarrantable Kihlstrom, J. F. (1992b). Dissociative and conver- impertinence” [Commentary on The Illusion of sion disorders. In D. J. Stein & J. Young (Eds.), Conscious Will by D.M.Wegner]. Behavioral & Cognitive science and clinical disorders (pp. 247– Brain Sciences, 27, 666–667, 270). San Diego: Academic Press. Kihlstrom, J. F., Barnhardt, T. M., & Tataryn, D. J. Kihlstrom, J. F. (1992c). Hypnosis: A sesquicen- (1992). Implicit perception. In R. F. Bornstein & tennial essay. International Journal of Clinical & T. S. Pittman (Eds.), Perception without aware- Experimental Hypnosis, 40(4), 301–314. ness: Cognitive, clinical, and social perspectives 17 54 Kihlstrom, J. F. (1994a). One hundred years of (pp. – ). New York: Guilford Press. hysteria. In S. J. Lynn & J. W. Rhue (Eds.), Kihlstrom, J. F., Brenneman, H. A., Pistole, D. D., Dissociation: Clinical and theoretical perspectives & Shor, R. E. (1985). Hypnosis as a retrieval cue (pp. 365–394). New York: Guilford Press. in posthypnotic amnesia. Journal of Abnormal 3 264 271 Kihlstrom, J. F. (1994b). The rediscovery of Psychology, 94( ), – . the unconscious. In H. Morowitz & J. L. Kihlstrom, J. F., & Edmonston, W. E. (1971). Singer (Eds.), The mind, the brain, and complex Alterations in consciousness in neutral hypno- adaptive systems (pp. 123–143). Reading, MA: sis: Distortions in semantic space. American Addison-Wesley. Journal of Clinical Hypnosis, 13(4), 243–248. Kihlstrom, J. F. (1996). Perception without aware- Kihlstrom, J. F., & Eich, E. (1994). Altering states ness of what is perceived, learning without of consciousness. In D. Druckman & R. A. Bjork awareness of what is learned. In M. Vel- (Eds.), Learning, remembering, and believing: mans (Ed.), The science of consciousness: Psycho- Enhancing performance (pp. 207–248). Wash- logical, neuropsychological and clinical reviews ington, DC: National Academy Press. (pp. 23–46). London: Routledge. Kihlstrom, J. F., & Evans, F. J. (1976). Recovery of Kihlstrom, J. F. (1998). Dissociations and dissoci- memory after posthypnotic amnesia. Journal of ation theory in hypnosis: Comment on Kirsch Abnormal Psychology, 85(6), 564–569. and Lynn (1998). Psychological Bulletin, 123(2), Kihlstrom, J. F., & Evans, F. J. (1977). Residual 186–191. effect of suggestions for posthypnotic amnesia: Kihlstrom, J. F. (2000, November 2). Hypnosis A reexamination. Journal of Abnormal Psychol- and pain: Time for a new look. Paper presented ogy, 86(4), 327–333. at the Annual meeting of the American Pain Kihlstrom, J. F., & Evans, F. J. (1979). Memory Society, Atlanta, GA. retrieval processes in posthypnotic amnesia. In Kihlstrom, J. F. (2001a). Dissociative disorders. In J. F. Kihlstrom & F. J. Evans (Eds.), Functional P. B. Sutker & H. E. Adams (Eds.), Compre- disorders of memory (pp. 179–218). Hillsdale, hensive handbook of psychopathology (3rd ed., NJ: Erlbaum. pp. 259–276). New York: Plenum. Kihlstrom, J. F., & McGlynn, S. M. (1991). Exper- Kihlstrom, J. F. (2001b, August). Hypnosis in imental research in clinical psychology. In M. surgery: Efﬁcacy, speciﬁcity, and utility. Paper Hersen, A. E. Kazdin, & A. S. Bellack (Eds.), presented at the the annual meeting of The clinical psychology handbook (2nd ed., the American Psychological Association, San pp. 239–257). New York: Pergamon Press. Francisco. Kihlstrom, J. F., Mulvaney, S., Tobias, B. A., & Kihlstrom, J. F. (2002a). Demand characteristics in Tobis, I. P. (2000). The emotional unconscious. the laboratory and the clinic: Conversations and In E. Eich, J. F. Kihlstrom, G. H. Bower, J. P. collaborations with subjects and patients. Retri- Forgas, & P. M. Niedenthal (Eds.), Cognition eved from http://journals.apa.org/prevention/ and emotion (pp. 30–86). New York: Oxford volume5/pre0050036c.html. University Press. P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

474 the cambridge handbook of consciousness

Kihlstrom, J. F., & Schacter, D. L. (2000). Func- Lang, E. V., Benotsch, E. G., Fick, L. J., Lut- tional amnesia. In F. Boller & J. Grafman (Eds.), gendorf, S., Berbaum, M. L., Berbaum, K. S., Handbook of neuropsychology (2nd ed., Vol. 2, Logan, H., & Spiegel, D. (2000, April 29). pp. 409–427). Amsterdam: Elsevier. Adjunctive non-pharmacological analgesia for Kihlstrom, J. F., Tataryn, D. J., & Hoyt, I. P. invasive medical procedures: A randomised (1993). Dissociative disorders. In P. J. Sutker & trial. Lancet, 355, 1486–1500. H. E. Adams (Eds.), Comprehensive handbook Lang, E. V., Joyce, J. S., Spiegel, D., Hamilton, D., of psychopathology (2nd ed., pp. 203–234). New & Lee, K. K. (1996). Self-hypnotic relaxation York: Plenum Press. during interventional radiological procedures: Killeen, P. R., & Nash, M. R. (2003). The four Effects on pain perception and intravenous causes of hypnosis. International Journal of Clin- drug use. International Journal of Clinical & ical and Experimental Hypnosis, 51(3), 195–231. Experimental Hypnosis, 44, 106–119. Kirsch, I. (2000). The response set theory of hyp- Laurence, J. R., Perry, C., & Kihlstrom, J. F. nosis. American Journal of Clinical Hypnosis, (1983). Hidden observer phenomena in hyp- 42(3–4), 274–292. nosis: An experimental creation?Journal of Per- 163 169 Kirsch, I. (2001a). The altered states of hypnosis. sonality and Social Psychology, 44, – . Social Research, 68(3), 795–807. Leibowitz, H. W., Lundy, R. M., & Guez, J. R. Kirsch, I. (2001b). The response set theory of (1980). The effect of testing distance on sug- hypnosis: Expectancy and physiology. Ameri- gestion induced visual field narrowing. Inter- can Journal of Clinical Hypnosis, 44(1), 69–73. national Journal of Clinical and Experimental Hypnosis, , 409–420. Kirsch, I., & Lynn, S. J. (1995). Altered state of 28 hypnosis: Changes in the theoretical landscape. Leibowitz, H. W., Post, R. B., Rodemer, C. S., American Psychologist, 50(10), 846–858. Wadlington, W.L., & Lundy, R. M. (1980). Roll Kirsch, I., & Lynn, S. J. (1997). Hypnotic involun- vection analysis of suggestion induced visual tariness and the automaticity of everyday life. field narrowing. Perception and Psychophysics, 173 176 American Journal of Clinical Hypnosis, 40(1), 28, – . 329–348. Levy, B. L., & Anderson, M. C. (2002). Inhibitory Kirsch, I., & Lynn, S. J. (1998a). Dissociation the- processes and the control of memory retrieval. 299 305 ories of hypnosis. Psychological Bulletin, 123(1), Trends in Cognitive Sciences, 6, – . 100–115. Ludwig, A. M. (1966). Altered states of con- Kirsch, I., & Lynn, S. J. (1998b). Social-cognitive sciousness. Archives of General Psychiatry, 15, 225 234 alternatives to dissociation theories of hypnotic – . involuntariness. Review of General Psychology, MacLeod, C. M. (1991). Half a century of research 2 (1), 66–80. on the Stroop effect: An integrative review. 2 163 203 Knox, V. J., Morgan, A. H., & Hilgard, E. R. Psychological Bulletin, 109( ), – . (1974). Pain and suffering in ischemia: The MacLeod, C. M. (1992). The Stroop task: The paradox of hypnotically suggested anesthesia “gold standard” of attentional measures. Journal as contradicted by reports from the “hidden of Experimental Psychology: General, 121(1), 12– observer.”Archives of General Psychiatry, 30, 14. 840–847. MacLeod-Morgan, C., & Lack, L. (1982). Hemi- Kosslyn, S. M., Thompson, W. L., Costantini- spheric specificity: A physiological concomi- Ferrando, M. F., Alpert, N. M., & Spiegel, D. tant of hypnotizability. Psychophysiology, 19, (2000). Hypnotic visual hallucination alters 687–690. brain color processing. American Journal of Mallard, D., & Bryant, R. A. (2001). Hypnotic Psychiatry, 157(8), 1279–1284. color blindness and performance on the Stroop Kuhn, T. (1962). The structure of scientific revolu- test. International Journal of Clinical and Exper- tions. Chicago: University of Chicago Press. imental Hypnosis, 49, 330–338. Kunzendorf, R., Spanos, N., & Wallace, B. (Eds.). Maquet, P., Faymonvi., M. E., DeGuelder, C., (1996). Hypnosis and imagination. New York: DelFiore, G., Franck, G., Luxen, A., & Lamy, Baywood. M. (1999). Functional neuroanatomy of hyp- LaBerge, D., & Samuels, S. J. (1974). Toward a notic state. Biological Psychiatry, 45, 327–333. theory of automatic information processing in Marcel, A. J. (1983). Conscious and unconscious reading. Cognitive Psychology, 6, 293–323. perception: Experiments on visual masking P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

consciousness in hypnosis 475

and word recognition. Cognitive Psychology, 15, perle, M., & Buchser, E. (1991). Mechanism 197–237. of analgesia induced by hypnosis and acupunc- McClelland, D. C., Koestner, R., & Weinberger, ture: Is there a difference? Pain, 45, 135–140. J. (1989). How do self-attributed and implicit Nace, E. P., & Orne, M. T. (1970). Fate of an motives differ?Psychological Review, 96, 690– uncompleted posthypnotic suggestion. Journal 702. of Abnormal Psychology, 75, 278–285. McConkey, K. M., Bryant, R. A., Bibb, B. C., & Nace, E. P., Orne, M. T., & Hammer, A. G. (1974). Kihlstrom, J. F. (1991). Trance logic in hypnosis Posthypnotic amnesia as an active psychic pro- and imagination. Journal of Abnormal Psychol- cess. Archives of General Psychiatry, 31, 257– ogy, 100(4), 464–472. 260. McConkey, K. M., Bryant, R. A., Bibb, B. C., Nash, M. (1987). What, if anything, is regressed Kihlstrom, J. F., & Tataryn, D. J. (1990). Hyp- about hypnotic age regression: A review of the notically suggested anaesthesia and the circle- empirical literature. Psychological Bulletin, 102, touch test: A real-simulating comparison. 42–52. British Journal of Experimental & Clinical Hyp- Nash, M. R., Lynn, S. J., Stanley, S., & Carlson, V. nosis, 7, 153–157. (1987). Subjectively complete hypnotic deaf- McConkey, K. M., & Sheehan, P. W. (1980). ness and auditory priming. International Jour- Inconsistency in hypnotic age regression and nal of Clinical and Experimental Hypnosis, 35, cue structure as supplied by the hypnotist. 32–40. International Journal of Clinical and Experimen- Nelson, T. O. (1992). Metacognition: Core read- tal Hypnosis, 38, 394–408. ings. Boston: Allyn and Bacon. McGlashan, T. H., Evans, F. J., & Orne, M. T. Nelson, T. O. (1996). Consciousness and meta- (1969). The nature of hypnotic analgesia and cognition. American Psychologist, 51(2), 102– placebo response to experimental pain. Psycho- 116. 227 246 somatic Medicine, 31, – . Nelson, T. O., & Narens, L. (1990). Metamem- Melzack, R. (1975). The McGill Pain Question- ory: A theoretical framework and some new naire: Major properties and scoring methods. ﬁndings. In G. H. Bower (Ed.), The psychology Pain, 1, 277–299. of learning and motivation Vol. 26 (pp. 125–173). Metcalfe, J., & Shimamura, A. P. (1994). Metacog- New York: Academic Press. nition: Knowing about knowing. Cambridge, Nogrady, H., McConkey, K. M., & Perry, C. MA: MIT Press. (1985). Enhancing visual memory: Trying hyp- Metzinger, T. (Ed.). (2000). Neural correlates of nosis, trying imagination, and trying again. consciousness. Cambridge, MA: MIT Press. Journal of Abnormal Psychology, 94, 195– Miller, M. E., & Bowers, K. S. (1986). Hypnotic 204. analgesia and stress inoculation in the reduc- Nolan, R. P., & Spanos, N. P. (1987). Hypnotic tion of pain. Journal of Abnormal Psychology, analgesia and stress inoculation: A critical reex- 95, 6–14. amination of Miller and Bowers. Psychological Miller, M. E., & Bowers, K. S. (1993). Hypnotic Reports, 61, 95–102. analgesia: Dissociated experience or dissoci- Oakley, D. A. (1999). Hypnosis and conversion ated control?Journal of Abnormal Psychology, hysteria: A unifying model. Cognitive Neu- 102, 29–38. ropsychiatry, 4, 243–265. Miller, R. J., Hennessy, R. T., & Leibowitz, H. W. Obstoj, I., & Sheehan, P. W. (1977). Aptitude (1973). The effect of hypnotic ablation of the for trance, task generalizability, and incon- background on the magnitude of the Ponzo per- gruity response in hypnosis. Journal of Abnor- spective illusion. International Journal of Clini- mal Psychology, 86, 543–552. cal and Experimental Hypnosis, 21, 180–191. Orne, M. T. (1959). The nature of hypnosis: Arti- Montgomery, G. H., DuHamel, K. N., & Redd, fact and essence. Journal of Abnormal and Social W. H. (2000). A meta-analysis of hypnotically Psychology, 58, 277–299. induced analgesia: How effective is hypno- Orne, M. T. (1962). On the social psychology of sis?International Journal of Clinical and Experi- the psychological experiment: With particular mental Hypnosis, 48(2), 138–153. reference to demand characteristics and their Moret, V., Forster, A., Laverriere, M.-C. , Gail- implications. American Psychologist, 17, 776– lard, R. C., Bourgeois, P., Haynal, A., Gem- 783. P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

476 the cambridge handbook of consciousness

Orne, M. T. (1969). On the nature of the Rainville, P., Hofbauer, R. K., Paus, T., Bushnell, posthypnotic suggestion. In L. Chertok (Ed.), M. C., & Price, D. D. (1999). Cerebral mech- Psychophysiological mechanisms of hypnosis anisms of hypnotic induction and suggestion. (pp. 173–192). Berlin: Springer-Verlag. Journal of Cognitive Neuroscience, 11, 110–125. Orne, M. T. (1971). The simulation of hypno- Ray, W. J., & Tucker, D. M. (2003). Evolution- sis: Why, how, and what it means. International ary approaches to understanding the hypnotic Journal of Clinical and Experimental Hypnosis, experience. International Journal of Clinical and 19, 183–210. Experimental Hypnosis, 51(3), 256–281. Orne, M. T. (1973). Communication by the total Raz, A., Fan, J., Shapiro, T., & Posner, M. I. experimental situation: Why it is important, (2002, November). fMRI of posthypnotic sugges- how it is evaluated, and its significance for tion to modulate reading of Stroop words. Paper the ecological validity of findings. In P. Pliner, presented at the Society for Neuroscience, L. Krames, & T. Alloway (Eds.), Communi- Washington, DC. cation and affect (pp. 157–191). New York: Raz, A., Landzberg, K. S., Schweizer, H. R., Academic. Zephrani, Z., Shapiro, T., Fan, J., & Posner, M. I. Orne, M. T., Sheehan, P. W., & Evans, F. J. (1968). (2003). Posthypnotic suggestion and the mod- Occurrence of posthypnotic behavior outside ulation of Stroop interference under cyclople- the experimental setting. Journal of Personality gia. Consciousness & Cognition, 12, 332–346. & Social Psychology, 9, 189–196. Raz, A., Shapiro, T., Fan, J., & Posner, M. I. Otto-Salaj, L. L., Nadon, R., Hoyt, I. P., Regis- (2002). Hypnotic suggestion and the modula- ter, P. A., & Kihlstrom, J. F. (1992). Lateral- tion of Stroop interference. Archives of General ity of hypnotic response. International Journal Psychiatry, 59, 1155–1161. of Clinical & Experimental Hypnosis, 40, 12– Reber, A. S. (1967). Implicit learning of artificial 20. grammars. Journal of Verbal Learning & Verbal Pattie, F. A. (1937). The genuineness of hypnoti- Behavior, 6, 855–863. cally produced anesthesia of the skin. American Reber, A. S. (1993). Implicit learning and tacit Journal of Psychology, 49, 435–443. knowledge: An essay on the cognitive unconscious. Peterson, M. A., & Hochberg, J. (1983). Opposed Oxford: Oxford University Press. set measurement procedure: A quantitative Reder, L. M. (1996). Implicit memory and metacog- analysis of the role of local cues and intention nition. Mahwah, NJ: Erlbaum. in form perception. Journal of Experimental Psy- Register, P.A., & Kihlstrom, J. F. (1987). Hypnotic chology: Human Perception and Performance, 9, effects on hypermnesia. International Journal of 183–193. Clinical & Experimental Hypnosis, 35(3), 155– Posner, M. I., & Snyder, C. R. R. (1975). Atten- 170. tion and cognitive control. In R. L. Solso (Ed.), Reyher, J. (1967). Hypnosis in research on psy- Information processing and cognition: The Loyola chopathology. In J. E. Gordon (Ed.), Handbook Symposium (pp. 55–85). New York: Wiley. of clinical and experimental hypnosis (pp. 110– Rainville, P., Carrier, B., Hofbauer, R. K., Bush- 147). New York: Macmillan. nell, M. C., & Duncan, G. H. (1999). Disso- Reyher, J., & Smyth, L. (1971). Suggestibility dur- ciation of sensory and affective dimensions of ing the execution of a posthypnotic suggestion. pain using hypnotic modulation. Pain, 82(2), Journal of Abnormal Psychology, 78, 258–265. 159 171 – . Roche, S. M., & McConkey, K. M. (1990). Rainville, P., Duncan, G. H., Price, D. D., Carrier, Absorption: Nature, assessment, and corre- B., & Bushnell, M. C. (1997, August 15). Pain lates. Journal of Personality and Social Psychol- affect encoded in human anterior cingulate but ogy, 59, 91–101. 968 not somatosensory cortex. Science, 277, – Rosenthal, R., & Rubin, D. B. (1978). Interper- 971 . sonal expectancy effects: The first 345 studies. Rainville, P., Hofbauer, R. K., Bushnell, M. C., Behavioral & Brain Sciences, 3, 377–415. 2002 Duncan, G. H., & Price, D. D. ( ). Hypno- Sabourin, M., Brisson, M. A., & Deschamb, A. sis modulates the activity in cerebral structures (1980). Evaluation of hypnotically suggested involved in the regulation of consciousness. selective deafness by heart-rate conditioning Journal of Cognitive Neuroscience, 14(Suppl), and reaction time. Psychological Reports, 47, 887 901 – . 995–1002. P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

consciousness in hypnosis 477

Sackeim, H. A. (1982). Lateral asymmetry in bod- nications in Psychology, Psychiatry & Behavior, ily response to hypnotic suggestions. Biological 7, 357–272. Psychiatry, 17, 437–447. Sheehan, P. W. (1988). Memory distortion in Sackeim, H. A., Nordlie, J. W., & Gur, R. C. hypnosis. International Journal of Clinical and (1979). A model of hysterical and hypnotic Experimental Hypnosis, 36, 296–311. blindness: Cognition, motivation, and aware- Sheehan, P. W., & McConkey, K. M. (1982). Hyp- ness. Journal of Abnormal Psychology, 88, 474– nosis and experience: The exploration of phenom- 489. ena and process. Hillsdale, NJ: Erlbaum. Sarbin, T. R. (1973). On the recently reported Sheehan, P. W., Obstoj, I., & McConkey, K. M. physiological and pharmacological reality of (1976). Trance logic and cue structure as sup- the hypnotic state. Psychological Record, 23, plied by the hypnotist. Journal of Abnormal 501–511. Psychology, 85, 459–472. Sarbin, T. R., & Coe, W. C. (1972). Hypnosis: A Sheehan, P. W., & Orne, M. T. (1968). Some com- social psychological analysis of inﬂuence commu- ments on the nature of posthypnotic behavior. nication. New York: Holt, Rinehart, & Winston. Journal of Nervous & Mental Disease, 146, 209– Sarbin, T. R., & Slagle, R. W. (1979). Hyp- 220. nosis and psychophysiological outcomes. In. Shiffrin, R. M., & Schneider, W. (1977). Con- In E. Fromm & R. E. Shor (Eds.), Hypnosis: trolled and automatic human information Developments in research and new perspectives processing: II. Perceptual learning, automatic (pp. 273–303). New York: Aldine. attending and a general theory. Psychological Schacter, D. L. (1987). Implicit memory: History Review, 84(2), 127–190. and current status. Journal of Experimental Psy- Shiffrin, R. M., & Schneider, W. (1984). Auto- chology: Learning, Memory, and Cognition, 13, matic and controlled processing revisited. Psy- 501–518. chological Review, 91(2), 269–276. Schacter, D. L., & Graf, P. (1986). Effects of Shimamura, A. P., & Squire, L. R. (1987). A elaborative processing on implicit and explicit neuropsychological study of fact memory and memory for new associations. Journal of Exper- source amnesia. Journal of Experimental Psy- imental Psychology: Learning, Memory, and chology: Learning, Memory, and Cognition, 13, Cognition, 12. 464–473. Schacter, D. L., Harbluk, J. L., & McClachlan, D. Shor, R. E. (1979a). The fundamental problem R. (1984). Retrieval without recollection: An in hypnosis research as viewed from historic experimental analysis of source amnesia. Jour- perspectives. In E. Fromm & R. E. Shor (Eds.), nal of Verbal Learning and Verbal Behavior, 23, Hypnosis: Developments in research and new per- 593–611. spectives. New York: Aldine. Schacter, D. L., & Kihlstrom, J. F. (1989). Func- Shor, R. E. (1979b). A phenomenological method tional amnesia. In F. Boller & J. Graffman for the measurement of variables important to (Eds.), Handbook of neuropsychology (Vol. 3, an understanding of the nature of hypnosis. pp. 209–231). Amsterdam: Elsevier. In E. Fromm & R. E. Shor (Eds.), Hypnosis: Scheibe, K. E., Gray, A. L., & Keim, C. S. (1968). Developments in research and new perspectives Hypnotically induced deafness and delayed (pp. 105–135). New York: Aldine. auditory feedback: A comparison of real and Shor, R. E., Orne, M. T., & O’Connell, D. simulating subjects. International Journal of N. (1962). Validation and cross-validation Clinical & Experimental Hypnosis, 16, 158–164. of a scale of self-reported personal experi- Schneider, W., & Shiffrin, R. M. (1977). Con- ences which predicts hypnotizability. Journal trolled and automatic human information proof Psychology, 53, 55–75. cessing: I. Detection, search, and attention. Snyder, M., & Swann, W. B. (1978). Behav- Psychological Review, 84(1), 1–66. ioral conﬁrmation in social interaction: From Seger, C. A. (1994). Criteria for implicit learn- social perception to social reality. Journal ing: De-emphasize conscious access, empha- of Experimental Social Psychology, 14, 148– size amnesia. Behavioral & Brain Sciences, 17, 162. 421–422. Spanos, N. P. (1983). The hidden observer as Sheehan, P. W. (1982). Imagery and hypnosis: an experimental creation. Journal of Personal- Forging a link, at least in part. Research Commu- ity and Social Psychology, 44, 170–176. P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

478 the cambridge handbook of consciousness

Spanos, N. P. (1986a). Hypnosis, nonvolitional hidden pain during hypnotic analgesia. Journal responding, and multiple personality: A social of Abnormal Psychology, 92, 479–488. psychological perspective. In B. A. Maher & Spanos, N. P., & Hewitt, E. C. (1980). The hid- W. B. Maher (Eds.), Progress in experimental den observer in hypnotic analgesia: Discovery personality research (pp. 1–62), New York: Aca- or experimental creation. Journal of Personality demic Press. & Social Psychology, 39, 1201–1214. Spanos, N. P. (1986b). Hypnotic behavior: A Spanos, N. P., Jones, B., & Malfara, A. (1982). social psychological interpretation of amne- Hypnotic deafness: Now you hear it – Now you sia, analgesia, and trance logic. Behavioral and still hear it. Journal of Abnormal Psychology, 91, Brain Sciences, 9, 449–467. 75–77. Spanos, N. P. (1991). A sociocognitive approach Spanos, N. P., Menary, E., Brett, P. J., Cross, W., to hypnosis. In S. J. Lynn & J. W. Rhue (Eds.), & Ahmed, Q. (1987). Failure of posthypnotic Theories of hypnosis: Current models and per- responding to occur outside the experimental spectives (pp. 324–361). New York: Guilford setting. Journal of Abnormal Psychology, 96, 52– Press. 57. Spanos, N. P., & Barber, T. X. (1968). “Hyp- Spanos, N. P., Radtke, H. L., & Dubreuil, D. notic” experiences as inferred from auditory L. (1982). Episodic and semantic memory in and visual hallucinations. Journal of Experimen- posthypnotic amnesia: A reevaluation. Journal tal Research in Personality, 3, 136–150. of Personality and Social Psychology, 43, 565– Spanos, N. P., & Chaves, J. F. (1970). Hypnosis 573. research: A methodological critique of experi- Spanos, N. P., & Saad, C. L. (1984). Prism adap- ments generated by two alternative paradigms. tation in hypnotically limb-anesthetized sub- American Journal of Clinical Hypnosis, 13(2), jects: More disconﬁrming data. Perceptual and 108–127. Motor Skills, 59, 379–386. Spanos, N. P., & Chaves, J. F. (1991). History Spiegel, D., & Albert, L. H. (1983). Naloxone fails and historiography of hypnosis. In S.J. Lynn & to reverse hypnotic alleviation of chronic pain. J.W. Rhue (Eds.), Theories of hypnosis: Current Psychopharmacology, 81, 140–143. 43 78 models and perspectives (pp. – ). New York: Stern, J. A., Brown, M., Ulett, G. A., & Sletten, Guilford Press. I. (1977). A comparison of hypnosis, acupunc- Spanos, N. P., Cobb, P. C., & Gorassini, D. R. ture, morphine, valium, aspirin, and placebo (1985). Failing to resist hypnotic test sugges- in the management of experimentally induced tions: A strategy for self-presenting as deeply pain. Annals of the New York Academy of Sci- hypnotized. Psychiatry, 48, 282–292. ence, 296, 175–193. Spanos, N. P., DeGroot, H. P., & Gwynn, M. I. Stoyva, J., & Kamiya, J. (1968). Electrophysiolog- (1987). Trance logic as incomplete responding. ical studies of dreaming as the prototype of Journal of Personality and Social Psychology, 53, a new strategy in the study of consciousness. 911–921. Psychological Review, 75, 192–205. Spanos, N. P., Dubreuil, D. L., Saad, C. L., & Stroop, J. R. (1935). Studies of interference in Gorassini, D. (1983). Hypnotic elimination of serial verbal reactions. Journal of Experimental prism-induced aftereffects: Perceptual effect or Psychology, 18, 643–662. responses to experimental demands. Journal of Sutcliffe, J. P. (1960). “Credulous” and “skeptical” 216 222 Abnormal Psychology, 92, – . views of hypnotic phenomena: A review of cer- Spanos, N. P., Gorassini, D. R., & Petrusic, W. tain evidence and methodology. International (1981). Hypnotically induced limb anesthesia Journal of Clinical and Experimental Hypnosis, and adaptation to displacing prisms: A failure 8, 73–101. to conﬁrm. Journal of Abnormal Psychology, 90, Sutcliffe, J. P. (1961). “Credulous” and “skeptical” 329 333 – . views of hypnotic phenomena: Experiments Spanos, N. P., Gwynn, M. I., et al. (1988). in esthesia, hallucination, and delusion. Jour- Social psychological factors in the genesis of nal of Abnormal & Social Psychology, 62, 189– posthypnotic source amnesia. Journal of Abnor- 200. 322 329 mal Psychology 97, – . Szechtman, H., Woody, E., Bowers, K. S., & Nah- Spanos, N. P., Gwynn, M. I., & Stam, H. J. (1983). mias, C. (1998). Where the imaginal appears Instructional demands and ratings of overt and real: A positron emission tomography study P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

consciousness in hypnosis 479

of auditory hallucination. Proceedings of the induced affect. Motivation and Emotion, 11, National Academy of Sciences USA, 95, 1956– 67–81. 1960. Weitzenhoffer, A. M. (1974). When is an Tellegen, A., & Atkinson, G. (1974). Open- “instruction” an “instruction”?International ness to absorbing and self-altering experiences Journal of Clinical & Experimental Hypnosis, (“absorption”), a trait related to hypnotic sus- 22, 258–269. ceptibility. Journal of Abnormal Psychology, 83, White, R. W. (1941). A preface to the theory 268–277. of hypnotism. Journal of Abnormal & Social Terao, T., & Collinson, S. (2000). Imaging hyp- Psychology, 36, 477–505. notic paralysis. Lancet, 356(9224), 162–163. Wiggins, J. S., & Trapnell, P. D. (1997). Person- Timm, H. W.(1981). The effect of forensic hypno- ality structure: The return of the big ﬁve. In sis techniques on eyewitness recall and recog- R. Hogan, J. A. Johnson, & S .R.dr Briggs nition. Journal of Police Science, 9, 188–194. (Eds.), Handbook of personality psychology (pp. Tulving, E. (1983). Elements of episodic memory. 737–765). San Diego: Academic Press. Oxford: Oxford University Press. Williamsen, J. A., Johnson, H. J., & Eriksen, C. W. Veith, I. (1965). Hysteria: The history of a disease. (1965). Some characteristics of posthypnotic Chicago: University of Chicago Press. amnesia. Journal of Abnormal Psychology, 70, 123 131 Wallace, B. (1980). Factors affecting proprio- – . ceptive adaptation to prismatic displacement. Woody, E. Z., & Bowers, K. S. (1994). A frontal Perception & Psychophysics, 28, 550–554. assault on dissociated control. In S. J. Lynn & Wallace, B., & Fisher, L. E. (1982). Hypnoti- J. W. Rhue (Eds.), Dissociation: Clinical and cally induced limb anesthesia and adaptation to theoretical perspectives (pp. 52–79). New York: displacing prisms: Replication requires adher- Guilford Press. ence to critical procedures. Journal of Abnormal Woody, E. Z., & McConkey, K. M. (2003). Psychology, 91, 390–391. What we don’t know about the brain and Wallace, B., & Fisher, L. E. (1984a). Prism adap- hypnosis, but need to: A view from the tation with hypnotically induced limb anes- Buckhorn Inn. International Journal of Clinical thesia: The critical roles of head position and and Experimental Hypnosis, 51(3), 309–337. prism type. Perception and Psychophysics, 36, Woody, E. Z., & Sadler, P. (1998). On reinte- 303–306. grating dissociated theories: Commentary on Wallace, B., & Fisher, L. E. (1984b). The roles Kirsch and Lynn (1998). Psychological Bulletin, of target and eye motion in the production of 123, 192–197. the visual shift in prism adaptation. Journal of Woody, E. Z., & Szechtman, H. (2000a). General Psychology, 110, 251–262. Hypnotic hallucinations and yedasentience. Wallace, B., & Garrett, J. B. (1973). Reduced Contemporary Hypnosis, 17(1), 26–31. felt arm sensation effects on visual adaptation. Woody, E. Z., & Szechtman, H. (2000b). Hyp- Perception and Psychophysics, 14, 597–600. notic hallucinations: Towards a biology of epis- Wallace, B., & Garrett, J. B. (1975). Perceptual temology. Contemporary Hypnosis, 17(1), 4–14. adaptation with selective reductions of felt sen- Woody, E. Z., & Szechtman, H. (2003). How sation. Perception, 4, 437–445. can brain activity and hypnosis inform each Wegner, D. M. (2002). The illusion of conscious other? International Journal of Clinical and will. Cambridge, MA: MIT Press. Experimental Hypnosis, 51(3), 232–255. Weiss, F., Blum, G. S., & Gleberman, L. (1987). Yzerbyt, V., Lories, G., & Dardenne, B. (1998). Anatomically based measurement of facial Metacognition: Cognitive and social dimensions. expressions in simulated versus hypnotically Thousand Oaks, CA: Sage Publications. P1: KAE 0521857430c17 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:47

480 P1: KAE 0521857430c18 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:51

CHAPTER 18 Can We Study Subjective Experiences Objectively? First-Person Perspective Approaches and Impaired Subjective States of Awareness in Schizophrenia

Jean-Marie Danion and Caroline Huron

Abstract where I did a thing, and how I felt when I did it. There are all the things that I remem- One of the main challenges of scientific ber, either having experienced them myself research in psychiatry and clinical psychol- or been told about them by others. Out of ogy is to take account of subjectivity, as the same storehouse, with these past impres- defined by the experiential sense of existing sions, I can construct now this, now that, as a subject of experience, or the first-person image of things that I either have experi- perspective of the world (Sass & Parnas, enced or have believed on the basis of expe- 2003). Such clinical symptoms as halluci- rience – and from these I can further con- nations, delusions of alien control, feelings struct future actions, events, and hopes; and of guilt, thoughts of worthlessness, dereal- I can meditate on all these things as if they were present. [. . .]. I speak to myself in ization, and depersonalization are subjec- this way; and when I speak, the images tive experiences that have to be studied in of what I am speaking about are present themselves if research in clinical psychology out of the same store of memory; and if the and psychiatry is not to be excessively sim- images were absent I could say nothing at plistic. First-person approaches, such as the all about them. [. . .]. Here also is all, learnt remember/know procedure (Tulving, 1985), of the liberal sciences and as yet unfor- make it possible to study subjective expe- gotten; removed as it were to some inner riences objectively. We show how results place, which is yet no place: nor are they the from studies exploring conscious awareness images thereof, but the things themselves. in schizophrenia using first- and third-person St Augustine, Confessions, Book X perspective approaches provide new evidence for the validity of using first-person First-Person Perspective Approaches perspective approaches. to Conscious Awareness All this I do within myself, in that huge hall of my memory. [. . .]. There also I meet Since it became an object of scientific myself and recall myself – what, when, or investigation, conscious awareness has been 481 P1: KAE 0521857430c18 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:51

482 the cambridge handbook of consciousness

studied using the so-called third-person a film but fail to remember anything else perspective approaches. Typically, these about it. These examples from everyday approaches contrast performance in tasks life suggest that recognition memory may that rely heavily on conscious processes to be based either on feelings of familiarity that in tasks that do not rely, or rely less, accompanied by no recollection of contex- on conscious processes. Demonstration of an tual information or alternatively on the con- impaired performance only in tasks that rely scious recollection of details from a past heavily on conscious processes is taken as event. Tulving (1985) was the first to pro- evidence of a specific impairment of these pose a first-person perspective approach to conscious processes. These approaches to measure these two subjective experiences conscious awareness are described as third- (see Chapter 10). This approach hypothe- person perspective approaches because the sizes that consciousness is not a unitary phe- workings of consciousness are inferred by nomenon and it has to be fragmented to investigators from performance patterns in be accessible to experiments. Thus, Tulving selected tasks. However, as the investigator’s (1985) distinguishes two subjective experi- interpretation is based on indirect data, alter- ences, referred to as autonoetic and noetic native explanations sometimes have to be awareness, which are characterized by dis- considered. Indeed, the two selected tasks tinct phenomenological attributes. Auto- may differ in terms of parameters other than noetic awareness is the kind of conscious the involvement of conscious processes, in awareness that is experienced by normal which case these different parameters may subjects who consciously recollect personal account for the dissociation of performance. events by reliving them mentally. It makes First-person perspective approaches have it possible to be aware of one’s own expe- been developed recently by cognitive sci- riences across subjective time and to have entists as a means of studying conscious- a feeling of individuality, uniqueness, and ness directly as a subjective experience, self-direction. It is intimately associated with rather than indirectly as a function. These our awareness of ourselves as persons with approaches are not aimed at explaining the a past and a future. Noetic awareness, on individual subjective experience of a par- the other hand, corresponds to the knowl- ticular subject, a goal that remains beyond edge that an event occurred but without the realms of science. Rather, the goal is to any conscious recollection. It conveys a more account for populations of subjective expe- abstract sense of the past and future, based riences that may be experienced by numer- on feelings of familiarity (Tulving, 1985). ous subjects. Thus, first-person perspective It does not entail time travel but aware- approaches are aimed at defining these pop- ness of knowledge that one has about the ulations of subjective experiences as pre- world in which one lives. Unlike autonoetic cisely as possible and measuring them in a awareness, noetic awareness does not enable reproducible way (Gardiner, 2000). us to re-experience personal events in a self-reflective way (Gardiner, 2000). Tulving suggests that memory systems should be First-person Perspective Approaches to redefined in accordance with the related Recognition Memory: The Distinction subjective experience at retrieval. In this Between Autonoetic and Noetic context, autonoetic awareness stems from an Awareness episodic system, whereas noetic awareness All of us, at least once in our lives, have stems from a semantic system. recognized someone as being familiar but have not been able to remember who he The Remember/Know Procedure or she was or been able to recollect anything about the person and our previous To investigate the distinction between auto- encounter with him or her. Similarly, we can noetic and noetic awareness experimentally, know that we have read a book or watched Tulving (1985) developed the remember/ P1: KAE 0521857430c18 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:51

can we study subjective experiences objectively? 483

know procedure, an experiential procedure influence remember and know responses in in which the states of awareness related an opposite way, and finally, some variables to memory recognition are measured. Typi- influence remember and know responses in cally, participants are asked to report their a parallel way (for a review, see Gardiner subjective state of awareness at the time & Richardson-Klavehn, 2000). These results they recognize each individual item. They show that remember and know responses make a remember response if recognition is are not only dissociable but also function- accompanied by the conscious recollection ally independent. They indicate that remem- of some specific feature of the item’s pre- ber responses involve strategic, intentional, sentation (where it was, what they thought, and goal-directed processes, whereas know etc.). Thus, remember responses are associ- responses are based on more perceptual ated with a qualitatively rich mental expe- processes. rience, including perceptual, spatial, tempo- The second type of evidence comes ral, semantic, emotional, and other details from studies carried out in brain-damaged that are attributed to the past learning phase patients and neuroimaging studies. These (Johnson, Hashtroudi, & Lindsay, 1993). studies show that remember and know These remember responses index auto- responses are associated with the activation noetic awareness. Participants make a know of distinct neural substrates (Eldridge et al., response if recognition is associated with 2000; Henson et al., 1999; Yonelinas, 2002). feelings of familiarity in the absence of con- Broadly speaking, remember responses are scious recollection. Thus, know responses associated with activations of left prefrontal are associated with the simple knowledge and hippocampal regions, whereas know that an item has been seen previously. They responses are associated with activations index noetic awareness. of right prefrontal and parahippocampal Recent studies using the remember/know regions. Taken together, these findings lend procedure suggest that some know responses much weight to the view that remember and are not in fact based on feelings of familiar- know responses index two distinct subjec- ity but are simply guesses (Gardiner, Java, tive states of conscious awareness. & Richardson-Klavehn, 1996): Participants guess that they studied an item previously but do not experience familiarity (know- First-Person Perspective Approaches ing) or recollect any details from the learn- to Conscious Awareness in ing phase (remembering). To distinguish Schizophrenia between knowing and guessing, a third category of responses, namely guess responses, Henry Ey (1963) was the first person to has been added in some studies. postulate that schizophrenia is primarily a Following the first study by Tulving in disorder of consciousness. He argued that 1985, the remember/know procedure has an impairment of consciousness is associ- been used widely in numerous recognition ated with the typical impairment of the self memory studies. Findings from these stud- in patients with schizophrenia. However, ies provide evidence that the scientific study his view of schizophrenia as a disorder of of subjective experiences is relevant. The consciousness was based on a philosophical first type of evidence comes from reports of premise, and concepts and methods to assess systematic and replicable dissociations and consciousness empirically did not exist at associations between remember and know the time. Recently, several theoretical mod- responses as a function of various experi- els of schizophrenia have reformulated the mental manipulations. A fourfold patterns hypothesis of schizophrenia as a disorder of of outcomes has been observed: Some vari- consciousness with reference to the concep- ables influence remember but not know tualization of consciousness as a function. responses, some variables influence know Nancy Andreasen (1999) argues that the dis- but not remember responses, some variables ruption of the fluid, coordinated sequences P1: KAE 0521857430c18 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:51

484 the cambridge handbook of consciousness

of thought and action that underlie con- the validity of using first-person perspective sciousness in normal subjects is the funda- approaches seems to be especially critical in mental deficit in schizophrenia. Frith (1992) these patients. This question is so crucial that regards schizophrenia as a disorder of con- some psychiatrists and psychologists deny sciousness, impairing the ability to think the scientific interest of these approaches using metarepresentations, which are repre- in schizophrenia. It has to be said, though, sentations of mental states. that this denial seems somewhat paradoxi- Several studies using third-person per- cal as it implies that first-person perspective spective approaches to consciousness approaches should not be applied to subjects provide consistent experimental evidence for which they are the most likely to be inter- for an impairment of consciousness in esting. We argue that the only way to deal schizophrenia. Patients with schizophrenia satisfactorily with the issue of the validity exhibit a dissociation between impaired per- of using first-person perspective approaches formance in explicit tasks, such as recall and in patients with schizophrenia is to exam- recognition ones, in which participants are ine available empirical data. In the next part required to retrieve information from mem- of the chapter, we present the results of ory consciously (Clare, McKenna, Mortimer, our studies using the remember/know pro- & Baddeley, 1993; Gras-Vincendon et al., cedure in patients with schizophrenia. 1994), and preserved performance in implicit tasks, such as perceptual priming Impairment of Autonoetic Awareness 1994 tasks (Gras-Vincendon et al., ) and in Schizophrenia procedural memory tasks (Goldberg, Saint- Cyr, & Weinberger, 1990; Michel, Danion, A set of studies using the remember/know Grange, & Sandner, 1998), for which sub- procedure to assess subjective states of jects are not required to retrieve material awareness in patients with schizophre- consciously. Performance of patients with nia showed that autonoetic awareness is schizophrenia is also intact in implicit impaired. learning tasks in which the acquisition of knowledge is also implicit (Danion, Gokals- impairment of word frequency ing, Robert, Massin-Krauss, & Bacon, 2001). effect in remember responses Furthermore, patients with schizophrenia Huron et al. (1995) used a recognition experience a dissociation between preserved memory task including high- and low- automatic subliminal priming and impaired frequency words. The results show that conscious control (Dehaene et al., 2003). the level of remember responses is reduced Together with evidence of impaired for low-frequency words in patients with awareness of self-generated action (Franck schizophrenia, whereas the number of et al., 2001), these results converge to sug- remember responses for high-frequency gest that an impairment of conscious aware- words and the number of know responses ness might be a core deficit in schizophrenia. for both high- and low-frequency words However, this assumption is drawn from an do not differ between groups. There- inference based on indirect evidence, and the fore, patients with schizophrenia do not use of first-person perspective approaches in exhibit the word frequency effect (more patients with schizophrenia to measure con- remember responses for low-frequency scious awareness directly might be a particu- than high-frequency words) observed in larly relevant and informative way of finding normal subjects (see also Gardiner & Java, out more about the cognitive mechanisms of 1990). The word frequency effect has been this mental disease. accounted for in normal subjects by encod- But if it is conceded that mental disor- ing differences in information processing ders may impair not only subjective experi- that appear during the study phase: The dis- ences but also the ability of patients to assess tinctive low-frequency words undergo more these subjective experiences, the question of strategic processing than the less distinctive P1: KAE 0521857430c18 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:51

can we study subjective experiences objectively? 485

high-frequency words. Therefore, the nition and related states of awareness in absence of a word frequency effect on schizophrenia (Huron & Danion, 2002). The remember responses in patients with schi- results show that patients with schizophre- zophrenia suggests that the impairment of nia recognize fewer critical lures (false recog- autonoetic awareness may be attributed to nition) and studied words (correct recogni- a failure of strategic processes engaged at tion) than normal subjects. This deficit is encoding. restricted to items associated with remember responses. The proportion of know impairment of false memories responses does not differ between groups. associated with remember The results confirm the selective impair- responses ment of autonoetic awareness associated We have also studied subjective states of with true memories and extend these find- awareness associated with false memories – ings to false memories. They are consis- that is, memories for events that never hap- tent with an impairment of strategic pro- pened – in schizophrenia. The most widely cesses in schizophrenia. They also indicate used experimental procedure to induce false that the mere construction of memories memories in normal subjects is that initially and autonoetic awareness is defective in this introduced by Deese and subsequently mod- pathology. ified by Roediger and McDermott (1995). In this procedure, subjects study lists of impairment of contents of 15 words that are semantically related to a autonoetic awareness non-presented theme word or critical lure. As well as studying the frequency of For instance, the words presented for the autonoetic awareness in schizophrenia, we critical word mountain are hill, valley, climb, investigated the content of autonoetic summit, top, molehill, peak, plain, glacier, goat, awareness (Sonntag et al., 2003). More pre- bike, climber, range, steep, ski. A subsequent cisely, we used a remember/know procedure recognition test includes both previously together with a directed forgetting paradigm presented words and non-presented critical to investigate the contents of awareness at words, along with unrelated new items. In retrieval depending on whether information normal subjects, this procedure induces a has been identified as relevant or irrelevant robust false recognition effect for the crit- at encoding. In this paradigm, patients with ical lures (mountain, in this case). More- schizophrenia and comparison subjects are over, when subjects are asked to report, for presented with words and instructed to learn each item they identify as being old, whether half of them and forget the other half. The they remember or know that the item was instruction “to be learned” or “to be forgot- on the list they studied, the false recogni- ten” occurs just after each word is presented tion of critical lures is most often accompa- during the study phase. The recognition task nied by an experience of remembering. It is carried out on all the words presented has been hypothesized that this false recol- previously, mixed with new words. Partici- lection involves strategic processes (Holmes pants are instructed to identify all the words et al., 1998; Mather et al., 1997). On the from the study list irrespective of whether whole, studies of false memories in normal the words were to be learned or forgotten subjects provide direct evidence that memo- and to report their subjective state of aware- ries and associated awareness are not a literal ness at the time they recognize a word. This reproduction of the past but depend instead approach tells us about the strategic regula- on constructive and reconstructive processes tion of the content of awareness for relevant that are sometimes prone to errors and dis- information, which is beneficial to recollect, tortions (Conway, 1997; Holmes et al., 1998; and irrelevant information, which is benefi- Schacter et al., 1998). cial to forget. The results show that both nor- We used the Deese/Roediger-McDer- mal subjects and patients with schizophre- mott approach to investigate false recog- nia recognize more to-be-learned than P1: KAE 0521857430c18 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:51

486 the cambridge handbook of consciousness

to-be-forgotten words, indicating that both investigator when necessary. All participants groups exhibited a directed forgetting effect. performed a practice test on 10 items, 5 of However, whereas the effect was observed which were presented just after the items to both for remember and know responses in be studied in the main test and 5 of which comparison subjects, it was observed for were new items. For each item, subjects were know, but not for remember, responses in asked whether they recognized it as hav- patients. This experiment provides evidence ing been presented previously or not. When that schizophrenia impairs the relevance of they recognized an item, they were asked to the content of autonoetic awareness. It is select a remember, know, or guess response. possible that patients, unlike comparison At the end of the practice test, they were subjects, fail to engage the strategic regu- asked to explain each response to check that lation of encoding that makes the relevant they had correctly interpreted the instruc- information easier to retrieve than the irrel- tions. Throughout all of this, there was no evant information. indication that patients had any difficulty understanding or remembering the instruc- Is the Remember/Know Procedure tions. The very few participants who failed Valid in Schizophrenia? to perform the practice test properly were Evidence for the validity of using the left out of the experiment. They represent 5 remember/know procedure in schizophre- less than % of the overall participants in nia is provided by checking that patients our studies and include both patients with with schizophrenia properly understand and schizophrenia and normal subjects. These apply the task instructions, demonstrating findings confirm that the remember/know that some experimental variables induce distinction is psychologically relevant not 2000 the same patterns of responses in patients only in normal subjects (Gardiner, ) but and in controls, and showing the consis- also in patients with schizophrenia. tency of findings from first-person and third- Another possibility is to ask participants, person perspective approaches to recogni- at the end of the main recognition task, tion memory. to explain their remember responses by reporting exactly what they remembered. do patients with schizophrenia Like comparison subjects, patients with properly understand and apply schizophrenia explain these responses by the instructions during the recollection of highly specific details from remember/know procedure? the learning phase. However, a more precise When using the remember/know procedure analysis of these explanations shows that, in patients with schizophrenia, it is par- in some experimental conditions, patients ticularly important to check carefully that with schizophrenia report fewer associa- patients fully understand the instructions tions between words from the study list given for the task and apply them prop- than comparison subjects, whereas they erly. Because a proper understanding of recollect as many associations with per- the distinction among remember, know, and sonal events (Huron et al., 1995). This guess responses is critical to the task, we finding does not raise any doubts about took numerous precautions to ensure that the accuracy of the remember responses the subjects fully understood the meanings reported by patients with schizophrenia. of these responses in all our studies using Indeed, it is likely that these differences this procedure in patients with schizophre- reflect the failure of strategic processes in nia. Instructions were presented orally and schizophrenia: Associations between stud- then in written form. Some examples from ied words require intentional, strategic orga- everyday life were described, and subjects nization of the information to be learned, were asked whether they would choose a whereas the spontaneous evocation of a per- remember, know, or guess response for each sonal event may be triggered automatically instance. Corrections were made by the by a studied word. This interpretation is P1: KAE 0521857430c18 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:51

can we study subjective experiences objectively? 487

consistent with the view expressed by jective experiences of awareness that are Yonelinas (2002) that autonoetic awareness qualitatively similar to those experienced by sometimes depends on strategic processes comparison subjects. and sometimes depends on more automatic Evidence that the memory of the source processes. Thus, schizophrenia might specif- of an item and, more generally, the mem- ically impair autonoetic awareness based ory of an association is better for a con- on strategic processes, but appears to spare sciously recollected item than for a familiar autonoetic awareness involving more auto- item also seems to demonstrate the valid- matic processes. ity of using remember and know responses A further way of assessing the validity of (Conway & Dewhurst, 1995; Perfect, Mayes, using remember, know, and guess responses Downes, & Van Eijk, 1996). Such evidence is to compare patients with schizophrenia has been found in patients with schizophre- and comparison subjects in terms of the nia (Danion et al., 1999). We have used a qualitative characteristics or, in other words, source recognition memory task to measure the perceptual, spatial, temporal, semantic, the relation between defective autonoetic and emotional attributes of the subjective awareness and impaired source memory in experience for each reported response. This schizophrenia. During the study phase, par- kind of assessment has been performed in ticipants are presented with a set of com- a study of the picture superiority effect, mon objects (e.g., a candle, a toothbrush, a in which participants were instructed not handkerchief, a battery, and a tire pump). only to report a remember, know, and guess They are instructed to make pairs of objects response for each recognized item but also by positioning an object next to another to rate the specific qualitative characteris- (e.g., the subject has to put the candle next tics of their memory on visual analog scales to the tire pump) or to watch the experi- (Huron et al., 2003). The picture superi- menter perform the action (e.g., the exper- ority effect describes the finding that it is imenter puts the toothbrush next to the typical for normal subjects to recognize pic- battery). In this way, participants have to tures more readily than words in a subse- study complex events, which each consist quent recognition memory task. Moreover, of target information (a pair of objects) and this effect is mainly related to recognition source contextual information (who paired accompanied by remember responses. Our the two objects). results show that patients with schizophre- In a recognition task, participants are pre- nia exhibit a lower picture superiority effect sented with pairs of objects. All the objects selectively related to remember responses have been presented during the study phase, than comparison subjects. Most importantly, so that the recognition of objects has no they show that the qualitative characteristics influence on the recognition of pairs. As of memories do not differ between patients presented, the pairs consist of old pairs of with schizophrenia and controls. Despite two old objects occurring in their previ- the lower frequency of remember responses ous combination and new pairs of two old in patients with schizophrenia, when they objects occurring in a new combination. report a remember response, the qualita- Accordingly, correct recognition of old pairs tive characteristics of this subjective state depends on specific associations between of awareness seem to be similar to those objects made by participants during the reported by comparison subjects. Moreover, study session, which make the pairs distinc- in both groups the qualitative character- tive. Participants are asked to identify old istics of subjective experiences associated pairs (recognition of pairs of objects) and to with remember responses are quite different make a remember or know response for the from those associated with know and guess pair. They then have to say whether they responses. These findings suggest that the performed the action or watched it (source remember and know responses of patients recognition) and to make a remember or with schizophrenia index two distinct sub- know response for the action. P1: KAE 0521857430c18 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:51

488 the cambridge handbook of consciousness

The performance of patients with patients and controls. Evidence that an schizophrenia is particularly impaired in experimental variable has the same impact observed actions, as both recognition of on remember responses in both groups pairs of objects and recognition of source would show that patients with schizophre- are impaired. This impairment is associated nia and comparison subjects behave similarly with a reduction in the frequency of remem- during this first-person perspective task. ber, but not know, responses. Comparison Such evidence has been provided in a study subjects make few errors in source recogni- in which the effect of the affective valence tion when recognition of pairs of objects is (positive, negative, or neutral) of words on accompanied by remember responses. They subjective states of awareness was compared make significantly more errors when recog- between patients with schizophrenia and nition is accompanied by know responses. comparison subjects. The results show lower Patients with schizophrenia make numer- levels of remember responses in patients. ous source recognition errors. However, However, like comparison subjects, patients their performance is better for remember report more remember responses for emo- responses than for know responses, albeit to tional words than for neutral words. In con- a lesser degree than in comparison subjects. trast, the level of know responses is not Both groups perform better when they influenced by emotional words. Evidence paired the objects themselves: Pair recogni- that both patients and comparison subjects tion, source recognition, and the frequency consciously recollect emotional words bet- of remember responses increase. Source ter than neutral words suggests that the recognition performance improves when impact of the emotional valence of words recognition of pairs of objects is accompa- on autonoetic awareness is preserved in nied by remember, but not know, responses. schizophrenia. Therefore, in patients with schizophrenia, as in comparison subjects, subjective reports Are the Results from First-Person and of awareness and objective measures of Third-Person Perspective Approaches memory (recognition of pairs of objects to Recognition Memory Consistent and source recognition) are consistent in in Schizophrenia? all experimental conditions. Evidence that source recognition is higher for consciously The distinction between two subjective recollected pairs of objects than for familiar states of consciousness proposed by Tulving pairs provides a powerful argument for (1985) is similar to the distinction between the validity of using remember and know two types of memory processes or sys- responses in schizophrenia. Moreover, tems, generally referred to as conscious rec- these results indicate that patients with ollection and familiarity,1 reported in dual schizophrenia are less able than comparison recognition memory models. These mod- subjects to link the separate aspects of els, which have been developed mainly by events together into a cohesive, memorable, Atkinson and colleagues, Mandler, Jacoby, and distinctive whole. and Yonelinas (reviewed in Yonelinas, 2002), have been tested using a vari- do some experimental variables ety of third-person perspective methods, induce the same pattern of including recall/recognition comparisons, remember/know responses in item/associative recognition comparisons patients with schizophrenia and the process-dissociation procedure. as in controls? Yonelinas (2002) took advantage of the sim- Most of the studies that have used the ilarity between the remember/know and the remember/know procedure in schizophre- conscious recollection/familiarity distinc- nia have shown that the effect on remember tions to compare findings from first-person responses of an experimental variable (i.e., and third-person perspective approaches to word frequency, picture) differs between recognition memory in normal subjects and P1: KAE 0521857430c18 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:51

can we study subjective experiences objectively? 489

brain-lesioned patients. He showed that recall fewer items than control subjects findings from the remember/know proce- (Culver et al., 1986; Gerver, 1967;Koh dure are consistent with those from third- & Kayton, 1974; Koh et al., 1973, 1980; person perspective approaches, providing McClain, 1983; Russel & Beekhuis, 1976; evidence to support the validity of using Sengel & Lovallo, 1973; Truscott, 1970). The first-person perspective approaches in these robustness of this deficit has been confirmed populations. In keeping with this line of by a recent meta-analysis (Aleman et al., reasoning, we review the studies that used 1999) including 70 studies on long-term these third-person perspective approaches memory in schizophrenia that reported a to recognition memory in schizophrenia. large effect size for recall. This deficit occurs Doing so enables us to draw inferences about for verbal and non-verbal stimuli. recollection and familiarity in schizophre- In studies using recognition tasks, it nia and to compare the results from these is sometimes reported that patients with studies with those from first-person perspec- schizophrenia perform worse than normal tive studies. Consistent findings will pro- controls (Barch et al., 2002; Russel et al., vide further arguments for the validity of 1975; Sullivan et al., 1997; Traupman, 1975), using first-person perspective approaches but sometimes there seems to be no dif- to recognition memory in patients with ference (Koh et al., Exp. 2, 1973; Rushe schizophrenia. et al., 1999). A meta-analysis by Aleman et al. (1999) indicates that recognition is recall/recognition comparison less impaired than recall. Studies that assess The rationale underlying third-person per- both recall and recognition performance in spective approaches to conscious recog- the same patients with schizophrenia lead nition is similar to that underlying the to the same conclusion. Some of these stud- above-mentioned third-person perspective ies found impaired recall along with nor- approaches to conscious awareness. How- mal recognition in patients compared to ever, instead of using tasks selected to controls (Bauman, 1971; Bauman & Muray, compare conscious and unconscious pro- 1968; Beatty et al, 1993; Koh & Peterson, cesses, these approaches use tasks selected to 1978; Nachmani & Cohen, 1969). Other compare recollection and familiarity. They studies have shown that, even if patients compare performance in task conditions with schizophrenia exhibit a recognition assumed to require one of the two recogni- deficit, they are nevertheless more impaired tion memory processes more than the other. in recall tasks than in recognition mem- For instance, performance in a recall task ory tasks (Brebion´ et al., 1997; Calev, 1984; is assumed to rely more on conscious rec- Calev et al., 1983; Chan et al., 2000; Gold ollection than on familiarity. On the other et al, 1992; Goldberg et al., 1989; Paulsen hand, whereas familiarity has few effects et al., 1995; Russel et al., 1975). In the few on recall performance, it contributes to per- studies in which recognition and recall tasks formance in an item recognition task to a have been matched for difficulty (e.g., Calev, greater extent. Accordingly, recall perfor- 1984), it has also been reported that performance is taken as an index of conscious mance is less impaired in recognition than recollection, whereas item recognition per- in recall. Because conscious recollection is formance is taken as an index of familiar- assumed to contribute more to recall than to ity. If a condition influences recall perfor- recognition performance, a greater impair- mance to a greater extent than recognition ment in recall provides evidence for a deficit performance, this condition is assumed to in recollection. influence conscious recollection more than familiarity. item/associative recognition On the whole, studies on recall in comparison schizophrenia show defective performance: A complementary third-person perspec- Patients with schizophrenia consistently tive approach contrasts performance in a P1: KAE 0521857430c18 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:51

490 the cambridge handbook of consciousness

single-item recognition task and an associa- to rely primarily on recollection, than in tive recognition task. In an associative recog- item recognition, which may reflect both nition task participants have to recollect a recollection and familiarity. They are conspecific association between a target infor- sistent with an impairment of conscious mation and contextual information. Per- recollection in schizophrenia. Direct evi- formance depends mainly on a conscious dence of a link between defective associa- recollection process because this process tive recognition memory and impaired con- requires the binding together of the dis- scious recollection has been provided by the tinct aspects of an event to be remem- study by Danion et al. (1999) referred to bered. Conversely, performance in an item- above using a combination of the remem- recognition task relies more on familiarity. ber/know procedure and source judgements Several studies have compared the perfor- in schizophrenia. mance of patients with schizophrenia in a single-item recognition task and an associa- process-dissociation procedure tive recognition task requiring memory of Both recall/recognition and item/associative contextual information. They have shown recognition comparisons provide consistent that patients with schizophrenia exhibit dis- findings about conscious recollection in proportionate deficits in associative recog- schizophrenia. However, these third-person nition tests compared to item-recognition perspective approaches are not designed tests. This finding has been confirmed by to estimate quantitatively the respective a meta-analysis (Achim & Lepage, 2005) contribution of conscious recollection and of 23 studies of recognition memory in familiarity processes. Jacoby (1991) has schizophrenia that observed an impairment developed the process-dissociation proce- 20% greater for associative recognition rela- dure to overcome this limitation (see Chap- tive to item recognition. In comparison to ter 10). This procedure was devised to sep- single-item recognition tests, patients per- arate mathematically the respective contri- form poorly in tests that require them to butions of consciously controlled and auto- remember when or where an item was matic memory processes to performance in presented (Rizzo et al., 1996a,b; Schwartz a single memory task by combining inclu- et al., 1991; Sullivan et al., 1997; Waters sion and exclusion test conditions. The pro- et al, 2004; but see Shoqueirat & Mayes, cedure initially proposed by Jacoby involves 1998), which modality it was presented in a source memory task. During the study (for instance, verbally or visually; Brebion´ phase, subjects are shown words that they et al., 1997), or how frequently it was pre- are not instructed to learn and then asked to sented (Gold et al., 1992; Gras-Vincendon learn words that they hear. During the test et al., 1994). Patients with schizophrenia phase, under the inclusion condition, partic- have also been found to exhibit deficits in ipants are instructed to give a yes response reality-monitoring tasks in which they have to all previously presented words; that is, to discriminate (1) self-generated informa- both to words they have seen and to words tion from information generated by an exter- they have heard. Under this condition, con- nal source (Brebion´ et al., 1997; Harvey, sciously controlled and automatic memory 1985; Keefe et al., 1999; Moritz et al., 2003; processes act in concert to facilitate per- Vinogradov et al., 1997; Waters et al, 2004), formance. Under the exclusion condition, (2) information from two external sources – participants are instructed to give a yes a male and a female voice, and (3) informa- response only to words that they have heard. tion from two internal sources – i.e., words Under this condition, consciously controlled they imagine themselves saying and words and automatic memory processes act in they imagine the experimenter saying (Keefe opposition: The controlled use of memory et al., 1999) or imagined answers and verbal- both increases the number of correct yes ized answers (Henquet et al., 2005). These responses (yes responses to words heard) and results suggest a greater deficit in associa- decreases the number of false alarms (yes tive recognition memory, which is assumed responses to words seen), whereas automatic P1: KAE 0521857430c18 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:51

can we study subjective experiences objectively? 491

influences increase the number of false cess of familiarity are less clear. Evidence alarms. Estimates of recollection and famil- from third-person perspective approaches iarity are derived from equations based on that recognition memory is sometimes intact performance (correct responses and false in schizophrenia is consistent with pre- alarms) under each condition. served familiarity. However, some studies Two studies (Kazes` et al., 1999; Linscott have shown lowered recognition perfor- & Knight, 2001) have used the process- mance, and the two studies in which the dissociation procedure to assess recollection process-dissociation-procedure was used to and familiarity in schizophrenia directly. In measure familiarity directly have produced a source memory task, Linscott and Knight discrepant findings: Familiarity is preserved (2001) reported lower estimates of conscious in one of these studies and increased in recollection in patients with schizophrenia the other. Studies using the remember/know than in comparison subjects, but no differ- procedure show no impairment of noetic ence between groups for the estimates of awareness as measured by know responses. familiarity. Using a version of the process- However, results from studies by Danion dissociation procedure that involves a word- et al. (2003) show a decrease in familiarity stem completion task, both Kazes` et al. when familiarity is estimated using an inde- (1999) and Linscott and Knight observed pendence model applied to the proportions reduced levels of conscious recollection in of remember and know responses (Yonelinas schizophrenia. This impaired conscious rec- et al., 1998). ollection was associated with a spared famil- These discrepancies prompted us to iarity in the study by Kazes` et al. (1999) review the results of all our remem- and an increased familiarity in Linscott and ber/know studies using the framework of the Knight’s study (2001). These two studies independence model devised by Yonelinas confirm the impairment of conscious recol- et al. (1998). The underlying assumption lection and do not provide any evidence of of this model is that conscious recollection a deficit in familiarity. and familiarity are independent processes, To summarize, findings from studies whereas the typical remember/know proce- of schizophrenia using first- and third- dure is based on a mutually exclusive rela- person perspective approaches are con- tion. The results of these reviews, which are cordant, showing a consistent impairment presented in Table 18.1, show that familiar- of autonoetic awareness and the underly- ity decreases in two studies (Danion et al., ing process of conscious recollection. In 2003; Huron et al., 1995), but not in all the combination with empirical evidence that others. patients with schizophrenia properly under- Overall, when results from both first- stand and use task instructions, and that person perspective and third-person pers- some experimental variables induce the pective approaches in schizophrenia are con- same response patterns in patients and in sidered, there is no evidence of a deficit comparison subjects, these findings provide in noetic awareness as measured by know substantial evidence of the validity of using responses and little evidence of an impair- the remember/know procedure in schizo- ment in the process of familiarity. Therefore, phrenia. we can conclude that, although a deficit in familiarity may be observed under certain experimental conditions, this deficit is much Does Schizophrenia Impair Noetic less pronounced than the deficit of conscious Awareness and Familiarity? recollection. However, this conclusion can- Whereas there is converging evidence not be regarded as definitive because no that autonoetic awareness and conscious study to date has taken the experimental recollection are consistently impaired in variables that are known to influence know schizophrenia, the results from both first- responses specifically in normal subjects and third-person perspective studies of and applied them to patients with schizo- noetic awareness and the underlying pro- phrenia. P1: KAE 0521857430c18 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:51 2 12 13 51 19 41 42 52 66 45 64 46 ...... F 0 0 0 0 0 0 * * * * * * 1 13 0 49 55 39 56 35 46 28 0 32 0 24 0 67 0 ...... 10 0 05 0 07 0 07 0 nses...... It has to be noted that old 17 0 13 0 10 0 10 0 04 0 09 0 ...... 01 0 16 0 02 0 02 0 02 0 05 0 03 0 03 0 03 0 ...... was subtracted from F 11 0 11 0 10 0 14 0 18 0 18 0 29 0 02 0 36 0 34 0 33 0 ...... ) from the proportion of remember responses for old new . . . . . new was estimated by the proportions of know responses to new 01 0 01 0 02 0 02 0 02 0 02 0 03 0 . . . . 002 0 . . 002 0 . new . . was estimated by the proportions of know responses to old items old 17 0 29 0 25 0 56 0 43 0 39 0 74 0 36 0 68 0 47 0 50 0 Remember Know Guess R ...... ). new 2 13 0 22 0 32 0 29 0 42 0 20 0 49 0 45 0 48 0 37 0 50 0 ...... F -R ). To calculate familiarity (F), F 1 . new 05 -R 1 1 ). Similarly, for new items, F <. 21 0 21 0 15 0 31 0 19 0 20 0 25 0 32 0 23 0 32 0 73 0 ...... p old -R 1 01 0 08 0 04 0 09 0 . . . . 13 0 14 0 06 0 07 0 07 0 07 0 ...... 01 0 02 0 02 0 06 0 04 0 04 0 05 0 03 0 03 0 ...... ) ) 2003 2003 13 0 19 0 16 0 19 0 14 0 28 0 07 0 38 0 33 0 36 0 40 0 ...... ) a) ) 01 0 01 0 01 0 01 0 01 0 02 0 03 0 07 0 04 0 . . . . . 2003 . . . . Patients with Schizophrenia Comparison Subjects 2002 1995 31 0 14 0 22 0 22 0 26 0 26 0 24 0 20 0 37 0 56 0 33 0 Remember Know Guess R ...... Old New Old New Old New Old New Old New Old New 0 0 0 0 0 0 0 0 0 0 0 Mean values of recollection (R) and familiarity (F) of patients with schizophrenia and comparison subjects, condition by condition, computed from ) and then dividing by the opportunity to assign a remember response to old items ( . old 1 . 18 words words items items Significant differences between patients with schizophrenia and comparison subjects, when the studies allow both know and guess responses, the proportion of know responses was replaced by the sum of the proportions of know and guess respo divided by the probability that an old item did not receive a remember response ( items (R items divided by the probability that a new item did not receive a remember response ( For each participant, estimates of familiarity were computed separately for old and new items. For old items, F For each participant, recollection (R) was calculated by subtracting the proportion of remember responses for new items (R individual values High-frequency Table Word frequency (Huron et al., ∗ 1 2 Negative Low-frequency Directed forgetting paradigm (Sonntag et al., To-be-learned Pictures False recognition (Huron et al., Studied Positive Affective valence (Danion et al., To-be-forgotten Words The picture superiority effect (Huron et al., Neutral Critical lures

492 P1: KAE 0521857430c18 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:51

can we study subjective experiences objectively? 493

Conclusions recollection might explain some behavioral abnormalities associated with schizophre- Throughout this chapter, we have argued nia, notably inadequate functional outcomes that the concepts and methods needed to in everyday life. Because autonoetic aware- study some subjective experiences objec- ness is severely disrupted in schizophrenia, a tively are already available. For instance, the past event cannot be used with great flexibil- first-person perspective approach proposed ity to guide and control behavior, affects, and by Tulving to assess subjective states of con- beliefs, which in turn are likely to be inap- scious awareness at retrieval seems to be propriate inasmuch as they can be driven valid not only in normal subjects (Yonelinas, only by noetic awareness or implicit mem- 2002) but also in patients with schizophre- ory. This probably explains why memory nia. The distinction between remember and impairments of patients with schizophre- know responses is relevant from a psycholog- nia are so consistently related to inade- ical viewpoint and is, more often than not, quate functional outcome in their daily lives appropriately applied by normal subjects (Green, 1996). and by patients with schizophrenia. More- over, as is indicated by numerous experimental and neuropsychological dissocia- Future Prospects tions, remember and know responses index two distinct subjective states of awareness: We conclude this chapter by looking at autonoetic and noetic awareness. These sub- some of the outstanding questions that jective states are not a literal reproduction of first-person perspective approaches have the past but instead a reconstruction of the opened up in the research field of psychi- past, which takes into account the present atry and clinical psychology. From a clin- time. Finally, this review shows a strong ical viewpoint, patients with schizophre- consistency of results from first-person and nia frequently experience perplexity about third-person perspective approaches. In the their own identity, which can take the future, it will be of interest to develop form of derealization and depersonalization. integrative and multidisciplinary approaches These symptoms are taken to be a distur- that combine first-person and third-person bance of the subjective sense of self. It has perspective methods in the same studies. been argued that the sense of self is sup- First-person perspective approaches com- ported by autobiographical events associ- bined with brain imaging (e.g., fMRI) will ated with autonoetic awareness (Conway also be required. & Pleydell-Pearce, 2000). Therefore, the The use of the remember/know proce- use of the remember/know procedure to dure in patients with schizophrenia pro- study subjective states of awareness associ- vides a better understanding of the cogni- ated with autobiographical memories makes tive impairments associated with the disease. it possible to study the subjective sense of It makes it possible to present a coherent the self. However, until now, studies using and accurate picture of the various recall the remember/know procedure in patients and recognition disturbances that have been with schizophrenia have been performed reported in these patients. Schizophrenia under conditions that have had little to seems to be characterized by an impairment do with real life and so prevent the gen- of autonoetic awareness and its underlying eralization of their conclusions to autobio- conscious recollection process. This impair- graphical memory. Indeed, the stimuli have ment results from a failure of strategic pro- been words and pictures, and the learn- cessing at encoding (e.g., Brebion et al., 1997; ing and test phases have been measured Gold et al., 1992; Koh & Peterson, 1978), in minutes or hours. These stimuli are not but an impairment of strategic processing at comparable with complex and meaning- retrieval cannot be ruled out. The impair- ful autobiographical events that have reten- ment of autonoetic awareness and conscious tion intervals measured in weeks, months, P1: KAE 0521857430c18 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:51

494 the cambridge handbook of consciousness

and years. Using an autobiographical mem- interact. The world subjectively experienced ory inquiry in combination with the is distinct from the external physical world. remember/know procedure, we showed a It is an active construction arising from our lower frequency and consistency among mind that we project outside it. This sub- patients of remember responses associated jective representation of the world has to be with autobiographical memories (Danion constrained by external physical realities in et al., 2005). This finding is consistent with order to be adaptative. Realism and ideal- the impairment of the sense of self reported ism lead to diametrically opposed views of by patients. There is preliminary evidence psychopathological manifestations, as illus- that this impairment might result from a trated by hallucinations. According to the defective construction of personal identity realism point of view, hallucinations are false occurring during adolescence or early adult- perceptions that arise in the absence of an hood (Riutort, Cuervo, Danion, Peretti, & external object or event. According to tran- Salame, 2003). scendental idealism, both normal percep- First-person perspective approaches tions and hallucinations are subjective expe- have also been developed to investigate riences subserved by the same internal pro- metamemory; that is, subjects’ knowledge cess. They only differ in respect to the extent about their own memory capabilities. to which they are constrained by sensory They make it possible to study subjective input from the external world. Such a posi- experiences related to the knowledge that tion might open up new prospects for the subjects possess about the functioning of understanding of hallucinations (Behrendt & their memory. These experiences can be Young, 2004), and more generally of the sub- evaluated either qualitatively, such as the jectivity impairments associated with men- phenomenon of something being on the tal disorders. tip of the tongue, or quantitatively, such as the Feeling of Knowing or the Judgment of Confidence. Applying these approaches to Note the study of schizophrenia or other mental diseases opens up a new field of research, 1. Conscious recollection and familiarity are which is still virtually unexplored (but sometimes used to describe subjective states of see Bacon, Danion, Kauffmann-Muller, & awareness, as well as cognitive processes under- Bruant, 2001; Danion et al., 2001). lying recognition memory. The use of the same From a more theoretical point of view, terms to describe separate concepts is confusing. Indeed, in the former case, conscious rec- because first-person perspective methods ollection and familiarity refer to experimen- focus on the subjective dimension of psy- tal data in the form of remember and know chopathological manifestations, they may be responses, whereas in the latter case, they refer the first step in a major conceptual change to hypothetical constructs – processes – arising that would modify our understanding of from a theoretical model and its underlying these manifestations. Behrendt and Young hypotheses. To avoid any confusion, we use (2004) point out that psychiatry has usually autonoetic and noetic awareness to qualify adopted a philosophical position of realism subjective states of awareness associated with that assumes that the world that we perceive recognition memory and conscious recollec- is an objective reality. This world is thought tion and familiarity to qualify cognitive pro- to exist independently of those who perceive cesses underlying recognition memory. it and not to be a product of their mind. In contrast, Gestalt psychologists, in keep- References ing with the philosophical position of transcendental idealism of Kant, consider that Achim, A. M., & Lepage, M. (2005). Episodic a clear distinction has to be made between memory-related activation in schizophrenia: the world that we subjectively perceive and meta-analysis. British Journal of Psychiatry, the external physical world with which we 187(12), 500–509. P1: KAE 0521857430c18 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:51

can we study subjective experiences objectively? 495

Aleman, A., Hijman, R., de Haan, E. H., & nia: What is impaired and what is pre- Kahn, R. S. (1999). Memory impairment in served?Neuropsychologia, 31(11), 1225–1241. schizophrenia: a meta-analysis. American Jour- Conway, M. A. (1997). Past and present: Recov- nal of Psychiatry, 156(9), 1358–1366. ered memories and false memories. In M. A. Andreasen, N. C. (1999). A unitary model of Conway (Ed), Recovered memories and false schizophrenia: Bleuler’s “fragmented phrene” memories. (pp. 150–191). New York: Oxford as schizencephaly. Archives of General Psychia- University Press. try, 56(9), 781–787. Conway, M. A., & Dewhurst, S. A. (1995). Bacon, E., Danion, J. M., Kauffmann-Muller, Remembering, familiarity, and source moni- F., & Bruant, A. (2001). Consciousness in toring. The Quarterly Journal of Experimental schizophrenia: A metacognitive approach to Psychology A, Human Experimental Psychology, semantic memory. Consciousness and Cogni- 48(1), 125–140. tion, 10(4), 473–484. Conway, M. A., & Pleydell-Pearce, C. W. (2000). Barch, D. M., Csernansky, J. G., Conturo, T., The construction of autobiographical memo- & Snyder, A. Z. (2002). Working and long- ries in the self-memory system. Psychological 107 2 261 288 term memory deficits in schizophrenia: Is there Review, ( ), – . a common prefrontal mechanism? Journal of Culver, L. C., Kunen, S., & Zinkgraf, S. A. (1986). Abnormal Psychology, 111(3), 478–494. Patterns of recall in schizophrenics and normal Bauman, E. (1971). Schizophrenic short-term subjects. Journal of Nervous and Mental Dis- 10 620 623 memory: A deficit in subjective organization. eases, 174( ), – . Canadian Journal of Behavioral Sciences, 3(1), Danion, J. M, Cuervo, C., Piolino, P., Huron, C., 2005 55–65. Riutort, M., Peretti, S., et al. ( ). Abnor- mal subjective sense of self in patients with Bauman, E., & Murray, D. J. (1968). Recognition schizophrenia. Consciousness and Cognition, versus recall in schizophrenia. Canadian Jour- 14(3), 535–47. nal of Psychology, 22(1), 18–25. Danion, J. M., Gokalsing, E., Robert, P., Massin- Beatty, W. W., Jocic, Z., Monson, N., & Staton, Krauss, M., & Bacon, E. (2001). Defective rela- R. D. (1993). Memory and frontal lobe dysfunc- tionship between subjective experience and tion in schizophrenia and schizoaffective dis- behavior in schizophrenia. American Journal of order. Journal of Nervous and Mental Diseases, Psychiatry, 158(12), 2064–2066. 181(7), 448–453. Danion, J. M., Kazes,` M., Huron, C., & Behrendt, R. P., & Young, C. (2004). Hallucina- Karchouni, N. (2003). Do patients with tions in schizophrenia, sensory impairment and schizophrenia consciously recollect emotional brain disease: A unifying model. Behavioral & events better than neutral events? American Brain Sciences, 27(6), 771–787. Journal of Psychiatry, 160(10), 1879–1881. Brebion, G., Amador, X., Smith, M. J., & Gor- Danion, J. M., Rizzo, L., & Bruant, A. (1999). 1997 man, J. M. ( ). Mechanisms underlying Functional mechanisms underlying impaired memory impairment in schizophrenia. Psycho- recognition memory and conscious awareness 2 383 393 logical Medicine, 27( ), – . in patients with schizophrenia. Archives of Gen- Calev, A. (1984). Recall and recognition in mildly eral Psychiatry, 56(7), 639–644. disturbed schizophrenics: the use of matched Dehaene, S., Artiges, E., Naccache, L., Martelli, 2 425 429 tasks. Psychological Medicine, 14( ), – . C., Viard, A., Schurhoff, F., et al. (2003). Calev, A., Venables, P. H., & Monk, A. F. Conscious and subliminal conflicts in normal (1983). Evidence for distinct verbal memory subjects and patients with schizophrenia: The pathologies in severely and mildly disturbed role of the anterior cingulate. Proceedings of schizophrenics. Schizophrenia Bulletin, 9(2), the National Academy of Science USA, 100(23), 247–264. 13722–13727. Chan, A. S., Kwok, I. C., Chiu, H., Lam, L., Eldridge, L. L., Knowlton, B. J., Furmanski, C. Pang, A., & Chow, L. Y. (2000). Memory and S., Bookheimer, S. Y., & Engel, S. A. (2000). organizational strategies in chronic and acute Remembering episodes: a selective role for the schizophrenic patients. Schizophrenia Research, hippocampus during retrieval. Nature Neuro- 41(3), 431–445. science, 3(11), 1149–1152. Clare, L., McKenna, P. J., Mortimer, A. M., & Ey, H. (1963). La conscience [Consciousness]. Baddeley, A. D. (1993). Memory in schizophre- Oxford: Presses Universitaires France. P1: KAE 0521857430c18 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:51

496 the cambridge handbook of consciousness

Franck, N., Farrer, C., Georgieff, N., Marie- schizophrenia? American Journal of Psychiatry, Cardine, M., Dalery, J., d’Amato, T., et al. 153(3), 321–330. (2001). Defective recognition of one’s own Harvey, P. D. (1985). Reality monitoring in mania actions in patients with schizophrenia. Amer- and schizophrenia. The association of thought ican Journal of Psychiatry, 158(3), 454–459. disorder and performance. Journal of Nervous Frith, C. D. (1992). The cognitive neuropsychol- and Mental Diseases, 173(2), 67–73. ogy of schizophrenia. Hove: Lawrence Erlbaum Henquet, C., Krabbendam, L., Dautzenberg, J., Associates. Jolles, J., Merckelbach, H. (2005). Confus- Gardiner, J. M. (2000). On the objectivity of sub- ing thoughts and speech: source monitoring jective experiences of autonoetic and noetic and psychosis. Psychiatry Research, 133(1), 57– consciousness. In E. Tulving (Ed.), Memory, 63. consciousness, and the brain: The Tallinn Con- Henson, R. N., Rugg, M. D., Shallice, T., Josephs, ference (pp. 159–172). Philadelphia: Psychology O., & Dolan, R. J. (1999). Recollection and Press. familiarity in recognition memory: an event- Gardiner, J. M., & Java, R. I. (1990). Recollective related functional magnetic resonance imaging experience in word and nonword recognition. study. Journal of Neurosciences, 19(10), 3962– Memory and Cognition, 18(1), 23–30. 3972. Gardiner, J. M., Java, R. I., & Richardson- Holmes, J. B., Waters, H. S., Rajaram, S. Klavehn, A. (1996). How level of process- (1998). The phenomenology of false memo- ing really influences awareness in recognition ries: episodic content and confidence. Journal memory. Canadian Journal of Experimental of Experimental Psychology: Learning, Memory, Psychology, 50(1), 114–122. and Cognition, 24(4), 1026–40. Gardiner, J. M., & Richardson-Klavehn, A. Huron, C., & Danion, J. M. (2002). Impair- (2000). Remembering and knowing. In E. ment of constructive memory in schizophre- Tulving & F. I. M. Craik (Eds.), Hand- nia. International Clinical Psychopharmacology, book of memory. Oxford: Oxford University 17(3), 127–133. Press. Huron, C., Danion, J. M., Giacomoni, F., Grange,´ Gerver, D. (1967). Linguistic rules and the per- D., Robert, P., & Rizzo, L. (1995). Impairment ception and recall of speech by schizophrenic of recognition memory with, but not without, patients. The British Journal of Social and Clin- conscious recollection in schizophrenia. Amer- ical Psychology, 6(3), 204–211. ican Journal of Psychiatry, 152(12), 1737–1742. Gold, J. M., Randolph, C., Carpenter, C. J., Gold- Huron, C., Danion, J. M., Rizzo, L., Killofer, berg, T. E., & Weinberger, D. R. (1992). Forms V., & Damiens, A. (2003). Subjective quali- of memory failure in schizophrenia. Journal of ties of memories associated with the picture Abnormal Psychology, 101(3), 487–494. superiority effect in schizophrenia. Journal of Goldberg, T. E., Saint-Cyr, J. A., & Wein- Abnormal Psychology, 112(1), 152–158. berger, D. R. (1990). Assessment of procedural Huron, C., Giersch, A., & Danion, J. M. (2002). learning and problem solving in schizophrenic Lorazepam, sedation, and conscious recollec- patients by Tower of Hanoi type tasks. Journal tion: a dose-response study with healthy vol- of Neuropsychiatry and Clinical Neuroscience, unteers. International Clinical Psychopharma- 2 (2), 165–173. cology, 17(1), 19–26. Goldberg, T. E., Weinberger, D. R., Pliskin, Jacoby, L. L. (1991). A process dissociation frame- N. H., Berman, K. F., & Podd, M. H. (1989). work – separating autonomic from intentional Recall memory deficit in schizophrenia. A pos- uses of memory. Journal of Memory and Lan- sible manifestation of prefrontal dysfunction. guage, 30, 513–541. Schizophrenia Research, 2 (3), 251–257. Johnson, M. K., Hashtroudi, S., & Lindsay, D. S. Gras-Vincendon, A., Danion, J. M., Grange, D., (1993). Source monitoring. Psychological Bul- Bilik, M., Willard-Schroeder, D., Sichel, J. P., letin, 114(1), 3–28. et al. (1994). Explicit memory, repetition prim- Kazes,` M., Berthet, L., Danion, J. M., Amado, I., ing and cognitive skill learning in schizophre- Willard, D., Robert, P., et al. (1999). Impair- nia. Schizophrenia Research, 13(2), 117–126. ment of consciously controlled use of memory Green, M. F. (1996). What are the functional in schizophrenia. Neuropsychology, 13(1), 54– consequences of neurocognitive deficits in 61. P1: KAE 0521857430c18 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:51

can we study subjective experiences objectively? 497

Keefe, R. S., Arnold, M. C., Bayen, U. J., & Har- mental Psychology A, Human Experimental Psy- vey,P.D.(1999). Source monitoring deficits chology, 49(3), 797–813. in patients with schizophrenia: a multino- Riutort, M., Cuervo, C., Danion, J. M., Peretti, C. mial modelling analysis. Psychological Medicine, S., & Salame, P. (2003). Reduced levels of spe- 29(4), 903–914. cific autobiographical memories in schizophre- Koh, S. D., & Kayton, L. (1974). Memoriza- nia. Psychiatry Research, 117(1), 35–45. tion of “unrelated” word strings by young Rizzo, L., Danion, J. M., Van Der Linden, M., & nonpsychotic schizophrenics. Journal of Abnor- Grange,´ D. (1996a). Patients with schizophre- mal Psychology, 83(1):14–22. nia remember that an event has occurred, but Koh, S. D., Kayton, L., & Berry, R. (1973). not when. British Journal of Psychiatry, 168(4), Mnemonic organization in young nonpsychotic 427–431. schizophrenics. Journal of Abnormal Psychol- Rizzo, L., Danion, J. M., Van Der Linden, ogy, 81(3), 299–310. M., Grange,´ D., & Rohmer, J. G. (1996b). Koh, S. D., Marusarz, T. Z., & Rosen, A. J. (1980). Impairment of memory for spatial context in Remembering of sentences by schizophrenic schizophrenia. Neuropsychology, 10(3), 376– young adults. Journal of Abnormal Psychology, 384. 89(2), 291–294. Roediger, H. L., & McDermott, K. B. (1995). Koh, S. D., & Peterson, R. A. (1978). Encod- Creating false memories: Remembering words ing orientation and the remembering of not presented in lists. Journal of Experimental schizophrenic young adults. Journal of Abnor- Psychology: Learning, Memory, and Cognition, mal Psychology, 87(3), 303–313. 21(4), 803–814. Linscott, R. J., & Knight, R. G. (2001). Auto- Rushe, T. M., Woodruff, P. W., Murray, R. M., matic hypermnesia and impaired recollection & Morris, R. G. (1999). Episodic memory and in schizophrenia. Neuropsychology, 15(4), 576– learning in patients with chronic schizophre- 585. nia. Schizophrenia Research, 35(1), 85–96. Mather, M., Henkel, L. A., & Johnson, M. K. Russell, P. N., Bannatyne, P. A., & Smith, J. F. (1997). Evaluating characteristics of false mem- (1975). Associative strength as a mode of orga- ories: Remember/know judgments and mem- nization in recall and recognition: a compari- ory characteristics questionnaire compared. son of schizophrenics and normals. Journal of Memory and Cognition, 25(6), 826–837. Abnormal Psychology, 84(2), 122–128. McClain, L. (1983). Encoding and retrieval in Russell, P. N., & Beekhuis, M. E. (1976). Organi- schizophrenics’ free recall. Journal of Nervous zation in memory: a comparison of psychotics and Mental Diseases, 171(8), 471–479. and normals. Journal of Abnormal Psychology, Michel, L., Danion, J. M., Grange, D., & 85(6), 527–534. Sandner, G. (1998). Cognitive skill learning Sass, L. A., & Parnas, J. (2003). Schizophrenia, and schizophrenia: Implications for cognitive consciousness, and the self. Schizophrenia Bul- remediation. Neuropsychology, 12(4), 590–599. letin, 29(3), 427–444. Moritz, S., Woodward, T. S., & Ruff, C. C. (2003). Schacter, D. L., Norman, K. A., & Koutstaal, Source monitoring and memory confidence in W. (1998). The cognitive neuroscience of con- schizophrenia. Psychological Medicine, 33(1), structive memory. Annual Review of Psychology, 131–139. 49, 289–318. Nachmani, G., & Cohen, B. D. (1969). Recall Schwartz, B. L., Deutsch, L. H., Cohen, C., War- and recognition free learning in schizophrenics. den, D., & Deutsch, S. I. (1991). Memory for Journal of Abnormal Psychology, 74(4), 511–516. temporal order in schizophrenia. Biological Psy- Paulsen, J. S., Heaton, R. K., Sadek, J. R., Perry, chiatry, 29(4), 329–339. W., Delis, D. C., Braff, D., et al. (1995). The Sengel, R. A., & Lovallo, W. R. (1983). Effects nature of learning and memory impairments in of cueing on immediate and recent memory in schizophrenia. Journal of the International Neu- schizophrenics. Journal of Nervous and Mental ropsychological Society, 1(1), 88–99. Diseases, 171(7), 426–430. Perfect, T. J., Mayes, A. R., Downes, J. J., & Shoqeirat, M. A., & Mayes, A. R. (1988). Spa- Van Eijk, R. (1996). Does context discrimi- tiotemporal memory and rate of forgetting in nate recollection from familiarity in recogni- acute schizophrenics. Psychological Medicine, tion memory? The Quarterly Journal of Experi- 18(4), 843–853. P1: KAE 0521857430c18 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 12, 2007 23:51

498 the cambridge handbook of consciousness

Sonntag, P., Gokalsing, E., Olivier, C., Robert, Vinogradov, S., Willis-Shore, J., Poole, J. H., P., Burglen, F., Kauffmann-Muller, F., et al. Marten, E., Ober, B. A., & Shenaut, G. K. (2003). Impaired strategic regulation of con- (1997). Clinical and neurocognitive aspects tents of conscious awareness in schizophrenia. of source monitoring errors in schizophrenia. Consciousness and Cognition, 12(2), 190–200. American Journal of Psychiatry, 154(11), 1530– Sullivan, E. V., Shear, P. K., Zipursky, R. B., 1537. Sagar, H. J., & Pfefferbaum, A. (1997). Patterns Waters, F. A. V., Maybery, M. T., Badcock, J. of content, contextual, and working mem- C., Michie, P. T. (2004). Context memory ory impairments in schizophrenia and nonam- and binding in schizophrenia. Schizophrenia nesic alcoholism. Neuropsychology, 11(2), 195– Research, 68(2-3), 119–125. 206 . Yonelinas, A. P. (2002). The nature of recollec- Traupmann, K. L. (1975). Effects of categorization and familiarity: A review of 30 years of tion and imagery on recognition and recall by research. Journal of Memory & Language, 46(3), process and reactive schizophrenics. Journal of 441–517. 4 307 314 Abnormal Psychology, 84( ), – . Yonelinas, A. P., Kroll, N. E. A., Dobbins, I., Truscott, I. P. (1970). Contextual constraint and Lazzara, M., & Knight, R. T. (1998). Recollec- schizophrenic language. Journal of Consulting tion and familiarity deﬁcits in amnesia: Con- and Clinical Psychology, 35(2), 189–194. vergence of remember-know, process dissocia- Tulving, E. (1985). Memory and consciousness. tion, and receiver operating characteristic data. Canadian Psychology, 26(1), 1–12. Neuropsychology, 12(3), 323–339. P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

CHAPTER 19 Meditation and the Neuroscience of Consciousness: An Introduction

Antoine Lutz, John D. Dunne, and Richard J. Davidson

Abstract on the brain and body of long-term practitioners. After an overview of the mecha- The overall goal of this chapter is to explore nisms of mind-body interaction, this section the initial findings of neuroscientific research addresses the use of first-person expertise, on meditation; in doing so, the chapter especially in relation to the potential for also suggests potential avenues of further research on the neural counterpart of sub- inquiry. It has three sections that, although jective experience. In general terms, the sec- integral to the chapter as a whole, may tion thus points to the possible contributions also be read independently. The first sec- of research on meditation to the neuro- tion, “Defining Meditation,” notes the need science of consciousness. The final section, for a more precise understanding of med- “Neuroelectric and Neuroimaging Correla- itation as a scientific explanandum. Argu- tes of Meditation,” reviews the most relevant ing for the importance of distinguishing the neuroelectric and neuroimaging findings of particularities of various traditions, the sec- research conducted to date, including some tion presents the theory of meditation from preliminary correlates of the previously dis- the paradigmatic perspective of Buddhism, cussed Buddhist practices. and it discusses the difficulties encountered when working with such theories. The section includes an overview of three prac- Introduction tices that have been the subject of research, and it ends with a strategy for developing This chapter discusses possible contributions a questionnaire to define more precisely a of meditation to the neurobiological study of practice under examination. The second sec- consciousness and to cognitive and affective tion, “The Intersection of Neuroscience and neurosciences in general. Empirical research Meditation,” explores some scientific motion meditation started in the 1950s, and as vations for the neuroscientific examination much as 1,000 publications on meditation of meditation in terms of its potential impact already exist.1 Despite such a high number 499 P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

500 the cambridge handbook of consciousness

of scientific reports and inspiring theoret- wide range of practices. Thus, in a typical ical proposals (Austin, 1998; Shapiro & discussion of this kind, West (1987) argues Walsh, 1984; Varela, Thompson, & Rosch, that practices as diverse as the ritual dances 1991; Wallace, 2003; West, 1987), one still of some African tribes, the spiritual exer- needs to admit that little is known about cises of the desert fathers, and the tantric the neurophysiological processes involved practices of a Tibetan adept are all forms in meditation and about its possible long- of meditation. Historically, this attempt to term impact on the brain. The lack of sta- categorize diverse practices under the same tistical evidence, control populations, and rubric reflects some intellectual trends in the rigor of many of the early studies; the het- early 20th century, most especially “perenni- erogeneity of the studied meditative states; alism,” that argue unequivocally for a cer- and the difficulty in controlling the degree tain genre of mystical experience as the of expertise of practitioners can in part essence of religion (Proudfoot, 1985; Sharf, account for the limited contributions made 1998). From the standpoint of the neuro- by neuroscience-oriented research on medi- sciences, the problem with such a position tation. Thus, instead of providing a complete is that it begins from a set of hypotheses review of this empirical literature (Austin, that are difficult to test because they assume 1998; Cahn & Polich, 2006; Delmonte, 1984, that the common element in mystical expe- 1985; Fenwick, 1987; Holmes, 1984; Pagano rience necessarily transcends thought, lan- & Warrenburg, 1983) we choose to address guage, reason, and ordinary perception – our central question from three directions. most of which are required for any reli- The purpose of this first section is to clar- able neuroscientific procedure to test the ify conceptually what the term “meditation” hypotheses. means and to propose an operational def- In addition to the problem of unverifiable inition. We focus on Buddhist meditative hypotheses, the generic use of meditation practices as a canonical example. We pro- as applying to such a wide range of diverse vide a short presentation of the main tenets practices inevitably trivializes the practices of Buddhist psychology and epistemology, as themselves. For example, the unique tech- well as a description of the standard techniques and context of Sufi zikr must be niques used in many Buddhist practices. ignored if they are to be considered the From these standard claims, we then derive same as the Taoist practice of T’ai Chi. In the possible contributions of meditation to short, to make zikr and T’ai Chi describ- neurosciences and develop tentative propos- able with the same term, one must ignore als for a neuroscientific understanding of a good deal of what makes them radically the cognitive and affective processes that different from each other. This would be are altered by training in meditation. In the akin to the use of the word “sport” to last section, we review existing neuroelectric refer to all sports as if they were essen- and neuroimaging findings on meditation, as tially the same. A typical result of such an well as some preliminary correlates of these approach is the extremely general model Buddhist practices. proposed by Fischer (1971) in which all forms of meditation – exemplified by Zazen and some unspecified “Yoga” practice – fall along 1. Defining Meditation the same trophotropic scale of hypoarousal, even though attention to the details of Although widely used, the term “medita- many Buddhist practices, including Zazen tion” is often employed in a highly impre- (Austin, 1998), makes a description in terms cise sense such that its descriptive power is of hypoarousal extremely problematic. greatly decreased. One underlying reason for An alternative approach to research on the term’s inadequacy is that, in its typical meditation is to attend more closely to the usage, it refers generically to an extremely particularity of the individual practices in P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 501

question. An apt metaphor in this case Sorting Claims and Descriptions might be the interaction between traditional medical systems and researchers seek- In emphasizing the particularity of each tra- ing to develop new pharmaceuticals. In dition’s approach to meditation, one need their search for plants whose active ingre- not discount the possibility that highly dis- dients might yield effective new medica- parate traditions may have independently tions, some researchers have begun exam- developed techniques that lead to similar ining traditional medical systems in various and measurable outcomes.2 Nevertheless, it cultures in order to narrow their search seems best not to begin with an assump- based on traditional claims about the medic- tion about any such innate similarity in inal properties of local plants (Jayaraman, disparate meditative traditions. One reason 2003; Schuster, 2001). In that collaboration, for avoiding such assumptions is the issue attention to the particularity of the heal- of particularity above, but another reason ing tradition is crucial, for it is the local is that similarities among traditions tend knowledge about specific, local plants that to appear primarily in claims about the will aid the search for new medications. ultimate meaning or nature of the state Clearly, such a project would be gravely hin- attained (e.g., “pure consciousness”) or in dered if researchers were to assume that, for metaphysically charged phenomenological example, an Amazonian healer’s traditional descriptions (e.g., ineffability) that do not herbal lore would somehow amount to the lend themselves to easy measurement or same traditional knowledge about medicinal interpretation. herbs that one would hear from a Himalayan Because similarities among traditions healer. The value of consulting a specific tra- often rest on such issues, an emphasis on dition is precisely that – through accident or those similarities tends to exaggerate a prob- expertise – the tradition may have gleaned lem that all researchers on meditation must some valuable knowledge or developed face; namely, the need to discern which some practice that is not found elsewhere. parts of a traditional account of meditation This importance of particularity supports are useful in formulating a neuroscientific the need to preserve local traditions, but it research strategy, as opposed to parts of an also speaks to the need to heed their bound- account that are not suitable for that pur- aries. A common problem with the literature pose. The problem here is the need to inter- on meditation is a tendency to ignore those pret traditional discourse about meditation, boundaries in order to emphasize some especially in terms of meditative techniques vague universality in human experience. and resultant states. In short, traditional acc- This attention to the particularity of con- ounts often describe techniques and resul- templative traditions is related to another tant states that are measurable and repeat- aspect of the approach we adopt; namely, able; nevertheless, parts of the same account that it is also strongly consistent with our may also focus on issues that can neither be knowledge of the neurosciences. Specifi- measured nor repeated. In many traditions, cally, cognitive and affective neuroscience the distinction between these parts of an has matured over the past decade, and we account reflects a tension between (1) close now understand something about the brain descriptions of meditative techniques and mechanisms that subserve different atten- states and (2) the metaphysical or soteriolog- tional and affective processes. Meditation ical requirements that must be met by those techniques that target specific underlying states, often expressed in textual sources that processes are thus likely to engage different the tradition considers inviolable. neural circuitry. If, however, the particular- Let us take as an example the Tibetan ity of a tradition’s claims and practices are practice of “Open Presence,” which we dis- not examined, the possibility that a practice cuss further below. In describing Open Pres- targets a specific process will not be noted. ence, traditional authors, such as Wangchug P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

502 the cambridge handbook of consciousness

Dorje(´ 1989) and Thrangu (Thrangu & John- dhism, most traditions use a term for med- son, 2004), offer typically detailed descrip- itation that correlates with the Sanskrit tions both of the techniques that induce that term bh¯avan¯a, literally, “causing to become.” state and also of the experiences that should In Tibetan traditions, the usual translation occur when the techniques are applied for bh¯avan¯a is gôm (sgom), which roughly properly.3 For example, discursive tech- means “to become habituated to” or “to niques for de-emphasizing the objectifica- become familiar with.” The meditative tration of sensory content are described in ditions of Tibetan Buddhism often employ detail, and in terms of resultant states, the the term in a generic fashion, and as a consequent loss of a sense of subject-object result, it is often translated into English with duality is also articulated clearly. These parts the equally generic term “meditation.” The of the traditional account lend themselves generic usage of gôm or “meditation” reflects to investigation, inasmuch as they describe its application to a remarkably wide range techniques and results for which neural of contemplative practices: For example, the correlates may be plausibly postulated and visualization of a deity, the recitation of a tested. At the same time, however, Buddhist mantra, the visualization of “energy” flowing philosophical concerns also demand that the in the body, the focusing of attention on the state of open presence reflects the onto- breath, the analytical review of arguments logical foundation of all reality, and Bud- or narratives, and various forms of object- dhist notions of nirvan¯ . a also require that the less meditations would all be counted as realization of that state will lead the adept “meditation.” to attain inconceivable physical and mental Nevertheless, despite this variety, it is powers. Such claims often occur in texts that possible to identify some relevant features traditional scholars are obliged to defend common to the traditional descriptions of under all circumstances. From a neuroscien- these Buddhist practices, especially when tific perspective, however, these claims do one separates those descriptions from meta- not lend themselves readily to analysis or physical arguments or exigencies that stem description. Thus, from the vantage point of from defending a textual tradition. First, it the researcher who stands outside the tra- is assumed that each such practice induces dition, it is crucial to separate the highly a predictable and distinctive state (or set detailed and verifiable aspects of traditional of states) whose occurrence is clearly indi- knowledge about meditation from the tran- cated by certain cognitive or physical fea- scendental claims that form the metaphysi- tures or events phenomenally observable to cal or theological context of that knowledge. the practitioner. Second, the state induced is said to have a predictable effect on both mind and body in such a way that, by Meditation as Explanandum inducing that state repeatedly, a practitioner Attention to the particularity of each tra- can allegedly use it to enhance desirable dition and the careful examination of tra- traits and inhibit undesirable ones. Third, ditional knowledge about meditation both the practices are gradual in the sense that contribute to a main concern of this chapter: the ability to induce the intended state is the notion of meditation as an explanan- supposed to improve over time, such that an dum. Or, to put the issue another way, experienced practitioner should meditate in how does one define “meditation” in the a manner that is superior to a novice. From context of neuroscientific study? This ques- the traditional standpoint, this improvement tion is not answered easily in part because is marked especially by two phenomenally of the extremely wide variety of human reportable features: the acquisition of cer- activities to which the term “meditation” tain traits (cognitive, emotional, or physical) might be applied. And the situation may and/or the occurrence of certain events (cog- not be much improved even if one focuses nitive, emotional, or physical). Finally, the on just one tradition. In the case of Bud- practice used to induce the state must be P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 503

learned, usually from a meditation teacher into a highly detailed scholastic tradition who is said to be a virtuoso in the practice. known in Sanskrit as the Abhidharma – That teacher will also serve as a guide to a type of Buddhist “psychology” that also the practice so as to assist the practitioner includes discussions of epistemology, philos- in improving his or her ability to produce ophy of language, the composition of the the state. material world, and cosmology.4 Based on these features, these diverse Despite the variety of Buddhist traditions, forms of Buddhist meditation may be taken they share two axioms articulated in Abhid- as explananda in regard to three general harma texts: A central goal of Buddhist prac- issues: (1) the claimed production of a distin- tice is the elimination of suffering, and any ctive and reproducible state that is phenom- effective method to eliminate suffering must enally reportable, (2) the claimed relation- involve changes in one’s cognitive and emo- ship between that state and the development tional states, because the root cause of suffer- of specific traits, and (3) the claimed pro- ing is a set of correctable defects that affect gression in the practice from the novice to all the mental states of an untrained person the virtuoso. Although initially formulated (Gethin, 1998). Thus, any practice that is in terms of Tibetan practices, these features considered by the tradition to be an effec- seem to be a useful way of understanding tive method must involve the features noted how meditative practices in most contem- above, including some set of reliable tech- plative traditions may be construed as neuro- niques that induce mental states that will scientific explananda. induce the desired changes in behavioral and psychological traits. In this regard, the Bud- dhist contemplative traditions exhibit con- A Paradigmatic Framework: Buddhist siderable diversity, because they hold diver- Meditative Techniques gent opinions about the precise nature of Our use of Buddhist contemplative tradi- the defects to be eliminated, the traits to be tions to develop a theoretical framework induced, and the best methods for accom- for understanding meditation is not merely plishing all this. At the same time, both the a product of historical accident; rather, diversity and the continuity of Buddhist con- Buddhist contemplative traditions are par- templative practices also stem from the rich ticularly well suited to the development of cultural context in which Buddhism initially this kind of theoretical model. The rea- flourished. son, in brief, is that unlike many contemplative traditions, Buddhist traditions tend early history and basic forms to offer extensive, precisely descriptive, and When the historical Buddha S´ akyamuni¯ first highly detailed theories about their prac- set out on the religious life (ca. 500 bce), tices in a manner that lends itself readily he apparently encountered a large number to appropriation into a neuroscientific con- of meditative techniques that were already text. This emphasis on descriptive preci- being practiced by various contemplative sion stems from the central role that various traditions in South Asia. Although histor- forms of meditation play in Buddhist prac- ical sources from this period are generally tice. That is, from the standpoint of nearly vague in their descriptions of contempla- every Buddhist tradition, some type of med- tive practices, one can identify some com- itative technique must be employed if one mon trends. Broadly speaking, these tra- is to advance significantly on the Buddhist ditions maintained that the contemplative spiritual path, and because Buddhism ini- life should be focused on the search for tially developed in a cultural context where one’s true self (often called the atman¯ ), and a wide range of such techniques were avail- because this true self was generally assumed able, Buddhist theoreticians recognized the to be somehow obscured by one’s involve- need to specify exactly the preferred tech- ment in the world of the senses, many con- niques. Their analyses eventually develop templative techniques involved an inward P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

504 the cambridge handbook of consciousness

focus whereby one’s mind was retracted lated as both “mindfulness” and “awareness”; from the senses. In addition to this inward in simple terms, it is the mental function focus, most techniques from this period (caitt¯asika) that focuses the mind on an probably sought to reduce the occurrence object. At the same time, the meditation of other types of mental content – generi- involves a faculty that checks to see whether cally called “conceptuality” (kalpan¯a) – that the smr.ti is focused on the intended object were also thought to obscure one’s vision or whether it has lost the object. Thus, of the true self. Distractions caused by the this other faculty, often called samprajanya, fluctuation of the mind were commonly involves a type of meta-awareness that is thought to be linked to the fluctuation of not focused on an object per se, but rather the breath, and meditative techniques there- is an awareness of that intentional rela- fore often involved either breath control tion itself (Gethin, 1998; Silananda, 1990; (pr¯an. ay¯¯ ama) or at least some attention to Wallace, 1999). the disposition of the breath. And because Both as a state and as a style of prac- the mind was thought to be strongly influ- tice, samatha´ provides the practical and enced by the body, contemplative practices theoretical underpinnings of many other involved specific postures or corporeal exer- Buddhist practices, especially because it con- cises (Bronkhorst, 1986; Gethin, 1998). stitutes the basic paradigm for any prac- When these practices were appropriated tice that involves one-pointed concentra- by the historical Buddha S´ akyamuni,¯ their tion (ek¯agrat¯a) on a specific object. At the overall context was altered, inasmuch as same time, however, Buddhist theorists who the Buddha maintained that the belief in discuss samatha´ generally do not consider a “true self” (¯atman) was completely mis- it to be in and of itself Buddhist. That is, taken. Indeed, from the earliest days a cen- practices oriented toward attaining samatha´ tral goal of Buddhist contemplative practice must create a highly developed ability to sus- is precisely to demonstrate to the practi- tain intense focus on an object, and whereas tioner that no such fixed or absolute iden- the development of that ability does lead to tity could ever be possible (Gethin, 1998). some trait changes, it does not lead to all of Nevertheless, although the Buddha altered the changes that Buddhists seek, most espe- the context of the contemplative practices cially in regard to the regulation of emotions. that he encountered, the Buddhist medi- Hence, although a samatha´ -oriented prac- tative techniques that he and his followers tice may be a necessary ingredient of most developed retained some of the same basic Buddhist contemplative traditions, it must principles of inward focus, reduction of con- be accompanied by another fundamental ceptuality, the importance of the breath, and style of Buddhist practice; namely vipa´syan¯a the relevance of the body. or “insight.” (Gethin, 1998; Silananda, 1990; Perhaps the most ubiquitous style of Bud- Wallace, 1999). dhist meditation that exhibits these features As with the samatha´ style of practice, is meditation aimed at improving concentra- vipa´syan¯a is also one of the earliest and tion – a style of meditation that is rooted in most fundamental forms of meditation. For practices aimed at obtaining samatha´ . Trans- Buddhist theorists, vipa´syan¯a is a style of latable literally as “quiescence,” samatha´ is meditation that, in combination with the a state in which the practitioner is able focus or stability provided by cultivating to maintain focus on an object for a the- samatha´ , enables the practitioner to gain oretically unlimited period of time. As a insight into one’s habits and assumptions term, samatha´ therefore can also describe about identity and emotions. In general, one of the historically earliest and most basic this insight includes especially the real- styles of Buddhist meditation that aims at ization of “selflessness” (nair¯atmya) – that attaining that state. In such a practice, the is, realizing that one’s belief in a fixed, practitioner augments especially a mental essential identity is mistaken and hence faculty known as smr.ti, confusingly trans- that the emotional habits that reflect that P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 505

belief are baseless (Dalai Lama XIV, 1995; titioner. This basic technique of recitation Gethin, 1998; Silananda, 1990). Neverthe- and visualization is representative of a wide less, although every Buddhist contemplative range of similar Buddhist practices that tradition would agree that such a realiza- evolved during the first millennium. Chief tion must be part of vipa´syan¯a, one again among these is the practice of visualiz- encounters considerable diversity in the pre- ing deities and paradisiacal environments, cise way in which vipa´syan¯a is defined and a technique especially important in most the way it is developed in practice. For exam- forms of Buddhist tantra (Beyer, 1977). ple, in some traditions reasoning and a type Alongside Recollection (and later, visu- of internal conceptual discourse are critical alization) practices, Lovingkindness medita- to the practice, but other traditions maintain tion was also a widespread practice in both that reason and concepts are of only limited early and later Buddhism, where it is thema- use in obtaining vipa´syan¯a. Likewise, some tized as the cultivation of “great compassion” traditions maintain that a vipa´syan¯a medi- (mah¯akarun. a)¯ . The practice aims to culti- tation must have an object toward which vate an emotional state; namely, a sense of some type of analysis is brought to bear love and compassion toward all living things. (Dalai Lama XIV, 1995; Silananda, 1990), Representative of a wide range of practices whereas others maintain that the meditation that promote or inhibit traits by repeatedly must eventually become completely object- inducing a particular emotional state, some less (Wangchug Dorje,´ 1989). Perhaps the forms of the practice involve the recita- sole theme that runs throughout all Buddhist tion/visualization techniques employed in traditions is that, in vipa´syan¯a meditation, Recollection meditation. Some discursive the type of meta-awareness mentioned ear- strategies, such as thinking through the steps lier plays an especially important role – an of an argument for compassion, may also issue that we examine in the section on the be employed (Dalai Lama XIV, 1991;P. theory of meditation. Williams, 1989). Last to develop (toward the end of the further historical developments first millennium in India) are a variety of Although the basic combination of samatha´ practices that may be called tantric Wind and vipa´syan¯a provides both a theoretical meditations. These practices aim to manip- and historical touchstone for the develop- ulate the various forms of energy, metaphor- ment of Buddhist contemplative practices, ically called Wind (v¯ayu), that are alleged a number of other forms of meditation to flow in channels throughout the body. were developed in the various Buddhist This model is roughly analogous to the con- communities of Asia. Three practices ini- temporary understanding of the nervous sys- tially developed in India are especially tem, where the notion of Wind is analo- emblematic of the range of developments: gous to the propagation of neural impulses. “Recollection of the Buddha” meditations In the Buddhist model, the mind itself is (buddh¯anusmr.ti), Lovingkindness medita- thought to consist of such Wind energy, tion (maitr¯ıbh¯avan¯a), and tantric “Wind” and practices that manipulated that energy (v¯ayu) meditations. were therefore intended to induce or inhibit The practice of Recollection of the Bud- mental states or traits. The many techniques dha is probably, along with Lovingkindness employed include the visualization of vari- meditation, one of the oldest Buddhist prac- ous syllables or other items at specific points tices. Recollection involves the recitation of in the body as a means to alter the flow of the Buddha’s attributes, and in its earliest mental energy; physical exercises, including form it may have involved nothing more breathing exercises; and an array of other than that. At some point, however, the techniques, including manipulation of the recitation of the Buddha’s physical attributes diet. An example of this style of practice was linked with the visualization of the that later becomes important in Tibet is the Buddha in the space in front of the prac- “Tummo” (gtum mo) practice – a method P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

506 the cambridge handbook of consciousness

that, in manipulating the Wind, is also said to stood to describe two aspects of the same med- generate considerable body heat as a byprod- itative state. In the interest of both applica- uct (Cozort, 1986; Dalai Lama XIV, 1995; bility and simplicity, we derive our account English, 2002; Snellgrove, 2002). primarily from a specific and living contem- As Buddhism spread from its initial loca- plative tradition: Tibetan Buddhism. tion in the Gangetic plain, Buddhist practi- Drawing on a maxim developed by tioners developed and enhanced the above their Indian predecessors, Tibetan theorists practices, along with many other related maintain that the highest forms of Bud- forms of meditation. Eventually, Buddhism dhist meditation must integrate the quali- spread to other regions of Asia, and vari- ties of samatha´ and vipa´syan¯a into a sin- ous traditions arose that persist to this day. gle practice. As described by the most Each tradition elaborated its own particular common traditional metaphor, the prac- interpretation of techniques that, although titioner cannot make significant spiritual likely inherited from Indian Buddhist tradi- advancement without the combination of tions, always acquired a local flavor. Never- samatha´ and vipa´syan¯a, just as a cart cannot theless, most extant practices reflect the var- move without two wheels. Another tradi- ious styles of meditation noted above. tional metaphor is perhaps more descriptive: When attempting to see the murals on analysis of meditation: samatha´ and the wall of a dark cave, one must use a vipasyan´ a¯ as paradigm lamp that is both well shielded and bright. To aid in the mastery of meditative tech- If the lamp is not well shielded, then its niques – and also to respond to critics out- flame will flicker or even become extin- side their traditions – Buddhist theoreticians guished, and if its flame is not sufficiently in India and elsewhere developed detailed intense, the lamp’s light will be insufficient accounts of their contemplative practices. for the task at hand (Tsongkhapa, 2002). These accounts are often extremely com- This very basic metaphor describes the qual- plex, and as noted above, they sometimes ities that are indicated by the terms samatha´ raise metaphysical issues that are not eas- and vipa´syan¯a: The former primarily con- ily addressed in neuroscience. Likewise, to cerns the stability (gnas cha) of the medita- some degree the accounts are shaped by the tive state, whereas the latter concerns that need to defend a particular textual tradition state’s phenomenal or subjective intensity or line of argumentation, and as a result, (gsal cha) (Thrangu & Johnson, 2004). some statements that seem to be descrip- To state these features more precisely, in tive are not known to be exemplified by any meditations that involve an object, stability actual Buddhist practice. Nevertheless, other refers to the degree to which the practitioner aspects of the accounts seem more empiri- is able to retain focus on the object with- cal in their approach, and attention to those out interruption. In such meditations, clar- aspects may prove useful when examining ity refers to the sharpness or vividness of the meditation in a laboratory context. With appearance of the object in awareness. For this in mind, we have sketched the follow- example, in the visualization of a colored ing practical and simplified account of Bud- disc, a completely stable meditation would dhist meditation theory, aimed especially at be one in which the meditator’s focus on the researchers interested in studying contem- object is not perturbed at all by other phe- plative practices in Buddhism and other tra- nomenal events, such as emotions, thoughts, ditions. As one might expect, the numerous or sensory perceptions. In such a meditation, forms of Buddhist meditation are accompa- the clarity would be constituted by the disc’s nied by an equally wide range of theoreti- vividness of color and sharpness of shape. cal accounts. Nevertheless, the central issues Generally, these two aspects of a medi- can be addressed in terms of the theories that tative state are understood to work some- undergird samatha´ and vipa´syan¯a, especially what at odds with each other in the case when these styles of meditation are under- of novice meditators. That is, in the case P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 507

of a novice, the greater the stability of the worth noting that, just as the tradition con- meditative state, the more likely is it to lack tains techniques to ease mental or physical intensity. And the greater its intensity, the tension, it also espouses methods to coun- more likely is its lack of stability. This tension teract an excess of relaxation or dullness between stability and clarity is expressed in (Thrangu & Johnson, 2004; Tsongkhapa, the two main flaws that hinder a meditation: 2002; Wangchug Dorje,´ 1989). “dullness” (Tib., bying ba) and “excitement” Although the balance of clarity and sta- (Tib., rgod pa). When dullness first arises, the bility as described above forms an overall focus on the object will be retained, but as paradigm for Tibetan Buddhist practices, it dullness progresses, the clarity of the object is important to recognize the ways in which becomes progressively hindered, and a sen- that paradigm is modified for each practice. sation of drowsiness overtakes the medita- For example, novices hoping to develop the tor. If dullness continues, the dimness of the meditative state of Rigpa Chôgzhag or “Open object will cause the meditator to lose focus Presence” may be taught to emphasize one on it, or in the case of gross dullness, the or another feature in order to make ini- meditator will simply fall asleep. In contrast, tial headway in the practice. In short, they when excitement occurs, the clarity of the are encouraged to err on the side of clar- object will often increase, but the intensity ity, because it is more important to avoid of the mental state perturbs the meditation dullness than excitement in the early stages such that distractions easily arise and focus of that practice (Thrangu & Johnson, 1984; on the object is lost (Thrangu & Johnson, Wangchug Dorje,´ 1989). 2004; Tsongkhapa, 2002; Wangchug Dorje,´ The practice of Open Presence raises 1989). another issue: the applicability of this model In most practices, the ideal meditative to meditations that do not focus on an state – one beyond the novice stage – is a object. In objectless practice, the loss of state in which neither dullness nor excite- focus on the object or its degree of phenom- ment occurs; in short, stability and clarity are enal vividness obviously cannot be taken as balanced perfectly. Hence, for the Tibetan criteria for the degree of stability or clarity. contemplative traditions (and indeed, for Instead, stability becomes a marker for the nearly every other Buddhist tradition), it ease and frequency with which the medita- would be incorrect to interpret Buddhist tor is perturbed out of the state the medita- meditation as “relaxation.” This is not to tion is intended to induce, and clarity refers deny the importance of mental and phys- to the subjective intensity of that induced ical techniques that help the practitioner state. Thus, after a session of Open Presence, relax. Without such techniques, an excess a meditator who reports that the medita- of physical or mental tension may develop, tion was unstable but very clear would mean and when such tension occurs, excitement that, although the intended state was inter- will almost certainly arise. If, however, such rupted repeatedly, the subjective experience relaxation techniques are overused, they of the state was especially intense when it are likely to propel the practitioner into occurred. dullness and hence hinder the meditation. A final aspect of the basic theory of med- Indeed, from a Buddhist perspective a prac- itation concerns the distinction between the tice that only relaxes the mind might even- actual meditative state (Tib., dngos gzhi) and tually prove harmful. That is, such a practice the post-meditative state (Tib., rjes thob). would develop a great deal of dullness, and In brief, the states developed in medita- as a result the practitioner might become tion are usually thought to create a post- withdrawn, physically inactive, and mentally meditative effect. In some cases, some phe- depressed. Overall, then, Buddhist medi- nomenal aspect of the meditation persists tations avoid an excess of relaxation, and in the post-meditative state. For example, it is for this reason that very few prac- after a meditation in which one cultivates tices are done while lying down. It is also the experience of phenomenal content as P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

508 the cambridge handbook of consciousness

seeming dreamlike, one’s perceptions in the larized set of meditation instructions avail- post-meditative state are also said to have able to a wide population that is not lim- a dreamlike quality for at least some period ited to celibate monastics (Coleman, 2002). after arising out of meditation. In other cases, In its most typical form, the early stage of the post-meditative state involves a trait Vipassana¯ practice consists largely of a basic change. Meditation on love and compassion, samatha´ style of meditation focused on the for example, is alleged to inhibit the occur- sensation made by the breath as it flows in rence of anger between meditative sessions. and out of the nostrils, although sometimes From the Buddhist theoretical perspective, another aspect of the breath may be taken such post-meditative changes are often at as the object of meditation (Gunaratana, least as important as the states induced dur- 2002). In the early stages, the aim of the ing the meditation itself, and success in a meditation is to keep the attention focused practice is often measured by the strength of on the breath without distraction – that is, the effects that occur after meditation (Dalai without the attention wandering to some Lama XIV, 1995; Thrangu & Johnson, 1984; other object, such as a sensation or a mem- Tsongkhapa, 2002; Wangchug Dorje,´ 1989). ory. For beginners (and even for advanced practitioners), the attention inevitably wan- contemporary practice and ders, and the usual instruction is to recog- problems of terminology nize that the mind has wandered – for exam- In the laboratory setting neuroscientific ple, to see that it is now focused on the pain researchers are likely to encounter Bud- in one’s knee, rather than on one’s breath – dhists who engage in contemplative prac- and then to “drop” or “release” the distrac- tices located in three overall traditions: the tion (the knee pain) and return to the breath. Vipassana¯ or Insight Meditation movement Part of the aim is not only to develop focused located within Theravada¯ Buddhism, the attention on the breath but also to develop Zen tradition of Japan, and the Tibetan tra- two other faculties: a meta-awareness that dition. One might encounter practices from recognizes when one’s attention is no longer other Buddhist traditions, but the medi- on the breath and an ability to redirect tations of the aforementioned three tradi- the attention without allowing the meta- tions are by far the most widespread. They awareness to become a new source of distrac- are also the most likely to be practiced tion, as when one berates oneself for allow- by persons, such as Europeans and North ing the mind to wander (Gunaratana, 2002; Americans, who are not native to the cul- Kabat-Zinn, 2005). tures in which the practices have devel- Given the description thus far, practices oped (Coleman, 2002). Of these three tra- very similar to Vipassana¯ meditation are also ditions, the style of meditation taught in the found among contemporary practitioners of Vipassana¯ traditions is especially emblem- Zen and Tibetan Buddhism. Indeed, it is pos- atic, because the basic meditative style of sible that the Vipassana¯ approach to medi- Vipassana¯ closely resembles some founda- tation on the breath has led Zen and Tibetan tional practices in the Zen and Tibetan practitioners to employ a similar style of traditions. breath-meditation to a much greater extent The Vipassana¯ or Insight Meditation than they have in the past. Certainly, it is movement consists of several loosely allied clear that in contemporary Zen and Tibetan institutions and individuals that teach a practice, focusing the attention on the breath style of meditation rooted in the older (or sometimes another static object) is often contemplative traditions of Theravada¯ Bud- used as a means to develop the basic dhism in Myanmar, Thailand, and Sr´ ¯ıLanka.˙ level of concentration required for more Although it draws on older traditions, the advanced forms of meditation. In many Vipassana¯ movement is “modern” in that cases, these more advanced meditations aim it makes a somewhat simplified and regu- to enhance the type of meta-awareness that P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 509

is cultivated during the Vipassana¯ style of selves used as a means to thoroughly develop practice, and because the traditions have meta-awareness (samprajanya or praj˜n¯a) at different ways of understanding and enhanc- a later stage in the practice (Gunaratana, ing that meta-awareness, all three kinds of 2002). It is therefore not surprising that traditions – Vipassana,¯ Zen, and Tibetan smr.ti becomes closely associated with the Buddhism – diverge in their practices from meta-awareness, but this imprecise use of this point forward. smr.ti has contributed to the confusion con- Although Vipassana¯ meditation may be cerning the English terms “mindfulness” and especially representative of a widespread “awareness.” and foundational style of practice in contem- To restate the problem using the terms porary Buddhism, any discussion of Vipas- discussed earlier, we should note that such sana¯ meditation must address a problem authors as Thrangu (Thrangu & Johnson, of terminology: the often confusing use of 2004) employ the term “mindfulness” to the terms “mindfulness” and “awareness.” refer to the samatha´ aspect of a meditative In the Mindfulness-Based Stress Reduction practice – that is, the stability of the medita- (MBSR) designed by Jon Kabat-Zinn (2005), tion. And these authors then use “awareness” for example, the term “mindfulness” is used for the vipa´syan¯a aspect of the practice; that primarily to refer not to the focusing aspect is, the meta-awareness that is especially asso- of mind, but rather to the meta-awareness ciated with the clarity of the meditation. that surveys that focus and its relation to Turning then to its usage in, for example, the intended object. Likewise, in MBSR MBSR, one ﬁnds that mindfulness refers pri- the term “awareness” sometimes seems to marily to the vipa´syan¯a aspect of the prac- stand primarily for attention or the focusing tice, not the samatha´ aspect. aspect of mind. In contrast, popular works Although this problem is simply one of on Tibetan Buddhist meditation, such as terminology, it can prove quite confusing in a the work of Thrangu (Thrangu & Johnson, laboratory setting. In the case of Open Pres- 2004), use these same two terms, but their ence practice, Tibetan meditators will usu- meaning is reversed: “mindfulness” refers ally deny that their practice is of mindfulness to attention or focus, whereas “awareness” (i.e., dran pa’i nyer gzhag), whereas in fact, refers to a faculty of mind that surveys the they mean to say that they are emphasizing mental state at a meta-level. the development of some meta-awareness in The confusion in English terminology a way that has many parallels with the mind- is in part due to some confusion in the fulness practice of Vipassana¯ meditators or proper usage of the Buddhist technical terms persons trained in MBSR. For the researcher, themselves. Strictly speaking, smr.ti – liter- one solution to this problem is again to be ally, “memory” – is the focusing aspect of attentive to the particularities of the practice mind, and historically it is often translated in question while keeping track of the fact as mindfulness when used in the context that we have yet to standardize the English of meditation. An obvious case is the com- lexicon of technical terms for the analysis of mon technical term smr.tyupasth¯ana (in Pali,¯ meditation. satipat..th¯ana), usually rendered as “founda- Another problem of terminology comes tion of mindfulness.” The problem, how- with the use of the term samatha´ itself, espe- ever, is that even though smr.ti should stand cially in its Tibetan context. Our discus- only for the focusing aspect of mind, in both sion thus far has used the term samatha´ in popular and technical Buddhist literature on three basic meanings: (1) a particular state in meditation it is not infrequently assimilated which one can allegedly focus on an object to the meta-awareness mentioned above. for an unlimited period of time, (2) a style One reason for this is that in the Vipassana¯ of practice aimed at attaining that state, and tradition, meditations that initially empha- (3) the aspect of any meditative state that size smr.ti as the focusing faculty are them- constitutes its maximal stability. Already, P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

510 the cambridge handbook of consciousness

these three meanings can lead to consid- tioner is doing during meditation, it may erable ambiguity, but the second meaning be more useful to use other Buddhist terms is particularly troublesome. One problem is as labels for a particular practice – or per- simply that the expression “samatha´ medi- haps researchers will develop new terms in tation” is not sufficiently clear. That is, when dialogue with practitioners. Otherwise, if Buddhist theorists are being precise, they one is not careful, one may be misled into recognize that samatha´ meditation should believing that a wide set of disparate prac- be rendered more properly as “meditation tices are the same because, for one reason aimed at obtaining samatha´ .” or another, they may all be called “samatha´ But even after this clarification, a problem meditation.” remains: When one practices samatha´ meditation, which kind of samatha´ is one trying Three Meditative States in to obtain? In other words, is one attempt- Tibetan Buddhism ing to cultivate the ability to concentrate on an object for an unlimited period? Or is one We have mentioned repeatedly the impor- trying to cultivate some other kind of maxi- tance of attending to the particularity of mal stability? The main problem here is that, the contemplative tradition whose practices in the Tibetan context, the term samatha´ is might become part of a research agenda, and used to refer to stability in meditations that with this in mind, we now discuss briefly do not even have an object; hence, in those three specific forms of meditation found in cases samatha´ cannot relate to concentra- Tibetan Buddhism. Part of our aim is to set tion on an object. To make the matter even the ground for a discussion of these prac- more complicated, there are Tibetan practices, because they have already been the tices in which samatha´ is used in connection subjects of some preliminary neuroscientific with both kinds of meditations (i.e., with research, as is presented below. an object and without an object). Finally, All three styles of meditation come from even when samatha´ is related to practices a particular strand of contemplative prac- with an object, the object in question may tice in contemporary Tibetan Buddhism. differ considerably; an example with neuro- In Tibetan, a term for this style of prac- scientific import is the difference between tice is Chag-zôg (phyag rdzogs), a compound focus on the breath and focus on a visualized that refers to two traditions of meditation: object. Indeed, traditional scholars, such as the “Great Seal” or Chag-chen (phyag chen; Thrangu (Thrangu & Johnson, 2004,p.21), Skt. Mah¯amudr¯a) of the Kargyu¨ (bka’ with all these issues in mind, caution their brgyud) school and the “Great Perfection” students about the potential for confusion or Dzôg-chen (Karma Chagme,´ 2000)ofthe caused by the ambiguity of technical terms Nying-ma (Rnying ma) or “Ancient” school. such as samatha´ . Although historically and institutionally dis- As with the case of mindfulness and tinct, for the last 200 years the Great Seal awareness, the problems with the term and Great Perfection traditions have become “samatha´ ” should remind researchers that allied so closely that it is now exceedingly the particularities of a practice may rare to find a practitioner who employs the be obscured by ambiguous terminology, techniques of only one style in complete whether in the source language or in English isolation from the techniques of the other translation. As a practical matter, one may style. This is not to say, however, that there even wish to avoid the term samatha´ as are no important differences between these a description of a practice. This is not to two traditions. They differ especially in the say that the term should be abandoned: details of their most advanced practices, and Clearly, for both historical and theoreti- they also propose slightly different tech- cal reasons, samatha´ must remain in the niques for the three meditations discussed lexicon on Buddhist meditation. But when below. Nevertheless, for the purposes of the seeking to specify exactly what a practi- brief descriptions below, they may be treated P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 511

as constituting a single overall style of con- ment in developing the ability to concentrate templative practice. on an object (Thrangu & Johnson, 2004; The three practices in question are Tsé- Tsongkhapa, 2002; Wangchug Dorje,´ 1989). cig Ting-ngé-dzin (rtse gcig ting nges ‘dzin) One aspect of setting is the context formed or Focused Attention, Rig-pa Chôg-zhag (rig by the other practices in which a medita- pa cog bzhag) or Open Presence, and Mˆıg- tor is engaged. These practices include espe- mé Nying-jé (dmigs med snying rje) or Non- cially formal guidance received from one’s Referential Compassion. All meditators in preceptor, the study of Buddhist thought, the Chag-zôg style receive at least some a wide range of devotional practices, and instruction in all three of these practices, and the observance of a basic moral code based all advanced meditators will be thoroughly upon non-harm (ahim. s¯a) and compassion. familiar with them. Another aspect of setting concerns the site where one is to meditate. In this regard, tra- focused attention (tse-cig´ ditional accounts speak at length about the ting-nge-dzin´ ) need for a quiet place with few distractions The Tibetan term Tsé-cig Ting-ngé-dzin or and adequate access to food and water. So “Focused Attention” refers to a mental state too, the spot to be used for meditation is in which the mind is focused unwaveringly prepared by the meditator on a daily basis and clearly on a single object. This state, by cleaning it and preparing it through vari- which literally translates as “one-pointed ous ritual activities. concentration,” occurs in many practices, Once preparations for the session are and it is a typical goal for novices in the complete, the meditator adopts the posture Chag-zôg traditions. The relevant Buddhist for meditation. Various styles of Tibetan theories and techniques are usually drawn meditation involve different postures, but in from a generic account of practices that the context of developing Focused Atten- seek to develop samatha´ in the sense of tion, the general rule is that the spine the ability to focus on an object for an must be kept straight and that the rest unlimited time. This generic account, some- of the body must be neither too tense times called “common samatha´ ” (thun mong nor too lax (Thrangu & Johnson, 1984; gi zhi gnas), differs from the present con- Tsongkhapa, 2002; Wangchug Dorje,´ 1989). text. That is, in actual practice, Chag-zôg At this point, another element of the practitioners of Focused Attention usually setting – the use of memorized formulas develop a lesser (and often unspecified) state to induce the proper conceptual attitude – of concentration before being instructed by is invoked, and depending on the practice their teachers to move on to other prac- in question, a number of other practices or tices, which no longer involve focusing on ritual activities may precede the portion of an object (Thrangu & Johnson, 2004). Nev- the practice in which one seeks to develop ertheless, the practice of cultivating Focused Focused Attention. Attention draws heavily on the theories and When actually engaged in the practice techniques of common samatha´ . Perhaps of Focused Attention, the meditator focuses the most important principles drawn from the mind on the object to be meditated that generic account can be summarized upon. This object may be a sensory object, under six overall issues: the setting, the body such as a visible object in front of the med- posture, the object, the flaws that hinder itator, or it might be mental, such as a visu- progress, the “antidotes” to the flaws, and alized image (Thrangu & Johnson, 2004; the stages of development in meditation Wangchug Dorje,´ 1989). In general, Tibetan (Tsongkhapa, 2002). practitioners do not use the breath as an Although sometimes neglected in schol- object of meditation, except perhaps for rel- arly work on Buddhist meditation, the set- atively brief periods as a means to settle the ting for meditation is clearly considered by mind (Thrangu & Johnson, 2004). This trend traditional authors to be an important ele- may be changing, however, in part because P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

512 the cambridge handbook of consciousness

of the modern encounter with other Bud- counteract excitement and allow the medi- dhist traditions. To a great extent, the par- tator to return to the original object of med- ticular object chosen depends largely on the itation (Wangchug Dorje,´ 1989). particular practice (such as tantric visual- In terms of dullness, methods to counter- ization or Open Presence) that forms the act it are often related to those that coun- overall context for the development of Focu- teract excitement. For example, just as one sed Attention. might counteract excitement by meditating Having placed the attention on the object, in a dimly lit room, one can counteract dull- the meditator then seeks to avoid two over- ness by meditating in a brightly lit setting. all flaws: dullness and excitement. As men- So too, adding tension to the body or inten- tioned above, in the early stages, these flaws sity to a visualized object can also coun- manifest in a straightforward fashion. Dull- teract dullness. And in terms of affective ness is detected by, for example, a dim- methods, the meditator might temporarily ming or blurring of the object and, in its contemplate joy or compassion (and some- most gross form, a sensation of drowsi- times even fear) so as to energize the mind ness. The main symptom of excitement is enough to return to the original object. As distraction (i.e., the intensity of the focus with excitement, visualizations may also be causes one to be hyperaroused, and as a employed; for example, one may visualize result, attention wanders to other mental a white dot on one’s forehead at the point content or phenomena; Thrangu & Johnson, between the eyes (Wangchug Dorje,´ 1989). 1984; Tsongkhapa, 2002; Wangchug Dorje,´ For advanced meditators, many of the 1989). “antidotes” mentioned here are too coarse, In practice, the usual technique to coun- and they would lead to an overcorrection in teract excitement is to become aware of the the meditation. For these practitioners, the occurrence of the distracting content or phe- subtle degree of dullness or excitement that nomenon – that is, one notes the fact that the they encounter is corrected by equally sub- mind is now attending to another object, and tle adjustments to the clarity (for dullness) then one returns the mind to the intended or the stability (for excitement) of the med- object without allowing the original distrac- itation state until both stability and clarity tion to produce more mental distractions, reach their maximal, balanced state. such as the thought, “It is not good to be dis- The notion that advanced meditators tracted” (Thrangu & Johnson, 2004). Some- employ different responses to flaws in med- times excitement is also caused by physical itation raises the final relevant issue in or environmental factors – too much ten- traditional accounts; namely, the theories sion in the body or too much bright light in about the progression of stages in medita- the meditation area, for example. Or, excite- tion. Many contemplative traditions speak ment may be caused by applying too much of ascending stages through which the prac- effort to the meditation (i.e., being too rigid titioner passes; a typical account speaks of in one’s focus). Similarly, in the case of visu- nine levels of progressively higher degrees alized objects, excitement might be caused of concentration along with corresponding by too much intensity or brightness in the changes in the meditator’s response to dull- visualized object. Sometimes an affective ness and excitement (Thrangu & Johnson, remedy is used; for example, the meditator 1984; Tsongkhapa, 2002; Wangchug Dorje,´ might temporarily switch to a contempla- 1989). This schema, however, is far more tion of suffering, and the affective impact complicated than it seems, and as Apple of that contemplation will reduce excite- (2003) demonstrates, the Buddhist pen- ment enough that one can return to the orig- chant for scholasticism makes this topic an inal object of meditation. Various visualiza- extremely complicated one when it is con- tions – such as visualizing a small black drop sidered in its fullest form. Without going behind the navel – may also be employed to into great detail, it is important to note that, P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 513

according to these schemas, a single prac- counteracts flaws by seeing the true nature of tice may progress gradually through a num- identity and objects. And because the prac- ber of meditative states, but some of those tice involves many discursive strategies that states might differ significantly from each are based upon an underlying theory, one other both phenomenally and in terms of the must have some sense of those theoretical appropriate technique to be applied. Like- underpinnings. Hence, even though our pre- wise, the mental and physical effects of a sentation aims to focus on empirical descrip- practice may build gradually; for example, tions of what practitioners actually do, a brief as one’s level of concentration improves, foray into more abstract theory is neces- mental and physical well-being is also said sary. Our theoretical discussion is based on to increase But some effects occur only at three authors who are typical of the Chag-zôg some stages and do not progress further traditions: Karma Chagme(´ 2000), Thrangu (Tsongkhapa, 2002). (Thrangu & Johnson, 2004), and Wangchug In terms of the most relevant effects that Dorje(´ 1989). Our concise and thematic pre- are traditionally expected to arise from this sentation, however, might not be satisfac- practice, the main result of Focused Atten- tory to a strict traditionalist, in part because tion is a greater ability to concentrate and the issues involved are notoriously difficult a concomitant decrease in susceptibility to to explain. Nevertheless, from an academic being perturbed out of a concentrated state. and anthropological standpoint, this presen- The practice is also thought to increase not tation should suffice to convey the main the- only the stability of one’s concentration but oretical issues relevant to this style of con- also its intensity. At the higher levels of prac- templation as it is practiced currently. tice, this type of meditation is also said to Theoretical Background. When justifying reduce the need for sleep, and during the and explaining the types of practices that meditation it is thought to induce pleasur- induce Open Presence, Chag-zôg theorists, able sensations, including a lightness or pli- such as Karma Chagme(´ 2000), Thrangu ancy of mind and body. (Thrangu & Johnson, 2004), and Wangchug Dorje(´ 1989), argue that, properly speaking, open presence (rig-pa chog-zhagˆ ) objects are only known through experience; Open presence or Rig-pa Chôg-zhag is one it is nonsensical to speak of objects sepa- of the main meditative states that practition- rate from experience. Likewise, experience ers following the Chag-zôg style of practice of an object necessarily involves a subject attempt to cultivate. The basic motivation that experiences the object, and it is there- for the practice is rooted in a Buddhist axiom fore nonsensical to speak of objects without mentioned earlier: Namely, that one’s neg- speaking of a subject. The theoretical linch- ative emotional habits and behaviors arise pin is that the nature of both objects and from a set of mental flaws that cause one subjects is that which characterizes them to consistently misconstrue both one’s iden- under any circumstances – it must be essen- tity and also the objects toward which those tial to them, rather than accidental. And emotions and behaviors are directed. As what is essential to them is that they always noted above, those flaws are meant to be occur within experience. Hence, to know the corrected by vipa´syan¯a meditation through nature of objects and subjects is to know the which one cultivates an accurate under- nature of experience (Thrangu & Johnson, standing of the nature of one’s identity and 2004). the nature of objects in the world. In this Whatever may be the philosophical mer- sense, Open Presence may be considered its of such an analysis, Chag-zôg practi- a particular version of vipa´syan¯a. Chag-zôg tioners are thus aiming to understand the theorists, however, have a unique under- nature of experience – that which is essen- standing of what it means to gain the under- tial to any instance of experience, regard- standing or “wisdom” (Tib. ye shes) that less of the accidental and changing features P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

514 the cambridge handbook of consciousness

of the objects or subjectivities involved. To the process or occurrence of that experi- do so, they employ a set of techniques ence either passively or involuntarily. For that are intended to make the practitioner Chag-zôg theorists, this faculty of appercep- aware of the invariable feature of all expe- tive presentation is a derivative form of the riences. They speak of this invariable fea- more fundamental Awareness that is the ture using various descriptions, including basic nature or structure of consciousness Rigpa, “Awareness,” or, using the metaphor itself. Hence, a meditative technique that of light, Selwa (gsal ba), “Luminosity” or removes the cognitive features that usually “Clarity.” But whether called Awareness, obscure the implicit reflexivity of experience Clarity, or some other synonym, the point is one that moves that practitioner closer is that the invariant element in experience is to an understanding of that fundamental that which, from a phenomenal standpoint, Awareness. makes it possible for the subject-object rela- Above we noted that even the earli- tion to be presented in experience. est forms of vipa´syan¯a meditation seem to As the alleged invariant in all states of involve some form of meta-awareness that knowing, Awareness contrasts with features surveys the meditative state in such a way that are accidental (i.e., not essential) to that it enables one to know whether one any given cognition; namely, the particu- has lost the focus on the object. Likewise, lar features of the object and subject occur- that same type of meta-awareness serves ring within the cognition. What is accidental to determine whether or not dullness and about the object are its characteristics, such excitement are occurring. Thus, even in as color or shape. And what is accidental these other forms of vipa´syan¯a practice, one about a subject is, for example, its temporal encounters a type of reflexivity, inasmuch location in the narrative of personal iden- as the meditative state is meant to involve tity or the particular emotional state that an awareness of the state itself. With this in is occurring with the subjectivity. Hence, a mind, one can think of the Chag-zôg prac- meditative technique that enables the practice of cultivating Open Presence as empha- titioner to know Awareness or Clarity must sizing this aspect of vipa´syan¯a practice to somehow avoid attending to the particular- its furthest possible point. This practice dif- ities of object and subject and grant access fers from other meditations, however, in that instead to the fact of knowing itself. The theoretically it is taking an implicit aspect of problem, according to Chag-zôg theorists, is all cognitions – a fundamental form of reflex- that untrained persons are deeply entangled ivity – and making it phenomenally accessi- in the accidental features of experience; gen- ble to the practitioner. erally, they focus especially on the features of On this theoretical understanding of the object, and occasionally they are explic- Open Presence, two features of the practice itly aware of themselves as subjects. But in are especially salient. First, in other medi- either case, untrained persons are not aware tations that fall under the general rubric of of what is invariant in those experiences. vipa´syan¯a, one cultivates a faculty of mind To overcome this problem, the various that is best described as a meta-awareness; lineages of contemplative practice that fall it is “meta” in that it is dependent upon the under the rubric of Chag-zôg propose dis- mindfulness (smr.ti) that is focusing the mind tinct techniques, but one common approach on the object at hand. As noted above, this is based upon a move toward subjectivity meta-awareness surveys the mind itself so as in meditation. The notion here is that the to determine, for example, whether it is dull invariant aspect of experience is closely tied or excited. Thus, inasmuch as it focuses on to the reflexive awareness (Tib. rang rig) that cognition itself, the meta-awareness is reflex- enables one to have memories of oneself as ive, and to this degree it resembles the type an experiencing subject. On this theory, as of state cultivated in the practice of Open an object is being presented to an experienc- Presence. The difference, however, is that ing subject, reflexive awareness also presents in Open Presence the prefix “meta” would P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 515

be inappropriate; instead, it is assumed that, rather then being attendant upon the basic faculty of mindfulness – i.e., the faculty that focuses on an object – the reflexive aspect of mind is actually more fundamental than mindfulness. In other words, mindfulness must occur with an object, but the possibility of objects being presented in experience is itself rooted in a more fundamental reflexivity. The second distinctive aspect of Open Presence is that, unlike other meditations, at advanced stages of the practice there is no attempt either to suppress or to cultivate any particular mental content. One does not fo- Figure 19.1. cus, for example, on a visualized image or on a sensory object, such as a sensation made by This diagram is based especially on the the breath. In this sense the state of Open styles of Chag-zôg practice exemplified by Presence is objectless. Nevertheless, even Karma Chagme(´ 2000), Thrangu (Thrangu though higher levels of the practice do not & Johnson, 2004), and Wangchug Dorje´ involve any particular content or object, it (1989, and their works are the main tex- also is important for content to be occurring tual sources for the presentation below. in the mind because to cultivate an aware- It is important to note, however, that ness of the invariant nature of experience, the diagram suggests a trajectory of actual one must be having experiences. Indeed, for practice that, although clearly implicit in beginners it is preferable that the experi- these authors’ writings, is not explicit. ences be especially striking or clear. Thus, Nevertheless, this way of presenting the even though the meditation is objectless, it flow of the practice has the advantage of is not a state of blankness or withdrawal. Sen- being far less complicated than traditional sory events are still experienced, sometimes presentations. even more vividly. In terms of technique, this As Figure 19.1 illustrates, in this style of facet of the meditation is indicated by the practice the overall trajectory begins with a fact that one meditates with the eyes open meditation that develops concentration on and directed somewhat upward. an object. One then employs techniques Basic Practice. The actual state of Open that cultivate an awareness of subjectivity Presence is one in which the meditator in a manner that de-emphasizes the object. is aware of the Clarity or Awareness that In doing so, one gains phenomenal access makes all cognitions possible. This state is to the reflexive awareness that is thought a relatively advanced one, and even experi- to be invariant in cognition. One then de- enced practitioners may not be able to sus- emphasizes subjectivity as well so as to fur- tain it for more than a short period of time. ther enhance that access to reflexivity, and There are, however, a series of practices that finally one practices so as to move to the train inexperienced meditators to cultivate point where the invariant aspect of aware- Open Presence, and even experienced prac- ness is realized fully in meditation. Through- titioners sometimes modulate their practice out this entire process the close guidance of so as to move up or down the scale of prac- an instructor is considered essential. tices, depending on how well the particular This style of practice generally begins session is proceeding. with the development of Focused Attention Schematically, we use the diagram in Fig- (i.e., concentration on a particular object ure 19.1 to summarize the stages of the style as described previously). Initially retaining of practice that leads to Open Presence: some focus on the object, one then cultivates P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

516 the cambridge handbook of consciousness

attention to the state of the subjectivity often employed in oral instructions, one is observing the object. This is in part accom- not to “follow along” (rjes su ’brang ba) a plished by discursive strategies that are chain of thoughts. A much subtler indica- implemented after a certain level of concen- tion of grasping, however, is simply the fact tration and mental calm has been reached. that, in phenomenal terms, the appearance In one such strategy, one is instructed to ask or event seems separate from the subjectivity questions about the object, such as, “Is it in the experience. Thus, when one “releases” inside the mind or outside the mind?” Or, objects, one must do so with the understand- when an appearance arises, one observes it ing that the objects actually are not sepa- and asks, “Where did it arise from?” Or as rate from awareness itself, of which the sub- it abides in the mind, one asks, “Where is jectivity is also just a facet. This attitude is it abiding?” Or as it disappears, one asks, initially developed through discursive strate- “Where did it go?” These questions and sim- gies, which seem to play a crucial role in ilar discursive strategies are used to train developing Open Presence. one to see the object as just a phenom- Having become adept at emphasizing enal appearance and to create the subjec- subjectivity – attending to the state of one’s tive impression that the appearance is not awareness without construing its contents something separate from one’s mind. As a as separate from the subjectivity – the next means to heighten one’s awareness of subjec- stage of the practice involves techniques tivity, the same types of questions are then that de-emphasize subjectivity itself. The- applied to the phenomenal appearance as oretically, this is accomplished in part by mind (with questions such as, “What color one’s facility at releasing objects. That is, is the mind?” “What shape is the mind?”), because awareness is construed as subjec- which has the effect of pointing out the man- tivity in relation to objects, the practice of ner in which the phenomenal content is acci- releasing objects will also erode subjectivity. dental to the experience. Along with or in But another important aspect of this stage lieu of such strategies, a deliberate perturba- of practice is not to grasp onto subjectivity tion – such as a sudden shout – may be intro- itself as an object. That is, as one is attempt- duced into the meditation so that the effects ing to abide in a state that is aware and yet on subjectivity will be especially salient. not focused on an object, one may still have The move to an emphasis on subjectivity a sense of subjectivity that is caught up in an is further encouraged by dropping any delib- identity that extends beyond the particular erate focus on an object. As a sensory con- moment – as such, that sense of an identity is tent or mental event occurs, one observes considered an accidental feature of the state it (sometimes along with the momentary because it changes over time. use of a discursive strategy), and then one One of the many remedies employed here releases any focus on it. This is similar to the is the repeated (and somewhat paradoxical) Vipassana¯ practice discussed above, except instruction not to make an effort to med- that after releasing the content or event one itate. In other words, for the meditator a does not return to any object. Instead, one persistent way that the sense of “I” man- releases the mind into its “natural state” (rang ifests would be in the form of a thought, babs), which one understands to be the state such as, “I am meditating.” Such a thought reﬂecting only the invariant nature of con- involves conceptual and linguistic structures sciousness and not the accidental properties that connect to a sense of “I” located in the of subject and object. past and the future. And because that way of One is also repeatedly reminded by one’s locating subjectivity – essentially as a narra- instructor that “grasping” (’dzin pa) – taking tive agent – changes from one cognitive con- the mental content as an object – is to be text to the next, it is a type of subjectivity avoided. Here, the gross symptoms of grasp- that is thought to obscure the invariant fea- ing include indications that one has begun ture of consciousness. According to most tra- to focus on or examine the content or event ditional accounts, it is extremely difﬁcult to and then elaborate upon it – in a phrasing de-emphasize subjectivity to this degree. As P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 517

a result, most beginner and mid-range med- continue to experience phenomena with- itators engaged in the cultivation of open out objectifying them and, ideally, with- presence are likely to be actually meditating out having a sense of an agentive or nar- at the level below this stage; that is, the stage rative subjectivity. The state thus seems to at which subjectivity is still emphasized. cultivate a type of ipseity or bare aware- Through the above techniques – along ness. After a session, for advanced medi- with other methods that involve visualiza- tators the objects of perception will phe- tions and breathing exercises – advanced nomenally appear to be less fixed and more meditators are thought to eventually induce like appearances in a dream or a mirage a particular phenomenal experience: The for at least some period afterward. And as experience’s content does not appear as an one advances further, the state between ses- object over against a subject, and the expe- sions begins to seem more like Open Pres- rience also does not involve a sense of sub- ence itself. The relevant longer-term traits jectivity that is articulated by conceptual that are expected to arise from cultivating or linguistic structures, even if those struc- Open Presence include most prominently tures are only implicit. It is worth reiterat- a facility to regulate one’s emotions, such ing, however, that in de-emphasizing both that one is disturbed less easily by emotional object and subject, the aim of the practice is states. The mind is also said to be more not to become withdrawn from experiences, sensitive and flexible, and the cultivation whether perceptual or mental. Instead, the of other positive states and traits is there- aim is for experiences to continue to occur fore greatly facilitated. All three of our main even though the state de-emphasizes the traditional sources (Karma Chagme,´ 2000; particularity of the object and subject. It is Thrangu & Johnson, 2004; Wangchug Dorje,´ in this way that, according to Chag-zôg theo- 1989) make it clear that these and other rists, one will become aware of the invariant indications are thought to be observable in feature of all states of consciousness. behavior, because it is through observing and Finally, at the highest level of practice, interviewing students that the meditation what we have described as a de-emphasis master is able to guide them in this difficult of both object and subject moves, at least practice. theoretically, to a point where no elements of objectivity or subjectivity – whether in non-referential compassion the form of conceptual structures, categories (mıg-mˆ e´ nying-je´) of time and space, or some other feature – Unlike practices oriented toward gener- remain in the experience. At this point, the ating Focused Awareness or Open Pres- invariant feature of cognition is said to be ence, the practice of Non-Referential realized fully by the meditator, and this is the Compassion aims to produce a specific full-blown state of Open Presence. It seems emotional state; namely, an intense feeling that because this state is extremely advanced of lovingkindness.5 The state is necessar- in each generation of practitioners the Chag- ily other-centered, but it is non-referential zôg traditions recognize only a small number (dmigs med) in that it does not have any spe- of practitioners as having truly reached this cific object or focus (dmigs pa), such as a level of practice. specific person or group of persons. Thus, In terms of the effects of the practice, in effect this meditation has two aspects: one ability developed through cultivating the cultivation of compassion and the cul- Open Presence is the stability of the state – tivation of objectless awareness (i.e., Open that is, one is not easily perturbed out of the Presence). Hence, this practice may be con- state. The difference, however, is that unlike sidered a kind of variation on Open Pres- in Focused Attention, in Open Presence ence, but it also differs somewhat from Open the stability is not constituted by the fact Presence. That is, except for the earliest that other phenomena do not pull one stages of the practice, in Open Presence away from the object on which one focuses. the meditator does not usually require any Instead, stability consists of one’s ability to particular mental content or event as the P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

518 the cambridge handbook of consciousness

context for the cultivation of Open Presence. a person toward whom one feels indiffer- But in the cultivation of Non-Referential ence. With a visualization of these persons Compassion, one does require a particular in place, one then employs discursive strate- mental event – the emotion of compassion – gies – such as the argument that all beings are that forms the context for the cultivation equal in wanting to be happy and wishing to of the objectless awareness that is Open avoid suffering – that are designed to elimi- Presence. nate one’s biases toward these persons. In the The two aspects of Non-Referential Sevenfold Causal Instructions, one is then Compassion – compassion and Open Pres- encouraged not only to see all beings as equal ence – must occur together for the med- but also to take one’s mother as paradig- itation to be successful (Wangchug Dorje,´ matic of all beings. Another set of discur- 1989), but although precise descriptions of sive contemplations – sometimes including this practice are not readily available, it specific visualizations – are then used to disappears that for many practitioners this prac- place one’s preferential treatment of oneself tice requires a sequence within the session. over others. One contemplates, for example, In some cases, a meditator may first cultivate how despicable one would be to prefer one’s Open Presence and then cultivate compas- own happiness over the well-being of one’s sion while retaining the state of Open Pres- mother; here, the practitioner might recall ence to the greatest degree possible. After a memorized aphorism or the admonitions compassion has been evoked, the meditator of his or her teacher. Finally, by recalling may then emphasize Open Presence once or visualizing the intense suffering experi- again, because the techniques for cultivat- enced by others – i.e., “all sentient beings ing compassion may have led the medita- who are as if one’s mother” (ma sems can tor to stray from an objectless state. In other thams cad) – one becomes motivated empa- cases, a meditator may begin by first evok- thetically to eliminate that suffering. Toward ing compassion, and then, while the mind the endpoint of this process one experiences is suffused with compassion, the meditator a visceral, emotional reaction that is said will cultivate Open Presence. to involve especially a feeling of opening at The sequentiality of the practice, which the center of the chest, sometimes accom- does not apply to the most advanced prac- panied by horripilation and the welling of titioners, stems largely from the methods tears in the eyes. This state involves both love that are initially used to evoke a compas- (matr¯ı) – the aspiration that other beings be sionate mental state. These methods often happy – and compassion (karun. a)¯ – the aspi- combine multiple techniques, most espe- ration that other beings be free of suffer- cially a discursive strategy (usually the steps ing. At this point the state might involve a of a memorized argument), a set of visu- degree of sentimentality, and the final phase alizations, and sometimes a litany or other of developing compassion is meant to go recitation. In all the Tibetan traditions, three beyond that state to one that is both more such meditations are widely practiced: the stable and also more engaged with aiding “Sevenfold Causal Instructions” (sems bskyed others (Dalai Lama XIV, 1991). rgyu ’bras man ngag bdun), the “Equanimous Most Tibetan practitioners are trained Exchange of Self and Other” (bdag bzhan intensively in this type of contemplation mnyam brjes), and the practice of “Giving for generating compassion. It is evident, and Taking” (gtong len) (Dalai Lama XIV, however, that these techniques for induc- 1991, 1995). ing compassion are not objectless, inas- All three of these practices, which them- much as they involve visualizations, argu- selves may be combined in various ways, typ- ments, aphorisms, litanies, and so on that are ically begin with an evocation of equanimity focused on objects of one kind or another. (btang snyoms, Skt, upeks.a)¯ . Often a visu- Nevertheless, having generated compassion, alization of three persons is used: a beloved the practitioner can then cultivate Open person (most especially one’s mother), a per- Presence from within that state. Indeed, as son for whom one has some enmity, and a phenomenally intense state, compassion P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 519

is well suited to the early stages of cul- Generating a Description tivating Open Presence, because compassion’s intensity lends itself well to an aware- As the discussion above indicates, even ness of subjectivity and, hence, reflexivity. within Buddhism a large number of distinct And if the emotional state of compassion contemplative practices continue to be prac- can be sustained even while one is cultivat- ticed in living traditions. For this reason, ing Open Presence, the meditator is in the significant changes in meditative style are state of Non-Referential Compassion. As a found even for a basic samatha´ style of med- meditator becomes more adept at cultivat- itation focused on the sensation made by ing compassion through the various tech- the breath. For instance, practitioners from niques mentioned above, the mind becomes the Vipassana¯ or Insight Meditation move- more habituated to the state such that an ment may practice this meditation with the advanced practitioner can induce the state eyes closed so as to de-emphasize the impor- of compassion almost effortlessly. At this tance of the visual modality. In contrast, clos- stage the practice would no longer require ing the eyes is rarely encouraged in the Zen a sequence; that is, compassion can be cul- and Tibetan traditions, in part because it is tivated directly within a state of Open Pres- assumed to induce dullness or drowsiness. ence itself (Wangchug Dorje,´ 1989). The difference of opinion concerning the In general the cultivation of compassion closing of the eyes is illustrative of the dif- is thought to grant the meditator numerous ficulties that researchers face when speci- beneficial effects between sessions, such as fying the exact nature of a meditation to creating a general sense of well-being and be studied. One main problem is that the aiding in counteracting anger or irritation. traditional accounts of these practices move Long-term practitioners of this practice are far beyond the sketches given here; instead, also said to have an effect on others around those accounts are usually highly detailed them, in that other persons nearby may also and extremely complex. Likewise, the ter- feel a greater sense of well-being and happi- minology used to describe the practices ness. Compassion is also thought to provide is sometimes unreliable, either because of benefits when one is in a meditative session ambiguities in the traditions themselves or involving other practices. It is especially use- problems in translation. Practices from dif- ful for counteracting torpor in meditation; ferent traditions may in fact overlap signifi- that is, it is a considered a strong antidote cantly – the overlap between contemporary for dullness, as mentioned earlier. Likewise, Mindfulness practice and Open Presence because compassion is other-centered, it is meditation is one case in point. Finally, sub- considered to develop traits that are essential jects from traditionally Buddhist cultures for the successful cultivation of Open Pres- sometimes are reluctant to depart from tex- ence. That is, in developing Open Presence tual descriptions of meditative practices. To one must eliminate the mind’s “grasping” do so would require a practitioner to assert directed toward objects and also toward sub- some authoritative experience as a medita- jectivity itself. Grasping, moreover, is rooted tor, and it is usually thought to be inappro- in a persistent trait within the mind that priate to claim that degree of accomplish- absolutizes the standpoint of the subject. By ment in meditation. persistently orienting the meditator toward Researchers may also encounter problems others, compassion lessens this fixation on when attempting to assess the degree of self and makes it possible for grasping to training and practice over the life of a par- be eliminated through the practice of Open ticular practitioner. One of the main diffi- Presence. In this way, the cultivation of com- culties is that, traditionally, contemplative passion is thought to train the mind in a way practice involves many varieties of medita- that is essential to the success of some prac- tion, each of which mutually influence each tices (Dalai Lama XIV, 1991, 1995; Karma other and differentially affect the mind. The Chagme,´ 2000; Thrangu & Johnson, 2004; quantification of the total hours of medita- Wangchug Dorje,´ 1989). tion throughout a practitioner’s life is thus P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

520 the cambridge handbook of consciousness

not straightforward and will require further through the senses, or is it imagined methodological development. in the mind through visualization or In any case, before tackling the prob- another technique? r lem of quantification, an important task for If the meditation does not include an any researcher is generating a precise and object, then does one direct one’s atten- concrete account of the practices in which tion to something else? meditating subjects claim to be engaged. The 3. Concerning meditative techniques, one may best way to proceed is probably to develop ask the following: r a list of questions that can be used to Is the practice done with the eyes help prompt descriptions from practitioners opened or closed? r without getting caught either in traditional Does the practice employ any discur- categories or issues of cultural translation. To sive strategies, such as recitations, mem- assist researchers in this task, we close this orized descriptions, or arguments that section on Buddhist meditation with a prac- one reviews? r tical series of sample questions that would Does the practice use breath mani- aid both the meditator and researcher in pulation? r defining more clearly the relevant facets of Does the meditation involve focus- the practice to be studied. The questions ing on different parts of the body by address five overall issues: (1) the relative means of a visualization or some other degree of stability and clarity appropriate to technique? r the practice; (2) the “intentional modality” Does the practice require a specific pos- (i.e., whether the meditation has an object); ture or set of physical exercises? 3 ( ) the techniques, such as breath manipu- 4. Concerning expected effects during medita- 4 lation, that are employed; ( ) the expected tion, one may ask the following: r effects of the practice during meditation; Is the meditation expected to pro- 5 and ( ) the expected effects after a session. duce any physical sensations or Although based especially on the practice mental events, either constantly or and theory of meditation found in Tibetan intermittently? r Buddhism, questions of this type are likely Does one expect the meditation to to be useful when examining other contem- produce subjectively noticeable alter- plative traditions. ations in cognition, either constantly or 1. Concerning stability and clarity, one may intermittently? One example would be ask the following: the impression that one’s perceptions r In view of the practitioner’s level, seem to be like the appearances in a should the meditation favor stability, dream. r clarity, or a balance? Is the meditation expected to cause b What are the indications that stability any emotions, either constantly or needs adjustment? intermittently? b What are the indications that clarity 5. Concerning expected effects after medita- needs adjustment? tion, one may ask the following: r 2. Concerning intentional modality, one may Does one expect the meditation to alter ask the following: one’s cognitions? One example would r If the meditation includes an object, be the impression that one’s percep- then, tions are more vivid. b r Is there one object or many objects in Does one expect the meditation to alter the meditation? one’s behavior? One example would be b For each object, is the object dynamic a tendency to sleep less. r or static? Does one expect the meditation to alter b If the object includes or consists of one’s emotions? One example would be a visual form, a sound, or a sen- the tendency to recover more quickly sation, then is the object perceived from emotional disturbances. P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 521

By adapting or adding to this list of ques- pass either meditative states or traits indu- tions for the particular practice in question, ced by meditation. Meditative states refer meditators and researchers may be able to to the transient alterations of experience collaborate more readily so as to describe in voluntarily cultivated by a given medita- a straightforward way the major features of tion practice (i.e., bodily awareness, relax- a practice that are relevant to a particular ation, emotions, and so on). Traits refer neuroscientific research agenda. With such to the lasting changes in these dimen- descriptions in place, the dialogue between sions that persist in the practitioner’s daily meditators and researchers can be far more experience irrespective of being actively precise, and the interaction between neuro- engaged in meditation. science and meditation is therefore likely to 2. Advanced practitioners can robustly be more fruitful. reproduce specific features in experience as cultivated in given meditative practice. This reproducibility makes those features The Intersection of Neuroscience scientifically tractable. and Meditation 3. Advanced practitioners provide more refined first-person descriptions of their This section briefly explores some possi- experiences than na¨ıve subjects. Thus, the ble scientific motivations for the neuroscien- neurophysiological counterpart of these tific examination of meditative practices and first-person accounts can be defined, their possible impact on the brain and body identified, and interpreted more easily by in advanced practitioners. The aim here is to the experimentalist. clarify further the distinguishing features of this approach compared to other empirical We now discuss these claims in relation to strategies described in this handbook. Before three neuroscientific agendas: neuroplastic- we move forward with this discussion, how- ity, the interaction of mind and body, and the ever, two points of clarification need to be possibility of neural counterparts to subjec- made. First, because of the relative paucity tive experience. In the course of this discus- of currently available empirical data in this sion, specific techniques from the Buddhist field, this section remains largely program- tradition serve as illustrations. matic. Second, we discuss some studies that involve novice meditators, but the set of Transforming the Mind and issues examined here are most relevant to Brain Neuroplasticity advanced practitioners of meditation. Nev- ertheless, emphasis on advanced practition- From a neuroscientific perspective, the first ers should not minimize the importance of promising claim made by Buddhist con- studying meditation in less practiced individ- templative traditions is that experience is uals. Indeed, some of us have already done so not a rigid, predetermined, and circum- (see, e.g. Davidson et al., 2003). For progress scribed entity, but rather a flexible and trans- in this general area to advance, we believe formable process. On this view, emotions, that research on practitioners at all levels attention, and introspection are ongoing and should be encouraged, but one must also rec- labile processes that need to be understood ognize that the goals of studying individuals and studied as skills that can be trained, sim- at different levels of accomplishment differ ilar to other human skills like music, math- somewhat. ematics, or sports. This principle is founda- Turning now to the question of advanced tional for Buddhist contemplative practice, practitioners, we begin by noting three fre- because such practices are based upon the quently advanced hypotheses: notion that the mind is malleable in this way. As a result, the methods employed by Bud- 1. Advanced practitioners can generate new dhist contemplative practices resonate with data that would not exist without susta- widely accepted developmental models of ined mental training. These data encom- basic cognitive processes; according to these P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

522 the cambridge handbook of consciousness

models, cognitive functions are skills that as a cab driver predicted the size of the critically depend upon learning from envi- posterior hippocampus (Maguire et al., ronmental input (e.g., McClelland & Rogers, 2000). Further work by this group suggests 2003; Saffran, Aslin, & Newport, 1996). This that these differences in hippocampal size basic stance reflects another well-accepted are the results of experience and training and well-documented theory; namely, that as a cab driver and not a consequence of experience changes the brain. Interest in pre-existing differences in hippocampal this feature, known as neuroplasticity, has structure (Maguire et al., 2003). prompted an explosion of research over the Whether similar structural alterations in past decade. different regions of the brain occur also as a As a result of ongoing research on neu- consequence of affective – rather than sen- roplasticity, we now have a detailed under- sorimotor – experience is not definitively standing of many of the molecular and known, but all of the extant work at both system-level changes that are produced by the animal and human levels indicates that specific types of experiential input. For it does. Certainly there are good animal example, neonatal rodents exposed to vary- data to suggest that such changes occur, as ing levels of maternal licking and groom- indicated by the study on neonatal rodents. ing develop very different behavioral pheno- In humans, a variety of research indicates types. Those animals that receive high levels that deleterious conditions, such as chronic of licking and grooming (the rodent equiva- stress, neglect, and abuse, produce func- lent of highly nurturing parenting) develop tional changes in the brain that are likely into more adaptable and relaxed adults. Of subserved by structural alterations (Glaser, great interest is the fact that the brains of the 2000). Likewise, research on depression animals are critically affected by this differ- indicates that patients with mood disorders ential rearing. Indeed, gene expression for exhibit structural differences in several brain the gene that codes for the glucocorticoid regions, including the hippocampus and ter- receptor is actually changed by this experi- ritories with the prefrontal cortex; signif- ence, and the detailed molecular pathways icantly, at least some of these differences by which experience can alter gene expres- are associated strongly with the cumulative sion have now been worked out in this model number of days of depression in the patients’ (Meaney, 2001). This program of research lifetimes (Sheline, 2003). illustrates the profound ways in which neu- These findings raise the possibility that roplasticity can unfold, and it demonstrates training and practices that are specifically that experience-induced alterations in the designed to cultivate positive qualities, such brain can occur all the way down to the level as equanimity and lovingkindness, will proof gene expression. duce beneficial alterations in brain function Meaney’s work on alterations in brain and structure. Presumably, these alterations gene expression implies that similar would be most prominent in long-term, experience-induced alterations might occur advanced practioners, but we have already in humans. Currently, however, there are shown that even very brief short-term train- no direct measures of localized neuronal ing (30 minutes) in emotion regulation can gene expression that can be non-invasively produce reliable alterations in brain function obtained in humans. Nevertheless, other (Urry et al., 2003). So too, we have observed research suggests that such changes do that a 2-month course in Mindfulness-Based indeed occur. For instance, the brain of Stress Reduction (MBSR) can produce alter- an expert, such as a chess player, a taxi ations in patterns of prefrontal brain activ- driver, or a musician, is functionally and ity that we have previously found to structurally different from that of a non- accompany positive affect (Davidson et al., expert. London taxi cab drivers have larger 2003). hippocampi than matched controls, and the The findings concerning MBSR may be amount of time the individual has worked especially relevant, because MBSR is likely P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 523

to provide a large pool of persons for the ine possible anatomical changes induced by study of neuroplasticity and meditation. An meditation. 8-week program that was originally developed in a hospital setting for patients with Mechanisms of Mind-Body Interaction chronic illnesses, MBSR is now applied across an extremely wide range of popu- In addition to the study of neuroplasticity, lations (Kabat-Zinn & Chapman-Waldrop, one of the most potentially fruitful ques- 1988; Kabat-Zinn, Lipworth, & Burney, tions in the study of meditation is the impact 1985). The method, based primarily on of training the mind on peripheral biolog- Buddhist practices, seems to be effective ical processes that are important for phys- for chronic pain, anxiety disorders, gen- ical health and illness. Quite literally, the eral psychological well-being, psoriasis, and question here is whether mental training can recurrent depression (Grossman, Niemann, affect the body in a way that will have a sig- Schmidt, & Walach, 2004). The program nificant impact on physical health. Although seems to work by helping the patient dis- there are many popular claims about the tinguish primary sensory experience (e.g., health benefits of meditation and contem- chronic pain, physical symptoms of anxiety) plative practice, there is relatively little that from the secondary emotional or cognitive is solidly known about this potentially cru- processes created in reaction to the primary cial issue. Even more importantly, there are experience. Individuals are trained to use the preciously few attempts to mechanistically mindfulness practice to elicit the details of link the changes that are occurring in the their experience and to directly perceive the brain with alterations that may be produced unstable and contingent nature of the feel- in peripheral biology. It is beyond the scope ings and sensations that are associated with of this chapter to review the basic research aversion and withdrawal; as a result, indvid- relevant to these questions, but it provides uals are better able to counter any propensity some general guidelines and examples. toward withdrawal and aversion in response It is established that there is bidirectional to physical or psychological pain. communication between the brain and the From a neuroscientific perspective, the periphery and that this communication pro- apparent effectiveness of MBSR practice ceeds along three basic routes: the auto- raises the question of neuroplasticity – that, nomic nervous system, the endocrine sys- is, does it produce alterations in brain func- tem, and the immune system. In each of tion and structure? Recent data indicate these systems, specific pathways and sig- a possible relationship between medita- naling molecules enable this bidirectional tion training and changes in brain structure communication to occur. These structural (Lazar et al., 2005). In this study, corti- characteristics are highly relevant to the cal thickness was assessed using magnetic possibility that meditation may influence resonance imaging. Increased cortical thick- physical health. That is, some conditions ness could be caused by greater arboriza- of peripheral biology may be potentially tion per neuron, increased glial volume, or affected by meditative practices because increased regional vasculature, all of which those conditions – such as an illness – are important for neural function. Corti- are susceptible to modulation by the auto- cal brain regions associated with attention, nomic, endocrine, and/or immune pathways interoceptive, and sensory processing were involved in brain-periphery communication. found to be thicker for a group of mid-range Thus, because there is bidirectional com- practitioners than for matched controls (the munication between the brain and periph- meditator participants had, on average, 40 ery, it is theoretically possible to affect those minutes of daily practice of Insight medita- types of conditions by inducing changes tion for an average of 9 years). We antici- in the brain through meditation. At the pate that research conducted over the next same time, however, other conditions or ill- 2 years by several groups will further exam- nesses may not be influenced in this way by P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

524 the cambridge handbook of consciousness

2.35 eek to

W 2.3

accine 2.25 ost-V P

eek 2.2 8–9 W med Antibody Rise from 3–5 or 2.15 ansf r T Log 2.1 Control Meditation Figure 19.2 . Effect of meditation training on the immune system during a Mindfulness-Based Stress Reduction program (8-week program) with novice practitioners. Means ± SE antibody increase from the 3-to5-week to the 8-to 9-week blood draw in the Meditation and Control groups. The ordinate displays the difference in the log-transformed antibody rise between the 3-to5-week and the 8-to9-week blood draws derived from the hemagglutination inhibition array. (From Davidson et al., 2003.)

meditation because the peripheral biological body response to the influenza vaccine. We processes in question cannot be affected by found that individuals with high levels of any pathway involved with brain-periphery left prefrontal activation, a pattern that we communication. have previously found to predict more pos- With this in mind, the strategy that we itive dispositional mood, had higher levels have adopted in some of our work is to of antibody titers in response to the vac- examine a proxy measure that we know can cine (Rosenkranz et al., 2003). The spe- be modulated by central nervous changes cific question of whether mental training and that is health-relevant. In a seminal could improve the immune response was study, Kiecolt-Glaser, Glaser, Gravenstein, addressed in our study of MBSR (David- Malarkey, and Sheridan (1996) reported son et al., 2003; Figure 19.2). In this study, that caregivers of dementia patients had an individuals who had been randomly assigned impaired response to influenza vaccine com- to an MBSR group were compared to a pared with a matched control group. Several wait-list control group. Subjects were tested groups have now independently replicated just after the MBSR group had completed this basic finding and have examined some their 8-week training. We found that the of the details regarding the mechanism by meditators exhibited a significantly greater which stress might impair humoral immu- antibody response to the influenza vaccine. nity (e.g., Miller et al., 2004). Of most importance in this study was an In our group, we have investigated analysis we conducted that examined rela- whether individual differences in patterns of tions between brain and immune function prefrontal brain activity that we have pre- changes with meditation. We found that for viously found to be associated with affec- the individuals assigned to the meditation tive style (i.e., individual differences in pro- group, those who showed the greatest shift files of affective reactivity and regulation) toward left-sided activation also exhibited are also associated with differences in anti- the largest increase in antibody titers to the P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 525

CONTROL MEDIT ATION RECOVERY 36 CALF NAVEL LUMBAR 34 TOE 32 STERNUM FOREARM 30

° 28

ature ( 26 FINGER 24 100

emper HEAR T RATE T 80 − 22 (beats min 1 ) 20 60 AIR 18

01020 3040 102030

CONTROL MEDIT ATION 0 RECOVERY Time(min) Figure 19.3. Effect of Tummo meditation on the regulation of body temperature. Skin and air temperature and heart rate changes before, during, and after the meditation of long-term practitioner L.T. (adapted from Benson et al., 1982).

vaccine. These findings suggest some associ- practices can affect the body in a way that ation between the magnitude of neural and improves health. immune changes. A variety of other findings in the litera- Using First-Person Expertise to Identify ture suggest that autonomic changes occur the Neural Counterpart of during specific types of meditation. As Subjective Experience noted in the first part of this chapter, the Tibetan Tummo practice has as its byprod- As discussed in the first section of this chap- uct the production of heat. Benson and his ter, various meditative practices induce a colleagues reported on three practitioners wide variety of altered states of conscious- and found that, by using this practice, they ness. It is thus frequently claimed that the were indeed able to voluntarily increase tem- study of meditation will contribute to our perature in their toes and fingers by as much general understanding of the neural basis of as 8◦C (Figure 19.3; Benson et al., 1982). consciousness. Here we aim to move beyond Takahashi et al. in a study of Zen medita- this general claim and illustrate how specific tion in a fairly large sample (N = 20) found Buddhist practices might provide research changes in heart rate variability (reflect- opportunities to glean new insights about ing parasympathetic nervous system activ- some of the brain mechanisms contributing ity) that were associated with changes in to consciousness. More precisely, we discuss specific EEG frequencies (Takahashi et al., how the collaboration with long-term Bud- 2005). In future studies, the combination of dhist practitioners is of great interest in the both brain and peripheral measures will be study of (1) the physical substrate of sub- important in helping understand the mecha- jectivity or the self, (2) the physical prin- nisms by which such peripheral changes may ciples underlying the emergence of coher- be occurring. This type of data will be espe- ent conscious states from unconscious brain cially useful as research moves forward on processes, and (3) the functional role of the the practical question of whether meditative spontaneous brain baseline. P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

526 the cambridge handbook of consciousness

studying the substrate explained as involving a reflexive awareness of subjectivity of one’s lived body or embodied subjectiv- One of the useful points of intersection ity correlative to the experience of the object between Buddhist contemplative practice (Mearleau-Ponty, 1962; Wider, 1997). It is, in and the neuroscience of consciousness is the short, another way of speaking about ipseity. emphasis on understanding the nature of the If the above model is correct, then in self. A comparison between Buddhist and giving an account of consciousness, neuro- neuroscientific models of the self is beyond science needs to explain both “how the brain the scope of this chapter, but it is impor- engenders the mental patterns we experi- tant to note that both traditions distinguish ence as the images of an object” and “how, between a minimal subjective sense of “I- in parallel . . . the brain also creates a sense ness” in experience, or ipseity, and a narrative of self in the act of knowing . . . ” In other or autobiographical self. Ipseity is the mini- words, to give an account of consciousness, mal subjective sense of I-ness in experience, neuroscience must show how it is that one and as such, it is constitutive of a minimal or has a “sense of me” by demonstrating “how core self. By contrast, a narrative or autobio- we sense that the images in our minds are graphical self (Legrand, 2005) encompasses shaped in our particular perspective and categorical or moral judgment, emotions, belong to our individual organism” (Dama- anticipation of the future, and recollections sio, 1999,pp.136–137). As a number of of the past. This explicit sense of narra- cognitive scientists have emphasized, this tive or autobiographical self is often char- primitive self-consciousness might be fun- acterized as occurring in correlation with an damentally linked to bodily processes of life explicit content, or object, of experience. It regulation (Damasio, 1999; Panksepp, 1988). also appears to be dependent in some fash- This approach to the question of con- ion on ipseity, inasmuch as the narrative self sciousness suggests a research strategy that is in part based upon that minimal subjective might be aided by the use of experienced sense of I-ness. meditators. That is, to understand conscious- The notion of ipseity is further explained ness, it is presumably best to begin by exam- through Western phenomenological theory, ining it in its simplest form. And on this the- according to which one can speak of experi- ory, the simplest form of a conscious state is ence in terms of both transitive and intran- reducible to ipseity, which is required for or sitive modes of consciousness. Any experi- is prior to the narrative self. One experimen- ence “intends” (i.e., refers to) its intentional tal strategy would be to involve long-term object; this is its transitive aspect. At the practitioners of a practice such as Open Pres- same time, the experience is also reflexive- ence that allegedly induces a state in which ly manifest to itself, and this is its intransi- ipseity is emphasized and the narrative self tive aspect. On this theory, the intransitive is lessened or eliminated. By examining such aspect of experience is a form of self- practices, one may be able to find neural cor- consciousness that is primitive inasmuch as relates of a bare subjectivity, which in turn it (1) does not require any subsequent act may yield some insight into the neural cor- of reflection or introspection, but occurs relates of the most basic type of coherent simultaneously with awareness of the object; states that we call consciousness. (2) does not consist of forming a belief or making a judgment; and (3) is passive in studying the substrate the sense of being spontaneous and involun- of consciousness tary (Zahavi & Parnas, 1998). For instance, Empirical evidence clearly indicates that when one consciously sees an object, one only a selective set of neurons in the brain is also at the same time aware of one’s participates in any given moment of con- seeing; similarly, when one visualizes a men- sciousness. In fact many emotional, motor, tal image, one is also aware of one’s visualiz- perceptual, and semantic processes occur ing. This tacit self-awareness has often been unconsciously. These unconscious processes P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 527

are usually circumscribed brain activities in ally specialized brain regions. A common local and specialized brain areas (Dehaene theoretical proposal is that each moment of & Naccache, 2001). This result suggests that, conscious awareness involves the transient when a stimulus is phenomenally reportable selection of a distributed neural popula- from the standpoint of experience, it is tion that is both integrated or coherent, the result of translocal, large-scale mecha- and differentiated or flexible, and whose nisms that somehow integrate local func- members are connected by reciprocal and tions and processes. In other words, it has transient dynamic links. As we show in been hypothesized that the neural activ- the next section, neural synchronization ity crucial for consciousness most probably and de-synchronization between oscillating involves the transient and continual orches- neural populations at multiple frequency tration of scattered mosaics of functionally bands are popular indicators of this large- specialized brain regions, rather than any sin- scale integration (Engel, Fries, & Singer, gle, highly localized brain process or struc- 2001; Varela, et al. 2001). For instance, dur- ture (Dehaene & Naccache, 2001; Llinas, ing the binocular rivalry discussed above Ribary, Contreras, & Pedroarena, 1998;W. the alternation of perceptual dominance Singer, 2001; Tononi & Edelman, 1998). correlates with different ongoing patterns The issue at stake here can be illus- of distributed synchronous brain patterns trated by the example of binocular rivalry. (Cosmelli et al., 2004; Srinivasan, Russell, When the right eye and the left eye are pre- Edelman, & Tononi, 1999). sented with competing, dissimilar images, Thus it is hypothesized that large-scale the observer does not experience a stable integrative mechanisms play a role in con- superimposed percept of the images pre- scious processes. However, these large-scale sented to the two eyes, but instead per- brain processes typically display endoge- ceives an ongoing alternation between the nous, self-organizing behaviors that cannot images seen by each eye every couple of sec- be controlled fully by the experimenter and onds. When one percept is consciously per- are highly variable both from trial to trial and ceived, the other remains unconscious. Yet, across subjects. Linking these brain patterns even if a stimulus is not reportable, there to the experiential domain is notoriously dif- is evidence it is still processed by the brain ficult in part because the first-person reports in various ways. Activity in the amygdala, are readily inaccurate or biased (Nisbett & which is known to increase during the pre- Wilson, 1977). The problem, in a nutshell, sentation of facial expressions of fear and is that the experimenter is led to treat as anger, is still detectable even when the emo- noise the vast majority of the large-scale tional face is suppressed because of binoc- brain activity, as he or she can neither inter- ular rivalry (Williams, Morris, McGlone, pret it nor control it. Abbott, & Mattingley, 2004). It is in this context that meditation Binocular rivalry suggests that local neu- becomes relevant. We hypothesize that ral processes – such as those involved in the long-term practitioners of meditation can processing of visual stimuli – are not in them- generate more stable and reproducible men- selves sufficient to account for conscious- tal states than untrained subjects and that ness. In other words, the neural process they can also describe these states more occurs, but may or may not be consciously accurately than na¨ıve subjects. The prac- experienced. Some other process or mech- titioners’ introspective skills could provide anism must be involved. It is in response a way for experimenters to better control, to this type of issue that contemporary identify, and interpret the large-scale inte- researchers (Llinas et al., 1998; Singer, 2001; grative processes in relation to the subjective Tononi & Edelman, 1998) have hypothe- experience. sized some kind of integrative mechanisms This “neurophenomenology” approach or processes that, although transient, are (Varela, 1996) was tested in the context of able to orchestrate or coordinate function- a visual protocol with na¨ıve subjects. On P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

528 the cambridge handbook of consciousness

the day of the experiment, participants were tion in rivalry rate, in contrast with Focused first trained during a practice session to be Attention meditation, which led to extreme aware of subtle fluctuations from trial to increases in perceptual dominance durations trial in their cognitive context as defined by that were reported by 50% of monks after their attentive state, spontaneous state pro- a period of single-pointed meditation. The cesses, and the strategy used to carry out the monks reported additional stabilization of task. During the visual task, their electrical the visual switching when they did the visual brain activity and their own report about task during Focused Attention meditation. their cognitive context were recorded. Tri- In sum, the capacity of long-term prac- als were clustered according to the acquired titioners to examine, modulate, and report first-person data, and separate, dynamical their experience might provide a valuable analyses were conducted for each cluster. heuristic to study large-scale synchronous Characteristic patterns of synchrony in the brain activity underlying conscious activity frontal electrodes were found for each clus- in the brain. ter, depending in particular on the stability of attention and the preparation to meditation and physiological do the task (Lutz, Lachaux, Martinerie, & baselines Varela, 2002). A central goal of the practice of meditation is As discussed in the report by Lutz and to transform the baseline state of experience Thompson (2003), the collaboration with and to obliterate the distinction between long-term practitioners will be particularly the meditative state and the post-meditative relevant to the further extension of this state. The practice of Open Presence, for research strategy. Let us return again to the instance, cultivates increased awareness of binocular rivalry paradigm. The perceptual a more subtle baseline (i.e., ipseity) during selection in rivalry is not completely under which the sense of an autobiographical or the control of attention, but can be mod- narrative self is de-emphasized. Long-term ulated by selective attention (Ooi & He, training in Compassion meditation is said 1999). Evidence suggests in particular that to weaken egocentric traits and change the the frequencies of the perceptual switch emotional baseline. Practitioners of Mind- can be controlled voluntarily (Lack, 1978). fulness/Awareness meditation aim to expe- It is possible that long-term practitioners rience the present nowness, and this type of Focused Attention meditation can gain of meditation affects the “attentional base- a more thorough control of the dynamic line” by lessening distractions or daydream- of perceptual switch than na¨ıve subjects like thoughts. In this way, meditative prac- and that they can also refine their descrip- tices are generally designed to cultivate tions of the spontaneous flow of perceptual specific qualities or features of experience dominance beyond the mere categories of that endure through time relatively indepen- being conscious of one or another percept, dent of ongoing changes in somatosensory or thereby leading possibly to new brain cor- external events. These qualities are thought relates. Carter et al. (2005) recently pro- to gradually evolve into lasting traits. vided behavioral evidence that long-term From an empirical standpoint, one way to practitioners can indeed change the rivalry conceptualize these various meditative traits rate during meditation. They reported differ- is to view them as developmental changes ential effects on visual switching accompa- in physiological baselines in the organism. nying rivalry during Non-Referential Com- Finding ways of systematically characteriz- passion meditation and Focused Attention ing these baselines before, during, and after meditation (see the earlier discussion on mental training is thus crucial for the empir- meditative states in Tibetan Buddhism) ical examination of the long-term impact of among 23 Tibetan Buddhist monks varying meditation. A systematic characterization of in experience from 5 to 54 years of training. baselines in the context of meditation is also Compassion meditation led to no modifica- an important methodological issue because P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 529

an essential aspect of experimental research system even amid environmental perturba- is the identification and control of a base- tions. It is in this sense that the notion of a line against which the condition of inter- baseline is related to the ongoing identity of est can be contrasted. Underestimating this an organism. And given the sensitivity that issue is a potential source of confusion when most contemplative traditions show to this studying meditation practitioners, particu- type of issue, the search for baseline changes larly long-term practitioners, because the throughout the continuum of mental train- contrast between the initial baseline and the ing is a general strategy that can potentially meditative state might be biased by a base- be applied to clarify mind-brain-body inter- line difference between groups (i.e., novices actions at many explanatory levels, includ- versus adepts). ing brain chemical, metabolic, or electrical When conceptualizing the notion of a activity; the immune system; the cardiovas- baseline, perhaps the most useful approach cular system; and the hormonal system. is to consider a baseline in terms of the This idea can be illustrated with the case capacity for living systems to maintain their of metabolic and electrical brain baselines. identity despite the fluctuations that affect Investigation of the metabolic brain baseline them. Indeed, for many theorists this abil- began following the finding that in an awake ity is one of the most fundamental biologi- resting state, the brain consumes about 20% cal roots of individuality, if not subjectivity of the total oxygen used by the body, despite itself (Maturana & Varela, 1980). This basic the fact that it represents only 2%ofthe notion of homeostatic identity is the intu- total body weight (Clark & Sokoloff, 1999). ition that underlies the concept of a baseline This finding raises the question of the func- in various domains. In the context of medita- tional significance of this ongoing consump- tion, the notion of a baseline is clearly mean- tion of energy by the brain. Interestingly, ingful in relation to “raising the baseline” by brain imaging techniques that permit an developing traits that persist outside of any examination of baseline levels of brain activ- meditative state. In the scientific context, ity have suggested that this global activa- the concept of a baseline plays an impor- tion at rest is not homogeneously localized tant role in characterizing a broad variety of in the brain. There are a consistent set of biological phenomena. Similarly, in psychol- brain areas active at rest with eyes closed, as ogy a baseline state is defined as a resting well as during visual fixation and the passive state by comparison to a task-specific state viewing of simple visual stimuli. The role (say the task of remembering a succession of these activated areas is revealed from the of numbers). In an even broader, ecological attenuation of their activity during the per- context, psychologists attempt also to iden- formance of various goal-directed actions. tify regularities or traits in the average ongo- Because the activity in these areas is asso- ing states of an individual (e.g., mood and ciated with the baseline activity of the brain personality). Areas of biology that study liv- in these passive conditions, Raichle and col- ing processes at a systemic level – such as league have suggested that they are function- the level of the cell, the organism, or the ally active, although they are not “activated” immune system – also use the notion of the (Gusnard & Raichle, 2001). In contrast to baseline to convey the idea that something the transient nature of typical activations, remains constant in the system through time. the presence of this functional activity in the Such features would include, for example, baseline implies the presence of sustained the electrical charges of the neuron, the glu- information processing. cose level in the blood system, and body Our current understanding of the func- temperature. It is important to note that, tional role of these tonically active networks in all these contexts, the functional invari- is still limited. Evidence from brain imaging ance of a given baseline provides information indicates that the posterior part of this toni- about the homeostatic mechanisms that reg- cally activated network (posterior cingulate ulate and maintain the organization of the cortex, precuneus, and some lateral posterior P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

530 the cambridge handbook of consciousness

cortices) is important for the continuous Neuroelectric and Neuroimaging gathering of information about the world Correlates of Meditation around and possibly within us, whereas the anterior part (ventro- and dorsoventral pre- The search for the physiological correlates frontal cortices) is important for the ongo- of meditation has been centered essentially ing association among sensory, emotional, on three groups: Yogis and students of Yoga and cognitive processes that participate in in India, adherents of Transcendental Med- spontaneous self-referential or introspec- itation (TM; Becker & Shapiro, 1981)in tively oriented mental activity (Gusnard & the United States, and practitioners of Zen Raichle, 2001). One can speculate that, given and Tibetan Buddhism in Japan, the United the nature of many meditations, these brain States, and South Asia (India, Nepal). His- areas activated in the resting brain will be torically, the first studies took place in Asia functionally affected by long-term medita- in the 1950s with advanced yogic practi- tive practices. tioners in India (Das & Gastaut, 1955) and Similarly, the awake, resting brain is with long-term Zen practitioners in Japan associated with a well-defined neuroelec- (Kasamatsu & Hirai, 1966). Since the 1970s, tric oscillatory baseline different from the meditation research has been done almost one during sleep, anesthesia, or active exclusively in the United States on prac- mental activity. Changes in this baseline titioners of TM (Becker & Shapiro, 1981); index are implicated in developmental pro- over 500 studies have been conducted to cesses in children, aging, cognitive IQ, and date. Compared to the degree and sophis- mental disorders (Klimesch, 1999). Along tication of the training of practitioners in the these same lines, we found a group differ- early studies conducted in Asia, TM research ence in the initial premeditation baseline relied mainly on the experience of relatively between long-term Buddhist practitioners novice Western practitioners using mostly a and novices, suggesting some impact of long- single standard relaxation technique. term mental training (Lutz, Greischar, Rawl- Since the late 1990s, we have witnessed a ings, Ricard, & Davidson, 2004). The num- renewed interest in research on meditation. ber of hours of formal meditation during the Various researchers have begun the explo- lifetime of long-term Buddhist practition- ration of a broad range of meditative prac- ers correlates with some oscillatory proper- tices inspired by various traditions – such ties of the brain electrical baselines. Further- as Zen, Tibetan Buddhism, Yoga, and Qi- more, we showed that the post-meditation gong – involving novices, patients, or long- baseline was affected by the meditation ses- term practitioners. Academic research insti- sion, and this suggests a short-term effect tutions are starting to express an interest of meditation on the EEG baseline (Figure in this area as epitomized by the Mind 19.4). In any case, this question of a neuro- and Life meetings at MIT in 2003 and a electric baseline and its relation to mental meeting co-sponsored by the John Hop- training is a fruitful one, and it is currently kins School of Medicine and the George- being investigated, as is discussed in more town School of Medicine in November 2005 detail in the next section. between the Dalai Lama, along with other To summarize, our functional under- Buddhist scholars, and neuroscientists (Bari- standing of the brain and body baselines naga, 2003). remains largely incomplete, and given the This new, broad interest has been fostered importance played by some notion of a base- by several factors. First, the neurobiology of line in most meditative traditions, it is likely consciousness and cognitive, affective, and that our understanding will be significantly social neuroscience have become central and advanced by understanding those meditative accepted areas of research in neurosciences practices that, above all else, aim to trans- over the last two decades, which lends form the “baseline of the mind.” legitimacy to the research on meditation. P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 531

a % % b % 100 100 Controls * 80 * Practitioners * 45 45 * * 40 ** * ** * 0 0 0 12345678910 12345678 IB OB MS Controls Practitioners Figure 19.4. Relative EEG gamma power during non-referential compassion meditation in a group of novices and a group of long-term Buddhist practitioners. a–b. Intra-individual analysis on the ratio of gamma (25–42 Hz) to slow oscillations (4–13 Hz) averaged through all electrodes. a. The abscissa represents the subject numbers, the ordinate represents the difference in the mean ratio between the initial state and meditative state, and the black and red stars indicate that this increase is greater than two and three times, respectively, the baseline standard deviation. b. Interaction between the subject and the state factors for this ratio (ANOVA, F(2,48) = 3.5, p <.05). IB (initial baseline), OB (ongoing baseline) and MS (Fischer, 1971 #68). The relative gamma increase during meditation is higher in the post-meditation session. In the initial baseline, the relative gamma is already higher for the practitioners (p <.02) than the controls and correlates with the length of the long-term practitioners’ meditation training through life (adapted from Lutz et al. 2004).

Second, because meditative practices, such oscillatory neural synchrony and as Yoga or Mindfulness-Based Stress Reduc- consciousness tion (MBSR), are now used routinely in the In 1929, Hans Berger recorded for the first medical environment, clinicians recognize time a human brain’s EEG and reported the need to validate its impact on the brain the presence of several brain rhythms in and the body. these signals (Berger, 1929). Since Berger’s first observation, the various ongoing brain oscillations have been used successfully to Neuroelectric Correlates of characterize mental states, such as sleep, the Meditative States waking state, or vigilance, and mental Since the 1950s, the electrophysiological pathologies, such as epilepsy. Sensory evo- measure of brain or autonomic system has ked potentials (EEG signals triggered by an been the most popular imaging tool with external stimulation) or the Bereitschaftpo- which to study meditation. tential (a “readiness EEG potential” that can Electroencephalography (EEG: Cooper be recorded over motor areas up to 1 s before et al., 2003) is a non-invasive technique the execution of a movement) have demon- that measures the electrical potentials on the strated that such mental factors as sensation, scalp. EEG has an excellent temporal reso- attention, intellectual activity, and the plan- lution in the millisecond range that allows ning of movement all have distinctive elec- the exploration of the fine temporal dynamic trical correlates at the surface of the skull of neural processes. Below we discuss some (Zeman, 2001). basic findings about the nature and function Even though EEG results may be also of brain oscillatory processes as measured contaminated by muscle activity, EEG oscil- by electrophysiology, present some common lations are believed to reflect mostly the theoretical assumptions about the neurody- post-synaptic activity of neurons, in partic- namic basis of consciousness, and review the ular from the neocortex. More precisely, neuroelectric correlates of various medita- when a large population of neurons recorded tive styles. by a single electrode transiently oscillates at P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

532 the cambridge handbook of consciousness

the same frequency with a common phase, imately 1 cm, through monosynaptic con- their local electric fields add up to produce a nections with conduction delays of 4 to 6 burst of oscillatory power in the signal reach- ms. Most electrophysiological studies in ani- ing the electrode (Nunez, Wingeier, & Sil- mals have dealt with short-range synchronies berstein, 2001). The amplitude, or power, or synchronies between adjacent areas cor- of these EEG oscillations thus provides a responding to a single sensory modality. coarse way to quantify the synchronization These local synchronies have usually been of a large population of oscillating neurons interpreted as a mechanism of perceptual below an electrode. binding – the selection and integration of Oscillatory neural synchrony is a funda- perceptual features in a given sensory modal- mental mechanism for implementing coor- ity (e.g., visual Gestalt features). dinated communication among spatially dis- Large-scale integration concerns neural tributed neurons. Synchrony occurs in the assemblies that are farther apart in the brain at multiple spatial and temporal scales brain and are connected through polysy- in local, regional, and long-range neural naptic pathways with transmission delays networks. At the cellular level, oscillatory longer than 8 to 10 ms (Schnitzler & Gross, synchrony, or phase synchrony, refers to the 2005; Varela et al., 2001). In this case, syn- mechanism by which a population of oscil- chrony cannot be based on the local cellular lating neurons fires their action potentials architecture, but must instead reside in dis- in temporal synchrony with a precision in tant connections (cortico-cortical fibers or the millisecond range. At the population thalamocortical reciprocal pathways). These level, neuronal synchrony is best analyzed pathways correspond to the large-scale con- by looking at the common average oscil- nections that link different levels of the latory neural activity among the popula- network in different brain regions to the tion. This oscillatory activity can be mea- same assembly. The underlying mechanism sured either from the local field potentials of long-range synchrony is still poorly under- (the summated dendritic current of local stood (for a model see Bibbig, Traub, & Whit- neural groups) or from the macroscale of tington, 2002). Long-range synchronization scalp recordings in EEG (Becker & Shapiro, is hypothesized to be a mechanism for the 1981) and magnetoencephalography (MEG); transient formation of a coherent macro- Srinivasan et al., 1999). The emergence of assembly that selects and binds multimodal synchrony in a neural population depends networks, such as assemblies between occip- on the intrinsic rhythmic properties of indi- ital and frontal lobes, or across hemispheres, vidual neurons, on the properties of the net- which are separated by dozens of millisec- work (Llinas et al., 1998), and on the inputs onds in transmission time. The phenomenon delivered to the network. As a general princi- of large-range synchrony has received con- ple, synchrony has been proposed as a mech- siderable attention in neuroscience because anism to “tag” the spatially distributed neu- it could provide new insights about the rons that participate in the same process emergent principles that link the neuronal and, consequently, to enhance the salience and the mental levels of description. Sev- of their activity compared to other neurons eral authors have proposed that these mech- (Singer, 1999). anisms mediate several generic features The functional and causal roles of syn- of consciousness, such as unity (Varela & chrony are still an active area of research Thompson, 2001), integration and differen- and depend on the spatial scale at which tiation (Tononi & Edelman, 1998), transi- these phenomena are analyzed. In particular, toriness and temporal flow (Varela, 1999), it is useful to distinguish between two main and awareness of intentional action (Free- scales, short range and long range. Short- man, 1999). In this view, the emergence of a range integration occurs over a local net- specific coherent global assembly underlines work (e.g., columns in the primary visual the operation of any moment of experience. cortex), distributed over an area of approx- The transition from one moment to the next P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 533

would be subserved by desynchronization taut, 1955; Wenger & Bagchi, 1961). Since of some coherent assemblies and the syn- then, more than 100 studies have investi- chronization of new ones. It has also been gated the tonic changes in the ongoing EEG hypothesized that whether a local process from a restful state to a meditative state or participates directly in a given conscious the modulatory, or phasic, effect of medi- state depends on whether it participates tation on the electrical brain responses to in a coherent, synchronous global assembly external sensory stimuli (for reviews, see (Dehaene & Naccache, 2001; Engel et al., Andresen, 2000; Davidson, 1976; Delmonte, 2001). 1984; Fenwick, 1987; Pagano & Warrenburg, Neural synchronies occur in a broad range 1983; Schuman, 1980; Shapiro & Walsh, of frequencies. Fast rhythms (above 15 Hz) 1984; West, 1980, 1987; Woolfolk, 1975). in gamma and beta frequencies meet the The majority of these EEG studies requirement for fast neural integration and focused on the change in the brain’s oscil- thus are thought to play a role in conscious latory rhythms, particularly in the slow fre- processes on the time scale of fractions of quencies (alpha and theta rhythms). It is a second (Tononi & Edelman, 1998; Varela, important to keep in mind that such mea- 1995). Fast-frequency synchronies have been sures reflect extremely blurred and crude found during such processes as attention, estimates of the synchronous processes of sensory segmentation, sensory perception, the ∼1011 neurons in a human brain. Because memory, and arousal. slow oscillations have high electrical voltages Yet neural synchrony must also be under- that make them visually detectable, early stood in the context of the slower alpha and studies only reported coarse visual descrip- theta bands (4–13 Hz), which play an impor- tions of EEGs. Changes in fast-frequency tant role in attention, working memory oscillations during meditation have been (Fries, Reynolds, Rorie, & Desimone, 2001; rarely reported (with the notable exception von Stein, Chiang, & Konig, 2000), and sen- of Das & Gastaut, 1955, and more recently sorimotor integration (Burgess & O’Keefe, Lutz et al., 2004) possibly because the lower 2003; Rizzuto et al., 2003). This evidence voltage of these oscillations requires spec- supports the general notion that neural syn- tral analysis instead of simple visual inspec- chronization subserves not simply the bind- tion. The investigation of fast-frequency syn- ing of sensory attributes, but the overall inte- chrony during meditation has become more gration of all dimensions of a cognitive act, common since the 1990s following a devel- including associative memory, affective tone oping understanding of its functional role in and emotional appraisal, and motor plan- the “binding problem.” ning. In addition to spectral analysis, medita- So far, oscillatory synchrony has been tion has also been characterized with mea- investigated mostly on oscillatory signals sures of coherence or long-distance phase having the same rhythms. More complex synchrony (LDS) (Fries et al., 2001). These non-linear forms of cross-band synchroniza- measures quantify the dynamic coupling tion, so-called generalized synchrony (Schiff, between EEG channels over distant brain So, Chang, Burke, & Sauer, 1996), are also regions. Coherence is the frequency correla- expected and may indeed prove more rele- tion coefficient, and it represents the degree vant in the long run to understanding large- to which the frequency profiles of two dis- scale integration than strict phase synchro- tant areas of the head, as reflected in the nization (Le Van Quyen, Chavez, Rudrauf, electrical signals detected by scalp elec- & Martinerie, 2003). trodes, are similar. LDS measures the instan- Considering the general importance of taneous phase relationship between signals neural synchrony in brain processing, it is not at a given frequency (Lachaux, Rodriguez, surprising that scientists interested in medi- Martinerie, & Varela, 1999). LDS provides a tation have tried to study its electrical brain more direct measure of phase-locking than correlates as early as the 1950s (Das & Gas- coherence because it can separate the effects P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

534 the cambridge handbook of consciousness

of amplitude and phase in the interrelations Alpha oscillations are found primarily over between signals. Thus, LDS can test more occipital-parietal channels particularly when precisely the assumption that phase syn- the eyes are closed, yet alpha activity can chrony is involved in long-distance neu- be recorded from nearly the entire upper ronal integration. Because of the non-linear cortical surface. During wakefulness, it is a nature of brain processes, these linear anal- basic EEG phenomenon that the alpha peak ysis approaches are likely to characterize reflects a tonic large-scale synchronization of only partially the functional properties of a very large population of neurons. This low- synchronous activity. Yet, complementary frequency global neural activity is thought to non-linear analysis of brain dynamics dur- be elicited by reciprocal interactions among ing meditation has only just started to be the cortex, the reticular nucleus, and the tha- explored (Aftanas & Golocheikine, 2002). lamocortical (Delmonte, 1985) cells in other In the selective summary of the liter- thalamic nuclei (Klimesch, 1999; Nunez et ature below, we review only those EEG al., 2001; Slotnick, Moo, Kraut, Lesser, & studies published in top-tier journals and/or Hart, 2002) even if cortico-cortical mech- those that focused on the study of long-term anisms also play a possible role (Lopes da practitioners. Silva, Vos, Mooibroek, & Van Rotterdam, 1980). transcendental meditation Because an overall decrease in alpha TM is a passive meditation adapted for West- power has been related to increasing deman- erners from the Vedic or Brahmanical tradi- ds of attention, alertness, and task load, alpha tions of India. The subject sits quietly, with activity is classically viewed as an “idling the eyes closed, repeating a Sanskrit sound rhythm” reflecting a relaxed, unoccupied (mantra), while concentrating on nothing brain (Klimesch, 1999). Large-scale alpha and letting the mind “drift” (Morse, Martin, synchronization blocks information process- Furst, & Dubin, 1977). The continued prac- ing because very large populations of neu- tice of TM supposedly is said to lead to an rons oscillate with the same phase and expansion of consciousness or the attain- frequency; thus, it is a state of high integra- ment of “cosmic” or “pure, self-referral con- tion but low differentiation. Within a band- sciousness” (Maharishi, 1969). The techni- width of perhaps 2 Hz near this spectral que is described as “easy, enjoyable and does peak, alpha frequencies frequently produce not involve concentration, contemplation or spontaneously moderate to large coherence any type of control” (R. K. Wallace, 1970). (0.3–0.8 over large interelectrode distance The standard EEG correlate of TM is (Nunez et al., 1997). The alpha coher- an increase in alpha rhythm amplitude, fre- ence values reported in TM studies, as a trait quently followed by a slowing in frequency in the baseline or during meditation, belong by 1–3 Hz and a spreading of this pattern into to this same range. Thus a global increase the frontal channels (R. K. Wallace, 1970). of alpha power and alpha coherence might An increase in bursts of theta oscillations not reflect a more “ordered” or “integrated” (4–7 Hz) has also been reported. Global experience, as frequently claimed in TM lit- fronto-central increases in coherence in erature, but rather a relaxed, inactive mental alpha (6–12 Hz), as well as in theta frequency state (Fenwick, 1987). ranges between baseline and TM practice, In contrast, alpha desynchronization have been found frequently (for reviews and reflects actual cognitive processes in which for a model of TM practice see Travis, Are- different neuronal networks start to oscillate nander, & DuBois, 2004)). locally at different frequencies – typically The dominant frequency in the scalp EEG in higher frequencies (>15 Hz), as well as of human adults is the alpha rhythm. It slower rhythms (4–15 Hz) – and with differ- is manifest by a peak in spectral analysis ent phases, reflecting local processing of spe- around 10 Hz and reflects rhythmic alpha cialized neuronal circuitries, such as those waves (Klimesch, 1999; Nunez et al., 2001). for attention, vision, memory, emotion, P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 535

and so on. Large-scale synchrony between tions might best be interpreted as meta- distant neuronal assemblies oscillating at var- physical assertions rather than first-person ious frequencies reflects an active coordina- descriptions, they do suggest that this state tion of functionally independent networks; of absorption could also involve some form in short, it reflects a state of high inte- of meta-awareness. Nevertheless, despite the gration and high differentiation. Thus, the possibility of a more sophisticated phe- slow frequency activity (<13 Hz) found dur- nomenological interpretation and the need ing TM meditation, combined with the fre- to relate physiological data to subjective quent finding of decreased autonomic activ- data, it is still unclear whether and how ity, has been interpreted by many authors TM meditation practices produce increased as reflecting hypoarousal or a relaxed state alpha activity beyond a general arousal effect (Delmonte, 1984; Holmes, 1984; Pagano & or an inhibition of task-irrelevant cortical Warrenburg, 1983). zones. Other relaxation techniques have led Yet, the “idling” model of alpha activity to the same EEG profile, and studies that has been extended recently to account for employed counter-balanced control relax- new findings. Alpha oscillation can, paradox- ation conditions consistently found a lack ically, increase locally over specific regions of alpha power increases or even decreases or also across specific areas while the subject when comparing relaxation or hypnosis to is actively focusing his or her attention on TM meditation (Morse et al., 1977; Tebe- an object or while holding in mind infor- cis, 1975; Warrenburg, Pagano, Woods, & mation (memory load during retention, for Hlastala, 1980). Similarly, the initial claim instance). Slow rhythms (4–12 Hz) can thus that TM produces a unique state of con- also be involved in active mental states sciousness different from sleep has been requiring attention, working memory, or refuted by several EEG meditation studies semantic encoding (Ward, 2003). This alpha that reported sleep-like stages during this model still remains compatible with the technique with increased alpha and then idling model because on this view, alpha theta power (Pagano, Rose, Stivers, & War- rhythms during mental activity reflect active renburg, 1976; Younger, Adriance, & Berger, inhibition of non-task-relevant cortical areas 1975). (Klimesch, 1999). To summarize, alpha global increases and Because TM is described as a passive med- alpha coherence mostly over frontal electro- itation without active control or concentra- des are associated with TM practice when tive effort, the EEG picture found during meditating compared to baseline (Morse, TM meditation can still be interpreted as Martin, Furst, & Dubin, 1977). This global reflecting mainly hypoarousal or a relaxed alpha increase is similar to that produced state. Yet, it is also possible that the ongoing by other relaxation techniques. The pass- repetition of the mantra, which involves, for ive absorption during the recitation of the instance, some form of attention and work- mantra, as practiced in this technique, pro- ing memory, can lead to an active exclusion duces a brain pattern that suggests a decrease of some brain processes compatible with an of processing of sensory or motor informa- increase in alpha activity in non-task-related tion and of mental activity in general. Be- cortical territories. cause alpha rhythms are ubiquitous and fun- TM researchers further view this EEG ctionally non-specific, the claim that alpha picture as reflecting a single and original oscillations and alpha coherence are desir- state of “Transcendental pure consciousness” able or are linked to an original and higher (Maharishi, 1969; Travis et al., 2004). The state of consciousness seem quite premature. transcendental state is conceptualized as a “fourth state of consciousness,” a “wake- attention meditation ful hypometabolic state” that differs from with an object hypnosis and ordinary or sleep states (R. This section regroups EEG studies on med- K. Wallace, 1970). Although these descrip- itative practices having a component of P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

536 the cambridge handbook of consciousness

attention regulation. In all these practices, subjects displayed large alpha activity during the intentional structure of a subject/object periods of rest and increases during medita- remains. As mentioned earlier these tech- tion (Anand et al., 1961). niques lie somewhere on a continuum Yet, several exceptions deserve scrutiny. between two poles of practices: On the one Two early field studies of Yoga in India by hand, one-pointed attention techniques cul- Das and Gastaut (1955) and Wenger and tivate a form of voluntary, effortful, and sus- Bagchi (1961), reported a clear sign of auto- tained attention on an object, and on the nomic arousal with increased heart rate and other hand, vipa´syan¯a meditation cultivates skin conductance when advanced yogis med- a more broadly focused, non-judgmental itate. High-amplitude high-frequency EEG mode of bare attention. These medita- oscillations (beta and gamma) were found tions differ from relaxation techniques and were more pronounced during deep because they cultivate a balance between meditation (Das & Gastaut, 1955). In a hypoarousal (Becker & Shapiro, 1981) and well-controlled study, Corby et al. (1978) excitation. This balance is required, in studied a form of Tantric Yoga meditation particular, to maintain a sufficient clarity where the practitioners and controls focused or meta-awareness throughout the medi- on their breath and on the mantra. Unlike tative session. These practices encompass, previously reported studies, proficient med- for instance, Zazen meditation, Indian yogic itators demonstrated increased autonomic concentration, meditation in MBSR, and activation during meditation, whereas unex- one-pointed focused attention. The empha- perienced meditators demonstrated auto- sis on stabilizing the mind on an object or nomic relaxation. During meditation, profi- on the awareness of the intentional rela- cient meditators showed an increase in alpha tion itself depends not only on the given and theta power, minimal evidence of EEG- technique but also likely on the degree of defined sleep, and a decrease in autonomic the practitioner’s accomplishment in a given orienting to external stimulation. practice. These findings are consistent with the With some important exceptions, most view described above that alpha and theta studies on Zazen or India yogic concentra- activation can also index attentional pro- tion practices have revealed an EEG signa- cesses. Because one major feature of atten- ture similar to TM as characterized by low- tion is selection, it is likely that the local- ered autonomic arousal and slow-frequency ized increases in slow frequencies reflect EEG patterns (either an increase in alpha or cortical tuning such that those cortical an increase in theta activity; Austin, 1998; zones that are not required for task engage- Delmonte, 1984; Fenwick, 1987; Shapiro & ment are selectively inhibited to facilitate Walsh, 1984). This pattern was reported as a task performance (see e.g., Cooper et al., state and sometimes also a trait. For instance, 2003). Also consistent with this formula- Kasamatsu and Hirai measured the EEG tion are data on attentional anticipation. signals of 48 priests and disciples during Foxe, Simpson, and Ahlfors (1998) demon- Zazen practices (Kasamatsu & Hirai, 1966). strated that a cue indicating an upcoming All subjects exhibited visually an increase auditory stimulus induced increased alpha in alpha activity mostly over central and power over parieto-occipital (Blake & Logo- frontal electrodes immediately after begin- thetis, 2002) cortex, compared when the ning meditation. Less experienced subjects cue indicated an upcoming visual stimu- tended to maintain high-amplitude alpha lus. These findings are all consistent with activity throughout the meditative session, the idea that alpha synchronization dur- whereas the EEGs of those with more years ing attentional processes reflects inhibition of Zazen practice showed a rhythmical theta of non-relevant areas or process (Klimesch, wave pattern during the later stage of Zazen. 1999). Anand, Chhina, and Singh (1961) visually It would be misguided to identify alpha or compared the EEG activity of four advanced theta activity as the sole index of mindful- yogis during rest and during meditation. All ness/awareness meditation. Numerous data P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 537

suggest that synchronized gamma activity is the specific neural processes that are altered also specifically involved in selective atten- by these practices. tion. In the literature on meditation, one Finally there is some evidence that early study on advanced yogic practition- alpha/theta oscillations during Zazen or ers reported spectacular generalized high- Samadhi practices differ functionally from amplitude beta/gamma oscillations during the alpha/theta activity during a relaxed intense internal concentration of attention non-meditative state. An early model of (Das & Gastaut, 1955). We also found an meditation proposed that “de-automization” increase in fast-frequency oscillations dur- was induced, such that each stimulus trial ing samatha practice (unpublished data). was perceived as “fresh” during medita- Numerous studies of humans, as well as ani- tive states cultivating a receptive and open mals, have demonstrated an enhancement awareness (Deikman, 1966; Kasamatsu & of the gamma activity when subjects were Hirai, 1966). A possible indication of this actively attending to a certain stimulus or process is EEG alpha blocking, which is simply perceived an object (Tallon-Baudry & defined as a decrease in ongoing alpha (8– Bertrand, 1999). Such synchronized gamma 12 Hz) power when comparing prestimu- activity during attention participates not lus to post-stimulus activity. Typically alpha only in bottom-up processes (e.g., sen- activity is reduced from closed eyes to open sory segmentation, feature extraction) but eyes or when discrete stimuli are presented also in top-down processes (Engel et al., and is thought to reflect cortical process- 2001). ing. This response habituates after repeti- Slow and fast-frequency rhythms inter- tive stimulus presentations (Morrell, 1966). act in the brain. For instance, in intracra- Early field studies on yogis reported no alpha nial recordings from area V4 in monkeys, blocking in response to auditory, thermal, increased gamma range synchronization but and visual stimuli (Anand et al., 1961; Das reduced low-frequency synchronization is & Gastaut, 1955; Wenger & Bagchi, 1961). observed among neurons activated by the Subsequent Zen studies found alpha block- attended stimulus as compared to neurons ing to auditory sounds but without habitua- activated by an identical but non-attended tion (Kasamatsu & Hirai, 1966). Early TM stimulus (Fries et al., 2001). These impor- studies produced conflicting results, with tant results, as well as event-related data, one finding an absence of alpha blocking lead to the notion of a surround inhibi- whereas the other reported habituation of tion wherein active cortical areas, indexed alpha blocking to auditory stimuli (Banquet, by alpha desynchronization and/or fast- 1973; R. K. Wallace, 1970). frequency synchronies, are surrounded by a A replication and extension of these find- “doughnut” of alpha synchronization or inhi- ings were attempted (Becker & Shapiro, bition (Suffczynski, Kalitzin, Pfurtscheller, 1981). Experienced Zen, Yoga, and TM med- & Lopes da Silva, 2001). This balance itators, and “attend” and “ignore” groups of between slow and fast frequencies can be controls were presented with auditory clicks detected under specific experimental condi- during mediation. The attend group was tions, such as intracranial recording or sim- asked to “pay strong attention” to each click, ple event-related tasks, over motor or sen- notice all of its sound qualities and subtleties, sory areas. Yet, this distinction is likely to be and count the number of clicks; the ignore blurred in general while recording ongoing group was told “try not to let the clicks dis- EEG signals because of volume conduction turb your relaxed state.” EEG alpha sup- (i.e., a single neural source is likely to influ- pression and skin conductance response both ence the signal in many recording channels). showed clear habituation, which did not dif- Despite this limitation, the combined char- fer among groups, thus failing to replicate acterization of fast-frequency synchronies, the earlier studies. in addition to the slow frequencies, over var- As a summary, these meditation prac- ious topographical regions of the scalp is tices that feature focused attention on likely to provide increased understanding of objects most frequently are accompanied by P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

538 the cambridge handbook of consciousness

increases in alpha and theta power, but also had a peak-to-peak amplitude of the order by fast frequencies (beta and gamma) during of dozens of microvolts for several prac- deep meditation. The slow-frequency activ- titioners. High-amplitude oscillations were ity overlaps notably with early drowsiness continuous during the meditation over sev- and sleep stages even if these oscillations eral dozens of seconds and gradually increase potentially differ functionally. The neuro- during the practices. These EEG patterns electric signatures of these various medi- differ from those of controls, in particular tative techniques (Focus Attention, Zazen, over lateral fronto-parietal electrodes. Some Vipasyan´ a¯ meditation) have not yet been preliminary data further suggest that these firmly established. Our current understand- ongoing high-amplitude gamma oscillations ing of attention suggests that the selection are correlated with self-reports of the clarity or the exclusion from attention of particu- (see the section on samatha and vipa´syan¯a lar contents (sensory, motor, internal tasks) as paradigms of meditation; Lutz et al., is correlated with the activation or inhi- 2005). bition of specific brain areas, as indexed These new findings are similar to the early by specific changes in selective brain oscil- report of Das and Gastaut (1955) during the latory patterns. The combination of topo- Samadhi of advanced Indian Yogis. Samadhi graphical information with spectral informa- was defined as a state during which “the tion seems a promising method by which perfectly motionless subject is insensible to to delineate further these various meditative all that surrounds him and is conscious of techniques. nothing but the subject of his meditation.” Das and Gastaut (1955) reported an accel- objectless meditation eration of the cardiac rhythm during med- During objectless meditation, such as Open itation almost perfectly parallel to that of Presence or Non-Referential Compassion the EEG. The EEG showed progressive and meditation, it is said that the practitioner spectacular modifications during the deep- does not focus on a particular object but est meditations in those subjects who had rather cultivates a state of being. Object- the longest training: acceleration of the alpha less meditation does so in such a way that, rhythm and decrease in the amplitude and according to reports given after meditation, appearance of faster oscillations (>20 Hz). the intentional or object-oriented aspect of These fast frequencies (beta (25–30 Hz) experience appears to dissipate in medita- and sometimes even gamma activity (40–45 tion along with the explicit sense of being a Hz) became generalized during the Samadhi perceiver or an agent (autobiographical self). meditation, with high amplitude reaching One working hypothesis is that some form of between 30–50 mV. meta-awareness or, more precisely, of some In our study (Lutz et al., 2004), we fur- mere ipseity still remains or is enhanced dur- ther showed that during this objectless med- ing these states. itation the ratio of fast-frequency activity These types of meditation have been (25–42 Hz) to slow oscillatory activity (4–13 poorly investigated so far. We studied a Hz) over medial fronto-parietal electrodes is group of long-term practitioners who under- initially higher in the resting baseline before went mental training in the same Tibetan meditation for the practitioners than the Nyingmapa and Kargyupa traditions for controls (Figure 19.4). During meditation, 10,000 to 50,000 hours over time periods this difference increases sharply over most of ranging from 15 to 40 years. We found the scalp electrodes and remains higher than that these long-term Buddhist practitioners the initial baseline in the post-meditative self-induced sustained EEG high-amplitude baseline. The functional and behavioral con- gamma-band oscillations and phase syn- sequences of sustained gamma activity dur- chrony during Non-Referential Compassion ing objectless meditation are not currently meditation (Lutz et al., 2004, Figure 19.5). known, and such effects clearly need further These fast-frequency oscillations (>20 Hz) study. P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 539

a μV 20 F3 −20 20 Fc5 −20 20 Cp5 −20 20 F4 −20 20 Fc6 −20 20 Cp6 −20 b μV2 500 300 100 c % 14 12 10 0 50 100 150 Time (s)

Resting state Meditative state Block colors: (1) (2) (3) (4) Figure 19.5. Example of high-amplitude gamma activity during the non-referential compassion meditation of long-term Buddhist practitioners. a. Raw electroencephalographic signals. At t = 45 s, practitioner S4 started generating a state of non-referential compassion, block 1. b. Time course of gamma activity power over the electrodes displayed in “a” during four blocks computed in a 20-s sliding window every 2 s and then averaged over electrodes. c. Time course of their cross-hemisphere synchrony between 25–42 Hz. The density of long-distance synchrony above a surrogate threshold was calculated in a 20-ssliding window every 2 s and for each cross-hemisphere electrode pairs and, then, was averaged across electrode pairs (adapted from Lutz et al. 2004).

Neuroimaging Correlates of Meditation neural basis of the peak experience of a meditative state termed “kensho” or “satori” At this stage neuroimaging studies on in Japanese Zen Buddhism. In this state, meditation are typically more exploratory the sense of selfhood is allegedly dissolved than hypothesis driven. Nevertheless, some and an “unattached, self-less, impersonal” progress has been made in the identiﬁca- awareness remains (this state shares a strong, tion of structural-functional brain relation- descriptive similarity with the Open Pres- ships of meditative states and traits using a ence state discussed above). After examin- variety of neuroimaging modalities. In paring the precise experiential changes induced ticular, some theoretical efforts have been by this state, he reviewed the various phys- made to localize the neural circuitry selec- iological subsystems that might participate tively engaged during a meditative state. in this state. Austin speciﬁcally introduces Austin (1998), for instance, elegantly com- the distinction between “egocentric” neural bined his insight as a Zen practitioner with networks involved in the generation of a his neuroanatomical knowledge of the brain multifaceted self situated in time and space as a medical doctor to speculate about the and “allocentric” neural networks involved P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

540 the cambridge handbook of consciousness

in the mere processing of the external envi- istry of the working brain. SPECT (single ronment. For Austin, neural networks par- photon emission computed tomography) is ticipating in the construction of the narra- another neuroimaging method that is similar tive self could be shut down during Kensho, to, though less sophisticated than, PET, and specifically through thalamic gating origi- it produces images of neurochemical func- nating from the reticular formation. At the tion that have less spatial resolution than same time, he proposes that the state of PET. MRI uses magnetic fields and radio hyperawareness during this practice is medi- waves to produce high-quality two- or three ated by intralaminar nuclei of the thala- dimensional images of brain structures with- mus that can increase the fast-frequency syn- out injecting radioactive tracers. Using MRI, chrony in other cortical regions. These nuclei scientists can see both surface and deep brain could shape the resonance of the cortico- structures with a high degree of anatomi- thalamo-cortical loops and functionally alter cal detail (millimeter resolution). MRI tech- the neural processing in these egocen- niques can also be used to image the brain as tric/allocentric networks. These proposals it functions. (Austin, 1998, 2006) are clearly specula- Functional MRI (fMRI) relies on the mag- tive, and further discussion is beyond the netic properties of blood to enable the scope of this review. Nevertheless, Austin’s researcher to measure the blood flow in the work amply illustrates the potential bene- brain as it changes dynamically in real time. fits that may come when the neuroscience of Thus researchers can make maps of changes meditation and first-person descriptions are in brain activity as participants perform var- brought into a dynamic dialogue that com- ious tasks or are exposed to various stim- bines their findings in a manner that places uli. An fMRI scan can produce images of fruitful constraints on each. brain activity as fast as every second or two, Although there is considerable potential whereas PET usually takes several dozens for advancement in neuroscience through of seconds to image brain function. Thus, neuroimaging studies of meditation, the with fMRI, scientists can determine pre- number of published studies remains sparse. cisely when brain regions become active and To illustrate the range of methods and ques- how long they remain active. As a result, tions utilized thus far, we now review briefly they can see whether brain activity occurs the published research in this area. simultaneously or sequentially in different brain regions as a participant thinks, feels, or brain imaging techniques used reacts to external stimuli. An fMRI scan can in meditation research also produce high-quality images that can Positron emission tomography (PET; Blake identify more accurately than PET which & Logothetis, 2002) and functional magnetic areas of the brain are being activated. In sum- resonance imaging (fMRI) are two func- mary, fMRI offers better image clarity along tional brain imaging methods that have been with the ability to assess blood flow and brain used to study meditation. PET measures function in seconds. So far, however, PET emissions from radioactively labeled chemi- retains the advantage of being able to pin- cals that have been injected into the blood- point which neurochemicals are involved in stream and uses the data to produce two- or functional brain alterations. three-dimensional images of the distribution of the chemicals throughout the brain and early neuroimaging studies on body. Using different tracers, PET can reveal relaxation practice and meditation blood flow, oxygen and glucose metabolism, The studies from Lou et al. (1999) and and neurotransmitter concentrations in the Newberg et al. (2001) provide the first evi- tissues of the working brain. Blood flow and dence of functional brain changes using PET oxygen and glucose metabolism reflect the or SPECT during a relaxation practice and amount of brain activity in different regions, a meditative practice, respectively. Even if and this type of data enables scientists to these studies offer new insights about these characterize the physiology and neurochem- states, they speak also for the need to more P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 541

precisely develop descriptions of the prac- role not only in motivation and resolution tices to better understand just what the func- of conflict but also skeleto-motor control tional neural changes are reflecting. and executive attention, and the cerebellum Yoga Nidra,¯ literally “Yoga-Sleep,” is a is implicated in cognitive functions such as state in the Yoga tradition in which con- attention. sciousness of the world and consciousness Lou et al. (1999) interpreted these results of action are meant to be dissociated: The as reflecting dissociation between two com- mind “withdraws” from wishing to act and is plementary aspects of consciousness: the not associated with emotions or the power conscious experience of the sensory world of will. The practitioner allegedly becomes and the “fact or illusion of voluntary control, a neutral observer who experiences the loss with self regulation.” Unfortunately, the lack of conscious control, concentration, or judg- of a control population makes it difficult to ment, yet maintains an equal and impar- interpret whether the brain patterns reflect tial attention to sensory awareness, which specific meditative qualities or the cognitive 15 is said to be enhanced. A PET ( O-H2 O) processes induced by the instructions. study of blood flow changes during Yoga Using SPECT Newberg et al. (2001) Nidra¯ practice was carried out while subjects measured changes in regional blow flow listened to a tape recording, with guided (rCBF) while eight relatively experienced instructions on the different phases of the Tibetan Buddhist practitioners meditated. practice (Lou et al., 1999). The relaxation The practitioners practiced daily for at least tape contained focusing exercises on body 15 years and underwent several 3-month sensation, abstract joy, visual imagery of a retreats. In the scanner, the practitioners summer landscape, and symbolic represen- were instructed to “focus their attention on a tation of the self. Participants listened to visualized image and maintained that focus the tape and followed the instructions of with increasing intensity.” In constrast to Lou the guided meditation. The baseline con- et al. (1999), Newberg and colleagues (2001) dition was obtained by replaying the tape reported an increase in activity in orbital while participants remained neutral (i.e., frontal cortex, dorsolateral prefrontal cortex they did not follow the instructions). Each (DLPFC), and thalamus. They also found of the guided meditation phases was associ- a negative correlation between the DLPFC ated with different regional activations dur- and the superior parietal lobe, which was ing meditation relative to the control con- interpreted as reflecting an altered sense of ditions. Yet, during all meditative phases, space experienced during meditation. The overall increases in bilateral hippocampus, difference in the frontal areas between their parietal, and occipital sensory and associa- finding and that of Lou et al. (1999)was tion regions were found compared to control viewed as reflecting a difference between an conditions. This pattern suggests an increase active and a passive form of meditation. of activity in areas involved in imagery. In addition to the fact that no control Deactivation was found during medita- participants were involved in the Newberg tion in orbitofrontal, dorsolateral prefrontal, study, there is regrettably a lack of descrip- anterior cingulate cortices, temporal and tive precision of the meditative state that inferior parietal lobes, caudate, thalamus, was studied. This limitation will hamper the pons, and cerebellum. This differential activ- future comparison of this study with others. ity was interpreted as reflecting a “tonic” More precisely, a broad variety of Tibetan activity during normal consciousness in the meditative techniques could encompass the baseline condition. The areas decreasing dur- provided meditative descriptions. These ing the meditation state are known to par- practices include, for instance, Focused ticipate in executive function or control Attention on a mental object, or any medita- of attention. More particularly, dorsolateral tion on the visualization of a deity, or indeed prefrontal cortex participates in working the visualization of one’s guru. Unfortu- memory and the preparation for voluntary nately, these practices can differ or even movement, the anterior cingulate plays a be opposite in terms of their motivations P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

542 the cambridge handbook of consciousness

or emotional qualities. For instance, the visu- became distracted or sleepy. Control sub- alization of deities could involve some invo- jects with no prior meditative training were cation of anger or lust, whereas the visu- given instruction in concentration medita- alization of the guru is meant to induce tion with daily practice a week before the a strong devotional affect in the medita- fMRI scan. fMRI of concentration meditator. Because the independent variable (i.e., tion in both the experienced meditators and the specific meditative practice) was only the controls showed common areas of acti- vaguely described in this study, its impact vation in the traditional attention network, is limited. including such areas as the intraparietal sulci, frontal eye fields (FEF), thalamus, insula, focused attention/mindfulness- lateral occipital, and basal ganglia. How- awareness meditation ever, experienced meditators showed more A form of Kundalini Yoga using mantra rep- activation, especially in the frontal-parietal etition combined with breath awareness was network. The increased activation in these assessed with fMRI (Lazar et al., 2000). The regions for experienced practitioners may control state entailed the mental enuncia- represent a neural correlate for these sub- tion of animal names. Five Yoga adepts who jects’ expertise in sustained attention. The had practiced Kundalini Yoga for at least fact that controls show greater activation in 4 years served as subjects. An increase in the anterior cingulate compared with the the Blood Oxygenation Level Dependent adepts may reflect greater error proneness (BOLD) signal was found from baseline to (i.e., distraction) and conflict monitoring in meditation in the putamen, midbrain, ante- the controls than the adepts; the conflict rior cingulate cortex, and the hippocam- would be between the instructions to focus pal/parahippocampal formation, as well as and the difficulty of complying with such in regions in the frontal and parietal cor- instructions. tices. The comparison of early versus late Taken together these two brain imaging meditation states showed activity increase studies show that concentration meditation in these regions, but within a greater area enhances processing in regions similar to and with larger signal changes later in the those found in other attentional paradigms. practice. Because the pattern of brain activ- The group differences between long-term ity increased with meditation time, it may practitioners and novices support the view index the gradual changes induced by medi- that attention processing could be affected tation. This pattern of activity encompassed by mental training. areas subserving attention (fronto-parietal cortices) and areas subserving arousal and pure compassion and autonomic control (limbic regions, mid- lovingkindness meditation brain, and anterior cingulate cortex). Using functional imaging, we assessed In another attention-related study, we brain activity while novice and long-term recently studied experienced Buddhist med- practitioners generated a Lovingkindness- itators (>10,000 hours of cumulative med- Compassion meditation, alternating with itation practice) and newly trained control a resting state (Brefczynski-Lewiset et al., subjects while they performed a Focused 2004). As described in the section on Non- Attention meditation (Tsé-cig Ting-ngé-dzin; Referential Compassion meditation, this see the section on Focused Attention), alter- standard Buddhist meditation involves the nating with a passive state, while undergo- generation of a state in which an uncon- ing block-design fMRI (Brefczynski-Lewis, ditional feeling of lovingkindness and com- Lutz, & Davidson, 2004). During this stan- passion pervades the whole mind as a dard meditation, the participants concen- way of being, with no other considera- trated their attention on an external visual tion, reasoning, or discursive thoughts. This object (a white dot on the screen), gently state is called in Tibetan “pure” or “non- bringing attention back to the object if they referential” compassion, as the practitioner P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 543

is not focused upon particular objects dur- hence, a common view is that the very ing this state. In the resting state the subjects regions subserving one’s own feeling states were asked to be in the most ordinary state also instantiate one’s empathic experience without being engaged in an active mental of other’s feelings. This framework derives state or being in a pleasant or unpleasant from perception-action models of motor emotional state. Subjects were eight long- behavior and imitation. The key proposal term Buddhist practitioners and eight age- is that the observation and imagination of matched healthy control volunteers who another person in a particular emotional were interested in learning to meditate. state automatically activate a similar affec- Buddhist practitioners underwent mental tive state in the observer, with its associated training in the Tibetan Nyingmapa and Kar- autonomic and somatic responses. Thus, gyupa traditions for 10,000 to 50,000 hours experienced and empathic pain commonly over time periods ranging from 15 to 40 activated the anterior insula and rostral ante- years. During the meditative state, we found rior cingulate cortex (Singer et al., 2004). a common activation in the striatum, ante- The activation in the anterior insula was rior insula, somatosensory cortex, anterior stronger for the practitioners, an area that cingulate cortex, and left-prefontal cortex some scientists have found to be involved in and a deactivation in the right interior pari- feelings. These data are consistent with the etal. This pattern was robustly modulated view that our experience of another’s suf- by the degree of expertise, with the adepts fering is mediated by the same brain regions showing considerably more enhanced acti- involved in the experience of our own pain. vation in this network compared with the We further found that brain activity for novices. the long-term practitioners was greater than These data provide evidence that this the novices in several of the commonly acti- altruistic state involved a specific matrix of vated regions. These analyses indicate that brain regions that are commonly linked to the degree of training, as reflected in the feeling states, planning of movements, and hours of cumulative meditation experience, positive emotions. Maternal and romantic modulates the amplitude of activation in the love have been linked in humans to the acti- brain areas commonly involved in this state. vation of the reward and attachment cir- To summarize, our study of Compas- cuitries, such as the substantia nigra and sion meditation found activation in brain the striatum (caudate nucleus, putamen, regions thought to be responsible for moni- globus pallidus; Bartels & Zeki, 2004). Pos- toring one’s feeling state, planning of move- itive and negative emotions are expected to ments, and positive emotions. This pattern differentially activate the left and right pre- was robustly modulated by the degree of frontal cortices, respectively, as suggested by expertise. These data suggest that emotional lesion and electrophysiological data (David- and empathic processes are flexible skills son, 2000). More generally, feeling states that can be enhanced by training and that are thought to be mediated by structures such training is accompanied by demonstra- that receive inputs regarding the internal ble neural changes. milieu and musculoskeletal structures and include the brainstem tegmentum, hypothalamus, insula, and somatosensory and cin- General Conclusion gulate cortices (Damasio, 1999). This view has received some neuroimaging support in Overall, this chapter aimed to summa- a task where subjects self-generate emo- rize the state of knowledge in neuro- tional states and more recently in studies scientific research on meditation and to using pain experience or interoceptive tasks suggest potential avenues of inquiry illu- (Craig, 2002). minated by these initial findings. The first Finally, love and compassion require an section discussed the need for more pre- understanding of the feelings of others; cise descriptions of meditative practices P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

544 the cambridge handbook of consciousness

so as to define properly the practices that the neural counterpart of subjective expe- are the objects of scientific study. Fol- rience. These intersections between neuro- lowing this recommendation, the Buddhist science and meditation were separated here contemplative tradition was presented in mainly for analytical purposes, but these detail as a canonical example. The main Bud- heuristic distinctions implicitly suggest an dhist theories of meditation were reviewed important area of further research; namely, as well as the basic parameters that define the interactions among the various themes of most forms of Buddhist contemplative prac- research. For instance, one question of inter- tice. In addition to suggesting an approach est will be to explore whether it is mean- to defining and categorizing meditation, this ingful to study the alleged therapeutic or section also aimed to underscore the diffi- healing virtues of meditation as a variable of culty of separating well-defined first-person research in isolation from other issues. The descriptions of meditative states from other interest in this question stems from the pos- claims that, although apparently descriptive, sibility that the beneficial changes found in are best understood as reflecting particu- practitioners of meditation are intrinsically lar cultural or religious exigencies that are dependent on other practices or virtues cul- not strictly rooted in scientifically tractable tivated in their tradition, such as compas- observations. The choice to view a Bud- sion, ethical behavior, or a first-person explo- dhist claim as a first-person description of ration of the nature of the self and external an actual state or as primarily a prod- perception. Having suggested, in any case, uct of some religious and cultural rhetoric the potentially fruitful exploration of medi- is certainly subject to debate and inter- tation from a neuroscientific perspective, in pretation. Further developments will def- the final section, we reviewed the most rel- initely be needed to delineate these dis- evant neuroelectric and neuroimaging find- tinctions. With these difficulties in mind, ings of research conducted to date. We antic- three standard Buddhist meditative states ipate that the renewed interest in research were described in detail, as well as the on meditation will probably extend and pos- rationale for the cultivation of these states sibly modify this section within the near and the expected post-meditative effects. future. Some general guidelines were then pro- As noted earlier, we chose to emphasize posed for developing a questionnaire to the practice of long-term Buddhist practi- define more precisely a practice under exam- tioners, in part because of the potential that ination. It is our hope that this first sec- a study of such practitioners might have tion will provide researchers with some to enhance our understanding of conscious- theoretical and methodological principles ness. Already we have some indication that to clarify and enhance future research on experienced practitioners are able to pro- meditation. vide repeatable subjective reports that are The second section explored some sci- more reliable than those from untrained per- entific motivations for the neuroscientific sons, and this opens the door to wide-ranging examination of meditation in terms of its research into the neural correlates of those potential impact on the brain and body of reportable states. More particularly, the pos- long-term practitioners or its possible role in sibility that some meditators may be able the neuroscientific study of subjective expe- to induce a state approaching some form of rience. After an overview of the mechanisms bare consciousness or ipseity raises the tan- of neuroplasticity and mind-body interac- talizing (if contentious) hypothesis that the tion, we argued that mental training might neural correlates of such a state would bring have a long-term impact on the brain and us closer to understanding what we mean body in a way that is beneficial for physical by consciousness from a neuroscientific per- health, illness, and possibly well-being. We spective. then suggested how the use of first-person Our decision to focus on long-term Bud- expertise might foster our understanding of dhist practitioners, however, should not P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 545

diminish the importance of future research 3. To facilitate further inquiry by readers unfa- on novices, of longitudinal studies of changes miliar with the relevant Asian languages, only over time in novice or mid-range practi- sources available in English have been used to tioners, or of research involving other con- present the pertinent Buddhist theories and templative traditions. This point is crucial practices. It is important to note, however, that if one believes that some of these medita- many of the most relevant Tibetan texts in particular have yet to be translated reliably into tive practices have the potential to evolve any European language. into a more secular form of mental train- 4. Gethin (1998) provides an excellent overview ing, with alleged therapeutic, pedagogical, of the Abhidharma and its context. It is impor- and/or health value. Most importantly, the tant to note that the two classical South collective evidence showcased in this review Asian languages most relevant to the his- underscores the fact that many of our core tory of living Buddhist traditions are San- mental processes, such as awareness and skrit and Pali.¯ Sanskrit is relevant especially attention and emotion regulation, including to Tibetan, Chinese, Japanese, and Korean our very capacity for happiness and com- Buddhism. Pali¯ is still a scholarly language passion, should best be conceptualized as of the Theravada¯ Buddhist traditions that are trainable skills. The meditative traditions active, especially in Sr´ ¯ıLanka,˙ Thailand, and provide a compelling example of strategies Myanmar. For consistency, we have used San- and techniques that have evolved over time skrit for technical terms that apply generally to Buddhist traditions, but some academic to enhance and optimize human potential sources will favor the Pali¯ equivalents. In such and well-being. The neuroscientific study of sources, Abhidharma would be rendered as these traditions is still in its infancy, but Abhidhamma. the early findings promise both to reveal 5. In English, the term “lovingkindness” is often the mechanisms by which such training may used in lieu of “compassion” because it more exert its effects and underscore the plastic- accurately translates the Sanskrit compund, ity of the brain circuits that underlie com- maitr¯ıkarun. a¯. This compound consists of two plex mental functions. It is our fervent hope terms: maitr¯ı, translated as “loving,” is defined that this review will stimulate additional as the aspiration for another to be happy, and research and will lead to the increased use of karun. a¯, translated as “kindness,” is defined as these practices in a wide range of everyday the aspiration that another be free of suf- contexts. fering. The term karun. a¯ is also translated as “compassion,” and in Tibetan it is rendered as snying rje, the term that occurs in “non- Acknowledgments referential compassion” (dmigs med snying rje; Skt., niralambanakarun. a)¯ . Nevertheless, even though the most accurate translation of Support for writing this chapter and the this compound should include only the word research from the authors’ lab that is “compassion,” the actual practice of generating reported herein was provided by NIMH this state involves both love and compassion; P50-MH069315 to RJD, gifts from Adrianne that is, both maitr¯ı and karun. a.¯ and Edwin Cook-Ryder and from Bryant Wangard, NCCAM U01AT002114-01A1 and the Fyssen Foundation to A. L. References

Aftanas, L. I., & Golocheikine, S. A. (2002). Notes Non-linear dynamic complexity of the human EEG during meditation. Neuroscience Letters, 1. Number of articles indexing the term “medi- 330(2), 143–146. tation” in Medline in 2005. Anand, B., Chhina, G., & Singh, B. (1961). Some 2. For a fruitful and pragmatic development of aspects of electroencephalographic studies in this hypothesis see Depraz, Varela, & Vermer- yogis. Electroencephalography and Clinical Neu- sch (2003). rophysiology, 13, 452–456. P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

546 the cambridge handbook of consciousness

Andresen, J. (2000). Meditation meets behav- Bronkhorst, J. (1986). The two traditions of medi- ioural medicine: The story of experimental tation in ancient India (28th ed.). Stuttgart: F. research on meditation. Journal of Conscious- Steiner Verlag Wiesbaden. ness Studies, 7, 17–73. Burgess, N., & O’Keefe, J. (2003). Neural repre- Apple, J. (2003). Twenty varieties of the Samgha: sentations in human spatial memory. Trends in A typology of noble beings (arya)¯ in Indo- Cognitive Science, 7(12), 517–519. Tibetan scholasticism. Part 1. Journal of Indian Cahn, R., & Polich, J. (2006). Meditation states Philosophy, 31(5–6), 503–592. and traits: EEG, ERP, and neuroimaging stud- Austin, J. H. (1998). Zen and the brain: Toward ies. Psychological Bulletin, 132, 180–211. an understanding of meditation and conscious- Carter, O. L., Presti, D. E., Callistemon, C., ness (2nd ed.). Cambridge, MA: MIT Press. Ungerer, Y., Liu, G. B., & Pettigrew, J. D. Austin, J. H. (2006). Zen-brain reﬂections. Review- (2005). Meditation alters perceptual rivalry ing recent developments in meditation and con- in Tibetan Buddhist monks. Current Biology, sciousness. Cambridge, MA: MIT Press. 15(11), R412–413. Banquet, J. P. (1973). Spectral analysis of the EEG Clark, D. D., & Sokoloff, L. (1999) Circulation in meditation. Electroencephalography & Clini- and energy metabolism of the brain. In G. J. cal Neurophysiology, 35(2), 143–151. Siegel, B. W. Agranoff, R. W. Albers, S. K. Barinaga, M. (2003). Buddhism and neuro- Fisher, & M. D. Uhler (Eds.), Basic neurochem- science. Studying the well-trained mind. Sci- istry. Molecular, cellular and medical aspects. ence, 302(5642), 44–46. Philadelphia: Lippincott-Raven. 2002 Bartels, A., & Zeki, S. (2004). The neural corre- Coleman, J. W. ( ). The new Buddhism: The lates of maternal and romantic love. Neuroim- Western transformation of an ancient tradition. age, 21(3), 1155–1166. New York: Oxford University Press. Becker, D. E., & Shapiro, D. (1981). Physiological Cooper, N. R., Croft, R. J., Dominey, S. J., 2003 responses to clicks during Zen, Yoga, and TM Burgess, A. P., & Gruzelier, J. H. ( ). Para- meditation. Psychophysiology, 18(6), 694–699. dox lost? Exploring the role of alpha oscillations during externally vs. internally directed Benson, H., Lehmann, J. W., Malhotra, M. S., attention and the implications for idling and Goldman, R. F., Hopkins, J., & Epstein, M. D. inhibition hypotheses. International Journal of (1982). Body temperature changes during the Psychophysiology, 47(1), 65–74. practice of Tum-mo yoga. Nature, 295(5846), 234–236. Corby, J. C., Roth, W. T., Zarcone, V. P., Jr., & Kopell, B. S. (1978). Psychophysiological corre- Berger, H. (1929). Uber das Elektrenke- lates of the practice of Tantric Yoga meditation. laphogramm des Menschen. Archiv fur ¨ Psy- Archives of General Psychiatry, 35, 571–577. chiatrie und Nervenkrankheiten, 87, 527–570. Cosmelli, D., David, O., Lachaux, J. P., Mar- Beyer, S. (1977). Notes on the vision quest tinerie, J., Garnero, L., Renault, B., & Varela, in early Mahay¯ ana.¯ In L. Lancaster (Ed.), F. (2004). Waves of consciousness: Ongo- Praj˜n¯ap¯aramit¯a and related systems (pp. 329– ing cortical patterns during binocular rivalry. 340). Berkeley, CA: University of California Neuroimage, 23(1), 128–140. Press. Cozort, D. (1986). Highest yoga tantra: An intro- Bibbig, A., Traub, R. D., & Whittington, M. A. duction to the esoteric Buddhism of Tibet. Ithaca, (2002). Long-range synchronization of gamma NY: Snow Lion Publications. and beta oscillations and the plasticity of exci- 2002 tatory and inhibitory synapses: A network Craig, A. D. ( ). How do you feel? model. Journal of Neurophysiology, 88(4), 1634– Interoception: The sense of the physiological 1654. condition of the body. Nature Reviews Neuro- science, 3(8), 655–666. Blake, R., & Logothetis, N. K. (2002). Visual com- = petition. Nature Reviews Neuroscience, 3(1), 13– Dalai Lama XIV [ Bstan-’dzin-rgya-mtsho]. 1991 21. ( ). Path to bliss. Ithaca, NY: Snow Lion Publications. Brefczynski-Lewis, J. A., Lutz, A., & Davidson, R. = J. (2004). A neural correlate of attentional exper- Dalai Lama XIV [ Bstan-’dzin-rgya-mtsho]. 1995 tise in long-time Buddhist practitioners. (Report ( ). The world of Tibetan Buddhism: An no. 715.8.). San Diego: Society for Neuro- overview of its philosophy and practice. Boston: science. Wisdom Publications. P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 547

Damasio, A. R. (1999). The feeling of what Fischer, R. (1971). A cartography of the ecstatic happens: Body and emotion in the making of con- and meditative states. Science, 174(12), 897– sciousness. New York: Harcourt Brace. 904. Das, N. N., & Gastaut, H. C. (1955). Variations de Foxe, J. J., Simpson, G. V., & Ahlfors, S. P. (1998). l’activite electrique du cerveau, du coeur et des Parieto-occipital approximately 10 Hz activ- muscles squelletiques au cours de la medita- ity reﬂects anticipatory state of visual attention et de l’extase yogique. Electroencephalogra- tion mechanisms. Neuroreport, 9(17), 3929– phy & Clinical Neurophysiology, 6(suppl.), 211– 3933. 219. Freeman, W. (1999). Consciousness, intentional- Davidson J. M. (1976). The physiology of med- ity, and causality. Journal of Consciousness Stud- itation and mystical states of consciousness. ies, 6, 143–172. Perspect Biol Med. 19(3):345–79. Fries, P., Reynolds, J. H., Rorie, A. E., & Des- Davidson, R. J. (2000). Affective style, psy- imone, R. (2001). Modulation of oscillatory chopathology, and resilience: Brain mecha- neuronal synchronization by selective visual nisms and plasticity. American Psychologist, attention [see comment]. Science, 291(5508), 55(11), 1196–1214. 1560–1563. Davidson, R. J., Kabat-Zinn, J., Schumacher, J., Fries, P., Roelfsema, P. R., Engel, A. K., Konig, P., Rosenkranz, M., Muller, D., Santorelli, S. F., & Singer, W. (1997). Synchronization of oscil- Urbanowski, F., Harrington, A., Bonus, K., latory responses in visual cortex correlates with & Sheridan, J. F. (2003). Alterations in brain perception in interocular rivalry. Proceedings of and immune function produced by mindful- the National Academy of Sciences USA, 94(23), ness meditation [see comment]. Psychosomatic 12699–12704. Medicine, 65(4), 564–570. Gethin, R. (1998). The foundations of Buddhism. Dehaene, S., & Naccache, L. (2001). Towards a Oxford: Oxford University Press. cognitive neuroscience of consciousness: Basic Glaser, D. (2000). Child abuse and neglect and evidence and a workspace framework. Cogni- the brain – a review. Journal of Child Psychol- tion, 79(1–2), 1–37. ogy & Psychiatry & Allied Disciplines, 41(1), 97– Deikman, A. J. (1966). Implication of experimen- 116. tally induced contemplative meditation. Jour- Grossman, P., Niemann, L., Schmidt, S., & nal of Nervous and Mental Disease. 142(2):101– Walach, H. (2004). Mindfulness-based stress 16. reduction and health beneﬁts. A meta-analysis. Delmonte, M. M. (1984). Electrocortical activity Journal of Psychosomatic Research, 57(1), 35– and related phenomena associated with med- 43. itation practice: A literature review. Interna- Gunaratana, H. (2002). Mindfulness in plain tional Journal of Neuroscience, 24, 217–231. English. Boston: Wisdom Publications. Delmonte, M. M. (1985). Biochemical indices Gusnard, D. A., & Raichle, M. E. (2001). Search- associated with meditation practice: A lit- ing for a baseline: Functional imaging and the erature review. Neuroscience & Biobehavioral resting human brain. Nature Reviews Neuro- Reviews, 9(4), 557–561. science, 2 (10), 685–694. Depraz, N., Varela, J. F., & Vermersch, P. (2003). Holmes, D. S. (1984). Meditation and somatic On becoming aware: A pragmatics of experienc- arousal reduction. A review of the experimen- ing. Amsterdam: John Benjamins. tal evidence. American Psychologist, 39, 1–10. Engel, A. K., Fries, P., & Singer, W. (2001). Jayaraman, K. S. (2003). Technology, tradition Dynamic predictions: Oscillations and syn- unite in India’s drug discovery scheme. Nature chrony in top-down processing. Nature Reviews Medicine, 9(8), 982. 10 704 716 Neuroscience, 2 ( ), – . Kabat-Zinn, J. (2005). Wherever you go, there English, E. (2002). Vajrayogin¯ı, a study of her visu- you are: Mindfulness meditation in everyday life. alizations, rituals & forms: A study of the cult of New York: Hyperion. Vajrayogin¯ı in India. Boston: Wisdom Publica- Kabat-Zinn, J., & Chapman-Waldrop, A. (1988). tions. Compliance with an outpatient stress reduc- 1987 Fenwick, P. B. ( ). Meditation and the EEG. tion program: Rates and predictors of program In A. West (Ed.), The psychology of meditation completion. Journal of Behavioral Medicine, 104 117 (pp. – ). New York: Clarendon Press. 11(4), 333–352. P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

548 the cambridge handbook of consciousness

Kabat-Zinn, Lipworth, & Burney (1985). The ear dynamics of the brain. Journal of Physiology, clinical use of mindfulness meditation for the Paris, 97(4–6), 629–639. self-regulation of chronic pain. J Behav Med. Llinas, R., Ribary, U., Contreras, D., & 8(2):163–90. Pedroarena, C. (1998). The neuronal basis for Kabat-Zinn, J., Massion, A. O., Kristeller, J., Peter- consciousness. Philosophical Transactions of the son, L. G., Fletcher, K. E., Pbert, L., Lenderk- Royal Society of London – Series B: Biological ing, W. R., & Santorelli, S. F. (1992). Effec- Sciences, 353(1377), 1841–1849. tiveness of a meditation-based stress reduction Lopes da Silva, F. H., Vos, J. E., Mooibroek, J., program in the treatment of anxiety disorders. & Von Rotterdam, A. (1980). Relative con- American Journal of Psychiatry, 149(7), 936– tributions of intracortical and thalamo-cortical 943. processes in the generation of alpha rhythms, Karma Chagme[´ = Karma-chags-med]. (2000). revealed by partial coherence analysis. Elec- Naked awareness: Practical instructions on the troencephalography and Clinical Neurophysiol- union of Mahamudra and Dzogchen. With com- ogy, 50(5–6), 449–456. mentary by Gyatrul Rinpoche (A. B. Wallace, Lou, H. C., Kjaer, T. W., Friberg, L., Wilds- Trans.). Ithaca, NY: Snow Lion Publications. chiodtz, G., Holm, S., & Nowak, M. (1999). Kasamatsu, A., & Hirai, T. (1966). An elec- A 15O-H2 O PET study of meditation and the troencephalographic study of Zen meditation resting state of normal consciousness. Human (Zazen). Folia Psychiatrica et Neurologica Japon- Brain Mapping, 7(2), 98–105. 315 336 ica, 20, – . Lutz & Davidson 2005. Kiecolt-Glaser, J. K., Glaser, R., Gravenstein, S., Lutz, A., Greischar, L. L., Rawlings, N. B., Ricard, 1996 Malarkey, W. B., & Sheridan, J. ( ). Chronic M., & Davidson, R. J. (2004). Long-term med- stress alters the immune response to influenza itators self-induce high-amplitude gamma syn- virus vaccine in older adults. Proceedings of chrony during mental practice. Proceedings of 7 the National Academy of Sciences USA, 93( ), the National Academy of Sciences USA, 101(46), 3043 3047 – . 16369–16373. 1999 Klimesch, W. ( ). EEG alpha and theta oscil- Lutz, A., Lachaux, J. P., Martinerie, J., & Varela, F. lations reflect cognitive and memory perfor- J. (2002). Guiding the study of brain dynamics mance: A review and analysis. Brain Research – by using first-person data: Synchrony patterns 2 3 169 195 Brain Research Reviews, 29( – ), – . correlate with ongoing conscious states during Lachaux, J. P., Rodriguez, E., Martinerie, J., & a simple visual task. Proceedings of the National Varela, F. J. (1999). Measuring phase synchrony Academy of Sciences USA, 99(3), 1586– in brain signals. Human Brain Mapping, 8(4), 1591. 194–208. Lutz, A., Rawlings, N., & Davidson, R. J. (2005). Lack, L. C. (1978). Selective attention and the con- Changes in the tonic high-amplitude gamma oscil- trol of binocular rivalry. Adelaide, The Flinders lations during meditation correlates with long- University of South Australia. term practitioners’ verbal reports. Paper pre- Lazar, S., Bush, G., Gollub, R. L., Fricchione, G. sented at the Society for Neuroscience, Wash- L., Khalsa, G., & Benson, H. (2000). Functional ington, DC. brain mapping of the relaxation response and Lutz, A., & Thompson, E. (2003). Neuropheno- meditation. Neuroreport, 11(7), 1581–1585. menology: Integrating subjective experience Lazar, S., Kerr, C., Wasserman, R., Gray, J., and brain dynamics in the neuroscience of Greve, D., Treadway, M., McGarvey, M., consciousness. Journal of Consciousness Studies, Quinn, B., Dusek, J., Benson, H., Rauch, S., 10(9–10), 31–52. Moore, C., & Fischl, B. (2005). Meditation Maguire, E. A., Gadian, D. G., Johnsrude, I. S., experience is associated with increased corti- Good, C. D., Ashburner, J., Frackowiak, R. S., cal thickness. Neuroreport, 16(17):1893–7. & Frith, C. D. (2000). Navigation-related struc- Legrand, D. (2005). The bodily self: The sensori- tural change in the hippocampi of taxi drivers motor roots of pre-reflexive self-consciousness. [see comment]. Proceedings of the National Phenomenology and the Cognitive Sciences, 5(1): Academy of Sciences USA, 97(8), 4398– 89–118. 4403. Le Van Quyen, M., Chavez, M., Rudrauf, D., Maguire, E. A., Spiers, H. J., Good, C. D., & Martinerie, J. (2003). Exploring the nonlin- Hartley, T., Frackowiak, R. S., & Burgess, N. P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 549

(2003). Navigation expertise and the human human alpha rhythms: Theory, microcurrent hippocampus: A structural brain imaging anal- sources, multiscale measurements, and global ysis. Hippocampus, 13(2), 250–259. binding of local networks. Human Brain Map- Maharishi, M. Y. (1969). Maharishi Mahesh Yogi ping, 13(3), 125–164. on the Bhagavad Gita. New York: Penguin. Ooi, T. L., & He, Z. J. (1999). Binocular rivalry Maturana, H. R., & Varela, F. J. (1980). Autopoiesis and visual awareness: The role of attention. and cognition: The realization of the living, Vol. Perception, 28(5), 551–574. 42. Dordrecht: D. Reidel. Pagano, Rose, Stivers, & Warrenburg, (1976) McClelland, J. L., & Rogers, T.T.(2003). The par- Sleep during transcendental meditation. Sci- allel distributed processing approach to seman- ence. 23;191(4224):308–10. tic cognition. Nature Reviews Neuroscience, Pagano, R. R., & Warrenburg, S. (1983). 4(4), 310–322. Meditation: In search of a unique effect. In R. J. Meaney, M. J. (2001). Maternal care, gene expres- Davidson, G. E. Schwartz, & D. Shapiro (Eds.), sion, and the transmission of individual dif- Consciousness and self-regulation (Vol. 3). New ferences in stress reactivity across generations. York: Plenum Press. Annual Review of Neuroscience, 24, 1161–1192. Panksepp, J. (1988). Affective neuroscience: Merleau-Ponty, M. (1962). Phenomenology of per- The foundations of human and animal ception (C. Smith, Trans.). London: Routledge. emotions.Oxford: Oxford University Press. Miller, G. E., Cohen, S., Pressman, S., Barkin, A., Proudfoot, W. (1985). Religious experience. Rabin, B. S., & Treanor, J. J. (2004). Psycholog- Berkeley: University of California Press. ical stress and antibody response to influenza Rizzuto, D. S., Madsen, J. R., Bromfield, vaccination: When is the critical period for E. B., Schulze-Bonhage, A., Seelig, D., stress, and how does it get inside the body? Aschenbrenner-Scheibe, R., & Kahana, M. J. Psychosomatic Medicine, 66(2), 215–223. (2003). Reset of human neocortical oscillations Morrell, L. K. (1966). Some characteristics during a working memory task. Proceedings of of stimulus-provoked alpha activity. Elec- the National Academy of Sciences USA, 100(13), troencephalography & Clinical Neurophysiology, 7931–7936. 21(6), 552–561. Rosenkranz, M. A., Jackson, D. C., Dalton, K. M., Morse, D. R., Martin, J. S., Furst, M. L., & Dubin, Dolski, I., Ryff, C. D., Singer, B. H., Muller, L. L. (1977). A physiological and subjective D., Kalin, N. H., & Davidson, R. J. (2003). evaluation of meditation, hypnosis, and relax- Affective style and in vivo immune response: ation. Psychosomatic Medicine, 39(5), 304– Neurobehavioral mechanisms. Proceedings of 324. the National Academy of Sciences USA, 100(19), 11148 11152 Newberg, A., Alavi, A., Baime, M., Pour- – . dehnad, M., Santanna, J., & d’Aquili, E. (2001). Saffran, J. R., Aslin, R. N., & Newport, E. The measurement of regional cerebral blood L. (1996). Statistical learning by 8-month-old flow during the complex cognitive task of infants. Science, 274(5294), 1926–1928. meditation: A preliminary SPECT study. Psy- Schiff, S. J., So, P., Chang, T., Burke, R. chiatry Research, 106(2), 113–122. E., & Sauer, T. (1996). Detecting dynamical Nisbett, R. E., & Wilson, T. D. (1977). Telling interdependence and generalized synchrony more than we can know: Verbal reports on through mutual prediction in a neural ensem- mental processes. Psychological Review, 84, 231– ble. Physical Review. E. Statistical Physics, Plas- 259. mas, Fluids, and Related Interdisciplinary Topics, 6 6708 6724 Nunez, P. L., Srinivasan, R., Westdorp, A. F., 54( ), – . Wijesinghe, R. S., Tucker, D. M., Silberstein, Schnitzler, A., & Gross, J. (2005). Normal and R. B., & Cadusch, P. J. (1997). EEG coherency. pathological oscillatory communication in the I: Statistics, reference electrode, volume con- brain. Nature Review Neuroscience, 6(4), 285– duction, Laplacians, cortical imaging, and inter- 296. pretation at multiple scales. Electroencephalog- Schuman, M. (1980). The psychophysiological raphy & Clinical Neurophysiology, 103(5), 499– model of meditation and altered states of 515. consciousness: A critical review. In J. M. David- Nunez, P. L., Wingeier, B. M., & Silberstein, son & R. J. Davidson (Eds.), The psychobiology 333 378 R. B. (2001). Spatial-temporal structures of of consciousness, – . New-York: Plenum. P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

550 the cambridge handbook of consciousness

Schuster, B. G. (2001). A new integrated program Tallon-Baudry, C., & Bertrand, O. (1999). Oscil- for natural product development and the value latory gamma activity in humans and its role in of an ethnomedical approach. Journal of Alter- object representation. Trends in Cognitive Sci- native & Complementary Medicine, 7(6), 61–72. ence, 3(4), 151–162. Shapiro, D. H., & Walsh, R. N. (Eds.). (1984). Tebecis, A. K. (1975). A controlled study of the Meditation: Classical and contemporary perspec- EEG during transcendental meditation: Com- tives. New York: Aldine. parison with hypnosis. Folia Psychiatrica et Sharf, R. H. (1998). Experience. In M. Taylor Neurologica Japonica, 29(4), 305–313. (Ed.), Critical terms for religious studies (pp. 94– Thrangu, & Johnson, C. (2004). Essentials of 116). Chicago: University of Chicago Press. Mahamudra: Looking directly at the mind. Sheline, Y. I. (2003). Neuroimaging studies of Boston: Wisdom Publications. mood disorder effects on the brain. Biological Tononi, G., & Edelman, G. M. (1998). Conscious- Psychiatry, 54(3), 338–352. ness and complexity. Science, 282(5395), 1846– Silananda, U. (1990). The four foundations of mind- 1851. fulness. Boston: Wisdom Publications. Travis, F., Arenander, A., & DuBois, D. (2004). Singer, T., Seymour, B., O’Doherty, J., Kaube, H., Psychological and physiological characteristics Dolan, R. J., & Frith, C. D. (2004). Empathy of a proposed object-referral/self-referral con- for pain involves the affective but not sensory tinuum of self-awareness. Consciousness & Cog- components of pain. Science, 303(5661), 1157– nition, 13(2), 401–420. 1162. Tsongkhapa [ = Tsong kha pa blo bzang grags pa]. Singer, W. (2001). Consciousness and the binding (2002). The great treatise on the stages of the path problem. Annals of the New York Academy of to enlightenment (31st ed.). Ithaca, NY: Snow Sciences, 929, 123–146. Lion Publications. Singer, W. (1999). Neuronal synchrony: A versa- Urry, H. L., van Reekum, C. M., Johnstone, T., tile code for the deﬁnition of relations? Neuron, Thurow, M. E., Burghy, C. A., Mueller, C. J., 24(1), 49–65, 111–125. & Davidson, R. J. (2003). Neural correlates of Slotnick, S. D., Moo, L. R., Kraut, M. A., voluntarily regulating negative affect. (Report no. 725 18 Lesser, R. P., & Hart, J., Jr. (2002). Interactions . .). San Diego: Society for Neuroscience. between thalamic and cortical rhythms during Varela, F. (1995). Resonant cell assemblies: A new semantic memory recall in human. Proceedings approach to cognitive functions and neuronal 1 81 95 of the National Academy of USA, 99(9), 6440– synchrony. Biological Research, 28( ), – . 6443. Varela, F. (1996). Neurophenomenology: A Snellgrove, D. L. (2002). Indo-Tibetan Buddhism: methodological remedy to the hard problem. 330 350 Indian Buddhists and their Tibetan successors. Journal of Consciousness Studies, 3, – . Boston: Shambhala. Varela, F. (1999). The specious present: A neu- Srinivasan, R., Russell, D. P., Edelman, G. M., & rophenomenology of time consciousness. In J. Tononi, G. (1999). Increased synchronization Petitot, F. J. Varela, J.-M. Roy, & B. Pachoud 266 of neuromagnetic responses during conscious (Eds.), Naturalizing phenomenology (pp. – 314 perception. Journal of Neuroscience, 19(13), ). Stanford, CA: Stanford University 5435–5448. Press. Suffczynski, P., Kalitzin, S., Pfurtscheller, G., & Varela, F., Lachaux, J. P., Rodriguez, E., & Mar- 2001 Lopes da Silva, F. H. (2001). Computational tinerie, J. ( ). The brainweb: Phase synchro- model of thalamo-cortical networks: Dynami- nization and large-scale integration. Nature 4 229 239 cal control of alpha rhythms in relation to focal Reviews Neuroscience, 2 ( ), – . attention. International Journal of Psychophysi- Varela, F., & Thompson, E. (2001). Radical ology, 43(1), 25–40. embodiment: Neural dynamics and conscious- 10 418 Takahashi, T., Murata, T., Hamada, T., Omori, ness. Trends in Cognitive Science, 5( ), – 425 M., Kosaka, H., Kikuchi, M., Yoshida, H., & . Wada, Y. (2005). Changes in EEG and auto- Varela, F. J., Thompson, E., & Rosch, E. (1991). nomic nervous activity during meditation and The embodied mind. Cambridge, MA: MIT their association with personality traits. Inter- Press. national Journal of Psychophysiology, 55(2), 199– von Stein, A., Chiang, C., & Konig, P. (2000). 207. Top-down processing mediated by interareal P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

meditation and the neuroscience of consciousness 551

synchronization. Proceedings of the National West, M. A. (1980). Meditation and the EEG. Academy of Sciences USA, 97(26), 14748– Psychological Medicine, 10(2), 369–375. 14753. West, M. A. (Ed.). (1987). The psychology of med- Wallace, A. B. (Ed.). (2003). Buddhism and sci- itation. New York: Clarendon Press. ence. New York: Columbia University Press. Wider, K. V. (1997). The bodily basis of conscious- Wallace, B. A. (1999). The Buddhist tradition of ness. Ithaca, NY: Cornell University Press. Samatha: Methods for reﬁning and examining Williams, M. A., Morris, A. P., McGlone, F., consciousness. Journal of Consciousness Studies, Abbott, D. F., & Mattingley, J. B. (2004). 6(2–3), 175–187. Amygdala responses to fearful and happy facial Wallace, R. K. (1970). Physiological effects of expressions under conditions of binocular sup- transcendental meditation. Science, 167(926), pression. Journal of Neuroscience, 24(12), 2898– 1751–1754. 2904. Wangchug Dorje[´ = Dbang-phyug-rdo-rje]. Williams, P. (1989). Mah¯ay¯ana Buddhism: The (1989). The Mahamudra eliminating the dark- doctrinal foundations. London: Routledge. ness of ignorance (3rd rev. ed.). Dharamsala, Woolfolk R. L. (1975). Psychophysiological cor- India: Library of Tibetan Works and Archives. relates of meditation. Arch Gen Psychiatry. 32 10 1326 33 Ward, L. M. (2003). Synchronous neural oscilla- ( ): – . tions and cognitive processes. Trends in Cogni- Younger, J., Adriance, W., & Berger, R. J. (1975). tive Science, 7(12), 553–559. Sleep during Transcendental Meditation. Per- 3 953 954 Warrenburg, Pagano, Woods, & Hlastala, (1980). ceptual & Motor Skills, 40( ), – . A comparison of somatic relaxation and Zahavi, D., & Parnas, J. (1998). Phenom- EEG activity in classical progressive relaxation enal consciousness and self-awareness: A and transcendental meditation. J Behav Med. phenomenological critique of representational 3(1):73–93. theory. Journal of Consciousness Studies, 5, 687– 705 Wenger, M. A., & Bagchi, B. K. (1961). Studies of . autonomic functions in practitioners of Yoga in Zeman, A. (2001). Consciousness. Brain, 124(Pt India. Behavioral Sciences, 6, 312–323. 7), 1263–1289. P1: KAE 0521857430c19 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:48

552 P1: KAE 0521857430c20 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:41

G. Anthropology/Social Psychology of Consciousness

553 P1: KAE 0521857430c20 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:41

554 P1: KAE 0521857430c20 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:41

CHAPTER 20 Social Psychological Approaches to Consciousness

John A. Bargh

Abstract any given phenomenon. However, because these studies focus on the relative influence A central focus of contemporary social psy- of both conscious and automatic processes, chology has been the relative influence there has been a strong influence within of external (i.e., environmental, situational) social psychology of dual-process models versus internal (i.e., personality, attitudes) that capture these distinctions (e.g., inten- forces in determining social judgment and tional versus unintentional, effortful versus social behavior. But many of the classic find- efficient, aware versus unaware). Another ings in the field – such as Milgram’s obe- reason that dual-process models became dience research, Asch’s conformity studies, popular in social psychology is that the dis- and Zimbardo’s mock-prison experiment – tinction nicely captured an important truth seemed to indicate that the external forces about social cognition and behavior: that swamped the internal ones when the chips people seem to process the identical social were down. Where in the social psycholog- information differently depending on its rel- ical canon was the evidence showing the evance or centrality to their important goals internal, intentional, rational control of one’s and purposes. own behavior? Interestingly, most models of a given phenomenon in social psychology have started with the assumption of a Introduction major mediational role played by conscious choice and intentional guidance of judg- Historically, social psychology has been con- ment and behavior processes. Then, empiri- cerned with the determinants of social beha- cal work focuses on the necessity or valid- vior; specifically, the relative influence of ity of this assumption. As a consequence external (i.e., environmental, situational) there has been a greater research focus on the versus internal (i.e., personality, attitudes) non-conscious than the conscious aspects of causal forces. Many of the most famous

555 P1: KAE 0521857430c20 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:41

556 the cambridge handbook of consciousness

studies in social psychology focused on this Conscious by Default issue of internal versus external determinants of behavior (Wegner & Bargh, 1998). These findings were surprising at the time For example, early attitude research was (to social psychologists as well as to the driven by the belief that attitudes would general public) because they violated peo- prove to be a strong predictor of actual ple’s strongly held assumption that one’s behavior. Yet it wasn’t long after Thur- own behavior was under one’s internal and stone (1928) first demonstrated that inter- intentional control. Compared with cogni- nal, private attitudes could be measured that tive psychology or cognitive neuroscience, LaPiere (1934) caused great consternation by social psychology tends to focus on psy- seeming to show that one’s stated attitudes chological processes of a relatively high toward a social group did not predict one’s order of complexity: for instance, judgment, actual behavior toward that group very well goal pursuit over extended time periods, at all. and behavior in social interaction. Going Asch’s (1952) famous conformity studies back at least to Descartes (1633) there is were surprising at the time because they a deep philosophical tradition of assigning seemed to show that a person’s publicly such complex processes to an agentic “mind” made judgments of the relative lengths of instead of the mechanical “body.” That is, lines presented clearly on a chalkboard were for any given process of such complexity, swayed by the (stage-managed) opinions of the initial assumption tends to be that the the other “subjects” present in the experi- individual plays an active, agentic role in mental session. Thus even in cases where its instigation and operation, as opposed the judgment or behavior should have been to it being a purely mechanical, deter- determined entirely by internal perceptual mined phenomenon (see Bargh & Ferguson, experience, external forces still played a role. 2000). Milgram’s (1963) obedience experiments, in Perhaps as an inheritance or vestige of this which subjects believed they were admin- long-standing philosophical stance, then, istering painful shocks to another subject, social psychology tends to begin its analysis were disturbing and controversial because of any complex, important phenomenon by they demonstrated the power of a situa- assuming a central role for conscious (intentional influence over the subject’s behavior tional, effortful, and aware; see next section) (i.e., the experimenter’s authority) to over- choice and monitoring processes. Research ride presumed internal influences (i.e., the then has the effect of discovering the extent subject’s presumed personal values not to and role of non-conscious components of cause pain or harm to another). the process or phenomenon. Note how, in Darley and Latane’s (1968) seminal stud- the classic studies above, the initial starting ies of bystander intervention showed how assumption is that the judgment or behav- the simple presence of other people in the ior is under internal, strategic (i.e., con- situation seemed to inhibit individuals from scious) control. This pattern can be found helping another person in clear distress. And in other traditional areas of social psycho- last but not least, the well-known Stanford logical inquiry as well. Early attribution the- Prison Study (Haney, Banks, & Zimbardo, ories began with a model of humans as ratio- 1973) provided a powerful demonstration nal scientists, using effortful and intentional of situational forces (social roles, as pris- “analysis-of-variance” methods to draw infer- oner versus guard) swamping dispositional ences of causality (Kelley, 1967). However forces (values, good intentions) in determin- as the research evidence started to come in, ing the behavior of the participants in a real- attribution theory then moved to a more istic prison simulation. automatic and less deliberative model (e.g., Where oh where, in all of these findings, Gilbert, 1989; Taylor & Fiske, 1978). Sim- was the internal, intentional, rational control ilarly, the phenomena of stereotyping and of one’s own behavior? prejudice were initially assumed to be driven P1: KAE 0521857430c20 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:41

social psychological approaches to consciousness 557

by motivated, conscious processes (see Nis- dual-process models that capture these dis- bett & Ross, 1980), but then were shown by tinctions (e.g., intentional versus uninten- a considerable amount of research to have tional, effortful versus efficient, aware ver- a significant automatic, non-conscious com- sus unaware). Cognitive social psychology ponent (Brewer, 1988; Devine, 1989). has emphasized the study of non-conscious This is how research related to issues processes, whereas motivational social psy- of consciousness has proceeded in social chology is still mainly the study of conscious psychology. The initial models start with processes. But clearly, conscious and non- the default assumption that the phe- conscious components of a complex psycho- nomenon under investigation involves con- logical process are two sides of the same coin. scious, aware, intentional appraisals or By testing the default initial assumptions of behavior on the part of the participants, and a necessary and pivotal role for conscious then this set of presumed necessary condi- processes – showing where conscious pro- tions is whittled down as the research find- cesses are needed versus where they are not – ings warrant. As a consequence there has we learn a great deal about the role and been a greater research focus on the non- function of consciousness. In this subtrac- conscious than the conscious aspects of any tive manner, the social psychological study given phenomenon. of non-conscious judgment and behavioral The main exceptions to this rule are mod- phenomena adds to our understanding of the els of self-regulation and goal pursuit (see purpose of conscious processes. Bandura, 1986; Carver & Scheier, 1981; Deci & Ryan, 1985; Locke & Latham, 1990; Mis- chel, Cantor, & Feldman, 1996), in which Dual-Process Models: Automatic conscious choice and willpower are fea- Versus Controlled Processes tured as mediating, explanatory variables. This is probably because, even among the Cognitive approaches to social psychology relatively complex phenomena studied in were greatly influenced by the dual-process social psychology, self-regulatory processes models of the 1970s (Posner & Snyder, are the most complex, dynamic, and inter- 1975; Shiffrin & Schneider, 1977) that dis- active with the shifting, uncertain environ- tinguished between conscious and auto- ment (see Baumeister, 1998; Fitzsimons & matic modes of information processing (see Bargh, 2004). Because of the level of abstrac- Chaiken & Trope, 1999). Conscious or con- tion and complexity of these processes, it is trolled processes were said to be inten- understandable that it has taken longer to tional, controllable, effortful, and in awareness, find and isolate their mechanisms and com- whereas automatic processes were charac- ponents. Yet even in the domain of self- terized by the opposite set of features: They regulation research, studies are beginning to were unintended, uncontrollable, effortless, identify non-conscious, automatic compo- and outside of awareness (Johnson & Hasher, nents (see Fitzsimons & Bargh, 2004). For 1987). However, at least for psychological example, complex goal pursuit can be put processes of the level of complexity studied into motion by situational features instead of by social psychologists, these qualities did exclusively by consciously made intentions not seem to co-vary together in an all-or- or choices, and it can operate in a flexible nothing fashion (Bargh, 1989). For instance, manner, interacting with the changing envi- stereotypes might become activated auto- ronment over time, just as can conscious goal matically (unintentionally and efficiently), pursuit (Chartrand & Bargh, 2002). but their influence on judgment was con- All of these domains of research, then, trollable (Devine, 1989; Fiske, 1989); mak- recognize the influence and importance of ing dispositional attributions might be the both conscious and automatic processes. It efficient and reflexive default process, but is not surprising then that there has been a still required the intention to understand strong influence within social psychology of the causes of the person’s behavior (Gilbert, P1: KAE 0521857430c20 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:41

558 the cambridge handbook of consciousness

1989). Consequently, social psychologists mation moderates whether the individual have tended to study the separate and dis- will deal with it in an effortful and systematic tinct aspects of automatic and controlled manner versus in an off-hand and efficient processes, and I have organized the brief way – holds across many different social review below in these lower-level terms (for psychological phenomena. One of the most more complete reviews see Bargh, 1994, important is impression formation, in particu- 1996; Wegner & Bargh, 1998). lar the degree to which the perceiver will pay attention to and be influenced in his or her judgment by the target person’s individual Effortful Processing: Only When characteristics, as opposed to more superfi- it Matters cial (but less effort-requiring) features, such One reason that dual-process models as race, gender, age, or ethnicity. If a person’s became popular in social psychology is that own outcomes depend on the target person the distinction nicely captured an important (i.e., there is goal relevance), a more indi- truth about social cognition and behavior: viduated and less stereotype-based impres- that people seem to process the identical sion is formed (Erber & Fiske, 1984; Neu- social information differently depending on berg & Fiske, 1987) than if the target person its relevance or centrality to their impor- does not have control over the perceiver’s tant goals and purposes. For example, peo- outcomes. ple process a persuasive message differently Taylor and Fiske (1978) coined the if it concerns or affects them directly ver- term “cognitive miser” to refer to this sus when it does not. If a proposed com- human tendency not to think effortfully prehensive exam requirement is allegedly about other people or attitude issues unless to be instituted next year at a student’s really necessary. Underlying this idea is the own university, she will spend more time notion that effortful processing is limited and think more effortfully about the vari- in its capacity at any given moment and ous arguments for versus against it; if it is to so should be used only sparingly, to be occur 5 years from now or at another univer- reserved for the most important stimuli sity, she will not expend the same degree of and events. Consistent with this notion, effort. Instead she will tend to rely on heuris- recent research by Baumeister, Bratslavsky, tics or shortcuts – such as the attractiveness Muraven, and Tice (1998) has confirmed the or expertise of the source of the message – limited-capacity nature of effortful social- to decide her position (Chaiken, 1980; Petty information processing: They found that & Cacioppo, 1984). using it in one domain – even just to make People also were found to use short- a simple choice – seems to severely limit cut heuristics in making causal attributions its availability for other tasks, for some time (Taylor & Fiske, 1978) and even behavioral thereafter. choices in social interaction settings (Langer, 1978 Blank, & Chanowitz, ). Langer and col- Efficient Processing: Automatic leagues argued that people develop men- Components of Social Perception tal representations of common situations (‘scripts’; see Abelson, 1981) that then guide Another important variable moderating their behavior “mindlessly” within those sit- whether people will engage in effortful uations. Heuristic cues such as the size of versus heuristic or superficial processing is a request (e.g., for 1 minute versus 10 min- whether they are able to do so under the cur- utes of your time) determined whether peo- rent circumstances. Conscious, effortful pro- ple would assent to it, not the quality of the cessing is relatively slow and so takes time; reason given for the request. often the individual does not have the time, This basic principle – that the personal as when under time pressure or when there importance or goal relevance of target infor- are multiple people or events to attend to P1: KAE 0521857430c20 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:41

social psychological approaches to consciousness 559

at once. Indeed, such information overload who did not) showed much higher corre- conditions are not unusual in the busy, noisy spondence between those attitudes and their “real world” outside the psychology labora- actual voting behavior (several months later) tory. Under these conditions, efficient, auto- in the election. matic forms of information processing have greater influence than usual, because they causal attribution are not constrained as much by capacity or Because of the immediacy and fluency of time limitations or by the current focus of automatic forms of information process- conscious thought. ing, such as stereotyping, several researchers have proposed sequential or stage mod- attitude activation els of phenomena in which the first or An excellent example of this can be found default stage is relatively reflexive or auto- in Fazio’s (1986, 1990) model of the relation matic, with the second, more controlled between attitudes and behavior. The extent stage occurring only if the person has both to which one’s attitudes determine one’s the ability (i.e., lack of time or overload con- behavior has long been a central research straints) and motivation to do so. Gilbert question in social psychology (see Eagly (e.g., 1989) argued that people have a default & Chaiken, 1993). Faced with evidence or automatic bias to locate causality for that the general correspondence between another person’s behavior “in” that person attitudes and behavior was weak at best him- or herself – in other words, making a (e.g., LaPiere, 1934; Wicker, 1969), attitude dispositional attribution about the reason for researchers began to look for the conditions that behavior. In his model, a causal attri- that supported or fostered the relation. One bution to situational factors is only made such proposal was Fazio’s (1986) automatic within a second, conscious processing stage – attitude activation model. but that stage only occurs if the person has In this model he contended that attitudes the time and processing resources available varied in strength, or the degree of association to engage in it. between the representation of the attitude In several studies, Gilbert and colleagues object and its evaluative tag (i.e., as good or showed that people did not take clear situ- bad). Strong attitudes are those character- ational influences into account when under ized by a relatively automatic association, conditions of distraction or attentional load. such that the mere perception of the atti- When watching a videotape of a woman tude object in the environment was suffi- being interviewed and being asked rather cient to also activate its associated attitude – embarrassing questions, those participants no intentional, effortful thinking (such as under attentional overload (performing a about how one feels about the object) secondary attention-demanding task while was necessary. Weak attitudes, on the other watching the tape) concluded she was a dis- hand, did not possess this automatic associa- positionally shy and anxious person. People tion and so did not become active unless the in the control condition, on the other hand, person happened to think about his or her who watched the tape without having to do feelings toward the object. the secondary task, did not draw that con- In several studies, in which attitude clusion – instead, they attributed the reason strength was either manipulated or mea- for her anxious behavior to the situation of sured, automatic attitudes showed a more having to answer embarrassing questions. consistent influence on behavior than did weak attitudes. Indeed, Fazio and Williams impression formation (1986) showed that those participants who Much research in social psychology has possessed strong, relatively automatic atti- focused on the immediate or spontaneous tudes toward the candidates in the 1984 effects of social stimuli – those that occur so U.S. presidential contest (compared to those efficiently that all it takes for the process to P1: KAE 0521857430c20 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:41

560 the cambridge handbook of consciousness

occur is the mere perception of the object, to become active as well to influence impres- person, or event in the environment (Bargh, sions of a target person. 2001). For example, in the case of Fazio’s Devine’s study stimulated a great deal of (1986) model of the attitude-behavior rela- research into the conditions under which tion, discussed above, seeing a rose activates the automatic stereotype activation effect not only the associated concept “rose” but is more or less likely to occur (see review also one’s feelings or attitude toward roses. in Devine & Monteith, 1999). The bottom The activation of the attitude occurs in an line seems to be that cultural stereotypes uncontrollable manner similar to how writ- can be picked up at a quite early age and ten words activate their meanings during can exert a biasing influence on social per- reading. ception, judgment, and even behavior (Fazio When forming initial impressions of other et al., 1995) without the person being aware people, certain forms of information about of such influence (Bargh, 1999). Fortunately, them appear to have a similarly privileged however, racial and gender stereotyping is status; we tend to detect and be influenced one form of unconscious bias that many peo- by these features in the course of percep- ple now seem to accept as a possibility (i.e., tion, in an automatic fashion, without being they have a correct “theory of influence” in aware of it. For example, Higgins, King, and this case; see next section) and so can adjust Mavin (1982) showed that each of us is and correct for it if they have the motivation chronically sensitive to certain kinds of social to do so. behavior but not others, with wide individual differences in the exact content of these Awareness and Control chronic sensitivities. Bargh and Thein (1985) then showed that under information over- People are often unaware of the reasons load (rapid presentation) conditions that and causes of their own behavior. In fact, prevented people in a control group from recent experimental evidence across several being able to differentiate in their impres- different areas of psychology points to a sions between a mainly honest and a mainly deep and fundamental dissociation between dishonest target person, those participants conscious awareness and the mental pro- who were chronically or automatically sen- cesses responsible for one’s behavior; many sitive to the dimension of honesty were still of the wellsprings of behavior appear to be able to differentiate the two target persons. opaque to conscious access. Although that This is because they were able to process and research has proceeded somewhat indepen- be influenced by the honest and dishonest dently in social psychology (e.g., Dijkster- behaviors in an automatic, efficient manner. huis & Bargh, 2001; Wilson, 2002), cognitive psychology (e.g., Knuf, Aschersleber, stereotyping and prejudice & Prinz, 2001; Prinz, 1997), and neuropsy- By far the most researched form of such chology (e.g., Frith, Blakemore, & Wolpert, spontaneous cognitive reactions to the social 2000; Jeannerod, 1999), using quite differ- environment is social stereotyping (see ent methodologies and guiding theoretical Bargh, 1999; Brewer, 1988; Devine, 1989). perspectives, all three lines of research have In a now classic study, Devine (1989) found reached the same general conclusions. that even non-prejudiced people (at least In social psychology, awareness of sources by one fairly explicit paper-and-pencil mea- of influence on judgment and social behavior sure of racism) show evidence of automatic has long been an important research topic. stereotype activation. In one study (1989, Beginning with the seminal work of Nisbett Experiment 2), she presented participants and Wilson (1977), researchers observed subliminally with stimuli related to positive that people were often unaware of actual aspects of the African-American stereotype strong influences on their choices and behav- (e.g., musical, athletic) and showed that this ior. In one study, for example, some experi- caused the negative aspects (e.g., hostility) mental participants watched a job interview P1: KAE 0521857430c20 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:41

social psychological approaches to consciousness 561

in which the interviewee spilled some to these stimuli is believed to “prime” or coffee; others saw the identical tape with- make that concept temporarily accessible in out the spill incident. Although the former memory. In such experiments, participants group rated the interviewee as significantly are of course aware of the critical word less qualified for the job, they also reported stimuli at the time of working on the test, that the coffee spill (among a list of many yet they have no awareness that such mere possible influencing factors) did not affect exposure to words could possibly influence their judgment. their judgments or behavior (they can and Wilson (2002; Wilson & Brekke, 1994) has do). However, when participants do become extended this line of research to document aware of a potential influence, such as when the many ways in which people seem out the priming stimuli are extreme and salient of touch with the actual determinants and (e.g., “Dracula” as a prime for hostility; Herr, influences of their judgments and behavior. Sherman, & Fazio, 1986), the usual prim- An emergent principle from this research is ing effects are no longer obtained. Thus, in that people have lay “theories” about what social psychology, an important distinction is influences their feelings and decisions, or that between awareness of the stimulus ver- causes them (and others) to behave in cer- sus awareness of its possible effects (Bargh, tain ways, and often if not usually these theo- 1992). The latter and not the former appears ries do not accurately reflect the actual influ- to be the key moderator of whether uncon- ences and causes. scious influences of that stimulus will occur. Priming research, in which social concepts (e.g., traits and stereotypes) are first Intentionality: What We Do activated in an off-hand, subtle manner and Without Meaning To then influence the person’s subsequent judgments or behavior (see reviews in Bargh & impression formation Chartrand, 2000; Higgins, 1996), provides Uleman and his colleagues (1989; Uleman, another example of the dissociation between Newman, & Moskowitz, 1996; Winter & important environmental influences and Uleman, 1984; see also Carlston & Skowron- the person’s awareness of those influences. ski, 1994) have documented a “spontaneous Across many studies, the critical variable as trait inference” effect, in which social per- to whether a person is able to control the ceivers tend to encode the behavior of oth- external effect is not whether the person is ers in trait-concept terms (e.g., as an honest, aware of the influencing stimulus per se (i.e., intelligent,orselfish behavior), automatically whether it was subliminal or supraliminal), and without intending to do so. Using Tulv- but rather whether the person is aware of the ing and Thompson’s (1973) encoding speci- potential influence of that stimulus. Thus, ficity paradigm, these researchers showed priming stimuli presented subliminally have that trait terms corresponding to the behav- the same quality of effect as those presented ior (e.g., generous for “she donated her stock supraliminally (i.e., visible, reportable), as gains to charity”) later served as effec- long as the person does not believe or appre- tive retrieval cues for the behavior, even ciate that the stimulus could have an effect though the experimental participants had on him or her (Bargh, 1992). not been instructed to form impressions of For example, in the popular scrambled the sentence actors (merely to remember the sentence test method of priming (Srull behaviors). Apparently, then, the trait term & Wyer, 1979; see Bargh & Chartrand, had been spontaneously encoded by the par- 2000), experimental participants complete ticipants when reading that behavior. an ostensible language test in which they Spontaneous attitude, trait-concept, and reorder strings of words into grammatically stereotype activation are three important correct sentences. Embedded in this test are ways in which people “go beyond the infor- some words semantically related to a cer- mation given” (Bruner, 1957), such that sem- tain social concept; merely being exposed antic and affective information not actually P1: KAE 0521857430c20 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:41

562 the cambridge handbook of consciousness

present in the current environment becomes For example, words related to achieve- activated automatically in the course of per- ment and high performance might be ception to then exert an “unseen” influence embedded along with other, goal-irrelevant on judgments and behavior. words in a puzzle, or words related to cooperation might be presented subliminally in imitative behavior and the course of an ostensible reaction time ideomotor action task. Just as with single forms of social behav- Two streams of research in social psychol- ior such as politeness or intelligence, pre- ogy have converged on the idea that com- senting goal-related stimuli in this fashion plex social behavior tendencies can be trig- causes the goal to become active and then gered and enacted non-consciously. One line operate to guide behavior over time toward of research focuses on ideomotor action or that goal. People primed with achievement- the finding that mental content activated in related stimuli perform at higher levels on the course of perceiving one’s social envi- subsequent tasks than do control groups, ronment automatically creates tendencies to those primed with cooperation-related stim- behave the same way oneself (Prinz, 1997). uli cooperate more in a commons-dilemma Thus, for example, one tends to mimic, with- game, and those primed with evaluation- out realizing it, the posture and physical related stimuli form impressions of other gestures of one’s interaction partners (Char- people while those in a control group do trand & Bargh, 1999). not (see reviews in Chartrand & Bargh, 2002; This “chameleon effect” has been found Fitzsimons & Bargh, 2004). to extend even to the automatic activation of Neither the ideomotor action nor the abstract, schematic representations of peo- automatic goal pursuit effects are restricted ple and groups (such as social stereotypes) to the laboratory environment; for example, in the course of social perception (see Dijk- merely thinking about the significant other sterhuis & Bargh, 2001). For example, sub- people in our lives (something we all do tly activating (priming) the professor stereo- quite often) causes the goals we character- type in a prior context causes people to score istically pursue when with them to become higher on a knowledge quiz (Dijksterhuis & active and to then guide our behavior with- van Knippenberg, 1998), and priming the out our choosing or knowing it, even when elderly stereotype makes college students those individuals are not physically present not only walk more slowly but have poorer (Fitzsimons & Bargh, 2003). And the non- incidental memory as well (Bargh, Chen, & conscious ideomotor effect of perception Burrows, 1996). Thus, the passive activation on action becomes a matter of widespread of behavior (trait) concepts in the natural social importance considering the mass course of social perception (as experimen- exposure of people to violent behavior on tally simulated by priming manipulations) television or in movies (see Anderson & increases the person’s tendency to behave in Bushman, 2002; Berkowitz, 1984). line with that concept him- or herself. unconscious motivation and Dissociations Between Intention automatic goal pursuit and Action The second stream of research has shown that social and interpersonal goals can also These findings within social psychologi- be activated automatically through external cal research of non-conscious control over means (as in priming manipulations). The higher mental processes, such as support individual then pursues that goal in the sub- behavior in social settings, may seem a bit sequent situation, but without consciously magical or mysterious without a consider- intending to or being aware of doing so ation of related recent findings in cogni- (Bargh, 1990; Bargh, Gollwitzer, Lee-Chai, tive psychology and cognitive neuroscience. Barndollar, & Troetschel, 2001; Chartrand & Together, though, these streams of research Bargh, 1996). tell a coherent story about the non-conscious P1: KAE 0521857430c20 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:41

social psychological approaches to consciousness 563

wellsprings and governing structures of of “mirror neurons” – first in macaque mon- social judgment, behavior, and goal pursuit. keys (Rizzolatti & Arbib, 1998) and then in humans (Buccino et al., 2001). In these studies, simply watching mouth, hand, and Non-Conscious Action Control foot movements causes the activation of Several lines of cognitive neuroscience the same functionally specific regions of the research support the idea of a dissociation premotor cortex as when performing those between conscious awareness and intention, same movements oneself. These mirror neu- on the one hand, and the operation of com- rons could be a neurological basis for the plex motor and goal representations on the “chameleon effect” of non-conscious imita- other (Prinz, 2003). One major area of such tion of the behavior of one’s interaction part- research focuses on the distinct and separate ners (Chartrand & Bargh, 1999). visual input pathways devoted to perception versus action. Non-Conscious Operation of Working The first such evidence came from a study Memory During Goal Pursuit of patients with lesions in specific brain regions (Goodale, Milner, Jakobsen, & Carey, Clearly, non-conscious goal pursuit must uti- 1991). Those with lesions in the parietal lobe lize the structures of working memory to region could identify an object but not reach guide behavior within the unfolding situ- for it correctly based on its spatial orienta- ation toward the desired goal (see Hassin, tion (such as a book in a horizontal versus 2004). Such complex behavior, which is con- vertical position), whereas those with lesions tinually responsive to ongoing environmen- in the ventral-visual system could not recog- tal events and coordinated with the behav- nize or identify the item but were nonethe- ior of others, has to involve the operation less able to reach for it correctly, when asked of the brain structures that support working in a casual manner to take it from the exper- memory – namely the frontal and prefrontal imenter. In other words, the latter group cortex. However, under the original concept showed appropriate action toward an object of working memory as that portion of long- in the absence of conscious awareness or term memory that was currently in con- knowledge of its presence. scious awareness (e.g., Atkinson & Shiffrin, Decety and Grezes` (1999) and Norman 1968), the idea of non-conscious operation of (2002) concluded from this and related evi- working memory structures is incoherent at dence that two separate cortical visual path- best. If working memory was a single men- ways are activated during the perception of tal “organ” that held both the current goal human movement: a dorsal one for action and the relevant environmental information tendencies based on that information, and on which that goal was acting (selecting rele- a ventral one used for understanding and vant information and transforming it accord- recognition of it. The dorsal system oper- ing to the requirements of the current goal; ates mainly outside of conscious awareness, see Cohen, Dunbar, & McClelland, 1990), whereas the workings of the ventral system then one should always be aware of the are normally accessible to consciousness. intention or goal that is currently residing Thus the dorsal stream (or activated prag- in active, working memory. matic representation) could drive behavior The answer to this apparent paradox, of in response to environmental stimuli in the course, is that working memory is not a sin- absence of conscious awareness or under- gle unitary structure. This idea was originally standing of that external information. It proposed by Baddeley and Hitch (1974; see could, in principle, support a non-conscious also Baddeley, 1986), who envisaged a sys- basis for ideomotor action effects that are tem comprising multiple components, not primed or driven by recent behavioral infor- just for the temporary storage of informa- mational input from other people. tion (the phonological loop and visuospatial Additional support for non-conscious scratchpad) but also for the direction and action initiation comes from the discovery allocation of limited attention (the “central P1: KAE 0521857430c20 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:41

564 the cambridge handbook of consciousness

executive”). In a parallel development, movements are normally represented in the psychiatrists working with patients with prefrontal and premotor cortex, but the rep- frontal lobe damage – the frontal lobes being resentations actually used to guide action are brain structures underlying the executive in the parietal cortex. In other words, inten- control functions of working memory (Bad- tions and the motor representations used to deley, 1986) – were noting how the behav- guide behavior seem to be held in anatomi- ioral changes associated with frontal lobe cally separate, distinct parts of the brain. This damage were exceedingly complex and vari- makes it possible for some patients to no able, depending on the exact locations of the longer be able to link their intentions to their damage (Mesulam, 1986,p.320). This too actions if there is impairment in the loca- was consistent with the notion that exec- tion where intended movements are repre- utive control was not a single resource but sented, but no impairment in the location rather comprised several distinct specialized where action systems actually operate. functions, located in different parts of the The finding that, within working mem- frontal and prefrontal cortex. ory, representations of one’s intentions (acc- If so, then at least in theory it becomes essible to conscious awareness) are stored possible that there are dissociations between in a different location and structure from consciously held intentions on the one hand the representations used to guide action (not and the goal-driven operation of working accessible) is of paramount importance to an memory structures on the other. This is understanding of the mechanisms underly- what is manifested in Lhermitte’s (1986) ing non-conscious social behavior and goal syndrome; “an excessive control of behav- pursuit. If it had been the case that inten- ior by external stimuli at the expense of tions and corresponding action plans were behavioral autonomy” (p. 342). Lhermitte’s stored in the same location, so that aware- patients had suffered a stroke which had ness of one’s intention was solely a mat- produced lesions in the same (inferior pre- ter of conscious access to the currently frontal) location of the brain in both cases. operative goal or behavior program, then it The behavior of these patients became con- would be difficult to see how non-conscious tinually driven by cues in the environment control over social behavior could be pos- and by little else. For example, bringing sible. Instead, as Posner and DiGirolamo the man onto a stage in front of a small (2000) recently remarked, the information- audience caused him to deliver an award processing and the neurophysiological levels acceptance speech; bringing the woman into of analysis of psychological processes have the (medical) doctor’s office caused her to achieved a level of mutual support greater give Dr. Lhermitte a physical exam com- than previously imagined. plete with injections of vaccines. Across these and several other situations, neither patient noticed or remarked on anything Implications for the Nature and unusual or strange about their behavior. Purpose of Consciousness Lhermitte (1986) concluded that they had suffered “a loss of autonomy: for the patient, There is a baffling problem about what the social and physical environments issue consciousness is for. It is equally baffling, the order to use them, even though the moreover, that the function of consciousness patient himself or herself has neither the idea should remain so baffling. It seems extraor- nor the intention to do so” (p. 341). dinary that despite the pervasiveness and Subsequent research in cognitive neuro- familiarity of consciousness in our lives, we science has largely supported Lhermitte’s are uncertain in what way (if at all) it is actually indispensable to us. (Frankfurt, deductions that this area of the prefrontal 1988,p.162) cortex is critical for the planning and control of action. Frith et al. (2000) concluded from Action tendencies can be activated and their review of this research that intended put into motion without the need for the P1: KAE 0521857430c20 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:41

social psychological approaches to consciousness 565

individual’s conscious intervention; even James (1890) argued, consciousness tends to complex social behavior can unfold without drop out of those processes where it is no an act of will or awareness of its sources. longer needed and thereby frees itself for Evidence from a wide variety of domains where it is. of psychological inquiry is consistent with In a very real sense, then, the purpose this proposition. Behavioral evidence from of consciousness – why it evolved – may patients with frontal lobe lesions, behavior be for the assemblage of complex non- and goal-priming studies in social psychol- conscious skills. In harmony with the gen- ogy, cognitive neuroscience studies of the eral plasticity of human brain development structure and function of the frontal lobes (see Donald, 2001), human beings – unlike as well as the separate actional and seman- even our nearest primate kindred – have tic visual pathways, cognitive psychological the capability of building ever more com- research on the components of working plex automatic “demons” that sublimely ﬁt memory and on the degree of conscious their own idiosyncratic environment, needs, access to motoric behavior – all of these and purposes. Intriguingly, then, one of the converge on the conclusion that complex primary objectives of conscious processing behavior and other higher mental processes at the level of the individual person may can proceed independently of the conscious be to eliminate the need for itself in the will. Indeed, the neuropsychological evi- future as much as possible, freeing itself up dence suggests that the human brain is for even greater things. It would be ironic designed for such independence. indeed, given the juxtaposition of automatic But this is not to say that conscious- and conscious processes in contemporary ness does not exist or is merely an epiphe- social psychology, if the evolved purpose nomenon. It just means that if all of these of consciousness turned out to be the cre- things can be accomplished without con- ation of ever more complex non-conscious scious choice or guidance, then the pur- processes. pose of consciousness (i.e., why it evolved) probably lies elsewhere. And the research described above points to one prime Acknowledgments candidate. That is, although we do not yet know Preparation of this chapter was supported much about how non-conscious goal pur- in part by Grant MH60767 from the U.S. suit capabilities develop, the most plausi- Public Health Service. ble guess is that they develop much as other automatic processes develop – out of frequent and consistent experience (Bargh, References 1990; Bargh & Gollwitzer, 1994; see Shiffrin & Dumais, 1981). This means in the case 1981 of automatic goal pursuit that the individ- Abelson, R. P. ( ). Psychological status of the script concept. American Psychologist, 36, 715– ual most likely consciously chose at one point 729. to pursue that particular goal in that par- Anderson, C. A., & Bushman, B. J. (2002, March ticular situation, then chose it again, and 29). The effects of media violence on society. so on until that goal representation became Science, 295, 2377–2379. associated so strongly with that situation Asch, S. E. (1952). Effects of group pressure representation that the former became auto- on the modiﬁcation and distortion of judg- matically associated with the latter. Then, ments. In G. E. Swanson, T. M. Newcomb, & entering the situation from then on causes E. L. Hartley (Eds.), Readings in social psychol- both the situation and the goal represen- ogy(2nd ed., pp. 2–11). New York: Holt. tations to become active, no longer with Atkinson, R. C., & Shiffrin, R. M. (1968). Human any need for conscious choice of that goal memory: A proposed system and its control (see Bargh & Chartrand, 1999). As William processes. In K. W. Spence & J. T. Spence P1: KAE 0521857430c20 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:41

566 the cambridge handbook of consciousness

(Eds.), The psychology of learning and motiva- research methods in social psychology (pp. 253– tion: Advances in research and theory (Vol. 2, 285). New York: Cambridge University Press. pp. 89–195). New York: Academic Press. Bargh, J. A., Chen, M., & Burrows, L. (1996). Baddeley, A. D. (1986). Working memory. New Automaticity of social behavior: Direct effects York: Oxford University Press. of trait construct and stereotype priming on Baddeley, A. D., & Hitch, G. J. (1974). Working action. Journal of Personality and Social Psychol- memory. In G. Bower (Ed.), Recent advances ogy, 71, 230–244. in learning and motivation (Vol. 8,pp.47–90). Bargh, J. A., & Ferguson, M. L. (2000). Beyond New York: Academic Press. behaviorism: On the automaticity of higher Bandura, A. (1986). Social foundations of thought mental processes. Psychological Bulletin, 126, 925 945 and action. Englewood Cliffs, NJ: Prentice- – . Hall. Bargh, J. A., & Gollwitzer, P. M. (1994). Bargh, J. A. (1989). Conditional automaticity: Environmental control over goal-directed Varieties of automatic influence in social per- action. Nebraska Symposium on Motivation, 41, 71 124 ception and cognition. In J. S. Uleman & J. A. – . Bargh (Eds.), Unintended thought (pp. 3–51). Bargh, J. A., Gollwitzer, P. M., Lee-Chai, A. Y., New York: Guilford. Barndollar, K., & Troetschel, R. (2001). The Bargh, J. A. (1990). Auto-motives: Preconscious automated will: Nonconscious activation and determinants of social interaction. In E. T. Hig- pursuit of behavioral goals. Journal of Personal- 1014 1027 gins & R. M. Sorrentino (Eds.), Handbook of ity and Social Psychology, 81, – . motivation and cognition (Vol. 2,pp.93–130). Bargh, J. A., & Thein, R. D. (1985). Individual New York: Guilford. construct accessibility, person memory, and the Bargh, J. A. (1992). Why subliminality does not recall-judgment link: The case of information matter to social psychology: Awareness of the overload. Journal of Personality and Social Psy- 1129 1146 stimulus versus awareness of its influence. In R. chology, 49, – . F. Bornstein & T. S. Pittman (Eds.), Perception Baumeister, R. F. (1998). The self. In D. T.Gilbert, without awareness (pp. 236–255). New York: S. T. Fiske, & G. Lindzey (Eds.), Handbook Guilford. of social psychology (4th ed., pp. 680–740). Bargh, J. A. (1994). The four horsemen of Boston: McGraw-Hill. automaticity: Awareness, intention, efficiency, Baumeister, R. F., Bratslavsky, E., Muraven, M., & and control in social cognition. In R. S. Wyer Tice, D. M. (1998). Ego-depletion: Is the active & T. K. Srull (Eds.), Handbook of social cogni- self a limited resource? Journal of Personality tion (2nd ed., Vol. 1,pp.1–40). Hillsdale, NJ: and Social Psychology, 74, 1252–1265. Erlbaum. Berkowitz, L. (1984). Some effects of thoughts on Bargh, J. A. (1996). Principles of automaticity. anti- and prosocial influences of media events: In E. T. Higgins & A. W. Kruglanski (Eds.), A cognitive-neoassociation analysis. Psycholog- Social psychology: Handbook of basic principles ical Bulletin, 95,410–427. (pp. 169–183). New York: Guilford. Brewer, M. B. (1988). A dual process model of Bargh, J. A. (1999). The cognitive monster: impression formation. In T. K. Srull & R. S. The case against controllability of automatic Wyer, Jr. (Eds.), Advances in social cognition stereotype effects. In S. Chaiken & Y. Trope (Vol. 1,pp.1–36). Hillsdale, NJ: Erlbaum. (Eds.), Dual process theories in social psychology Bruner, J. S. (1957). On perceptual readiness. Psy- 361 382 (pp. – ). New York: Guilford. chological Review, 64, 123–152. Bargh, J. A. (2001). The psychology of the mere. Buccino, G., Binkofski, F., Fink, G. R., Fadiga, L,, In J. A. Bargh & D. Apsley (Eds.), Unraveling the Fogassi, L., Gallese, V., Seitz, R. J., Zilles, K., complexities of social life: A Festschrift in honor of Rizzolatti, G., & Freund, H.-J. (2001). Action Robert B. Zajonc (pp. 25–37). Washington, DC: observation activates premotor and parietal American Psychological Association. areas in somatotopic manner: An fMRI study. Bargh, J. A., & Chartrand, T. (1999). The unbear- European Journal of Neuroscience, 13, 400–404. able automaticity of being. American Psycholo- Carlston, D. E., & Skowronski, J. J. (1994). Sav- gist, 54, 462–479. ings in the relearning of trait information as Bargh, J. A., & Chartrand, T. L. (2000). A practi- evidence for spontaneous inference generation. cal guide to priming and automaticity research. Journal of Personality and Social Psychology, 66, In H. Reis & C. Judd (Eds.), Handbook of 840–856. P1: KAE 0521857430c20 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:41

social psychological approaches to consciousness 567

Carver, C. S., & Scheier, M. F. (1981). Attention effects of social perception on social behavior. and self-regulation: A control-theory approach to In M. P. Zanna (Ed.), Advances in experimental human behavior. New York: Springer. social psychology (Vol. 33,pp.1–40). San Diego: Chaiken, S. (1980). Heuristic versus systematic Academic Press. information processing and the use of source Dijksterhuis, A., & van Knippenberg, A. (1998). versus message cues in persuasion. Journal of The relation between perception and behavior, Personality and Social Psychology, 39, 752–766. or how to win a game of Trivial Pursuit. Journal Chaiken, S., & Trope, Y. (Eds.). (1999). Dual of Personality and Social Psychology, 74, 865– process theories in social psychology. New York: 877. Guilford. Donald, M. (2001). A mind so rare. New York: Chartrand, T. L., & Bargh, J. A. (1996). Auto- Norton. matic activation of social information process- Eagly, A. H., & Chaiken, S. (1993). The psy- ing goals: Nonconscious priming reproduces chology of attitudes.New York: Harcourt Brace effects of explicit conscious instructions. Jour- Jovanovich. nal of Personality and Social Psychology, 71, 464– Erber, R., & Fiske, S. T. (1984). Outcome depen- 478. dency and attention to inconsistent informa- Chartrand, T. L., & Bargh, J. A. (1999). The tion. Journal of Personality and Social Psychol- chameleon effect: The perception-behavior ogy, 47, 709–726. link and social interaction. Journal of Personality Fazio, R. H. (1986). How do attitudes guide and Social Psychology, 76, 893–910. behavior? In R. M. Sorrentino & E. T. Hig- Chartrand, T. L., & Bargh, J. A. (2002). Noncon- gins (Eds.), Handbook of motivation and cog- scious motivations: Their activation, operation, nition (Vol. 1,pp.204–243). New York: and consequences. In A. Tesser, D. A. Stapel, & Guilford. J. V. Wood (Eds.), Self and motivation: Emerging Fazio, R. H. (1990). Multiple processes by which 13 41 psychological perspectives (pp. – ). Washing- attitudes guide behavior: The MODE model ton, DC: American Psychological Association. as an integrative framework. In M. P. Zanna Cohen, J. D., Dunbar, K., & McClelland, J. (1990). (Ed.), Advances in experimental social psychol- On the control of automatic processes: A paral- ogy (Vol. 23,pp.75–109). San Diego, CA: Aca- lel distributed processing account of the Stroop demic Press. 332 361 effect. Psychological Review, 97, – . Fazio, R. H., Jackson, J. R., Dunton, B. C., & Darley, J. M., & Latane, B. (1968). Bystander Williams, C. J. (1995). Variability in automatic intervention in emergencies: Diffusion of activation as an unobtrusive measure of racial responsibility. Journal of Personality and Social attitudes: A bona ﬁde pipeline? Journal of Per- Psychology, 8, 377–383. sonality and Social Psychology, 69, 1013–1027. Decety, J., & Grezes,` J. (1999). Neural mech- Fazio, R. H., & Williams, C. J. (1986). Attitude anisms subserving the perception of human accessibility as a moderator of the attitude- actions. Trends in Cognitive Sciences, 3, 172–178. perception and attitude-behavior relations: An Deci, E. L., & Ryan, R. M. (1985). Intrinsic moti- investigation of the 1984 presidential election. vation and self-determination in human behavior. Journal of Personality and Social Psychology, 51, New York: Plenum. 505–514. Descartes, R. (1972). Treatise of man (T. S. Hall, Fiske, S. T. (1989). Examining the role of intent: Trans.). Cambridge, MA: Harvard University Toward understanding its role in stereotyping Press. (Work originally published 1633) and prejudice. In J. S. Uleman & J. A. Bargh Devine, P. G. (1989). Stereotypes and prejudice: (Eds.), Unintended thought (pp. 253–283). New Their automatic and controlled components. York: Guilford. Journal of Personality and Social Psychology, 56, Fitzsimons, G. M., & Bargh, J. A. (2003). Think- 680–690. ing of you: Nonconscious pursuit of interper- Devine, P. G., & Monteith, M. J. (1999). Auto- sonal goals associated with relationship part- maticity and control in stereotyping. In S. ners. Journal of Personality and Social Psychol- Chaiken & Y. Trope (Eds.), Dual process the- ogy, 83, 148–164. ories in social psychology (pp. 339–360). New Fitzsimons, G. M., & Bargh, J. A. (2004). Auto- York: Guilford. matic self-regulation. In R. Baumeister & K. Dijksterhuis, A., & Bargh, J. A. (2001). The Vohs (Eds.), Handbook of self-regulation(pp. perception-behavior expressway: Automatic 151–170). New York: Guilford. P1: KAE 0521857430c20 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:41

568 the cambridge handbook of consciousness

Frankfurt, H. G. (1988). The importance of what we Knuf, L., Aschersleben, G., & Prinz, W. (2001). care about. New York: Cambridge University An analysis of ideomotor action. Journal of Exp- Press. erimental Psychology: General, 130, 779–798. Frith, C. D., Blakemore, S.-J., & Wolpert, D. M. Langer, E. J., Blank, A., & Chanowitz, B. (1978). (2000). Abnormalities in the awareness and The mindlessness of ostensibly thoughtful control of action. Philosophical Transactions action: The role of ‘placebic’ information in of the Royal Society of London, 355, 1771– interpersonal interaction. Journal of Personality 1788. and Social Psychology, 36, 635–642. Gilbert, D. T. (1989). Thinking lightly about LaPiere, R. T. (1934). Attitudes vs. actions. Social others: Automatic components of the social Forces, 13, 230–237. inference process. In J. S. Uleman & J. A. Bargh Lhermitte, F. (1986). Human anatomy and the (Eds.), Unintended thought (pp. 189–211). New frontal lobes. Part II: Patient behavior in com- York: Guilford. plex and social situations: The “environmental Goodale, M. A., Milner, A. D., Jakobsen, L. S., dependency syndrome.” Annals of Neurology, & Carey, D. P. (1991). Perceiving the world and 19, 335–343. grasping it: A neurological dissociation. Nature, Locke, E. A., & Latham, G. P. (1990). A theory 349, 154–156. of goal setting and task performance. Englewood Haney, C., Banks, C., & Zimbardo, P. G. (1973). Cliffs, NJ: Prentice-Hall. Interpersonal dynamics in a simulated prison. Mesulam, M.-M. (1986). Frontal cortex and International Journal of Criminology and Penol- behavior. Annals of Neurology, 19, 320–325. 69 97 ogy, 1, – . Milgram, S. (1963). Behavioral study of obedi- Hassin, R. R. (2004). Implicit working memory. ence. Journal of Abnormal and Social Psychology, In R. R. Hassin, J. S. Uleman, & J. A. Bargh 67, 371–378. (Eds.), The new unconscious. New York: Oxford Mischel, W., Cantor, N., & Feldman, S. (1996). University Press. Principles of self-regulation: The nature of Herr, P. M., Sherman, S. J., & Fazio, R. H. (1984). willpower and self-control. In E. T. Higgins On the consequences of priming: Assimilation & A. W. Kruglanski (Eds.), Social psychology: and contrast effects. Journal of Experimental Handbook of basic principles (pp. 329–360). Social Psychology, 19, 323–340. New York: Guilford. Higgins, E. T. (1996). Knowledge activation: Neuberg, S. L., & Fiske, S. T. (1987). Moti- Accessibility, applicability, and salience. In E. vational inﬂuences on impression formation: T. Higgins & A. T. Kruglanski (Eds.), Social Outcome dependency, accuracy-driven atten- psychology: Handbook of basic principles (pp. tion, and individuating processes. Journal of Per- 133–168). New York: Guilford. sonality and Social Psychology, 53, 431–444. Higgins, E. T., King, G. A., & Mavin, G. Nisbett, R. E., & Ross, L. (1980). Human infer- H. (1982). Individual construct accessibility ence: Strategies and shortcomings of social judg- and subjective impressions and recall. Jour- ment. Englewood Cliffs, NJ: Prentice-Hall. nal of Personality and Social Psychology, 43, Nisbett, R. E., & Wilson, T. D. (1977). Telling 35–47. more than we can know: Verbal reports on James, W. (1890). Principles of psychology. New mental processes. Psychological Review, 84, 231– York: Holt. 259. Jeannerod, M. (1999). To act or not to act: Norman, J. (2002). Two visual systems and two Perspectives on the representation of actions. theories of perception: An attempt to reconcile Quarterly Journal of Experimental Psychology, the constructivist and ecological approaches. 52A, 1–29. Behavioral and Brain Sciences, 24, 73–96. Johnson, M. K., & Hasher, L. (1987). Human Petty, R. E., & Cacioppo, J. T. (1984). The effects learning and memory. Annual Review of Psy- of involvement on responses to argument quan- chology, 38, 631–668. tity and quality: Central and peripheral routes Kelley, H. H. (1967). Attribution theory in to persuasion. Journal of Personality and Social 69 81 social psychology. In D. Levine (Ed.), Nebraska Psychology, 46, – . Symposium on Motivation (Vol. 15,pp. Posner, M. I., & DiGirolamo, G. (2000). Cogni- 192–241). Lincoln: University of Nebraska tive neuroscience: Origins and promise. Psycho- Press. logical Bulletin, 126, 873–889. P1: KAE 0521857430c20 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:41

social psychological approaches to consciousness 569

Posner, M. I., & Snyder, C. R. R. (1975). Atten- Thurstone, L. L. (1928). Attitudes can be mea- tion and cognitive control. In R. L. Solso sured. American Journal of Sociology, 33, 529– (Ed.), Information processing and cognition: The 554. Loyola symposium (pp. 55–85). Hillsdale, NJ: Tulving, E., & Thomson, D. M. (1973). Encoding Erlbaum. speciﬁcity and retrieval processes in episodic Prinz, W. (1997). Perception and action plan- memory. Psychological Review, 80, 352–373. ning. European Journal of Cognitive Psychology, Uleman, J. S. (1989). A framework for thinking 9, 129–154. intentionally about unintended thoughts. In J. Prinz, W. (2003). How do we know about our S. Uleman & J. A. Bargh (Eds.), Unintended own actions? In S. Maasen, W. Prinz, & G. thought(pp. 425–449). New York: Guilford. Roth (Eds.), Voluntary action: Brains, minds, Uleman, J. S., Newman, L. S., & Moskowitz, G. B. and sociality (pp. 21–33). New York: Oxford (1996). People as spontaneous interpreters: University Press. Evidence and issues from spontaneous trait Rizzolatti, G., & Arbib, M. A. (1998). Language inference. In M. P. Zanna (Ed.), Advances in within our grasp. Trends in Neurosciences, 21, experimental social psychology (Vol. 28,pp.211– 188–194. 279). San Diego, CA: Academic Press. Shiffrin, R. M., & Dumais, S. T. (1981). The Wegner, D. M., & Bargh, J. A. (1998). Control development of automatism. In J. R. Ander- and automaticity in social life. In D. T. Gilbert, son (Eds.), Cognitive skills and their acquisition S. T. Fiske, & G. Lindzey (Eds.), Handbook (pp. 111–140). Hillsdale, NJ: Erlbaum. of social psychology (4th ed., pp. 446–496). Shiffrin, R. M., & Schneider, W. (1977). Con- Boston: McGraw-Hill. trolled and automatic human information Wicker, A. W. (1969). Attitudes versus actions: processing: II. Perceptual learning, automatic The relationship of verbal and overt behavioral attending, and a general theory. Psychological responses to attitude objects. Journal of Social Review, 84, 127–190. Issues, 25 (4), 41–78. Srull, T. K., & Wyer, R. S., Jr. (1979). The role Wilson, T. D. (2002). Strangers to ourselves. Cam- of category accessibility in the interpretation bridge, MA: Harvard University Press. of information about persons: Some determi- Wilson, T. D., & Brekke, N. (1994). Mental con- nants and implications. Journal of Personality tamination and mental correction: Unwanted and Social Psychology, 37, 1660–1672. inﬂuences on judgments and evaluations. 117 142 Taylor, S. E., & Fiske, S. T. (1978). Salience, atten- Psychological Bulletin, 116, – . tion, and attribution: Top of the head phenom- Winter, L., & Uleman, J. S. (1984). When are ena. In L. Berkowitz (Ed.), Advances in experi- social judgments made? Evidence for the spon- mental social psychology (Vol. 11,pp.249–288). taneousness of trait inferences. Journal of Per- New York: Academic Press. sonality and Social Psychology, 47, 237–252. P1: KAE 0521857430c20 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:41

570 P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

CHAPTER 21 The Evolution of Consciousness

Michael C. Corballis

Abstract in our own species, Homo sapiens, perhaps around 100,000 years ago. This gave rise It is likely that many other species are con- to the “human revolution,” with dramatic scious of the physical world in much the advances in manufacture, art, and social same way that we humans are. It is often behavior. These developments may not have claimed, however, that only humans are self- altered the nature of consciousness itself, but conscious, or can represent the mental states they added to its content by vastly enriching of others, or can travel mentally through our mental and physical lives. time. I argue that the embracing characteristic that distinguishes such aspects of human cognition from that of other species Introduction is recursion. This feature not only permits us to mentally represent the mental states of Consciousness has so far managed to escape others but also underlies such activities as any universally agreed definition. Most mental time travel, manufacture, and lan- would agree that an earthworm is probably guage. Recursive operations are themselves not conscious and that we normally consider not necessarily conscious, but they permit a ourselves to be conscious at least some of the more advanced level of consciousness, pro- time. This relies partly on subjective experi- viding virtually infinite flexibility and gen- ence, from which we can conclude that we erativity. It is likely that recursion evolved are not only conscious but that we actually with the increase in brain size in the genus exist, as Descartes famously deduced. The Homo from about 2 million years ago, prob- introspective certainty that we are conscious ably as a result of increasing social demands. is of little help in understanding the evo- Generative language probably evolved dur- lution of consciousness, although Wilhelm ing the first million years of this period, and Wundt (1894), who built an experimen- switched gradually from a manual to a vocal tal psychology on the introspective method, mode, culminating in autonomous speech nevertheless was able late in his career to 571 P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

572 the cambridge handbook of consciousness

make some pronouncements about the ani- ing question then, however, was whether mal mind. Even so, the identification of con- the animals had the subjective experience sciousness in non-human species must surely of seeing the light in the blind field. When be based largely on behavioral rather than given the choice of three responses, one for introspective evidence, whatever conscious a left light, one for a right light, and one for empathy we may experience through gazing no light, the animals consistently chose “no into the soulful eyes of a chimpanzee. light” when the light was in the blind field. Historically, though, animal behaviorists That is, monkeys seem to be able to report have not been much help since J. B. Wat- on their own subjective experience of see- son’s (1913) famous manifesto of the behaving nothing when a light was presented in iorist revolution that banished all mention of the blind field, even though they were able consciousness, or even of mind. There was to respond to it accurately. at least a whiff of consciousness, however, in There is also evidence that monkeys can the later distinction between instrumental or perform such mental operations as mental operant conditioning, on the one hand, and rotation and memory scanning (Georgopou- respondent or classical conditioning, on the los & Pellizzer, 1995). In his classic study other. A dog classically conditioned to sali- of problem solving in chimpanzees, Kohler¨ vate at the sound of a bell presumably makes (1925) provided compelling evidence that no conscious decision to salivate, although the animals were able to solve problems it may arguably be consciously aware of its mentally before demonstrating the solutions unseemly drool. A pigeon pressing a key to in practice. For example, a chimpanzee may achieve a food reward, however, may well understand that one can haul in a banana be aware of the contingency that leads it that is out of reach by using a rake, or even by to peck, and there are well-articulated laws extending the length of a rake by adding an governing the rates of pecking under dif- extra piece to the handle. The chimpanzees ferent reinforcement regimes (Davison & often seemed to solve these problems sud- McCarthy, 1987; Ferster & Skinner, 1957). denly, as though through insight rather than Despite the traditional lack of interest trial and error. Tomasello (1996) has demon- in consciousness among behaviorists, there strated similar problem-solving abilities in have been recent attempts to develop prox- chimpanzees, but notes that chimpanzees ies for introspection in non-human ani- do not learn to solve problems by imi- mals. For example Logothetis (1998) studied tating others, whereas human children do. bistable percepts in macaques by teaching Povinelli (2001) suggests that chimpanzees them to give different responses to different are actually fairly limited in their ability to stimuli and then placing the two stimuli in solve physical problems. In one experiment, binocular rivalry. The animals then indicate chimpanzees were taught how to reach a which stimulus they perceive at any given hooked tool through holes in a plexiglass time by giving the response to that stim- screen and retrieve a banana that was just ulus, implying report of their own internal out of reach. The banana was on a piece of percepts. That is, the animals are indicating wood with a vertical post at one end and a what they see, rather than simply respond- ring at the other, and the animals learned ing to the presence or absence of a stimulus. to hook the ring and haul in the banana. Cowie and Stoerig (1995) have also devel- But when the ring was removed, they did oped a way of testing blindsight in mon- not seem to understand that they could use keys that seems tantamount to subjective the tool to hook the post and haul in the report. After unilateral removal of the striate banana that way. Whatever the limitations of cortex, monkeys were able to reach accu- these thought-provoking experiments, there rately for panels that were lit up in either seems no reason to doubt that great apes, visual field, thereby demonstrating blind- and probably other primates as well, have sight – just as humans with striate-cortex some internal computational abilities, and it lesions do (Weiskrantz, 1986). The interest- is probably also reasonable to suppose that P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

the evolution of consciousness 573

the implied mental activity is in some sense nation. This point was revived by Chom- subjective – although we can never know sky (1966), who similarly pointed out the for sure. generativity of language and emphasized its Nevertheless there seems good reason to uniqueness to humans. believe that other primates, at least, are con- The emphasis on language as the entry scious of the physical world in much the into the consciousness of others, or even as same ways as we humans are and that they a criterion of consciousness itself, and the can also think consciously about operations Chomsky-inspired insistence that language on the world, even if in rather primitive is uniquely human may have at least tem- fashion. It is sometimes suggested, however, porarily closed off the possibility of admit- that there are additional dimensions of con- ting consciousness in non-human animals, sciousness that only humans possess. As we although there was something of a chal- shall see, for example, it is claimed that only lenge from those who tried to demonstrate humans have an awareness of self or of the language-like behaviors in apes (e.g., Gard- thoughts, perceptions, and feelings of others. ner & Gardner, 1969; Miles, 1990; Patterson, The rest of this chapter is concerned with 1978; Savage-Rumbaugh, Shanker, & Taylor, these more complex, social aspects of con- 1998) and dolphins (e.g., Herman, Richards, sciousness and the claim that at least some & Wolz, 1984). The possibilities of animal of these are uniquely human. Again, though, consciousness were revived somewhat, and the problem is largely one of method. How broadened, when Premack and Woodruff can we know, for example, whether a chim- (1978), frustrated by the lack of progress in panzee or a bonobo can tell what another the ape-language research, opened the door individual ape is thinking? a little wider with the question: “Does the Of course, this problem applies as much chimpanzee have a theory of mind?” (p. 515). to other people as to animals. How can I Yet even if the answer is no, we might still be sure that you, the reader, are subjectively be reluctant to deny some form of conscious- aware of what I am trying to convey through ness to our cat or dog, or to the New Cale- these very words? How do I know that all my donian crow that manufactures tools so skill- friends are not simply zombies, programmed fully to extract foodstuffs from holes (Hunt, to act intelligently in my presence? Part of 1996), or to the above-mentioned chim- the answer comes from the fact that we can panzees with their problem-solving exploits. interrogate people as to what is going on in The question then is whether we can gain a their minds through language. Indeed, lan- better understanding of an animal’s behavior guage is not only the principal access into by attributing consciousness to it. How are the subjective lives of other people but it we to do this? is also sometimes taken as integral to consciousness itself. Bickerton (1995), for example, proposed that “the peculiar properties The Intentional Stance of a distinctively human intelligence as such are derived from the possession of language” One approach is that of the intentional and went on to suggest that “consciousness stance, which seems to offer a way out of rad- as we know it may arise from an identical ical behaviorism and its denial of conscious- source” (p. 7). The idea that language dis- ness without resorting to introspectionism tinguishes humans from other animals goes and the implied inaccessibility to animal back at least to Descartes (1985/1647), who minds (Dennett, 1983). The characteristics thought that non-human animals were mere that identify intentional as distinct from machines, operating according to mechani- non-intentional systems, at least as presented cal principles, whereas humans possessed a by Dennett, may seem counter-intuitive and freedom of thought and action that set them have to do with the nature of statements apart. Language, in particular, had an open- about the world. In most true statements, endedness that deﬁed any mechanical expla- one can replace a term in the statement by P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

574 the cambridge handbook of consciousness

another term that refers to exactly the same over, the intentional stance is not water- thing, and the statement remains true. For tight because it is always possible to devise example, if Jack is the farmer’s son, then the purely behavioral accounts that make no ref- sentence “This is the house that Jack built” erence to subjective terms (or more formally, can be replaced by “This is the house that that do not exhibit referential opacity). For the farmer’s son built.” The truth or falsity example, we may be willing to attribute of the ﬁrst sentence is not altered by the belief to the egregious pigeon if it does not substitution. peck the key if it happens to notice that there Statements of intentionality, in contrast, is no food in the food hopper. It now does exhibit referential opacity, such that substi- not believe that it will receive food if it pecks tuting a term with one that refers to the same the key, and being a rational bird it stops. thing can alter its truth or falsity. Dennett However, we could just as easily explain the gives an example taken from Bertrand Rus- behavior in terms of environmental contin- sell (1905). Consider the sentence “George gencies. The bird pecks if the hopper con- IV wondered whether Scott was the author tains food, but does not peck if the hopper of Waverly,” and suppose it to be true. It is is empty. also true that Sir Walter Scott did write the My guess is that it will always be possible novel Waverly, so one might consider substi- to convert an intentional statement about tuting “Scott” for “the author of Waverly”: an animal’s behavior into a statement that “George IV wondered whether Scott was does not include intentional clauses, but is Scott.” It is unlikely (though remotely possi- expressed rather in terms of environmen- ble) that George IV actually did wonder this, tal contingencies. This means that there will so in this case the substitution does change always be those who reject claims of ani- the sentence from a true one to a false one. mal consciousness – from Watson (1913)to In other words, referential opacity creates a Skinner (1957) to Heyes (1998, 2001). The barrier to logical analysis. two points of view may remain as implaca- To down-to-earth scientists these consid- bly opposed as the particle and wave theo- erations may seem far removed from con- ries of light. And there may never be a cru- sciousness, but referential opacity picks out cial experiment ( pace Heyes, 1998) that can precisely those clauses that include sub- distinguish them. jective terms like “believes (that),” “knows Does that mean we should abandon the (that),” “expects (that),” “wants (it to be quest for animal consciousness? For the pre- the case that),” “recognizes (that),” “won- sent, at least, I think it is worth taking the ders (whether),” “understands (that),” and intentional stance, because it offers a vocab- so forth. The question then is, do we get a ulary that can address the important issues better account of some behavioral act if we and suggest interesting experiments. This is include one of these terms, thereby creat- not to say that a purely behavioral account ing referential opacity? For example, do we might not eventually do as well, if not bet- gain a better account of a pigeon’s behavior ter, but it is the intentional stance that seems if we conclude not just that it pecks a key to be making the running, perhaps begin- and then receives food but also that it pecks ning with Premack and Woodruff’s (1978) the key in the belief that this will produce famous question about theory of mind in food? the chimpanzee. Thanks to Dennett’s for- By attributing subjective states, the inten- malism, the deﬁnition of intentional clauses tional stance does not appeal to introspec- removes the taint of introspectionism, and tion – you don’t actually ask the pigeon to questions about believing, wanting, recogni- introspect about its behavior and tell you zing, etc, can be translated into objective what it thinks. The important tests are still experiments. The fact that such terms map behavioral, and the question is whether the onto introspective states can even be consi- attribution gives a more compelling account dered an advantage because we are accus- of observed behavior than would otherwise tomed to thinking in such terms, and this be the case. As Dennett recognized, more- can help in setting up the important tests of P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

the evolution of consciousness 575

animal consciousness. In this chapter I therein a sequence involves one or more of the fore continue to use subjective terms like preceding terms. A simple example is the think, understand, feel, and imagine, but for use of integers for counting. So long as we the philosophically pure it should be under- have rules that allow us to generate a number stood (sic) that these terms owe their spe- from the preceding number in a sequence, cial status to referential opacity, and not to there is no limit to the number of objects we whatever it is that we experience in our own can count. The basic rule, of course, is sim- heads. In other words, these terms should be ply to add 1 to the preceding number. This is unqualiafied (again, sic). not an entirely trivial exercise because there One major reason to prefer the inten- are subrules that tell you how to do this by tional stance to the purely behavioral one starting with the rightmost digit, and deal- is that it allows us to define different orders ing with the special case where the right- of intentionality. This may be helpful, if not most digit or digits equal 9. The point is that crucial, in understanding what is different the procedure is recursive, and the sequence about the human mind and human con- can be continued indefinitely. Counting is sciousness and what may be different about thus an example of what Chomsky (1988) the minds of the great apes relative to other has called “discrete infinity.” primates. The example of counting illustrates that recursion is not restricted to intentional states – a computer can count without any- Orders of Intentionality, one suspecting it of consciousness. Con- and Recursion sider also the following well-known children’s story: Zero-order intentionality simply involves one subjective term, as in “Alice wants Fred This is the house that Jack built. to stop bugging her.” First-order intention- This is the malt that lay in the house that ality would involve two such terms, as in Jack built. “Ted thinks Alice wants Fred to stop bugging This is the rat that ate the malt that lay in the house that Jack built. her.” And so to second-order: “Alice believes This is the cat that killed the rat that ate the that Fred thinks she wants him to stop malt that lay in the house that Jack built. bugging her.” This is the dog that worried the cat that It should be clear from these exam- killed the rat that ate the malt that lay in ples that the different orders of intention- the house that Jack built. ality involve recursion and can be continued indefinitely. There seems some reason This increasingly recursive sequence does to believe, though, that we humans run out not involve intentional clauses, although in of grunt at about the fifth or sixth order telling the story there is perhaps implicit int- (Cargile, 1970), perhaps because of limited entionality. Bennett (1976) has argued that working memory capacity, rather than any genuinely communicative speech requires intrinsic limit on recursion itself. We can at least three orders of intentionality. In perhaps just wrap our minds around propo- telling you the story, I intend you to recognize sitions like “Ted suspects that Alice believes that I want you to understand that “This is that he does indeed suspect that Fred thinks the house that Jack built,” etc. Otherwise I that she wants him (Fred) to stop bugging might as well be talking to the moon. her.” That’s fourth-order, as you can tell It should be understood that stories like by counting the italicized words and sub- The House that Jack Built are not simply tracting one. You could make it fifth-order sequential events. It is not simply that a dog by adding “George imagines that . . . ” at the worried a cat, a cat killed a rat, and so forth; beginning. the recursion implies that we are specifying More formally, recursion is a mathemati- a rather specific dog. We are not referring to cal device for generating terms in a sequence, a dog that worried a cat that killed a rat that where the rule for generating the next term did not eat the malt that lay in the house P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

576 the cambridge handbook of consciousness

that Jack built. In a line-up of suspects, such S

a dog would not do. (Rule 1) We can also express the story using NP VP embedded phrases, which allows us to focus (Rule 2) (Rule 4) on a particular actor in the story, such as the art + noun + RC verb + NP unfortunate malt: (Rule 3) (Rule 2) The malt that the rat that the cat that the relpron + VP art + noun dog worried killed ate lay in the house that (Rule 4) Jack built. verb + NP And this, I think, again pushes the bound- (Rule 2) aries of human understanding, although I art + noun find that if I take it slowly I can get there. In presenting these examples, I have The dog that chased the cat killed the rat also illustrated the recursiveness of lan- Figure 21.1. This tree diagram shows how the guage itself – we need recursive language to recursive use of rules can produce a sentence express recursive thoughts. In linguistics, the with an embedded clause. recursiveness is often expressed in terms of rewrite rules in which the same term can and competition, of guess and second-guess. appear on either side of the expression defin- The recursive nature of social relations is ing a rule. For example, The House that Jack well conveyed in literature and the theater. Built involves a succession of relative clauses In Jane Austen’s Pride and Prejudice, Eliza- into which we can substitute more relative beth Bennet thinks that Darcy thinks that she clauses; thus thinks he thinks too harshly of her family. Or relative clause = relative clause in Shakespeare’s Twelfth Night, Maria fore- + relative clause. sees that Sir Toby will eagerly anticipate that Olivia will judge Malvolio absurdly imperti- Or we could break it down into a more com- nent to suppose that she wishes him to regard plex set of rules, as follows: himself as her preferred suitor.1 One might add a further order of recursion by prefix- (Rule 1)S NP+ VP ing each of these situations with the phrase, (Rule 2) NP article + noun + [RC] (Rule 3) RC relative pronoun + VP “The audience understands that . . . ” (Rule 4) VP verb + [NP] Recursion might be said to add a further dimension to consciousness (see Chap- where S = sentence, NP = noun phrase, ter 15). Tulving (2001) has distinguished RC = relative clause, and VP = verb noetic consciousness, which implies know- phrase, and the bracketed items are optional. ing, from what he calls autonoetic con- These rules are sufficient to generate sen- sciousness, which implies knowing that one tences like “The dog that chased the cat knows. According to Tulving, autonoetic killed the rat,” as shown in Figure 21.1. consciousness is uniquely human and underlies the concept of self. One example of autonoetic consciousness is episodic mem- The Recursive Mind ory, which is memory for events in one’s life that have specific locations in time Hauser, Chomsky, and Fitch (2002), dis- and space. Episodic memory is distinguished cussing the evolution of language, have from semantic memory, which is memory argued that recursion is a uniquely human for generic facts, and is noetic rather than mental accomplishment. People can count autonoetic. Episodic memory does seem to recursively, think recursively about the men- depend on a mental reconstruction or re- tal states of others (and of themselves), and enactment of some earlier experience and use recursive syntax to tell stories. Our social implies first-order intentionality of the form lives involve a subtle calculus of cooperation I experience that I experienced X. One can P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

the evolution of consciousness 577

remember that one saw something, remem- states. That is, we might easily concede zero- ber that one felt ill, or even remember that order intentionality, or noetic consciousness. one once remembered something now for- It may well be reasonable to suppose that gotten. a pigeon in some sense “believes” that if it Episodic memory may be regarded as pecks a key it will receive food, but it is part of a more general ability to engage in problematic as to whether it believes this mental time travel, which involves imagin- of another pigeon. Much of the discussion ing the future, as well as the past (Sudden- and research effort has been essentially con- dorf & Corballis, 1997). There is evidence, cerned with whether first-order intentional- for example, that the loss of episodic mem- ity can be demonstrated in any other species, ory due to brain injury also leads to a loss of and most particularly in the great apes. the ability to engage in what has come to be called episodic future thinking, and which Theory of Mind in Great Apes? probably depends on the prefrontal cortex (Atance & O’Neill, 2001; Wheeler, Stuss, & Premack and Woodruff (1978), who set the Tulving, 1997). Although we tend to think of ball rolling, suggested a number of exper- memory as factual and the future as hypo- iments that might demonstrate first-order thetical, both are probably constructed from intentionality and applied a few of them to a mixture of fact and fantasy. By including Sarah, a captive chimpanzee. One of their the future as well as the past, we essentially techniques was to show videos of a human create an extended sense of self that allows grappling with some problem and then us to create the entity that distinguishes us offer Sarah a choice of photographs, one from others. of which depicted a solution to the prob- Yet recursion runs deeper than autonoetic lem. One such test showed a woman try- consciousness and the concept of self. We ing to escape from a locked cage, and one can think recursively well beyond first-order of the photos showed a key, whereas oth- intentionality, as explained earlier. Recursive ers showed objects irrelevant to the task. thinking need not have anything to do with Sarah performed quite well at choosing the the self or even with levels of knowing. Our appropriate photo, although as Premack and counting system is recursive, as is language Woodruff recognized, this need not show itself. Human manufacture is also recursive, that she appreciated what was going on and in a review of evidence on toolmaking, in the mind of the person depicted. For Beck (1980,p.218) noted, “Unquestionably example, the key might have been selected man [sic] is the only animal that to date has through simple association with the cage. been observed to use a tool to make a tool.” One researcher who has taken up Pre- In addition, we create objects, or compo- mack and Woodruff’s challenge is Povinelli, nents, that are then used in the manufac- who has tried further tests to determine ture of higher-order objects, which are then whether a chimpanzee can understand what combined into yet higher-order ones. From is going on in the mind of a person (see bricks come houses, and from houses vil- Povinelli, Bering, & Giambrone, 2000, for lages. And there are wheels within wheels. a summary). The results have been largely negative. Chimpanzees readily approach humans to beg for food, and this behavior Recursion in Other Species provided an opportunity to check whether in so doing they are influenced by whether The claim that only humans are capable of the person can see or not. But when offered recursion does not mean that other animals the choice of two individuals to beg from, don’t possess consciousness. There seems no one with a blindfold over her eyes, the ani- reason to deny at least some other animals mals did not systematically choose the one pleasure or pain, or even belief, and we saw who could see. The same was true when one earlier that primates, at least, can appar- of the people had a bucket over her head or ently give overt reports on their own internal covered her eyes with her hands. Only when P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

578 the cambridge handbook of consciousness

one of the people was actually facing the There is some evidence, though, that the other way did the chimpanzees easily choose work of Povinelli and his colleagues under- the one facing toward the animal. Young estimates the social intelligence of the chim- children, on the other hand, quickly recog- panzee. Chimpanzees are by nature com- nize that they should approach the person petitive creatures, and one may wonder who can see them and understand that this why they should trust what humans are depends on the eyes. The failure of the chim- trying to indicate. Dogs, in contrast, have panzee to appreciate this fact does not arise been bred to cooperate with humans, and from failure to observe the eyes, because Hare and Tomasello (1999) have shown that they readily follow the gaze of a person con- dogs do seem to be able to choose food fronting them. Chimpanzees may eventually sources according to where either a person choose the person who can see them, but the or another dog is looking or pointing. It is behavior is more simply explained as being of course possible that some dogs, through due to associative learning, and not due to selective breeding, have acquired an order the understanding that eyes are for seeing. of intentionality beyond that achieved by Another test depends on the chim- the chimpanzee, although it is perhaps more panzee’s apparent understanding of point- likely that they have simply learned a skill ing. If a person sits in front of a chimpanzee that does not involve any understanding of and points to one of two boxes on the left what goes on in the mind of the pointer. or right, the chimpanzee understands read- Hare, Call, Agnetta, and Tomasello (2000) ily enough that if it wants food, it should have also shown that chimpanzees are aware go to the box that the person is pointing to. of what other chimpanzees can see and mod- But the choice breaks down if the person ify their behavior accordingly. For example, points from some distance away and is sys- a chimpanzee will approach food when a tematically reversed if the person sits closer more dominant chimpanzee cannot see the to the box that does not contain the food and food, but will be reluctant to do so when points to the other one. It seems that chim- they can see that the dominant chimpanzee panzees respond on the basis of how close is watching. the pointing hand is to the box containing This is an example of tactical deception. the food, and not on the basis of where the Deception itself is widespread in nature, hand is actually pointing. whether in the camouflage of a butterfly Yet chimpanzees do follow eye gaze wing or the uncanny ability of the Aus- (Tomasello, Hare, & Agnetta, 1999), just as tralian lyre bird to imitate the sounds of humans do, and this may suggest that they other species – including, it is said, the sound have at least some understanding that oth- of a beer can being opened. Tactical decep- ers can see. Povinelli argues, though, that tion, however, is that in which the decep- behaviors like following eye gaze have the tion is based on an appreciation of what same basis in humans as in other primates, the deceived animal is actually thinking or but that we “reinterpret” these behaviors as what it can see. From primate researchers being more sophisticated than they really are working in field settings, Whiten and Byrne (Povinelli & Bering, 2002). For example, we (1988) collected a database of anecdotal evi- may spontaneously follow the gaze of some- dence suggesting tactical deception. They one who seems to be gazing at something screened the reports to eliminate cases in in the sky without going through an intel- which the animals might have learned to lectual (and presumably conscious) exercise deceive through trial and error and con- along the lines of the following: “That fel- cluded that only the four species of ape occa- low must be able to see something up there sionally showed evidence of having deceived that’s interesting.” Gaze following may simon the basis of an understanding of what the ply be an adaptive response that alerts other deceived animal could see or know. Even so, animals to danger or reward, but we humans there were relatively few instances – only 12 have intellectualized it, often after the fact. from common chimpanzees and 3 each from P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

the evolution of consciousness 579

bonobos, gorillas, and orangutans – so there panzee who passes the test may recognize remains some doubt as to whether tactical its mirror image as its own body, but still not deception truly shows that great apes can attribute its mental states to that image. “read the minds” of others. What of mental time travel? Curiously, Leslie (1994) distinguished between two the main challenge to the idea that mental levels of what he called the Theory of Mind time travel is uniquely human has come, not Mechanism (ToMM). A child or a chimp from apes or even primates, but from birds. with ToMM-1 can attribute goal directed- Clayton, Bussey, and Dickinson (2003)have ness and self-generated motion to others, shown that scrub jays cache food in dif- whereas a child with ToMM-2 attributes ferent locations and then unerringly return full-blown intentional states, such as believ- to the appropriate locations to recover the ing, desiring, seeing, or remembering, to oth- food. Moreover, they calibrate food recovery ers. Tomasello and Rakoczy (2003) make a according to how long it has been cached. very similar distinction). ToMM-1 does not For example, if they cache worms in one seem to enable recursion; attributing self- location and peanuts in another, after 4 generated motion to another creature car- hours they will return to the location con- ries no implication that the other creature taining the worms, which they prefer to can also attribute self-generated motion. But peanuts. But if they have already learned that once desires and beliefs are attributed, as in worms become inedible if left buried for too ToMM-2, then recursion can follow because long, and if they are tested 24 hours later, one can attribute the attribution of such they will recover the peanuts in preference desires and beliefs. There seems no reason to the worms. According to Clayton et al., at present to suppose that the great apes this means that they remember what they can advance beyond ToMM-1 at most, and have cached, where they have cached it, and Hauser and Carey (1998) suggest that “the when they cached – thereby passing what intellectual tie breaker between humans and they call the “www test” for episodic mem- nonhumans is likely to lie within the power ory. But are we really to believe that these of ToMM-2.” A person with full recursive birds travel mentally in time to re-enact the capabilities would bestow upon others all of actual experience of caching food? It seems the mental states that they themselves expe- more likely that they have evolved simpler rience, and do so recursively. mechanisms for recovering cached food, perhaps attaching some internal timing device to each location that specifies how long ago Self-Awareness in Other Species? food was stored there. It has been shown that great apes (with the There is some evidence, reviewed by possible exception of the gorilla) pass the so- Schwarz and Evans (2001), that primates called mirror test, implying self-awareness. can also remember where specific events A rigorous form of the test is first to anes- occurred. In one example, a chimpanzee thetize the animal and then apply red dye taught to use lexigrams to represent objects to an eyebrow ridge, and then, when the was able to select a lexigram for a food animal awakes, to allow it to see itself in item and then point to a location where that a mirror. Chimpanzees with prior experi- item had been hidden some time beforehand ence of mirrors reach up and touch the red (Menzel, 1999). In a slightly more compli- marks on their faces, whereas those with- cated example, Schwarz, Colon, Sanchez, out experience with mirrors react as though Rodriguez, and Evans (2002) report evi- there is another chimpanzee in the mirror, dence that a gorilla had encoded both a food as do rhesus monkeys even after many years item and the person who had previously of experience (Gallup, 1998). (Gorillas, curi- given him the food. But in none of reviewed ously, do not seem able to pass the test – see cases was there any evidence that the ani- Gallup, 1997). But even this test need not mals had coded the time of the past event, imply more than ToMM-1. That is, a chim- so the www criteria were not fully satisfied. P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

580 the cambridge handbook of consciousness

Social Behaviors a disposition that has been selected in evolution, but that need not imply any under- Whether or not apes have ToMM-2, or first- standing of what goes on in the minds of order intentionality, or mental time travel, others. there seems no need to deny them rich social The evidence on theory of mind in the lives. Most primates, with some exceptions great apes seems to have left researchers such as the solitary orangutan, are social as divided as ever they have been on the creatures with established hierarchies of issue of continuity and discontinuity. In the dominance, who form coalitions and friend- Introduction to a special issue of Cogni- ships, show empathy, and help each other in tive Science devoted to primate cognition, distress. Indeed, de Waal (1996) has claimed Tomasello (2000) wrote as follows: that the basic ingredients of morality can be discerned in the behavior not just of pri- H]uman cognition is a specific (in the literal mates but also of other animals, including meaning of the word) instance of primate marine animals and dogs. Animals have often cognition, and evolution by means of nat- been observed to perform acts of seeming ural selection is a mostly conservative pro- generosity, in which there is a cost to the cess that preserves adaptations for as long as they work. Human cognition is thus not giver but a benefit to the receiver. Exam- just similar to nonhuman primate cogni- ples include food sharing, especially promi- tion, it is identical in many of its structures. nent among chimpanzees, and alliances to The study of nonhuman primate cognition ward off aggression, where one animal will should therefore play a more important role come to the aid of another despite the in cognitive science than is currently the risk of injury. The principle underlying case (p. 351). such behavior is thought to be reciprocal altruism, such that the giver eventually That says it for continuity, but it did not receives a reciprocal benefit from the orig- take long for the opposite view to emerge inal receiver. An important feature of recip- from other researchers in primate cogni- rocal altruism is that the payback is not tion. Povinelli and Bering (2002) decry the immediate. As de Waal points out, primates overzealous attempts “to dismantle argu- form friendships, just as humans do, with the ments of human uniqueness” while never- effective guarantee of reciprocal altruism in theless reaffirming the importance of com- future interactions. parative psychology: “A true comparative Evolutionary accounts of reciprocal altru- science of animal minds . . . will recognize ism had been assumed to apply only to the complex diversity of the animal king- genetically related animals, enhancing the dom, and will thus view Homo sapiens as survival of shared genes, until Trivers (1971) one more species with a unique set of adap- gave a plausible evolutionary account of tive skills crying out to be identified and reciprocal altruism among non-kin. Recipro- understood” (p. 115). cal altruism may well underlie the increase in My guess is that it is indeed recursion group size in primate evolution, culminating that sets the human mind apart. The great in the large groupings that we see in human apes may have evolved a limited repertoire societies – and making the perils of war that of recursive thought processes, but these are much more horrific. Yet reciprocal altruism largely situation-specific and perhaps more need not imply intentionality. We humans dependent on trial and error or on emo- form friendships, or go to the aid of strangers tional traits than on logical reasoning. Chim- in distress, usually without thinking explic- panzees may well have learned that it is a itly of any future benefit – although there are good idea to not to behave in particular ways of course exceptions, as in those who culti- when the dominant male is watching, and vate the rich. In that respect, the behaviors this is sometimes elevated (or degraded) into that reflect reciprocal altruism may be like deliberate deception. But there is no con- the response to pointing, referred to above – vincing evidence, as far as I know, for levels of P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

the evolution of consciousness 581

intentionality beyond zero-order intention- mately recursively, in the calculus of social ality and therefore no convincing evidence of relations. recursive thought in any of the great apes, or When is this likely to have occurred? indeed in any non-human species. But once It was probably late in hominid evolution, recursive thinking is established, then the because brain size remained approximately step beyond zero-order to higher orders of constant from the divergence from the chim- intentionality may follow quite easily, but it panzee line until around 2 million years ago. is a step that has massive implications. It is Further, there was little evidence for sophis- a bit like the discovery that, having learned ticated tool use during that period. There to count to 10, one can then use recursion to are a number of reasons to believe that the carry on counting forever. progression from protolanguage and recur- What does this have to do with con- sive thought may have begun with the emer- sciousness itself? As we have seen, recur- gence of Homo erectus some 2 million years sion itself does not imply consciousness. ago. Around that time, hominids began to However, the ability to think recursively migrate out of Africa into Asia and later into may have endowed our species with an Europe (Tattersall, 1998), and the Acheulian extra dimension of consciousness that under- industry emerged, with large bifacial tools lies much of our complex social behavior, and handaxes that seemed to mark a signif- including the ability to express that com- icant advance over the simple flaked tools plexity in the form of language. of the earlier Oldowan industry (Gowlett, 1992) – although, as we see, the advancement of manufacturing techniques remained Evolution of the Recursive Mind relative slow until comparatively recently. Brain size also began to increase dra- The question now arises as to how and when matically (Deacon, 1997). Indeed Chomsky the mind began to take on this recursive (1975) has suggested that language may have aspect. If chimpanzees do not think recur- emerged simply as a consequence of possess- sively, then recursion must have evolved in ing an enlarged brain, without the assistance the 6 or 7 million years since the hominids of natural selection: split from the line leading to modern chimpanzees and bonobos. We know very little about what happens 10 As noted earlier, it is often supposed that when 10 neurons are crammed into some- recursion was driven by language, and there thing the size of a basketball, with further is little doubt that language has a recur- conditions imposed by the specific manner in which this system developed over time. sive structure. Yet there seems little reason It would be a serious error to suppose that to believe that recursive syntax evolved de all properties, or the interesting structures novo. It must surely have evolved as a vehi- that evolved, can be ‘explained’ in terms of cle for the expression of recursive thought. natural selection (p. 59). Bickerton (2000) has argued that it evolved from reciprocal altruism, which requires Although this seems to ignore the selective keeping track of the variety of events that processes that must have led to the increase take place within an animal’s social group. in brain size in the first place, it is plausible to It implies a distinction between individu- suppose that an enlarged brain provided the als (you need to know who your friends extra circuitry required, if not for language are), a recognition of actions, and a knowl- alone, then for recursive thought generally. edge of payback structures. According to Another factor may have been the pro- Bickerton, this leads to setting up the cate- longation of childhood, which was probably gories of agent, theme, and goal. As the com- an indirect result of the increase in the size of plexity of social life increased the demands the brain, and therefore of the head. To con- on these categories must have grown, so form to the general primate pattern, human that they could be used flexibly, and ulti- babies should be born at around 18 months, P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

582 the cambridge handbook of consciousness

not 9 months (Krogman, 1972), but as any allows for the cumulative development of mother will know this would be impossible, culture, because information can be trans- given the size of the birth canal. The brain mitted verbally from generation to genera- of a newborn chimpanzee is about 60%of tion, bypassing the slow route of evolution- its adult weight; that of a newborn human is ary change. only about 24%. This feature of prolonged As suggested earlier, another possible childhood probably emerged in evolution manifestation of recursive consciousness is with the increase in brain size, and there mental time travel. Perhaps it was this that is evidence that it was present in Homo led to an appreciation of impending death erectus by 1.6 million years ago (Brown, and the emergence of burial rituals. There is Harris, Leakey, & Walker, 1985). Our pro- some evidence that the Neanderthals buried longed childhood means that the human their dead (Smirnov, 1989), perhaps even brain undergoes most of its growth while with offerings of flowers (Solecki, 1975), exposed to external influences and is there- although the evidence remains controversial fore tuned more finely to its environment. (Noble & Davidson, 1996). There is also evi- There is some reason to believe that learn- dence of cut marks inflicted on the crania of ing during growth may be the key to recur- the dead from the earliest Homo (Pickering, sion. Elman (1993) has devised a network White & Toth, 2000) to early Homo sapi- with recurrent loops that can apparently ens (Clarke et al., 2003), suggesting deflesh- learn something resembling grammar. Given ing, but it is not clear whether its purpose a partial sequence of symbols, analogous to was decorative or cannibalistic. It may have a partial sentence, the network can learn more to do with hunger than with empathy. to predict events that would follow accord- There is clear evidence of ritual burial, often ing to rules of grammar. In a very limited involving the use of red ochre for decora- way, then, the network “learns” the rules of tion, dating from some 40,000 years ago in grammar. At first, the network was unable Europe and Australia (Wreschner, 1980). to handle the recursive aspects of grammar, in which phrases are embedded in other phrases, so that words that go together may Evolutionary Psychology and be separated by several other words. This the Pleistocene problem was at least partially surmounted when Elman introduced a “growth” factor, The four critical evolutionary markers men- which he simulated by degrading the sys- tioned above – migrations, enhanced man- tem early on so that only global aspects of ufacture, increased brain size, and prolon- the input were processed, and then gradu- gation of childhood – presumably took ally decreasing the “noise” in the system so place slightly before or during the early that it was able to process more and more part of the era known as the Pleistocene, detail. When this was done, the system was which is usually dated from about 1.8 mil- able to pick up some of the recursive quality lion years ago until about 10,000 years ago of grammar and so begin to approximate the (e.g., Janis, 1993). The new-found disci- processing of true language (Elman, 1993; pline of evolutionary psychology (see Chap- Elman, Bates, & Newport, 1996). ter 22) has typically emphasized the Pleis- Language is not critical to conscious- tocene era, and the so-called hunter-gatherer ness, or even to recursive consciousness, phase of hominid evolution, as the source but it surely enriches our conscious lives. of most distinctively human characteristics The unboundedness of language means that (e.g., Barkow, Cosmides, & Tooby, 1992). there is no limit to the number of sentences The basic research strategy, as also popular- we can create, and therefore no limit to the ized by Pinker (1994), is to “reverse engineer” number of images and scenarios we can cre- present-day cognitive and social characteris- ate in the minds of others. Consciousness tics to discover their roots in the Pleistocene. thus becomes a shared experience. This also This approach is allied with the view that the P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

the evolution of consciousness 583

mind consists of independent, encapsulated was no danger of obesity; we like landscapes “modules” (Fodor, 1983). with trees because trees provide shade and One example is the so-called cheater- escape from dangerous carnivores on the detection module, inferred from perfor- Africa savanna; flowers please us because mance on a test devised by Wason (1966), they are markers for edible fruits, nuts, or which runs as follows. You are shown four tubers amid the greenery of the savanna; and cards, bearing symbols such as A, C, 22, and so on – “there are modules,” he writes, “for 17. You are then asked which two cards you objects and forces, for animate beings, for should turn over to test the truth of the fol- minds, and for natural kinds like animals, lowing statement: “If a card has a vowel on plants, and minerals” (p. 315). one side, then it has an even number on the other side.” Most people will choose the A, which is rational, because if there is an odd Becoming Modern number on the other side, then the proposition is false. For their second choice, most Whether or not the mind is quite as modu- people turn over the 22, but this is in fact lar as the evolutionary psychologists suggest, not very revealing, because whatever is on there can be little doubt that it was during the other side cannot disconfirm the state- the Pleistocene era that our forebears put ment. The better strategy is to turn over the away apish things and became human. Even 17, because the presence of a vowel on the so, developments were in many respects other side would disconfirm the statement. slow. Our own wise species, Homo sapi- Suppose now that the task is translated so ens, arrived late in the Pleistocene. Estimates that A stands for ale and C for coke, the num- based on mitochondrial DNA place the most bers stand for people’s ages, and the state- recent common ancestor of modern humans ment to be verified is “If a person is drinking at some 170,000 years ago (Ingman, Kaess- ale, then he or she must be over 20 years old.” mann, Paäbo,¨ & Gyllensten, 2000), and there People still turn over the A to check the age is fossil evidence that modern human cranial of the drinker, but most now understand that structure had emerged by at least 154,000 they should turn over the 17, and if the kid is years ago (White et al., 2003). But so- drinking ale he should be thrown out of the called modern behavior may not have sur- pub. In short, there is a module for detect- faced until even more recently. As Bickerton ing people who don’t obey the rules, and (2002) roguishly puts it, “For the first 1.95 its logic is better understood in social con- million years after the emergence of erec- texts than in an abstract formulation (after tus almost nothing happened: The clunky Cosmides, 1989). stone tools became less clunky and slightly The danger with this approach is that more diversified stone tools, and everything it becomes too easy to postulate modules beyond that, from bone tools to supercom- and to tell “just so” stories as to how they puters, happened in the last one-fortieth of evolved in the Pleistocene, so that there is a the period in question” (p. 104). risk of returning to the instinct psychology The “last one fortieth” refers to what has of the early 20th century (e.g., McDougall, been called a “human revolution” (Mellars 1908). Instinct psychology perished under & Stringer, 1989), beginning a mere 40,000 the sheer weight of numbers – the author years ago. That period heralded a num- of one text counted 1,594 instincts that ber of developments suggestive of “mod- had been attributed to animals and humans ern” behavior. A sudden flowering of art (Bernard, 1924) – and evolutionary psychol- and technology in Europe around 30,000 to ogy may also drown in a sea of modules, if 40,000 years ago includes a dramatic expan- not of mixed metaphors. Pinker (1997) sug- sion of manufactured objects to include pro- gests that we like potato chips because fatty jectiles, harpoons, awls, buttons, needles, foods were nutritionally valuable during the and ornaments (Ambrose, 2001). Cave draw- Pleistocene, but scarce enough that there ings in France and Northern Italy, depicting P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

584 the cambridge handbook of consciousness

a menagerie of horses, rhinos, bears, lions, India, with one branch continuing on and and horses, date from the same period eventually reaching Australia as early as (Knecht, Pike-Tay, & White, 1993). The first 74,000 years ago. Another branch eventu- unequivocal musical instruments are bird- ally turned northward from somewhere near bone flutes from the early Upper Paleolithic western India to reach Europe some 46,000 in Germany (Hahn & Munzel,¨ 1995), and to 50,000 years ago. According to Oppen- there is widespread evidence across Russia, heimer, this view makes better sense of the France, and Germany for the weaving of evolutionary trees based on mtDNA and fibers into clothing, nets, bags, and ropes, Y-chromosome sequences than does the dating from some 29,000 years ago (Soffer, view that the migrants from Africa pro- Adovasio, Illingworth, Amirkhanov, Praslov, ceeded directly through the Levant to & Street, 2000). Europe (see also Macaulay et al., 2005). The suddenness of the human revolution The seeds of the human revolution may may actually be somewhat illusory, because therefore have been sown in Africa and it may not have been entirely indigenous developed by the early emigrants well before to Europe. The conventional view is that they arrived in Europe. Stone tool indus- it was probably instigated by the arrival of tries from the Klasies River Mouth near the a migratory wave of H. sapiens that began southern tip of Africa are very similar to from Africa around 50,000 years ago, reach- those of the Upper Paleolithic in Europe, ing Europe some 10,000 years later (Klein, but have been provisionally dated at around 2001). This is consistent with a study of 70,000 years ago (Mellars, 2002). Evidence present-day variation in mitochondrial DNA for a bone-tool industry, as well as deliber- (mtDNA), which confirms other evidence ate engraving of abstract forms in ochre and that H. sapiens dates from some 170,000 other artifacts indicative of “modern” behav- years ago in Africa and also indicates that ior, has been discovered at Blombos Cave in non-Africans share a most recent common South Africa and dated at around 70,000 ancestor with Africans dated at an esti- years ago (Henshilwood, d’Errico, Marean, mated 52,000 years ago (Ingman et al., Milo, & Yates, 2001). An even earlier bone 2000). This in turn is consistent with the industry, which probably included the man- idea that all present-day non-Africans are ufacture of harpoons to catch fish, has been descended from people who migrated from discovered in the republic of Congo and Africa relatively late in the history of our dates from about 90,000 years ago (Yellen, species. These migrants replaced, to put it Brooks, Cornelissen, Mehlman, & Stewart, politely, the Neanderthals in Europe, who 1995). became extinct some 28,000 years ago, and On the basis of such evidence, it has may also have overlapped with, and eventu- been argued that there was really no human ally replaced, Homo erectus in Java (Swisher revolution at all, but simply the accumula- et al., 1996). tion of shared knowledge and expertise in There is now reason to believe, how- Africa, perhaps originating as far back as ever, that the exodus of H. sapiens from 250,000 to 300,000 years ago (McBrearty. Africa that gave rise to all modern humans & Brooks, 2000). Stringer (2003) suggests, took place around 83,000 years ago, rather however, that “this may be too simple a than 52,000 years ago (Macaulay et al., story” (p. 28), and it is perhaps more likely 2005; Oppenheimer, 2003). This claim is that modern behavior evolved in Africa after based in part on an analysis by Oppen- the earlier migration of H. sapiens around heimer (2003) of the complete mtDNA 125,000 years ago. Stone tools discovered on sequences collected by Ingman et al. Accord- the Red Sea coast of Eritrea and dated at ing to this more recent scenario, as con- around 125,000 years ago are characteristic structed by Oppenheimer, the emigrants of the Middle Stone Age and are more primi- crossed the mouth of the Red Sea to the tive than the African industries dated within Yemen and proceeded round the coast to the past 100,000 years (Walter et al., 2000). P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

the evolution of consciousness 585

The Eritrean tools were probably associated as modern sports teams do. The invading with a migration of early H. sapiens to coastal sapients may have been simply wiser, as regions – a migration that may have led to their name suggests, although their ascen- subsequent coastal migrations into the Lev- dancy was probably not simply a matter of ant by 100,000 years ago (Grun¨ & Stringer, brain size, because the Neanderthal brain 1991). These early migrants probably did not appears to have been slightly larger than the survive the Ice Age that descended from the brain of H. sapiens. Tattersall (2003) refers north from about 90,000 years ago (Oppen- to “the emergence of modern cognition – heimer, 2003). The evidence from the ear- which, it is reasonable to assume, is equiv- liest known remains of H. sapiens in Africa, alent to the advent of symbolic thought” dating from around 154,000 to 160,000 years (p. 27). Yet Kanzi, the bonobo, learned to ago, also suggests a tool culture that was a use abstract symbols to refer to objects and mix of Acheulian and Middle Stone Age cul- actions (Savage-Rumbaugh et al., 1998), so tures (Clarke et al., 2003). it seems unlikely that symbol use per se was The likely scenario, then, is that mod- the decisive development. ern human behavior evolved in Africa some The cognitive ingredient that distingui- time within the period after the first expan- shed modern H. sapiens from the Neander- sion of H. sapiens out of Africa some thals, and even from earlier sapiens, is indeed 125,000 years ago, but before the later one of the puzzles of human evolution. A co- expansion of around 83,000 years ago, and mmon solution is simply to suppose that was carried and developed in the coastal some capacity for complex thought, be it migrations to Asia and Australia and in language or some deeper property of thou- the eventual migrations north to Europe ght, was potentially present in large-brained and China. There is some dispute as to hominids, but was not realized until the whether this development was seeded by human revolution. In speculating about the some biological change or was purely cul- emergence of syntax, for example, Bickerton tural. Klein (2001) has argued that there was (1995) suggests that the essential elements indeed some biological change, and White were already present, and all it took was a that al. (2003) suggest that the Ethiopian row of a switch to get them up and running: fossils dating from over 150,000 years ago represent an archaic form, which they Imagine a newly constructed factory lying label H. sapiens idaltu. Stringer (2003) dis- idle because someone neglected to make a agrees, suggesting that these fossils are “the crucial connection in the electric wiring. The making of that single connection is all that oldest definite record of what we currently 693 is needed to turn a dark and silent edifice think of as modern H. sapiens”(p. ). into a pulsating, brilliantly lit workplace. Whether biological or cultural, whatever This could have been how syntax evolved it was that H. sapiens cooked up in Africa in the brain (p. 83). between 125,000 and 83,000 years ago was a powerful brew. Not only did it lead to the Similarly, Tattersall (2003) suggests that the human revolution, but there was a darker modern human capacity “lay fallow until it side as well, resulting in the extinction of was activated by a cultural stimulus of some the other species of Homo – namely, the kind” and goes on to suggest that this capac- Neanderthals and Homo erectus – outside of ity was “the invention of language” (p. 27). Africa. It may have had to do simply with In evolutionary terms, it is difficult to superior weaponry or advances in technol- understand how there could be natural selec- ogy that allowed our forebears to adapt bet- tion for unrealized capacities, although it is ter to new environments – although it should of course possible that complex abilities in be remembered that the conquerors were one domain might be generalized to another. moving into a new environment, and one Some humans perform prodigiously on the might have expected the indigenous Nean- piano, for example, but this can scarcely have derthals to have held the home advantage, been a factor in the evolution of manual P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

586 the cambridge handbook of consciousness

dexterity. But language seems too embed- their eventual extinction. This work remains ded biologically to be the result of a late controversial (e.g., Gibson & Jessee, 1999), invention, or the late throwing of a switch, or but there is other evidence that the cranial a generalization from some other ability – structure underwent critical changes sub- or even, as Chomsky (1975) proposed, an sequent to the split between anatomically adventitious consequence of increased brain modern and earlier “archaic” Homo, such as size. I think it more likely that recursive the Neanderthals, Homo heidelbergensis, and thought evolved over the past 2 million years Homo rhodesiensis. One such change is the as a consequence of social pressures and was shortening of the sphenoid, the central bone selected in overt behavioral contexts involv- of the cranial base from which the face grows ing social dynamics, including language. In forward, resulting in a flattened face (D. E. what follows, I want to suggest that the new Lieberman, 1998). D. E. Lieberman specu- ingredient that led to the dominance of our lates that this is an adaptation for speech, species was not language per se, nor the abil- contributing to the unique proportions of ity to think recursively, but was the emer- the human vocal tract, in which the hori- gence of autonomous speech. zontal and vertical components are roughly equal in length – a configuration, he argues, that improves the ability to produce acous- Did We Talk Our Way tically distinct speech sounds. into Modernity? Articulate speech also required radical change in the neural control of vocalization. Attempts over the past half-century to com- The species-specific and largely involuntary municate with the great apes have taught calls of primates depend on an evolutionar- us at least one thing – these animals are ily ancient system that originates in the lim- a long way from being able to speak. For bic system, but in humans this is augmented example Viki, the chimpanzee raised from by a separate neocortical system operating infancy in a human household, was never through the pyramidal tract and synapsing able to utter more than about four indistinct directly with the brainstem nuclei for the words (Hayes, 1952). The limitation was vocal cords and tongue (Ploog, 2002). The at least partly articulatory, because much evidence suggests that voluntary control of greater success was later achieved by teach- vocalization in the chimpanzee is extremely ing apes to communicate using manual ges- limited at best (Goodall, 1986). The devel- tures (e.g., Gardner & Gardner, 1969)or opment of cortical control must surely have pointing to symbols displayed on a keyboard occurred gradually, rather than in an all- (e.g., Savage-Rumbaugh et al., 1998). Even or-none fashion, and may have reached its then, though, their communicative abilities final level of development only in anatomi- remained at the level of what Bickerton has cally modern humans. An adaptation unique called “protolanguage,” which is essentially to H. sapiens is neurocranial globularity, language without grammar. defined as the roundness of the cranial vault The evidence suggests that the articula- in the sagittal, coronal, and transverse planes, tory and neural adaptations necessary for flu- which is likely to have increased the relative ent speech emerged late in the evolution of size of the temporal and/or frontal lobes rela- Homo and may have been complete only tive to other parts of the brain (D. E. Lieber- with the emergence of our own species – man, McBratney, & Krovitz, 2002). These and possibly even later still. According to changes may reflect more refined control of P. Lieberman (1998; Lieberman, Crelin, & articulation and/or more accurate percep- Klatt, 1972), the lowering of the larynx, tual discrimination of articulated sounds. an adaptation that increased the range of One clue as to the recency of articulate speech sounds, was incomplete even in the speech comes from a speech and language Neanderthals of 30,000 years ago. Their disorder that afflicts a large family, known as resultant poor articulation would have kept the KE family, in England. Over three gen- them separate from H. sapiens, leading to erations, half of the members of this family P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

the evolution of consciousness 587

have been affected by the disorder, which with a model in which the expansion of is evident from the affected child’s first modern humans was driven by the appear- attempts to speak and persists into adult- ance of a more-proficient spoken language” hood (Vargha-Khadem, Watkins, Alcock, (p. 871). That is, the FOXP2 mutation may Fletcher, & Passingham, 1995). The precise have been the most recent step in the evo- nature of the deficit remains somewhat con- lution of articulate speech. troversial. Some have argued that the deficit If recursive thinking and syntax emerged is primarily linguistic, mainly (but not exclu- before articulate speech in the evolution sively) affecting the ability to use inflectional of Homo, as suggested earlier in this chap- morphosyntactic rules, such as changing the ter, then how did syntactic language itself endings of words to mark tense or num- evolve? There has been a long history ber (Gopnik, 1990; Gopnik & Goad, 1997; of speculation that language evolved from Pinker, 1994). Other, more recent work sug- manual gestures (e.g., Armstrong, 1999; gests that the core deficit is one of articula- Armstrong, & Wilcox, 1995; Corballis, 1992, tion rather than syntax, with morphosyntax 1999, 2002; Givon,´ 1995; Hewes, 1973; Riz- a secondary casualty (Alcock, Passingham, zolatti & Arbib, 1998), and syntax may there- Watkins, & Vargha-Khadem, 2000; Vargha- fore have evolved in the context of manual Khadem et al., 1998; Watkins, Dronkers, & gesture, not of speech. My own view is that Vargha-Khadem, 2002). the vocal component was introduced only It is now known that the disorder is gradually, perhaps culminating in the muta- due to a point mutation on the FOXP2 tion of the FOXP2 gene (Corballis, 2003)– gene (forkhead box P2) on chromosome 7, and even today most people gesture as they and for normal speech to be acquired, two speak (McNeill, 1985). Evidence for the functional copies of this gene seem to be gestural theory is necessarily indirect, but necessary (Fisher, Vargha-Khadem, Watkins, comes from a variety of sources. First, as Monaco, & Pembrey, 1998). FOXP2 has been we have seen, attempts to teach language sequenced in humans, chimpanzees, goril- to great apes have achieved much more suc- las, orangutans, rhesus monkeys, and mice cess by using gestural or visual means than by (Enard et al., 2002). The sequences reveal using vocalization. Second, the homologue changes in amino-acid encoding and the of Broca’s area in the monkey includes so- pattern of nucleotide polymorphism that called mirror neurons that respond both to emerged after the split between human the production and the perception of spe- and chimpanzee lineages and were therefore cific reaching and grasping movements, sug- probably selected for their beneficial effect gesting that the foundations for language on vocal communication. were manual rather than vocal (Rizzolatti In attempting to estimate the date of & Arbib, 1998). Broca’s area is now con- fixation of the mutated gene, Enard et al. sidered part of a more general “mirror sys- (2002) obtained the surprising answer that tem” involved in the understanding of bio- the most likely date was zero years ago, but logical action (Rizzolatti, Fogassi, & Gallese, suggested that this should be adjusted by 2001). Third, it is now clear that the signed some 10,000 to 100,000 years to correct languages invented by the deaf have all the for rapid population growth. Although the hallmarks of true language, including syn- estimated standard error was some 120,000 tax (e.g., Armstrong et al., 1995; Neidle, years, which could place the mutation as Kegl, MacLaughlin, Bahan, & Lee, 2000). much as 200,000 years into the past, the Fourth, the fact that the left hemisphere most likely estimate places it closer to the in most people plays the dominant role in present, and not implausibly in the win- both speech and manual praxis (as well as in dow between the earlier exodus from Africa signed languages) suggests a common neu- some 125,000 years ago and the more recent rological basis (Corballis, 1991, 2003). (and cataclysmic) exodus of 83,000 years The conversion from manual and facial ago. Enard et al. (2002) write that their dat- gestures to vocal speech is perhaps best ing of the FOXP2 mutation “is compatible understood if speech itself is considered a P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

588 the cambridge handbook of consciousness

gestural system, as a number of authors may have been assimilated to add diver- have proposed (Liberman & Whalen, 2000; sity to facial gestures, eventually creating Studdert-Kennedy & Goldstein, 2003). the distinction between voiced and unvoiced Browman and Goldstein (1991) developed a speech sounds. Voicing would also have ren- gestural theory of speech, based on “exactly dered accessible facial gestures that would the model used for controlling arm move- otherwise be inaccessible, such as those ments, with the articulators of the vocal tract involving the positioning of the tongue simply substituted for those of the arm” (p. inside the mouth. 314). The McGurk effect, with false dubbing Speech confers other advantages. It allows of speech syllables onto videos of mouths communication in the dark or when obsta- uttering different syllables, illustrates that cles intervene between sender and receiver. the perception of a speech sound depends There is some evidence that the telling of as much on what is seen as on what is stories, critical for the transmission of cul- heard (McGurk & MacDonald, 1976), which ture, may have been accomplished at night. explains the disturbing effect of movies in In describing life among the San, a mod- which the sound track is out of sync. Neu- ern hunter-gatherer society, Konner (1982) roimaging has also shown that both Broca’s writes, and Wernicke’s areas are active when people War is unknown. Conflicts within the “read” speech from facial gestures, consistent group are resolved by talking, sometimes with the view that “the core perceptual pro- half the night or all the night, for nights, cesses for speech are embodied” (Calvert & weeks on end....When we slept in a grass 2003 67 Campbell ,p. ). hut in one of the villages, there were many A recent fMRI study suggests a link nights when its flimsy walls leaked charged between the FOXP2 mutation and the mir- exchanges from the circle around the fire, ror system. Unaffected members of the frank expressions of feeling and contention KE family showed the expected activation beginning when the dusk fires were lit and of Broca’s area while they covertly gener- running on until the dawn (p. 7). ated verbs, but affected members of the Konner later notes that pedagogy is also a family showed underactivation of Broca’s night-time activity among the San: area (Liegeois,´ Baldeweg, Connelly, Gadian, Mishkin, & Vargha-Khadem, 2003). This Not only stories, but great stores of knowl- finding might be interpreted to mean that edge are exchanged around the fire among the mutation of the FOXP2 gene some the !Kung and the dramatizations – per- 100,000 years ago had to do with the incor- haps best of all – bear knowledge critical poration of vocal control into the mirror sys- to survival. A way of life that is difficult enough would, without such knowledge, tem (Corballis, 2004). It provides further become simply impossible (p. 171). evidence that the FOXP2 mutation was the final stage in a series of adaptations that Pinker and Bloom (1990) suggest that vocal allowed speech to become autonomous. It oratory (especially at night, one is tempted was the mutation, perhaps, that held the key to think) might have been subject to sex- to modernity. ual selection, citing Symons’s (1979) observation that tribal chiefs are often both gifted orators and highly polygynous. What Advantages Did Speech Confer Speech would have freed the hands for over Manual Gesture? other activities. In a species that increasingly made use of the hands and arms for One possible scenario is that the develop- manufacture, and probably also for carry- ment of manufacturing and tool use gave ing things (Lovejoy, 1981), there were surely selective advantage to shifting the burden advantages to shifting the communicative of communication from the hands to the load to the face and ultimately to the vocal face. Grunts associated with facial gestures channel. Carrying would presumably have P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

the evolution of consciousness 589

been important in an increasingly peripatetic Homo erectus in Asia and the Neanderthals existence, as exemplified in migrations from in Europe were probably capable of recur- Africa well before the emergence of H. sapi- sive thought, theory of mind, and grammat- ens, beginning a little under 2 million years ical language well before Homo sapiens burst ago (Tattersall, 1998), perhaps leading to the onto the scene. But the final accomplish- pressure to add voicing. Speech and the free- ment of a form of language that could be ing of the hands may have enhanced the purely vocal led to an explosion of art, man- development of technologies, enabling manufacture, and creative accomplishment that ual crafts to be at once demonstrated and would have added to the contents of con- verbally explained (Corballis, 2002). More- sciousness, if not to its fundamental nature. over, the products of manufacture are cumulative, so that a slight initial advantage may have multiplied to the point that our African Conclusions forebears were able to dominate and eventually replace all other hominids. Technol- Consciousness itself is almost certainly not ogy has continued to advance in exponential unique to humans. The ability to repre- fashion, allowing us unprecedented domin- sent aspects of the world internally and ion over and beyond the planet, but it may to carry out mental manipulations, such as have been seeded by a simple mutation that mental rotation and problem solving, prob- enabled us to talk. ably evolved independently in mammals, These arguments in favor of speech must cetaceans, and birds, driven in part by selec- of course be qualified by the growing evi- tion to survive in increasingly complex social dence that the signed languages of the deaf worlds (although ants may have found an have all of the semantic and syntactic sophis- alternative solution). In the case of primates, tication of spoken languages. Powerful testi- especially, mental life would have included mony to this is Gallaudet University, located representations of conspecifics, and their in Washington DC – the only university in dominance relations, moods, and the like. the world that caters exclusively for the deaf, Their mental life may include at least zero- with American Sign Language as the lan- order intentionality, such that a chimpanzee guage of instruction. The advantages asso- may believe that another chimpanzee is ciated with vocal language were therefore angry or that chimpanzee A is dominant probably to do with factors unrelated to lan- over chimpanzee B. There is little convinc- guage per se, but rather to the extended ing evidence, that non-human primates – or circumstances under which language could indeed any non-human species – are capable operate and the effects of freeing the hands of higher levels of intentionality. for the various other activities that have Nevertheless, it is still something of an changed the very nature of the physical open question whether the great apes can world. Indeed, there must surely have been achieve first-order intentionality. A gorilla strong evolutionary pressure to add vocaliza- well appreciates that another gorilla is angry, tion to the communicative repertoire, des- but this could be simply be based on the pite the fact that the lowering of the larynx behavioral demeanor of the other animal, increases the chances of death by choking. and not on the attribution of an emotional The switch to autonomous speech prob- state comparable to its own state when ably did not alter the fundamental nature angry. Similarly, a chimpanzee might avoid of consciousness itself. There is no reason certain actions while a more dominant chim- to believe that deaf signers, or the affected panzee is watching, but it remains unclear members of the KE family, are incapable whether this is simply a consequence of lear- of recursive thought, whether conscious or ning or whether the animal somehow under- not. Recursive thought is more likely to have stands what is going on in the mind of the evolved from 2 million years earlier, when other. It is clear, however, that humans can brain size underwent its dramatic increase. easily attribute mental states to others and P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

590 the cambridge handbook of consciousness

do so recursively, so that I can understand It is probably safe to assume, however, that you understand what I am getting at, ri- that our forebears of 80,000 to 100,000 years ght? Recursive thought also allows us to ima- ago were fully anatomically modern. Sub- gine ourselves in the past or future, thus gen- sequent developments have been profound, erating a sense of self as distinct from others. but can be attributable to cultural inven- Recursive thought may have evolved with tion and accurate transmission between gen- the genus Homo from around 2 million years erations rather than to changes in biol- ago. The earliest selective pressures leading ogy. Although chimpanzees display quite to recursive thought may have been primar- marked cultural variations (Whiten et al., ily social. Evolutionary psychologists have 1999), transmission of culture is undoubt- speculated at length about social charac- edly much more efficient in humans, creat- teristics, such as cooperation, mate choice, ing a ratchet effect culminating in the com- parental investment strategies, and language plex lives we lead today. As Tomasello (1999) itself, as having emerged in the Pleis- points out, “[The] key adaptation is one that tocene (Barkow et al., 1992; for an updated enables individuals to understand other indi- review, see Barrett, Dunbar, & Lycett, 2002). viduals as intentional agents like the self” Whereas evolutionary psychology is typi- (p. 509) – or in other words, first-order cally based on the premise that the mind intentionality. It is not simply a question is modular, recursion is a property that cuts of language; young children have a capac- across different mental attributes, including ity for joint attention that enables rapid cul- language, theory of mind, mental time travel, tural learning and effective imitation. We are and manufacture. wired to assimilate in a way that no other Human activity took a “Great Leap For- primate is. ward,” to borrow Diamond’s (1997) phrase, Nevertheless changes in the medium of with the human revolution of around language have undoubtedly had a profound 40,000–50,000 years ago – although the effect on cultural transmission. The emer- dramatic events in Europe may have been gence of speech would have permitted com- the culmination of changes that took place munication at night and enabled the oral prior to the exodus from Africa some 83,000 transmission of culture that persists in many years ago and in the long trek into South societies today. The invention of writing Asia and then into Europe. These devel- some 5,000 years ago would have greatly opments are generally considered to mark increased the precision and duration of the beginnings of so-called modern human transmitted culture (see Diamond, 1997) behavior. There is controversy as to whether and contributed to the profound variation they were due to some biological change among human cultures. The invention of the or whether they were purely cultural inno- Internet is likely to have equally profound vations. One possibility is that syntactic implications. Another major factor in human language itself emerged during this period cultural variation simply has to do with the (Bickerton, 1995), either as a cultural inven- pattern of migrations. Following the migration or as the result of some structural tions out of Africa some 50,000 years ago, change in the brain. In this chapter, I our forebears distributed themselves around have argued instead that syntactic language the globe in vastly different conditions of cli- emerged more gradually over the preceding mate and geography. Around 13,000 years 2 million years, in concert with the devel- ago, domestication of wild plants and ani- opment of recursive thought. The revolu- mals in different regions, notably the Fertile tionary events of the past 100,000 years or Crescent, China, and Mesoamerica, began so may have been instead a consequence traditions of control over the environment of the emergence of autonomous speech as and the eventual emergence of what we are the primary vehicle for linguistic expression. pleased to call civilizations. The anatomical change that finally permit- The growth of culture, especially in man- ted speech to dominate may have been a ufacturing societies, has vastly increased the mutation of the FOXP2 gene. number of material objects, ideas, and words P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

the evolution of consciousness 591

to describe them, which has surely had a Atance, C. M., & O’Neill, D. K. (2001). Episodic major impact on the contents of human con- future thinking. Trends in Cognitive Sciences, 5, sciousness, if not its essential nature. It is 533–539. unlikely that the mechanisms of conscious- Barkow, J., Cosmides, L., & Tooby, J. (Eds.) ness have changed much since the emer- (1992). The adapted mind: Evolutionary psychol- gence of Homo sapiens some 170,000 years ogy and the generation of culture. New York: ago or even since the large-brained hominids Oxford University Press. of a million years ago, but the things we are Barrett, L., Dunbar, R., & Lycett, J. (2002). conscious of have changed dramatically, at Human evolutionary psychology. New York: least in Western societies. Yet has civiliza- Palgrave. 1980 tion made us smarter? After working for 33 Beck, B. B. ( ). Animal tool behavior: The use years among technologically primitive New and manufacture of tools by animals. New York: Garland STPM Press. Guineans, Diamond (1997) argues that they 1976 are in fact smarter than Westerners. It may be Bennett, J. ( ). Linguistic behavior. Cam- bridge: Cambridge University Press. appropriate then to end this chapter with the 1924 question posed to Diamond by Yali, a local Bernard, L. L. ( ). Instinct: A study in social psychology. New York: Holt. New Guinean: “Why is that you white peo- 1995 ple developed much cargo and brought it to Bickerton, D. ( ). Language and human behavior. Seattle, WA: University of Washington New Guinea, but we black people had little Press/UCL Press. cargo of our own?” The answer, as Diamond Bickerton, D. (2000). Reciprocal altruism as demonstrates so convincingly, has to do with the predecessor of argument structure. In geography, not biology, and Yali reminds us W. H. Calvin & D. Bickerton (Eds.), Lingua ex that we are all people beneath the veneer of machina: Reconciling Darwin and Chomsky with culture. the human brain (pp. 123–134). Cambridge, MA: MIT Press. Bickerton, D. (2002). From protolanguage to lan- Note guage. In T. J. Crow (Ed.), The speciation of modern Homo Sapiens (pp. 103–120). Oxford: Oxford University Press. 1. I am indebted to Brian Boyd for these exam- 1991 ples. They are taken from a talk entitled “Evo- Browman, C., & Goldstein, L. ( ). Gestural lution, Cognition, Narration, Fiction,” which structures: Distinctiveness, phonological pro- he presented at the Interdisciplinary Sympo- cesses, and historical change. In I. G. Mattingly sium on the Nature of Cognition, Tamaki & M. Studdert-Kennedy (Eds.), Modularity and 313 Campus, University of Auckland, on October the motor theory of speech perception (pp. – 338 13, 2001. ). Hillsdale, NJ: Erlbaum. Brown, F., Harris, J., Leakey, R., & Walker, A. (1985). Early Homo erectus skeleton from 788 References west Lake Turkana, Kenya. Nature, 316, – 792. Calvert, G. A. , & Campbell, R. (2003). Reading Alcock, K. J., Passingham, R. E., Watkins, K. E., speech from still and moving faces: The neural & Vargha-Khadem, F. (2000). Oral dyspraxia substrates of visible speech. Journal of Cognitive in inherited speech and language impairment Neuroscience, 15, 57–70. and acquired dysphasia. Brain & Language, 75, Cargile, J. (1970). A note on “iterated knowings.” 17 33 – . Analysis, 30, 151–155. Ambrose, S. H. (2001). Paleolithic technology and Chomsky, N. (1966). Cartesian linguistics: A chap- human evolution. Science, 291, 1748–1753. ter in the history of rationalist thought. New York: Armstrong, D. F. (1999). Original signs: Gesture, Harper & Row. sign, and the source of language. Washington, Chomsky, N. (1975). Reﬂections on language. New DC: Gallaudet University Press. York: Pantheon. Armstrong, D. F., Stokoe, W. C., & Wilcox, S. Chomsky, N. (1988). Language and problems of E. (1995). Gesture and the nature of language. knowledge: The Managua lectures. Cambridge, Cambridge: Cambridge University Press. MA: MIT Press. P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

592 the cambridge handbook of consciousness

Clarke, J. D., Beyenne, Y., WoldeGabriel, G., Elman, J. (1993). Learning and development in Hart, W. K., Renne, P. R., Gilbert, H., et al. neural networks: The importance of starting (2003). Stratigraphic, chronological and behav- small. Cognition, 48, 71–99. ioral contexts of Pleistocene Homo sapiens from Elman, J., Bates, E., & Newport, E. (1996). Middle Awash, Ethiopia. Nature, 423, 747– Rethinking innateness: A connectionist perspec- 752. tive on development. Cambridge, MA: MIT Clayton, N. S., Bussey, T. J., & Dickinson, A. Press. (2003). Can animals recall the past and plan Enard, W., Przeworski, M., Fisher, S. E., Lai, C. S., for the future? Nature Reviews, 4, 685–691. Wiebe, V., Kitano, T., et al. (2002). Molecular Corballis, M. C. (1991). The lopsided ape: Evolution evolution of FOXP2, a gene involved in speech of the generative mind. Oxford: Oxford Univer- and language. Nature, 418, 869–872. sity Press. Ferster, C. B., & Skinner B. F. (1957). Schedules Corballis, M. C. (1992). On the evolution of lan- of reinforcement. New York, Appleton-Century- guage and generativity. Cognition, 44, 197–226. Crofts. Corballis, M. C. (1999, March-April). The gestu- Fisher, S. E., Vargha-Khadem, F., Watkins, K. E., ral origins of language. American Scientist, 87, Monaco, A. P., & Pembrey, M. E. (1998). Local- 138–45. ization of a gene implicated in a severe speech Corballis, M. C. (2002). From hand to mouth: and language disorder. Nature Genetics, 18, 168– The gestural origins of language. Princeton, NJ: 170. Princeton University Press. Fodor, J. (1983). The modularity of mind. Cam- Corballis, M. C. (2003). From hand to mouth: bridge, MA: MIT Press. The gestural origins of language. In M. H. Gallup, G. G., Jr. (1997). The rise and fall of self- Christiansen & S. Kirby (Eds.), Language evolu- conception in primates. Annals of the New York tion (pp. 201–218). Oxford: Oxford University Academy of Sciences, 118, 73–84. Press. Gallup, G. G., Jr. (1998). Self-awareness and Corballis, M. C. (2004). FOXP2 and the mirror the evolution of social intelligence. Behavioural system. Trends in Cognitive Sciences, 8, 95–96. Processes, 42, 239–247. Cosmides, L. (1989). The logic of social exchange: Gardner, R. A., & Gardner, B. T. (1969). Teach- Has natural selection shaped how humans ing sign language to a chimpanzee. Science, 165, reason? Studies with the Wason selection task. 664–672. Cognition, 31, 187–276. Georgopoulos, A. P., & Pellizzer, G. (1995). The Cowie, A., & Stoerig, P. (1995). Blindsight in mental and the neural – psychological and neu- monkeys. Nature, 373, 247–249. ral studies of mental rotation and memory scan- Davison, M. C., & McCarthy, D. C. (1987). The ning. Neuropsychologia, 33, 1531–1547. matching law: A research review. Hillsdale, NJ: Gibson, K. R., & Jessee, S. (1999). Language evo- Erlbaum. lution and expansions of multiple neurological Deacon, T. (1997). The symbolic species: The co- processing areas. In B. J. King (Ed.), The origins evolution of language and the brain. New York: of language: What nonhuman primates can tell Norton. us. Santa Fe, NM: School of American Research Dennett, D. C. (1983). Intentional systems in cog- Press. nitive ethology: The “Panglossian paradigm” Givon,´ T. (1995). Functionalism and grammar. defended. Behavioral & Brain Science, 6, 343– Philadelphia: Benjamins. 390. Goodall, J. (1986). The chimpanzees of Gombe: Descartes, R. (1985). The philosophical writings Patterns of behavior. Cambridge, MA: Harvard of Descartes (J. Cottingham, R. Stoothoff, & University Press. D. Murdock, Eds. & Trans.). Cambridge: Cam- Gopnik, M. (1990). Feature-blind grammar and bridge University Press. (Work originally pub- dysphasia. Nature, 344, 715. lished 1647) Gopkin, M., & Goad, H. (1997). What underlies de Waal, F. B. M. (1996). Good natured: The origins inﬂectional error patterns in genetic dysphasia? of right and wrong in humans and other animals. Journal of Neurolinguistics, 10, 109–137. Cambridge, MA: Harvard University Press. Gowlett, J. A. J. (1992). Early human mental abili- Diamond, J. (1997). Guns, germs, and steel. New ties. In S. Jones, R. Martin, & D. Pilbeam (Eds.), York: Norton. The Cambridge encyclopedia of human evolution P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

the evolution of consciousness 593

(pp. 341–345). Cambridge: Cambridge Univer- tion and the origin of modern humans. Nature, sity Press. 408, 708–713. Grun,¨ R., & Stringer, C. B. (1991). Electron spin Janis, C. (1993). Victors by default: The mam- resonance dating and the evolution of modern malian succession. In S. J. Gould (Ed.), humans. Archaeometry, 33, 153–199. The book of life (pp. 169–217). New York: Hahn, J., & Munzel,¨ S. (1995). Knochenfloten¨ Norton. aus dem Aurignacien des Geissenklosterle¨ bei Klein, R. G. (2001). Southern Africa and mod- Blaubeuren, Alb-Donau-Kreis. Fundberichte ern human origins. Journal of Anthropological Aus Baden-Wurttemberg, 20, 1–12. Research, 57, 1–16. Hare, B., Call, J., Agnetta, B., & Tomasello, M. Knecht, H., Pike-Tay, A., & White, R. (1993). (2000). Chimpanzees know what conspecifics Before Lascaux: The complex record of the early do and do not see. Animal Behaviour, 59, 771– Upper Paleolithic. Boca Raton: CRC Press. 785. Kohler,¨ W. (1925). The mentality of apes. New Hare, B., & Tomasello, M. (1999). Domestic dogs York: Routledge & Kegan Paul. (Canis familiaris) use human and conspecific Konner, M. (1982). The tangled wing: Biologi- cues to locate hidden food. Journal of Compar- cal constraints on the human spirit. New York: ative Psychology, 113, 173–177. Harper. Hauser, M., & Carey, S. (1998). Building a cog- Krogman, W. M. (1972). Child growth. Ann nitive creature from a set of primitives. In D. Arbor: University of Michigan Press. D. Cummins & C. Allen (Eds.), The evolution Leslie, A. M. (1994). ToMM, ToBY, and agency: 51 106 of mind (pp. – ). New York: Oxford Uni- Core architecture and domain specificity. In L. versity Press. A. Hirschfeld & S. A. Gelman (Eds.), Mapping Hauser, M. D., Chomsky, N., & Fitch, W. T. the mind: Domain specificity in cognition and cul- (2002). The faculty of language: What is it, ture (pp. 119–148). New York: Cambridge Uni- who has it, and how did it evolve? Science, 298, versity Press. 1569 1579 – . Liberman, A. M., & Whalen, D. H. (2000). On Hayes, C. (1952). The ape in our house. London: the relation of speech to language. Trends in Gollancz. Cognitive Sciences, 4, 187–196. Henshilwood, C. S., d’Errico, F., Marean, C. W., Lieberman, D. E. (1998). Sphenoid shortening Milo, R. G., & Yates, R. (2001). An early bone and the evolution of modern cranial shape. tool industry from the Middle Stone Age at Nature, 393, 158–162. Blombos Cave, South Africa: Implications for Lieberman, D. E., McBratney, B. M., & Krovitz, the origins of modern human behaviour, sym- G. (2002). The evolution and development of bolism and language. Journal of Human Evolu- cranial form in Homo sapiens. Proceedings of the 631 678 tion, 41, – . National Academy of Sciences USA, 99, 1134– Herman, L. M., Richards, D. G., & Wolz, 1139. 1984 J. P. ( ). Comprehension of sentences Lieberman, P. (1998). Eve spoke: Human language 129 by bottle-nosed dolphins. Cognition, 16, – and human evolution. New York: Norton. 219. Lieberman, P., Crelin, E. S., & Klatt, D. H. (1972). Hewes, G. W. (1973). Primate communication Phonetic ability and related anatomy of the and the gestural origins of language. Current new-born, adult human, Neanderthal man, and 5 24 Anthropology, 14, – . the chimpanzee. American Anthropologist, 74, Heyes, C. M. (1998). Theory of mind in nonhu- 287–307. man primates. Behavioral & Brain Sciences, 21, Liegeois,´ F., Baldeweg, T., Connelly, A., Gadian, 101–148. D. G., Mishkin, M., & Vargha-Khadem, F. Heyes, C. M. (2001). Theory of mind and other (2003). Language fMRI abnormalities associ- domain-specific hypotheses – Response. Behav- ated with FOXP2 gene mutation. Nature Neu- ioral & Brain Sciences, 24, 1143–1145. roscience, 6, 1230–1237. Hunt, G. R. (1996). Manufacture and use of hook- Logothetis, N. K. (1998). Single units and con- tools by New Caledonian crows. Nature, 379, scious vision. Proceedings of the Royal Society of 249–251. London, Series B, 353, 1801–1818. Ingman, M., Kaessmann, H., Paäbo,¨ S., & Gyllen- Lovejoy, O. C. (1981). The origin of man. Science, sten, U. (2000). Mitochondrial genome varia- 221, 341–350 P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

594 the cambridge handbook of consciousness

Macaulay, V., Hill, C., Achilli, A., Rengo, C., Pinker, S., & Bloom, P. (1990). Natural language Clarke, D., Meehan, W., et al. (2005). Single, and natural selection. Behavioral & Brain Sci- rapid coastal settlement of Asia revealed by ences, 13, 707–784. analysis of complete mitochondrial genomes. Ploog, D. (2002). Is the neural basis of vocal- Science, 308, 1034–1036. ization different in non-human primates and McBrearty, S., & Brooks, A. S. (2000). The revo- Homo sapiens? In T. J. Crow (Ed.), The speci- lution that wasn’t: A new interpretation of the ation of Modern Homo Sapiens (pp. 121–135). origin of modern human behavior. Journal of Oxford: Oxford University Press. Human Evolution, 39, 453–563. Povinelli, D. J. (2001). Folk physics for apes. New McDougall, W. (1908). An introduction to social York: Oxford University Press. psychology. London: Methuen. Povinelli, D. J., & Bering, J. M. (2002). The men- McGurk, H., & MacDonald, J. (1976). Hearing tality of apes revisited. Current Directions in lips and seeing voices. Nature, 264, 746–748. Psychological Science, 11, 115–119. McNeill, D. (1985). So you think gestures are Povinelli, D. J., Bering, J. M., & Giambrone, nonverbal?’ Psychological Review, 92, 350– S. (2000). Toward a science of other minds: 371. Escaping the argument by analogy. Cognitive Mellars, P. A. (2002). Archaeology and the ori- Science, 24, 509–541. gins of modern humans: European and African Premack, D., & Woodruff, G. (1978). Does the perspectives. In T. J. Crow (Ed.), The speciation chimpanzee have a theory of mind? Behavioral of modern Homo Sapiens (pp. 31–47). Oxford: & Brain Sciences, 4, 515–526. Oxford University Press. Rizzolatti, G., & Arbib, M. A. (1998). Language Mellars, P. A., & Stringer, C. B. (Eds.) (1989). within our grasp. Trends in Neuroscience, 21, The human revolution: Behavioral and biologi- 188–194. cal perspectives on the origins of modern humans. Rizzolatti, G., Fogassi, L., & Gallese, V. (2001). Edinburgh: Edinburgh University Press. Neurophysiological mechanisms underlying Menzel, C. R. (1999). Unprompted recall and the understanding and imitation of action. reporting of hidden objects by a chimpanzee Nature Reviews, 2 , 661–670. (Pan troglodytes) after extended delays. Jour- Russell, B. (1905). On denoting. Mind, 479–493. 1 9 nal of Comparative Psychology, 113, – . Savage-Rumbaugh, S., Shanker, S. G., & Taylor, Miles, H. L. (1990). The cognitive foundations for T. J. (1998). Apes, language, and the human reference in a signing orangutan. In S. T. Parker mind. New York: Oxford University Press. & K. Gibson (Eds.), ‘Language’ and intelligence Schwarz, B. L., Colon, M. R., Sanchez, I. C., in monkeys and apes: Comparative developmen- Rodriguez, I. A., & Evans, S. (2002). Single- 511 539 tal perspectives (pp. – ). New York: Cam- trial learning of ‘what’ and ‘who’ information bridge University Press. in a gorilla (gorilla gorilla gorilla): Implications Neidle, C., Kegl, J., MacLaughlin, D., Bahan, B. & for episodic memory. Animal Cognition, 5, 85– Lee, R. G. (2000). The syntax of American Sign 90. Language. Cambridge, MA: MIT Press. Schwarz, B. L., & Evans, S. (2001). Episodic mem- Noble, W., & Davidson, I. (1996). Human evo- ory in primates. American Journal of Primatol- lution, language, and mind. Cambridge: Cam- ogy, 55, 71–85. bridge University Press. Skinner, B. F. (1957). Verbal behavior. New York: Oppenheimer, S. (2003). Out of Eden: The peo- Appleton-Century-Crofts. pling of the world. London: Robinson. Smirnov, Y. (1989). Intentional human burial. Patterson, F. (1978). Conversations with a gorilla. Journal of World Prehistory, 3, 199–233. National Geographic, 154, 438–465. Soffer, O. J. M., Adovasio, J. S., Illingworth, H. A., Pickering, T., White, T. D., & Toth, N. (2000). Amirkhanov, N. D., Praslov, N. D., & Street, Cutmarks on a Plio-Pleistocene hominid from M. (2000). Palaeolithic perishables made per- Sterkfontein, South Africa. American Journal of manent. Antiquity, 74, 812–821. Physical Anthropology, 111, 579–584. Solecki, R. S. (1975). Shanidar IV, a Neanderthal Pinker, S. (1994). The language instinct. New York: ﬂower burial in Northern Iraq. Science, 190, William Morrow. 880–881. Pinker, S. (1997). How the mind works. New York: Stringer, C. (2003). Out of Ethiopia. Nature, 423, Norton. 692–695. P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

the evolution of consciousness 595

Studdert-Kennedy, M., & Goldstein, L. (2003). language disorder. Proceedings of the National Launching language: The gestural origin of Academy of Sciences USA, 92, 930–933. discrete infinity. In M. H. Christiansen & S. Vargha-Khadem, F., Watkins, K. E., Price, C. Kirby (Eds.), Language evolution (pp. 235–254). J., Ashburner, J., Alcock, K. J., Connelly, A., Oxford: Oxford University Press. et al. (1998). Neural basis of an inherited Suddendorf, T., & Corballis, M. C. (1997). Men- speech and language disorder. Proceedings of the tal time travel and the evolution of the human National Academy of Sciences USA, 95, 12695– mind. Genetic, Social, and General Psychology 12700. Monographs, 123, 133–167. Walter, R. C., Buffler, R. T., Bruggemann, J. H., Swisher, C. C., III, Rink, W. J., Anton, H. P., Guillaume, M. M. M., Berhe, S. M., Negassi, B., Schwarcz, H. P., Curtis, G. H., Suprijo, A., et al. (2000). Early human occupation of the & Widiasmoro (1996). Latest Homo erectus of Red Sea coast of Eritrea during the last inter- Java: Potential contemporaneity with Homo glacial. Nature, 405, 65–69. sapiens in Southeast Asia. Science, 274, 1870– Wason, P. (1966). Reasoning. In B. M. Foss 1874. (Ed.), New horizons in psychology (pp. 135–151). Symons, D. (1979). The evolution of human sexu- London: Penguin. ality. Oxford: Oxford University Press. Watkins, K. E., Dronkers, N. F., & Vargha- Tattersall, I. (1998). Becoming human: Evolution Khadem, F. (2002). Behavioural analysis of an and human uniqueness. New York: Harcourt inherited speech and language disorder: Com- Brace. parison with acquired aphasia. Brain, 125, 452– Tattersall, I. (2003). Once we were not alone. Sci- 464. entific American, 13(2), 20–27. Watson, J. B. (1913). Psychology as a behaviorist Tomasello, M. (1996). Do apes ape? In J. Galef views it. Psychological Review, 20, 1–14. & C. Heyes (Eds.), Social learning in animals: Weiskrantz, L. (1986). Blindsight. A case study The roots of culture (pp. 319–346). New York: and implications. Oxford: Oxford University Academic Press. Press. Tomasello, M. (1999). The human adaptation Wheeler, M. A., Stuss, D. T., & Tulving, E. (1997) for culture. Annual Review of Anthropology, 28, Toward a theory of episodic memory: Auto- 509–529. noetic consciousness and the frontal lobes. Psy- Tomasello, M. (2000). Primate cognition: Intro- chological Bulletin, 121, 331–354. duction to the issue. Cognitive Science, 24, 351– White, T. D., Asfaw, B., DeGusta, D., Gilbert, 361. H., Richards, G. D., Suwa, G., et al. (2003). Tomasello, M. Hare, B., & Agnetta, B. (1999). Pleistocene Homo sapiens from Middle Awash, Chimpanzees, Pan troglodytes, follow gaze Ethiopia. Nature, 423, 742–747. direction geometrically. Animal Behaviour, 58, Whiten, A., & Byrne, R. W. (1988). Tactical 769–777. deception in primates. Behavioral & Brain Sci- Tomasello, M., & Rakoczy, H. (2003). What ences, 11, 233–244. makes human cognition unique? From individ- Whiten, A., & Byrne, R. W. (1997). Machiavel- ual to shared to collective intentionality. Mind lian intelligence II: Extensions and evaluations. and Language, 18, 121–147. Cambridge: Cambridge University Press. Trivers, R. L. (1971). The evolution of reciprocal Whiten, A., Goodall, J., McGrew, W.C., Nishida, altruism. Quarterly Review of Biology, 46, 35– T., Reynolds, V., Sugiyama, Y., et al. (1999). 57. Culture in chimpanzees. Nature, 399, 682– Tulving, E. (2001). Origin of autonoesis in 685. episodic memory. In H. L. Roediger, III, J. S. Wreschner, E. E. (1980). Red ochre and human Nairne, I. Neath, & A. M. Surprenant (Eds.), evolution. Current Anthropology, 21, 631–644. The nature of remembering: Essays in honor of Wundt, W. (1894). Lectures on human and animal Robert G. Crowder (pp. 17–34). Washington, psychology. London: Swan & Sonnenschein. DC: American Psychological Association. Yellen, J. E., Brooks, A. S., Cornelissen, E., Mehl- Vargha-Khadem, F., Watkins, K. E., Alcock, K. J., man, M. J., & Stewart, K. (1995). A Middle Fletcher, P., & Passingham, R. (1995). Praxic Stone Age worked bone industry from and nonverbal cognitive deficits in a large fam- Katanda, Upper Semliki Valley, Zaire. Science, ily with a genetically transmitted speech and 268, 553–556. P1: KAE 0521857430c21 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:48

596 P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

CHAPTER 22 The Serpent’s Gift: Evolutionary Psychology and Consciousness

Jesse M. Bering and David F. Bjorklund

Abstract tive in human ancestral history suddenly became maladaptive when consciousness As a higher-order cognitive system enabling appeared. access to intentional states, and one that few (if any) other species even marginally possess, consciousness or, more appropriately, The Serpent’s Gift: Evolutionary self-consciousness has likely been both selec- Psychology and Consciousness tively advantageous and the source of adaptive conflict in human evolutionary history. Consciousness or, more properly, self- Consciousness was likely advantageous to consciousness (or self-awareness) has long early human beings because it built on more been one of the features, along with culture, ancient primate social adaptations. Individ- tool use, and language, that have been used uals likely profited by having the capacity to set human beings apart from other ani- to track the intentions of the self and of mals (at least in the minds of many Homo social others in that consciousness permit- sapiens). Yet, the erection of a species bar- ted behavioral strategies involving decep- rier only serves as a target for others to top- tion and declarative communication. How- ple (see, for example, evidence of tool use ever, consciousness was likely also a source and cultural transmission in chimpanzees; of adaptive conflict in that it interfered with Whiten et al., 1999), and consciousness is no the functioning of more ancient social adap- exception. Although we have our own views tations, such as infanticide and male sexual about the possible uniqueness of human coercion of females. Having access to the consciousness (see below), our primary task epistemic states of others meant that knowl- is this chapter is not a discussion of the evo- edge of social transgressions could be rapidly lution of consciousness per se as much as it is conveyed between parties. For many evolved about the role of consciousness in determining psychological mechanisms, what was adap- particular human adaptations. Of course, we

597 P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

598 the cambridge handbook of consciousness

are concerned with the evolutionary origins approach for investigations of the anatom- of these adaptations, and we do not neglect ical and physiological aspects of conscious- this topic. Nonetheless, we see human beings ness, particularly if the goal of explanation as possessing a degree of self-awareness that is at, say, the level of sensory experience has no parallel in the animal world and that, and motor planning (e.g., Cotterill, 2001; once evolved, drastically modified the nature Humphrey, 2000; Jeannerod, 1999; Searle, of the beast, comparable to the effect that 2000). But our goal in the current chapter flight must have had on the biological line is somewhat different – although we ack- that led to birds. Consciousness, however, nowledge the phylogenetic continuity of the was a mixed blessing, for it provided not biological substrates of consciousness, we only new opportunities for the species that believe that there is now sufficient evidence possessed it but also new challenges, and it to show that human beings are operating required the seemingly rapid acquisition of with a mental representational system that a suite of cognitive adaptations to deal with can find no analogy in the central nervous this new level of self-awareness. In this chap- systems of other species. Thus, our approach ter, we examine the role of consciousness here is to highlight the likely consequences, in human functioning from the perspective both the good and the bad, of this evolu- of evolutionary psychology. As we demon- tionary innovation on the lives of hominid strate, evolutionary psychology has had rel- ancestors. We further propose that modern atively little to say about this role to date, humans have inherited behavioral propensi- but has the tools to contribute significantly ties to act in ways that enabled these ances- to our understanding of this phenomenon. tors to capitalize on the consequences of this As a first step, we feel it necessary to system’s presence, and thus also the psycho- define consciousness in terms that are amen- logical mechanisms that made these behav- able to empirical science. There will there- iors likely to occur in an adaptive context. fore be no zombies joining us at the table nor In the sections that follow, we first intro- qualia to occupy our thoughts (cf. Chalmers, duce the reader to the basic concepts of 1996) Rather, we define consciousness as evolutionary psychology. We then provide a that naturally occurring cognitive representa- brief description of human brain evolution, tional capacity permitting explicit and reflective along with speculations as to how human accounts of the – mostly causative – contents of consciousness emerged. We then examine a mind, contents harbored by the psychological related topic; namely, evidence for higher- frame of the self and, as a consequence, also order cognition in our closest primate rel- the psychological frames of others. Our view of atives, chimpanzees, which serve as imper- consciousness is therefore not one of a solely fect models for what the common ancestor autonoetic (cf. Tulving, 1985) nature, nor of apes and human beings may have been does it remove the self from consciousness, like. Finally, we discuss the impact that con- but rather seeks to integrate the concept into sciousness made on human evolution – an the empirical tradition of cognitive science impact that was felt in three ways. First, by holding it as a system enabling higher- consciousness, as a domain-general mecha- order representations of abstract causes of nism, provided direct benefits to the species behaviors. because it expanded on more ancient pri- This definition of consciousness will mate adaptations (e.g., deception, coopera- almost certainly strike some readers as too tion, reciprocal altruism) that had evolved narrow. Commonly, the topic of conscious- to cope with living in large social groups but ness is handled by scholars in the fields were not necessarily dependent on an aware- of cognitive neuroscience, philosophy, and ness of other minds. Second, the emergence comparative psychology, and within these of consciousness posed a new series of adap- areas consciousness is frequently viewed in tive challenges because it disrupted fitness- shades of increasing complexity both within maximizing categories of primate behavior, and across species. This is certainly the right particularly social behaviors involving the P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

the serpent’s gift: evolutionary psychology and consciousness 599

adaptive exploitation of other members of behaviors as mate-guarding and reciprocal the species (e.g., sexual coercion and infanti- altruism have greater inclusive fitness over cide). These conflicting challenges between those who lacked them? (Inclusive fitness more ancient adaptations and the new prob- not only refers to producing offspring, as in lems encountered by consciousness created the case of the more traditional Darwinian new behavioral algorithms that served to concept of reproductive fitness, but also con- reduce, but not eliminate, the incidence of siders the influence that an individual may socially proscribed behaviors in the species. have in getting other copies of his or her And, third, we propose that human con- genes into subsequent generations, through sciousness has been responsible for the evo- grandchildren or nieces and nephews, for lution of a suite of novel psychological adap- example [Hamilton, 1964].) tations that are unshared, even in precursory A central assumption of evolutionary form, with other species (e.g., the psycholog- psychology is that the psychological struc- ical mechanisms responsible for suicide). tures that evolved are adaptive, information- processing mechanisms designed to deal with recurrent problems faced by our ancestors. According to evolutionary psycholo- Evolutionary Psychology gists, “the causal link between evolution and behavior is made through psychologi- Underlying Assumptions of cal mechanisms” (Cosmides & Tooby, 1987, Evolutionary Psychology p. 277). Individuals who did not possess For the past two decades, investigators have adaptive psychological traits were unable been carefully reconstructing the evolution- to reliably engage in behaviors that were ary history of specific human psycholog- adaptive in ancestral environments. As a ical systems (see e.g., Barkow, Tooby, & result, they failed to disperse their genes Cosmides, 1992; Buss, 1995, 2005; Daily & as much as those who did engage in these Wilson, 1988). Based on the central tenets adaptive behaviors, and eventually only of Darwinian natural selection, evolution- those individuals whose behaviors were sup- ary psychology is a subfield of psychology ported by these psychological systems were that seeks to understand the adaptive func- represented in the population. What get tion of the diverse universal cognitive abili- selected, according to this rationale, are not ties and human behaviors that were selected the adaptive behaviors per se, but rather in the environment of evolutionary adapted- those psychological systems undergirding ness, usually defined as the Pleistocene, the and enabling these adaptive behaviors (e.g., last 2 million years or so when humans Buss, 1995; Tooby & Cosmides, 1992). emerged as a species. Evolutionary psychol- Moreover, these mechanisms are domain- ogy is not concerned with how human beings specific in nature. Human beings (and pre- are similar to or different from other species, sumably other animals) did not evolve but rather with how the human mind was general learning or information-processing shaped over the course of its recent evolu- abilities that could be applied to the wide tion. Similar to the way evolutionary biol- range of problems they encounter as they go ogists attempt to explain the emergence about their lives. Instead, what evolved were and contemporary appearance of morpho- a host of relatively specific mechanisms, each logical structures, such as the human hand sculpted by natural selection to deal with rel- or digestive tract, evolutionary psycholo- atively specific and recurring problems, such gists are concerned with the emergence and as language, detecting cheaters, or gaining contemporary appearance of psychological and maintaining mates. As an analogy, the structures, such as those involved in mate- mind is compared to a Swiss Army Knife, guarding or reciprocal altruism. That is, with different tools designed for different to what extent did those individuals who tasks (Tooby & Cosmides, 1992), rather than possessed psychological traits driving such a broad ax, which may be powerful but P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

600 the cambridge handbook of consciousness

too wieldy to be useful for many complex anisms living in complex socioecologies (e.g., problems. Despite claims to the contrary Tooby & Cosmides, 1992). (e.g., Lickliter & Honeycutt, 2003), evolutionary psychology does not advocate a form Self-Consciousness as an Epiphenomenon of genetic determinism, but emphasizes that evolved, adaptive mechanisms are sensi- Still, the role of self-consciousness, which tive to environmental context (see Tooby, appears to be a very general mechanism Cosmides, & Barrett, 2003). This is espe- permitting reflective awareness of the self’s cially true for human beings, who live in proximate motivational states, continues to diverse physical and social environments and remain very unclear in evolutionary mod- require a flexible intelligence to survive. els of human cognition. Many evolutionary Nevertheless, the plasticity of human psychologists consider consciousness to be thought and behavior is not infinite. Human an epiphenomenon that shadows the intu- infants are prepared by evolution for a struc- itive operations of psychological adaptations tured world that includes sights and sounds, and that has played no important role in a lactating mother, social support, and lan- the evolutionary emergence of these adap- guage, among many other things. There are tations. To support this position, such the- constraints on what they can process and orists cite people’s na¨ıve, explicit explana- how they will interpret experience. These tions for the causes of their own adaptive enabling constraints (Gelman & Williams, behaviors, explanations that are far removed 1998) should not be viewed negatively, for from plausible selection-based explanations they make it easier for children to master for their actions (French, Kamil, & Ledger, the ways of a human world, facilitating the 2001). In addition, similar behaviors that acquisition of language, for example, but occur under similar ecological conditions in making it impossible to learn to navigate different societies are often interpreted in via echolocation. Over the course of devel- very different ways (e.g., for an application opment, children’s information-processing of this principle to the subject of infanti- biases are modified by experience, but cide, see Daly & Wilson, 1988). This suggests inevitably result in behaviors that are gen- that, although there was selection for cog- erally well suited to the social environments nitive programs that prompt specific types in which they live. of responses when encountering particular In an effort to disentangle such com- environmental conditions, people’s causal plex issues, evolutionary psychologists have interpretations of these identical behaviors established empirical programs with the may vary considerably. Among those fac- explicit purpose of identifying and explain- tors contributing to attributional differences ing the ultimate function of human think- between societies and between individuals ing in different problem-solving domains, are cultural traditions, narratives, religious such as those found in the social, physical, indoctrination, and education. and biological environments (see examples The key point is that such causal inter- in Barkow et al., 1992; Buss, 2005). The uni- pretations of behavior matter little in the que metatheoretical perspective of evolutio- long run – so long as an adaptive behavior nary psychology has also contributed to an occurs, it makes no difference whether peo- understanding of human behavior that goes ple believe that the gods made it so or that well beyond that of solely proximate expla- it was triggered by the state of the economy. nations proffered by many social learning This informs us that the cognitive systems theories. Importantly, evolutionary psychol- supporting many human behaviors appeared ogy argues that human behavior is motivated earlier in evolutionary phylogeny than did not by the conscious interests of people in the conscious awareness that currently over- infinitely malleable social environments, but sees and interprets them. According to this rather by the genetic interests of human org- rationale, if self-consciousness were integral P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

the serpent’s gift: evolutionary psychology and consciousness 601

in causing adaptive behaviors, then it is rea- cation of this fact is that, regardless of both sonable to expect that all individuals, irre- cross-cultural and individualistic differences spective of population or individual differ- in how people explain the causes of their ences, would provide the same type of causal own behaviors, what ultimately determines explanation when interpreting these behav- behavior are the implicit, evolved psycho- iors. In this light, self-consciousness is rightly logical mechanisms that instantiate a given considered an epiphenomenon with respect course of action whenever an individual is to these strategies, in that it is ostensibly confronted with a problem that the human inconsequential to the standard operations mind is designed to solve. of many psychological adaptations. As a gen- We also suspect, as do others (e.g., Crook, eral rule of thumb, whenever a behavior can 1980; Donald, 2000; Humphrey, 1976), that be reliably predicted to arise in response this is only part of the picture and that self- to a definable set of environmental fac- consciousness in fact played an enormously tors, and whenever post-hoc explanations important role in the evolution of psycholog- for this identical response vary from person ical adaptations that are specific to human to person or from culture to culture, self- beings. This is because self-consciousness consciousness has served at most a periph- seems to have meaningfully disrupted many eral role in the evolution of the psychological ancient psychological adaptations that hu- adaptation supporting this behavior. Homo man beings share with other species and to sapiens, like any other extant species, has have presented a new series of challenges a deep history; in addition, it also has one that our distant human ancestors were never that is characterized by only a recent split forced to confront. These challenges, we be- from the other primate clades. We there- lieve, were initially focused in the social fore suspect that a significant proportion of realm, with self-consciousness producing human psychological adaptations fit into this individuals who were more keenly aware category. That is, we believe that much of of their own knowledge and motivations human behavior is likely governed by uncon- and those of others. Such awareness could scious decision-making strategies that led to have provided great advantages, but with it genetic fitness throughout the course of pri- great problems. With an onslaught of new mate evolution. For such adaptations, self- dynamical problems caused by conscious- consciousness principally serves a spectator ness, human beings evolved a fundamentally role, allowing explanatory searches for the novel suite of adaptive solutions designed to causes of adaptive behaviors but not insert- redress these problems. ing itself into the decision making in any Along these lines, evolutionary psychol- meaningful way. ogy distinguishes the proximate causes of Research in evolutionary psychology has human behaviors from their distal causes. amassed considerable support for the idea The proximate level of behavioral causa- that there exists an underlying genotypic tion consists of motivational causes, such as structure in human beings that leads, in affective, perceptual, and epistemic states interaction with the environment over the that the individual experiences subjectively. course of development, to the phenotypic In contrast, the ultimate level of behavioral expression of psychological systems spe- causation consists of the intuitive, domain- cially designed to solve recurrent environ- driven psychological processes promoting mental problems. Across human societies, adaptive behavior that are barred from the and across the human life cycle, individu- individual’s conscious access. Because any als encounter the same set of basic chal- given adaptive behavior has both a prox- lenges in the social and physical environ- imate cause that must gear the individ- ments – challenges that, if gone unmet, ual toward engaging in a specific course of would directly threaten the successful prop- action, as well as a distal cause that strives to agation of their genes. One important impli- ensure that this course of action is in the best P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

602 the cambridge handbook of consciousness

interests of the individual’s genes, these two large head makes birth difficult, due to lim- levels of causation are inseparable. Psycho- its on how broad a woman’s hips can be logical adaptations are complex, rule-driven and still afford bipedality. As a result, many processing systems that respond to domain- women and infants have died in childbirth. specific environmental factors. However, the cognitive benefits of a large brain must have been greater than the cost in maternal and neonatal mortality; otherwise Adaptations, Byproducts, Noise, selective factors would have worked against and Exaptations such costly anatomical constraints and alter- At this point, some comment should be native fitness-conferring mechanisms would made about evolutionary psychology’s adop- likely have evolved. tion of an “adaptationist program.” Evolu- A related concept popular with many tionary psychology has often falsely been evolutionary biologists (e.g., Tattersall, accused of assuming that any species- 1998) is that of exaptation, defined as “a universal contemporary behavior must be an feature, now useful to an organism, that did adaptation. But this is not so. Many features not arise as an adaptation for its present of the modern human mind and behavior role, but was subsequently co-opted for are byproducts of other adaptations, or are its current function.” Further, exaptations simply noise. Some may actually be mal- are “features that now enhance fitness, but adaptive, just not so maladaptive as to have were not built by natural selection for their caused the elimination of the genes underly- current role” (Gould, 1991,p.47). The ing these features from the species’ genome. classic example of an exaptation is the case David Buss and his colleagues (Buss, Hasel- of avian feathers, which evolved initially to ton, Shackelford, Bleske, & Wakefield, 1998) serve a thermoregulatory function but were defined adaptations as reliably developing, co-opted to facilitate flight in birds. inherited characteristics, produced by nat- Although the concept of exaptations has ural selection, that served to solve recur- not generally been accepted by evolution- rent problems in the environment of evolu- ary psychologists (Buss et al., 1998), we tionary adaptedness and resulted in greater believe that the basic idea is solid – many inclusive fitness. Buss et al. used the umbil- of the products of evolution arose based ical cord as an example of an adaptation, on byproducts of other adaptations or fea- as it solved the problem of how to get tures that, initially, had no inherent function nutrients from a mammal mother to her for an organism. This is likely to be espe- fetus. In contrast, byproducts are features cially true for brain evolution, with parts that have not been shaped by natural selec- of the brain being co-opted for functions tion and did not solve some recurring prob- they were not originally selected to per- lem, but are a consequence of being associ- form. Yet, once co-opted, any new function ated with some adaptation. The belly button must pass through the sieve of natural selec- would be an example of a byproduct. Finally, tion. For instance, although feathers may not noise refers to random effects that may have initially evolved for flight, they became be attributed to mutations, changes in the necessary for birds to fly, making them an environment, or variations of development, adaptation. Likewise, even if many human such as the shape of one’s belly button. cognitive abilities are the products of the Thus, not all evolved characteristics should co-opting of brain tissue originally used for be viewed as adaptations. The belly button other purposes, it is reasonable to ask what clearly evolved, but it cannot accurately be new problems (if any) these abilities solved viewed as an adaptation. Moreover, some and if they, too, may be adaptations (albeit adaptations may have negative side effects co-opted ones). We thus treat exaptations (byproducts). For instance, the fetus’s large as special cases of adaptations. Because nat- skull is surely an adaptation, housing the ural selection is not forward looking but large brain that resides within it. Yet, this serves only to adapt organisms to their local P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

the serpent’s gift: evolutionary psychology and consciousness 603

environments, it is likely that many contem- be a series of species, some of which were porary and ancient adaptations may have surely our ancestors and others of which been co-opted from other seemingly unre- likely lead to evolutionary dead ends. Fig- lated functions. This perspective may be ure 22.1 presents one possible phylogeny for of special importance to human intelli- human evolution, dating back about 5 mil- gence and for functions associated with lion years. These species differed in many the expanding neocortex that characterized physical characteristics (we can only guess members of the hominid line over the past at what their behavior might have been 5 million years. Consciousness may be the based on brain size and some artifacts), per- product of our big brain, but we can only haps most prominently being brain size in guess at the selection pressures, if any, that relation to body size. Figure 22.2 presents generated this ability. the encephalization quotients for contemporary chimpanzees (Pan troglodytes), modern human beings (Homo sapiens), and three species believed to be ancestral to Homo The Evolution of the Human Brain sapiens: Australopithecus afarensis, which roamed Africa about 3 million years before Changes in Brain Size over Hominid present; Homo habilis, the first member of Evolution the Homo genus that first appears in the Human beings are noted for their big brains fossil record about 2.5 million years ago; relative to their body size. Primates, in gen- and Homo erectus, who left Africa to pop- eral, have large brain-to-body size ratios, but ulate Asia and Europe about 1.5 million this trend is exaggerated in human beings years ago. As can be seen, the encephal- (Jerrison, 1973; Rilling & Insel, 1999). Jer- ization quotient of Australopithecus afarensis rison (1973) developed the encephalization was only slightly greater than that of mod- quotient (EQ) to evaluate the expected brain ern chimpanzees, with this value increasing weight/body weight ratio for animals within sharply over the next 3 million years in the a family. For instance, given the typical pat- genetic line that presumably led to Homo tern of changes in brain and body weight in sapiens (Tobias, 1987). The modern human mammals, brain weight should increase at brain, then, is a reflection of a more gen- a certain rate relative to increases in body eral pattern shown in primates and particu- weight. If a species’ brain is smaller than that larly in hominids, those big-brained, bipedal expected for its body weight, the encephal- animals of which Homo sapiens is the only ization quotient will be less than 1.0, and extant species. it will be greater than 1.0 if its brain is Building bigger brains, at least in pri- larger than expected for its body weight. mates, seems to be the result of extending Most primates have encephalization quo- the time the brain can grow; delaying the tients greater than 1.0, with chimpanzees offset of brain growth results in the produc- being 2.3, meaning their brains are, on aver- tion of more neurons (Finlay & Darlington, age, more than twice the size expected for 1995; Finlay, Darlington, & Nicastro, 2001) a mammal of their size. This impressive and greater dendritic and synaptic growth. brain/body ratio is dwarfed by that of human However, in human beings, the brain can beings, however, which is more than three only get so large before the skull that con- times greater still (EQ = 7.6, Jerrison, 1973; fines it becomes too big to fit through the Rilling & Insel, 1999). birth canal. As a result, human infants are Of course, human beings did not evolve born prematurely, and much of brain devel- from chimpanzees, but last shared a com- opment occurs postnatally. Were human mon ancestor with modern chimps between gestation to correspond to what would be 5 and 7 million years ago. In between our expected for their brain and body size, chimp-like common ancestor1 and contem- women would be pregnant between 18 and porary people, paleoanthropologists descri- 24 months (Gould, 1977). P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

604 the cambridge handbook of consciousness

7/8

Unknown Ancestor

6 Sahelanthropus tchadensis

5 Ardipithecus kadabba Ardipithecus ramidus

4 Australopithecus anamensis

s Ago Australopithecus afarensis ear Y

3 Australopithecus africanus Australopithecus gahri

Millions of Homo habilis

Australopithecus 2 (Parathropus) Homo erectus (Homo ergaster) robustus

1 Archaic Homo sapiens (Homo heidelbergensis) Homo floresiensis Homo neanderthalensis Homo sapiens 0

Figure 22.1. One possible phylogenetic tree of human evolution.

Although the canonical perspective of malian evolution can be attributed to delay- evolutionary psychology is that adaptive ing “neuronal birthdays” (when precursor mechanisms are domain-specific in nature, nerve cells stop dividing symmetrically and a more parsimonious interpretation, we be- begin their migration within the neural lieve, is that increased brain size afforded tube), and not to changes in specific areas greater general information-processing capa- of the brain associated with particular func- city that may have permitted the evolu- tions, which would be indicative of domain- tion or execution of more domain-specific specific selection pressures (Finlay et al., mechanisms, particularly in the social realm 2001). This is similar to claims made by (see Bjorklund & Harnishfeger, 1995; Bjork- Gould (1991), who argued that many aspects lund & Kipp, 2001; Bjorklund & Pellegrini, of modern human intelligence do not rep- 2002; Geary, 2005). Consistent with this resent domain-specific evolved psychologi- domain-general perspective is the claim that cal mechanisms, but rather are the byprod- most of the increase in brain size over mam- ucts of an enlarged brain (see also Finlay P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

the serpent’s gift: evolutionary psychology and consciousness 605

8 7.6 ily functions. But if an animal is going to invest the time necessary to learn impor- 7 tant aspects of its environment, it should 5.5 6 be relatively long-lived. Brains are metabol- 5 1995 4 ically expensive (Aiello & Wheeler, ), 4 and an animal whose existence on this earth 3.1 3 is measured in weeks or months, rather than

Encephalization Quotient 2 years or decades, would be better off investing its time and caloric resources in pursuits 1 that do not require substantial learning. And 0 the more an animal has to learn to achieve A. afarensis H. habilis H. erectus H. sapiens inclusive fitness, the longer its prereproduc- Species tive period needs to be. Human beings, of Figure 22.2 . Encephalization quotients for course, fit this bill well, not reaching repro- several hominid species (from Tobias, 1987). ductive maturity until the teen years, with Reprinted with permission. some anthropologists suggesting that this was likely closer to 20 years of age for our et al., 2001). Although this argument sug- ancestors (e.g., Bogin, 1999; Kaplan, Hill, gests that not all features of the human brain Lancaster, & Hurtado, 2000). and mind were specifically targeted for selec- In particular, we, and others, have argued tion (see Geary, 2005; Geary & Huffman, that big brains and slow development 2002), it does not mean that differences evolved in primates to deal with the com- between chimpanzees and human beings, plexity of their social group (e.g., Alexan- for example, are due only to differences in der, 1989; Bjorklund & Bering, 2003a; Bjork- the total volume of brain tissue between lund & Harnishfeger, 1995; Byrne & Whiten, these species. There are many differences 1988; Dunbar, 1992, 2001; Geary & Flinn, in the microcircuitry of many parts of the 2001; Humphrey, 1976). Our remarkable brains of monkeys, chimpanzees, and human technological and abstract reasoning skills beings, for example, suggesting that spe- have been co-opted from the “intelligence” cific brain areas and cognitive functions have evolved to deal with cooperating, compet- indeed undergone selective pressure (Preuss, ing, and understanding conspecifics, and this 2001). This implies that, even if much of is a trend observed in primates in general. brain evolution within the hominid line can This trend is seen in research by Dunbar be attributed to a general mechanism asso- (1992, 1995, 2001), who reported a signif- ciated with the delay of neuronal birthdays, icant relation between measures of brain subsequent specialization of brain and cogni- size and social complexity among primates tive functions, which are relatively domain- (correlation between size of neocortex and specific in their application, could still have group size = .76). Moreover, larger brain taken place (Geary, 2001). size is also negatively associated with length of the juvenile period (Bonner, 1988), suggesting that both brain size and delayed Big Brains, Slow Development, development are important interdependent and Social Complexity factors that are related to success in com- What are big brains for? For one thing, large plex societies. This triadic relationship was brains are greatly beneficial for learning. A empirically demonstrated by Joffe (1997), species that lives in varied environments or who reported that the proportion of the requires sophisticated memory abilities to lifespan spent as a juvenile among 27 differ- navigate its environment or to remember ent species of primates was positively corre- the location of hidden caches of food can- lated with group size and the relative size not achieve these feats with a brain that is of the nonvisual neocortex. We make no barely large enough to control its basic bod- claims that any one of these factors is the P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

606 the cambridge handbook of consciousness

cause for another; surely, brain size, length of Higher-order cognition in great apes has juvenile period, and social complexity inter- been an area of great contention, with acted synergistically, with large brains and an some scientists arguing that chimpanzees are extended juvenile period being necessary for “almost human,” possessing, in rudimentary mastering the ways of one’s group, and social form, nearly all the intellectual abilities seen complexity in turn exerting selection pres- in Homo sapiens (e.g., Fouts, 1997; Goodall, sures for increased brain size and an exten- 1986; de Waal, 1986); whereas others con- sion of the juvenile period. tend that chimpanzees are merely clever As suggested above, the increasing brain “behaviorists,” able to accomplish feats of volume over the course of primate and social and technological complexity with- hominid evolution likely is the result of con- out the need for abstract (i.e., self-conscious) straints in neural development (e.g., Finlay cognition (e.g., Povinelli, 2000; Povinelli & et al., 2001). Yet, it seems that the increased Bering, 2002). Although a detailed discus- computing power that larger brains afforded sion of this literature is beyond the scope was put to good use, specifically to deal with of the present chapter (see Bjorklund & Pel- the complexity of primate social groups. It legrini, 2002; Suddendorf & Whiten, 2001; also suggests that consciousness is not nec- Tomasello & Call, 1997, for reviews), we essary for life in a complex primate social now present briefly some of the evidence for group. However, we believe that when neu- and against higher-order cognition in chim- ral organization produced consciousness, it panzees and the implications it may have for did so in a context in which it could be put the evolution of human consciousness. to good use, specifically in social cognition. Mirror-Self-Recognition Precursors to Consciousness: Many comparative psychologists have The Comparative Psychology argued that evidence of mirror “self- of Consciousness recognition” in great apes is diagnostic of self-consciousness, and this, in turn, has led If the common ancestor of human beings and them to infer that such species must have contemporary members of the Pan genus empathic social cognition as well (Gallup, were anything like extant chimpanzees, they 1982, 1985; Jolly, 1991). This position was lived a highly complex social life. For exam- initially advanced by Gallup (1979), whose ple, chimpanzees in the wild have been original mirror self-recognition procedure shown to possess at least crude culture, as of placing a dye mark on a hidden portion reflected by the transmission from one gen- of an animal’s body and then recording its eration to the next of complex behaviors responses to the mark when confronted involved in grooming, nut cracking, and ter- with a mirror has become the litmus test for mite fishing, for example (Whiten et al., self-awareness in other species. If the animal 1999). Many of these behaviors are unique to reaches up to touch the mark, then it is said a particular chimpanzee troop and so cannot to understand that the mirror image is a be attributed to species-universal behavioral representation of itself. If it does something features. Human beings and chimpanzees else, however, such as touch the surface of also show considerable overlap when it the mirror or threaten its own image, then comes to social behaviors, such as, among it is said to have “failed” the mark test and many others, status striving, coalition for- have shown no understanding of its own mation, reconciliation, and tit-for-tat strate- subjective existence. gies (e.g., de Waal, 1982, 1986; Goodall, To date, other than human beings, who 1986). But do chimpanzees achieve these show mirror self-recognition at about 18 complex social feats using higher-order cog- months of age (see Brooks-Gunn & Lewis, nition involving consciousness, or can their 1984), only chimpanzees, orangutans, and accomplishments be explained otherwise? a few gorillas have “passed” the mark test P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

the serpent’s gift: evolutionary psychology and consciousness 607

(see Suddendorf & Whiten, 2001; Swartz, & Kruger 1993). However, it is still unclear Sarauw, & Evans, 1999), although vari- whether such atypical rearing experiences ants of the test that have been devised endow these animals with an understanding for use with the anatomical constraints of of intentionality or whether they simply be- dolphins (Reiss, & Marino, 2001) suggest come more sensitive to human behavioral that this species may demonstrate mirror contingencies that co-occur with specific self-recognition as well. Interpretation of intentional states (Bering, 2004a; Bjorklund these findings vary widely, however, with & Bering, 2003b; Bjorklund & Rosenberg, some researchers arguing that such mirror- 2005; Call & Carpenter, 2003). contingent behaviors are clear evidence of Boesch (1991, 1993) has argued that one self-consciousness and others arguing that phenomenon that would seemingly require they demand only an ability to learn how cer- conscious cognition is explicit teaching. He tain kinesthetic-proprioceptive experiences reasons that teaching requires the under- map onto a mirror image (for a review of standing of others as not possessing infor- this complex debate, see Parker, Mitchell, & mation, and thus behaviors that appear Boccia, 2006). designed to change the epistemic content of others’ minds would (in principle) be evidence of the instructor’s metarepresen- Social Learning tational capacities. There have been several The impressive social learning of chim- observations of female chimpanzees in the panzees, on the surface, would seem to wild teaching their young offspring how to involve an appreciation on the part of the crack nuts (e.g., Boesch, 1991, 1993; Green- observer of the goal, or intent, of the field, Maynard, Boehm, & Schmidtling, model, a form of secondary representa- 2000). For example, mother chimps were tion and seemingly a characteristic of con- observed to position the anvil and hammer scious creatures. But not all social learn- rocks and the nut in such a way that all an ing requires such mental representation. For infant had to do was strike the nut to open it. instance, Tomasello and his colleagues (e.g., At other times, the female would move espe- Tomasello, 1996, 2000; Tomasello, Kruger, cially slowly in the presence of her infant. & Ratner, 1993) have proposed that only Although these are impressive demonstra- true imitation requires an understanding of tions and consistent with the interpreta- the goals, or intentions, of the model, in tion that mother chimpanzees actively teach addition to replication of important aspects their offspring complex technological skills, of the model’s behavior. Such imitation they have been observed only rarely and do requires the ability to take the perspec- not seem to be a common way in which tive of another, apparently requiring con- “cultural” information is transmitted (see scious awareness. Although great apes often Tomasello et al., 1993). master complicated tasks after watching a model perform similar problems, such Understanding the Perspective of Others learning usually occurs over multiple trials and involves significant trial-and-error learn- In other research, subordinate chimpanzees ing (e.g., Whiten, 1998; Whiten, Custance, seem to realize when and when not a dom- Gomez,¨ Teixidor, & Bard, 1996); in general, inant chimp can see a valued food item there is little evidence for true imitation and will only “compete” for the food when of actions on objects in chimpanzees (e.g., it is out of the dominant animal’s sight Tomasello, Savage-Rumbaugh, & Kruger, (Hare, Call, Agentta, & Tomasello, 2000; 1993). The exception seems to be for Hare, Call, & Tomasello, 2001). This behav- great apes that have been enculturated by ior implies that chimpanzees understand human caregivers (e.g., Bering, Bjorklund, & that “sight implies knowledge,” a seemingly Ragan, 2000; Bjorklund, Yunger, Bering, & rudimentary (but valuable) ability in social Ragan, 2002; Tomasello, Savage-Rumbaugh, cognition. Yet, research by Povinelli and his P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

608 the cambridge handbook of consciousness

colleagues (e.g., Povinelli & Eddy, 1996a; We cannot say that the primate litera- Reaux, Theall, & Povinelli, 1999) questions ture paints an easily interpretable picture this interpretation. In Povinelli’s research, of higher-cognitive abilities in great apes. It chimpanzees face two experimenters, one seems that chimpanzee social cognition is, with her eyes open and the other with in many ways, very much like that of human her eyes somehow occluded (e.g., her eyes beings. Yet, the understanding they demon- are closed or she’s wearing a mask). When strate of the knowledge of others – perhaps given the opportunity to make a reaching the critical component in human social cog- response to either of the experimenters to nition and one for which self-consciousness get food located between them, the chim- is required – is limited at best. We feel com- panzees respond randomly. Unlike the apes fortable saying that chimpanzees and the in the food-competition studies of Hare and other great apes do not experience a sense of his colleagues (2000, 2001), they seem not self-awareness and do not possess an under- to appreciate that the eyes are a source of standing of belief/desire reasoning compara- knowledge. ble to the degree characterizing nearly every With respect to complex social behav- normal 5-year-old human child. But the ior, chimpanzees do occasionally seem to roots of such awareness may be visible, and engage in deception; for example, females in the absence of controlled conditions, one occasionally place a hand over a subordinate can easily misinterpret the actions of these male’s mouth during furtive mating, serving animals as being based on an understanding to prevent the male’s screams from reach- of the knowledge and motivations of others ing the ears of a dominant male and alerting (as we often do with our pets). Assuming him to the behavior (see Whiten & Byrne, that our common ancestor had a social orga- 1988, for other examples). Yet such decep- nization and social learning abilities similar tion does not necessarily require that an ani- to those of extant chimpanzees, when con- mal knows what another animal is think- sciousness did first appear, it was in a species ing, but could have been acquired via pro- and context that would readily capitalize on cesses of trial-and-error learning. In related its attributes. research, when chimpanzees are given false- belief tasks under laboratory control, so that the only way to solve the problem is to Evolutionary Psychology appreciate what another individual knows, and Consciousness they fail (Call & Tomasello, 1999). Taken together, these findings and others from a Adaptive Information-Processing variety of controlled investigations lend sup- Mechanisms: The Assumption port to the idea that chimpanzees have, of Implicit Cognition at best, a vastly impoverished understanding of intentionality (Bering & Povinelli, For the most part, when evolutionary psy- 2003; Povinelli & Bering, 2002; Tomasello chologists talk about information-processing & Call, 1997). They have failed to distin- mechanisms or strategies used to solve spe- guish between ignorant and knowledgeable cific problems (e.g., mate-selection strate- social partners (e.g., Call & Tomasello, 1999), gies), they do so in the same way that evo- to understand seeing as a psychological state lutionary biologists speak of physiological (e.g., Povinelli & Eddy, 1996a), to distinguish mechanisms or behavioral strategies. Such between intentional and accidental actions “strategies” do not imply self-awareness. For (e.g., Povinelli, Perriloux, Reaux, & Bier- example, the mating strategy of a smaller- schwale, 1998), to instruct a na¨ıve conspe- than-average male fish to mingle among cific on how to go about a novel task that the females and mate inconspicuously with requires cooperative effort (e.g., Povinelli & them is quite different from the strategy O’Neill, 2000), and to understand the psy- used by a larger, more dominant fish. Nei- chological state of attention (e.g., Povinelli & ther strategy, of course, is conscious. Nei- Eddy, 1996b,c). ther animal has reflected upon the best P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

the serpent’s gift: evolutionary psychology and consciousness 609

way to get its genes into the next genera- else” (p. 10). Clearly, any behaviors that tion and chosen, after careful deliberation, are directly caused by consciousness can which course to take. These are unconscious, become targeted by natural selection. But a implicit strategies and of the same sort that novel twist to this logic, and one we think evolutionary psychologists propose underlie should be underscored in discussions of human behavior. Such an approach affords human cognitive evolution, is that, in a great explanatory benefits, for it permits psy- gossiping society, the inhibition of selfish chologists to explain the actions of human behaviors would also have become tar- beings in the same way that the actions of geted by natural selection (Bering, 2004b; non-human animals are explained. No spe- see also Bjorklund & Harnishfeger, 1995; cial pleading for the uniqueness of humans Bjorklund & Kipp, 2001; Dunbar, 1993). is necessary. That is, not engaging in a behavior that has As a result, higher-order cognition, such been conditioned by natural selection to as that requiring consciousness, is not a occur in response to specific environmental required topic for evolutionary-minded psy- contingencies, but instead overriding this chologists. This is reflected by examining tendency through higher-order cognitive the indexes of popular books on evolution- means, may under certain conditions be ary psychology. For example, the term “con- the more adaptive response. We argue that sciousness” is not found in the indexes of this is often the case for human beings, an Barkow, Tooby, and Cosmides’s (1992) semi- organism whose genetic fitness hinges on nal edited volume The Adapted Mind, Buss’s social reputation, impression management, (1999) textbook Evolutionary Psychology, and the advertisement of its altruistic traits. Cartwright’s (2000) textbook Evolution and The behavioral dispositions that we have Human Behavior, or in Pinker’s deservedly inherited from our genetically gratuitous popular books The Language Instinct (1994) prehominid ancestors, ancestors that did and The Blank Slate (2002). Yet the topic not have to worry about the existence of of consciousness receives considerable atten- other minds, must be suppressed very often. tion in other evolutionary psychology texts This is new psychology shaking hands with (e.g., Gaulin & McBurney, 2001; Palmer & old psychology. Palmer, 2002) and in popular books such as We also see an important overlap here Pinker’s (1997) How the Mind Works, reflect- between our definition of consciousness as ing, we think, the recognition by some evo- a system permitting metarepresentational lutionary psychologists of the central role thought and the more conventional view of consciousness plays in what it means to be consciousness, and the one often referred to human. But, in general, consciousness has in evolutionary terms as having conferred a been out of the mainstream of evolution- selective advantage, as a system that permits ary psychology, and certainly conscious cog- simulated outcomes on motoric planning. nition is not necessary to explain adaptive According to Cotterill (2001), “The adequa- behavior. tely endowed system conjures up a simula- Although evolutionary psychologists ted probable outcome of the intended motor need not postulate conscious cognition pattern, and vetoes it if the motion is ad- when trying to explain the adaptive value verse” (p. 8). Indeed, Jeannerod (1999) pres- of some cognitive mechanisms, they cannot ents evidence showing that overt movement totally ignore it. Cognitive scientists have and sheer imagery of these actions acti- recognized for some time the complex vate the same brain regions. Povinelli and interplay between consciousness and behav- Cant (1995) have speculated that a large- ior. In human beings, the relation between bodied arboreal primate, such as that envi- higher-order thought and action is a com- sioned as the common ancestor of humans plicated affair. As Cotterill (2001) writes, and great apes, would have profited enor- “We can think without acting, act without mously from symbolic representational abil- thinking, act while thinking about that act, ities enabling such foresight (a wrong move and act while thinking about something in the tree canopy could prove deadly.) We P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

610 the cambridge handbook of consciousness

would simply extend this line of reasoning ent species than we are today without con- by arguing that mental simulations of the pr- scious awareness, that does not preclude the obable outcomes of intended actions would question of what adaptive benefits, if any, have given human beings significant strate- consciousness afforded our ancestors. More- gic leverage in a variety of both physical and over, consciousness was a mixed blessing, social contexts. Social simulations would ins- perhaps producing as many problems as it tantiate alternative psychological construals solved. In this section, we look at some of that included representations not only of ot- the benefits and challenges brought about hers’ likely reactions to the self’s behaviors by consciousness. but also mental state attributions to these behaviors. Actions can thus be tailored to The Benefits of Consciousness the unique demands of the social situation. Although it may be the case that other know thyself, and know others primate species are capable of simulating Perhaps the dominant perspective of the behavioral outcomes of their actions in social “reason” for the evolution of consciousness contexts, we argue that only humans can was first presented by Nicholas Humphrey simulate the way that their actions will be (1976), who argued that consciousness interpreted and understood by social others. played a critical role when dealing with As we mentioned earlier, evolutionary members of our own species. For animals liv- psychology deals with both distal and ing in complex social groups, the ability to proximate causes of behavior, expressed predict, and possibly control, the behavior of through adaptive information-processing others would provide great advantage. Social mechanisms that are responsive to factors in primates, which surely included our ancient the local environment. Because, in human ancestors, already formed coalitions, coop- beings, psychological adaptations occur in erated and competed with one another, and conscious agents, these rule-driven process- sometimes used deception to obtain valuing systems are also tethered to the subjec- able resources or to avoid detection when a tive states of biological organisms. Environ- social rule was broken. Consciousness arose mental information that is detected and per- in this arena and provided an immediate ceived by the human brain is handled in ways social, and therefore reproductive, advan- that maximize biological success by either tage to its possessor. stimulating or inhibiting specific behavioral Note that it is implicitly assumed here responses. At least in human beings, how- that the ability to attribute mental states ever, the perception of these sources of to others is somehow linked to self- information is often accessible to conscious consciousness. This requires making the awareness, and any behaviors occurring in conceptual inference that one’s privileged response to environmental input may be access to one’s own proximate causal states closely monitored, if not controlled, by an feeds an understanding of others’ causal executive cognitive system. states. Specifically, it is assumed that any evolved social algorithms in human beings that necessitate the representation of what The Adaptive Benefits and Challenges others do or do not know first required the of Consciousness ability to reflect upon the epistemic contents of one’s own mind. Indeed, we believe On the surface, the benefits of conscious- that the ability to detect the intentions of ness are self-evident. Self-consciousness and other agents, as well as their emotional, epis- higher-order cognition are so integrated into temic, and perceptual status, has likely been the fabric of what it means (phenomenolog- an enormously influential factor in the evo- ically) to be human that there is no need lution of social adaptations in the human to look for its benefits. Yet, although it is species. Such a general representational sys- unquestionable that we’d be a very differ- tem, which falls under the rubric of “theory P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

the serpent’s gift: evolutionary psychology and consciousness 611

of mind” in the developmental and compar- where the confederate will think the treat ative psychological literatures, was capable is hidden. By age 4, most children solve the of transforming already existing ancient pri- problem correctly, stating that the confeder- mate adaptations in the social domain, such ate will have the false belief that the treat as reciprocal altruism and mate-guarding, is hidden in the first container. Most chil- into more complex adaptations demanding dren much younger than this age, however, the rapid inferential processing of informa- erroneously state that the confederate will tion dealing with mental states. believe the treat is hidden in container B. From this perspective, self-consciousness This, indeed, is where the treat is hidden, was first applied to social cognition and but this fact is known to the child and not may have permitted the evolution of the confederate. Results from experiments more domain-specific adaptations devoted using this and other variants of false-belief to dealing with conspecifics. For example, tasks indicate that young children do not Baron-Cohen (1995) proposed four separate behave as if they possess belief-desire reason- modules involved with theory of mind. The ing (see Wellman, Cross, & Watson, 2001). first module, the intentionality detector (ID), We would not want to declare these chil- permits one to infer that a moving object dren to be unconscious, but they fail to be may have some intent toward the individual able to take the perspective of the confed- (it may be trying to catch me), and the eye- erate and to understand that other people direction detector (EDD) interprets eye gaze. have knowledge and desires different from These modules develop by around 9 months their own that guide their behavior. People in in infancy and likely do not require full- all societies generally behave kindly toward blown consciousness. The shared-attention preschool children, so their lack of belief- mechanism (SAD) module is involved in tri- desire reasoning rarely causes them much adic interactions (e.g., person A and per- trouble. But it is difficult to imagine any son B can each be looking at object C and adult (or child beyond the age of 7 or 8) who understand that they see the same thing) lacks such reasoning functioning well in any and develops at around 18 months in human human culture, modern or ancient. beings. The theory-of-mind module (TOMM) develops around age 4 in children and is sim- consciousness and the ilar to belief-desire reasoning as described by development of technological Wellman (1990), in which children under- skills stand that their behaviors are governed by According to Humphrey’s hypothesis, con- what they believe, or know, and what they sciousness evolved to play a central role in want, or desire, and so is the behavior of social cognition. But human consciousness other people. Such reasoning requires sec- extends beyond the social realm into tech- ondary representation and what is conven- nology (e.g., tool making) and educability. tionally referred to as self-awareness or self- Human beings are not the only tool mak- consciousness. Theory of mind has been one ers in the animal kingdom, but the tools of the most investigated topics in cognitive made by chimpanzees, for instance prepar- development over the past 20 years, in part ing sticks for use in termite fishing (e.g., because it is at the core of social functioning McGrew, 1992), are simple, uncomplicated in any human society. devices compared to the tools made by mod- The most frequently used tasks to assess ern human beings (even those possessing theory of mind involves children’s under- “stone age” technology, see Stout, 2002). standing of false belief. In one version of the Moreover, Homo sapiens are the only species task, a child and a confederate watch as a that make tools to make tools. Is conscious- treat is placed in one container (A). The con- ness required for these accomplishments, federate then leaves the room, and the treat and if so, how did it evolve? is moved to a second container (B) while Mithen (1996) proposed that, with the the child watches. The child is then asked advent of language, “social intelligence starts P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

612 the cambridge handbook of consciousness

being invaded by non-social information, with culture can be considered to be biolog- the non-social world becomes available for ically secondary abilities. reflexive consciousness to explore” (p. 190). As defined, human beings may not be the In other words, people are able to repre- only species to display biologically secondary sent their thoughts and actions to them- abilities; chimpanzees, as we noted previ- selves (or re-representation, following the ously, pass on unique cultural knowledge arguments of Karmiloff-Smith, 1992). For from one generation to the next (Whiten Mithen, with consciousness, general intel- et al., 1999), and such accomplishments as ligence now serves to integrate the various nut cracking and termite fishing can be con- modules of the mind (e.g., social, techno- sidered to reflect biologically secondary abil- logical, natural history) and, with this inte- ities (Bjorklund & Bering, 2003a). But it gration, permits the construction of tools is human beings who have made the most and the transmission of knowledge in a way of culturally acquired cognitions, and this unprecedented in the animal world. With has been achieved, we argue, through con- consciousness, our ancestors could reflect on sciousness. The acquisition of such skills what they knew, using information acquired as reading, arithmetic, navigation, coordi- in one domain to bring to bear on issues in nated hunting, and external forms of mem- other domains. Learning can extend beyond ory (such as pictures, or even intentionally the immediate context and be applied to sit- placed cues designed to prompt memory) all uations only imagined or in one’s memory. required a degree of self-regulation and sec- The extension of learning is likely an ondary representation not available without important consequence of consciousness. consciousness. Self-awareness, by itself, may Self-awareness is, of course, not necessary not be sufficient for the successful execu- for complex learning to occur. But much tion of biologically secondary abilities. Indi- of what makes human beings unique is our viduals required the ability to sustain atten- educability, our ability to acquire information and to avoid distraction (i.e., to stay tion and procedures for solving problems “on task”) and to have sufficient working that our ancestors never encountered (see memory to achieve many secondary skills. Bjorklund & Bering, 2000). Relevant here These are abilities that likely evolved with is Geary’s (1995) distinction between bio- increased brain size (e.g., Bjorklund & Har- logically primary and biologically secondary nishfeger, 1995) and may have been neces- cognitive abilities. Biologically primary abil- sary for the emergence of consciousness and ities refer to those cognitive mechanisms higher-order cognition. But the conscious- that have undergone selection pressure over ness initially applied to social intelligence the course of evolution. The abilities them- was eventually applied to other domains, and selves and their developmental timetable are the result was an animal that created a com- species universal, they are acquired via rou- plex and rapidly changing culture, which tine interaction with the environment, and affected its members’ inclusive fitness more children are highly motivated to execute than its biology. them. Language is a prime example of a biologically primary ability. In contrast, biolog- The Challenges of Consciousness ically secondary abilities are those a culture “invents” to solve particular recurrent prob- other minds, new problems lems. They have not undergone selective As we have commented earlier, however, pressure and thus are not universal; they are possessing consciousness may have its draw- based on biologically primary abilities, but backs. For one thing, if consciousness pro- often require substantial practice to achieve, vides a window to one’s own thoughts and and children may need external motivation educated guesses to the thoughts of others, to execute them. Reading is a good exam- it is highly likely that those “others” possess ple of a biologically secondary ability. Most the same insight. This makes social inter- of the cognitive accomplishments associated course all the more complicated, particularly P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

the serpent’s gift: evolutionary psychology and consciousness 613

for one who may be less proficient at “mind ance of eye contact, and defensive postur- reading” than others. Because intentional ing. Because individuals who leaked these states are actually the causes of behavior, cues in the environment of evolutionary any ineffectiveness in taking these states into adaptedness were less able to deceive oth- account when strategically interacting with ers in the interest of their own genetic fit- others who are also able to detect inten- ness, individuals who were able to deceive tionality would be highly detrimental. If themselves about their own selfish motives other individuals do employ such knowl- may have been at a select advantage. These edge effectively (i.e., to adaptive ends) while latter individuals possessed a new type of the self merely represents others’ mental psychological adaptation – self-deception – states and perseveres with its old uncon- that essentially militated against the behav- scious devices, then unavoidably the self’s ioral effects of having knowledge of their genetic fitness would become reduced sig- own socially maligned intentions. As a con- nificantly. For the successful organism, how- sequence, they could deceive others more ever, the structured and organized use of effectively in the service of their own genes intentionality – again, a categorically new (Nesse & Llloyd, 1992; Trivers, 1981, 1985). brand of social information perhaps unde- Trivers (1981) therefore reasons that “the tectable by any other species – may have mind must be structured in a very complex elaborated these already existing psychologi- fashion, repeatedly split into public and pri- cal adaptations by applying strategic inferen- vate portions, with complicated interactions tial mechanisms involved in assessing other between the subsections” (p. 35). agents’ motives (Haslam, 1997). Interestingly, implicit in this rationale is the suggestion that the “unconscious” of the counter-intuitive challenges Freudian variety has a more recent phy- of consciousness logenetic history than does consciousness, In addition to the problems of dealing essentially emerging to serve as a repos- with other self-aware conspecifics, there may itory for those self-epistemological states also be some counter-intuitive detriments that may seriously hazard people’s biologi- of self-consciousness. According to Trivers cal success. In addition to Freud, the exis- (1981, 1985), being self-conscious of certain tential philosopher Kierkegaard (1946/1849) proximate, socially proscribed motivations seemed to capture these ideas when he could put someone at a disadvantage rela- wrote in The Sickness Unto Death that “with tive to those who were able to mask these every increase in the degree of conscious- intentions through either self-deception or ness, and in proportion to that increase, the through the targeted loss of conscious access intensity of despair increases: the more con- to such motives. In effect, self-consciousness sciousness, the more intense the despair” began to interfere with adaptive functioning (p. 345). For the contemporary evolutionary in the social domain. The ability to repre- psychologist, this despair is probably defined sent one’s own selfish intentions and motives in terms of intrusive higher-order thoughts can seriously disrupt the efficiency of adap- that interfere with an individual’s adaptive tive behaviors because these motives must information processing. Although interpret- be well hidden from others for adaptive out- ing correlations is an inherently difficult comes to occur. Because emotions are closely task, neurobiological findings of shrinkage tied to intentions, the ability to engage in in hippocampus volume in adult women adaptive social behaviors may be affected who experienced childhood sexual abuse negatively by having these selfish intentions and who have also been diagnosed with “leak” through behavior (Ekman & Friesen, post-traumatic stress disorder may support 1975). Social partners may be able to detect hypotheses arguing for an adaptive role of ulterior motives through a variety of subtle, unconscious processes (Bremner et al., 1999; affectively induced behavioral cues, such as but see Gilbertson et al., 2002). Repression a higher than normal tone of voice, avoid- of such traumatic experiences so that they P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

614 the cambridge handbook of consciousness

are inaccessible to declarative memory may than a trait that evolved directly through help sustain brain regions devoted to short- natural selection. As we have suggested preterm and autobiographical memory. viously, self-consciousness may, for exam- In another hypothesis, Burley (1979) has ple, have been an outgrowth of a dramatic argued that, with the onset of conscious- expansion of the frontal cortex during ness, women who could detect the physio- hominid evolution, a highly adaptive trend logical cues signaling ovulation in their own that greatly increased the general intelli- bodies might begin consciously regulating gence and planning abilities of these species their birth cycles. The likely result of this (Banyas, 1999; Luria, 1973). In line with purposeful regulation is fewer pregnancies, this, self-consciousness may rightly be con- which, although satisfying the “egocentric” sidered an “all or nothing” phenomenon, interests of individual women (e.g., avoid- rather than appearing in stages and degrees ing pain, having a manageable family size, over the course of phylogeny. As Humphrey accruing enough resources before having (1992) puts it, “There must have been a children), would also be necessarily detri- threshold where consciousness quite sud- mental to the genetic fitness of these same denly appeared – just as there is a threshold women. Indeed, some women, according to that we ourselves cross in going from sleep Burley, would fully exploit these physiolog- to wakefulness” (p. 206). ical indices and avoid pregnancy altogether If self-consciousness emerged as an by continually abstaining from intercourse unavoidable byproduct of such a cortical near the time of ovulation. Natural selection expansion, it may not have been easily might have therefore concealed ovulation “removed” by natural selection because such from women’s conscious awareness, elimi- a reorganization of the brain might have nating estrus and sharply reducing their sen- come with costly adaptive tradeoffs. Instead, sitivity to this critical reproductive period. it may have been more economical for In fact, evolutionary psychological research nature to have “allowed” human conscious- indicates that ovulating women, who are not ness by keeping the species’ evolved neu- taking oral contraceptives, are more sexu- ral organization intact while hammering out ally receptive, as reflected by their emotional a series of constrained, novel psychological ratings of the smell of androstenone (e.g., adaptations that were specifically designed Grammer, 1993) and also their dress and to handle the new adaptive problems of social signals, than are nonovulating women. consciousness. What is interesting about these accounts This approach to consciousness differs is that consciousness is considered to pose substantially from those who have been adaptive challenges to human beings, rather searching for an adaptive explanation for the than to facilitate their genetic fitness. This presence of such a representational capacity. argument flies in the face of any the- For instance, in a discussion on the adaptive orist who has ever boasted that human value of introspection, McGuire and Troisi consciousness represents the pinnacle of (1998) note, evolutionary achievement. In contrast, in many regards, self-consciousness seems to There are good theoretical and empiri- be a maladaptive trait, providing individu- cal reasons for doubting the accuracy of als with access to information that causes introspections; the workings of most infras- them to engage in biologically poor deci- tructural systems are not available to sion making that disrupts their ability to awareness despite often heroic efforts to make them so. Nevertheless, it is reasonable carry out adaptive behaviors. Because self- to argue that a capacity that can so influ- consciousness tampered with evolutionarily ence how we think, feel, and act is unlikely stable behavioral patterns, it is very possible to have appeared by chance. Most per- that self-consciousness was an evolutionary sons introspect; most give credence to their byproduct of some other adaptive system or introspections; and introspections often trig- adaptive trend in the human brain, rather ger strong emotions (e.g., shame) (p. 125). P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

the serpent’s gift: evolutionary psychology and consciousness 615

Consciousness as a Double-Edged Sword ancestral tapestry containing more ancient psychological adaptations (Povinelli, Bering, As we noted above, consciousness had its & Giambrone, 2000). It is this synchronic advantages, despite its apparently consider- existence of the old and the new that char- able disadvantages. The evolution of con- acterizes human behavior and the break- sciousness might have been responsible for down of which results in dysfunctional con- a fundamentally novel set of psychological sequences. Failure to adequately defend the adaptations – specialized cognitive systems self from knowledge of its own biologically that went through the sieve of natural selec- oriented intentional states may lead to psy- tion and that made human beings quali- chopathy (Becker, 1974;Fabrega,´ 2002). For tatively distinct from other closely related instance, human males appear to have inher- species, such as chimpanzees. Once armed ited ancestral adaptations for female sexual with these psychological adaptations that coercion (Thornhill & Palmer, 2000), but were specially ﬁtted to the problem of con- these ancient adaptations must peacefully sciousness, human beings were poised to coexist with more recently evolved psycho- become an enormously successful species, logical systems that enable others to infer ultimately capable of radiating widely across males’ sexual intentions and to rapidly trans- both hemispheres and easily outcompeting mit information dealing with socially pro- competitor species. To this end, conscious- scribed behaviors. Under these social con- ness was a “blessing in disguise” because it ditions, the inhibition of certain ancestrally forced heritable potentialities in individual adaptive behaviors, such as sexual coer- members of the species that, in concert with cion, becomes adaptive, and the evolution consciousness, enabled these organisms to of psychological mechanisms (e.g., moral harness ﬁtness advantages that occurred at emotions such as guilt and shame) capable unprecedented evolutionary rates. That is, of disengaging such phylogenetically older once consciousness became “manned” with responses becomes essential. This coexis- an ensemble of psychological adaptations tence of ancient primate adaptations with that were functionally designed to operate recent human psychological adaptations is it, the information made accessible by self- not well understood, as demonstrated by the consciousness (i.e., the proximate causes of recent statements of primatologist de Waal behavior) could be systematically controlled (2002): and exploited by the species. We believe that this interpretation of the A major problem with the strategy of sin- evolution of consciousness can go some dis- gling out rape for evolutionary explanation tance in answering the following important is that the behavior is shown by only a question posed by Pinker (1997,p.132): “If small minority. The same criticism applies consciousness is useless – if a creature with- to Daly and Wilson’s (1988) well-known out it could negotiate the world as well as a work on infanticide by stepparents . . . If child abuse by stepfathers is evolutionarily creature with it – why would natural selec- explained, why do so many more stepfa- tion have favored the conscious one?” thers lovingly care for their children than abuse them? And if rape is such an advantageous reproductive strategy, why are there On the Reorganization of More so many more men who do not rape than Ancient Primate Adaptations who do? (p. 189, italics in original). In fact, the relative rate of sexual coer- How Previously Adaptive Social cion should drop off substantially once Behaviors Became Socially Maladaptive encroached upon by a representational sys- in Modern Human Beings tem capable of tracking the self’s intentions We propose that new psychological adap- and also the intentions of others. In other pri- tations emerging in response to the evolu- mate species, such behaviors as forced copu- tion of consciousness were woven into an lation and infanticide may lead to retaliatory P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

616 the cambridge handbook of consciousness

attacks, sometimes lethal, by offended so; whether others know about or what they higher-status parties who have direct per- believe about Agent A engaging in similar ceptual access to such incidents (de Waal, behavior in the past; whether others believe 1982; Goodall, 1986; Kummer, 1971). How- Agent B “deserved” such treatment; whether ever, (a) the inability of potential “victims” others believe Agent B experienced phys- to perceive the hidden, aggressive intentions ical or psychological pain from Agent A’s of potential “perpetrators”; (b) the inability behavior; whether others believe Agent A’s of perpetrators to track others’ knowl- behavior is diagnostic of a stable personal- edge of their behaviors; and (c) the inabi- ity characteristic and is thus likely to occur lity of observers to intentionally communi- again; whether Agent A knows something cate the occurrence of these transgressions of relevance about those who know about to na¨ıve others who did not witness the the behavior and can use this information event foster a high level of frequency of such strategically; whether others believe Agent behaviors in non-human primates. Indeed, A’s behavior was caused by his own intrin- by all accounts, such behaviors almost cer- sic traits or was governed by the circum- tainly will occur whenever the conditions stances surrounding the event; whether oth- are “right” – that is, when dominant animals, ers believe Agent A’s claims about the causes or those with connections to dominant ani- of his own behavior; whether others believe mals who may recruit others to the event Agent A’s displays of remorse over or regret through various alarm displays, are absent, about his behavior are sincere; whether oth- making retaliation unlikely to occur. ers believe Agent A possesses specialized This changes dramatically, however, with knowledge that makes him valuable; and a species such as Homo sapiens, for whom whether Agent B might have possessed such social information is capable of being trans- knowledge. In contrast, individual members mitted rapidly between parties far removed of non-human primate species may have from the actual behavioral incident (Dun- “witnesses” to their social transgressions in bar, 1993), and individuals (any one of whom the technical sense of the term, but such wit- is a potential perpetrator) are knowledge- nesses pose minimal risks to genetic fitness able to this extent. In such cases, retalia- given their inability to represent the epis- tion for social transgressions is likely to ensue temic states of those who did not perceive as a direct consequence of others gaining the proscribed incident. knowledge of the proscribed behavior. What This functional synchrony between old is defined as a transgression is going to be and new psychological adaptations should determined by the various socioecologies of not be terribly surprising when considering different groups. However, in general such the species’ recent phyletic history. In terms judgments will be made for those behav- of their general morphological characteris- iors that pose a clear and present danger tics, human beings seem to have undergone to the fitness interests of individual mem- what Mayr (2001) refers to as mosaic evo- bers of a community such that group func- lution, which is “evolutionary change that tioning is adversely affected and may not occurs in a taxon at different rates for dif- adequately sustain the needs of individuals ferent structures, organs, or other compo- within the group as long as the behavior is nents of the phenotype” (p. 288). That is, allowed to occur. It is difficult to imagine aside from a handful of trademark charac- any human socioecology where rape, homi- teristics, Homo sapiens has remained largely cide, and child abuse would not meet these unchanged at the level of its structural criteria. But the real confound is the fact appearance from the time it last shared that, for human beings, the possibility of a common ancestor with the great apes. retaliation is no longer just a matter of who Derived traits distinguishing the species was physically present at the time of the from other primates evolved mostly inde- transgression but also who else knows what pendently of those ancestral traits – the Agent A did to Agent B; what these others “morphological bulk” – that are responsible believe Agent A’s intentions were in doing for the taxonomic classification of human P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

the serpent’s gift: evolutionary psychology and consciousness 617

beings as primates. Human beings continue on any referential meaning. That is, although to share with the African apes a basic bau- it serves adaptive ends (e.g., promoting phys- plan, or body plan, where the only distinctly ical contact such as grooming and conse- human characteristics are differences in the quently allowing subordinate chimpanzees proportions of the arms and legs, the mobil- to avoid future conflict with rivals), those ity of the thumb, amount of body hair, skin chimpanzees receiving this communicative pigmentation, the size of the central nervous sign will fail to recognize the “aboutness” system and, related to this, the reduction in of the gestural display. This can explain prognathic facial features (Mayr, 2001). By why chimpanzees in the wild have not been all accounts, human beings are primates first observed to engage in referential gesturing and hominids second; some scholars have (e.g., Plooij, 1978). To understand the ref- even argued that the molecular differences erential nature of communicative displays, between human beings and chimpanzees are organisms must first be capable of rep- too minimal to warrant classification as sep- resenting those unobservable causal states arate genera (Diamond, 1992). that are behind intentional actions (Baldwin, It is unlikely that gross similarities and a 1991, 1993). small subset of novel derivations are lim- For instance, the holding-out-a-hand ges- ited to physical characteristics, however. ture that human beings inherited from an Consciousness-based psychological adapta- ancestral primate species is reinterpreted as tions in human beings have continued to being an intentionally communicative dis- interact with many of the cognitive pro- play (Povinelli et al., 2003; see also Franco & grams supporting adaptive behaviors in non- Butterworth, 1996; Vygotsky, 1962). Adult human primates. This explains why human caregivers, for example, who witness their beings share so many behavioral patterns young infants extending their hand toward with chimpanzees and the other great apes, an out-of-reach object on the ground will as we noted above. According to Povinelli automatically attribute the gesture to the and his colleagues (Bering & Povinelli, 2003; infant’s wanting the object and will subse- Povinelli, 2000; Povinelli et al., 2000), quently retrieve it for them. Some theorists the ability to represent the underlying have even speculated that indexical point- intentional states promoting these adaptive ing naturally emerges in ontogeny because behavioral patterns (e.g., those involved in there is a differential extension of this finger reconciliation, coalition formation) enabled in the human hand (Itakura, 1996; Povinelli human beings to reinterpret the separable & Davis, 1994). The index finger is essentially actions comprising them in fundamentally “pulled out” by parental response during novel ways. This reinterpretation process in the course of early reaching attempts; those turn led human beings to adopt new sets reaches that contain more explicit indexi- of behaviors that were qualitatively differ- cal extensions are interpreted more readily ent from the evolved action configurations by caregivers as communicatively meaning- upon which they were based. ful. Such indexical extensions may then be For example, human beings share with co-opted to provide a more accurate ref- chimpanzees a number of gestural displays erential trajectory when engaging in both that are morphologically identical between imperative and declarative communicative the two species (see Povinelli, Bering, & attempts using pointing. Giambrone, 2003). One of these displays is Comparative experimental analyses of the holding-out-a-hand gesture that is used chimpanzees’ and human beings’ com- by chimpanzees to recruit allies, to solicit prehension of pointing provide support reconciliation, and to seek physical contact for Povinelli’s reinterpretation hypothesis with conspecifics (Bygott, 1979; de Waal, (Povinelli et al., 1998). Chimpanzees that are 1982). Without an attending higher-order confronted with a human experimenter who cognitive system that enables the represen- is pointing to the correct location of a hid- tation of the intentional states causing such den food reward fail to understand the com- behavior, however, the gesture cannot take municative intention of this action. Rather, P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

618 the cambridge handbook of consciousness

they choose a location to which the exper- tion seems to have also constructed several imenter’s hand is physically nearest, even fundamentally novel psychological adapta- though the pointing gesture clearly refer- tions that are entirely based and depen- ences a distal location. In contrast, 2-year- dent upon this competency. For instance, old children easily interpret the referential although there are several species, partic- intent of the experimenter and are able to ularly insects, where individual organisms find the hidden prize in the distal location. systematically increase their own risk of This principle is captured by Dennett’s mortality in the face of threats of inter- (1987) concept of the intentional stance, species predation to their larger colonies which essentially describes human beings’ (e.g., Andrade, 1996; Holmes & Bethel, intuitive causal reasoning about the under- 1972; McAllister, Roitberg, & Weldon, 1990; lying reason, or purpose, that the designer of McAllister & Roitberg, 1987; O’Connor, the action had in mind. Specifically, human 1978; Poulin, 1992), human beings seem beings appeal to the mental states of oth- to be the only species where individual ers when attempting to explain their actions. members commit suicide in response to There is an intentionality that underlies the negative social appraisal of conspecifics. all purposeful behavior; this intentionality Although suicide is a leading cause of death consists of the range of proximate causes in human beings, it is completely unheard (e.g., emotions, cognitions, perceptions) to of in other primate species beyond highly which the agent in question has conscious questionable anecdotal accounts. According access. Povinelli reasons that, with the evo- to some theorists, shame is both the best pre- lution of human beings, an awareness of dictor of suicide and its primary determi- this underlying intentionality engendered a nant (Lester, 1997). The capacity to expe- new way of thinking about others’ behav- rience shame requires self-consciousness in iors, where explanatory searches are intu- that it is an emotional reaction to negative itively launched in pursuit of the prox- self-appraisal (Gilbert, 1998; Tangney, 2001). imate causes of behaviors. This explana- Indeed, Baumeister (1990) has even referred tory drive is extremely powerful in human to suicide as “escape from self.” In addi- beings and seems to extend to the physical tion, shame assumes metarepresentational domain as well (Baillargeon, 1994; Spelke, abilities because it is a secondary social emo- Breinlinger, Macomber, & Jacobson, 1992). tion, centering on others’ perceptions and Gopnik (2000) has argued that the need knowledge about the self’s negative traits to understand the causes of events is auto- (Tangney, 2001). matic, compulsive, and affectively based. These factors are interesting in light of the This argument puts intuitive explanatory current discussion because suicide has been theorizing about the causes of events in line implicated as a probable adaptation facili- with Geary’s biologically primary abilities. A tating inclusive fitness. In his “mathematical biologically secondary ability in this domain model of self-destruction,” de Catanazaro might be scientific explanation, which builds (1986, 1987, 1991, 1992; see also Brown, on the natural explanatory drive but which Dehlen, Mils, Rick, & Biblarz, 1999) has is cognitively effortful, requires extensive shown that suicide is positively correlated practice and training, and is mastered by with genetic burdensomeness to close kin; only a subset of the species’ population individuals whose lives negatively affect the (McCauley, 2000). reproductive opportunities of family members are significantly more likely to commit suicide than others. Adaptations subserving The Evolution of Qualitatively Unique inclusive fitness are fairly common among Psychological Adaptations various species (Hamilton, 1964); however, in Human Beings psychological adaptations involving human In addition to adding increasing complexity suicide promote inclusive fitness in a quali- to pre-existing adaptations, metarepresenta- tatively different manner because they are P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

the serpent’s gift: evolutionary psychology and consciousness 619

dependent on access to the informational so I’m going to discourage her from seeing reservoir of intentionality. Human beings him again”). seem to be unusual among other primates Although there is disagreement over in that the representation of another con- the precise developmental mechanisms by specific’s mere thoughts, or the mispercep- which children come to understand the exis- tion of their thoughts, can engender affec- tence of other minds (e.g., Wellman, 1990), tive reactions in the self that are translated it seems reasonable to assume that gaining into actual behavior. One such behavior is access to the type of information provided suicide, and it is therefore difficult to reason by self-consciousness (i.e., the self’s own that consciousness serves only the role of an intentions), at the very least, would facil- epiphenomenon in this case. itate an understanding that others’ behaviors are caused by similar means. It seems implausible to us that an ensemble of adap- The Evolutionary Significance of the tive heuristics concerning various relation- Mechanism of Consciousness ships between other people’s behaviors and To state that the mechanism of conscious- the causal states generating them could be ness is responsible for certain psychological adequately developed by an organism that adaptations is to say that, without having does not first have the intellectual device the means to access the type of information required to conceptualize the general cate- we are calling intentionality, many human gory of causal states (i.e., intentionality) in behavioral patterns could simply not have question. Theories are only as useful as the evolved. This is the position that we endorse concepts that they contain; without an abil- and also one that, we believe, most evolu- ity to conceptualize those mental constructs tionary psychologists would accept as well. that are correlated to specific types of behav- However, we make no claims that people’s iors (e.g., “Jakob opened the cabinet because explanations for their own behaviors must he thought that’s where the bananas were”), be correct (i.e., biologically relevant) for such theories, whether wrong or right, sim- self-consciousness to have been a meaning- ply could not be constructed. ful component of natural selection. Because Children’s conceptual knowledge about there is overwhelming evidence that peo- mental states might become progressively ple do not understand the biological rele- enriched through language (see Tomasello & vance of their own adaptive behaviors (see Bates, 2001) and also through personal expe- Buss, 1999; French et al., 2001), evolution- rience involving intuitive hypothesis testing ary psychologists rightly reject this notion of (Gopnik & Meltzoff, 1997), but such elab- propositional veracity. This is typically what oration can only build on a basic capacity evolutionary psychologists are referring to to represent such states to begin with. As when they state that self-consciousness has Bloom (1998) writes, “Language is a tool had little to do with the evolution of adap- for the expression and storage of ideas, but tive human behaviors. People do not need not a mechanism that could give rise to the to know why they do what they do in order capacity to generate and appreciate these to behave adaptively (e.g., “I find my wife ideas” (p. 215). This is supported by strong sexually attractive because she possesses fea- evidence of preverbal infants’ abilities to tures that indicate our mutual offspring will attribute goals and intentions (Carpenter, likely be resistant to parasites”). The capacity Akhtar, & Tomasello, 1998; Meltzoff, 1995; to ascribe intentions and beliefs to the self but see Huang, Heyes, & Charman, 2001). and to others, however, is something alto- By 9 to 12 months of age, human beings gether different from this use of the term seem to be sensitive to the fact that inten- “consciousness” and is a pivotal element that tional agents engage in goal-directed actions, is required for many adaptive human behav- that their behavior is teleological, and that iors to occur (e.g., “I don’t think Kevin is sin- their actions are self-generated. Throughout cere about his intentions to marry my sister, early ontogeny, children’s understanding of P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

620 the cambridge handbook of consciousness

intentionality becomes increasingly enrich- embedded in an electrolyte solution, the ed, so that by the age of 4 years, they are combination of which creates a sensor that able to represent the beliefs and knowledge is designed to react with carbon monoxide states of other individuals and predict and molecules and to sound an auditory alarm explain their behavior on these grounds (see when the gas is present. There is also, of Wellman et al., 2001). Although slight differ- course, the proverbial canary in the coal ences may exist among societies in the rel- mine, whose odd behavior or death in poorly ative rate of acquisition of such social cog- ventilated mining shafts serves to alert work- nitive skills, as well as the emphasis that is ers of dangerous levels of carbon monoxide. placed on the types of mental state attribu- To some extent, self-consciousness is sim- tions that are made, the development of this ilar to this canary in the coal mine, in that “theory of mind” system runs a standard epi- it provides us with conceptual access to a genetic track across human societies (Tardif hidden source of information that has con- & Wellman, 2000; Wellman et al., 2001; see sequences for our genetic fitness. The canary also Lillard, 1998). provides us with access to information (car- To help illustrate, consider a case where, bon monoxide) in the external environment unlike mental states, the ability to natura- (the coal mine), whereas self-consciousness lly detect adaptively relevant information provides us with access to information is an impossibility for human beings. The (proximate causal states) in the internal human sensory system is unable to detect environment (the mind). What is adaptive dangerous levels of carbon monoxide, an is not simply having access to these types of odorless, colorless, tasteless gas that results information, however, but rather how that from the incomplete combustion of hydro- information is translated into the produc- carbon fuels. Carbon monoxide binds with tion of actual behavioral responses. Merely hemoglobin with an affinity about 250 times “being” self-conscious without having func- that of oxygen, interfering with oxygen tra- tional psychological adaptations designed to nsport, delivery, and utilization. At high lev- respond to and control the flow of informa- els of exposure, carbon monoxide can lead to tion dealing with the self’s proximate causal loss of consciousness, coma, and death. states is like ignoring the dead canary in the From an evolutionary perspective, hydro- coal mine. Information is only as useful as the carbon fuels are apparently too recent an psychological adaptations that are designed innovation for human beings to have evolved to harness and exploit it (for complemen- sensory capabilities designed to detect high tary accounts of consciousness, see Damasio, levels of carbon monoxide. As a result, vic- 2002; Frank, 1988). tims receive no obvious sensory warning that The analogy falls short, however, in that dangerous levels are present in the envi- self-consciousness provides human beings ronment. If such information were accessi- with access to information without requiring ble (e.g., through olfaction) and also present an artificial means of detection. The detec- in ancestral conditions, natural selection tion mechanism is naturally entrenched would have likely favored those individuals in the human cognitive system and does who responded to the presence of carbon not rely on metaphors or external devices. monoxide in adaptive fashion. Without such Rather, the same concepts that are used ability, however, no adaptive mechanisms to construct adaptive theories about other associated with such toxic environmental peoples’ behaviors are also the ones rep- conditions have evolved, and contemporary resented through first-order psychological individuals are seriously threatened by this experiences. poisonous gas. Fortunately, human beings have developed fairly effective strategies of detect- The Serpent’s Gift ing high levels of carbon monoxide by artificial means. Electrochemical devices, “But the serpent said to the woman, ‘You will for instance, contain platinum electrodes not die. For God knows that when you eat P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

the serpent’s gift: evolutionary psychology and consciousness 621

of it your eyes will be opened, and you will So consciousness did not provide a moral be like God, knowing good and evil.’ code, but it did provide the ability to see So . . . she took of its fruit and ate; and she things from another’s perspective, and with also gave some to her husband, and he ate. it the knowledge of good and evil, at least as Then the eyes of both were opened, and defined within a particular society. they knew they were naked” (Genesis 3: Consciousness also provided a view of 4–7, Revised Standard Version). the mind of the gods. With consciousness The serpent’s gift was self-knowledge, and comes the need for explanation. We look with it the knowledge of others and the for causes, for intentionality, not only in the ability to acquire a knowledge of right and actions of others but also in the events that wrong. Nature is amoral. We may shud- surround us. If Homer’s behavior is moti- der at the way chimpanzees tear apart a vated by his knowledge or his wishes, might colobus monkey they have caught or feel not the behavior of lightning or rain be revulsion that a male lion kills the cubs sired motivated similarly? There is a reason for by another when it acquires a new mate. everything, our consciousness tells us, and But those emotions are uniquely human, in prescientific days it was the gods or spir- and we do not judge as immoral the chim- its that made things happen – gods or spir- panzee or the lion for acting in its own best its we imbued with human-like motivation genetic interest. With consciousness, how- through our theory of mind. ever, comes social proscriptions. Actions can now be right or wrong, moral or immoral, The Role of Consciousness in Religion: even if they are executed in, what in ancient Spirits, Gods, and Morality times, would have been our own best interest. Stepfathers still murder their stepchil- Our ancestors’ behaviors were mediated not dren at rates many times greater than that only by the social forces that be, as are chim- of biological fathers (see Daly & Wilson, panzees’ behaviors, but also by the assump- 1988), something the male lion might under- tions that people made about what was stand (if he were conscious). But with the appropriate and inappropriate, moral and advent of consciousness this behavior is now immoral, evil and righteous. These beliefs wrong, and because it is not socially sanc- were supported by intuitions about fairness tioned, it is rare in an absolute sense. Theft, and injustice, but were also strongly enforced adultery, assault, and murder still happen by the community’s shared belief in super- with high frequency in all societies (although natural agents who were envisioned as hav- rates differ considerably among societies), ing a vested interest in moral affairs. Some and most societies have proscriptions against of these deontological assumptions, such as them and punishment for transgressors who “one should never steal,” were universal in are “caught in the act.” Although most of us nature because they contravened the fit- believe that there are some universal human ness interests of individual members of any rights, a look at the variety of behaviors that social group. Other deontological assump- are judged as moral and immoral, legal and tions, such as “one should never disobey illegal, in cultures around the world sug- one’s maternal grandfather,” were limited to gests that this may not be so. For exam- individual cultures, because in some soci- ple, although modern Westerners are aghast eties following such orders was adaptive by infanticide, it is expected, under some whereas in other socioecologies it was either circumstances, in some cultures (see Hrdy, maladaptive or not sufficiently adaptive to 1999); the treatment of women varies con- be supported through custom (Reynolds & siderably across cultures and history, each Tanner, 1995). culture believing that its view is the morally Although subscribing to these moral and correct one. (According to Genesis, it was conventional rules led to social harmony in woman who first ate of the tree of knowl- the group, which ultimately subserved the edge, and this has influenced her status in individual interests of in-group members, many cultures over the past two millennia.) people were easily tempted to go astray, P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

622 the cambridge handbook of consciousness

especially when they felt that they were various psychological adaptations. Although not being observed by social others and the works of Trivers on the origins of defense when they could benefit through transgress- mechanisms and that of Burley on the loss ing. Unfortunately, human beings have never of human estrus, for instance, are regarded been especially good at avoiding detection, as highly plausible, such models have yet to and they are prone to overestimating their be validated through empirical means. Com- ability to deceive others. Such errors can parative analyses between human beings and be genetically catastrophic, because the pay- closely related species, such as chimpan- offs for a successful social transgression (e.g., zees, can be extraordinarily informative in stealing a neighbor’s food) are not necessar- this regard. ily worth the risk of getting caught and fac- Although there may be few classes of ing the punitive actions of the other ingroup behavior that are truly unique to human members. beings, there are numerous categories of Under conditions where individuals are natural social behaviors that occur cross- uncertain of the presence of social others and culturally in human beings that are rarely are presented with opportunities to increase observed, or altogether absent, in closely their own genetic fitness at the expense of related species. Suicide is one such case, others, there was likely selective pressure for but feral chimpanzees have also not been a heuristic strategy leading to the inhibition observed to manufacture symbolic artifacts; of the socially proscribed behavior. If indi- to partake in non-functional, group-specific viduals are inclined to represent the pres- ritualistic behavior; to translate their devel- ence of some supernatural agency that has opmentally canalized repertoire of vocal- “privileged” perceptual access to the trans- izations into new strings of communicative gressor’s behaviors and that is also capable meaning unique to certain populations; to of responding to these actions in the form physically care for ill, maimed, or other- of aversive life events, then this might facili- wise importuned conspecifics; to engage in tate the inhibition of the social transgression, juvenile pretend play; to construct mate- therefore promoting an adaptive outcome. rial (i.e., clothing) designed to cover their In fact, Pascal Boyer (1994, 2001; also Atran, anogenital region; to kill or conspire to kill 2002; but see Bering, 2002) has shown that others who possess damaging knowledge; or human cognition is naturally susceptible to to cooperate to solve novel problems. Even supernatural agent concepts because such pointing behavior (for either declarative or concepts violate people’s intuitive ontolog- imperative purposes) and direct teaching of ical assumptions. For example, gods and novel tasks have only rarely been observed in spirits are represented as being essentially chimpanzees (e.g., Boesch, 1991), and even human and as such activate our folk psy- these have been questioned (see Povinelli chology systems (e.g., they can see and hear et al., 2000). and think), but gods and spirits also violate All of these behaviors (and others) are our intuitive assumptions about other agents considered standard fare in human groups, (e.g., they are invisible). Such scholars as whether small hunter-gatherer societies or Boyer (2001) and Atran (2002) have argued large industrialized nation-states. What is that religious concepts gain their entrance important is that each of these categories to mundane cognitive mechanisms through of adaptive behaviors requires the presence such attention-grabbing properties. of a functionally organized intentionality system that not only provides individuals with access to the hidden causes of behav- Concluding Remarks ior but also leads them to engage in adaptive behaviors. We have proposed through- A central challenge for evolutionary appro- out this chapter that consciousness endowed aches to consciousness is devising empirical human beings with information sui generis procedures that address its functional role in in the form of mental states, and that once P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

the serpent’s gift: evolutionary psychology and consciousness 623

consciousness became standardized in hu- Banyas, C. A. (1999). Evolution and phyloge- man cognition, a new suite of adaptive netic history of the frontal lobes. In B. L. behaviors, such as those listed above, evolved Miller & J. L. Cummings (Eds.), The human to satisfy its unique demands. There are frontal lobes: Functions and disorders (pp. 83– 106 no “precursory forms” of such behaviors in ). New York: Guilford Press. other primates because human beings alone Baron-Cohen, S. (1995). Mindblindness: An essay faced the adaptive challenges of conscious- on autism and theory of mind. Cambridge, MA: ness that led to their regular appearance in MIT Press. 1990 the species. Baumeister, R. F. ( ). Suicide as escape from self. Psychological Review, 97, 90–113. Becker, E. (1974). The denial of death. New York: Note The Free Press. Bering, J. M. (2002). The existential theory of 3 24 1. Paleontologists refer to chimpanzees as a “con- mind. Review of General Psychology, 6, – . servative species,” suggesting that they have Bering, J. M. (2004a). A critical review of changed relatively little over the past 7 million the “enculturation hypothesis”: The effects of years, making them a reasonably good model human rearing on great ape social cognition. for what our common ancestor may have been Animal Cognition, 7, 201–212. like. Bering, J. M. (2004b). Consciousness was a trouble-maker: On the general maladaptive- ness of unsupported higher-order mental rep- References resentation. Journal of Mind and Behavior, 25, 33–56. Aiello, L. C., & Wheeler, P. (1995). The Bering, J. M., Bjorklund, D. F., & Ragan, P. expensive-tissue hypothesis: The brain and the (2000). Deferred imitation of object-related digestive system in human and primate evolu- actions in human-reared juvenile chimpanzees tion. Current Anthropology, 36, 199–221. and orangutans. Developmental Psychobiology, Alexander, R. D. (1989). Evolution of the human 36, 218–232. psyche. In P. Mellers & C. Stringer (Eds.), The Bering, J. M., & Povinelli, D. J. (2003). Com- human revolution: Behavioural and biological paring cognitive development. In D. Maestrip- perspectives on the origins of modern humans (pp. ieri (Ed.), Primate psychology: Bridging the gap 455–513). Princeton, NJ: Princeton University between the mind and behavior of human and Press. nonhuman primates (pp. 205–233). Cambridge, Andrade, M. C. B. (1996). Sexual selection for MA: Harvard University Press. male sacriﬁce in the Australian redback spider. Bjorklund, D. F., & Bering, J. M. (2000). The Science, 271, 70–72. evolved child: Applying evolutionary develop- Atran, S. (2002). In gods we trust: The evolution- mental psychology to modern schooling. Learn- ary landscape of religion. Oxford: Oxford Uni- ing and Individual Differences, 12, 347–373. versity Press. Bjorklund, D. F., & Bering, J. M. (2003a). Baillargeon, R. (1994). How do infants learn Big brains, slow development, and social about the physical world? Current Directions complexity: The developmental and evolution- in Psychological Science, 3, 133–140. ary origins of social cognition. In M. Brune,¨ H. Baldwin, D. A. (1991). Infants’ contribution to the Ribbert, & W. Schiefenhovel˜ (Eds.). The social achievement of joint reference. Child Develop- brain: Evolutionary aspects of development and ment, 62, 875–890. pathology (pp. 113–151). New York: Wiley. Baldwin, D. A. (1993). Early referential under- Bjorklund, D. F., & Bering, J. M. (2003b). The standing: Infants’ ability to recognize refer- development of deferred imitation in juvenile ential acts for what they are. Developmental chimpanzees. Developmental Review, 23, 389– Psychology, 29, 832–843. 412. Barkow, J. H., Cosmides, L., & Tooby, J. (Eds.). Bjorklund, D. F., & Harnishfeger, K. K. (1995). (1992). The adapted mind: Evolutionary psychol- The role of inhibition mechanisms in the evo- ogy and the generation of culture. New York: lution of human cognition and behavior. In Oxford University Press. F. N. Dempster & C. J. Brainerd (Eds.), New P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

624 the cambridge handbook of consciousness

perspectives on interference and inhibition in Brown, R. M., Dehlen, E., Mils, C., Rick, J., cognition (pp. 141–173). New York: Academic & Biblarz, A. (1999). Evaluation of an evo- Press. lutionary model of self-preservation and self- Bjorklund, D. F., & Kipp, K. (2001). Social cogni- destruction. Suicide & Life-Threatening Behav- tion, inhibition, and theory of mind: The evo- ior, 29, 58–71. lution of human intelligence. In R. J. Sternberg Burley, N. (1979). The evolution of concealed & J. C. Kaufman (Eds.), The evolution of intelli- ovulation. American Naturalist, 114, 835–838. gence (pp. 27–53). Mahwah, NJ: Erlbaum. Buss, D. M. (1995). Evolutionary psychology. Psy- Bjorklund, D. F., & Pellegrini, A. D. (2002). The chological Inquiry, 6, 1–30. origins of human nature: Evolutionary develop- Buss, D. M. (1999). Evolutionary psychology: The mental psychology. Washington, DC: American new science of the mind. Boston: Allyn and Psychological Association. Bacon. Bjorklund, D. F., & Rosenberg, J. S. (2005). The Buss, D. M. (Ed.) (2005). Evolutionary psychology role of developmental plasticity in the evolu- handbook. New York: Wiley. tion of human cognition: Evidence from encul- Buss, D. M., Haselton, M. G., Shackelford, T. K., turated, juvenile great apes. In B. Ellis & D. F. Bleske, A. L., & Wakeﬁeld, J. C. (1998). Adap- Bjorklund (Eds.), Origins of the social mind: tations, exaptations, and spandrels. American Evolutionary psychology and child development Psychologist, 53, 533–548. (pp. 45–75). New York: Guilford. Bygott, J. D. (1979). Agonistic behavior, domi- Bjorklund, D. F., Yunger, J. L., Bering, J. M., nance, and social structure in wild chimpanzees 2002 & Ragan, P. ( ). The generalization of of the Gombe National Park. In D. A. Hamburg deferred imitation in enculturated chim- & E. R. McCown (Eds.), The great apes: Perspec- panzees (Pan troglodytes). Animal Cognition, 5, tives on human evolution (Vol. 5,pp.307–317). 49 58 – . Menlo Park: Benjamin/Cummings. 1998 Bloom, P. ( ). Some issues in the evolution Byrne, R., & Whiten, A. (Eds.) (1988). Machi- of language and thought. In D. Cummins & C. avellian intelligence: Social expertise and the evo- 204 Allen (Eds.), The evolution of mind (pp. – lution of intellect in monkeys, apes, and humans. 223 ). New York: Oxford University Press. Oxford: Clarendon. 1991 Boesch, C. ( ). Teaching among wild chim- Call, J., & Carpenter, M. (2003) On imitation in 530 532 panzees. Animal Behavior, 41, – . apes and children. Infancia y Aprendizaje, 26, Boesch, C. (1993). Toward a new image of culture 325–349. in chimpanzees. Behavioral and Brain Sciences, Call, J., & Tomasello, M. (1999). A nonverbal false 514 515 16, – . belief task: The performance of children and Bogin, B. (1999). Patterns of human growth (2nd great apes. Child Development, 70, 381–395. ed.). Cambridge: Cambridge University Press. Carpenter, M., Akhtar, N., & Tomasello, M. Bonner, J. T. (1988). The evolution of complexity (1998). Fourteen- through 18-month-old by means of natural selection. Princeton, NJ: infants differentially imitate intentional and Princeton University Press. accidental actions. Infant Behavior & Develop- Boyer, P. (1994). The naturalness of religious ideas: ment, 21, 315–330. A cognitive theory of religion. Berkeley: Univer- Cartwright, J. (2000). Evolution and human sity of California Press. behavior: Darwinian perspectives on human Boyer, P. (2001). Religion explained: The evolution- nature. Cambridge, MA: MIT Press. ary origins of religious thought. New York: Basic Chalmers, D. J. (1996). The conscious mind: In Books. search of a fundamental theory. New York: Bremner, J. D., Narayan, M., Staib, L. H., South- Oxford University Press. wick, S. M., McGlashan, T., & Charney, D. S. Cosmides, L., & Tooby, J. (1987). From evolu- (1999). Neural correlates of childhood sexual tion to behavior: Evolutionary psychology as abuse in women with and without posttrau- the missing link. In J. Dupre (Ed.), The latest matic stress disorder. American Journal of Psy- on the best essays on evolution and optimality chiatry, 156, 1787–1795. (pp. 277–306). Cambridge, MA: MIT Press. Brooks-Gunn, J., & Lewis, M. (1984). The devel- Cotterill, R. M. J. (2001). Evolution, cognition opment of early self-recognition. Developmen- and consciousness. Journal of Consciousness tal Review, 4, 215–239. Studies, 8, 3–17. P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

the serpent’s gift: evolutionary psychology and consciousness 625

Crook, J. M. (1980). The evolution of human con- Dunbar, R. I. M. (1995). Neocortex size and sciousness. Oxford: Clarendon Press. group size in primates: A test of the hypoth- Daly, M., & Wilson, M. (1988). Homicide. New esis. Journal of Human Evolution, 28, 287–296. York: Aldine. Dunbar, R. I. M. (2001). Brains on two legs: Damasio, A. R. (2002). A note on the neurobi- Group size and the evolution of intelligence. ology of emotions. In S. G. Post, L. G. Under- In F. B. M. de Waal (Ed.), Tree of origins: What wood, J. P. Schloss, & W. B. Hurlbut (Eds.), primate behavior can tell us about human social Altruism and altruistic love: Science, philosophy, evolution (pp. 173–191). Cambridge, MA: Har- and religion in dialogue (pp. 264–271). London: vard University Press. Oxford University Press. Ekman, P., & Friesen, W. V. (1975). Unmasking de Catanazaro, D. (1986). A mathematical the face. Englewood Cliffs, NJ: Prentice-Hall. model of evolutionary pressures regulating self- Fabrega,´ H. (2002). Origins of psychopathology: preservation and self-destruction. Suicide & The phylogenetic and cultural basis of mental ill- Life-Threatening Behavior, 16, 166–181. ness. New Brunswick, NJ: Rutgers University de Catanzaro, D. (1987). Evolutionary pres- Press. sures and limitations to self-preservation. In C. Finlay, B. L., & Darlington, R. D. (1995, June Crawford, M. Smith, & D. Krebs (Eds.), Socio- 16). Linked regularities in the development and biology and psychology (pp. 311–333). Hillsdale, evolution of mammalian brains. Science, 268, NJ: Erlbaum. 1579–1584. de Catanzaro, D. (1991). Evolutionary limits to Finlay, B. L., Darlington, R. B. , & Nicastro, N. self-preservation. Ethology & Sociobiology, 12, (2001) Developmental structure in brain evo- 13–28. lution. Behavioral and Brain Sciences 24, 263– de Catanzaro, D. (1992). Prediction of self- 308. preservation failures on the basis of quantita- Fouts, R. (1997). Next of kin: What chimpanzees tive evolutionary biology. In R. W. Maris, A. L. have taught me about who we are. New York: Berman, J. T. Maltsberger, & R. I. Yufit (Eds.), William Morrow. Assessment and prediction of suicide (pp. 607– Franco, F., & Butterworth, G. (1996). Pointing and 621). New York: Guilford Press. social awareness: Declaring and requesting in Dennett, D. (1987). The intentional stance. Cam- the second year. Journal of Child Language, 23, bridge, MA: Bradford Books/MIT Press. 307–336. de Waal, F. B. M. (1982). Chimpanzee politics: Frank, R. H. (1988). Passions with reason: The Power and sex among apes. London: Jonathan strategic role of the emotions. New York: Norton. Cape. French, J. A., Kamil, A. C., & Ledger, D. W. de Waal, F. B. M. (1986). The integration of domi- (2001). Evolutionary psychology and motivation. nance and social bonding in primates. Quarterly Vol. 47 of the Nebraska Symposium on Moti- Review of Biology, 61, 459–479. vation. Lincoln, NE: University of Nebraska de Waal, F. B. M. (2002). Evolutionary psycho- Press. logy: The wheat and the chaff. Current Direc- Gallup, G. G. Jr. (1979). Self-recognition in chim- tions in Psychological Science, 11, 187–191. panzees and man: A developmental and compar- Diamond, J. M. (1992). The third chimpanzee: The ative perspective. New York: Plenum Press. evolution and future of the human animal. New Gallup, G. G. Jr. (1982). Self-awareness and the York: Harper Collins. emergence of mind in primates. American Jour- Donald, M. (2000). The central role of culture in nal of Primatology, 2 , 237–248. cognitive evolution: A reflection on the myth of Gallup, G. G. Jr. (1985). Do minds exist in species the “isolated mind.” In L. P. Nucci, G. B. Saxe, other than our own? Neuroscience & Biobehav- & E. Turiel (Eds.), Culture, thought, and devel- ioral Reviews, 9, 631–641. opment (pp. 19–38). Mahwah, NJ: Erlbaum. Gaulin, S. J. C., & McBurney, D. H. (2001). Dunbar, R. I. M. (1992). Neocortex size as a con- Psychology: An evolutionary approach. Upper straint on group size in primates. Journal of Saddle River, NJ: Prentice Hall. Human Evolution, 20, 469–493. Geary, D. C. (1995). Reflections of evolution and Dunbar, R. I. M. (1993). Coevolution of neocor- culture in children’s cognition: Implications tical size, group size and language in humans. for mathematical development and instruc- Behavioral and Brain Sciences, 16, 681–735. tion. American Psychologist, 50, 24–37. P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

626 the cambridge handbook of consciousness

Geary, D. C. (2001). Sexual selection and Greenfield, P., Maynard, A., Boehm, C., & sex differences in social cognition. In A. V. Schmidtling, E. Y. (2000). Cultural appren- McGillicuddy-DeLisi & R. DeLisi (Eds.), Biol- ticeship and cultural change: Tool learning and ogy, society, and behavior: The development of sex imitation in chimpanzees and humans. In S. T. differences in cognition (pp. 23–53). Greenwich, Parker, J. Langer, & M. L. McKinney (Eds.), CT: Ablex. Biology, brains, and behavior: The evolution of Geary, D. C. (2004). The origin of mind: Evo- human development (pp. 237–277). Santa Fe, lution of brain, cognition, and general intelli- NM: School of American Research Press. gence. Washington, DC: American Psychologi- Hamilton, W. D. (1964). The genetical theory of cal Association. social behavior. Journal of Theoretical Biology, 7, Geary, D. C., & Flinn, M. V. (2001). Evolution of 1–52. human parental behavior and the human fam- Hare, B., Call, J., Agentta, B., & Tomasello, M. ily. Parenting: Science and Practice, 1, 5–61. (2000). Chimpanzees know what conspecifics Geary, D. C., & Huffman, K. (2002). Brain and do and do not see. Animal Behaviour, 59, 771– cognitive evolution: Forms of modularity and 785. functions of mind. Psychological Bulletin, 128, Hare, B., Call, J., & Tomasello, M. (2001). Do 667–698. chimpanzees know what conspecifics know? Gelman, R., & Williams, E. M. (1998). Enabling Animal Behaviour, 61, 139–151. constraints for cognitive development and Haslam, N. (1997). Four grammars for primate learning: Domain-specificity and epigenesis. In social relations. In J. A. Simpson & D. T.Kenrick D. Kuhn & R. S. Siegler (Eds.), Cognition, per- (Eds.), Evolutionary social psychology (pp. 297– ception, and language (Vol. 2,pp.575–630). 316). Mahwah, NJ: Erlbaum. New York: Wiley. Holmes, J. C., & Bethel, W. M. (1972). Modifi- Gilbert, P. (1998). What is shame? Some core cation of intermediate host behavior by para- issues and controversies. In P. Gilbert & B. sites. In E. U. Canning & C. A. Wrighs (Eds.), Andrews (Eds.), Shame: Interpersonal behavior, Behavioural aspects of parasite transmission psychopathology, and culture. Series in affective (pp. 123–149). London: Academic Press. science (pp. 3–38). New York: Oxford Univer- Hrdy, S. B. (1999). Mother Nature: A history sity Press. of mothers, infants, and natural selection. New Gilbertson, M. W., Shenton, M. E., Ciszewski, A., York: Pantheon Books. Kasai, K., Lasko, N. B., Orr, S. B., & Pittman, Huang, C.-T., Heyes, C., & Charman, T. (2002). R. A. (2002). Smaller hippocampal volume Infants’ behavioral reenactment of “failed predicts pathologic vulnerability to psycholog- attempts”: Exploring the roles of emulation ical trauma. Nature Neuroscience, 5, 1242–1247. learning, stimulus enhancement, and under- Goodall, J. (1986). The chimpanzees of Gombe. standing of intentions. Developmental Psychol- Cambridge, MA: Belknap. ogy, 38, 840–855. Gopnik, A. (2000). Explanation as orgasm and Humphrey, N. K. (1976). The social function of the drive for causal knowledge: The function, intellect. In P. P. G. Bateson & R. A. Hinde evolution, and phenomenology of the theory (Eds.), Growing points in ethology (pp. 303– formation system. In F. C. Keil & R. A. Wil- 317). Cambridge: Cambridge University Press. son (Eds.), Explanation and cognition (pp. 299– Humphrey, N. K. (1992). A history of the mind: 323). Cambridge, MA: MIT Press. Evolution and the birth of consciousness. New Gopnik, A., & Meltzoff, A. N. (1997). Words, York: Simon & Schuster. thoughts, and theories. Cambridge, MA: MIT Humphrey, N. K. (2000). How to solve the mind- Press. body problem. Journal of Consciousness Studies, Gould, S. J. (1977). Ontogeny and phylogeny. 7, 5–20. Cambridge, MA: Harvard University Press. Itakura, S. (1996). Manual action in infant Gould, S. J. (1991). Exaptation: A crucial tool chimpanzees: A preliminary study. Perceptual for evolutionary psychology. Journal of Social and Motor Skills, 83, 611–614. Issues, 47, 43–65. Jeannerod, M. (1999). To act or not to act: Grammer, K. (1993). 5-!a-androst-16en-3!a-on: A Perspectives on the representation of actions. male pheromone? A brief report. Ethology & Quarterly Journal of Experimental Psychology, Sociobiology, 14, 201–207. 52, 1–29. P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

the serpent’s gift: evolutionary psychology and consciousness 627

Jerrison, H. J. (1973). Evolution of the brain and McGuire, M. T., & Troisi, A. (1998). Darwinian intelligence. New York: Academic Press. psychiatry. New York: Oxford University Press. Joffe, T. H. (1997). Social pressures have selected Meltzoff, A. N. (1995). Understanding the inten- for an extended juvenile period in primates. tions of others: Re-enactment of intended acts Journal of Human Evolution, 32, 593–605. by 18-month-old children. Developmental Psy- Jolly, A. (1991). Conscious chimpanzees? A chology, 31, 838–850. review of recent literature. In C. A. Ristau Mithen, S. (1996). The prehistory of the mind: (Ed.), Cognitive ethology: The minds of other ani- The cognitive origins of art, religion and science. mals. Essays in honor of Donald R. Grifﬁn: Com- London: Thames and Hudson. parative cognition and neuroscience (pp. 231– Nesse, R. M., & Lloyd, A. T. (1992). The evo- 252). Hillsdale, NJ: Erlbaum. lution of psychodynamic mechanisms. In J. Kaplan, H., Hill, K., Lancaster, J., & Hurtado, H. Barkow, L. Cosmides, & J. Tooby (Eds.), A. M. (2000). A theory of human life history The adapted mind (pp. 601–624). New York: evolution: Diet intelligence, and longevity. Evo- Oxford University Press. lutionary Anthropology, 9, 156–185. O’Connor, R. J. (1978). Brood reduction in birds: Karmiloff-Smith, A. (1992). Beyond modularity: Selection for fratricide, infanticide and suicide. A developmental perspective on cognitive science. Animal Behaviour, 26, 79–96. Cambridge, MA: MIT Press. Palmer, J. A., & Palmer, L. K. (2002). Evolutionary Kierkegaard, S. (1946). A Kierkegaard anthology psychology: The ultimate origins of human behav- (R. Bretall, Ed.). New York: Modern Library. ior. Needham Heights, MA: Allyn & Bacon. (Work originally published 1849) Parker, S. T., Mitchell, R. W., & Boccia, M. L. Kummer, H. (1971). Primate societies: Group tech- (2006). Self-awareness in animals and humans: niques of ecological adaptation. Chicago: Aldine. Developmental perspectives. Cambridge, MA: Lester, D. (1997). The role of shame in suicide. Cambridge University Press. Suicide & Life Threatening Behavior, 27, 352– Pinker, S. (1994). The language instinct: How the 361. mind creates language. New York: Morrow. Lickliter, R., & Honeycutt, H. (2003). Devel- Pinker, S. (1997). How the mind works. New York: opmental dynamics: Towards a biologically Norton. plausible evolutionary psychology. Psychologi- Pinker, S. (2002). The blank slate: The modern cal Bulletin, 129, 819–835. denial of human nature. New York: Viking Lillard, A. S. (1998). Ethnopsychologies: Cultural Press. variations in theory of mind. Psychological Bul- Plooij, F. X. (1978). Some basic traits of language letin, 123, 3–32. in wild chimpanzees? In A. Lock (Ed.), Action, Luria, A. R. (1973). The working brain: An intro- gesture and symbol (pp. 111–131). London: Aca- duction to neuropsychology. New York: Basic demic Press. Books. Poulin, R. (1992). Altered behaviour in para- Mayr, E. (2001). What evolution is. New York: sitized bumblebees: Parasite manipulation or Basic Books. adaptive suicide? Animal Behaviour, 44, 174– McAllister, M. K., & Roitberg, B. D. (1987). 176. Adaptive suicidal behaviour in pea aphids. Povinelli, D. J. (2000). Folk physics for apes. New Nature, 328, 797–799. York: Oxford University Press. McAllister, M. K., Roitberg, B. D., & Weldon, Povinelli, D. J., Bering, J. M. (2002). The mental- K. L. (1990). Adaptive suicide in pea aphids: ity of apes revisited. Current Directions in Psy- Decisions are cost sensitive. Animal Behaviour, chological Science, 11, 115–119. 40, 167–175. Povinelli, D. J., Bering, J. M., & Giambrone, McCauley, R. N. (2000). The naturalness of reli- S. (2000). Toward a science of other minds: gion and the unnaturalness of science. In F. C. Escaping the argument by analogy. Cognitive Keil & R. A. Wilson (Eds.), Explanation and Science, 24, 509–541. cognition (pp. 61–85). Cambridge: Cambridge Povinelli, D. J., Bering, J., & Giambrone, S. University Press. (2003). Chimpanzee ‘pointing’: Another error McGrew, W. C. (1992). Chimpanzee mate- of the argument by analogy? In S. Kita (Ed.), rial culture: Implication for human evolution. Pointing: Where language, culture, and cognition Cambridge: Cambridge University Press. meet (pp. 35–68). Hillsdale, NJ: Erlbaum. P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

628 the cambridge handbook of consciousness

Povinelli, D. J., & Cant, J. G. H. (1995). Arbo- Searle, J. R. (2000). Consciousness. Annual real clambering and the evolution of self- Review of Neuroscience, 23, 557–578. conception. Quarterly Review of Biology, 70, Spelke, E. S., Breinlinger, K., Macomber, J., & 393–421. Jacobson, K. (1992). Origins of knowledge. Psy- Povinelli, D. J., Davis, D. R. (1994). Differences chological Review, 99, 605–632. between chimpanzees (Pan troglodytes) and Stout, D, (2002). Skill and cognition in stone tool humans (Homo sapiens) in the resting state production: An ethnographic case study from of the index finger: Implications for pointing. Irian Jaya. Current Anthropology, 43, 693–722. 134 Journal of Comparative Psychology, 108, – Suddendorf, T., & Whiten, A. (2001). Mental 139 . evolution and development: Evidence for sec- Povinelli, D. J., & Eddy, T. J. (1996a). What young ondary representation in children, great apes chimpanzees know about seeing. Monograph of and other animals. Psychological Bulletin, 127, the Society for Research in Child Development, 629–650. 3 247 61 ( ), no. . Swartz, K. B., Sarauw, D., & Evans, S. (1999). Povinelli, D. J., & Eddy, T. J. (1996b). Fac- Cognitive aspects of mirror self-recognition in tors influencing chimpanzees’ (Pan troglodytes) great apes. In S. T. Parker, R. W. Mitchell, & H. recognition of attention. Journal of Compara- L. Miles (Eds.), The mentalities of gorillas and tive Psychology, 110, 336–345. orangutans: Comparative perspectives (pp. 283– Povinelli, D. J., & Eddy, T. J. (1996c). Chim- 294). Cambridge: Cambridge University Press. panzees: Joint visual attention. Psychological Tangney, J. P. (2001). Self-conscious emotions: Science, 7, 129–135. The self as a moral guide. In A. Tesser, D. Povinelli, D. J., & O’Neill, D. K. (2000). Do A. Stapel, & J. V. Wood (Eds.), Self and chimpanzees use their gestures to instruct each motivation: Emerging psychological perspectives other? In S. Baron-Cohen, H. Tager-Flusberg, (pp. 97–117). Washington, DC: American Psy- & D. J. Cohen (Eds.), Understanding other chological Association. minds (pp. 459–487). Oxford: Oxford Univer- Tardif, T., & Wellman, H. M. (2000). Acqui- sity Press. sition of mental state language in Mandarin- Povinelli, D. J., Perilloux, H. K., Reaux, J. E., and Cantonese-speaking children. Developmen- & Bierschwale, D. T. (1998). Young and juve- tal Psychology, 36, 25–43. nile chimpanzees’ (Pan troglodytes) reactions to Tattersall, I. (1998). Becoming human: Evolution intentional versus accidental and inadvertent and human uniqueness. New York: Harcourt actions. Behavioural Processes, 42, 205–218. Brace. Preuss, T. M. (2001). The discovery of cerebral Thornhill, R., & Palmer, C. T. (2000). A natural diversity: An unwelcome scientific revolution. history of rape: Biological bases of sexual coercion. In D. Falk & K. Gibson (Eds.), Evolutionary Cambridge, MA: MIT Press. anatomy of the primate cerebral cortex (pp. 138– Tobias, P. V. (1987). The brain of Homo habilis:A 164). Cambridge: Cambridge University Press. new level of organization in cerebral evolution. Reaux, J. E., Theall, L. A., & Povinelli, D. J. Journal of Human Evolution, 16, 741–761 (1999). A longitudinal investigation of chim- Tomasello, M. (1996). Do apes ape? In C. Heyes panzee’s understanding of visual perception. & B. Galef (Eds.), Social learning in animals: Child Development, 70, 275–290. The role of culture (pp. 319–346). San Diego: Reiss, D., & Marino, L. (2001). Mirror self- Academic Press. recognition in the bottlenose dolphin: A case Tomasello, M. (2000). Culture and cognitive of cognitive convergence. Proceedings of the development. Current Directions in Psycholog- National Academy of Sciences USA, 98, 5937– ical Science, 9, 37–40. 5942. Tomasello, M., & Bates, E. (2001). Language Reynolds, V., & Tanner, R. (1995). The social ecol- development: The essential readings. Oxford: ogy of religion. New York: Oxford University Blackwell. Press. Tomasello, M., & Call, J. (1997). Primate cognition. Rilling, J. K., & Insel, T. R. (1999). The primate Oxford: Oxford University Press. neocortex in comparative perspective using Tomasello, M., Kruger, A. C., & Ratner, H. H. magnetic resonance imaging. Journal of Human (1993). Cultural learning. Behavioral and Brain Evolution, 37, 191–223. Sciences, 16, 495–511. P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

the serpent’s gift: evolutionary psychology and consciousness 629

Tomasello, M., Savage-Rumbaugh, S., & Kruger, Wellman, H. M. (1990). The child’s theory of mind. A. C. (1993). Imitative learning of actions on Cambridge, MA: MIT Press. objects by children, chimpanzees, and encul- Wellman, H. M., Cross, D., & Watson, J. (2001). turated chimpanzees. Child Development, 64, Meta-analysis of theory-of-mind development: 1688–1705. The truth about false belief. Child Develop- Tooby, J., & Cosmides, L. (1992). The psycho- ment, 72, 655–684. logical foundations of culture. In J. H. Barkow, Whiten, A. (1998). Imitation of sequential L. Cosmides, & J. Tooby (Eds.), The adapted structure of actions by chimpanzees (Pan mind: Evolutionary psychology and the genera- troglodytes). Journal of Comparative Psychology, 19 139 tion of culture (pp. – ). New York: Oxford 112, 270–281. University Press. Whiten, A., & Byrne, R. W. (1988). The manip- 2003 Tooby, J., Cosmides, L., & Barrett, H. C. ( ). ulation of attention in primate tactical decep- The second law of thermodynamics is the ﬁrst tion. In R. W.Byrne & A. Whiten (Eds.). Machi- law of psychology: Evolutionary developmen- avellian intelligence: Social expertise and the evo- tal psychology and the theory of tandem, coor- lution of intellect in monkeys, apes, and humans dinated inheritances: Comment on Lickliter (pp. 211–223). Oxford: Clarendon Press. and Honeycutt (2003). Psychological Bulletin, Whiten, A., Custance, D. M., Gomez,¨ J. C., 129, 858–865. Teixidor, P., & Bard, K. A. (1996). Imitative 1981 Trivers, R. ( ). Sociobiology and politics. In learning of artiﬁcial fruit processing in chil- E. White (Ed.), Sociobiology and human politics dren (Homo sapiens) and chimpanzees (Pan 1 43 (pp. – ). Lexington, MA: D. C. Health. troglodytes). Journal of Comparative Psychology, Trivers, R. (1985). Social evolution. Reading, MA: 110, 3–14. Benjamin/Cummings. Whiten, A., Goodall, J., McGrew, W.C., Nishida, Tulving, E. (1985). How many memory systems T., Reynolds, V., Sugiyama, Y., Tutin, C. E. G., are there? American Psychologist, 40, 385–398. Wrangham, R. W., & Boesch, C. (1999). Cul- Vygotsky, L. S. (1962). Thought and language. tures in chimpanzees. Nature 399, (Jun), 682– Cambridge, MA: MIT Press. 685. P1: KAE 0521857430c22 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 0:50

630 P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

CHAPTER 23 Anthropology of Consciousness

C. Jason Throop and Charles D. Laughlin

From the standpoint of human evolution, then, vision of an anthropology of conscious- both a social matrix of conduct and the expansion ness apparent in the writings of some of the cortex are among the necessary conditions of the discipline’s forefathers – including for the emergence of a human mind or a human Adolph Bastian, Franz Boas, and Emile personality structure. Just as bodily evolution and Durkheim, among others – before turn- mental evolution cannot be separated, neither can ing to examine the history of thought psychological structuralization and the social evo- among anthropologists living during vari- lution of mankind. To behave humanly as an adult the individual must become psychologically ous eras up to and including the present. organized in a socialization process. His biologi- The development of consciousness studies cal equipment is only one of the conditions neces- is tracked through mid- to late-20th cen- sary for this. Social or sensory isolation is a fatal tury structuralism, practice theory, neuroan- handicap. Hence, it seems reasonable to suppose thropology, and the symbolic-interpretative that the emergence of culture as a prime attribute approaches. These developments are then of human societies must be somehow connected shown to have culminated in contem- with a novel psychological structure rooted in the porary consciousness-related approaches, social behavior of the gregarious primate that gave including cognitive anthropology, psycho- rise to man. It is at this point that organic evo- cultural anthropology, cultural psychol- lution, behavioral evolution, and the old problem ogy, phenomenological anthropology, the of mental evolution come to a common focus. Irving Hallowell, Culture anthropology of the senses, and cultural neu- and Experience (1955) rophenomenology. Among the issues discussed are: the structures and vicissitudes of human experience; the extent to which Abstract culture inﬂuences perception, emotion, categorization, etc.; the biocultural constraints In this chapter we discuss the history upon development as manifest in differ- of anthropological studies of consciousness ent cultures; the relationship between the since the mid-19th century. We cover the evolution of the human nervous system 631 P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

632 the cambridge handbook of consciousness

and subjective experience; and how differ- In this chapter we discuss the history ent societies encourage, evoke, and interpret of anthropological studies of consciousness alternative states of consciousness. since the mid-19th century. We cover the vision of an anthropology of consciousness apparent in the writings of some of Introduction the discipline’s forefathers, before turning to examine the history of thought Anthropology is the study of humanity among anthropologists living during vari- from the broadest possible scope across ous eras up to and including the present. time and space. Anthropologists study our Most of our coverage is of American species, Homo sapiens, throughout its mil- anthropologists who, more than any of the lions of years of evolution to the present and other schools of anthropology, considered beyond. Anthropological research investi- consciousness-related issues to be impor- gates the variety of sociocultural forms from tant. We show how these early practical a cross-cultural perspective, utilizing data and theoretical ruminations centered on from hundreds of the estimated 4,000-plus questions of the evolution of human con- societies living either during the present or sciousness, the putative differences between in the not-so-distant past. Hence the range primitive versus civilized mentalities, and of perspectives within the discipline encom- the role of the group in establishing col- passes both biological and cultural con- lective states of consciousness. We track cerns. Moreover, the anthropological study the development of consciousness stud- of living peoples tends to be naturalistic – ies through mid- to late-20th century that is, anthropologists live with people structuralism, practice theory, neuroan- for extended periods of time, observing thropology, and the symbolic-interpretative and learning from them in the context approaches. We show how these develop- of everyday interaction and, with but rare ments have culminated in contemporary exceptions, seldom rely upon experimen- consciousness-related approaches, including tation. Rather, anthropologists are inter- cognitive anthropology, cultural psychol- ested in recording the way of life of ogy, phenomenological anthropology, the a people going about their normal daily anthropology of the senses, and cultural activities. neurophenomenology. We conclude by list- Examining consciousness through the ing some of the major findings pertaining broad lens of anthropology offers at least to consciousness provided by anthropolog- three major advantages to consciousness ical research and understanding to date. We studies. First, by combining evidence from believe the most crucial of these findings both the evolutionary/biological and the cul- is that anthropological studies of conscious- tural points of view, we are able to elim- ness have often set out to problematize the- inate the pernicious effects of mind-body oretical orientations that proffer accounts dualism that infest much of social scientific based upon unidirectional flows of causa- thinking. Second, by maximizing our pic- tion either from the external organization of ture of the range of states and contents of socioeconomically determined activity pat- consciousness occurring cross-culturally, we terns and cultural symbols to the formation are able to counter our own ethnocentrism – of contents and structures of consciousness that is, our natural tendency to experience or from pregiven structures of consciousness and think about things from our own cul- to those social and cultural forms that are tural point of view. And third, by consider- held to be their material externalizations. ing consciousness within its natural and local In the first section of this chapter we context, we are better able to evaluate how chose to dwell upon a few leading anthro- states and contents of consciousness dynam- pological thinkers in both theory and meth- ically relate to social interaction and envi- ods. This focus is easily achieved, for the ronmental adaptation. number of practicing anthropologists was P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

anthropology of consciousness 633

relatively small during the 19th and first languages would have no words that neatly half of the 20th centuries, and most field- gloss with the English term. Even among workers followed the theories and methods Indo-European languages the term does not of their mentors. Margaret Mead used to translate perfectly. say that in pre-World War II meetings of For instance, the Spanish word conscien- the American Anthropological Association cia connotes both awareness and what we everyone would fit into two buses when they would call “conscience” or “social conscious- would go on excursions, and had the buses ness,” whereas in German there are at least fallen off a bridge, there would have been no two words, Bewusstsein and Erkenntnis, that American anthropology left. But the disci- cover the semantic field covered by the pline has grown rapidly during the last half- English term. And depending upon how we century, and theoretical and methodologi- define consciousness, many of the world’s cal issues have become much more diverse. non-Indo-European languages would either For this reason, the latter portions of the have no word at all or a term or terms that chapter focus less upon influential individual would have to be transposed to make them thinkers and more upon general approaches fit. For example, it is far easier to find a to the study of consciousness. gloss for awareness or of the sensory or moral aspects of consciousness than it would be to find an exact gloss for, say, intentionality (see What Do We Mean by Consciousness? Duranti 1993, 2006), which in both Western Before we proceed any further, however, it philosophy and science is such a critical com- is necessary to examine just what the term ponent – or for consciousness defined either “consciousness” means. For Western scholars as a complex system of psychophysical func- without the broader cross-cultural perspec- tions (e.g., Baars, 1997) or as implying self- tive embraced in anthropology, this might be awareness as a necessary condition (Donald, deemed to be less of a problem (but see Hunt 1991). 1995), and accordingly, cognitive scientists It is fair to say then that most anthro- and philosophers today generally define con- pological research pertaining to conscious- sciousness in terms of such concepts as inten- ness focuses upon those contents and prop- tionality, qualia, and the like. The central erties of consciousness commonly encoun- problem for anthropologists is, however, that tered in the field – including such aspects such definitions are extremely ethnocentric, as sensation (Classen, 1997; Howes, 1991; and very few non-Western cultures would Stoller, 1989), time perception (Adam, 1994, view the matter in the way that Western Gell, 1992; Munn, 1992), perception of space consciousness researchers might conceive of (Pinxton, Van Dooren, & Harvey, 1983), it. From the anthropological perspective, the emotion (Hinton, 1999; Lutz & White, 1986, term “consciousness” presents particular dif- cognition (D’Andrade, 1995; Geertz, 1983), ficulties as an empirical term in science. This apperception (Hallowell, 1955), memory is because it is difficult to operationalize, (Antze & Lambek, 1996; Garro 2000, 2001) perhaps especially in cross-cultural research and symbolization and meaning (Foster, (see Laughlin, McManus, & d’Aquili, 1990, 1994; Foster & Botscharow, 1990), as well pp. 76–82; Peacock, 1975,p.6; Winkelman as questions of picture and illusion percep- 2000,pp.9–15 for the problems con- tion (Segall, Campbell, & Herscovits, 1966), fronted in operationalizing consciousness reason (Garro, 1998; Hamill, 1990; Tambiah, cross-culturally). Few anthropologists before 1990) creative imagination (Dissanayake, the late 20th century underwent fieldwork 1992), sense of the self (Heelas & Lock, intending to research consciousness per se. 1981; Morris, 1994; White & Kirkpatrick Why? Because few peoples on the planet 1986), cultural influence on experience would explicitly recognize the concept as (Jackson, 1989; Throop 2003a,b; Turner it has been developed in the context of & Bruner, 1986), dream states (Lincoln, Western philosophy and science, and their 1935; O’Nell, 1976), and other alternative P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

634 the cambridge handbook of consciousness

states of consciousness (visions, hallucina- most uniquely anthropological. Participant tions, drug-induced alternative states of con- observation is based on the assumption sciousness (Laughlin, McManus & d’Aquili, that, through participating in the day-to- 1990; Winkelman, 2000). Such Western sci- day activities of a particular culture or social entific and philosophical notions as stream of group, a researcher will be able to glean consciousness, qualia, sphere of conscious- insights from the effects of such participa- ness, intentional field, and the like are far tion on his or her own subjective states while less operationalizable cross-culturally, and also being granted access to observing social hence direct evidence about them is diffi- actors from the perspective of a co-actor cult to uncover in the ethnopsychologies and who is him- or herself enmeshed in the field ethnophilosophies of other peoples, save for of social action as it unfolds. those societies with religions that specialize Through participating in the everyday in one form of phenomenology or another activities of actors who are at times draw- (e.g., certain Buddhist and Hindu societies; ing from significantly different assump- see Chapter 5). tions about self, intersubjectivity, and real- That said, keeping the problem of eth- ity, anthropologists are thus often forced nocentricity squarely in mind and focusing to confront directly many of the assump- our attention on those attributes of con- tions that they otherwise take for granted sciousness that are in fact represented in the as “natural” attributes of human mentation, reports of peoples in widely disparate parts behavior, and social life. In addition to its of the world, we may say that the question function in highlighting the contours for of consciousness has in one way or another what Edmund Husserl (1913/1962) would been central to the anthropological enter- have termed the anthropologist’s “natural prise from its inception as a discipline. The attitude,” this methodology further allows domain of consciousness has been the focus anthropologists to explore how human con- of both temporal (historical or evolutionary) sciousness is manifest in the context of and spatial (cross-cultural) concerns for over everyday interaction. a century and a half. We may say with surety Although a great deal of anthropolog- that anthropologists have had much to say ical research has been grounded primar- about how observed differences and similar- ily upon this qualitative/descriptive foun- ities in the contents of consciousness cross- dation, cognitive anthropologists, psycho- culturally should inform theorizing about logical anthropologists, and evolutionary the structures and processes of conscious- psychological anthropologists have also, at ness as they are conceived in the social and times, incorporated standardized psycholog- physical sciences today. ical testing (e.g., Rorschach; see Hallow- ell, 1955), systematic data collection (e.g., triad testing, free-listing, paired compar- A Brief Note on Methodology isons, etc., see D’Andrade, 1995), and in in Anthropology some rare cases experimental designs (see Anthropological research on consciousness Barrett, 2004; Heinrich et al., 2004)in has traditionally been broadly qualitative the context of their investigations into the and descriptive, often relying on structured structures and contents of consciousness and semi-structured interviewing, observing cross-culturally. (and more recently videotaping) real-time social interaction, and participant observation. To this end, anthropologists have Structure and Content I: A History tended to rely upon a variety of methodolo- of Early Anthropological Thought gies that integrate third-person (i.e., observ- and Research ing), second-person (i.e., interviewing), and first-person (i.e., participant observation) Anthropology has been concerned with both perspectives. Of all of these methodolo- the structural underpinnings and content gies, participant observation is perhaps the variation in consciousness since at least the P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

anthropology of consciousness 635

mid-19th century. The linchpin idea that based upon this notion. He argued that permeated much of 19th-century thinking the mental acts of all people are the prod- about the origins of human culture and ucts of physiological mechanisms charac- consciousness was the “cultural idealist” teristic of the human species (what today (idealism from Latin idealis, “idea”; belief we might term the genetic loading on the that knowledge and experience about the organization and functioning of the brain). world derive from certain inherent struc- Every human mind inherits a comple- tures of mind) notion of the psychic unity ment of species-specific “elementary ideas” of humankind. This was the presumption (Elementargedanken), and hence the minds that there exists a single, overarching human of all people, regardless of their race or nature that permeates all peoples in all culture, operate in the same way. It is places on the planet, regardless of the par- the contingencies of geographic location ticularities of their sociocultural organiza- and historical background that we have tion. All people everywhere are born with to thank for the different elaborations of the same mental and physical potentialities, these elementary ideas, the different socio- which will cause them, when presented with cultural traditions, and the various levels of similar problems, to create similar solutions sociocultural complexity. According to Bas- to those problems. However, because peo- tian (1881, 1887), there also exists a law- ples are faced with quite unique circum- ful “genetic principle” by which societies stances by their locality and their history, develop over the course of their history from their societies and cultures will diverge in exhibiting simple sociocultural institutions time and appear, at least on the surface, to becoming increasingly complex in their to be quite different one from another. As organization. we see in the context of this section, the These elementary, inherited psycholog- interplay of similarity and difference that ical processes can be studied in a sys- is inherent in the 19th-century understand- tematic, objective, and comparative way, ing of psychic unity was itself, however, Bastian taught. Accordingly, he argued that articulated quite differently in the context the ethnographic project had to proceed of those early anthropologists whose writ- through a series of five analytical steps (see ings explored most explicitly both the struc- Koepping, 1983): ture and content of consciousness cross- culturally. 1. Fieldwork: Empirical description of cross- cultural data (as opposed to armchair phi- Adolf Bastian (1826–1905) and the losophy). Bastian himself spent much of Psychic Unity of Humankind his adult life among non-European peoples. Adolph Bastian was one of the leading pro- 2 ponents of this cultural idealist view in eth- . Deduction of collective representations: nology during the mid- to late-19th cen- From cross-cultural data we describe the tury. This was a period when Germany collective representations in a given soci- was a major player in science, and Bastian ety. was Germany’s leading social anthropolo- 3. Analysis of folk ideas: Collective rep- gist (Koepping, 1983; see also Lowie, 1937). resentations are broken down into con- Bastian (1881, 1887) is now credited with stituent folk ideas. Geographical regions having developed the notion of the “psychic often exhibit similar patterns of folk unity of mankind,” which played a critical ideas – he called these “idea circles” that role in defining many of the anthropological described the collective representations theories of consciousness that are explored of particular regions. below. 4. Deduction of elementary ideas: Resem- Bastian proposed a straightforward pro- blances between folk ideas and patterns ject for the long-term development of a sci- of folk ideas across regions indicate under- ence of human culture and consciousness lying elementary ideas. P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

636 the cambridge handbook of consciousness

5. Application of a scientific psychology: with the psychic unity question – a dialogue Study of elementary ideas defines the psy- that may very well have originated in the chic unity of humankind, which is due to context of Boas’s correspondences with Bas- the underlying psychophysiological struc- tian during the early 1880s and his later stud- ture of the species. ies under him in Berlin during 1886–1887 at the Royal Ethnographic Museum. What Bastian argued for was nothing less A key article highlighting some of Boas’s than what today we might call a psychobio- views on culture and sensation is his “On logically grounded, cross-cultural social psy- Alternating Sounds” (1889/1974). In this chology. Through ethnographic research, he article, Boas set out to demonstrate that most wrote, we can study the psychological laws forms of sensory perception are forms of of mental life as they reveal themselves in culturally mediated apperception – percep- diverse geographical settings. tion informed by some classificatory framework. Using the examples of “sound blind- Franz Boas (1858–1942) and the ness” and color perception, Boas attempted Anthropology of Sensory Impressions to demonstrate the effects of linguistically Although the investigation of consciousness mediated cultural categories on the pattern- per se may not have been an explicit object ing of auditory and visual sensations. The of attention for many anthropologists who phenomena of sound blindness – a condition followed Bastian in the context of their noted by the phonologists of the day wherein ethnographic research, practically since its the sounds uttered by a speaker are either inception as an organized field of study, not perceived at all or are misperceived by anthropology has been vitally interested in a hearer – provided the means for Boas to questions concerning the cultural pattern- lay the groundwork for a model of the cul- ing of sensory impressions. Indeed, a care- tural patterning of auditory sensations. This ful reading of the history of the discipline model was based on a suggestion that the reveals that much of modern anthropology probability that two resembling sounds will was founded upon an active interest in the be perceived as identical for any one given relation between culture and the differen- hearer is tied directly to the extent to which tial organization of various modes of sen- the two sounds correspond to a pregiven sory experience. Beginning with Franz Boas, classificatory frame (Boas, 1889/1974). whose dissertation in physics at the Univer- In the case of color perception, Boas sity of Kiel was focused on problems con- noted the growing evidence for cross- cerning the color of seawater and who later cultural variability in color terminology and spent time with Wilhelm Wundt at his lab- pointed to the fact that in languages where oratory in Leipzig studying the laws that there are no terms for “green,” the colors that account for the emergence, combination, fall within the spectrum of green become and organization of elementary sensations in classified according to the two closest sen- consciousness (Harkness, 1992; Laboratory sory frames; namely, blue and yellow. In the of Comparative Human Cognition, 1983), case of sound blindness, Boas focused upon we find an early and influential interest in how our culturally mediated apperceptive investigating the relationship between cul- frames serve to selectively filter out differ- ture and sensation. In fact, it was a grow- ences between discrete instances of auditory ing dissatisfaction with the psychophysical sense impressions. In the case of color per- approach of Wundt’s laboratory that eventu- ception, he noted how similar apperceptive ally led Boas to come “away convinced that frames work to selectively code for differ- ‘even elementary sensations were influenced ences between sense impressions that would by their contexts of occurrence” (Harkness, otherwise display a significant amount of 1992,p.103). This focus on the cultural pat- perceptual similarity. Borrowing from the terning of sensory experience also impor- Kantian tradition, Boas thus distinguished tantly fueled Boas’s long-standing dialogue apperception (cognitively mediated mental P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

anthropology of consciousness 637

representation of sense objects) from perception, auditory acuity, olfaction, taste, tac- ception (immediate sensation). And to this tile sensitivity, pain threshold, weight dis- extent, for “Boas, as Benedict put it, the see- crimination, and perceptual illusions tied ing eye was the organ of tradition” (Sahlins, to abilities to perceive differences in the 2000,p.152). relation between size and weight. With Boas is generally considered the father W. H. R. Rivers in charge of the exper- of cultural anthropology. His name and iments on vision; C. S. Myers focusing the methods he espoused are commonly on audition, olfaction, and gustation; and associated with relativist notions of human William McDougall investigating the hap- nature – that is, with the idea that in some tic senses and sensitivity to pain, the team way our species transcended biological con- sought to test ongoing debates over Her- ditioning when it became fully cultural in its bert Spencer’s (1886) hypothesis regarding patterns of adaptation. The na¨ıve assump- differences in the sensory acuity of “primi- tion here is that there was some kind of salta- tive” and “civilized” cultures (see Cole, 1996, tive leap from “nature” to “nurture” during p. 41; Richards, 1998,p.137). As Richards the evolution of humanity. But Boas himself (1998) notes, Spencer’s hypothesis was “that was neither so na¨ıve, nor was he completely ‘primitives’ surpassed ‘civilized’ people in comfortable with an extreme relativist (or psychophysical performance because more purely constructivist) view of human con- energy remained devoted to this level in the sciousness. One need only consider Boas’s former instead of being diverted to ‘higher discussion of psychic unity in his book, The functions,’ a central tenet of late Victorian Mind of Primitive Man (1910/1938), in which ‘scientiﬁc racism’” (p. 137). he wrestled with the tension between his The orientation of these researchers was conviction that all people everywhere are based upon the assumption, articulated by born with the same mind, and yet some Myers, that so-called primitive peoples were cultures appear to be more “advanced” in predisposed to pay greater attention to the terms of technological and social adapta- perception of “external stimuli” than Euro- tion than others. His overwhelming con- peans (cited in Titchener 1916,p.210). In line cern methodologically was with recording with this thesis, McDougall (1901) asserted the narratives of traditional peoples verba- that the people of Torres Strait showed tim and in using these texts as data upon greater sensitivity to the external impinge- which to eventually ground a solid science of ment of two points applied to the skin while human nature and consciousness – a research demonstrating a much greater insensitivity strategy he in fact shared with Bastian. to the internal impingement of pain sensation than their European counterparts. Find- ing that the Torres Straight islanders sub- The Torres Strait Expedition of 1898 jects were not fooled by a variety of opti- Boas was not alone in his interest in explor- cal illusions, Rivers (1901a) argued that the ing the relation between culture and sen- islanders had greater visual acuity than Euro- sation during this period. A year before peans. He also asserted that the islanders the publication of Boas’s pioneering discus- demonstrated a greater sensitivity to the sion of the inﬂuence of culture on auditory color red and a slight insensitivity to the and visual sensory modalities, the famous color blue, which was marked in terms of Cambridge Expedition to the Torres Strait the discernible prevalence or lack of lex- (between Australia and New Guinea) set ical items used to designate these partic- out to explore how culture affects “acu- ular ranges in the color spectrum. Myers ity for each of the basic senses” (Haddon, found that, although there were some obser- 1901; see also Cole, 1996; Richards, 1998; vational data suggesting that islanders had a Kuklick, 1991). Here Boas’s nascent inter- greater ability to identify faint, impercepti- est in color perception and audition was ble sounds, none of his tests on audition or extended to general visual acuity, spatial per- olfaction proved conclusive in this regard. In P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

638 the cambridge handbook of consciousness

fact, as Cole (1996) notes, Myers “found that of social organization that serve to shape auditory sensitivity was somewhat greater the categorical schemes mediating an indi- for European samples while there was no vidual’s perception of space. Expanding apparent difference in sensitivity to odors” on some of these ideas in his magnum (p. 43). opus, The Elementary Forms of Religious Life Although it seems that the expedition (1912/1995), Durkheim argued further that was able to garner some evidence support- socially derived collective representations ing the extent to which culture might serve play an important role in patterning the to pattern sensory experience, a number of individual’s various sensory modalities (see questions concerning the accuracy and valid- also Throop, 2003c). As Durkheim put it, ity of the methods and findings of the expe- collective representations “turn upon sen- dition were raised in a critical article by sation a beam that lights, penetrates, and Titchener (1916; see also Cole, 1996). transforms it” (1912/1995,p.437). More- over, Durkheim demonstrated that collective representations are reinforced through Emile Durkheim (1858–1918) and the Elementary Forms of Consciousness rituals that pair the society’s knowledge with certain altered states of consciousness – or The concern for the invariant properties what he called “collective effervescence.” of consciousness, which was first expli- This strain of thinking in Durkheim influ- cated in Bastian’s project, continued through enced the fundamental structuralist notions the history of the discipline. Although perpetuated by Mauss and later by Mauss’s the overly simplistic, evolutionist notion own pupil, Claude Levi-Strauss´ (1908-), of the “psychic unity of mankind” came founder of modern-day structuralism (see into serious question in the early 20th below). century, researchers were still interested Indeed, in his now famous essay, Les tech- in pan-human regularities in sociocultural niques du corps (1950), Mauss importantly phenomena, as well as the variation in extended some of these basic ideas regarding sensory experience influenced by culture. the relationship between culture and con- Well informed by these discussions, Emile sciousness to questions concerning the ways Durkheim (1858–1918), a renowned father in which culture influences body posture, of contemporary social theory, argued that balance, motility, and the kinesthetic sense. socially derived collective representations Using the term “habitus” to account for the play an important role in patterning the indi- acquired practical nexus that informs an vidual’s various sensory modalities. At the individual’s habitual use and perception of same time, he emphasized, in a critical dia- the body, Mauss viewed reproduction, con- logue with Immanuel Kant, that although sumption, care, movement, rest, and sleep all people come to know the world filtered as significant fields for investigating the vari- through the categories of time, space, num- ous ways that culture can pattern techniques ber, cause, class, person, and totality, these of somatic perception and action. As we see categories of understanding were empiri- below, this insight informed later approaches cally derived through social experience. in anthropology that sought to emphasize Emile Durkheim and his student, Marcel the significance of non-discursive, practical, Mauss (1872–1950), worked on the rela- and embodied aspects of consciousness. tions among classificatory schemes, social organization, and spatial perception, an Lévy-Bruhl (1857–1939) on Multiple effort that can also be viewed as an early Modes of Consciousness attempt at establishing the parameters for an anthropology of consciousness. In Prim- Building upon Durkheim’s earlier writings itive Classification (Durkheim & Mauss, on collective representations, Lucien Levy-´ 1903/1963), they suggested that it is forms Bruhl, both an influential and controversial P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

anthropology of consciousness 639

ﬁgure in the history of the discipline, set a nucleus surrounded by a layer of varying out to further Durkheim’s initial insight density of representations which are social in that “different mentalities will correspond to their origin” (1926,p.44). For Levy-Bruhl,´ different social types” (1926,p.27) by detail- it was precisely this collectively generated ing what he argued to be the fundamental representational saturation that accounted differences between “primitive” and “mod- for the differential functioning of primi- ern” mentalities (see Throop, 2003c). Where tive logic that operated without regard to the latter were held to be organized accord- the law of contradiction: the idea that a ing to logical modes of thought that were statement and its negation cannot be true primarily grounded in cognitive function- simultaneously. Because a “primitive’s” col- ing, the former were thought to be funda- lective representations directed his or her mentally “prelogical” and infused with imag- perception beyond immediate sensory data ination and emotional currents that often to the occult forces and the impercepti- served to distort the stability and coherence ble elements thought to operate beyond of the world as ﬁrst given to the senses. Care- the purview of our various sensory modal- ful not to follow Edward Tylor (1817/1958) ities, the primitive mind was understood by and Herbert Spencer’s (1886) view of prim- Levy-Bruhl´ to be driven to see connections itive mentality as an antecedent and infe- between otherwise logically disparate phe- rior stage to modern mentality, Levy-Bruhl´ nomena (e.g., between a man and his totem (1926) argued that both forms of mentality, animal). In Levy-Bruhl’s´ (1926) estimation, although organized differently, were to be therefore, categorical thought and instances understood as equally valued. of mutual exclusivity, which were the puta- In terms of the cultural patterning of tive hallmark of logical thought, were aban- conscious experience, Levy-Bruhl´ (1926) doned in primitive mentality. By contrast, asserted that, although for the modern mind primitive mentality was held to be orga- mental representations are cognitive phe- nized according to a “law of participation” nomena that are precise and well differ- whereby the mind is not merely presented entiated, for primitive mentality, mental with an object, but “communes with it and representations are far more complex, undif- participates in it, not only in the ideological, ferentiated, and infused with emotion, feel- but also in the physical and mystic sense of ing, and passion. Furthermore, he held that, the word” (p. 362). in addition to being suffused with such Although his perspective on putative dif- emotional and motor currents, the primi- ferences between these differing mentalities tive mentality is directed almost exclusively was criticized quite rightfully by his peers by culturally constituted collective represen- and successors, in all fairness to Levy-Bruhl,´ tations. Accordingly, it is “bound up with we should note that later on in his career preperceptions, preconceptions, preconnec- he came to argue that all people, be they tions, and we might almost say with pre- primitive or civilized, operate on both logical judgments” that alter the functioning of such thought and the “participation mystique,” mental capacities as reason, logic, and infer- depending upon social context (see Levy-´ ence (p. 108). Bruhl, 1975). In other words, Levy-Bruhl’s´ Levy-Bruhl´ believed that the primitive description of the primitive mentality can be mentality, driven by an alternative kind of understood as an accurate description of all logic, operated such that an individual would mentalities where cognitive, affective, and confront a world constituted by collective conative (i.e., volitional) elements are ever representations that were largely “impervi- present and often interfused. Indeed, much ous to experience.” As he put it, “Primitives of human experience is conditioned to a see with eyes like ours, but they do not per- great extent by “preperceptions,” “precon- ceive with the same minds. We might almost ceptions,” “preconnections,” and “prejudg- say that their perceptions are made up of ments.” P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

640 the cambridge handbook of consciousness

Language, Consciousness, and the temporal sense of “ever becoming later,” the Sapir-Whorf Hypothesis streams of sensory input are organized or patterned by habitual language use One of Boas’s students, Edward Sapir (1884– and the often covert grammatical forms – 1939), and one of Sapir’s students, Benjamin “cryptotypes” – that serve to give structure Lee Whorf (1897–1941), are famous both to that usage. Indeed, Whorf argued that the within and beyond the discipline of anthro- punctual and segmentative aspects of Hopi pology for revisiting Boas’s concern for cul- verbs actually supply Hopi speakers with a tural variation in consciousness. Each in their grammatical system that is better equipped own way explored the effects of culture to describe vibratile phenomena than the and language on the patterning of con- terminology we find in modern- day physics. tents of consciousness. Throughout their In this respect, Whorf (1956) argued that careers, they attempted to highlight the var- this “Hopi aspect-contrast which we have ious ways that cultural and linguistic sys- observed, being obligatory upon their verb tems serve to selectively direct individuals’ forms, practically forces the Hopi to notice attention to linguistically salient elements in and observe vibratory phenomena, and the perceptual field. Although there is still furthermore encourages them to find names much controversy over the extent to which for and to classify such phenomena” (p. 56; Sapir (1929/1958) would have supported emphasis ours). versions of linguistic relativity that pos- The combined efforts of Sapir and Whorf tulated any simple correspondence among to understand the influence of language the structures of language, the structures upon experience came to be known as the of thought, and the patterning of sensa- Sapir-Whorf Hypothesis. The tradition in tion and perception (e.g., see Gumperz & anthropology has been to speak of a hard Levinson, 1996), he did argue, however, that and a soft version of the hypothesis. The hard even “comparatively simple acts of percep- version would read something like this: Lan- tion are very much more at the mercy of guage categories determine the organization social patterns called words than we might of experience. However, virtually all of the suppose. ...We see and hear and otherwise empirical evidence over the last half-century experience very largely as we do because garnered in both anthropology and psychol- the language habits of our community pre- ogy appears to disconfirm this version of dispose certain choices of interpretation” the hypothesis. Moreover, neither Sapir, nor [emphasis in original] (p. 69). Whorf ever advocated such a simplistic and Following Boas’s and Sapir’s lead, deterministic relationship between language Benjamin Lee Whorf (1956) also examined and consciousness. What they did do, how- the relations among language, thought, and ever, was to hold fast to the notion, com- experience. For instance, in an exploration of mon in their day, that the structure of lan- how punctual and segmentative aspects of guage provides researchers with significant Hopi verbs come to characterize phenom- insight into an underlying cultural ontology ena according to “pulsative” and “vibrative” for reality. Consider one of Sapir’s (1931) characteristics, Whorf (1956) pointed out statements: “The ‘real world’ is to a large that, although we tend to view language in extent unconsciously built up on the lan- terms of its expressive function, “language guage habits of the group” (p. 580). Whorf first of all is a classification and arrangement (1956) took this a step further when he sug- of the stream of sensory experience which gested that “the categories and types that results in a certain world-order, a certain we isolate from the world of phenomena segment of the world that is easily express- we do not find there because they stare ible by the type of symbolic means that every observer in the face. On the con- language employs” (p. 55; emphasis ours). trary the world is presented in a kaleido- According to Whorf (1956), whether we are scopic flux of impressions which have to talking about visual perception, audition, or be organized in our minds. This means, P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

anthropology of consciousness 641

largely, by the linguistic system in our minds” sonal psychology and the influence of the (p. 128). culture in which he or she has been raised Although the extent to which either (1955,p.40). thinker would have supported the “hard” This understanding of the effects version of the hypothesis is still much of culture on perceptual processes – debated, it is safe to assume that both Sapir particularly those perceptual processes and Whorf would have definitely supported subserving vision – further informs what is a softer version, which might go something perhaps Hallowell’s most famous contribu- like this: The way people learn to speak tion to anthropological theory; namely, the about things influences how they are con- concept of the “behavioral environment.” ditioned to think about things (see Lucy, Borrowing the term from the Gestalt psy- 1985; Rumsey, 1990). The weight of the evi- chology of Koffka, Hallowell explained that dence over the last half-century would sup- the behavioral environment is not merely port some form of this view (Gumperz & the generalized physical or geographical Levinson, 1996; Hunt & Agnoli, 1991; Kay & environment, but rather that culturally Kempton, 1984; Lucy, 1992). ordered environment that is experienced by an individual social actor. That is, according to Hallowell, the behavioral environment Irving Hallowell (1892–1974) and consists of an ordered world of objects, the “Behavioral Environment” persons, and relations that is experienced by Irving Hallowell is another important pio- a social actor whose perceptual capacities neering figure in anthropology’s long- effectively suffuse those objects, persons, standing interest in the cultural patterning of and relations with personally and culturally sensation and consciousness. Much like Sapir constituted meanings, values, feelings, and and Whorf before him, Hallowell’s approach motivations. That said, in accord with to consciousness can be understood as an the ongoing dialogue in anthropology extension of the Boasian tradition. This is between the interplay of similarity and not surprising because Hallowell was a stu- difference in the structure and content of dent of Boas’s student, Frank Speck, at the consciousness cross-culturally, Hallowell University of Pennsylvania, and he also par- (1955) also argued that it is possible to ticipated regularly in Boas’s weekly semi- determine a number of basic orientations nars at Columbia (Bock, 1999). Moreover, that provide the structural underpinnings although never officially a student of Sapir’s, for the constitution of an individual’s world Hallowell was also clearly influenced by his of experience, regardless of the historical, writings (Regna Darnell, personal communi- cultural, and/or personal factors that shape cation). Hallowell (1955) argued that much siginificantly each of these orientations in a of what individuals take to be “direct per- multitude of ways. Outlining orientations ception” is in fact informed by personally for self, objects, space-time, motivation, and and culturally constituted “nonsensory” sym- norms, Hallowell demonstrated how each bolic forms. Citing Bruner’s (1951) claim orientation can be construed as a universal that perception must be understood as the and necessary structure of human con- expression of the entire personality structure sciousness that is also patterned significantly and Bartlett’s (1932) view that the mind’s according to unique cultural, social, and perceptual processes are organized accord- personal dictates. For instance, he pointed ing to an “effort after meaning,” Hallowell out that, although for all cultures it is nec- explained that much of what we under- essary to classify and discriminate among stand to be the transparent perception of an objects in relation to a culturally constituted external world is guided by the psychophys- understanding of the self, the ways in ical structure of the nervous system, which which objects are conceived varies tremen- has developed in each individual member dously among different cultural systems. of society according to that individual’s per- Hallowell explored this complex blend P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

642 the cambridge handbook of consciousness

of psychophysical, personal, and cultural from empirical and experiential diversity to influences upon perception through the use conceptual and intellectual simplicity that of traditional anthropological methods as guided much of Levi-Strauss’s´ structural- well as by administering Rorschach tests to ist vision. In this paradigm, cultural prod- subjects from Ojibwa and other societies ucts were therefore understood to be con- (1955,pp.32–74). strained by the limits imposed by the structural underpinnings of the human mind. These products were held to be expressions of the mind that generates them and were Structure and Content II: Modern intelligible on that accord. Structuralism, Practice Theory, In this light, Levi-Strauss´ argued that Hermeneutics, and the Anthropology phenomenological appeals to lived expe- of Experience rience in the context of anthropological explorations into the structures of con- Many of the themes and orientations present sciousness are in and of themselves always in the early development of the anthropol- inadequate. They are inadequate because ogy of consciousness continued to be devel- Levi-Strauss´ held that such an approach nec- oped during the mid- to late 20th century. essarily biases researchers to focusing exclu- However, research techniques became more sively on the particular, contingent, individ- sophisticated, and there were many more ual, and often inexpressible affective and ethnographers collecting data from among sensorial realm. He paid little attention the myriad peoples on the planet. Among to this “superficial,” “idiosyncratic” realm other things, the tension between univer- of individual consciousness, however, for salist (orientation toward the ways peoples although “ideas resulting from hazy and are alike) and relativist (orientation toward unelaborated attitudes ...have an experien- the ways people are different) perspectives tial character for each of us. ...[t]hese expe- became intensified. As we shall see, part of riences, however, remain intellectually dif- the tension is ideological, but part is also fuse and emotionally intolerable unless they methodological – it is next to impossible to incorporate one or another of the patterns perceive the structures subserving human present in the group’s culture. The assim- activity in the kind of naturalistic setting ilation of such patterns is the only means encountered by most ethnographers in the of objectivizing subjective states, of formu- field. We return to this point in the conclu- lating inexpressible feelings, and of integrat- sion of the chapter. ing inarticulated experiences into a system” (Levi-Strauss,´ 1958/1963,pp.171–172). In his analysis of myth, Levi-Strauss´ Claude Lévi-Strauss (1908–) and (1964/1969, 1966/1973, 1968/1978, 1971/ Structuralism 1981) attempted to show how, despite their An important contribution to anthropologi- seemingly non-linear, mystical, and irra- cal approaches to consciousness is found in tional exposition, myths are governed by a the work of the famous French structural- logical structure of binary oppositions, medi- ist, Claude Levi-Strauss.´ What was novel ators, and the reconciliation of the very about Levi-Strauss’s´ approach to conscious- logical contradictions (e.g. between nature ness was that he set out to ground the and culture) that Levy-Bruhl´ claimed to go generative source of particularistic cultural unnoticed by thinkers immersed in a par- givens in the universal structures of the ticipatory consciousness. For Levi-Strauss´ human brain (see Scholte, 1973; Throop, these seemingly irrational and culturally spe- 2003c). He held that it was unconscious cific narratives are actually one of the best mental structures that gave rise to the means of accessing the invariant structures experiential vicissitudes of culturally medi- of the human mind. Most simply put, Levi-´ ated consciousness. It was thus a movement Strauss’s project is predicated upon a model P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

anthropology of consciousness 643

of consciousness and culture that is rooted ongoing expressive behavior and the social in externalization. It is important to real- context within which that behavior arises. ize that Levi-Strauss´ was not interested in This description of expressive behavior in personal externalization (i.e., creativity), but the context of its occurrence is thus under- rather in a view that takes externalization to stood to serve as the foundation from which be tied primarily to the projection of the uni- researchers are later able to build an under- versal and invariant structures of the human standing of the interpretive frameworks that mind into the tangible medium of cultural actors use to give meaning to their action. forms, in this case mythological texts (and Following Max Weber (1864–1920), elsewhere kinship systems; see Levi-Strauss,´ Geertz (1973) asserted that his vision of an 1949/1969). “interpretive science,” whose purpose is to Levi-Strauss’s´ (1958/1963) approach to explore the “webs of significance” that serve consciousness has been faulted for failing to to constitute culture, is founded on the idea recognize the significance of emotion and that culture is a public, ideational, and yet motivation in all forms of mentality and for non-mentalistic system of construable signs. overemphasizing the importance of exter- Again paraphrasing Ryle, Geertz held that nalization in the patterning of cultural prod- it is precisely the inherently public nature ucts such as myth. In addition, many crit- of culture that makes problematic the ical assessments of Levi-Strauss’s´ work are notion that cultural symbols exist “in some- grounded in the contention that he never one’s head” – as operating within human systematically outlined a methodology for consciousness. Accordingly, he strongly how he went about discovering binary oppo- criticized the work of Goodenough (1971) sitions, mediators, etc., in the context of his and other “ethnoscientists” (e.g., Conklin, textual analysis of the underlying structural 1969; Frake, 1961; etc.) who have construed elements of myth. Although having impor- culture to be “composed of psychological tant roots in Bastian’s and Boas’s research structures by means of which individuals or paradigms, Levi-Strauss’s almost exclusive groups of individuals guide their behavior” methodological reliance on texts to the (p. 11). exclusion of other research methodologies In fact, Geertz (1973) believed with Ryle fueled many of the most trenchant critiques (1949/1984) that “mind” is a term that of structuralist approaches to consciousness denotes most accurately not some privately in the discipline as a whole. accessible “ghost in the machine,” but rather a publicly accessible “class of skills, propensities, capacities, tendencies and habits ...[in Clifford Geertz (1926–2006) and the short] an organized system of dispositions Interpretivist/Hermeneutic Approach which finds its manifestation in some actions In his now famous appropriation of philoso- and some things” (p. 58). He argued that pher Gilbert Ryle’s (1949/1984) notion it is these “external” (i.e., public) symbolic of the “thick description” of the behav- manifestations of complexes of skills and iors that reputedly evidence the workings habits that ultimately underlie all reflective of internal/mental acts, Clifford Geertz’s modes of consciousness. Indeed, what many (1973) approach to consciousness was clearly scholars assume to be mental processes are, rooted in his position that an “I-am- according to Geertz, construed more accu- a-camera, ‘phenomenalistic’” approach to rately as processes of “matching ...states anthropology is necessarily insufficient for and processes of [public] symbolic models any attempts to uncover the structures of against [equally public] states and processes signification that give meaning to observed of the wider world” (1973,p.78). phenomena (see Throop, 2003a). According Geertz also followed Ryle’s lead in argu- to Geertz, “thick description” as a methodol- ing that it is only by viewing symbols to ogy is predicated on anthropologists turning be “material vehicles of thought” that allow their attention to a detailed description of individuals to impress meaning upon objects P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

644 the cambridge handbook of consciousness

in experience that we can ever hope to (as well as philosophers and other social sci- legitimize the anthropological study of con- entists) to the significance of exploring non- sciousness. Geertz believed that it is only this conceptual, embodied aspects of human public view of symbols that can ensure the consciousness (see Throop & Murphy, possibility that anthropologists may eventu- 2002). His perspective, which has inspired ally discover properties of cultural and per- a long line of scholars interested in explor- sonal systems through systematic empirical ing consciousness from a practice theoretical analysis (1973). In Geertz’s view then, his approach, pivots on the concept of “habitus.” perspective provides anthropology with a According to Bourdieu, habitus is a durable way in which to uncover “what is given, set of culturally and socially inculcated pro- what the conceptual structure embodied in clivities to act, perceive, feel, appreciate, and the symbolic forms through which persons generally inhabit the world. He argued that are perceived actually is” (1973,p.364). It the concept of habitus enables him “to break was his opinion that this potential “method away from the structuralist paradigm (see of describing and analyzing the meaningful the above discussion of Levi-Strauss)´ with- structure of experience” can therefore pro- out falling back into the old philosophy of vide anthropology with the basis for estab- the subject or of consciousness” (Bourdieu, lishing a valid scientific “phenomenology of 1985,p.13). culture” (1973,p.364). Bourdieu’s (1977) understanding of habi- Building on Ryle, as well as on the work of tus is based on the assumption that “agents George Herbert Mead, Geertz argued that are possessed by their habitus more than the assumption that culture is both pub- they possess it” (p. 18). It is thus not lic and social leads inevitably to the insight through conscious attention to predeter- that cultural processes do not “happen in mined “roles,” “rules,” or “models” that the head,” but, consist, in contrast, of a traf- agents negotiate their interactions with the fic of significant symbols that “impose mean- social world, but through the unintentional ing upon experience”(1973,p.45; emphasis triggering of strategic patterns of thought ours). In short, Geertz’s view is that con- and action produced by habitus in its mutu- scious experience is a cultural artifact. With- ally informing relation to structure. Here out the medium of culturally infused signifi- lies what Bourdieu labeled the “fallacy of cant symbols Geertz asserted that conscious the rule,” which is based on the putatively experience would be relegated to incoher- false belief that agents are conscious of ent and impenetrable behaviors and sensa- how their practice is “objectively governed.” tions. For Geertz then, the imposition of Instead, Bourdieu argued that agents always meaning on an otherwise chaotic stream of lack explicit cognizance of the “mechanisms stimuli and responses is one of the key defin- producing this conformity in the absence ing components of human existence as cul- of intention to conform” (1977,p.29). tural beings. It is this externalized, socially According to Bourdieu, rules are only ever infused understanding of culture as a coher- functional if they serve to “awaken ...the ent system of symbols serving to inform schemes of perception and appreciation all moments of conscious experience that deposited, in their incorporated state, in allows Geertz to later establish his memo- every member of the group” (1977,p.17). rable metaphor of culture as text (1973). Rules are thus only ever intended to deal with “the collective enterprise of inculcation tending to produce habitus” (1977,p.17). Pierre Bourdieu (1930–2002) Connected to the “fallacy of the rule,” and Practice Theory and also crucial to understanding his per- Taking a significant step away from both spective on consciousness, is what Bourdieu Levi-Straussian´ structuralism and Geertzian terms the “finalist illusion.” In Bourdieu’s semiotics, Pierre Bourdieu may be credited opinion, practices are produced by habi- with turning the attention of anthropologists tus, which serves as a generative principle P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

anthropology of consciousness 645

allowing individuals to adjust and respond to habitus as a set of mental, perceptual, cog- ever-changing situations. He suggested that nitive, and behavioral dispositions that are it is a mistake, however, to believe that these inscribed in agents’ individual bodies early strategies are determined in accord with in their lifespans while being “constantly some future orientation that is explicitly reinforced by calls to order from the group, accessible in conscious experience. Instead, that is to say, from the aggregate of individ- he argued that habitus is “always tending to uals endowed with the same dispositions,to reproduce the objective structures of which whom each is linked by his dispositions and they are a product, [as] they are deter- interests” (1977,p.15; emphasis ours). mined by the past conditions which have Bourdieu also stated that the unconscious produced the principle of their production” patterning of perceptual, evaluative, and (1977,p.72). In this light, habitus “is that behavioral schemes functions in such a way presence of the past in the present which that conscious “quasi theoretical reflection” makes possible the presence in the present of in practice merely serves to conceal “the true the forth-coming” (2000,p.210). Bourdieu nature of practical mastery” (1977,p.19). therefore stresses the conservative nature Here we again find some inconsistencies in of habitus and the unconscious basis for his position, however, for he also explained the patterning of human behavior, for he that pedagogy can be seen as a process of asserts that habitus is predisposed to defend explication that entails the necessary dis- its integrity, resisting change by avoiding or tancing from the “natural” feel of practical rejecting information that serves to contra- mastery in the attempt to transfer skills from dict or question its previously acquired pre- master to student (1977,p.19). If it is indeed suppositions (1977,pp.60–61). As he puts it, the case that the transmission of skills from “habitus tends to favor experiences likely to an adept to a novice is often associated with reinforce it by[protecting] itself from crises direct explicit representation in conscious- and critical challenges by providing itself ness that is thus distanced from the expe- with a milieu to which it is as pre-adapted as rience of “practical mastery,” the question possible” (1990,p.61). In this light, Bourdieu that may be asked of Bourdieu is why he explained that habitus is implicated in pro- devotes so little attention to exploring the cesses of naturalizing the conventional in the relationship between representational and service of producing taken-for-granted (i.e., non-representational aspects of conscious- doxic) orientations to social life. In his words, ness and why he focused so much atten- “schemes of thought and perception can pro- tion on the unconscious acquisition of habi- duce the objectivity that they produce only tus. These questions and critiques aside, by producing misrecognition of the limits Boudieu’s formulation of habitus helped to of the cognition that they make possible, draw greater anthropological attention to thereby founding immediate adherence, in the significance of non-reflective, embodied, doxic mode, to the world of tradition expe- and sensory aspects of consciousness as key rienced as a ‘natural world’ and taken for sites for understanding social actors cultur- granted” (1977,p.164). allly defined participation in everday life. Bourdieu occasionally acknowledged, however, that we cannot in fact dispense Victor Turner (1920–1983) and completely with all forms of explicitness the Anthropology of Experience in consciousness, precisely because within the dialectical interaction of structure and Victor Turner has been held as one of the disposition, the inscription of structure in key figures in the development of both sym- habitus is seldom, if ever, a perfect match. bolic anthropology and the anthropology Even with these occasional cautions he of experience, and his approach to con- was not, however, always consistent in this sciousness has also influenced significantly regard. There are, in fact, a number of many current approaches to the topic in instances where we find Bourdieu defining the discipline (Peacock, 1975; Throop 2002, P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

646 the cambridge handbook of consciousness

2003a). In his earliest works, Victor Turner’s p. 15). In this respect, it can be said that every account of consciousness attempted to inte- “symbolic item is related to some empir- grate symbolic, physiological, bodily, sen- ical item of experience,” whether that be sory, and emotional dimensions of human “sensorily perceptible objects,” feeling states, existence. His distinction between the ideo- desires, a sense of purposiveness, or emo- logical/normative and sensory/orectic poles tions (pp. 42–43). In Turner’s early writings of meaning serves as a good case in point. (e.g., Turner, 1969), we find that conscious- In Turner’s (1969) framework, ritual sym- ness is most often characterized as having a bols are understood to be conventionalized, psychobiological base that is shaped by both multivocal (i.e., multi-referential) signs that culturally patterned and universally struc- instigate action and that share three basic tured properties that help give definition to properties: (1) the condensation of multi- an individual’s subjective take on his or her ple percepts, thought objects, and actions physical, social, and cultural worlds. within a “single formation”; (2) the unifi- Turner’s later writings on consciousness cation of disparate significata by means of (e.g., Turner 1982) were directly inspired “analogous qualities or by association in fact by the writings of the German philosopher or thought”; and (3) the polarization of Wilhelm Dilthey (1833–1911; see Throop, meaning into ideological (i.e., structurally 2002, 2003a). In the context of his dialogue normative) and sensory (i.e., physiological, with Dilthey, Turner extended his earlier orectic) referents (1969,p.28). As Turner understanding of consciousness to encom- put it, “At one pole cluster a set of referents pass a trichotomy of cognitive, affective, and of a grossly physiological character, relating volitional dimensions as they are directly to general human experience of an emotional lived through by individual actors. It is not kind. At the other pole cluster a set of ref- until 1982 that we find Turner’s first pub- erents to moral norms and principles gov- lished discussion of conscious experience erning social structure. ...Here we have an from the perspective of Dilthey’s frame- intimate union of the moral and material” work. Turner pointed out that in contrast (1969,p.54; emphasis ours). to Kant’s belief that it is only with the It is in the context of the polarization of conceptual pattering of sensation that the meaning that Turner argued that these two “raw data” of experience are given defi- basic poles become intimately interlinked. nite form, Dilthey holds that every distinct They are linked through processes of rit- unit of experience is given to an individual engagement with cultural symbols that ual’s awareness with a certain structure that shape individual consciousness according to is not merely the result of the categorical the dictates set by cultural and social ide- impositions of the human mind. Instead, als, morals, edicts, roles, and rules. In other “the data of experience are ‘instinct with words, in his earlier writings it is through form,’ and thought’s work is to draw out a ritually mediated engagement with sym- ‘the structural system’ implicit in every dis- bols that Turner grounded the canalization tinguishable Erlebnis or unit of experience” of an individual’s subjective, emotional, and (1982,p.13). It is important to point out physiological experiences according to pre- here that, although focusing upon struc- given social and cultural structures. More- tures of experience, Turner was also quite over, it is also through ritual that the imper- familiar with Dilthey’s attempt to discuss ceptible world of cultural representations is different moments of experience that are given coherence in reference to the percep- understood to be integral components of tible world of the senses. As he explains in the organization of any distinctive struc- reference to the Nbembu’s (dwelling in West ture of experience (see Throop, 2002). After Africa) cultural belief in “shades” (ghosts, outlining five such moments – the percep- ancestors), ritual “connects the known world tual core, the evocation of past images, the of sensorily perceptible phenomena with the revival of associated feelings, the emergence unknown and invisible realm” (Turner, 1969, of meaning and value, and the expression of P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

anthropology of consciousness 647

experience – Turner went on to argue, how- between culture and the contents of con- ever, that it is only really in the fifth moment sciousness, inasmuch as the latter variously of “expression” that the structured unit articulate with differing temporal orienta- of experience can be said to reveal itself tions in the stream of consciousness (see (1982,p.15). Throop, 2003a). Turner further recognized Dilthey’s important distinction between the immediate living through of experience as a Structure and Content III: sequence of events (Erleben) and the ret- Contemporary Developments in the rospective attribution of meaning tied to Anthropology of Consciousness the structuring of experience as a particular coherent unit or form. In this way, Turner Levi-Strauss’s structuralism, Geertz’s her- argued that in the context of its expression meneutics, Bourdieu’s practice theory, and “experience is both ‘living through’ and Turner’s anthropology of experience each ‘thinking back’ ...[and it] is also ‘willing set the stage for many of the contempo- or wishing forward’” (1982,p.18). Here, rary approaches to consciousness in the dis- Turner explores the multilayered structur- cipline that have arisen during the past 30 ing of “experience in terms of the temporal years. As we see below, whether it is in organization of meaning,value, and ends” terms of a critical reaction to, or an attempt (see also Turner, 1985,pp.214–215; Throop to build upon, the insights of these earlier 2003a). According to Turner’s reading of approaches to consciousness in the disci- Dilthey, “meaning” is essentially a cognitive pline, the themes, orientations, and method- structure oriented to the past. “Value” is, ological and theoretical tensions found in the by contrast, an affective structure tied to past century and a half continue to develop the vicissitudes of the present moment, in contemporary anthropology. Meanwhile, whereas “ends” are volitional structures new and intriguing approaches to conscious- tied to goal-directed behavior oriented ness have developed as well, all within the toward an emerging future. Again following basic naturalistic framework of ethnological Dilthey, he argues that there is little or no research. In this section, we explore some of cognitive coherence in the “unarticulated the major approaches to consciousness that quality of value.” Instead, it is only with have emerged in anthropology over the last the interconnection of what are other- 30 years. wise latent “tonal-affinities” that disparate “values” are able to be organized into a Transpersonal Anthropology and coherent structure through the “ligatures” Altered States of Consciousness provided by personal and cultural forms of meaning. In contrast to the transient It is perhaps anthropological attempts sequentiality that is characteristic of value to document evidence for alterations in as it manifests in its immediate immersion states of consciousness cross-culturally that in the “conscious present” – “[m]eaning have garnered the most attention from is apprehended by looking back over a researchers outside the discipline (see Chap- temporal process” (1982,p.76; emphasis in ter 19). And indeed, it is true that anthro- original). pologists have encountered, recorded, and In his dialogue with Dilthey, Turner thus thought about the importance of alternative highlighted for anthropologists the signifi- states of consciousness since the latter part of cance of exploring consciousness in terms the 19th century (e.g., see Lang, 1894, 1897, of its temporal structure (e.g., retrospection, 1898). Moreover, these observations have perception, and anticipation). As is evident challenged many of the assumptions about from Turner’s discussion of value, mean- consciousness that have appeared in the con- ing, and ends, it seems that there may very text of Western scientific approaches to the well be significant variations in the relation topic, including the prevalent “monophasic” P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

648 the cambridge handbook of consciousness

bias – the tendency to focus almost exclu- dent that few ethnographers had personally sively on exploring the contents and struc- experienced and reported on altered states tures of consciousness in normal waking. of consciousness related to the peoples they Such interest continued intermittently were studying. throughout the 20th century in a variety Methodological issues were part of the of reports of altered states of conscious- problem – it is very difficult to access altered ness. For example, this century saw reports states of consciousness from “outside” as it by Lincoln (1935) on dreams and culture, were (Braud & Anderson, 1998; Laughlin, Barnouw (1946) on the paranormal cross- 1989). Also, until the latter part of the 20th culturally, Aberle (1966) on the use of century, there had been an unwritten tradi- peyote in the Native American Church, tion in ethnology against using first-person Bourguignon (1973, 1976) on altered states narratives. Only during the last few decades of consciousness cross-culturally, Bharati have such reports begun to surface. Among (1975) on tantric meditation practices, and the reports are the following: Grindal (1983) Furst (1976), Dobkin de Rios (1984), and on a profound experience that occurred Dobkin de Rios and Winkelman (1989) to him while attending a Sisala funeral in on the use of hallucinogens and cultural Ghana in 1967; Katz (1982,pp.6ff ) on expe- practices. Moreover, Halifax (1975), Peters riences attendant to trance-dancing among (1982), and Peters and Price-Williams (1980) the Bushmen; Laughlin (1994) on altered have written on altered states of conscious- states of consciousness related to tantric ness related to shamanism, Long (1977) Buddhist practices; Laughlin, McManus, on extrasensory phenomena cross-culturally, and Webber (1984) on experiences arising Jilik (1982) on altered states of consciousness while meditating upon complex symbols in and healing, and Rouget (1985) on music and Tibetan tantric Buddhism, Lederman (1988, trance states. pp. 805–806) on ritually driven altered states However, it took developments in psy- of consciousness among Malay shamans, chology to make ethnology more aware MacDonald et al. (1988) on experiments that extraordinary experiences were not with the shaman’s mirror; Young-Laughlin only important for challenging assumptions and Laughlin (1988) and Webber, Stephens, about the nature of consciousness in North and Laughlin (1983) on the phenomenol- American and Western Europe but could ogy and structure of masked performance; also be of prime significance for doing ethno- Coult (1977) on the ethnological signif- graphic research. Hot on the heels of the icance of psychedelic experiences; Edith rise of transpersonal psychology in the late Turner (1996) on perceiving spirit entities; 1960s and early 1970s (Boucouvalas, 1980; Harner (1973) on hallucinogens and religion; Sutich, 1968), a number of anthropologists George (1995) on telepathic dreaming; began to realize that the experiences asso- Chagnon (1977,pp.154ff ) on an experiment ciated with some religious texts and ritual with hallucinogenics, shamanic dance, and practices encountered by ethnographers are chanting; and Young and Goulet (1994) for of an exceptional nature. They are in fact a review of such reports. altered states of consciousness – states not An excellent example of how an anthro- easily accessed by researchers using routine pological focus on alternative states of con- ethnographic field methods (see Campbell & sciousness has contributed to the scientific Staniford, 1978; Laughlin, 1989, 1990, 1994a; study of consciousness can be found in Laughlin, McManus, & Shearer, 1983; Lee, the comprehensive treatment of practices 1980; Long, 1976; MacDonald, 1981). More- of shamanism cross-culturally by Michael over, it became clear that such experiences Winkelman (2000). In response to a myriad are both very common cross-culturally and of positions in anthropology and elsewhere fundamental to most of the world’s tradi- that view the states of consciousness evoked tional religious systems (Bourguignon, 1973, by practicing shamans as evidencing var- 1976). At the same time, it also became evi- ious forms of psychopathology – ranging P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

anthropology of consciousness 649

from schizophrenia to epilepsy to dissoci- evaluations of those states of conscious- ation to hysterical neuroses – Winkelman ness that do not conform to what is largely provides an extensive review of the positive an unquestioned definition of normalcy as interpersonal, neurophysiological, and psy- calibrated according to the standards of choneuroimmunological effects of shamanic what Winkelman terms a “modern rational practice, ritual, and the induction of bureaucratic consciousness” (2000, p. xi). In concomitant non-ordinary states of con- highlighting the pervasiveness of this bias sciousness. Phenomenologically speaking, in anthropological assessments of shamanic Winkelman argues that a key difference practice, Winkelman thus notably calls our between shamanic and pathological states attention to the extent to which scientific of consciousness is found in the control of, assumptions pertaining to consciousness are and intentional entry into, those states of still often deeply permeated by unexam- consciousness that are often associated with ined cultural assumptions that we must con- shamanic practice (2000,p.79). Moreover, stantly struggle to “bracket” in the context of Winkelman holds that shamans are able our ongoing research and theorizing. to distinguish clearly between experiences had in non-ordinary states of consciousness and those had in everyday waking life; the Cognitive Anthropology absence of this ability is generally held to be a key defining characteristic of many forms of Cognitive anthropological approaches to psychopathology, including schizophrenia. consciousness have their roots in the ethno- All of these insights serve as an important scientific work of such scholars as Charles corrective to those scholars who view the Frake, Ward Goodenough, Brent Berlin, and intentional alteration of consciousness in the Paul Kay, among others. These approaches service of shamanic healing to be evidence arose, much like those of their psychological of psychopathology. Although it is certainly counterparts, in the context of what is now true that we must be careful not to fall prey understood to have been a general backlash to an unthinking relativism when exploring against behaviorism in the social sciences the relationship among culture, conscious- occurring during the mid-1960s and early ness, and psychopathology (see Spiro, 2001), 1970s (see D’Andrade, 1995). Indeed, ethno- it is also true, as Winkelman points out, scientific approaches in anthropology sought that we must, in searching for any transcul- to place consciousness back in the center of tural criteria for assessing psychopathology, anthropological research through exploring be careful not to fall prey to our own cultur- how the conceptual categorization of expe- ally shaped biases. rience in differing cultures could shed light Following Laughlin, Manus, and D’Aquili on both the flexibility and limitations inher- (1992), Winkelman suggests that there is in ent in human cognitive capacities. anthropology an all too often unexamined By the late 1970s and early 1980s, cog- monophasic bias when it comes to investi- nitive anthropological approaches to con- gating states of consciousness that fall out- sciousness focused a great deal of attention side the boundaries of normal waking states. on the significance of “cultural schemas” – In Husserlian terms, investigators are lim- a term that can ultimately be traced back ited by their taken-for-granted adherence to Bartlett’s “schemata” (1932) – in giv- to their culturally conditioned “natural atti- ing definition to the contents of human tude” (Husserl 1950/1993) – an attitude consciousness; in many ways, this focus con- that tends to privilege what Schutz and tributed significantly to the development of Luckmann (1973) have termed the everyday cognitive science as an independent field life-world of the “wide-awake and normal of study (see D’Andrade, 1995). In this adult.” Because of this bias, social scientists, section we discuss three recent, influen- psychotherapists, and medical practition- tial approaches to consciousness in cogni- ers are often prone to dispense negative tive anthropology. Out of all the approaches P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

650 the cambridge handbook of consciousness

to consciousness thus far discussed, cog- tioning, he has also played an important nitive anthropology has perhaps been the role in generating interest in how computer most innovative in developing methodolo- models based on parallel distributed pro- gies for exploring the impact of culture on cessing theory may inform our understand- consciousness and the impact of conscious- ing of the relation between culture and ness on culture. Indeed, from triad testing human mental processes (D’Andrade 1984, to paired comparisons to multidimentional 1987). scaling to cultural consensus measures, cog- Inspired by a number of D’Andrade’s nitive anthropology has been at the fore- insights, two other influential figures in the front of the effort to ground anthropologi- field, Claudia Strauss and Naomi Quinn cal research in systematic methods for data (1998), have recently set out to outline collection and analysis. a cognitive theory of cultural meaning Perhaps the single most influential fig- that focuses squarely upon the relation- ure in contemporary cognitive anthropol- ship between culture and consciousness in ogy is Roy D’Andrade. D’Adrande (1981) the context of a discussion of internaliza- has addressed the relation between culture tion and the analysis of “intrapersonal cul- and consciousness most explicitly through ture.” Arguing that philosophical and theo- an exploration of a number of differences retical discussions of cultural meaning have he notes between symbolically driven com- all too often left unexamined the psycho- puter programs and human cognitive abil- logical mechanisms underpinning the “inter- ities. According to D’Andrade, the most nalization” of cultural forms, Strauss and apparent difference is that whereas com- Quinn advocate a theory of signification in puter programs acquire information about which meaning is understood to arise from the world through explicit symbolic encod- the mutual interaction of intrapersonal and ing of the unambiguous linear ordering of extrapersonal realms. A central goal of their step-by-step instructions, human children work is therefore to maintain that cultural in most societies are seldom exposed to forms must be understood in the context explicit instruction and acquire what is of those mental processes that were mostly often context-specific knowledge and com- excised from the anthropological purview petencies through processes of observation, with the pervasive and uncritical acceptance peripheral participation, modeling, and trial- of Geertz’s Ryleian thesis that cultural pro- and-error activity. cesses are found primarily in public symbols In terms of the relation between culture (see discussion above). Also like D’Andrade, and consciousness, D’Andrade notes that, Strauss and Quinn are interested in explor- not only do cultural processes shape the ing how insights from connectionist theory – contents and processes of human conscious- a theoretical orientation in cognitive sci- ness but also the innate capacities of human ence where the encoding of representation is consciousness shape and constrain cultural thought to reside in the differential distribu- processes (e.g., see George A. Miller, 1956, tion of connection weights in a parallel dis- for findings with regard to the limits of tributed processing system that is organized human memory). D’Andrade attributes the according to levels of input and output units coexistence of the “flexibility” and “shared- (see Quartz & Sejnowski, 1997; Rumelhart, ness” of cultural forms to the basic ways Hinton, & McClelland, 1986;Way1997)– in which humans acquire knowledge and can shed light upon the psychological and competencies, which he characterizes as a neurophysiological processes that underlie “curious combination of self-initiated yet the acquisition, transformation, and trans- other dependent leaning” (D’Andrade, 1981, mission of cultural forms. For Strauss and p. 188). Although D’Andrade is convinced Quinn, an advantage to connectionist mod- that the computer analogy so prevalent in els is that they provide a view of encoding cognitive science often sheds light upon cultural schemas that is structurally coher- human mental functioning primarily by act- ent while still being sensitive to modifica- ing as a valuable contrast to that function through experience. Moreover, these P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

anthropology of consciousness 651

approaches seem to provide a way to ther recent significant contribution to cog- account for how schematic knowledge may nitive anthropological approaches to con- be encoded, without the concomitant neces- sciousness is found in the work of Bradd sity of explicitly stated rules (1998,p.53; Shore (1996). Shore (1996) has built a cogni- see also the discussion of Bourdieu above). tive theory of culture that links the anthro- According to Strauss and Quinn, these pologist’s concern with social action and models are thus able to generate rule-like institutions with the psychologist’s concept responses without the need for ever hav- of mental models. Grounding his discus- ing been exposed to the “rule as such” sion in both biogenetic structural theory (1998,p.75). Because of this feature, they (see below) and on the “neural Darwinism” argue that connectionist models do not of Jean-Pierre Changeux (see Changeux, privilege language-based learning and that 1985), Shore demonstrates that no simplistic “[e]mbodied ideas can be represented as well essentialist, hard-wired brain approach will as (and perhaps even more readily than) account for the psychic unity phenomenon highly abstract ones” (1998,p.53). (Bastian’s idea discussed above that the Although ultimately endorsing connec- minds of all people, regardless of their race or tionist approaches, Strauss and Quinn culture, operate in the same way). Instead, also utilize insights from anthropological what is required is a more complex devel- research to point to some limits of these opmental approach of the sort that the bio- models. First, they highlight the fact that genetic structuralists, as well as Jean Piaget, these models are often overly dependent Changeux, and others, have taken. upon “supervised” learning (e.g., through a Shore (1996) calls for conceiving of cul- back propagation learning algorithm). That ture as a system of models – “as an extensive is, researchers are needed to guide the train- and heterogeneous collection of ‘models,’ ing sessions carefully for the proper pattern- models that exist both as public artifacts ‘in ing of weighted distributions to take effect. the world’ and as cognitive constructs ‘in the As Strauss and Quinn make clear, this model mind’ of members of the community” (1996, does not account for those numerous “real- p. 44). Public models and mental models are life” learning situations where there is sel- not the same thing. Public models are anal- dom a teacher around to correct a novice’s ogous to knowledge and take on such myr- every mistake. Second, they reveal that, iad forms as dances, houses, paintings, songs, although these systems seem to be quite stories, clothing, pottery, and tools. Men- proficient for dealing with “implicit” knowl- tal models on the other hand are more like edge acquisition and “procedural” learning, historically patterned, received schemas – they do not seem to account for those forms the socially transmitted mental structure of of learning that proceed through “explicit” knowledge in the heads of people. A cul- statements (1988,p.77). Third, these sys- tural model is like an internalized script that tems make no attempt to account for the role is both socially derived and dwells within the that emotion and affect play in learning and head of cultural participants. memory. Finally, Strauss and Quinn argue The brain, of course, is the organ of men- that connectionism is limited in terms of its tal models – more than that, the brain is tendency to posit too little innate knowledge a model generator. There is no end to the (1988,p.79). Their solution is to argue that, variety of models that the human brain can although it may be true that we need to pos- produce. According to Shore, the model- tulate a small number of cognitive univer- ing capacity of the brain has two aspects: sals to account for the ability to acquire cul- the personal and the conventional. We can turally specific forms, the innate structuring develop our own models, but many of our necessary for humans is probably not overly models are derived from our culture. These detailed. socially shared models are generated under Using extensive ethnographic case studies the distinct constraints of social interaction and revisiting Bastian’s notion of the psychic (see also Rappaport 1979, 1999, on his notion unity of humankind (see above), a fur- of “cognized environment”). This is a theme P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

652 the cambridge handbook of consciousness

we have encountered earlier in the think- structuralist group, founded by Charles ing of Durkheim (see above). Consciousness D. Laughlin, Eugene G. D’Aquili, and then is formed from internalized models that John McManus, constructed a culturally are either generated by the person during his informed neuroanthropology of conscious- or her development or inculcated through ness (D’Aquili, 1982, 1983; D’Aquili & the process of enculturation. Both kinds of Laughlin, 1975; D’Aquili, Laughlin, & models are present in any and all moments McManus, 1979; D’Aquili & Newberg, 1996; of consciousness. Laughlin, 1991, 1996a; Laughlin & Brady, 1978; Laughlin & D’Aquili, 1974; Laugh- lin & McManus 1995; Laughlin, McManus, Neuroanthropological Approaches & D’Aquili 1990; Laughlin et al., 1986; to Consciousness Laughlin & Richardson, 1986). This body of As we have seen, the suggestion that both literature set out to recognize and explain the variance and the structural commonality the centrality of the brain in mediating cul- among peoples’ consciousness and cultures tural transmission, social action, and subjec- are due in some fashion to an underlying psy- tive experience. This perspective holds that chophysical basis has been bouncing around any and all states, structures, and contents in anthropology from its beginnings in the of consciousness are mediated by systems 19th century. And as we noted above, Levi-´ of cells in the nervous system. These sys- Strauss importantly suggested that structural tems of cells are termed “models” and are commonalities in expressive culture could determined in their initial organization by be traced to the structure of the human the DNA – they are “neurognositic,” and brain. That said, a full-blown neuroanthro- in Count’s terms, very much a part of the pology had to await the inspiration of Earl W. human biogram. Neural models, being made Count (1899–1996) in the mid-20th century up of living cells in communication with (1958, 1973). In his book Being and Becom- each other, continue to grow and elabo- ing Human (1973), Count argued that each rate, guided in their growth by inherent pro- and every species inherits the generalized cesses of development, by personal experi- pattern of adaptation of its “anlage”; that ences, and by culture (Laughlin, 1991). What is, its precursor species. Count called the it means to “enculturate” (anthropologists’ total package of species-typical adaptation term for the process of instilling culture into patterns that species’ “biogram.” The her- the minds of children) a member of society itable architectonics (or structures) of the is in fact to influence the organization of the neuroendocrine system and the brain are neural systems mediating mental processing very much a part of every species’ biogram. and action in any particular instance or rela- Within each species’ biogram are the lim- tive to any particular object. its of possible organizations of consciousness, Later research in biogenetic structuralism which then develop relative to the adapta- focused particularly upon the problem of tional circumstances of the individual and how publicly available symbolism has the the group. Thus, for Count, it was appropri- power to evoke subjective states of con- ate to speak in terms of the human “brain- sciousness and to channel social understand- mind” as inextricably a part of the human ing and action. The group first considered biogram and as the one and only organ the universal properties of ritual and how rit- of human consciousness. Indeed, he argued uals with their embedded techniques, neu- that only when anthropology reached the roendocrine system “drivers,” and symbol- point where the brainmind became its cen- ism operate to control experience in all tral object of study could the discipline cultures (D’Aquili, Laughlin, & McManus, be considered a mature science (see also 1979; Laughlin et al., 1986). They defined Laughlin, McManus, & d’Aquili, 1990). what they call a culture’s “cycle of meaning” Following in the footsteps of Earl Count, by which sociocultural practices operate to the pioneering work of the biogenetic integrate the society’s worldview (primarily P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

anthropology of consciousness 653

carried around in people’s head) and the anthropology in the late 1960s and early to consciousness of individual group members mid-1970s. In addition to questioning many by expressing that worldview through sym- of the universalist assumptions about the bolic texts and ritual activities in such a development of human consciousness from way that experiences arise that are then a Piagetian paradigm, cultural psychological typically interpreted in terms of the world- approaches to consciousness in anthropol- view – producing a kind of feedback loop. ogy and psychology often also draw (both For example, ritually enacted aspects of explicitly and implicitly) from early 20th- a culture’s worldview are able to engen- century Soviet psychology and the writings der experiences in participants that in turn of Lev Vygotsky and his students. are interpreted by people as both verify- Influenced by the writings of Levy-Bruhl´ ing and enlivening the society’s collective (see above), Vygotsky (1930/1978) posited representations – engendering an effect that that human consciousness consists of two Durkheim (1912/1995) long before termed forms of mental functioning: (1) elementary “collective effervescence” (see above). functions (i.e., unlearned biological func- This group of theorists had an influence tions, such as hunger) and (2) higher mental on Victor Turner’s later interest in explor- functions (i.e., language, memory, abstrac- ing the interconnection of culture, mind, tion). In Vygotsky’s estimation, what served and brain (see Turner, 1983). Other scholars to mediate the transition from elementary making important contributions to a neu- to higher mental functioning in the context roanthropological approach to conscious- of a child’s development are cultural sym- ness studies include Michael Winkelman bols/tools. Not unlike Geertz (see above), (1986, 1994, 2000) on the anthropology of Vygotsky understood these symbols/tools to consciousness and the neurophenomenol- be material manifestations of past genera- ogy of shamanism; Stephen Reyna (2002) tions of human activity and mentation, what on the brain, symbolism, and hermeneutics; some recent theorists have aptly termed and Dean Falk (1992) on the evolution “artifacts” (Cole, 1996). For Vygotsky, of all of the brain and consciousness. Warren of the artifacts that play a role in enabling the TenHouten (1978–79) has also contributed development of higher mental functions in much to the study on the brain and different humans, language is the most significant. As modes of thought cross-culturally. Finally, a result, the focus of much of Vygotskian- there is recent work exploring biocultural inspired research in cultural psychology is approaches to the emotions (see Hinton, exploring “language and activity-in-context” 1999), as well as research in evolution- as the means to understanding the interpen- ary psychology (see Barkow, 1989; Barkow, etration of culture and consciousness (see Cosmides, & Tooby, 1992; Pinker, 1997; Cole, 1996). Sperber, 1996; Sperber & Block, 2002) that Although the popularizing of Vygotsky’s holds that the structure of human conscious- writings in North American can largely be ness is at least partially informed by adap- attributed to the influential writings and tive modular neural structures that mediated research of such eminent psychologists as cognition during the long centuries of the Michael Cole, James Wertsch, Carl Ratner, Paleolithic (the Old Stone Age). Jerome Bruner, and others, one of the most influential proponents of a cultural psychological approach to consciousness in anthro- Cultural Psychology pology is Richard Shweder (1984). Although not relying as explicitly upon Soviet psy- Cultural psychological approaches to con- chology as did his peers in the discipline sciousness in anthropology largely emerged of psychology, Shweder has been a cham- as a form of critique of the prevalent appli- pion of a pluralistic contextualism in devel- cation of Piagetian research paradigms in oping his understanding of the intersection cross-cultural psychology and psychological among culture, activity, and consciousness P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

654 the cambridge handbook of consciousness

in human societies. Moreover, like his psy- behavior, which is forgotten all too eas- chological colleagues, he has been staunchly ily by experimental, cognitive, and cross- anti-Piagetian in his critique of Kohlberg’s cultural psychologists. Unfortunately, as is theories of moral development from a cross- typical of many of the more relativistic cultural framework. approaches to culture and consciousness Shweder (1984) predicates his version in anthropology and elsewhere, Shweder of cultural psychology on an understand- tends to overlook the very important dis- ing of consciousness as organized accord- tinction between structure (that is, the coming to a trichotomy of mental operations plexity of organization of information) and classified according to rational, irrational, content (the information being structured). and non-rational distinctions. In line with Moreover, he downplays ethnographic evi- a long history of debates in anthropology dence of social institutions designed with over putative differences between modern the developmental levels of maturation of and primitive mentalities (see Levy-Bruhl´ structure in mind. For instance, there are above), Shweder importantly emphasizes societies in which the telling of myths and that this trichotomy points to the coequal other sacred narratives is done with sensi- integration of these three forms of con- tivity to the maturity of the audience (see sciousness in all cultures, regardless of Jorgensen, 1980, on the Telefolmin of Papua their degree of “modernization.” To this New Guinea; Barth, 1975, on the Baktaman end, Shweder’s understanding of conscious- of New Guinea; Peters, 1982, on the Tamang ness is representative of an important shift shamans of Nepal; Reichel-Dolmatoff, 1971, in anthropological thinking concerning the on the Tukano of Amazonia; Griaule, 1965, structure and function of modern mental- covering the Dogon of West Africa; and ities. According to this perspective, mod- Beyer, 1973, on the instruction of Tibetan ern mentalities have come to be charac- lamas). terized as equally suffused with irrational and non-rational currents as primitive men- Psychocultural Anthropology talities were once held to be. Shweder’s discussion of the romantic contributions Another approach to consciousness in of “semantically induced memory drifts,” anthropology can be broadly termed “psy- “paradigms,” “cultural frames,” and “perfor- chocultural.” Whereas cultural psychologists matives” illustrates this point. Shweder is have focused much of their attention upon thus highly critical of what he characterizes activity and social context as a means to to be Piaget’s exaggerated stress upon self- understand the organization and functioning constructed knowledge, personal invention, of consciousness cross-culturally, psychocul- and predictable progressive development. tural approaches to consciousness in anthro- Shweder maintains that, on the contrary, pology (see Briggs 1998; Crapanzano, 1986, what cultural psychologists have discovered 1992; Levy, 1973, 1984; Hollan, 2000, 2001; is the significance of other-dependent learn- Obeyesekere, 1981, 1990; Parish, 1994; Spiro, ing, social interaction, and the internaliza- 1987, 1994, 1997) have generally empha- tion of collective representations in under- sized not only the significance of expand- standing the patterning and functioning of ing our understanding of the complexity of consciousness in any given culture or social the social matrix in defining consciousness group. but also how the multivariate nature of con- Once again we see the tension between sciousness influences social interaction and structuralist and relativist points of view the internalization of cultural meaning (see at play in the anthropology of conscious- Throop, 2003b). ness. Shweder (1984, 1991, 2003) can be At the heart of many of these approaches praised for doing what anthropologists have is an attempt to explore the articulation of always done best – that is, insisting upon personal and cultural processes in the cru- the cultural influence upon learning and cible of consciousness while also recognizing P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

anthropology of consciousness 655

that those contents, processes, and struc- account the interleaving of psychodynamic tures that scientists have glossed with the and cultural processes (e.g., the relation term “consciousness” actually consist of mul- between defense mechanisms and replica- tiple modalities that range from explicit con- tion of anxiety-provoking cultural ideolo- ceptual content to non-conceptual feelings gies) if one is going to adequately understand and images. In viewing consciousness from both the impact of culture on consciousness the perspective of multiple modalities that and the impact of consciousness on culture. may be co-present in any given moment He asserts that it is only in coming to under- in the stream of consciousness, the scholars stand the role of “precultural” (e.g., biolog- approaching the problem of culture and con- ical and social) contents and structures of sciousness from a psychocultural perspective consciousness that one can properly under- have tended to also highlight the significance stand how proclivities, dispositions, moti- of discrepancies and conflicts that may arise vations, and susceptibilities emerge for the between differing modes of consciousness, acquisition and “internalization” of specific each of which may in turn be affected dif- kinds of knowledge. ferently by cultural forms. Spiro contends that theories of cultural Much of this work is explicitly or implic- reproduction that do not address the interre- itly Freudian in its approach, inasmuch as it lation between culture and consciousness in relies upon a view of the mind as structured the context of psychological preadaptation according to non-conscious, preconscious, are at a significant disadvantage in explain- and conscious elements. In some of the most ing why it is that some cultural proposi- recent psychocultural approaches to con- tions are often still internalized by cultural sciousness, however, theorists have drawn participants. For example, Spiro interprets increasingly from insights formulated in rela- the “Ideology of the Superior Male” and tional psychoanalysis (e.g., Mitchell, 1988) the “Ideology of the Dangerous Female” in and dissociation theory (e.g., Stern, 1997). Burma as cultural resources that may be What all of these approaches share is the utilized by social actors in the construc- idea that culture can differentially influence tion of “culturally constituted defense mech- non-conscious, preconscious, and conscious anisms.” As he notes, although these two states and that there may be important ideologies are empirically false and anxi- divergences among feeling, representation, ety provoking, they are still readily internal- emotion, and motivation at these differing ized by Burmese men. According to Spiro, phases of consciousness. Methodologically, these ideologies are perpetuated in Burma, psychocultural approaches to consciousness as well as in other cultures, because they are in many instances also draw from psycho- ultimately beneficial both to the individu- analytic approaches to clinical interviewing als who internalize them and to the culture and as a result have often relied upon in- that constitutes them. Individuals who inter- depth, open ended, person-centered internalize these belief systems are able to par- viewing strategies (see Levy & Hollan, 1998). tially fulfill what would otherwise be frus- Melford Spiro (e.g., 1987) is one of the trated unconscious wishes and desires. At most influential figures in this tradition, and the same time, the culture of these individ- his writings on the relations between col- uals facilitates the reproduction of cultural lective representations and mental represen- forms by recruiting intrapsychic conflict in tations within the context of a Freudian the service of cultural propositions and in the theory of mind have for decades been con- process motivates the enactment of socially sidered essential reading for anthropolo- sanctioned roles. gists interested in the intersection of per- Robert Levy (1973, 1984) is another sonal and cultural aspects of consciousness. important figure in this tradition who Spiro’s (1997; see also Spiro,1994) most has written much about culture’s role recent formulation of this approach is based in differentially patterning the cognitive on the assumption that one must take into saliency of various states and processes of P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

656 the cambridge handbook of consciousness

consciousness. Although he explores many processes subserving the maintenance, trans- different aspects of consciousness from mission, and creation of cultural meaning. rationality to sensation, Levy is most often Through developing a complex model of referenced with regard to his discussion of the relationship between the conscious and the cultural patterning of emotion. He has unconscious dimensions of the mind, which coined the distinction between “hyper- and he stresses are seldom organized into dis- hypo-cognized” emotions, which serves to crete or independent realms but instead draw attention to the role that culture plays are held to coexist in ever-shifting and in differentially articulating patterns of fluid processes of interpenetration, Hol- attention, conceptualization, and sensation lan suggests that researchers interested in in the structuring of emotional experi- understanding the relationship between cul- ence (see Throop, 2005). Hypercognized ture and consciousness must explore how emotions are those that are culturally these various intrapsychic, interpsychic, and elaborated and thus highly salient, and as extrapsycic processes contribute to both a consequence they tend to be centers of the standardization and “personalization of recurrent attentional focus for individual meaning.” culture bearers. In contrast, due to their lack of culturally infused conceptual elabora- Anthropology of the Senses tion, hypocognized emotions tend to resist or defy explicit forms of representation Influenced by various semiotic and phe- because they tend not to evoke the same nomenological traditions, the cultural culturally attuned attentional focus. Levy patterning of olfactory, haptic, and auditory formulated these insights through conduct- sensations took on a renewed significance ing in-depth “person-centered” interviews during the mid-1970s and early 1980sin with informants in Tahiti, where he dis- the work of Alfred Gell (1977), Valentine covered, for instance, that there seemed Daniel (1983/1991), Steven Feld (1984), and to be “no unambiguous terms which rep- Paul Stoller (1984). Indeed, the wave of resent the concepts of sadness, longing, or interest in non-visual sensation during this loneliness. ...[In this light p]eople would period can be tied to more general trends name their condition, where I supposed in anthropology concerning the exploration that the context called for ‘sadness’ or of culture in relation to non-conceptual ‘depression,’ as ‘feeling troubled’ ...as ‘not modes of human existence. Here Bourdieu’s feeling an inner push’ as ‘feeling heavy’ writings on the concept of habitus as a as ‘feeling fatigued’ and a variety of other culturally inculcated generative structure terms all referring to a generally troubled or shaping fields of perception, motivation, subdued body state” (1973,p.305). judgment, and action (see above); a bur- Building upon the work of Spiro, Levy, geoning interest in the anthropology of and others, Douglas Hollan (2000) has emotion that largely set out to question recently argued that any adequate the- Cartesian-inspired distinctions between ory of culture must thus be predicated mind-body and thought-feeling (Lutz, upon an understanding of the relations 1988; Rosaldo, 1984); critiques of visualist between cultural processes and “the flu- and representationalist biases in modern idity and complexity of the psychological ethnography (Tyler, 1984); and a movement states that underlie” them. In line with psy- toward developing theories of practice in chocultural approaches generally in anthro- the discipline as a whole (Ortner 1984) pology, Hollan’s formulation pivots on the can all be seen as arising out of the same recognition of how both the fluctuating general Zeitgeist of resistance to what a nature of social interaction (self – “not me”- growing number of thinkers perceived to object relations) and the transitional nature be the overly intellectualist orientations of of consciousness (conscious-preconscious- Levi-Straussian´ structuralism and Geertzian unconscious relations) mutually affect those textualism (see above). P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

anthropology of consciousness 657

The early anthropological writings on cul- window upon the world and are thus in ture and sensation have played an important some way “precultural”(1997,p.402); (2) role in setting the stage for the great efflo- that all cultures organize their sensory orders rescence of work on sensation that emerged in terms of a structure similar to that of in the discipline during the 1990s. Sensa- the West with its bias toward visualism; and tion has clearly emerged as a significant area (3) the deeply ingrained “prejudice against of theorizing and research in the discipline smell, taste, and touch as ‘animal senses’” as a whole and now involves such ques- (1997,p.405) and the tendency to believe tions as how sensory modalities are impli- that there is little variation within sen- cated in healing practices (Csordas, 1994; sory modes cross-culturally. Instead, Classen Desjarlais, 1992; Landerman & Roseman, and her colleagues hold that it is time for 1996), cultural variations in senses of space anthropologists to recognize that culture not and place (Feld & Basso, 1996), the influ- only patterns the relative value assigned to ence of culture on ordering various “ratios of each sensory modality but also that within sense”(Classen, 1997; Howes, 1991; Synnott, each modality there are various experien- 1993), and the relation between sensation tial properties that can serve to fuel cul- and moral sensibilities (Geurts, 2002). In tural elaboration. For instance, the “vision addition, there is work exploring the cul- which is deemed rational and analytic in the tural patterning of varying somatic modes West ...may be associated with irrational- of attention (Csordas, 1993; Sobo, 1996; ity in another society” (1997,p.404). Or Throop, 2003a); the importance of recogniz- vision may very well be associated primar- ing differing sensory orders in the context of ily with aesthetic values that have little to ethnographic research (Stoller, 1989, 1997); do with the acquisition of “objective” forms the cultural elaboration of specific sensory of knowledge about the external world. modalities, such as audition, olfaction, or Central to the methodology proposed taste (Chuengsatiansup, 1999; Rasmussen, by Howes and Classen (1991) is a phe- 1999); and renewed interest in debates over nomenologically grounded effort to pene- culture and color perception (Goodwin, trate our taken-for-granted sensory attune- 1997; Hardin & Maffi, 1997). ments toward the world in an attempt to Although a great deal of important work better approximate the sensory orders most has emerged out of this tradition, there is valued in a particular culture under study. little doubt that the work of the Concor- To discover the various ways that a particu- dia Group – consisting of David Howes, lar culture sets out to differentially “empha- Constance Classen, and Anthony Synnott – size or repress” different varieties of sensory has provided much of the impetus behind experience, Howes and Classen suggest that the renewed interest in culture and sensa- it is first necessary for the researcher to work tion. In a series of books and articles, these to “overcome, to the extent possible, his or three scholars have argued for the histori- her own sensory biases” (1991,p.260). This cal rootedness of the senses. Anthropologists effort begins with working to evaluate and are being encouraged to move away from discover one’s own sensory predispositions. the strictly physicalist assumptions of sen- To this end, Howes and Classen suggest that sory psychology to a stance that not only researchers follow an exercise devised by explores the effect of cultural assumptions Galton that entails describing a past event in and values on the functioning of any one par- one’s life and noting which sensory modali- ticular sensory modality but also takes into ties are most heavily relied upon in recount- consideration “inter-sensory relations.” ing the details of that event. The second step According to Classen (1997), the research in this methodology consists of “training one- of the Concordia Group has challenged self to be sensitive to a multiplicity of sensory three prevalent assumptions in anthropo- expressions” (1991,p.260). Here the authors logical writing on sensation: (1) that the recommend following a basic phenomeno- senses in some way provide a transparent logical principle that is based on “disengaging P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

658 the cambridge handbook of consciousness

one’s attention from the object itself so as to or utilized by parents when raising their focus on how each of the sensory proper- children. This exploration can also be aug- ties would impinge on one’s consciousness mented by investigations that look at how were they not filtered in any way” (1991, the natural and built environment, rituals, p. 260). This disengagement is then comple- mythology, cosmology, and communication mented with cultivating the ability to be “of media potentiate or dispotentiate the vari- two sensoria about things”; that is, to be “able ous senses. to operate with complete awareness in two perceptual systems or sensory orders simul- Cultural Phenomenological and taneously (the sensory order of one’s own Neurophenomenological Approaches culture and that of the culture studied” and constantly comparing notes (1991,p.260). As evident in the methodological orien- Having worked to effectively bracket tation of sensorial anthropology outlined one’s “natural sensory attitude,” Howes and above, anthropology has in recent years dis- Classen then argue that there are several covered the significance of phenomenology, different avenues to explore in attempt- blending methods derived from various phe- ing to outline the sensory order of any nomenological approaches into the ethno- given culture. They advise researchers to graphic fieldwork “toolkit” (see Csordas begin by investigating the existing words 1990, 1993, 1994; Desjarlais, 1997, 2003; for the different senses, exploring which Jackson, 1989, 1996; Throop 2002, 2003a, of the various sensory modalities have the Throop & Murphy, 2002). Methodologically, largest vocabulary allotted to them, and a phenomenological ethnography requires determining how the various senses are used that the researcher focus on the experiences in metaphor and other forms of expressive of individual actors immersed in social inter- speech. This effort can be complemented action (see Jackson, 1996; Chapter 4). The by an examination of cultural aesthetics and shift is thus toward examining the world the realm of material artifacts. Here cul- of everyday experience and how individu- tural ideas regarding beauty can be eval- als in their everyday interactions are them- uated in terms of the relative contribu- selves conscious of their own lifeworlds. In tion of each of the senses. Moreover, an this approach, the stories and experiences investigator can examine how various cul- of individual social actors in the context of tural artifacts are used to represent or evoke everydayness form the field of investigation the senses of the creator/user/admirer. For for the researcher. This marks, of course, at instance, Howes and Classen (1991) assert least a partial return to the fieldwork strategy that a “Tsimshian ‘wraparound’ representa- of Franz Boas (see above). tion of Bear corresponds to the experience of An excellent example of phenomeno- sound, which also envelops and surrounds logical anthropology’s contributions to the one. ...[in other words the] ‘ear-minded’ study of consciousness and culture is found Tsimshian would thus seem to transpose in the influential work of Thomas Csordas visual imagery into auditory imagery in their (1990, 1993, 1994, 2002). A main thrust of visual art” (1991,p.265). Csordas’s writings (1990) is that anthropol- In addition to these two strategies, Howes ogy as a discipline is in the processes of shift- and Classen point out that the decoration ing toward a paradigm of embodiment. He and alteration of the form of the body suggests that such a view does not entail that can further serve as an effective means to all cultures should be understood to “have explore a culture’s sensory order. For exam- the same structures as bodily experience, ple, which of the sense organs are high- but that embodied experience is the start- lighted by decoration? Sensory orders may ing point for analyzing human participation also be explored through close examina- in a cultural world” (1993,p.135). To this tion of child-rearing practices where careful end, a key phenomenological insight (bor- attention is paid to which senses are stressed rowed from Merleau-Ponty, 1962) advanced P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

anthropology of consciousness 659

by Csordas (1990) is that perception does consciousness, however, has been to advance not begin, but, rather ends in objects. That the notion of “somatic modes of attention,” is, perceptual experience is understood to be which he defines as those “culturally elab- an active process of constitution whereby orated ways of attending to and with one’s otherwise ambiguous stimuli are able to body in surrounding that include the embod- be articulated into personally and culturally ied presence of others” (1993,p.138). In meaningful forms. Csordas’s cultural phe- other words, Csordas wishes to highlight the nomenology thus sets out to explore those various ways that culture can pattern indi- cultural processes implicated in the consti- vidual attention to bodily sensation in rela- tution of those objects confronting a social tion to both perception and motility. As he actor’s consciousness moment by moment. explains, to “attend to a bodily sensation According to Csordas, embodied expe- is not to attend to the body as an isolated rience as a mode of being-in-the-world object, but to attend to the body’s situation becomes a complementary perspective to, in the world” (1993,p.138). This approach what has historically been, an all too domi- should not be mistaken for a sensory-based nant textualist and representationalist per- empiricism because Csordas argues that spective in anthropological theorizing. A any attempt to compare the sensation of major shortcoming of such perspectives, heat, for example, in the context of heal- Csordas argues, is that they presume “an ing with the experience of heat associated unbridgeable gulf between language and with blushing fails to recognize the extent experience” that is “predicated on the notion to which a synthesis exists between inter- that language can only be about itself” (1994, pretive and experiential realms. Examples p. xii). On the contrary, Csordas maintains provided by Csordas drawn from his field that cultural phenomenological investiga- work among Catholic Charismatic groups tions into the dynamics and structures of include the interpretation of sensations of consciousness evidence the fact that “lan- “queasiness” as “indicating the activity of guage is not only a form of observable behav- evil spirits, and an unexpected sneeze or ior but a medium of intersubjectivity, so that yawn. ...[indicating] that a spirit is passing it is fair to say that language gives us authen- out of the supplicant through the healer” tic access to experience” (1994, :xii). (1993,p.141), This assertion is based on Csordas’ Also arguing for an embodied under- assumption, following Merleau-Ponty’s phe- standing of consciousness, we have incor- nomenology (1962) and the practice the- porated the earlier work of the biogenetic ory of Bourdieu (see above), that another structuralist group (see above) in develop- person’s intentions, emotions, thoughts, and ing a phenomenology that is grounded in desires are manifest preobjectively as an both the neurosciences and cultural anthro- intersubjectively accessible “co-presence” pology – what we term a “cultural neurophe- that can be immediately grasped “insofar nomenology.” Our approach seeks to wed as we share the same habitus.” By relying the application of a trained phenomenology upon Bourdieu’s concept of habitus – which that controls for cultural variation in per- serves to explain how it is that perceptual, ception and interpretation and is yet still affective, motivational, and evaluative ori- attentive to the personal and cultural influ- entations can become shared by individuals ences on everyday experience with informa- whose respective consciousnesses have been tion from the neurosciences about how the conditioned in similar social, economic, and organ of experience – the brain – is struc- cultural environments – Csordas argues that tured and functions. he is able to account for a dispositional reso- In previous writings we have suggested nance between actors that both precedes and that the influence of culture on conscious- informs the structure and use of language. ness is differentially articulated with differ- One of Csordas’s (1993) most important ing strata of the human nervous system. We contributions to anthropological studies of have also argued that the phenomenological P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

660 the cambridge handbook of consciousness

and neurophysiological structures underly- Conclusion ing many apparently diverse cultural institutions – institutions such as ceremonial rit- Anthropological approaches to conscious- ual, cosmological worldviews, myth, and the ness have traditionally been based on the fact almost ubiquitous drive among the world’s that anthropologists, like all human beings peoples to alter their states of conscious- on the planet, have been inclined by the very ness – may operate to “true-up” the life- structure of their own consciousness and its world of individual members of society in intentional nature to focus their attention on relation to an extra-mental reality within objects: thought objects, perceptual objects, which each person and group must adapt sensory objects, physical objects, cultural (Laughlin & Throop, 1999, 2001; Throop & objects, etc. In the act of focusing on these Laughlin, 2002). We also incorporate the objects – the contents of consciousness – notion of a “cycle of meaning” (see above) other processes and structures available for to explore how a culture’s myth and rit- scrutiny within the stream of consciousness ual purvey a given culture’s worldview in are necessarily occluded. Thus anthropolo- such a powerful way that they are able to gists have played an important role in high- engender experiences in individuals that in lighting the striking variation that is appar- turn verify and enliven the society’s collec- ent both cross-culturally and interpersonally tive representations – what Durkheim once in terms of the constitution and organization termed “collective effervescence.” of comparable objects. Perhaps most significant for this review Although it is true that some scholars, is our attempt to advance a corrective to like Levi-Strauss, have examined how cul- overly cognitivist and propositionally biased tural objects (i.e., texts, artifacts, behaviors, accounts of consciousness and its contents symbols) may evidence patterns that res- by exploring what we hold to be the multi- onate across cultures, thus working back plex nature of human consciousness. That from cultural product to the shared struc- is, we hold that consciousness is differen- tures of human consciousness, the over- tially organized according to the concep- all tendency in anthropology has been to tual and abstract contents of linguistically focus on content in terms of its variation mediated thought and the imagistic, percep- and fluctuation. Indeed, this tendency has tual, and somatosensory contents of presen- served anthropology well inasmuch as it has tational forms of awareness (see also Hunt, frequently forced scientists in other disci- 1995; Winkelman, 2000). For instance, in plines who are interested in exploring invari- an article investigating the cultural pattern- ant aspects of consciousness to take notice ing of human emotional experience (Lauglin of cross-cultural variation. Doing so injects & Throop, 1999), we suggest that differ- some humility into their otherwise sweep- ent phenomenologically accessible variants ing statements concerning the universal- of emotional awareness are grounded in the ity of these structures; a universality that differing structural and functional strata of all too often has arisen from the reifica- the human brain. The significance of these tion of unscrutinized cultural assumptions insights for the study of culture and con- about the nature of consciousness held by sciousness is tied to the fact that the var- a particular thinker whole cloth onto all of ious neurophysiological structures mediat- humanity. ing various conscious modalities may be That said, anthropology has also demon- affected differently by cultural resources and strated an equally long history of awareness as such may provide researchers with a and concern for the similarity, psychic unity, way to account for both interpsychic vari- and shared aspects of consciousness across ation and transcultural similarities in the cultural boundaries. Accordingly, one of the structuring of subjective experience cross- most important points to take away from culturally. this review of anthropological approaches to P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

anthropology of consciousness 661

consciousness is that it seems far too sim- behaviors and internal states, these interpre- plistic to assume a unidirectional flow of tations are [just] not acceptable as topics of causation from the external organization of talk” (1984:290). socioeconomically determined activity pat- There are a number of other key insights terns and cultural symbols to the forma- that have arisen from the anthropologi- tion of contents and structures of conscious- cal study of consciousness: (1) the abun- ness, just as it is too simplistic to assume dant evidence in the ethnographic record that social and cultural forms are merely for the interdependent and dynamic rela- the externalization of pregiven structures tionship among collective (cultural) knowl- of consciousness. The truth of the matter edge, enculturation, individual subjectivity, seems to lie somewhere between these two and transculturally shared somatic structures long-standing polar perspectives concern- in the constitution of consciousness; (2) the ing the relation between culture and con- suggestion that the drive to alter individ- sciousness. Indeed, the truth would appear ual states of consciousness by socially pre- to lie in attempting to develop perspectives scribed means is fundamental to both heal- within an anthropological frame that focus ing, health and producing a worldview that equally on exploring content and structure, better accords with reality; (3) the contents object and process, or what Edmund Husserl of consciousness are both structured by the termed noema (act) and noesis (content; see inherent organization of the body, and plas- Stroker,¨ 1993 and Chapter 2). tic and adaptively responsive to environmen- A second significant insight that read- tal, personal, and cultural influences; (4) the ers should take away from this discussion qualitative research devoted to the study of is the fact that anthropologists have long consciousness as it is manifest in the con- sought to highlight in their research how cul- text of everyday social interaction is cru- ture patterns the responses that social scien- cial to developing a more accurate under- tists receive from their informants. Here the standing of the structures and contents of questions that anthropologists have asked human consciousness cross-culturally; and concern the cultural frames of reference that (5) there still remains in contemporary sci- are embedded in what scientists perceive to entific and philosophical theorizing a num- be “culture-free” standardized tests used to ber of unexamined assumptions about con- gain insight into the structure and content sciousness that can be traced to particu- of consciousness. For instance, in the context larized cultural orientations to the relation of cross-cultural research exploring the cul- between mind and body. tural patterning of cognitive abilities, anthro- Considering consciousness from an pological research points to the importance anthropological point of view has the advan- of asking whether a refusal to answer a syl- tage of grounding consciousness research logism, which informants may attribute to in the naturalistic everydayness of human an “unfamiliarity” with the topic at hand, experience, a perspective all too easily lost indicates differences in cognitive functioning in the laboratory. We are often reminded or differences in communicative practices. of the story told by the great comparative For example, in their important work on psychologist, Emil Menzel (1967), who language socialization, Ochs and Schieffelin spent years working with olive baboons (1984) assert that, although Schieffelin’s in the laboratory, and had pretty much Kaluli informants often refused to speculate concluded they were a dull and unintel- about another individual’s thoughts, inten- ligent species of primate. Then chance tions, and feelings, this observation should took him to Kenya where he took a safari not simply be interpreted as an indica- out among olive baboon troops in their tion that the Kaluli have no theory of natural environment. He suddenly realized mind. Quite the contrary, “Kaluli obviously that his very erroneous picture of baboon interpret and assess one another’s available intelligence was based upon experimental P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

662 the cambridge handbook of consciousness

questions that had little to do with the Boas, F. (1974). On alternating sounds. In G. W. natural environment in which they evolved. Stocking, Jr. (Ed.), A Franz Boas reader: The Ever after Menzel argued very elegantly shaping of American anthropology, 1883–1911 for the necessity of blending experimental (pp. 72–76). Chicago: University of Chicago 1889 with naturalistic research. Much the same Press. (Original work published ) situation confronts consciousness research Boas, F. (1910/1938). The mind of primitive man. today, and the main role of anthropology New York: The Free Press. is to broaden our collective perspective by Bock, P. K. (1999). Rethinking psychological bringing the naturalistic framework back anthropology: Continuity and change in the study into our scope of inquiry. of human action. Prospect Heights, Illinois: Waveland Press, Inc. Boucouvalas, M. (1980). Transpersonal psy- References chology: A working outline of the field. Jour- nal of Transpersonal Psychology, 12(1), 37–46. 1977 Aberle, D. (1966). The peyote religion among the Bourdieu, P. ( ). Outline of a theory of practice. Navajo. Chicago: Aldine. Cambridge: Cambridge University Press. 1985 Adam, B. (1994). Perceptions of time. In T. Ingold Bourdieu, P. ( ). The genesis of the concepts 2 11 24 (Ed.), Companion encyclopedia of anthropology of habitus and field. Sociocriticism, 2 ( ): – . (pp. 503–526). London: Routledge. Bourdieu, P. (1990). Logic of practice. Stanford: Antze, P., & Lambek, M. (Eds.). (1996). Tense Stanford University Press. past: Cultural essays in trauma and memory. Bourdieu, P. (2000). Pascalian meditations. London: Routledge. Stanford: Stanford University Press. Baars, B. J. (1997). In the theater of consciousness: Bourguignon, E. (1973). Religion, altered states The workspace of the mind. New York: Oxford of consciousness, and social change. Columbus, University Press. OH: Ohio State University Press. Barkow, J. (1989). Darwin, sex and status: Bio- Bourguignon, E. (1976). Possession. San Francisco: logical approaches to mind and culture. Toronto: Chandler and Sharp. University of Toronto Press. Braud, W., & Anderson, R. (Eds.). (1998). Barkow, J., Cosmides, L., & Tooby, J. (Eds). Transpersonal research methods for the social sci- (1992). The adapted mind: Evolutionary psy- ences. Thousand Oaks, CA: SAGE Publications. chology and the generation of culture. Oxford: Briggs, J. (1998). Inuit Morality Play. New Haven: Oxford University Press. Yale University Press. Barnouw, V. (1946). Paranormal phenomena and Bruner, J. (1951). Personality dynamics and the culture. Journal of the American Society for Psy- process of perceiving. In R. V. Blake, & G. V. chical Research, 40, 2–21. Ramsey (Eds.), Perception: An approach to per- Barrett, H. C. (2004). Descent versus design in sonality (pp. 121–147). New York: Ronald Press. Shuar children’s reasoning about animals. Jour- Campbell, R. L., & Staniford, P. S. (1978). nal of Cognition and Culture, 4, 25–50. Transpersonal anthropology. Phoenix: The Jour- Barth, F. (1975). Ritual knowledge among the Bak- nal of Transpersonal Anthropology, 2 (1), 28–40. taman of New Guinea. New Haven: Yale Uni- Chagnon, N. (1977). Yanomamo: The fierce people. versity Press. New York: Holt, Rinehart and Winston. Bartlett, F. C. (1932). Remembering: A study in Changeux, J.-P.(1985). Neuronal man: The biology experimental and social psychology. Cambridge: of mind. Oxford: Oxford University Press. Cambridge University Press. Chuengsatiansup, K. (1999). Sense, symbol, and Bastian, A. (1881). Vorgeschichte der ethnologie. soma: Illness experience in the soundscape of Berlin. everyday life. Culture, Medicine, and Psychiatry, Bastian, A. (1887). Die Welt in ibren spiegelungen 23(3), 273–301. unter dem wandel des völkergedankens. Berlin. Classen, C. (1997). Foundation for an anthropol- Beyer, S. (1973). The cult of Tara: Magic and ritual ogy of the senses. International Social Science in Tibet. City, State: Publisher. Journal, 153, 401–412. Bharati, A. (1975). The Tantric tradition. New Cole, M. (1996). Cultural psychology. Cambridge, York: Samuel Weiser. MA: Harvard University Press. P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

anthropology of consciousness 663

Conklin, H. (1969). Lexicographical treatment of D’Aquili, E. G. (1983). The myth-ritual complex: folk taxonomies. In S. Tyler (Ed.), Cognitive A biogenetic structural analysis. Zygon, 18(3), anthropology. New York: Holt, Rinehart and 247–269. Winston. D’Aquili, E. G., & Laughlin, C. D. (1975). The Coult, A. (1977). Psychedelic anthropology: The biopsychological determinants of religious rit- study of man through the manifestation of the ual behavior. Zygon, 10(1), 32–58. mind. Philadelphia: Dorrance & Company. D’Aquili, E. G., Laughlin, C. D., & McManus, Count, E. W. (1958). The biological basis of J. (1979). The spectrum of ritual. New York: human sociality. American Anthropologist, 60, Columbia University Press. 1049–1085. D’Aquili, E. G., & Newberg, A. B. (1996). Con- Count, E. W. (1973). Being and becoming human. sciousness and the machine. Zygon, 31(2), 235– New York: Van Nostrand Reinhold. 252. Count, E. W. (1990). Interview with Earl W. Desjarlais, R. (1992). Body and emotion. Count. Neuroanthropology Network Newsletter, Philadelphia: University of Pennsylvania 3(1), 5–8. Press. Crapanzano, V. (1986). Tuhami: Portrait of a Desjarlais, R. (1997). Shelter blues. Philadelphia: Moroccan. Chicago: University of Chicago University of Pennsylvania Press. Press. Desjarlais, R. (2003). Sensory biographies. Crapanzano, V. (1992). Herme’s dilemma and Berkeley: University of California Press. Hamlet’s desire; On the epistomology of interpre- Dissanayake, E. (1992). Homo aestheticus: Where tation. Cambridge: Harvard University Press. art comes from and why. Seattle: University of Csordas, T. J. (1990). Embodiment as a paradigm Washington Press. for anthropology. Ethos, 18, 5–47. Dobkin de Rios, M. (1984). Hallucinogens: Cross- Csordas, T. J. (1993). Somatic modes of attention. cultural perspectives. Albuquerque: University Cultural Anthropology, 8, 135–156. of New Mexico Press. Csordas, T. J. (1994). The sacred self: A cul- Dobkin de Rios, M., & Winkelman, M. (Eds.). tural phenomenology of charismatic healing. Los (1989). Shamanism and altered states of con- Angeles: University of California Press sciousness. Special issue of the Journal of Psy- choactive Drugs, 1 Csordas, T. J. (2002). Body/meaning/healing. New 21( ). York: Palgrave. Donald, M. (1991). Origins of the modern mind D’Andrade, R. (1981). The cultural part of cogni- . Cambridge, MA: Harvard University Press. tion. Cognitive Science, 5, 179–195. Duranti, A. (1993). Truth and intentionality: An D’Andrade, R. (1984). Cultural meaning systems. ethnographic perspective. Cultural Anthropol- In R. Shweder & R. Levine (Eds.), Culture the- ogy, , 214–245. ory (pp.88–119). Cambridge: Cambridge Uni- 8 versity Press, Duranti, A. (2006). The social ontology of intentions. Discourse Studies, , 31–40. D’Andrade, R. (1987). A folk model of the 8 mind. In D. Holland & N. Quinn (Eds.), Durkheim, E., & Mauss, M. (1963). Primitive clas- Cultural models in language and thought (pp. siﬁcation (R. Needham, Trans.). Chicago: Uni- 112–148). Cambridge: Cambridge University versity of Chicago Press. (Original work pub- Press. lished 1903) D’Andrade, R. (1995). The development of cogni- Durkheim, E. (1995). The elementary forms of reli- tive anthropology. Cambridge: Cambridge Uni- gious life (K. Fields, Trans.). London: Allen. versity Press. (Original work published 1912) Daniel, V. E. (1991). The pulse as an icon in Sid- Falk, D..(1992). Braindance: What new ﬁndings dha medicine. In D. Howes (Ed.), The varieties reveal about human origins and brain evolution. of sensory experience (pp. 100–110). Toronto: New York: Henry Holt. University of Toronto Press. (Original work Feld, S. (1984). Sound structure as social struc- published 1983) ture. Ethnomusicology, 28(3), 383–408. D’Aquili, E. G. (1982). Senses of reality in science Feld, S., & Basso, K. H. (Eds.). (1996). Senses of and religion: A neuroepistemological perspec- place. Santa Fe: School of American Research tive. Zygon, 17(4), 361–384. Press. P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

664 the cambridge handbook of consciousness

Foster, M. L. (1994). Symbolism: The founda- ideas (R. Butler, Trans.). London: Oxford tion of culture. In T. Ingold (Ed.), Compan- University Press. ion encyclopedia of anthropology (pp. 366–395). Grindal, B. T. (1983). Into the heart of Sisala London: Routledge. experience: Witnessing death divination. Jour- Foster, M. L., & Botscharow, L. J. (Eds.). (1990). nal of Anthropological Research, 39(1), 60–80. The life of symbols. Boulder, CO: Westview. Gumperz, J. J., & Levinson, S. C. (Eds.). (1996). Frake, C. O. (1961). The diagnosis of disease Rethinking linguistic relativity. Cambridge: among the Subanun of Mindanao. American Cambridge University Press. Anthropologists, 63, 113–132. Haddon, A. C. (Ed.). (1901). Reports of the Cam- Furst, P. (1976). Hallucinogens and culture. San bridge expedition to Torres Straits: Vol. II: Phys- Francisco: Chandler and Sharp. iology and psychology. Cambridge: Cambridge Garro, L. (1998). On the rationality of decision University Press. making studies: Part 2: Divergent rationalities. Halifax, J. (1975). Shamanic voices. New York: Medical Anthropology Quarterly, 12(3). Dutton. Garro, L. (2000). Remembering what one knows Hallowell, A. I. (1955). Culture and experi- and the construction of the past. Ethos, 28(3), ence. Philadelphia: University of Pennsylvania 275–319. Press. Garro, L. (2001). The remembered past in a cul- Hamill, J. F. (1990). Ethno-logic: The anthropology turally meaningful life: Remembering as cul- of human reasoning. Urbana, IL: University of ture, social and cognitive process. In C. Moore Illinois Press. & H. Mathews (Eds.), The psychology of cultural Hardin, C. L., & Maffi, L. (Eds.). (1997). Color experience (pp. 105–150). Cambridge: Cam- categories in thought and language. Cambridge: bridge University Press. Cambridge University Press. Geertz, C. (1973). The interpretation of cultures. Harkness, S. (1992). Human development in New York: Basic Books. psychological anthropology. In T. Schwartz, Geertz, C. (1983). Local knowledge: Further essays G. M. White, & C. A. Lutz (Eds.), New in interpretive anthropology. New York: Basic directions in psychological anthropology (pp. Books. 102–124). Cambridge: Cambridge University Gell, A. (1977). Magic, perfume, dream. In Press. I. M. Lewis (Ed.), Symbols and sentiments: Harner, M. (ed.) (1973). Hallucinogens and Cross-cultural studies in symbolism (pp. 25–38). shamanism. New York: Oxford University London: Academic Press. Press. Gell, A. (1992). The anthropology of time: Cul- Heelas, P., & Lock, A. (Eds.). (1981). Indigenous tural constructions of temporal maps and images. psychologies: The anthropology of the self. New Oxford: Berg. York: Academic Press. George, M. (1995). Dreams, reality, and the Henrich, J., Boyd, R., Bowles, S., Camerer, C., desire and intent of dreamers as experienced Fehr, E., & Gintis, H. (2004). Foundations by a fieldworker. Anthropology of Consciousness, of human sociality: Economic experiments and 6(3), 17–33. ethnographic evidence from fifteen small-scale Geurts, K. L. (2002). Culture and the senses: societies. New York: Oxford University Press. Bodily ways of knowing in an African com- Hinton, A. L. (Ed.). (1999). Biocultural munity. Berkeley: University of California approaches to the emotions. Cambridge: Press. Cambridge University Press. Goodenough, W. (1971). Culture, language, and Hollan, D. (2000). Constructivist models of society. Reading, MA: Addison-Wesley. mind, contemporary psychoanalysis, and the Goodwin, C. (1997). The blackness of black: color development of culture theory. American categories as situated practices. In L. B. Resnick, Anthropologist, 102(3), 538–550. et al. (Eds.), Discourse, tools and reasoning: Hollan, D. (2001). Developments in person- Essays in situated cognition (pp. 111–140) Berlin: centered ethnography. In H. Mathews and Springer. C. Moore (Eds.), The psychology of cultural Griaule, M. (1965). Conversations with experience. New York: Cambridge University Ogotemmeli: an introductin to Dogon religious Press. P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

anthropology of consciousness 665

Howes, D. (ed.) (1991). The varieties of sen- child psychology: History, theory, and methods sory experience. Toronto: University of Toronto (pp. 295–356). New York: Wiley. Press. Landerman, C., & Roseman, M. (Eds.). (1996). Howes, D., & Classen, C. (1991). Conclusion: The performance of healing. London: Routledge. Sounding sensory profiles. In D. Howes (Ed.), Lang, A. (1894). Cock lane and common sense. The varieties of sensory experience. Toronto: Uni- New York: AMS Press. versity of Toronto Press. Lang, A. (1897). The book of dreams and ghosts. Hunt, E., & Agnoli, F. (1991). The Whorfian London. Longmans. hypothesis: A cognitive psychological per- Lang, A. (1898). The making of religion. New York: 3 377 spective. Psychological Review, 98( ), – AMS Press. 389. Laughlin, C. D. (1990). At play in the fields of the 1995 Hunt, H. T. ( ). On the nature of consciousness: lord: The role of metanoia in the development Cognitive, phenomenological, and transpersonal of consciousness. Play and Culture, 3(3), 173– perspectives. New Haven, CT: Yale University 192. Press. Laughlin, C. D. (1991). Pre- and perinatal brain Husserl, E. (1962). Ideas: General introduction to development and enculturation: A biogenetic pure phenomenology.New York: Collier Books. structural approach. Human Nature, 2 (3), 171– (Original work published 1913) 213. Husserl, E. (1950/1993). Cartesian meditations. Laughlin, C. D. (1994a). Transpersonal anthro- Dordrecht: Kluwer Academic Press. pology, then and now. Transpersonal Review, Husserl, E. (1970). The crisis of European sciences 1(1), 7–10. and transcendental phenomenology. Evanston, Laughlin, C. D. (1994b). Psychic energy and IL: Northwestern University Press. (Original transpersonal experience: A biogenetic struc- work published 1954) tural account of the Tibetan Dumo practice. Jackson, M. (1989). Paths toward a clearing: In D. E. Young & J.-G. Goulet (Eds.), Being Radical empiricism and ethnographic inquiry. changed by cross-cultural encounters. Peterbor- Bloomington: Indiana University Press. ough, Ontario: Broadview Press. Jackson, M. (ed.) (1996). Things as they are: Laughlin, C. D. (1996a). Archetypes, neurogno- New directions in phenomenological anthro- sis and the quantum sea. Journal of Scientific pology. Bloomington: Indiana University Exploration, 10(3), 375–400. Press. Laughlin, C. D. (1996b). Phenomenological Jilik, W. G. (1982). Indian healing. Surrey, British anthropology. In D. Levinson & M. Ember Columbia, Canada: Hancock House. (Eds.), Encyclopedia of cultural anthropology Jorgensen, D. (1980). What’s in a name? The (Vol. 3,pp.924–926). New York: Henry meaning of meaninglessness in Telefolmin. Holt. Ethos, 8, 349–66. Laughlin, C. D. (1989). Transpersonal anthro- Katz, R. (1982). Boiling energy: Community heal- pology: Some methodological issues. Western ing among the Kalahari Kung. Cambridge, MA: Canadian Anthropologist, 5, 29–60. Harvard University Press. Laughlin, C. D., & Brady, I. A. (1978). Extinction Kay, P., & Kempton, W. (1984). What is the and survival in human populations. New York: Sapir-Whorf hypothesis? American Anthropol- Columbia University Press. ogist, 86, 65–79. Laughlin, C. D., & D’Aquili, E. G. (1974). Bio- Koepping, K.-P. (1983). Adolf Bastian and the psy- genetic structuralism. New York: Columbia Uni- chic unity of mankind: The foundations of anthro- versity Press. pology in nineteenth century Germany. St. Lucia: Laughlin, C. D., & McManus, J. (1995). The rel- University of Queensland Press. evance of the radical empiricism of William Kuklick, H. (1991). The savage within: The history James to the anthropology of consciousness. 3 34 46 of British anthropology, 1885–1945. Cambridge: Anthropology of Consciousness, 6( ), – . Cambridge University Press. Laughlin, C. D., McManus, J., & D’Aquili, E. G. Laboratory of Comparative Human Cognition (1990). Brain, symbol and experience: Toward (LCHC). (1983). Culture and cognitive devel- a neurophenomenology of consciousness. New opment. In P. Mussen (Ed.), Handbook of York: Columbia University Press. P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

666 the cambridge handbook of consciousness

Laughlin, C. D., McManus, J., Rubinstein, R. A., Levy, R. I. (1973). Tahitians. Chicago: University & Shearer, J. (1986). The ritual control of expe- of Chicago Press. rience. In N. K. Denzin (Ed.), Studies in sym- Levy, R. I. (1984). Emotion, knowing and culture. bolic interaction (Part A). Greenwich, CT: JAI In R. Shweder & R. A. Levine (Eds.), Culture Press. theory: essays on mind, self, and emotion (pp. Laughlin, C. D., McManus, J., & Shearer, J.. 214–237). Cambridge: Cambridge University (1983). Dreams, trance and visions: What a Press. transpersonal anthropology might look like. Levy, R., & Hollan, D. (1998). Person-center inter- Phoenix: The Journal of Transpersonal Anthro- viewing and observation. In H. R. Bernard pology, 7(1/2), 141–159. (Ed.), Handbook of methods in cultural anthro- Laughlin, C. D., McManus, J., & Webber, pology (pp. 331–362). New York: Altamira M. (1984). Neurognosis, individuation, and Press. Tibetan arising yoga practice. Phoenix: The Jour- Levy-Bruhl,´ L. (1926). How natives think. nal of Transpersonal Anthropology, 8(1/2), 91– London: Allen & Unwin 106 . Levy-Bruhl,´ L. (1975). The notebooks on primitive Laughlin, C. D., & Richardson, S. (1986, June). mentality. New York: Harper & Row. The future of human consciousness. Futures, Lincoln, J. S. (1935). The dream in primitive cul- 401 419 – . ture. London: Cresset. Laughlin, C. D., & Throop, C. J. (1999). Emotion: Long, J. K. (1976). Shamanism, trance, hallucino- A view from biogenetic structuralism. In A. L. gens, and psychical events: Concepts, methods, Hinton (Ed.), Biocultural approaches to the emo- and techniques for ﬁeldwork among primitives. tions (pp. 329–363). Cambridge: Cambridge In A. Bharati (Ed.), Realm of the extra-human. University Press. The Hague: Mouton. Laughlin, C. D., & Throop, C. J. (2001). Imag- Long, J. K. (1977). Extrasensory ecology. ination and reality: On the relations between Metuchen, NJ: Scarecrow Press. myth, consciousness, and the quantum sea. Lowie, R. H. (1937). The history of ethnological 4 709 736 Zygon, 36( ), – . theory. New York: Rhinehart. 1988 Lederman, C. ( ). Wayward winds: Malay Lucy, J. (1985). Whorf’s view of the linguistic archetypes, and theory of personality in the mediation of thought. In E. Mertz, & R. Par- context of shamanism. Social Science and mentier (Eds.), Semantic mediation. New York: 8 799 810 Medicine, 27( ), – . Academic Press. 1980 Lee, S. ( ). Association for transpersonal Lucy, J. (1992). Language diversity and thought: A anthropology. Phoenix: The Journal of Transper- reformulation of the linguistic relativity hypothe- 1 2 2 6 sonal Anthropology, 4( / ), – . sis. Cambridge: Cambridge University Press. Levi-Strauss,´ C. (1963). Structural anthropology. Lutz, C. (1988). Unnatrual emotions. Chicago: New York: Basic Books. (Original work pub- University of Chicago Press. 1958 lished ) Lutz, C., & White, W. (1986). The anthropology Levi-Strauss,´ C. (1969a). The elementary struc- of emotion. Annual Review of Anthropology, 15, tures of kinship. Boston: Beacon Press. (Original 405–36. 1949 work published ) MacDonald, G. F., Cove, J., Laughlin, C. D., & Levi-Strauss,´ C. (1969b). The raw and the cooked. McManus, J. (1988). Mirrors, portals and mul- Chicago: University of Chicago Press. (Original tiple realities. Zygon, 23(4), 39–64. 1964 work published ) MacDonald, J. L. (1981). Theoretical continu- Levi-Strauss,´ C. (1973). From honey to ashes. New ities in transpersonal anthropology. Phoenix: York: Harper Row. (Original work published The Journal of Transpersonal Anthropology, 5(1), 1966) 31–47. Levi-Strauss,´ C. (1978). The origin of table man- Mauss, M. (1950). Les techniques du corps: Sociolo- ners. New York: Harper Row. (Original work gie et anthropologie. Paris: Presses Universitares published 1968) de France. Levi-Strauss,´ C. (1981). Naked man. Chicago: McDougall, W. (1901). Cutaneous sensations. In University of Chicago Press. (Original work A. C. Haddon (Ed.), Report of the Cambridge published 1971) anthropological expedition to the Torres Straits P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

anthropology of consciousness 667

(Vol. 2,pp.189–195). Cambridge: Cambridge Pinxton, R., Van Dooren, I., & Harvey, F. (1983). University Press. The anthropology of space. Philadelphia: Univer- Menzel, E. W. (1967). Naturalistic and experi- sity of Pennsylvania Press. mental research on primates. Human Develop- Quartz, S., & Sejnowski, T. J. (1997) The neu- ment, 10, 170–186. ral basis of cognitive development: A construc- Merleau-Ponty, M. (1962). Phenomenology of per- tivist manifesto. Behavioral and Brain Sciences, ception. London: Routledge and Kegan Paul. 20, 537–596 Miller, G. A. (1956). The magical number seven, Rappaport, R. A. (1979). Ecology, meaning and plus or minus two: Some limits on our capac- religion. Richmond: North Atlantic Books ity for processing information. Psychological Rappaport, R. A. (1999). Ritual and religion in the Review, 63, 81–97. making of humanity. Cambridge: Cambridge Miller, G. A., Galanter, E., & Pribram, K. H. University Press. (1960). Plans and the structure of behavior. New Rasmussen, S. (1999). Making better ‘scents’ York: Holt, Rhinehart & Winston. in anthropology: Aroma in Tuareg sociocul- Mitchell, S. (1988). Relational concepts in psycho- tural systems and the shaping of ethnography. analysis. Cambridge, MA: Harvard University Anthropological Quarterly, 72(2), 55–73. Press. Reichel-Dolmatoff, G. (1971). Amazonian cosmos: Morris, B. (1994). Anthropology of the self: The The sexual and religious symbolism of the individual in cultural perspective. London: Pluto Turkano Indians. Chicago: University of Press. Chicago Press. Munn, N. (1992). The cultural anthropology of Reyna, S. P. (2002). Connections: Brain, mind, time: A critical essay. Annual Review of Anthro- and culture in a social anthropology. New York: pology, 21, 93–123. Routledge. Obeyesekere, G. (1981). Medusa’s hair. Chicago: Richards, G. (1998). Getting a result: The expedi- University of Chicago Press. tions’s psychological research 1898–1913.InA. Obeyesekere, G. (1990). The work of culture. Herle & S. Rouse (Eds.), Cambridge and the Tor- Chicago: University of Chicago Press. res Strait: Centenary Essays on the 1898 anthro- 136 157 Ochs, E., & Schieffelin, B. (1984). Language pological expedition (pp. – ). Cambridge: acquisition and socialization: Three develop- Cambridge University Press. mental stories. In R. Shweder & R. LeVine Rivers, W. H. R. (1901a). Colour vision. In A. (Eds.), Culture theory: Essays in mind, self, and C. Haddon (Ed.), Report of the Cambridge emotion (pp. 276–320). Cambridge: Cambridge anthropological expedition to the Torres Straits, University Press. Vol 2. Cambridge: Cambridge University O’Nell, C. W.(1976). Dreams, culture and the indi- Press. vidual. San Francisco: Chandler and Sharp. Rivers, W. H. R. (1901b). The colour vision of the Ortner, S. (1984). Theory in anthropology since natives of Upper Egypt. Journal of the Royal the sixties. Comparative Studies in Society and Anthropological Institute, 31, 229–247. History, 37(1), 173–193. Rivers, W. H. R. (1902). The colour vision of the Parish, S. (1994). Moral knowing in a Hindu Eskimo. Proceedings of the Cambridge Philosoph- sacred city: an exploration of mind, emotion, and ical Society, 11(2), 143–149. self.New York: Columbia University Press. Rivers, W. H. R. (1903). Observations on the Peacock, J. L. (1975). Consciousness & change: vision of the Uralis and the Sholagas. Bulletin of Symbolic anthropology in evolutionary perspec- the Madras Government Museum, 5, 3–18. tive. Oxford: Blackwell. Rivers, W. H. R. (1905). Observations on the Peters, L. G. (1982). Trance, initiation, and psy- senses of the Todas. British Journal of Psychol- chotherapy in Tamang shamanism. American ogy, 93, 10–20. Ethnologist, 9(1), 21–46. Rivers, W. H. R., & Head, H. (1908). A human Peters, L. G., & Price-Williams, D. (1980). experiment in nerve division. Brain, 31, 323– Towards an experiential analysis of shamanism. 450. American Ethnologist, 7(3), 397–413. Rouget, G. (1985). Music and trance: A theory Pinker, S. (1997). How the mind works. New York: of the relations between music and possession. WW Norton & Company, Inc. Chicago: University of Chicago Press. P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

668 the cambridge handbook of consciousness

Rumelhart, D. E., Hinton, G. E., & McClelland, Sperber, D. (1996). Explaining culture: A natural- J. L. (1986). A general framework for parallel istic approach. Oxford: Blackwell. distributed processing. In D. E. Rumelhart, J. Sperber, D., & Block, M. (2002). Kinship L. McClelland, & The PDP Research Group and evolved psychological dispositions: The (Eds.), Parallel distributed processing: Explo- mother’s brother controversy reconsidered. rations in the microstructure of cognition, Vol. 1: Current Anthropology, 43(4), 723–748. Foundations. Cambridge, MA: MIT Press. Spiro, M. (2001). Cultural determinism, cultural Rumsey, A. (1990). Wording, meaning, and lin- relativism, and the comparative study of psy- guistic ideology. American Anthropologist, 92, chopathology. Ethos, 29, 218–234. 346 361 – . Spiro, M. (1987). Culture and human nature: The- Ryle, G. (1984). The concept of mind. London: oretical papers of Melford Spiro. Chicago: Uni- Hutchinson. (Original work published 1949) versity of Chicago Press. Sahlins, M. (2000). Colors and cultures. In M. Spiro, M. (1997). Gender ideology and psycho- Sahlins (Ed.), Culture in practice: Selected essays logical reality. New Haven: Yale University (pp. 139–162). New York: Zone Books. Press. Sapir, E. (1931). Conceptual categories in primi- Spiro, M. (1994). Collective representations and tive languages. Science, 74, 578–584. mental representations in religious symbol sys- Sapir, E. (1958). Culture, language and personal- tems. In B. Kilbourne & L. L. Langness (Eds), ity. Los Angeles: University of California Press. Culture and Human Nature (pp. 161–184). New (Original work published 1929) Brunswick, NJ: Transactions Publishers. (Orig- Scholte, B. (1973). The structural anthropology of inal work published 1984) Claude Levi-Strauss. In J. J. Honigmann (Ed.), Stern, D. B. (1997). Unformulated experience. Handbook of social and cultural anthropology Hillsdale, NJ: Analytic Press. 637 716 (pp. – ). Chicago: Rand McNally. Stoller, P. (1984). Sound in Songhay cultural Schutz, A., & Luckmann, T. (1973). The stuctures experience. American Ethnologist, 11(3), 559– of the LifeWorld. Vol I. Evanston: Northwestern 70. University Press. Stoller, P. (1989). The taste of ethnographic things: Segall, M. H., Campbell, D. T., & Herscovits, M. The senses in anthropology. Philadelphia: Uni- J. (1966). The inﬂuence of culture on visual per- versity of Pennsylvania Press. ception. Chicago: Bobbs-Merrill. Stoller, P. (1997). Sensuous scholarship. Shore, B. (1996). Culture in mind: Cognition, Philadelphia: University of Pennsylvania culture, and the problem of meaning. Oxford: Press. Oxford University Press. Strauss, C., & Quinn, N. (1998). Cognitive the- 1984 Shweder, R. A. ( ). Anthropology’s romantic ory of cultural meaning. Cambridge: Cambridge rebellion against the enlightenment, or there’s University Press. more to thinking than reason and evidence. In Stroker,¨ E. (1993). Husserl’s transcendental phe- R. A. Shweder & R. A. Levine (Eds.), Cul- nomenology (L. Hardy, Trans.). Stanford, CA: ture history: Essays on mind, self, and emotion Stanford University Press. (pp. 27–66). Cambridge: Cambridge Univer- 1968 sity Press. Sutich, A. J. ( ). Transpersonal psychology: An emerging force. Journal of Humanistic Psy- Shweder, R. A. (1991). Thinking through cultures. chology, 8, 77–79. Cambridge: Harvard University Press. Synnott, A. (1993). The body social: Symbolism, Shweder, R. A. (2003). Why do men barbecue?: self, and society. London: Routledge. Recipes for cultural psychology. Cambridge: Har- vard University Press. Tambiah, S. (1990). Magic, science, religion and Sobo, E. J. (1996). The Jamaican body’s role in the scope of rationality. Cambridge: Cambridge emotional experience and sense perception: University Press. Feelings, hearts, minds and nerves. Culture, TenHouten, W. (1978–79). Hemispheric interac- Medicine and Psychiatry, 20(3), 313–342. tion in the brain and the propositional, compo- Spencer, H. (1886). Principles of psychology, Vol. sitional, and the dialectical modes of thought. 2 5. New York: Appleton. Journal of Altered States of Consciousness, 4( ), 129–140. Spencer, H. (1997). Gender ideology and psycho- 2000 logical reality. Princeton, NJ: Princeton Univer- Throop, C. J. ( ). Shifting from a construc- sity Press. tivist to an experiential approach to the anthro- P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

anthropology of consciousness 669

pology of self and emotion. Journal of Con- Turner, V. (1985). On the edge of the bush. In V. sciousness Studies, 7(3), 27–52. Turner, & E. L. B. Turner (Eds.), On the edge Throop, C. J. (2002). Experience, coherence, of the bush. Tucson: The University of Arizona and culture: The signiﬁcance of Dilthey’s Press. “descriptive psychology” for the anthropology Tyler, S. (1984). The vision quest in the West or of consciousness. Anthropology of Conscious- what the mind’s eye sees. Journal of Anthropo- ness, 13(1), 2–26. logical Research, 40, 23–40. Throop, C. J. (2003a). Articulating experience. Tylor, E. B. (1958). Primitive culture, Vol. 1: The Anthropological Theory, 3(2), 219–241. origins of culture. New York: Harper and Row. 1871 Throop, C. J. (2003b). On crafting a cultural (Original work published ) mind – a comparative assessment of some Vygotsky, L. (1978). Mind in society: The devel- recent theories of “internalization” in psycho- opment of higher mental processes. Cambridge, logical anthropology. Transcultural Psychiatry, MA: Harvard University Press. (Original work 40(1), 109–139. published 1930) Throop, C. J. (2003c). Minding experience: En Walsh, R. N., & Vaughan, F. (1980). Beyond exploration of the concept of experience in ego: Transpersonal dimensions in psychology. Los the French anthropology of Durkheim, Levy-´ Angeles: J. P. Tarcher. Bruhl, and Levi-Strauss.´ Journal of the History Way, E. C. (1997). Connectionism and concep- of the Behavioral Sciences, 39(4), 365–382. tual structure. American Behavioral Scientist, Throop, C. J. (2005). Hypocognition, a ‘sense 40(6), 729–753. of the uncanny,’ and the anthropology of Webber, M., Stephens, C. D., & Laughlin, C. D. ambiguity: Reﬂections on Robert I. Levy’s con- (1983). Masks: A re-examination, or ‘masks? tribution to theories of ‘experience’ in anthro- You mean they affect the brain?’ In N. R. pology. Ethos, 33(4), 499–511. Crumrine & M. Halpin (Eds.), The power of Throop, C. J., & Laughlin, C. D. (2002). Rit- symbols (pp. 204–218). Vancouver, BC: Univer- ual, collective effervescence and the categories: sity of British Columbia Press. Toward a neo-Durkheimian model of the White, G., & Kirkpatrick, J. (Eds.). (1986). Per- nature of human consciousness, feeling and son, self, and experience. Berkeley: University of understanding. Journal of Ritual Studies, 16(1), California Press. 40–63. Whorf, B. L. (1956). Language, thought, and Throop, C. J., & Murphy, K. M. (2002). Bour- reality: Selected writings of Benjamin Lee Whorf dieu and phenomenology: A critical assess- (J. B. Carroll, Ed.). Cambridge, MA: MIT ment. Anthropological Theory, 2 (2), 185–207. Press. Titchener, E. B. (1916). On ethnological tests of Winkelman, M. (1986). Trance states: A theoreti- sensation and perception, with special refer- cal model and cross-cultural analysis. Ethos, 14, ence to tests of color vision and tactile dis- 174–203. crimination described in the reports of the Winkelman, M. (1994). Multidisciplinary per- Cambridge anthropological expedition to Tor- spectives on consciousness. Anthropology of res Straits. Proceedings of the American Philo- Consciousness, 5(2), 16–25. sophical Society, 55(3), 204–236. Winkelman, M. (2000). Shamanism: The neural Turner, E. (1996). The hands feel it: Healing and ecology of consciousness and healing. Westport, spirit presence among a northern Alaskan people. CT: Bergin & Garvey. DeKalb, IL: Northern Illinois University Press. Young, A. (1995). Harmony of illusions: Invent- Turner, V. (1969). The ritual process: Structure and ing posttraumatic stress disorder. Princeton, NJ: anti-structure. Chicago: Aldine. Princeton University Press. Turner, V. (1982). From ritual to theatre. New Young, D., & Goulet, J.-G. (Eds.). (1994). Being York: Performing Arts Journal Publications. changed by cross-cultural encounters: The anthro- Turner,V.(1983). Body, brain, and culture. Zygon, pology of extraordinary experience. Peterbor- 18(3), 221–245. ough, Ontario: Broadview Press. Turner, V., & Bruner, E. M. (1986) The anthro- Young-Laughlin, J. & Laughlin, C. D. (1988). pology of experience. Urbana, IL: University of How masks work, or masks work how? Jour- Illinois Press. nal of Ritual Studies, 2 (1), 59–86. P1: KAE 0521857430c23 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:0

670 P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

H. Psychodynamic Approaches to Consciousness

671 P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

672 P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

CHAPTER 24 Motivation, Decision Making, and Consciousness: From Psychodynamics to Subliminal Priming and Emotional Constraint Satisfaction

Drew Westen, Joel Weinberger, and Rebekah Bradley

Abstract ing. We describe experimental research in both areas, including recent neuroimaging This chapter describes the relevance of research attempting to identify the neural clinically derived concepts of conscious circuitry involved in motivated reasoning. and unconscious processes for contempo- We conclude by describing how psychody- rary theory and research. We first describe namic theory and research might inform models of consciousness that emerged from contemporary accounts of consciousness by psychoanalytic clinical observation at the disentangling three distinct meanings of acti- turn of the last century. We argue that, vation and by refining the implicit/explicit and although these models had many flaws, they declarative /non-declarative distinctions. were prescient in their postulation of unconscious thought, feeling, and motivation; consciousness as a limited-capacity system used Introduction for problem solving superimposed on a set of competing, collaborating, and conflicting Although von Helmholtz (1909) used the unconscious processes; expression of mem- term “unconscious inference” to describe the ory consciously through explicit recollection way the brain adjusts for distance when or unconsciously in behavior; the influence assessing the size of objects (and William of unconscious networks of association on James, 1890, certainly spilled some ink on consciousness and behavior; and multimodal the topic of conscious and unconscious men- representations associated with multiple tal events; see Weinberger, 2000), for much affects. We then describe two areas of psy- of its first century, psychology had little choanalytically influenced research that bear interest in the distinction between con- on contemporary concepts of consciousness: scious and unconscious processes. Behavior- (1) unconscious (subliminal) activation and ism ruled both conscious and unconscious (2) unconscious affect-regulation processes processes out of court for scientific study. that affect judgment and decision mak- The serial processing models of cognition 673 P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

674 the cambridge handbook of consciousness

that spurred the cognitive revolution in the porary theory and research. The clinical 1960s and 1970s rarely addressed the ques- database has three features that make it tion of consciousness, although they posited particularly useful in hypothesis generation. a short-term memory system that was essen- The first is its basis in the naturalistic study of tially its equivalent. Similarly, social psychol- people’s mental lives and behavior longitu- ogists did not distinguish until the 1990s dinally as they talk about things that matter whether the attitudes and stereotypes they to them in an ongoing way. The second is the were studying were conscious or uncon- focus on affectively charged, motivationally scious, and they tended to make method- significant cognition and behavior, which is ological decisions (exclusive reliance on self- difficult to approximate in the laboratory. reports) that assumed that such processes The third is the focus, since the beginning of are conscious or can readily be made con- psychoanalysis, on identifying and tracking scious by asking. implicit associative networks hypothesized The situation has radically changed in the to regulate behavior and provide the neural last decade. The question of consciousness – foundation for many forms of psychopathol- and with it, of unconsciousness – has ogy. Research on implicit processes tends become the Cinderella of contemporary cog- to focus on shared networks – and indeed nitive neuroscience and allied disciplines, presumes such networks (so that, for exam- as methodological advances and changes in ple, for most subjects robin should facili- paradigms (both conceptual and experimen- tate the recognition of bird ). In contrast, tal) have uncovered the glass slippers that clinical observation focuses on the unshared have finally allowed conscious and uncon- networks that make people different from scious (explicit and implicit) processes to one another and particularly on the idiosyn- dance at the scientific ball. Until the “sec- cratic, affect-laden networks hypothesized ond cognitive revolution” that ushered in to underlie many forms of psychopathology. the focus on implicit processes in the 1990s We would suggest that it was not accidental (see Westen, 2000a), however, the primary that clinicians struggling to confer explana- field of inquiry that focused on the dis- tory coherence (Thagard, 1989) on the data tinction between conscious and unconscious of clinical observation needed to assume the processes and attempted to outline their dif- existence of unconscious processes a century ferential functions was psychology’s wicked before data from the laboratory rendered the stepsister, psychoanalysis. For a variety of postulation of such processes indispensable. reasons, the hypotheses about conscious- With different vantage points come different ness that emerged from clinical observation discoveries. of psychopathology over the last hundred Our aim in this chapter, although partly years (no different, in principle, from clin- historical, is not primarily to describe the ical observations of neurological cases that fossil record of prehistoric (i.e., prescientific) are treated routinely as important sources of thought on consciousness in psychoanalysis. hypotheses in contemporary neuroscience) Rather, our goal is to focus on the theory and never entered into mainstream theory and research on psychodynamic processes that research. Indeed, many students of con- may contribute to our current understanding sciousness today take the view that, if any- of conscious and unconscious processes. We thing, the views of consciousness formulated begin by describing the models of conscious- by this wicked stepsister actually set back ness that emerged from psychoanalytic clin- the study of conscious and unconscious pro- ical observation at the turn of the last cen- cesses by decades (Kihlstrom, 1999, 2000; tury. We argue that these models were not Kihlstrom, Barnhardt, & Tataryn, 1992). only prescient in multiple respects but that Our goal in this chapter is to highlight they also point to phenomena that would clinically derived conceptions of conscious be important to integrate with contempo- and unconscious processes that we believe rary views of consciousness that have their could be profitably integrated into contem- roots in the laboratory (and hence emerged P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

motivation, decision making, and consciousness 675

to answer different questions). We then most striking about these patients was that describe two areas of psychoanalytically they could not offer compelling (conscious) influenced research that bear on contempo- explanations for what they were doing. rary concepts of consciousness: (1) uncon- Freud thus made a simple deduction. scious (subliminal) activation and (2) uncon- If their behavior was motivated (e.g., not scious affect-regulation processes that affect just a tic or physiological event without judgment and decision making. We con- psychological meaning) but was not conclude by describing two ways that psycho- sciously intended, there was only one other dynamic theory and research might inform possibility: It must have been unconsciously contemporary accounts of consciousness, by motivated (see Erdelyi, 1985). He made distinguishing among different meanings of the corollary assumption that motives may activation and between implicit/explicit and be kept from conscious awareness because declarative/nondeclarative processes. they would be threatening to acknowledge (e.g., hostility toward significant others, expressed instead as passive-aggressive A History of Psychoanalytic Views of behavior). He referred to such processes Conscious and Unconscious Processes as dynamically unconscious and contrasted them with the descriptively unconscious From the start, Freud (1900/1953) consid- processes described above (e.g., phenomena ered the theory of unconscious processes to that emerge in priming studies, in which be the cornerstone of psychoanalytic theory. implicit activation of a network influences He argued that if there is a discontinuity reaction time to semantically related words). in consciousness – something the person is Freud proposed his first comprehensive doing but cannot report or explain – then model of the relation between conscious the relevant mental processes necessary to and unconscious process in 1900, but his “fill in the gaps” must be unconscious (see views changed over the years, and he never Rapaport, 1944/1967). Early in his career completely reconciled what he considered, Freud studied aphasia, and he knew then until the end of his life, works in progress. that we cannot be aware of the processes Developments in psychoanalysis since Freud that generate our capacity to speak fluently. have also been associated with somewhat He included such processes, along with acti- different models of consciousness. To cover vated networks that are not currently con- all (or even most) of these models in any scious, among the rubric of mental events detail would be impossible in a brief chap- that are descriptively unconscious, by which ter (see Weinberger, in press). Our review he meant simply that they were active but of the history of psychoanalytic views of not accessible to introspective awareness. unconscious processes is therefore neces- Today we might say that such processes are sarily short and selective. We focus first unconscious by virtue of mental architecture on Freud’s topographic model (conscious, rather than by motivation, a rendering with preconscious, unconscious), which was his which Freud would likely have been quite first and most systematic formulation of the comfortable. relation between conscious and unconscious The starting point for Freud’s develop- processes. We then briefly describe his struc- ment of a theory of conscious and uncon- tural model (id, ego, superego), followed by scious processes, however, was the clinical a brief tour through the conceptualizations observation of psychopathology. He repeat- of conscious and unconscious processes in edly encountered patients who explicitly the psychoanalytic literature since Freud. wanted to overcome their symptoms, but could not do so despite their best conscious Freud’s Topographic Model efforts; for example, obsessive-compulsive patients who would wash their hands repet- Freud’s first systematic attempt at a model of itively until they bled. What was perhaps the mind (Freud, 1900/1953) was called the P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

676 the cambridge handbook of consciousness

topographic model. This model proposed a memory) has severe processing constraints tripartite division of mental processes into imposed by the limited amount of informa- conscious, preconscious, and unconscious tion that can be held in conscious aware- systems. These systems were distinguished ness. This means that the vast majority of by the person’s capacity to be aware of their potentially available information is not in the mental contents. According to the topo- conscious system. Instead, it is represented graphic model, mental processes can be clas- elsewhere in the mind (brain), where it sified by their level of awareness, from fully can remain potentially accessible for lengthy conscious (conscious system) to capable of periods of time. In its latent state, such infor- becoming conscious (preconscious system) mation is not conscious and is therefore to never in awareness (unconscious system). unconscious by definition, although uncon- Like contemporary cognitive neuroscien- scious networks of association can influ- tists, Freud, who was influenced both by his ence conscious thought and behavior. Freud experience as a neurologist and by the Dar- described this latent information as residing winian thinking of his time (the Origin of in a “preconscious” system or state. Species was published in the year of Freud’s This preconscious system consists of men- birth), believed that mental processes could tal contents (memories, experiences, and so be understood as systems and defined by on) that a person can usually bring into con- what they did (i.e., by their functions). In sciousness when needed but that are not cur- his topographic model, these functions were rently conscious. Once preconscious mate- defined by their level of awareness. rial (e.g., a phone number) is brought into consciousness, it becomes part of the con- the conscious system scious system. Once it is no longer actively The conscious system includes what a per- conscious, it returns to the preconscious sys- son is aware of at any particular moment in tem. Although the preconscious system as time. The reader is, for example, conscious of we have just described it shares many of the reading this sentence or is perhaps distracted more static features of long-term memory as by a smell from the kitchen. The contents of described in cognitive theories of the 1970s this system are therefore constantly chang- and 1980s, Freud actually proposed a suring and are selected by their perceptual, prisingly contemporary model of represen- emotional, or motivational significance. A tations as distributed processes that vary in unique feature of Freud’s theory of the selec- their potential for reactivation. In his paper tion of mental contents for conscious atten- on “The Unconscious,” published 90 years tion was his suggestion that the emotional ago, he described representations as significance of representations could either nothing static, but something in the nature activate or inhibit them, with some represen- of a process ...start[ing] from a particular tations inhibited as a way of managing anx- point in the cortex and spread[ing] from iety (e.g., the representation that a darken- there over the whole cortex or along cer- ing mole could be cancerous). Empirically, tain tracts. When this process is completed, this aspect of Freud’s theory has stood the it leaves a modification behind in the cortex test of time (e.g., Ditto, Scepansky, Munro, that has been affected by it – the possibility Apanovitch, & Lockhart, 1998) and could be of remembering. Our consciousness shows usefully incorporated into more cognitively nothing of the sort to justify, from the physi- inspired models of consciousness, a point to cal point of view, the name of a ‘latent mne- which we return. mic image.’ But whenever the same state of the cortex is provoked again, the psychical the preconscious system aspect comes into being once more as a mnemic image. (Freud, 1915/1957,p.208) People are necessarily only aware of a very small part of the information potentially The topographic model posits a relatively available to them at any one moment in time. free and easy exchange between the two In modern terms, consciousness (or working systems, conscious and preconscious. What P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

motivation, decision making, and consciousness 677

is preconscious can usually be brought into are symbolic, imagistic, and associative. At consciousness, and what is conscious easily the risk of reification, one might say that moves to a preconscious state once it is no the unconscious system “thinks” in similes, longer needed. To bring a preconscious con- metaphors, and the images of poetry (cf. tent into consciousness, the person has to Schimek, 1975). In the language of con- focus attention on it. The nature of the orga- temporary theories of persuasion (Eagly & nization of both the conscious and precon- Chaiken, 1998; Petty & Cacioppo, 1986), it scious systems, according to Freud, is ratio- is this system that is involved in “periph- nal and linguistic. That is, their contents eral” routes to persuasion that influence con- can be expressed sensibly and in language. sumers and voters when they are not attend- The only difference between the two sys- ing to and rationally weighing the arguments. tems is that of phenomenal awareness, con- The beautiful women in the beer com- trolled by attention. As a result, Freud some- mercial “speak” the language of the uncon- times referred to them as a single system, the scious, as sirens who promise that Budweiser preconscious/conscious system. (We address will bring something more gratifying than a below the place of information encoded in beer gut. sensory modalities other than language in The unconscious system differs from Freud’s topographic model.) the conscious/preconscious in yet another way: Because it is not bound by ratio- the unconscious system nality, opposing desires can exist side by Freud’s first two systems, conscious and pre- side without contradiction (Freud, 1933). conscious, yield a model of the mind that The mind in this model is often in con- looks a great deal like the serial processing flict. Unconscious desires press for satis- models that dominated cognitive psychology faction, without concern for reality, safety, for 30 years. What was perhaps most dis- morality, or other motives that may exert tinctive about his topographic model, how- equal and opposite pressure for satisfac- ever, was the third system, the unconscious tion. The conscious/preconscious system, system. Based on his clinical observation as in contrast, is connected to reality, moral- well as his naturalistic observation of the ity, and adaptation and hence must weigh “psychopathology of everyday life” (1901; the merits of alternative desires, find com- e.g., slips of the tongue), Freud identified promises among them, and keep others in a set of mental processes that operate in check. ways that differ radically and qualitatively In some cases, desires that are consciously from those of the conscious/preconscious unacceptable slip through the cracks, find and are, by the standards of those systems, alternative outlets, or emerge in disguised “irrational.” The unconscious system in the form. Freud would likely have looked with topographic model is a reservoir of desires, a combination of consternation and the needs, and urges. As a scientist steeped in delight of recognition at the televangelists of evolutionary theory, as well as an observer of the 1980s, such as Jimmy Swaggart and Jim the motives that seemed to get people into Baker, who constantly preached about the trouble in their lives, Freud first posited that evils of sex yet (or perhaps consequently) these desires are primarily self-preservative seemed to have sex on their minds much of (egoistic) and reproductive. Eventually he the time. Apparently, preaching about the came to believe that they are primarily sex- evils of sex was not enough to satisfy their ual and aggressive.1 sexual desires in the face of highly repres- The unconscious system of the topo- sive sexual attitudes, as both men engaged in graphic model is primarily conative rather all kinds of colorful variants of it (e.g., with than cognitive. It differs, however, from prostitutes) when the cameras (and their the conscious and preconscious systems in consciences) were turned off. another way: Rather than being organized Freud termed the strategies people use to by logic and language, its thought processes keep threatening desires at bay or to get their P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

678 the cambridge handbook of consciousness

needs met when acknowledging them would emotional significance of stimuli is often set be too emotionally threatening defenses. (In to zero. the topographic model he tended to use the Nevertheless, the topographic model of term “repression” more generically to refer consciousness has many limitations, some to defenses until he and others came to of which Freud recognized (hence the understand better the complex ways peo- development of his next model of the ple can regulate their emotions; A. Freud, mind, the structural model), and others of 1936; Vaillant, 1977; Vaillant & McCullough, which he did not. Of particular impor- 1998). Defenses often permit some form tance was his tendency to force multiple of satisfaction of unconscious desires with- dichotomies (e.g., conscious/unconscious, out giving them full expression. Frequently rational/irrational, linear/associative, cog- this requires substantial compromise. If such nitive/emotional, linguistic/imagistic) into a compromise allows for enough satisfac- a single one, defined by the distinction tion, all is well. For example, when the between the conscious/preconscious and first two authors look at themselves in the unconscious systems (Westen, 1999b). Here mirror each morning, they do not typi- he was clearly wrong. For example, he cally retch, but neither do they imagine insisted that emotions and language must that they are Brad Pitt. (In their most self- be conscious, whereas irrationality must congratulatory moments, they might set- bespeak unconscious processes. Emotional tle for Dustin Hoffman.) When such com- processes can be conscious or unconscious, promises do not work so well, the per- just as semantic representations can be pro- son suffers in some way, often through a cessed linearly or associatively (as in prim- psychological symptom (e.g., a preoccupa- ing). Interestingly, similar confusions have tion with appearance characteristic of many emerged in the history of thinking about patients with histrionic personality disor- consciousness, as in the widespread con- der, body dysmorphic disorder, or anorexia fusion of declarative and explicit memory nervosa). (described below).

Foresight and Hindsight: Evaluating Consciousness in Later Psychoanalytic Freud’s First Model of Consciousness Models With the benefits of a century of hindsight, Psychodynamic theorists of all stripes gen- theory, and data, we can appreciate both the erally adhere to the proposition that much strengths and weaknesses of the topographic of mental life – including thoughts, feelings, model as a model of the relation between and motives – is unconscious, which means conscious and unconscious processes. As that people can behave in ways or develop noted above, the model has many features symptoms that are inexplicable to them. that should be recognizable to contemporary Likewise, psychodynamic theorists believe cognitive scientists, such as a serial process- that mental processes, including affective ing system superimposed on a parallel archi- and motivational processes, operate in par- tecture. Freud’s model of associative net- allel, so that individuals can have conflict- works was also remarkably similar to the ing feelings toward the same person or situ- spreading activation theories that emerged ation that motivate them in opposing ways in the 1960s and 1970s (see Blum, 1960; and often lead to compromise solutions. The Pribram & Gill, 1976; Westen, 1985). And psychoanalytic view of these processes is as argued below, his understanding of the not, however, monolithic. Here we describe dynamics of unconscious affective and moti- three developments in psychoanalytic the- vational influences on conscious thought ory since Freud’s original model that bear on and behavior continues to offer insights the nature of consciousness: the structural worth incorporating into theories derived model, ego psychology, and object relations primarily from the laboratory, where the theory. P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

motivation, decision making, and consciousness 679

freud’s structural model experience led him to believe that many Whereas Freud’s first model categorized aspects of moral experience are unconscious, mental processes according to the extent as when a person with a harsh superego to which they were or could become con- (i.e., unrealistically stringent moral stan- scious, his second (and final) systematic dards or ideals) berates himself or rumi- model of the mind, the structural model nates on things he wishes he had done dif- (see Freud, 1926, 1933), categorized mental ferently. In Freud’s structural model, much processes by their functions (Jahoda, 1977). of personality reflects the unconscious inter- With the introduction of the familiar tri- nalization of such functions as morality, partite model of id, ego, and superego in self-restraint, and self-soothing from sig- the structural model, Freud’s understand- nificant others. We are rarely aware of ing of the dynamics of the mind shifted such internalizations, which tend to operate from conflict between conscious and uncon- automatically. scious to conflict between desires and the The unconscious system of the topo- dictates of conscience and/or reality. In graphical model was relatively unchanged in the structural model, all sides of the con- the structural model. It was given the new flict could have conscious and unconscious name, “id” (the “it” in the original German, elements. reflecting the way Freud talked to patients Perhaps the most central change from the about seemingly “foreign” aspects of them- topographic to the structural model can be selves, such as symptoms, that felt like “not seen in what had once been the precon- me”), and was no longer the sole locus of scious/conscious system(s), now renamed unconscious processes. It continued, how- the “ego” (the “I” in the original German; ever, to be the reservoir for ontogenetically Bettelheim, 1983). No longer is everything (and phylogenetically) primitive, largely in this system potentially available to aware- sexual and aggressive, needs, desires, and ness. The ego is the part of the mind wishes. that must somehow balance the demands The structural model offered a more of desire, reality, and morality. To do this, accurate view of the relation between con- it marshals mechanisms of defense as well scious and unconscious processes than the as creative compromises among compet- topographic model because it recognized ing forces, many of which are unconscious. that many of the functions ascribed to Thus, unconscious processes are no longer the ego and superego, particularly those the province of a single mental system (in involved in cognition and self-regulation, are the topographical model, the unconscious unconscious. Nevertheless, few psychoana- system). All the systems of the mind, now lysts today rely on the structural model or functionally defined, have aspects that are use the terms id, ego, and superego. This unconscious.2 Mental processes range from largely reflects a recognition of the dan- completely unconscious to completely con- gers of reifying “structures” as real entities scious, with conscious and unconscious pro- or homunculi, rather than treating them, as cesses (e.g., conscious coping strategies and Freud intended, as constructs denoting func- unconscious defenses) often serving similar tionally related processes (see Klein, 1976; functions. Schafer, 1976). The structural model placed greater emphasis on morals, values, and ideals (the ego psychology “superego”; literally, in the original German, A significant shift in psychoanalytic the- “above-me,” as in standing in judgment ory began at about the time of Freud’s on oneself ). Whereas many theorists have death with the development of ego psy- assumed that our values are largely con- chology. Whereas Freud’s primary focus was scious or accessible to consciousness (or have on motivation and conflict (the province not considered the question of level of con- of the topographic unconscious and the sciousness of moral values), Freud’s clinical structural id), ego psychology focused on P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

680 the cambridge handbook of consciousness

the nature and development of the functions wish that something bad happen to even Freud ascribed to the ego, such as problem the score. solving, self-regulation, impulse regulation, The complexity and ambivalence of such and affect regulation (see Blanck & Blanck, a seemingly simple “fantasy” are hallmarks 1974, 1979). Heinz Hartmann (1939/1958) not only of ego psychology but also of and his colleagues (e.g., Hartmann, Kris, the psychoanalytic understanding of mean- & Loewenstein, 1946) were, to a signif- ing more generally. The notion that men- icant extent, cognitive psychologists, and tal representations of this sort are often they actively read and attempted to inte- complex and ambivalent is another aspect grate into psychoanalytic theory the then- of the psychoanalytic theory of the struc- current work of Piaget and Werner. Hart- ture of unconscious networks and their mann (1939/1958) discussed means-end conscious expressions or concomitants that problem solving and the impact of auto- could be profitably integrated with contem- maticity of thought processes on cognitive porary, more cognitive views of conscious- development and adaptation in ways that ness. Consider, for example, a not infre- would be familiar to contemporary cognitive quently encountered clinical phenomenon, psychologists. the survivor guilt often seen in children Later ego psychologists (Arlow & Bren- with a mentally retarded sibling, who may ner, 1964; Brenner, 1982; Gill, 1967) argued as adults unconsciously sabotage their own that “fantasies” (affect-laden beliefs) form success as a way of “making up for” hav- people’s templates for understanding and ing had an intact intellect. Although they reacting to the world. What they termed fan- are generally aware of some of their guilt, tasy functioned much as schemas later came they may be less aware of its patterns of acti- to do in cognitive and social psychology (cf. vation and the ramifications on their ongo- Schimek, 1975; Weinberger, in press). The ing thought, feeling, and behavior. From difference was that, unlike most conceptions a theoretical point of view, in such cases, of schemas, the fantasy templates described networks of association activated by an by ego psychologists tended to be highly impending intellectual success may activate affective and largely unconscious and often an associatively linked network representing included a combination of wishes, fears, and self-in-relation-to-retarded sibling. This in cognitive constructions.3 turn activates unconscious guilt and efforts Weiss, Sampson, and their colleagues to regulate it, for example by failing. We (1986) developed the notion of unconscious address below some of the empirical data fantasy in a more empirical direction, focus- bearing on unconscious emotion and emo- ing on “pathogenic beliefs” in psychopathol- tion regulation of this sort (see also Westen, ogy and psychotherapy. For example, they 1998b, 1999a). have studied the phenomenon of survivor guilt (O’Connor, Berry, Weiss, Schweitzer, object relations theory & Sevier, 2000), first emphasized by psy- Perhaps the most important development choanalytic clinicians describing the expe- in psychoanalysis since Freud is object rela- rience of survivors of concentration camps, tions theory (Greenberg & Mitchell, 1983; who often suffered with the fantasy (i.e., Guntrip, 1971; Mitchell, 1988; Scharf & belief/fear/wish) that they should have died Scharf, 1998). The term “object relations” instead of their brother, sister, or parent. refers to enduring patterns of interpersonal Of note is that the “fantasy” underlying sur- functioning in intimate relationships and the vivor guilt may include not only a pathogenic cognitive and affective processes mediating belief that somehow the survivor could have those patterns (Westen, 1991). Object rela- done something different to save the per- tions theories emerged from the observa- son who died or that doing penance will tion of patients with serious interpersonal somehow undo the unjustice but also a problems, such as people who rapidly attach fear that the survivor is indeed bad and a to someone they have just met and then P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

motivation, decision making, and consciousness 681

feel desperate, betrayed, enraged, or suicidal sonal functioning (Ackerman, Hilsenroth, when the other person does not reciprocate Clemence, Weatherill, & Fowler, 2000; Blatt, their affection or begins to back off when Auerbach, & Levy, 1997; Huprich & Green- confronted with what feels like an unrealistic berg, 2003; Westen, 1991). Perhaps the level of emotional investment in a relation- most widely known object relations the- ship that does not really exist. Whereas ory outside of psychoanalytic circles, and classical psychoanalysis focused primarily on the most generative of research, is attach- sexual and aggressive motives, object rela- ment theory, which emerged from the work tions theories (and related developments, of the psychoanalyst and ethologist John such as relational psychoanalysis; see Aron, Bowlby (1969, 1973, 1982). A core propo- 1996) tend to focus on relational motives and sition of attachment theory is that expe- on what can happen when these motives go riences in the close relationships begin- awry. For example, children who are abused ning with primary caregivers in infancy by a parent often cling tenaciously to the and early childhood coalesce into internal very person who is abusing them (DiLillo, representations or “internal working mod- Long, & Russell, 1994), just as victims of els” of attachment relationships (see Main, childhood sexual abuse are more likely to Kaplan, & Cassidy, 1985). These working find themselves repeatedly in abusive rela- models are cognitive-affective representa- tionships or dangerous situations as adults tions that influence the ways people think, (Classen, Palesh, & Aggarwal, 2005). From feel and behave, particularly in close inter- an object relations standpoint, people can personal relationships. They form the core have conflicting relational motives and are of patterns of attachment-related behav- frequently unaware of both these motives iors, often referred to as attachment styles. and the interpersonal patterns that they Two aspects of attachment style of partic- engender (Fonagy, Gergely, Jurist, & Target, ular significance here are that they oper- 2002; see Westen, 1990b, 1997). ate primarily outside of awareness, similar For the present purposes, the major inno- to the representations postulated by other vation of object relations theory lies in its object relations theorists, and that they are emphasis on the complex representations of intertwined with unconscious or implicit self, others, and relationships hypothesized affect-regulation strategies (defenses) that to mediate interpersonal functioning in close emerge through interactions (particularly relationships. In the early 1960s, Sandler and aversive interactions) with attachment fig- Rosenblatt (1962) described the cognitive- ures. These affect-regulation strategies have affective structure of the “representational observable manifestations in the first 12 to world” – that is, of people’s representations 18 months of life and throughout the lifes- of the self, others, and relationships. They pan (see Mikulincer, Shaver, & Pereg, 2003; distinguished between people’s conscious Shaver & Mikulincer, 2005). representations of self and the complex net- Kernberg (1975, 2004) has focused on works of unconscious self-representations the structure of representations in patients and episodic memories that could influence with personality disorders, such as the ten- the “shape” of conscious self-representation dency of patients with borderline personality at any moment. disorder to have difficulty maintaining bal- Unlike many areas of psychoanalytic the- anced representations of significant others. ory, object relations theory has amassed a Empirically, patients with borderline pathol- substantial body of empirical research, as ogy are prone to “splitting” their represen- researchers have attempted to track down tations into emotionally one-sided, all-good and operationalize the processes that medi- or all-bad views of the self and others, parate the capacity for intimacy, including ticularly when emotions are strong (Baker, the implicit representations, expectations, Silk, Westen, Nigg, & Lohr, 1992; Conklin and ways of construing relationships that & Westen, 2005; Shedler & Westen, 2004). guide both normal and pathological interper- In the present context, what is particularly P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

682 the cambridge handbook of consciousness

relevant about splitting in borderline per- (i.e., associated with both positive and neg- sonality disorder is that in many respects ative affect) but also multivalent (i.e., asso- it reflects a deficit in the self-regulation ciated with a range of different affects). (If of consciousness, whereby attributes of a the reader has any doubts, try answering person are selected for integration into con- the question, “How do you feel about your scious representations only to the extent that mother?” We suspect the answer will vary they share the prevailing emotional tone. In substantially depending on when the ques- some ways, splitting can be understood as tion is asked and who is asking it.) an extreme form of mood-dependent mem- Second, representations of the self and ory and cognition, although at times it can significant others tend to be densely inter- also be motivated, as when a person ideal- connected with wishes, fears, interpersonal izes a troubled spouse or parents and system- patterns (implicit relational procedures; see atically filters out information that might Westen, 1997), and patterns of affect reg- “tarnish” the representation. ulation. As a result, they tend to be slow Both Kernberg (1975) and Kohut (1966, to change, and efforts to alter maladaptive 1971) focused on the structure of self- explicit representations of the self or oth- representations in narcissistic personality ers may have little enduring impact because disorder. Of particular relevance to the they do not address the broader networks understanding of consciousness is a charac- in which they are embedded (see Westen, teristic feature of many patients with severe 2000b; Westen, Novotny, & Thompson- narcissistic disturbances, whose conscious Brenner, 2004). grandiosity often rests on a foundation of Third, a central point of psychoanalysis unconscious self-devaluation or desperate since its start, and implicit in most object fear that they are not who they want to relations theories, is that it is not inciden- think they are (see Russ, Bradley, Shedler, tal whether a representation is conscious & Westen, 2005; Westen, 1990a). These or unconscious. Much of the work of con- patients often are at their most (consciously) temporary dynamic psychotherapy involves grandiose when they feel most vulnera- identifying unconscious representations of ble to devaluation, as when they respond self, others, and relationships and implicit with rage, narcissistic tantrums, or grandiose relational procedures that lead people to indignation when someone questions their experience repetitive, unsatisfying interac- judgment or ability. Clinically, the constant tions with other people and the nega- need for mirroring, approbation, and brag- tive affect states those interactions typically ging in narcissistic patients often appears to entail. Some of these unconscious represen- belie a powerful need to bolster unrealistic tations and relational procedures are uncon- explicit views of the self (for empirical data, scious by virtue of defense (i.e., kept from see Baumeister & Vohs, 2001; Rhodewalt & consciousness because they are threatening), Sorrow, 2003; Shedler & Westen, 2004). whereas others are likely unconscious sim- More broadly, the psychoanalytic concept ply because they never became the object of of representations, which has been refined introspection. through object relations and relational theorists over the last 50 years, has three features worthy of note. First, representations are Psychoanalytically Inspired Research multidimensional and multivalent. A per- on Unconscious Processes: son has a large repertoire of implicit repre- Subliminal Activation sentations of the self and significant others that may be mutually inconsistent, encoded As with the psychoanalytic theory of uncon- in multiple modes, and activated at dif- scious processes, it would be impossible to ferent times and in different combinations review psychoanalytic research concerning under different circumstances. Representa- unconscious processes comprehensively in a tions of significant social others are also chapter. (The interested reader is referred always affect-laden and not only ambivalent to a series edited by Bornstein & Masling, P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

motivation, decision making, and consciousness 683

1998.) Instead, we describe two lines of tral stimulus. Perceptual vigilance was said research relevant to contemporary theories to occur if an affectively arousing stimulus of consciousness. (To illustrate the psycho- required a shorter presentation than did a analytic concept of narcissism, we focus on neutral stimulus. research programs in which we have been Both defense and vigilance were easy centrally involved.) to find. For example, taboo words took The first line of research regards sublimi- longer to identify than did neutral words. nal activation. Psychoanalytic theorists and This was perceptual defense. Threatening clinicians from Freud through the present words sometimes took longer to identify have taken as axiomatic that affective pro- (perceptual defense) and sometimes were cesses can influence conscious thought and identified more easily (perceptual vigilance) behavior outside of awareness and that than were neutral words. These differ- affects themselves can be generated by ences seemed to make sense psychoana- events that occur outside of awareness lytically. For example, hysterics, who are (see Weinberger, in press; Westen, 1998c). said to avoid threat, showed perceptual Research on subliminal activation supports defense, whereas paranoids, who see threat both suppositions. everywhere, showed perceptual vigilance. In many respects, this work foreshadowed research conducted decades later using such The New Look paradigms as the emotional Stroop task to The first systematic effort to examine psy- measure implicit attentional biases (simi- choanalytic views of unconscious processes larly assessed via reaction time) in different emerged in the 1950s in what was termed psychopathological groups (e.g., spider pho- the “New Look” in perception. Dixon (1971, bics, depressed patients;Williams, Mathews, 1981) provides a comprehensive review and & MacLeod, 1996). critique of this work (see also Weinberger, in Methodological critiques of these studies press). The basic premise of the New Look abounded. Many were justified. For exam- was that perception does not just involve ple, some pointed out that taboo words may a neutral mirroring of environmental stim- take longer to identify than neutral words ulation. Instead, perception is influenced simply because people are reluctant to say by psychological processes, including moti- them out loud until they are absolutely vation and emotion. To test this premise, certain of them so as to avoid embarrass- investigators measured detection and recog- ment. No such hesitancy prevented them nition thresholds of affectively meaningful from saying neutral words out loud. The and neutral stimulation. This work quickly same could be said of threatening words. developed into efforts to investigate psy- Hysterics tend to deny threat and so would choanalytic propositions about what should not admit to it until absolutely forced to affect perceptual processes. Two constructs do so. Paranoids see threat everywhere and directly derived from psychoanalytic theory so will guess a threatening word on mini- soon became central to New Look research: mal evidence. The advent of signal detection defense and repression. Defense was stud- theory gave researchers the ability to differ- ied in what New Look researchers termed entiate between perceptual and psychologi- perceptual defense and vigilance; repression cal threshold setting. However, the literature was studied in what was termed subception. dealing with these controversies is too large Perceptual defense and vigilance were and complex to review in any detail here (see studied by presenting a word or picture Dixon, 1971, 1981; Weinberger, in press), and tachistoscopically and determining at what the issue was never definitively resolved. speed (measured in milliseconds) it could Another argument was theoretical. For be accurately identified. Perceptual defense perceptual defense to work, the person was said to occur if detection of an affec- must first perceive the word or image and tively arousing stimulus required a longer then had to not see it (consciously). This presentation than did detection of a neu- process was sometimes called perceiving P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

684 the cambridge handbook of consciousness

for the purpose of not perceiving and was cognitive psychology. Psychoanalytic theory seen as a logical impossibility. This argu- did not have either the empirical base or the ment was based on the major model of intellectual “cachet” in the field to maintain perception of the time, a serial processing its competing view of a brain characterized model wherein a stimulus was first regis- by parallel, unconscious processing. tered, then recognized, and then reacted to. Erdelyi (1974) pointed out that such a Subliminal Psychodynamic Activation model is not necessary because several perceptual processes can occur simultaneously. In the 1960s, Lloyd Silverman began a Today, most researchers recognize that men- series of investigations termed subliminal tal processes often occur in parallel (and that psychodynamic activation, based explic- affective processing can be quicker than cog- itly on psychoanalytic theories. Reviews of nitive processing; LeDoux, 1986). It could the early work can be found in Silver- therefore make sense for someone to react man, Bronstein, and Mendelsohn (1976); emotionally but show no recognition of a later work in this area was reviewed by stimulus. These parallel processing and neu- Silverman and Weinberger (1985) and Siegel roscience models were, unfortunately, not and Weinberger (1998). The basic paradigm available at the time, and as a result, the cri- for this work involved presenting a stimu- tiques of the New Look were perceived as lus intended to represent a psychoanalytic definitive. construct subliminally and seeing if it had Subception did not fare much better than the effects predicted by the theory. Much of did perceptual defense/vigilance. In subcep- the research was supportive. The subliminal tion, a neutral stimulus was associated with stimulus was typically presented four times, electrical shock. It was then presented sub- via a tachistoscope, for 4 ms. Tests for subliminally, and the person was then asked liminality followed the presentations. Two to say what he or she saw. The person stimuli were usually compared, one selected would deny seeing the stimulus while, at for its presumed psychodynamic meaning the same time, evidencing arousal, in the and the other a control stimulus. Studies form of an electrodermal response to it. Such employed both within-subject and between- electrodermal responding was held to indi- subject designs, and results were similar for cate anxious arousal. Thus it seemed that both. affective responding (anxiety) occurred in Silverman’s best-known research focused the absence of conscious perception. This, on the effects of stimulating a wish or too, was disputed, most elegantly by Eriksen representation of being closely connected (1959). He pointed out that recognition of a with the mother of early childhood, on stimulus is an either/or proposition, whereas the theory that the infant-mother con- electrodermal responding is continuous. It nection can be one of the most nurtu- is therefore quite easy to imagine a per- rant and non-conflictual periods of life. son denying seeing a stimulus when uncer- The phrase MOMMY AND I ARE ONE, tain, but showing some physiological activity presented subliminally, operationalized this indicating perception of the stimulus. Once representation; the control stimulus was again, this critique seemed definitive at the usually PEOPLE ARE WALKING (Siegel time, although today we recognize that cog- & Weinberger, 1998; Silverman, Lachmann, nitive and affective processes may occur in & Milich, 1982; Silverman & Weinberger, parallel and rely on different (though inter- 1985). dependent) neural circuitry. Initial experiments with schizophrenia As a result of these criticisms, the New patients revealed that what was termed Look disappeared from the literature after their ego pathology (a variable closely akin 1960. Behaviorism was in its heyday, and to thought disorder) could be temporar- serial processing models of information pro- ily reduced by presenting them with this cessing were dominant in then-emergent subliminal message MOMMY AND I ARE P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

motivation, decision making, and consciousness 685

ONE. Subsequent experiments expanded cific explanation; namely, that it is the acti- the approach from laboratory demonstra- vation of a representation of an impor- tions of transient effects to a more direct test tant significant other, the mother in this of the general adaptation-enhancing effects case, that underlies these effects. Baldwin of subliminal MOMMY AND I ARE ONE (1994) found that exposing participants to (MIO) stimulation in therapeutic and edu- the name of a supportive other led to more cational settings. In most of these studies, positive self-evaluations. Pierce and Lydon treatment was more effective when pre- (1998) primed proximity-related words and ceded by MIO than by control stimulation. thereby increased participants’ reliance on This does not mean that MIO stimulation support seeking when faced with stress. effected cure; rather, treatment was found Cohen, Towbes, and Flocco (1988) found to be more efficacious when preceded by that priming memories of attachment secu- MIO than by a control stimulus. Of particu- rity caused participants to perceive others in lar importance, however, a range of semanti- more supportive terms. cally similar control stimuli (e.g., MOMMY The most compelling research in support AND I ARE GOING) did not have the same of the specificity of attachment figures over a effects in any of these studies. general mood-enhancing effect comes from Meta-analyses confirmed the effects of the work of Mikulincer and Shaver. Mikulin- these seemingly fantastic findings (see cer et al. (2000) subliminally primed partic- Weinberger, 1992, for a discussion of the ipants with either a stressful word (failure) counterintuitive nature of these findings and or a neutral word (hat). They then asked the resultant resistance to them). Hardaway participants whether a series of word strings (1990) conducted a meta-analysis of all MIO were words or not and measured their reac- studies and obtained what Cohen (1977) tion time in responding (a lexical decision would call a small to moderate effect size task). The logic behind a lexical decision of d = .41. Moreover, the effect sizes of task of this sort is that words related to the studies conducted by or in conjunction prime become more accessible and therewith Silverman were identical to the effect fore should evoke a shorter reaction time. sizes conducted in independent laboratories. The investigators made a creative (or what Weinberger and Hardaway (1990) further Popper would call “risky”) prediction based demonstrated that published and unpub- on attachment theory. In attachment the- lished studies yielded equivalent effect sizes. ory (Bowlby, 1969), threat leads to a need They also conducted what Rosenthal (1979) for security, which activates the attachment has termed a counter-null analysis aimed at system. Mikulincer and colleagues there- addressing the “file drawer” problem. Such fore hypothesized that threat words would an analysis estimates how many null find- make attachment-related words more acces- ings (presumably unreported and stashed in sible. The hypothesis was supported. The various file drawers) would be required to threat prime led to faster reaction times for cancel out the obtained effects and render words related to proximity to attachment them non-significant. This analysis revealed figures than for neutral words. More impor- that it would require 2,287 studies with tantly, participants responded more quickly null results to abrogate the significant MIO to proximity words than to generally pos- results reported in the literature. Thus, the itive or negative words. Thus, the threat results are real and reliable. prime uniquely affected proximity-related Weinberger (1992) suggested that MIO representations, and the effects could not results might be mediated by mood effects. be attributed to the general activation of That is, MIO stimulation might generate affect. These findings were replicated when positive mood, which would then mediate the threat prime was changed to “death” or positive outcomes in therapeutic and edu- “illness.” Thus, the effects were not unique cational venues. Recent research in related to any particular threat word, but held for areas suggests an alternative and more spe- threat in general. P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

686 the cambridge handbook of consciousness

Mikulincer, Gillath, and Shaver (2002) resulted in reduction of out-group bias. The later conducted a study that paralleled the out-groups were homosexuals and Arabs as MIO paradigm more closely. In this study, evaluated by heterosexual Israelis. Because the words used in the lexical decision task one of the primes employed as a control (LDT) were not general proximity words, was evocative of positive affect, the find- but were instead people’s names. The names ings could not be attributable to mood vari- included current attachment figures, close ations. Priming positive affect had no more friends, a current romantic partner, peo- effect than negative or neutral affect. Simi- ple participants knew of but did not know larly, Mikulincer et al. (2001) found in mul- personally (e.g., famous actors), and peo- tiple studies that activation or priming of ple whom participants did not know; other attachment led participants to respond to stimuli were non-words. The investigators the needs of others more empathically. Once once again primed participants with either a again, a positive affect prime had no effect. threat word or a neutral word. Priming facil- These results and others (Mikulincer et al., itated processing only of attachment figures 2002) suggest that there is something spe- (evidenced in reduced reaction time). The cific about priming attachment relationships investigators once again replicated the effect that has different effects than priming other using a different threat word and switching positively valenced relationships or, more from a within-subject to a between-subject generally, positive affect. design. Together, this body of research work It may be that these studies explain the strongly suggests that unconscious activation robust effects of what seems like an out- of representations of attachment leads to landish stimulus, MOMMY AND I ARE multiple positive effects, from threat reduc- ONE. Mikulincer et al. (2000, 2002) demon- tion to treating other people with more strated the specificity of their effects to understanding. The question of whether attachment figures and proximity seeking. mother is somehow special or is just one Close relationships did not have the same (albeit the most common) instance of a pos- effects if they were not attachment rela- itive attachment representation remains to tionships, just as would be predicted by be studied. attachment theory. Because mothers are typically people’s most important child- Subliminal Politics hood attachment figures, it may be that MIO effects are attributable to participants’ Three recent studies by Weinberger and attachment to their mothers. In support of Westen (2005) used subliminal priming to this, Mikulincer and Shaver (personal com- examine affective influences on political munication to J. W., February, 2003) noted evaluations and choices. One was based on that the most common attachment figure an ad run during the 2000 presidential elec- identified by their participants was their tion by the Bush campaign; it criticized mother. Al Gore containing what appeared to be The studies by Mikulincer and colleagues the subliminal word RATS (Berke, 2000; may also help explain the well- repli- Crowley, 2000). Gore supporters suspected cated but hard-to-believe findings using foul play. Bush supporters insisted the sub- MIO studies; namely, that the MIO stimu- liminal appearance was inadvertent. Adver- lus led to increases in adaptive behavior in tising executives were generally skeptical, many samples. For example, Mikulincer and likening subliminal effects to belief in astrol- Arad (1999) reported that priming memories ogy and alien abduction (Egan, 2000)or of attachment security increased cognitive alligators in the sewers of New York City openness in response to belief-discrepant (Shapiro, 2000). information. Mikulincer and Shaver (2001) In our experiment, the word RATS was found in five studies that priming with presented subliminally prior to the supralim- stimulation evocative of positive attachment inal presentation of a photo of an individual P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

motivation, decision making, and consciousness 687

ostensibly running for office. Participants susceptible to manipulation of unconscious were asked to evaluate this individual associative processes. according to several affect-laden qualities. These three experiments had particularly Participants stimulated with RATS rated the high ecological validity because they were purported candidate more negatively than conducted on the Internet. Although peo- did participants stimulated with sublimi- ple participated at the times of their choos- nal control stimuli, including STAR (RATS ing, and distracting stimuli were not con- spelled backwards). This experiment sug- trolled for, significant effects were obtained gested that the presentation of the sub- nonetheless. This suggests that subliminal liminal message RATS by Republicans in stimulation can have effects in the real the 2000 presidential campaign, whether world. intended or not, could have affected political evaluations. A second experiment was designed to Emotional Constraint Satisfaction test the hypothesis that the Gore campaign in Judgment and Decision Making shot itself in the foot in the 2000 election by trying to dissociate itself from Bill A second body of psychodynamically Clinton. In this study, preceding the same inspired research of relevance to the under- ersatz candidate was a subliminal photo of standing of conscious and unconscious pro- either the candidate himself or Bill Clin- cesses focuses on implicit and explicit emo- ton. Evaluations were less negative after tional influences on judgment and decision the photo of Clinton than after the photo making – from the kind of judgments people of the candidate. (In this and the next make in everyday life (e.g., about whether experiment, only negative evaluations were to take one job or another or whether a affected by subliminal stimulation. Posi- comment had a hostile “twist”) to political, tive evaluations were unaffected. Results are judicial, or boardroom decisions of tremen- therefore described in terms of reduced or dous significance (e.g., whether Iraq had increased dislike.) weapons of mass destruction prior to the A third experiment had potentially U.S. invasion in 2003). Across a number of greater political ramifications. Here the sub- fields – cognitive science, psychology, eco- liminal Clinton photo preceded a photo of nomics, political science, and business – the former Governor of California Gray Davis. most widely held models of judgment and The study was conducted during the recall decision making today are bounded ratio- election that removed Davis from office. nality models, or almost-rational models. The results were complex, but theoretically These models suggest that people are largely coherent. Republicans and Democrats were rational but that they rely on (frequently barely affected by the subliminal stimula- adaptive) shortcuts and have some cognitive tion, likely reflecting their strong prior atti- quirks (e.g., in making decisions in high- vs. tudes toward Davis. Independents, however, low-risk situations) that can sometimes lead were a different story. Although Indepen- to divergences from rational choice. dents reported disliking Davis, when Davis These bounded rationality models gener- was preceded by the subliminal Clinton, ally begin with some version of expectancy- their ratings were dramatically less negative. value theories, which focus on the explicit The implication is that subliminal stimula- (conscious) processes by which people tion in politics is unlikely to change attitudes weigh various options and draw conclusions in people whose opinions are strongly held designed to maximize utility. Central to but can have real effects on those whose these models is the idea that, when peo- opinions are not set in concrete. Elections ple make decisions, they consider both the often hinge on the voting decisions made utility or value to them of different options by such individuals, and politicians therefore and the probability or estimated likelihood of target them. Apparently such individuals are obtaining the outcomes associated with each P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

688 the cambridge handbook of consciousness

option (Edwards, 1977; Edwards & New- rationality models suggesting that people man, 1986). A rational judgment, accord- devote conscious cognitive resources to ing to such models, requires comparing each problems of significance to them, but tend to potential option on its expected utility, calcu- use cognitive shortcuts in matters of less con- lated by multiplying the utility and expected sequence (Gigerenzer & Goldstein, 1996; probability of a given outcome. Gigerenzer & Selten, 2001; Simon, 1990). Consider the decision-making process Although some researchers (e.g., Mellers, that confronted typewriter manufacturers in 2000) now posit a greater role of affect in the 1970s and early 1980s, as the prospect decision making, models of judgment, deci- of personal computers (PCs) with word- sion making, and problem solving in cogni- processing capabilities began to emerge. In tive psychology, political science, and eco- weighing the extent to which they should nomics have always emphasized cognition devote capital to the development of then- over affect in accounting for both optimal nonexistent PCs, they had to consider the and suboptimal judgment and decision mak- utility (in terms of potential sales) of devot- ing (see also Marcus, Newman, & MacKuen, ing resources to this new technology ver- 2000; Simon, 1967, 1984). sus the utility of continued development of Like a number of other emotion theo- their current products (which were begin- rists (e.g., Panksepp, 2005; Plutchik, 1980; ning to have some word-processing capa- Tomkins, 1962), Westen (1985, 1994) has bility). Even if the potential utility was argued for a much more substantial role astronomical, they also had to assess the of emotion – and emotional regulation – probabilities that (a) R&D devoted to com- in judgment and decision making. Con- puter technology would lead somewhere – temporary views of motivation emphasize which was certainly not a sure thing in the approach and avoidance systems motivated 1970s – and that (b) a market for home by positive and negative affect4 (Carver, computers would ultimately arise that might 2001; Davidson, Jackson, & Kalin, 2000; capture some of the market share then occu- Gray, 1990). Westen (1985) proposes that pied by typewriters. If either utility or prob- the “value” component of expectancy-value ability were judged to be low, executives theories is primarily affective, as peo- would have been likely to select the more ple implicitly and explicitly respond with conservative option of continuing to develop approach and avoidance to various alterna- their current product lines. Judging from the tives based on elicited, associated, or antic- relative accessibility today of the names IBM ipated affect (see contemporary research and Smith Corona, we can make some edu- on emotional forecasting; Gilbert, Pinel, cated guesses about the judgments made in Wilson, Blumberg, & Wheatley, 2002). In different boardrooms. (Smith Corona even- this view, operant conditioning can be under- tually went bankrupt.) stood as implicit emotion-based decision With respect to the “almost” or “bounded” making, in which humans and other ani- part of the almost-rational or bounded ratio- mals gravitate toward or away from actions nality models, researchers recognized over associatively linked to rewarding or aversive 30 years ago that the idealized views of ratio- consequences (for relevant data, see Westen, nal decision making in expectancy-value 1985). In explicit judgment and decision theories left something to be desired as making, people similarly approach and avoid descriptions of how people actually make alternative conclusions based on their emo- decisions. Kahneman, Tversky, and others tional consequences, although these per- have identified numerous heuristics and ceived consequences may be much more cognitive biases that influence judgments strongly cognitively mediated. (Kahneman & Tversky, 2000; Tetlock & Of particular relevance to the present dis- Mellers, 2002; Tversky & Kahneman, 1974). cussion is Westen’s assertion that precisely Other theorists have developed bounded the same processes of affect-based approach P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

motivation, decision making, and consciousness 689

and avoidance that motivate “rational” the meeting and how he anticipates feeling decision making also frequently distort the many years down the road, will determine conclusions people reach, the data they con- his decisions both about what he perceives sider, and the way they weigh those data. To as the best options and how much he speaks return to our example of IBM and Smith up about his opinion. Corona, imagine the events in the mind Now let us factor in the role of implicit of an executive vice president participat- affect-regulatory processes in the way he ing in a discussion of whether and how thinks about both the probability and much to devote resources to PC research and likely consequences – financial, career, and development. A traditional decision-making emotional – of various possibilities. Let us account might describe the events as fol- suppose, for example, that he is high on the lows. The vice president collects data on personality dimension of harm avoidance what is known about PC technology, esti- (Ben-Porath, Almagor, Hoffman-Chemi, mates the likelihood that personal comput- & Tellegen, 1995; Cloninger, Przybeck, ing could actually emerge as a significant Svrakic, & Wetzel, 1994); that is, he is anx- “player” in the emerging word-processing ious and highly sensitive to potential dangers market, estimates the level of investment and tends to be drawn more to avoid nega- required for his organization to become a sig- tive outcomes and affect states than to seek nificant player in this new technology, esti- pleasure, excitement, and risk. And sup- mates the costs of diverting resources from pose he is interpersonally relatively astute continuing to develop its current product and can see that the CEO is leaning in a line of word-processing typewriters, calcu- risk-averse direction. As he collects infor- lates the costs and benefits of each alter- mation about the emerging PC industry, he native multiplied by their probabilities, and pays particularly close attention to infor- makes his presentation to his colleagues. mation suggesting that the technology is Now consider the same scenario, adding many years away, likely to be too expen- in, first, some of the explicit (conscious) sive for the reach of most households for affective processes hypothesized by the decades, and so forth. At the same time, model to occur in decision making. To this he questions the motivations of “yaysayers” executive, the affective value of getting in who seem to him overly enthusiastic while on the ground floor of the PC industry may giving particular credence to internal combe enormous, in two senses. First, to the pany reports about the likely market for its extent that he is invested in his company impressive new word-processing typewrit- (not only emotionally but also financially, ers. The meeting makes him anxious, as e.g., through stock options and retirement his colleagues are expressing strong opin- income tied to the value of the company’s ions in opposing directions, and “part of stock), an investment in PC technology, if it him” responds to the excitement in the room pays off, will mean a tremendous windfall. about this new technology and what it could Second, more personally, suppose he is its mean for the company’s success and his own most vocal champion among the people at financial security. He is thus not the first to the table in the boardroom. If his bet pays speak, but he follows up on one of the com- off, he is a visionary, and his personal stock ments by the CEO with his careful estimates will rise astronomically in the organization, of probabilities – careful estimates that have perhaps landing him the CEO position he been substantially biased by the interaction has fantasized about for years. On the other of his personality style with just the kind of hand, if he is wrong, the costs might also situation in which such a style can either by be substantial if it becomes clear that PCs very helpful or very detrimental. will never take off. These various emotional The kinds of biases in reasoning that costs and benefits, which will likely affect entered the judgments of the executive both how he feels moment to moment in in this last account were first described P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

690 the cambridge handbook of consciousness

systematically in the psychoanalytic liter- the psychoanalytic construct of “compro- ature on defense (which today we might mise formation,” which refers to the pro- think of in terms of implicit forms of affect cess whereby the mind generates compro- regulation). This literature, like research on mise solutions in the face of multiple, often implicit forms of affective activation, is actu- competing motives (Brenner, 1982; Westen, ally quite well developed empirically as well 1985, 1998b, 1999a; Westen, Blagov, Feit, as theoretically (see Vaillant, 1992; Westen, Arkowitz, & Thagard, 2005). Readers with a 1998a; in press). An emerging research lit- background in cognitive science will readily erature on motivated reasoning has also recognize the analogy to equilibration and begun to put such processes on firm empiri- constraint satisfaction in connectionist net- cal ground (e.g., Ditto, Munro, Apanovitch, works, whereby the brain equilibrates to a Scepansky, & Lockhart, 2003; Ditto et al., solution that optimizes goodness of fit to the 1998; Dunning, 1999; Jost, Glaser, Kruglan- available data. Like connectionist models, ski, & Sulloway, 2003; Kunda, 1990). For models of compromise formation, first pro- example, years ago, Mahoney (1977) demon- posed by Freud a century ago to account for strated experimentally the effect of motiva- dream and symptom formation and applied tional biases on manuscript reviewing. He more widely by later psychoanalytic theo- found that scientists identified many more rists (Brenner, 1982), reject a primarily serial methodological limitations in studies that processing view of the mind (in this case, of refuted rather than supported their pre- motives that must come into consciousness existing beliefs even though the methods one at a time to influence behavior; see Olds, sections were actually identical. 1994; Westen & Gabbard, 2002). They sug- As this last rendering of the vice pres- gest instead that multiple motives can opera- ident’s decision-making processes makes tive simultaneously and influence thought or clear, people are often called upon not to action outside of awareness, leading to com- solve one problem at a time but to solve sev- promise solutions. eral affective-motivational problems simul- Westen and colleagues have proposed a taneously – balancing multiple competing model of emotion regulation, judgment, and and collaborating emotional “pulls.” In this decision making that includes the connec- example, the executive had motivations to tionist focus on constraints imposed by the “get it right” – to process information accu- data as well as a second set of constraints rately, given the decisive importance of mak- imposed by the hedonic implications of differing a good decision – as well as to manage ent solutions. According to this model, judg- his anxiety, to protect his investment in his ments about emotionally meaningful issues company, to protect his career, to maintain (which, in everyday life, include most judg- his relationship with the CEO, to respond ments and decisions) reflect the simultane- appropriately to people’s realistic excite- ous satisfaction of two sets of constraints: ment about the possibilities of computer cognitive constraints (imposed by data and technology, and so forth. (For a connection- their logical entailments) and emotional ist account of complex group situations such constraints (imposed by emotional associ- as this, see Thagard, in press.) The psycho- ations and anticipated emotions). Just as logical “task” facing our executive was to information provides constraints on the equilibrate to a number of solutions (about equilibrated solutions that people reach – by how to weigh the data, how to present what spreading activation to networks (and ulti- he had learned, and how to manage his rela- mately conclusions) that make sense of the tionships in the boardroom) in ways that “gestalt” of available data and spreading inhi- would be both as faithful as possible to the bition to alternatives that make less sense data and emotionally tolerable to him. of the totality of the data – feelings and In attempting to model complex acts of emotion-laden goals provide constraints on cognition and emotion regulation of this the equilibrated solutions people reach by sort, Westen and colleagues have drawn on spreading activation to the neural networks P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

motivation, decision making, and consciousness 691

or units that lead people toward desired con- “high crimes and misdemeanors” as intended clusions and inhibiting those that increase by the framers of the U.S. Constitution) the likelihood of undesired ones. Thus, the would reflect some combination of cognitive brain equilibrates to solutions designed not constraints (what participants knew about only to maximize goodness of fit to the data Clinton and the scandal) and affective con- but also to maximize positive and mini- straints (what they felt about the Demo- mize negative affect. Where emotional con- cratic and Republican parties, Clinton, infi- straints are relatively strong, even in the face delity, and feminism). We assessed cognitive of strong cognitive constraints, we would constraints by asking participants ten factual predict that excitation and inhibition of questions about Clinton’s life that would units of relevant networks based on their indicate long-standing knowledge about his associations with positive and negative affect political and personal history and ten factual will exert a strong influence on judgments, questions about the scandal. We assessed leading to solutions that maximize what emotional constraints by asking multiple Thagard (2000, 2003) has called emotional questions about participants’ feelings toward coherence. the Democrats and Republicans, Clinton, We have tested this model in a series of infidelity, and feminism, which we predicted five studies involving three crises over the would provide overlapping but in many last 6 years in U.S. politics (the impeachment cases competing emotional constraints on of Bill Clinton, the disputed presidential judgment. election of 2000, and the discovery of torture As predicted, participants’ judgments at by the U.S. military at Abu Ghraib prison in all three time points reflected a complex bal- Iraq; Westen, Blagov et al., 2005). In all five ancing act, in which their decisions could be studies, we collected data from community predicted by a combination of their feelings samples, assessing or manipulating cognitive toward the parties, Clinton, infidelity, and constraints (the information that would con- feminism (with each set of emotional constrain their judgments) and assessing com- straints contributing significantly after hold- peting emotional constraints, with the goal ing the others constant) and the constraints of predicting judgments that most decision imposed by their knowledge. However, the theories would explain in more “cold” cog- emotional constraints at all three time points nitive terms. The gist of the findings are, swamped cognition. For example, we could first, that people’s political judgments do predict people’s judgments about whether indeed reflect an interaction of multiple cog- the President’s actions crossed the constitu- nitive and emotional constraints; and sectional threshold for impeachment from cog- ond, that competing and collaborating emo- nitive and emotional constraints measured 6 tional pulls dominate judgment and decision to 9 months earlier with remarkable accu- making in high-stakes, emotion-laden polit- racy (88% of the time). However, we could ical situations, generally overriding even rel- predict the same judgments with 85% accu- atively strong cognitive constraints. racy when we included only emotional con- For example, in three studies conducted straints in the model. People’s knowledge at different points of the Clinton-Lewinski about the scandal placed a statistically signif- crisis, we used cognitive and emotional icant but practically insignificant constraint constraints to predict people’s judgments on people’s judgments longitudinally. about whether the President likely groped We obtained similar findings regarding Kathleen Willey in the Oval Office, whether people’s beliefs about the validity of man- he lied to the grand jury investigating the ual versus machine ballot counts in the dis- case, and whether his actions constituted an puted presidential election of 2000. (Recall impeachable offense as defined by the U.S. that Democrats argued for the importance Constitution. We hypothesized that peo- of a manual recount to assess as accurately ple’s answers (e.g., about the extent to which as possible the intent of each voter, whereas lying about sex to a grand jury constitutes Republicans argued that such a recount P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

692 the cambridge handbook of consciousness

would be less valid than the machine counts jointly predicted people’s judgments in a rel- that had rendered George W. Bush the vic- atively “cold” cognitive domain (the state of tor by a handful of votes in Florida.) We the economy). could predict participants’ judgments 83.6% We recently completed a study testing of the time from the combination of two this model using functional neuroimaging sets of cognitive constraints (their knowl- with a sample of committed Democrats edge of relevant data and an experimen- and Republicans during the 3 months tal manipulation) and two sets of emo- prior to the U.S. presidential election of tional constraints (their feelings toward the 2004 (Westen, Kilts, Blagov, Harenski, & two candidates and the two parties). Elim- Hamman, 2006). We presented participants inating both sets of cognitive constraints with 18 sets of stimuli, 6 each regarding from the model (i.e., including only emo- President Bush, his challenger Senator John tional constraints) decreased our ability to Kerry, and neutral male control figures. For predict – but only to 83.0%. Thagard (2003; each set of stimuli, participants first read a Westen, Blagov, et al., 2005) has success- statement from the target (e.g., Bush), fol- fully modeled these processes computation- lowed by a second statement documenting ally using his HOTCO 2 (hot cognition) a clear contradiction between the target’s program. words and deeds that would be threaten- Similar findings emerged in an experi- ing to a partisan (generally suggesting that mental study in which we manipulated cog- the candidate was dishonest or pandering). nitive constraints, providing participants in Next, participants were asked to consider different experimental conditions with more the discrepancy and then to rate the extent or less evidence in an alleged case of abuse to which the target’s words and deeds were at Abu Ghraib. The experimental manip- contradictory. Finally, they were presented ulation had a small, marginally significant with an exculpatory statement that might effect, but this effect was less than one- explain away the apparent contradiction and sixth the magnitude of the effect of feel- asked to reconsider and again to rate the ings toward the parties and was dwarfed as extent to which the target’s words and deeds well by the effects of participants’ feelings were contradictory. toward the U.S. military and toward human Behavioral data (participants’ ratings of rights as a goal of U.S. policy – none of which, the extent to which the first and sec- logically, should have affected people’s judg- ond statements were contradictory) showed ments. In all of these studies, data from com- the expected motivated reasoning, with munity samples mirrored decisions made by partisans denying obvious contradictions elected officials and judges, suggesting that made by their own candidate that they the same processes occur in political elites as had no difficulty detecting in the oppos- in the general electorate. The only study thus ing candidate. Importantly, in both their far in which we have found a more substan- behavioral and neural responses, Republi- tial impact of cognitive constraints (although cans and Democrats did not differ in the still less powerful than the effects of feelings way they responded to apparent contradic- toward the two parties) is in a study just tions for the neutral control targets, but completed in which we predicted people’s Democrats responded to Kerry as Repub- beliefs about the state of the national econ- licans responded to Bush (not, as would omy from an objective set of economic indi- be predicted by most decision models, sim- cators (e.g., gross domestic product, unem- ilarly weighing similar evidence). While ployment), their personal finances, and their weighing apparent contradictions for their feelings toward the political parties (Westen, own candidate, partisans showed expected Kelley, & Abramowitz, 2005). Once again, activations throughout the orbital frontal however, and supporting the model, these cortex, indicating affective processing and various cognitive and emotional constraints presumably affect-regulatory strategies, as P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

motivation, decision making, and consciousness 693

well as lateral orbital and insular cortex consciousness and behavior even when they activations suggesting the experience of are not readily apparent. They argued that negative affect. They also showed large acti- representations are multimodal and usually vations in the anterior and posterior cingu- associated with multiple affects, both posi- late cortices, suggesting emotion processing, tive and negative. conflict monitoring, and perhaps judgments A second way to view psychoanalytic con- of forgivability or moral accountability (see, ceptions, however, is in terms of their poten- e.g., Farrow et al., 2001). After apparently tial contributions to contemporary theory and having found a way to resolve the contra- research on consciousness. We have already diction, while asked to “consider” (and then noted a number of ways in which psychoan- rate) the contradiction, partisans showed alytic views of consciousness emerging from a large activation in the ventral striatum clinical observation point to important phe- and nucleus accumbens – indicating reward nomena that need to be addressed by any processing. We suspect this reflects essen- theory of consciousness, such as differences tially the operant conditioning of a defen- between implicit and explicit representa- sive response, associating the participant’s tions that may be active simultaneously, the “revisionist” account of the data with posi- activation and influence of implicit affective tive affect or relief. and motivational processes, and the multivalent nature of most networks representing emotionally significant concepts (e.g., atti- Conclusion tudes, representations of significant others). Further, psychoanalysis has argued since its One can think about the relevance of psy- inception that meaning is not always mani- choanalytic views of consciousness to the fest in the ways assumed by the widespread contemporary understanding of conscious- use of self-report methods and brief treat- ness in one of two ways. The first is his- ments for complex psychological disorders torical. Historically, psychoanalysis made that build in minimal time or techniques for some claims, based on clinical data, that exploration of associative networks. In this not only differed from the dominant per- view, understanding complex behavior and spectives in psychology that emerged over emotionally significant actions and commu- the next 90 years but also turned out to nications in everyday life (as in the consult- be correct. Most importantly, psychoanalysts ing room) requires an ability to frame and argued that much of mental life is uncon- test hypotheses about the implicit networks scious and that this extends to thought, feel- regulating a person’s behavior in ways that ing, and motivation (Shevrin & Dickman, maximize what Thagard (2002) has called 1980; Westen, 1998c). They viewed con- explanatory coherence. Although clinical sciousness as a serial processing system used inference and interpretation of meaning cer- for problem solving that is superimposed tainly have their pitfalls (e.g., Dawes, Faust, on a mental processing system that operates & Meehl, 1989), recent data suggest that outside of awareness. They argued that men- clinical inferences can be remarkably reli- tal events that occur in parallel can compete, able and valid in assessing such subtle con- collaborate, and conflict in various ways and structs as narcissism from people’s narra- that nothing guarantees their smooth coor- tives (Westen & Weinberger, 2004, 2005). dination. They argued that memory can be Indeed, we decode meaning all the time expressed consciously through explicit rec- in everyday life, and much of the time ollection or unconsciously in behavior. They observers reach considerable consensus on argued that much of psychological experi- what they have observed. The same is true ence is mediated by networks of associa- when researchers apply coding systems to tion that encode not only our experience but narratives to quantify both thematic con- also our feelings and fears and can influence tent and structural aspects (e.g., complexity, P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

694 the cambridge handbook of consciousness

syntax) that provide insight into implicit following unit A. What we have suggested is meanings (e.g., Main et al., 1985). that repeated association of unit A with neg- We conclude with two final ways ative affect can lead to cognitive avoidance that psychoanalytically inspired theory and (inhibition) of unit A, and repeated associ- research may contribute to contemporary ation of unit B with positive affect can lead theory and research on conscious and to approach to, or hyperactivation of, unit B. unconscious processes: by distinguishing In judgment and decision making, these pro- among three types of activation and by cesses of hedonic activation (excitation and distinguishing implicit/explicit from declar- inhibition) can influence the arguments for ative/procedural knowledge. or against a given position that reach consciousness, the amount of time spent thinking about different arguments or pieces of Three Types of Activation data, the conscious processes by which peo- In our discussion of cognitive and emotional ple weigh evidence or rationalize their deci- constraint satisfaction, we have essentially sions, the amount of activation or inhibition argued for the importance of distinguishing spread to units in a network of association, two forms of activation, one cognitive and and so forth (see, e.g., Ditto et al., 2003). one hedonic (Westen, 1985). Cognitive acti- We would note a final distinction vis-a-vis` vation (including both excitation and inhibi- activation, namely between conscious and tion) reflects principles of association famil- unconscious activation. Cognitive theorists iar since Aristotle and refined over the last frequently use the term “activation” without 40 years by cognitive psychologists. Cog- specifying the level of consciousness, on the nitive excitation occurs as one unit in a assumption that if enough activation spreads network increases the level of activation of to a representation it will become conscious. another by virtue of some form of sen- Both psychoanalytic clinical observation and sory, mnemonic, or logical conjunction (i.e., neuroimaging data on the nature of con- the two units have been activated together scious thought processes (usually described before because of some sensory or cognitive in terms of activation of working memory) relation, such as similarity or temporal asso- render this assumption unlikely. ciation, and hence are more likely to be coac- From a neuroimaging standpoint, peo- tivated in the future). Conversely, cognitive ple process a considerable amount of infor- inhibition occurs as one unit in a network mation using posterior networks without (or one network, depending on the extent involvement of the dorsolateral prefrontal to which one wants to model more specific circuits involved in working memory. From processes) inhibits another by virtue of the a clinical standpoint, as we have suggested, fact that its presence suggests the absence or a person could have a highly active repre- low probability of the other.5 sentation that is nonetheless inhibited from A second form of activation is hedo- conscious awareness because of its emo- nic, as affect activates approach or avoid- tional significance. For example, as noted ance, either toward stimuli associated with earlier, narcissistic patients often appear to pleasure or pain (behavioral activation and be at their most consciously grandiose when inhibition; see, e.g., Carver, 2001; Corr, 2002; their self-esteem is at its most unconsciously Davidson et al., 2000; Elliot & Thrash, precarious. Similarly, people who fancy 2002;Gray,1990) or toward mental pro- themselves non-prejudiced may have highly cesses (e.g., inferences, representations, attri- active negative representations of minor- butions) associated or expected to be asso- ity group members that motivate subtle ciated with positive or negative feelings. forms of devaluation while being consciously Cognitive theories have focused on the way aware of nothing but positive feelings toward repeated associations between two units in the target of their implicit prejudice. (This a network can strengthen their association, latter example likely explains the oft-heard leading to heightened activation of unit B sentiment among African-Americans that P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

motivation, decision making, and consciousness 695

the only thing as aversive as an old- of procedural knowledge, involves activation fashioned bigot is a patronizing liberal.) of implicit networks, yet the networks on In each of these cases, neural networks which priming operates are quintessentially appear to have settled simultaneously into declarative. This is particularly apparent in two solutions, one conscious and one uncon- priming procedures involving semantic net- scious. (Alternatively, one might consider works linking words with related mean- the combination of the two incongruent rep- ings. Second, subliminal priming leads to resentations at different levels of awareness robust effects, suggesting that a “declara- to be a single equilibrated solution.) tive” prime (i.e., a prime with semantic con- What this analysis suggests is that content) can activate implicit networks with- scious activation is not simply a function of out itself ever being explicitly processed. the level of cognitive activation, in which Third, from a clinical standpoint, many crossing a threshold brings representations declarative memories or representations – into consciousness. Rather, conscious activa- e.g., warded-off self-representations, beliefs tion is probably a joint function of (1) cog- about what happened at Abu Ghraib that a nitive activation, (2) affective attention to supporter of the U.S. administration would emotionally significant events, (3) affective rather not acknowledge – are not consciously inhibition of unpleasant information (which declarable because of their affective content. can run counter to attention to emotionally Squire’s distinction, and its subse- significant events), and (4) affective excita- quent equation in the literature with the tion of pleasurable information. In the case implicit/explicit distinction, created a of defended-against information, represen- confusion between type of knowledge (how- tations selected for consciousness are not to knowledge, which is procedural, and those with the highest level of unconscious declarative knowledge, which has content) cognitive activation. and the way that knowledge is expressed (explicitly, through conscious recollection, or implicitly, through behavior). The Implicit/Explicit and Procedural/ declarative/procedural dichotomy refers Declarative: One Dichotomy or Two? to the type of knowledge or representa- A second domain in which psychoan- tional mode (facts versus skills), whereas the alytic theory, research, and data might explicit/implicit distinction refers to the way prove useful is in distinguishing two ways this knowledge is retrieved and expressed of classifying mental events or forms (with or without conscious awareness). of memory that are frequently used as There is no necessary relation between synonyms: implicit/explicit and declara- these two distinctions, which represent tive/procedural (or non-declarative). In orthogonal axes. We know many things 1986, Squire introduced the landmark dis- unconsciously, and we exercise many tinction between declarative and procedural choices and instigate and regulate many knowledge. Declarative knowledge referred actions consciously. Declarative knowledge to conscious knowledge, typically verbal or (representations of facts or events) can be semantic, that could be “declared.” Pro- either explicit (as when a person remem- cedural knowledge referred to knowledge bers a recent event) or implicit (as in seman- expressed in behavior that did not require tic priming). Procedural knowledge (repre- conscious recollection. Squire placed several sentations of internal or external actions) phenomena under the rubric of procedural can be either explicit (e.g., conscious deci- knowledge, including priming, skill learning, sion making, coping skills, or deliberate and conditioning. attempts to remember or forget) or implicit Although this distinction was extraordi- (e.g., skills such as reading facial emotions, narily useful in moving the field forward, responses learned through operant condi- it was problematic in two respects. First, tioning, defenses against unpleasant feel- priming, which Squire defined as a form ings). For example, coping strategies are, like P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

696 the cambridge handbook of consciousness

Mode of expression (explicit vs. implicit)

Explicit/declarative (e.g., Implicit/declarative (e.g., recall, recognition) subliminal priming, warded- Type of off views of self) knowledge (declarative vs. procedural) Explicit/procedural (e.g., Implicit/procedural (e.g., problem solving, coping operant conditioning, strategies) defensive strategies)

Figure 24.1. Declarative/procedural and implicit/explicit knowledge.

other habits, procedural, but many of them able only indirectly). Other examples of are quite conscious, as when we remind quasi-declarative knowledge are feelings of ourselves that something painful will be knowing and feelings of familiarity. These over soon. Defenses, in contrast, are implicit feelings inform us of something factual and procedures whose function is similarly to hence have quasi-declarative content (e.g., regulate affect. Thus, a given mental event that we can retrieve this information if we can be declarative and explicit, declarative try hard enough or we have seen it before, and implicit, procedural and explicit, or even though the content is unavailable). procedural and implicit. To broaden the We fear at this point that if we engage distinction to include mental events other in any further distinction-drawing, the clini- than “cold” cognitions, we might distinguish cally sophisticated reader will begin to infer between type of mental event (whether the obsessional dynamics on the part of the event is an idea, feeling, or motive on the authors, and hence we stop here. We hope, one hand, or a mental or behavioral opera- however, that we have made the point tion, action, or strategy on the other) and its convincingly argued that the clinically in- mode of expression (whether the idea, feeling, formed approaches to consciousness that motive, or action can be consciously recog- first emerged from psychoanalytic observa- nized or recalled or whether it is expressed tion a century ago may be not only of histori- in behavior without conscious intention or cal interest but may also continue to provide the necessity of conscious awareness). The insight into aspects of consciousness in the type of mental event usually reflects its mode next century. Just as it would be remarkable or modes of representation (e.g., words or if neuroscientists had learned nothing about images vs. action tendencies). Figure 24.1 consciousness from the recent explosion describes the four quadrants defined by these of neuroimaging research, it seems equally two axes. unlikely that clinicians learned nothing from Alongside episodic and generic forms of a century of listening to people struggle with declarative knowledge, we might also pos- the things that matter most to them. tulate another kind of knowledge that is quasi-declarative; namely, memory for feelings and motives. Emotions and motives Acknowledgments can be considered quasi-declarative because they have content (a “what”). They can Preparation of this article was supported be explicit (conscious emotions, wishes, or in part by NIMH grants MH62377 and fears) or implicit (unconscious and observ- MH62378 to the first author. P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

motivation, decision making, and consciousness 697

Notes Aron, L. (1996). A meeting of minds: Mutuality in psychoanalysis (Vol. 4). New York: Analytic 1. Interestingly, his first dual-motive model of Press. motivation was closer to contemporary evolu- Baker, L., Silk, K. R., Westen, D., Nigg, J. T., & tion thinking than his late model of sex and Lohr, N. E. (1992). Malevolence, splitting, and aggression as fundamental drives, which per- parental ratings by borderlines. Journal of Ner- haps applies best to movies and frat parties. As vous and Mental Disease, 180, 258–264. in his theories of consciousness, Freud’s the- Baldwin, M. W. (1994). Primed relational ories of motives always followed his clinical schemas as a source of self-evaluative reactions. observation, in this case, the observation that Journal of Social and Clinical Psychology, 13, these two motives tend to be the ones that get 380–403. people into the most trouble. Baumeister, R. F., & Vohs, K. D. (2001). Nar- 2. Like the conscious/preconscious, the ego also cissism as addiction to esteem. Psychological draws upon what Freud termed secondary pro- Inquiry, 12, 206–210. cesses (e.g., rational thought, memory, volun- Ben-Porath, Y. S., Almagor, M., Hoffman-Chemi, tary behavior) in its efforts to balance the afore- A., & Tellegen, A. (1995). A cross-cultural mentioned demands. study of personality with the Multidimensional 3. Interestingly, this concept of affect-laden rep- Personality Questionnaire. Journal of Cross- resentations is much closer to contemporary Cultural Psychology, 26, 360–373. views of representation, stereotypes, and atti- Berke, R. L. (2000, Sept. 12). The 2000 campaign: tudes in the social psychology literature than The ad campaign: Democrats see, and smell, were the initial schema theories. Because of rats in G.O.P. ad. New York Times,p.1. the emergence of this concept of fantasy from Bettelheim, B. (1983). Freud and man’s soul. New the clinic, however, the primary focus was York: Knopf. on fantasies that are maladaptive and often Blanck, G., & Blanck, R. (1974). Ego psychology: rooted in painful or conflictual developmen- Theory and practice. New York: Columbia Uni- tal experiences (e.g., the common fantasy of versity Press. young children that they caused their parents’ 1979 divorce). Blanck, G., & Blanck, R. ( ). Ego psychology II: Theory and practice. New York: Columbia 4. We use the terms “affect” and “emotion” University Press. throughout interchangeably. Blatt, S. J., Auerbach, J. S., & Levy, K. N. (1997). 5. Emotional processes can influence this form of Mental representations in personality develop- activation in ways that do not require different ment, psychopathology, and the therapeutic principles. For example, affect orients people process. Review of General Psychology, 1, 351– to phenomena of adaptive significance, which 374. in turn spreads activation to networks repre- 1960 senting those phenomena. In this sense, affect Blum, G. S. ( ). Psychoanalytic behavior (both positive and negative) may provide one theory: A conceptual framework for research. more source of excitatory cognitive (associa- In H. P. David & J. C. Brengelmann (Eds.), Per- 107 138 tional) activation. spectives in personality research (pp. – ). Oxford: Springer. Bornstein, R. F., & Masling, J. M. (Eds.). (1998). Empirical perspectives on the psychoanalytic References unconscious. Washington, DC: American Psy- chological Association. 1969 1 Ackerman, S. J., Hilsenroth, M. J., Clemence, Bowlby, J. ( ). Attachment (Vol. ). New York: A. J., Weatherill, R., & Fowler, J. C. (2000). The Basic Books. effects of social cognition and object represen- Bowlby, J. (1973). Separation (Vol. 2). London: tation on psychotherapy continuation. Bulletin Hogarth Press. of the Menninger Clinic, 64, 386–408. Bowlby, J. (1982). Attachment and loss: Ret- Arlow, J. A., & Brenner, C. (1964). Psychoana- rospect and prospect. American Journal of lytic concepts and the structure model. Jour- Orthopsychiatry, 52, 664–678. nal of the American Psychoanalytic Association, Brenner, C. (1982). The mind in conflict. New (Monograph No. 3). York: International Universities Press. P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

698 the cambridge handbook of consciousness

Carver, C. S. (2001). Affect and the functional inconsistent information. Journal of Personality bases of behavior: On the dimensional struc- & Social Psychology, 75, 53–69. ture of affective experience. Personality and Dixon, N. F. (1971). Subliminal perception: The Social Psychology Review, 5, 345–356. nature of a controversy. London: McGraw-Hill. Classen, C. C., Palesh, O. G., & Aggarwal, R. Dixon, N. F. (1981). Preconscious processing. New (2005). Sexual revictimization: A review of the York: Wiley. empirical literature. Trauma, Violence, & Abuse, Dunning, D. (1999). A newer look: Motivated 103 129 6, – . social cognition and the schematic representa- Cloninger, C. R., Przybeck, T. R., Svrakic, D. M., tion of social concepts. Psychological Inquiry, 10, & Wetzel, R. D. (1994). The Temperament and 1–11. Character Inventory (TCI): A guide to its devel- Eagly, A. H., & Chaiken, S. (1998). Attitude struc- opment and use. St. Louis: Washington Univer- ture and function. In D. T. Gilbert, S. T. Fiske, sity Center for Psychobiology of Personality. & G. Lindzey (Eds.), The handbook of social Cohen, J. (1977). Statistical power analysis for the psychology (4th ed., pp. 269–322). New York: behavioral sciences. Hillsdale, NJ: Erlbaum. McGraw-Hill. Cohen, L. H., Towbes, L. C., & Flocco, R. (1988). Edwards, W. (1977). How to use multiattribute Effects of induced mood on self-reported life utility measurement for social decisionmaking. events and perceived and received social sup- IEEE Transactions on Systems, Man, & Cyber- port. Journal of Personality and Social Psychol- netics, 7, 326–340. ogy, 55, 669–674. Edwards, W., & Newman, J. (1986). Multiat- Conklin, C. Z., & Westen, D. (2005). Borderline tribute evaluation. In H. R. Arkes & K. R. Ham- personality disorder in clinical practice. Ameri- mond (Eds.), Judgment and decision making: An can Journal of Psychiatry, 162, 867–875. interdisciplinary reader (pp. 13–37). New York: Corr, P. J. (2002). J. A. Gray’s reinforcement sen- Cambridge University Press. sitivity theory: Tests of the joint subsystems Egan, J. (2000, Sept. 13). RATS ad: Sub- hypothesis of anxiety and impulsivity. Person- liminal conspiracy? Retrieved from www. ality and Individual Differences, 33, 511–532. news.bbc.co.uk. Crowley, C. (2000, Sept. 12). Bush says “RATS” ad Elliot, A. J., & Thrash, T. M. (2002). Approach- not meant as subliminal message; Gore calls ad avoidance motivation in personality: Approach “disappointing development.”Retrieved from and avoidance temperaments and goals. Journal www.CNN.com of Personality and Social Psychology, 82, 804– 818 Davidson, R. J., Jackson, D. C., & Kalin, N. H. . (2000). Emotion, plasticity, context, and Erdelyi, M. (1974). A new look at the New Look: regulation: Perspectives from affective neuro- Perceptual defense and vigilance. Psychological science. Psychological Bulletin, 126, 890–909. Review, 81, 1–25. Dawes, R. M., Faust, D., & Meehl, P. E. (1989). Erdelyi, M. H. (1985). Psychoanalysis: Freud’s cog- Clinical versus actuarial judgment. Science, nitive psychology. New York: W. H. Freeman. 243(4899), 1668–1674. Eriksen, C. W. (1959). Discrimination and learn- DiLillo, D. K., Long, P. J., & Russell, L. M. ing without awareness: A methodological sur- (1994). Childhood coping strategies of intrafa- vey and evaluation. Psychological Review, 67, milial and extrafamilial female sexual abuse 279–300. victims. Journal of Child Sexual Abuse, 3, 45– Farrow, T. F. D., Zheng, Y., Wilkinson, I. D., 65. Spence, S. A., Deakin, J., Tarrier, N., et al. Ditto, P. H., Munro, G. D., Apanovitch, A. M., (2001). Investigating the functional anatomy Scepansky, J. A., & Lockhart, L. K. (2003). of empathy and forgiveness. Neuroreport, 12, Spontaneous skepticism: The interplay of 2433–2438. motivation and expectation in responses to Fonagy, P., Gergely, G., Jurist, E. L., & Target, favorable and unfavorable medical diagnoses. M. (2002). Affect regulation, mentalization, and Personality & Social Psychology Bulletin, 29(9), the development of the self. New York: Other 1120–1132. Press. Ditto, P. H., Scepansky, J. A., Munro, G. D., Freud, A. (1936). The ego and the mechanisms of Apanovitch, A. M., & Lockhart, L. K. defense. New York: International Universities (1998). Motivated sensitivity to preference- Press. P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

motivation, decision making, and consciousness 699

Freud, S. (1901). Psychopathology of everyday life. Huprich, S. K., & Greenberg, R. P. (2003). London: T. Fischer Unwin. Advances in the assessment of object relations Freud, S. (1953). The interpretation of dreams. In in the 1990s. Clinical Psychology Review, 23, J. Strachey (Ed. & Trans.), The standard edition 665–698. of the complete psychological works of Sigmund Jahoda, M. (1977). Freud and the dilemmas of psy- Freud (Vol. 4,pp.1–338). London: Hogarth chology. London: Hogarth Press. 1900 Press. (Original work published ) James, W. (1890). The principles of psychology. Freud, S. (1957). The unconscious. In J. Stra- New York: Holt. chey (Ed. & Trans.), The standard edition of the Jost, J. T., Glaser, J., Kruglanski, A. W., & Sul- complete psychological works of Sigmund Freud loway,F.J.(2003). Political conservatism as 14 159 215 (Vol. ,pp. – ). London: Hogarth Press. motivated social cognition. Psychological Bul- 1915 (Original work published ) letin, 129(3), 339–375. 1926 Freud, S. ( ). The ego and the id. Honolulu, HI: Kahneman, D., & Tversky, A. (Eds.). (2000). Hogarth Press. Choices, values, and frames. New York: Cam- Freud, S. (1933). New introductory lectures on bridge University Press. psycho-analysis. New York: W. W. Norton. Kernberg, O. (1975). Borderline conditions and Gigerenzer, G., & Goldstein, D. G. (1996). Rea- pathological narcissism. Northvale, NJ: Jason soning the fast and frugal way: Models of Aronson. bounded rationality. Psychological Review, 103, Kernberg, O. F. (2004). Borderline person- 650 669 – . ality disorder and borderline personality Gigerenzer, G., & Selten, R. (2001). Rethinking organization: Psychopathology and psy- rationality. In G. Gigerenzer & R. Selten (Eds.), chotherapy. In J. J. Magnavita (Ed.), Handbook Bounded rationality: The adaptive toolbox of personality disorders: Theory and practice (pp. 13–36). Cambridge, MA: MIT Press. (pp. 92–119). New York: Wiley. Gilbert, D. T., Pinel, E. C., Wilson, T. D., Blum- Kihlstrom, J. F. (1999). The psychological uncon- berg, S. J., & Wheatley, T. P. (2002). Durabil- scious. In L. A. Pervin & O. P. John (Eds.), ity bias in affective forecasting. In T. Gilovich, Handbook of personality: Theory and research D. Grifﬁn, & D. Kahneman (Eds.), Heuristics (2nd ed., pp. 424–442). New York: Guilford and biases: The psychology of intuitive judgment Press. 292 312 (pp. – ). New York: Cambridge Univer- Kihlstrom, J. (2000). Freud’s inﬂuence on psy- sity Press. chology has been that of a dead weight. In R. Gill, M. M. (1967). The collected papers of David I. Atkinson, R. C. Atkinson, E. E. Smith, D. Rapaport. New York: Basic Books. J. Bem, & S. Nolen Hieksema (Eds.), Hilgard’s Gray, J. A. (1990). Brain systems that mediate introduction to psychology (13th ed., pp. 480). both emotion and cognition. Cognition and New York: Harcourt Brace. Emotion, 4, 269–288. Kihlstrom, J. F., Barnhardt, T. M., & Tataryn, D. J. Greenberg, J. R., & Mitchell, S. (1983). Object (1992). The psychological unconscious: Found, relations in psychoanalytic theory. Cambridge, lost, and regained. American Psychologist, 47, MA: Harvard University Press. 788–791. Guntrip, H. (1971). Psychoanalytic theory, therapy, Klein, G. S. (1976). Freud’s two theories of sexu- and the self. New York: Basic Books. ality. In L. Breger (Ed.), Clinical-cognitive psychology (pp. 136–181). Englewood Cliffs, NJ: Hardaway, R. A. (1990). Subliminally activated Prentice-Hall. symbiotic fantasies: Facts and artifacts. Psycho- logical Bulletin, 107, 177–195. Kohut, H. (1966). Forms and transformations of narcissism. Journal of the American Psychoana- Hartmann, H. (1958). Ego psychology and the lytic Association, (2), 243–272. problem of adaptation. Oxford: International 14 Universities Press. (Original work published Kohut, H. (1971). The analysis of the self: A sys- 1939) tematic approach to the treatment of narcissistic personality disorders. New York: International Hartmann, H., Kris, E., & Loewenstein, R. M. Universities Press. (1946). Comments on the formation of psychic structure. Psychoanalytic Study of the Child, 2 , Kunda, Z. (1990). The case for motivated reason- 11–38. ing. Psychological Bulletin, 108, 480–498. P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

700 the cambridge handbook of consciousness

LeDoux, J. E. (1986). Sensory systems and Non-interpretive mechanisms in psychoanalytic emotion: A model of affective processing. Inte- therapy: The ‘something more’ than interpre- grative Psychiatry, 4, 237–243. tation. (1998). International Journal of Psycho- Mahoney, M. J. (1977). Reflections on the Analysis, 79, 903–921. cognitive-learning trend in psychotherapy. O’Connor, L. E., Berry, J. W., Weiss, J., American Psychologist, 32, 5–13. Schweitzer, D., & Sevier, M. (2000). Survivor Main, M., Kaplan, N., & Cassidy, J. (1985). Secu- guilt, submissive behaviour and evolutionary rity in infancy, childhood, and adulthood: A theory: The down-side of winning in social move to the level of representation. Mono- comparison. British Journal of Medical Psychol- graphs of the Society for Research in Child Devel- ogy, 73, 519–530. opment, 50(1–2), 66–104. Olds, D. D. (1994). Connectionism and psycho- Marcus, G. E., Newman, W. R., & MacKuen, M. analysis. Journal of the American Psychoanalytic (2000). Affective intelligence and political judg- Association, 42, 581–611. ment. Chicago: University of Chicago Press. Panksepp, J. (2005). Affective consciousness: Mellers, B. (2000). Choice and the relative plea- Core emotional feelings in animals and sure of consequences. Psychological Bulletin, humans. Consciousness & Cognition: An Inter- 126, 910–924. national Journal, 14, 30–80. Mikulincer, M., & Arad, D. (1999). Attachment Petty, R., & Cacioppo, J. (1986). Communication working models and cognitive openness in and persuasion: Central and peripheral routes to close relationships: A test of chronic and tem- attitude change. New York: Springer-Verlag. porary accessibility effects. Journal of Personal- Pierce, T., & Lydon, J. (1998). Priming relational ity and Social Psychology, 77, 710–725. schemas: Effects of contextually activated and Mikulincer, M., Birnbaum, G., Woddis, D., & chronically accessible interpersonal expecta- Nachmias, O. (2000). Stress and accessibility tions on responses to a stressful event. Journal of proximity-related thoughts: Exploring the of Personality and Social Psychology, 75, 1441– normative and intraindividual components of 1448. attachment theory. Journal of Personality and Plutchik, R. (1980). Emotions: A psychoevolution- 509 523 Social Psychology, 78, – . ary synthesis. New York: Harper & Row. Mikulincer, M., Gillath, O., Halevy, V., Avihour, Pribram, K., & Gill, M. (1976). Freud’s project 2001 N., Avidan, S., & Eshkoli, N. ( ). Attach- reassessed. New York: Basic Books. ment theory and reaction to others’ needs: Evi- 1967 dence that activation of the sense of attach- Rapaport, D. ( ). The scientific method- ment security promotes empathic responses. ology of psychoanalysis. In M. Gill (Ed.), The collected papers of David Rapaport. New Journal of Personality and Social Psychology, 81, 1205–1224. York: Basic Books. (Original work published 1944) Mikulincer, M., Gillath, O., & Shaver, P. R. 2003 (2002). Activation of the attachment system Rhodewalt, F., & Sorrow, D. L. ( ). Interper- in adulthood: Threat-related primes increase sonal self-regulation: Lessons from the study the accessibility of mental representations of of narcissism. In M. R. Leary & J. P. Tangney 519 attachment figures. Journal of Personality & (Eds.), Handbook of self and identity (pp. – 535 Social Psychology, 83, 881–895. ). New York: Guilford Press. Mikulincer, M., & Shaver, P. R. (2001). Attach- Rosenthal, R. (1979). The file drawer problem ment theory and intergroup bias: Evidence that and tolerance for null results. Psychological Bul- priming the secure base schema attenuates neg- letin, 86, 638–641. ative reactions to out-groups. Journal of Person- Russ, E., Bradley, R., Shedler, J., & Westen, ality and Social Psychology, 81, 97–115. D. (2005). Refining the narcissistic diagnosis: Mikulincer, M., Shaver, P. R., & Pereg, D. (2003). Defining criteria, subtypes, and endophenotypes. Attachment theory and affect regulation: The Unpublished manuscript, Emory University, dynamics, development, and cognitive conse- Atlanta, GA. quences of attachment-related strategies. Moti- Sandler, J., & Rosenblatt, B. (1962). The concept vation & Emotion, 27, 77–102. of the representational world. Psychoanalytic Mitchell, S. A. (1988). Relational concepts in Study of the Child, 17, 128–145. psychoanalysis: An integration. Cambridge, Schafer, R. (1976). A new language for psychoanal- MA: Harvard University Press. ysis. New Haven, CT: Yale University Press. P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

motivation, decision making, and consciousness 701

Scharf, J. S., & Scharf, D. (1998). Object relations Thagard, P. (1989). Explanatory coherence. individual therapy. New York: Jason Aronson. Behavioral and Brain Sciences, 12, 435–467. Schimek, J. G. (1975). A critical re-examination Thagard, P. (2000). Coherence in thought and of Freud’s concept of unconscious mental rep- action. Cambridge, MA: MIT Press. resentation. International Review of Psycho- Thagard, P. (2002). The passionate scientist: Analysis, 2 , 171–187. Emotion in scientific cognition. In P. Car- Shapiro, W. (2000, Sept. 12). Fear of sublim- ruthers, S. Stich, & M. Siegal (Eds.), The cog- inal advertising is irrational. Retrieved from nitive basis of science (pp. 235–250). New York: www.USA.com Cambridge University Press. Shaver, P. R., & Mikulincer, M. (2005). Attach- Thagard, P. (2003). Why wasn’t O. J. convicted? ment theory and research: Resurrection of the Emotional coherence in legal inference. Cogni- psychodynamic approach to personality. Jour- tion and Emotion, 17, 361–385. nal of Research in Personality, 39, 22–45. Thagard (in press). Shedler, J., & Westen, D. (2004). Refining per- Tomkins, S. S. (1962). Affect, imagery, sonality disorder diagnoses: Integrating science consciousness: Vol. I. The positive affects. and practice. American Journal of Psychiatry, Oxford: Springer. 161, 1350–1365. Tversky, A., & Kahneman, D. (1974). Judgment Shevrin, H., & Dickman, S. (1980). The psycho- under uncertainty: Heuristics and biases. Sci- logical unconscious: A necessary assumption ence, 185, 1124–1131. for all psychological theory? American Psychol- Vaillant, G. (1977). Adaptation to life. Boston: Lit- ogist, 35, 421–434. tle Brown. Siegel, P., & Weinberger, J. (1998). Capturing the Vaillant, G. (Ed.). (1992). Ego mechanisms of “mommy and I are one” merger fantasy: The defense: A guide for clinicians and researchers. oneness motive. In R. F. Bornstein & J. M. Washington, DC: American Psychiatric Asso- Masling (Eds.), Empirical perspectives on the psy- ciation Press. choanalytic unconscious (pp. 71–97). Washing- Vaillant, G. E., & McCullough, L. (1998). The ton, DC: American Psychological Association. role of ego mechanisms of defense in the diag- Silverman, L. H., Bronstein, A., & Mendelsohn, nosis of personality disorders. In J. W. Barron E. (1976). The further use of the sublimi- (Ed.), Making diagnosis meaningful: Enhancing nal psychodynamic activation method for the evaluation and treatment of psychological disor- experimental study of the clinical theory of ders (pp. 139–158). Washington, DC: American psychoanalysis: On the specificity of the rela- Psychological Association. tionship between symptoms and unconscious von Helmholtz, H. (1909). Handbuch der Physiol- conflicts. Psychotherapy: Theory, Research & ogischen Optik (3rd ed.). Leipzig: Leopold Voss. 2 16 Practice, 13, – . Weinberger, J. (1992). Validating and demystify- Silverman, L. H., Lachmann, F. M., & Milich, R. ing subliminal psychodynamic activation. In R. (1982). The search for oneness. New York: Inter- F. Bornstein & T. S. Pittman (Eds.), Perception national Universities Press. without awareness: Cognitive, clinical, and social Silverman, L. H., & Weinberger, J. (1985). perspectives (pp. 170–188). New York: Guilford Mommy and I are one: Implications for psy- Press. chotherapy. American Psychologist, 40, 1296– Weinberger, J. (2000). William James and the 1308. unconscious: Redressing a century-old misun- Simon, H. A. (1967). Motivational and emotional derstanding. Psychological Science, 6, 439–445. controls of cognition. Psychological Review, 74, Weinberger, J. (in press). The unconscious. New 29–39. York: Guilford Press. Simon, H. A. (1984). Commentary. In L. S. Weinberger, J., & Hardaway, R. (1990). Separat- Sproull & P. D. Larkey (Eds.), Advances in infor- ing myth from reality in subliminal psycho- mation processing in organizations (pp. 169–171). dynamic activation. Clinical Psychology Review, Greenwich, CT: JAI Press. 10, 727–756. Simon, H. A. (1990). Invariants of human behav- Weinberger, J., & Westen, D. (2005). RATS, ior. Annual Review of Psychology, 41, 1–19. we should have used Clinton: Subliminal stim- Tetlock, P., & Mellers, B. (2002). The great ratio- ulation and presidential politics. Unpublished nality debate. Psychological Science, 13, 94–99. manuscript, Adelphi University. P1: KAE 0521857430c24 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:17

702 the cambridge handbook of consciousness

Weiss, J., & Sampson, H. (1986). The psychoana- Westen, D. (2000a). Commentary: Implicit and lytic process: Theory, clinical observation & empir- emotional processes in cognitive-behavioral ical research. New York: Guilford Press. therapy. Clinical Psychology: Science & Practice, Westen, D. (1985). Self and society: Narcissism, 7, 386–390. collectivism, and the development of morals. New Westen, D. (2000b). Integrative psychotherapy: York: Cambridge University Press. Integrating psychodynamic and cognitive- Westen, D. (1990a). The relations among nar- behavioral theory and technique. In C. R. Sny- cissism, egocentrism, self-concept, and self- der & R. Ingram (Eds.), Handbook of psychologi- esteem: Experimental, clinical and theoretical cal change: Psychotherapy processes and practices considerations. Psychoanalysis & Contemporary for the 21st century (pp. 217–242). New York: Thought, 13, 183–239. Wiley. Westen, D. (1990b). Towards a revised theory Westen, D., Blagov, P., Feit, A., Arkowitz, J., & of borderline object relations: Contributions Thagard, P. (2005). When reason and passion of empirical research. International Journal of collide: Cognitive and emotional constraint sat- Psycho-analysis, 71, 661–693. isfaction in high-stakes political decision mak- Westen, D. (1991). Social cognition and object ing. Manuscript submitted for publication. relations. Psychological Bulletin, 109, 429–455. Westen, D., & Gabbard, G. O. (2002). Devel- Westen, D. (1994). Toward an integrative model opments in cognitive neuroscience: I. Conflict, of affect regulation: Applications to social- compromise, and connectionism. Journal of the 53 psychological research. Journal of Personality, American Psychoanalytic Association, 50, – 98 62, 641–667. . Westen, D. (1997). Towards a clinically and Westen, D., Kelley, M., & Abramowitz, A. 2005 empirically sound theory of motivation. Inter- ( ). Is GDP in the eyes of the beholder? Cog- national Journal of Psycho-analysis, 78, 521– nitive and emotional constraint satisfaction in 548. voters’ perceptions of the economy. Unpublished manuscript, Emory University, Atlanta, GA. Westen, D. (1998a). Affect regulation and psychopathology: Applications to depression Westen, D., Kilts, C., Blagov, P., Harenski, K., & 2006 and borderline personality disorder. In W. Flack Hamman ( ). The neural basis of motivated & J. Laird (Eds.), Affect and psychopathology. reasoning: An fMRI study of emotional con- New York: Oxford University Press. straints on political judgment during the U.S. presidential election of 2004. Journal of Cogni- Westen, D. (1998b). The scientific legacy of tive Neuroscience, 18, 1947–1958. Sigmund Freud: Toward a psychodynamically informed psychological science. Psychological Westen, D., Novotny, C. M., & Thompson- 2004 Bulletin, 124, 333–371. Brenner, H. ( ). The empirical status Westen, D. (1998c). Unconscious thought, feel- of empirically supported psychotherapies: ing, and motivation: The end of a century- Assumptions, findings, and reporting in con- Psychological Bulletin, long debate. In R. F. Bornstein & J. M. Masling trolled clinical trials. 130, 631 663 (Eds.), Empirical perspectives on the psychoana- – . lytic unconscious (Vol. 7,pp.1–43). Washington, Westen, D., & Weinberger, J. (2004). When clin- DC: American Psychological Association. ical description becomes statistical prediction. 595 613 Westen, D. (1999a). Psychodynamic theory and American Psychologist, 59, – . technique in relation to research on cognition Westen, D., & Weinberger, J. (2005). In and emotion: Mutual implications. In T. Dal- praise of clinical judgment: Meehl’s forgot- gleish & M. Power (Eds.), Handbook of cognition ten legacy. Journal of Clinical Psychology, 61, and emotion (pp. 727–746). New York: Wiley. 1257–1276. Westen, D. (1999b). The scientific status of Williams, J. M. G., Mathews, A., & MacLeod, unconscious processes: Is Freud really dead? C. (1996). The Emotional Stroop Task and Journal of the American Psychoanalytic Associ- psychopathology. Psychological Bulletin, 120, ation, 47, 1061–1106. 3–24. P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

Part II THE NEUROSCIENCE OF CONSCIOUSNESS

703 P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

704 P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

A. Neurophysiological Mechanisms of Consciousness

705 P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

706 P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

CHAPTER 25 Hunting the Ghost: Toward a Neuroscience of Consciousness

Petra Stoerig

Abstract This chapter reviews evidence pertaining to the neural basis of state and trait con- Consciousness is a term with many mean- sciousness (Stoerig, 2002). As we (still?) lack ings. In one sense, we use the term to indian objective marker of conscious represen- cate whether or not an organism is in a con- tations that is independent of the subjects’ scious state. In this sense, consciousness is overt behaviour, we can assess conscious what is altered, reduced, or even lost when representations only when our subjects can we faint or undergo deep general anaes- access and (verbally or non-verbally) express thesia. In a second sense, it is a trait, an them. The same applies to the conscious attribute of a psychological process; we may state, because it is only observable conscious think, desire, hear, see, and feel consciously, access that precludes a diagnosis of uncon- thereby becoming conscious of thoughts, sciousness. If conscious representations as wishes, voices and music, of colours and tex- well as state consciousness are attributed on tures. Local anaesthesia, by abolishing con- the basis of evidence for conscious access, scious sensations of touch in the affected the neuronal processes that mediate con- limb, thus interferes with the consciousness scious access in its many forms are likely of something. Within trait consciousness, I to contaminate what we learn about those draw a further distinction between conscious that mediate conscious representations and representations and conscious access. Con- states. A prominent candidate for mediating scious representations are of objects or con- conscious access is a network of frontopari- tents of perceptions, desires, or actions; they etal cortical regions that play an important are usually phenomenal, like a strawberry role in attentional and behavioural selec- that is red, sweet. and heart-shaped and that tion of incoming and stored information. As I see, smell, and desire to eat. To describe these regions allocate processing resources the strawberry’s looks and taste or to resist and guide behavioural selection, it is not sur- its temptation, conscious access to its repre- prising that they are activated both when sentation is required (Block, 1995). vegetative state patients recover and when 707 P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

708 the cambridge handbook of consciousness

healthy subjects perform demanding per- wakefulness, characterized by activation of ceptual tasks. Although the frontoparietal both; and REM sleep (rapid eye movement network has been implicated in the medi- sleep) during which the cortex becomes ation of the conscious state, of conscious regionally activated, but the spinal fibres are representations, and of conscious access, I actively inhibited. This inhibition is crucial argue that these different manifestations of to prevent the motor signals that the cortex consciousness may well depend on different generates while we sleep from reaching neuronal processes. the muscles. A small lesion in the locus coeruleus interferes with this inhibition, so that a sleeping cat, albeit unresponsive to Consciousness as a State sensory stimulation, exhibits walking or run- of an Organism ning movements or even full-fledged fight- ing behaviour (Jouvet & Delorme, 1965); Is it possible to identify a system of neuronal the intact animal will execute only small structures that mediate all conscious states twitching movements in REM sleep. and whose destruction or downregulation In humans, REM sleep behaviour dis- abolishes consciousness? It should be a non- orders that are a consequence of dysfunc- specific system because we are in a conscious tions of the balance between endogenous state regardless of whether we listen to a activation and inhibition come in different concert, indulge in French cuisine, or suffer degrees of severity. In severe cases, sleep- from heartburn afterward; we are conscious ers can be propelled to perform violent (‘bei Bewusstsein’) regardless of what we actions that endanger themselves as well are conscious of. Being in a conscious state as their partners and can be most difficult is prerequisite to any conscious experience, to stop because external sensory input is and alterations in the state cause changes inhibited during REM sleep and so fails to in the experience. The following sections wake them. Upon awakening and recon- discuss sleep as an example of circadian necting with the environment, the patients endogenous state changes; general anaesthe- may report wild dreams that seem consis- sia as an example of chemically induced tent with the behaviour they displayed, indi- state changes; and comatose, vegetative, and cating dream enactment (Schenck, Bundlie, minimally conscious states as examples of Ettinger, & Mahowald, 1986). A young man pathology-induced state changes. The pur- recurrently dreamt he was a big cat let out of pose is to provide up-to-date information his cage by a zookeeper who offered him raw on several aspects of consciousness, as well meat that he could not snatch. His behaviour as their interdependence, and to point out during these periods reflected the dream: some critical confounds. He would prowl about the house, open the refrigerator with his mouth, or lift a mattress with his jaws and drag it about (Schenck et Sleep al., 1989,p.195). Sleep is regularly and actively induced by These REM sleep behaviour disorders dif- a shift in neuronal activity and neurotrans- fer from non-REM sleep disorders of which mitter balance in brainstem nuclei. In their sleep walking is probably the best known; seminal work, Morruzzi and Magoun (1949) sleep eating and sleep sex are other exam- discovered a network of neurons extending ples (Schenck & Mahowald, 1994). In all of from the medulla oblongata through the these disorders, the patients initiate seem- midbrain up to the diencephalon that see- ingly goal-directed activities; they will for med to regulate both the activity of the brain instance get up and walk to the fridge, and the spinal cord. We now know that three carry the food they find back to bed, and states are balanced within this network: non- devour it. Clearly, the behaviours are com- REM sleep, characterized by inhibition of plex, and sequences may last up to 1 or 2 the spinal cord and large parts of the cortex; hours. More tragic results than waking up P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

hunting the ghost: toward a neuroscience of consciousness 709

Table 25.1. The incidence of dream reports is lower when subjects are woken from non-REM sleep; absolute values depend on which non-REM stage preceded the waking, and what is counted as a dream. Authors % Dream Reports From REM Non-REM Aserinsky & Kleitmann, 1953 74 22 Dement & Kleitmann, 1957 79.67 Foulkes, 1962 87 74 Hobson et al., 1965 87.237.2 Kales et al., 1967 83 35 Foulkes and Schmidt, 1983 93 67 Cavallero et al., 1992 89.264.5 Bosinelli, 1995 89/93 65/77 (SWS) Cicogna et al., 2000 93.33 62.62 (SWS) Nielsen, 2000 81.9+/−.943+/−20.8

satiated in a bed messed up with food are also movie-like dreams (see Bosinelli, 1995, evidenced by a series of crimes committed – for example). The higher incidence of dream to the best judgement of the specialists and reports from REM sleep is a probability jury – in this altered state of consciousness effect, not an absolute difference. (Broughton et al., 1994; Hartmann, 1983). Currently, the EEG is used to classify sleep Unlike people awakening from REM sleep stages (Rechtschaffen & Kales, 1968), but behaviours, the non-REM disorder patients no physiological marker for the absence or usually recall neither a dream nor another presence of dream mentation has been iden- aspect of behaviour related to the episodes, tiﬁed. Therefore functional neuroimaging indicating that the neurochemistry govern- studies performed on sleeping subjects can ing REM and non-REM sleep is likely to provide insights into the regional brain acti- differentially affect recall after waking (see vation patterns that characterize sleep and Chapter 16 for more information on the neu- its different stages, but cannot differentiate rochemistry of sleep disorders). between periods with and without menta- The electroencephalographic characteris- tion. During both REM and non-REM sleep, tics of REM sleep – a low-amplitude high- the prefrontal and parietal cortical regions frequency EEG resembling that recorded are deactivated in comparison to the wakeful during waking with eyes open – are sugges- ‘resting state’ (Braun et al., 1997; Maquet, tive of dream experience. As dynamic, mul- 2000; Maquet et al., 1996). ‘Resting state’ tisensory dreams are reported more often describes a situation in which subjects lie, when sleepers are woken from REM rather with their eyes closed, in the scanner and, than non-REM phases, sleep researchers despite uncontrolled differences in inter- since Kleitman (1963) have tended to equate nally ongoing emotional and cognitive activ- dream and REM sleep. However, studies ity, are attentive to the environment. The that have used more sensitive measures to most active regions in the resting state are assess reports of mental activity prior to wak- the (left) dorsolateral and medial prefrontal ing (Foulkes, 1962) have shown that menta- areas, the inferior parietal cortex, and the tion is reported quite commonly after wak- posterior cingulate/precuneus (see Maquet, ing from REM as well as non-REM phases 2000; Shulman et al., 1997, for reviews). In (see Table 25.1). Even when woken from the slow-wave sleep (SWS, non-REM stages 3 deepest of the non-REM sleep stages, stage and 4 with high-amplitude low-frequency 4, which is characterized by a low-frequency, EEG), regional cerebral blood ﬂow decreases high-amplitude EEG, subjects may report in these as well as most other parts of the not only some thought-like mentation but brain including the thalamus and brainstem; P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

710 the cambridge handbook of consciousness

awake asleep

Figure 25.1. The relative contributions of exogenous stimulation (white) and endogenous activity (stippled) to mental life change during the waking period. Here, permeability is considerably higher than during sleep where the contribution of externally induced activity is systematically reduced.

only the perirolandic (sensorimotor) and subjects were found to be significantly more occipital (visual) cortices were reported to likely to detect hidden rules to a task when remain unaffected (Kajimura et al., 1999). they were allowed to sleep before continuing In REM sleep, despite overall increases in in its execution (Wagner et al., 2004). The cerebral blood flow and energy demands, activation of limbic structures and sensory relatively low regional cerebral blood flow association cortices seen during REM sleep (rCBF) persists in prefrontal and parietal cor- may provide the emotional colour and the tex, whereas relative activations are seen in multisensory phenomenal content of REM the brainstem, hypothalamus, and thalamus sleep dreaming; both higher auditory and where the nuclei that regulate sleep stages visual cortical areas appear quite reliably are housed (Hobson, 1989; Steriade, 2000; activated in this sleep stage. Chapter 16) and in the amygdalae and poste- A further and major difference between rior association cortices (Braun et al., 1997; waking and dreaming experience/mentation Nofzinger et al., 1997). Interestingly, the lies in its origin. In the waking state, inter- posterior cingulate-precuneal region that is nal (thoughts, imagery, feelings, memories) significantly deactivated in both SWS and and external (sensations) components of REM sleep (e.g. Braun et al., 1997; Maquet mental life shift in their relative contri- et al., 1996) was selectively activated in butions throughout the day, but always a SPECT-study of sleepwalking (Bassetti remain penetrable to each other, whereas et al., 2000). The increased activity in this in the sleeping state, the external contribu- posterior midline area in the face of persis- tion is consistently reduced (see Fig. 25.1). tent frontal deactivation suggests that this Electrophysiological recordings show that non-REM sleep, and possibly other sleep sensory stimuli generate only the early com- behaviour disorders, results from pathologi- ponents of the waking state evoked poten- cal combinations of neurobiological features tials (Yamada et al., 1988), and neuroimag- of sleep and wakefulness (Mahowald & ing of responses to auditory stimulation (text Schenck, 2001). being read to the sleeper) during non-REM Activation patterns observed during wak- sleep produced less auditory cortical activa- ing and sleeping indicate that the mental tion than in the waking state. In addition, activity during sleep differs from that during in the visual cortex a pronounced negative waking. The observed deactivation of pre- BOLD response was seen not only during frontal cortex seems to preclude neither the text reading (Czisch et al., 2002) but also in sometimes fantastic creativity expressed in response to visual stimulation (8 Hz flicker- dreams nor the cognitive insights that anec- ing light), and it corresponded to a decrease dotally occur during sleep. Only recently in rCBF in the same visual cortical areas (ros- has the role that sleep plays for the latter tromedial occipital cortex) in additional vol- 15 been demonstrated experimentally, when unteers who underwent H2 O PET during P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

hunting the ghost: toward a neuroscience of consciousness 711

SWS (Born et al., 2002). One of the possible interpretations of this finding is that external stimulation interferes with endogenous visual cortex activation. At present we cannot get information about the experiences of sleepers without waking them and thereby changing their state, and therefore we have no means to determine confidently whether a lack of report indicates a lack of experience or a lack of recall. At the same time, reports need not necessarily reflect mentation just before waking; if there were periods of true Figure 25.2. The most pronounced effects of nothingness, the reports could also reflect propofol (Fiset et al., 1999) are shown together the last event that left a reportable trace. with those induced by halothane and isoflurane Although the amount and movie-like char- inhalation (Alkire et al., 2000). Regional cerebral acter of dreams may differ during differ- blood flow was measured in the first and regional ent sleep stages and unconscious periods cerebral glucose metabolism in the second study. cannot be ruled out, sleep as such is not (With kind permission from Michael Alkire.) an unconscious state, but rather one that (See color plates.) alters consciousness by closing the shut- ters on the largest part of external sensory their central placement, be accessed in very input. different ways, or whether there are no core structures and the conscious state instead depends on a functional property of the net- Anaesthesia work. General anaesthesia is induced by a vari- Although earlier neuroimaging data that ety of very different chemical compounds. compared general anaesthesia with pre- This is probably the major reason why, at or post-anaesthetic states showed that present, we have no generally accepted the- the tested pharmaceutical agents produce ory of general anaesthesia. The major con- an overall reduction in brain metabolism tenders include (1) inhibition of excitatory (Alkire et al., 1995), more sensitive analy- receptor channels together with potentia- sis revealed that in addition to the fronto-

tion of inhibitory ones that include GABAA parietal resting state network, cerebel- (γ -aminobutyric acid), glycine, nicotinergic lar, frontobasal, and thalamo-mesencephalic acetylcholine, and serotonin (Franks & Lieb, brain regions are relatively more deactivated 1994; 1996); (2) interference with NMDA (Alkire et al., 2000; Fiset et al., 1999; see (N-methyl-D-aspartate) receptor function Fig. 25.2). Despite some agent-specific dif- (Anis, Berry, Burton, & Lodge, 1983; Flohr, ferences, the tested anaesthetics – halothane, 1995, Flohr, Glade, & Motzko, 1998); and (3) propofol, and isoflurane – thus produced changes in polymerization of microtubules considerable overlap regarding the most (Allison & Nunn, 1968; Hameroff, 1998). affected brain regions, suggesting the possi- Although all these hypotheses focus on some bility of a central core in the mesencephalic specific cell sites, none implies that a partic- reticular formation and the thalamus (Alkire ular brain structure (or set of brain struc- et al., 2000). Support for a special role of the tures) is especially involved in the media- thalamus and its reciprocal network of con- tion of the conscious state. It is still open to nections to the cortex comes from a variety discussion whether this reflects the fact that of sources: (1) Functional thalamo-cortical core structures that really are indispensable connectivity is altered in anaesthesia (White for the conscious state can, possibly due to & Alkire, 2003); (2) the EEG, which at P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

712 the cambridge handbook of consciousness

light doses may either increase or decrease Although this effect is reliably induced its frequency depending on the anaesthetic, with such common agents as propofol and changes to slow, large-amplitude patterns isoflurane and indicates severely altered at deep levels, indicating thalamocortical processing of external stimuli, opioids and synchronization; (3) 3-Hz electrical stimu- anaesthetics like ketamine (a dissociative lation applied to thalamic nuclei like the anaesthetic that in small doses is hallu- nucleus reticularis thalami during neuro- cinogenic and in large ones causes general surgery causes behavioural arrest whose anaesthesia without immobilizing the sub- appearance resembles the absences observed jects or producing the appearance of deep during certain epileptic seizures; the patient sleep) do not flatten the mid- and long- stops responding, looks straight ahead, and latency components to the same extent; cannot remember anything about this 5-to they may in fact even enhance the ampli- 10-s period after abruptly coming to (Jasper, tude both of the auditory mid-latency and 1998; Jasper & Droogleever-Fortuyn, 1947); the 40-Hz steady-state response (Plourde, and (4) especially the non-specific thalamic Baribeau, & Bonhomme, 1997). Certainly nuclei entertain closely knit connections not neurons in cortical as well as subcortical only to brainstem nuclei but also to the cor- structures do not simply stop responding tex, the basal ganglia, and the striatum. In to sensory signals, as is also demonstrated view of their central position they are opti- unequivocally by the fact that a vast part of mally placed to transmit arousal signals as our knowledge on their response properties well as exerting both global and local influ- stems from anaesthetized animals (Hubel & ences on neocortical activity. Originally seen Wiesel, 1968). Recent neuroimaging data on as a dorsal extension of the reticular activat- visual processing in anaesthetized monkeys ing system (Morruzzi & Magoun, 1949), they extend these findings, showing activation have been suggested as central players in in numerous cortical and subcortical struc- the mediation of consciousness (e.g. Bogen, tures in response to visual stimuli (Leopold, 1995; Purpura & Schiff, 1997; Chapter 27). Plettenberg, & Logothetis, 2003; Logothetis, The thalamus, and especially its non-specific Guggenberger, Peled, & Pauls, 1999). Human nuclei (or neurons, see Jones, 1998), may subjects stimulated tactually while under- thus play an important role in maintaining going stepped propofol (Bonhomme et al., the conscious state. 2001) or isoflurance anaesthesia (Antognini, Although the overall reduction in brain Buonocore, Disbrow, & Carstens, 1997) metabolism effected by the majority of com- showed that activation in somatosensory cor- pounds suggests a loss of consciousness, all tex is decreased at a lower dosage than tha- the problems of ascertaining whether such lamic activation. Stroboscopic visual flicker loss occurs in sleep are exacerbated under stimulation under pentobarbital produced anaesthesia: Communication in this state is similar decreases of the BOLD signal impossible as the subjects do not respond in visual cortex, culminating in negative to verbal commands, and post-anaesthetic responses in subjects who received the high- recall of events is impeded not only by est doses relative to body weight (Martin the drugs and their effects on memory but et al., 2000). Although at increasingly also by the longer recovery period. Regard- higher concentration these specific stimulus- ing sensory information processing, audi- evoked responses decrease and eventually tory and somatosensory potentials evoked disappear (Leopold et al., 2003) and the during general anaesthesia in rats (Angel, EEG turns isoelectric, they may remain at 1991, 1993) and humans (Madler et al., 1991; the dosage used for clinical interventions. Madler & Poppel,¨ 1987) show that only Behavioural evidence, again from both the earliest deflections (∼10 ms) remain monkeys and humans, provides further evi- unaffected, whereas mid- and long-latency dence for information processing under gen- components are much reduced or absent. eral anaesthesia. Ketamine-anaesthetized P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

hunting the ghost: toward a neuroscience of consciousness 713

Alert, Behaving Ey

3 Ambiguous e V

0 elocity − 3 100 0 (deg/s) osition (deg) −100 e P

Ey 0102030405060 Time (sec) Ey

Anesthetized 20 Unambiguous e V

0 elocity − 20 200 −40 0 (deg/s) osition (deg) −200 e P

Ey 0 10 20 30 40 50 60 70 Ey

20 Ambiguous e V

10 elocity 0 200 − 100 10 0 (deg/s) osition (deg) −100 −

e P 200

Ey 0 10 20 30 40 50 60 70 Time (sec) Figure 25.3. (Top) When gratings that move in opposite directions are presented simultaneously, one to each eye, the awake monkey, who is trained to indicate motion direction this way, responds by moving the handle alternately into the one or other direction (white and black bars at bottom). The perceptual switches are also reﬂected in his optokinetic responses that follow the presently perceived grating, as shown both with respect to eye movement amplitude (left axis) and velocity (right axis). The responses consist of alternating slow phases where the eyes follow the perceived motion, and fast phases in the opposite direction. The changes in the direction of the fast phases are easily seen in the velocity plot (grey). (Bottom) In the ketamine-anaesthetized monkey, optokinetic nystagmus follows the grating in the unambiguous condition. In the rivalry condition, responses again alternate direction.

monkeys can show not only preserved ulus motion; as they correspond to manual optokinetic nystagmus in response to mov- responses given by monkey and human sub- ing visual patterns but also irregular alter- jects to indicate the perceived direction of nations in its direction when stimuli are motion, they have been used as indicators of presented in binocularly rivalrous condi- perceptual changes in monkeys (Logothetis tions (Leopold et al., 2003; see Fig. 25.3). & Schall, 1989). The presence of optokinetic When two patterns moving in opposite responses under unambiguous stimulation direction are presented, one to each eye, indicates that sensorimotor loops continue the awake subjects’ percept switches in a to function at least under relatively light quasi-regular manner between that of the ketamine anaesthesia. More surprising is the one and that of the other stimulus (Blake & ﬁnding that these responses switch direc- Logothetis, 2002). This perceptual effect is tion under ambiguous stimulation, because reﬂected in the eye movements that follow a behavioural pattern that is regarded as the perceptually dominant direction of stim- evidence for alternations between conscious P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

714 the cambridge handbook of consciousness

percepts in the awake organism (Fig. 25.3) hours post-surgery (Merikle & Daneman, has thus been demonstrated in anaesthetized 1996). monkeys. Together, the studies on information pro- The second example concerns memory cessing under general anaesthesia demon- processes. Numerous studies have investi- strate not only dose-dependent specific brain gated memory in patients undergoing sur- activity in response to stimulation but also gery under general anaesthesia with a wide significant perceptual (in binocular rivalry) variety of agents. Stimulation is usually and motor organization (OKN) in ketamine- auditory; the most common paradigm in- dissociated monkeys (Leopold et al., 2003), volves reading word lists to the patient intra- and memory formation in human patients operatively. Retrieval of this information is under general clinical anaesthesia. Whether tested postoperatively with indirect meth- this evidence ought to be seen in the con- ods, such as word-stem completion. In this text of unconscious cognition (Merikle & task, stems of used and new words are pre- Daneman, 1996) or in the context of infor- sented (as in ‘mem ...’), and the subjects are mation being processed in a state of severely asked to complete these stems with the first altered, dissociated consciousness still needs word that comes to mind. If overall they to be determined. In the former case – cogni- use words they have heard during surgery tion in a state of unconsciousness – we would (say ‘memory’ rather than ‘member’) sig- have to believe that (a) perceptual organi- nificantly more often than an alternative zation continues to a level at which alter- (and equally common) one, this is regarded native interpretations of ambiguous stimu- as evidence for implicit memory. To distin- lations are presented and (b) that memory guish this effect more clearly from explicit traces laid down in a state of unconsciousness recall, Jacoby’s inclusion-exclusion task has can be accessed when consciousness is recov- been adapted (Bonnebakker et al., 1996): ered. In the latter case – anaesthesia pro- Here, in addition to word-stem completion duces a dissociated state of consciousness – with the first word that comes to mind, both the perceptual organization and the in different series the patients are explic- memory formation would occur in a con- itly asked NOT to use the word they have scious but dissociated subject. heard before. Whereas an implicit memory trace may bias a subject’s responses toward Coma and Vegetative State repeating what was processed implicitly or explicitly (inclusion), exclusion of this same Coma is a state of unarousable unrespon- information requires intentional avoidance siveness resulting from severe pathology. of the facilitated word. As this cognitive The patient lies with the eyes closed and operation is thought to involve conscious may appear almost as if in deep sleep (see access to the memory, it provides evi- Table 25.2). However, if this state does not dence for explicit recall. Remarkably, both end with recovery or death, within weeks inclusion and exclusion procedures as well the coma will give way to a vegetative state as two-alternative forced choice (2AFC) (VS; Jennett & Plum, 1972) that may con- have revealed significant memory effects tinue for months or years (persistent VS). (Bonnebakker et al., 1996), although not In this state, patients “awake” from their every study has yielded this finding. A meta- coma and may be able to breathe inde- analysis of 44 studies encompassing 2,517 pendently. Importantly, sleep and waking patients conducted to elucidate the rea- phases alternate in VS (but not in coma), sons for the divergent results showed that although movies and the popular press often the effect of the stimulation decayed over wrongly portray ‘comatose’ patients who time. No longer significant 36 hours after sleep restfully for years. Not only does coma surgery, it was highly significant within 12 not last that long but also VS patients open hours, and still significant but less so when and close their eyes, react to strong or patients were tested between 12 and 36 painful stimulation with eye opening and P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

hunting the ghost: toward a neuroscience of consciousness 715

Table 25.2 . Overview of best-possible behaviours observed in pathology-induced state changes. State Coma VS MCS Locked-in Reactive eyeopening No Yes Yes Yes Cyclic eye opening No Yes Yes Yes Verbal utterances Grunts Jumble Intelligible No Motor behaviour Reﬂex Reﬂex In context No (eyes only) Affective behaviour No Yes In context No Breathing Variable Normal Normal Normal EEG δ, τδ, τ, slow αα, local γ Almost normal Brain metabolism 50% ∼50% ∼40% Small-medium (overall) reduction Pain experience No No Yes Yes Self-consciousness No No Yes Yes

In both VS and MCS patients, stimulation may evoke local brain activation patterns. Note that negative statements on pain experience and self-consciousness are only inferred from the behavioural response patterns.

faster breathing, and may grimace and move Among the causes for continued loss of con- their limbs during the waking phases. In sciousness are extensive fibre degeneration, addition to pupillary light, cornea, and gag- extensive necrosis of the cerebral cortex or ging reflexes, in the waking phases sponta- the thalamus, and thalamic-hypothalamic neous movements may be observed; chew- and brainstem lesions that involve the pons ing, grunting, swallowing, smiling, teeth and its tegmentum (Plum & Posner, 1982). gnashing, and brief pursuit eye movements Support for a special role of the thalamus can occur. The state is described as one of comes from a recent review of pathology ‘wakefulness without awareness’ (Jennett & data: Of 35 cases of VS caused by trauma, Plum, 1972). 80% had thalamic damage; in addition, of The vegetative state has recently been dif- 14 non-traumatic VS cases, all had severely ferentiated from the newly introduced cate- damaged thalami (Jennett, 2002). Brain gory of the minimally conscious state (MCS; metabolism was relatively most affected in Giacino et al., 2002). MCS patients may frontal and parietotemporal association cor- show islands of relatively preserved brain tices of VS patients (Kassubek et al., 2003; responses (Boly, Faymonville, & Peigneux, Laureys et al., 2002a,b, Laureys, Owen, 2004; Schiff et al., 2002), as well as frag- & Schiff, 2004; see Fig. 25.4); in addi- ments of behaviours interpretable as signs tion, functional connectivity to the intralam- of perception and voluntary movement that inar nuclei of the thalamus was altered preclude the diagnosis of vegetative state (Laureys et al., 2000a). In contrast, the (Zeman, 1997). Both the vegetative and the removal of an entire cerebral hemisphere minimally conscious state need to be dis- does not cause coma, indicating that the tinguished from the locked-in syndrome in system that mediates the conscious state is which the patient is fully conscious but, bilateral and that one half of a brain suffices due to a circumscribed brainstem lesion, is to sustain it. Neither does coma result from unable to communicate in any way other bilateral lesions that destroy both occipital, than by lid closure and vertical eye move- both parietal, both temporal, or both frontal ments. Overall brain metabolism is less lobes. reduced in locked-in patients (Levy et al., Cortical as well as subcortical activation 1987; see Table 25.2). has been observed in response to auditory Lesions causing coma, VS, and MCS may (Laureys et al., 2000b) and noxious stimu- be diffuse and metabolic as well as focal. lation of VS patients (Kassubek et al., 2003; P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

716 the cambridge handbook of consciousness

Laureys et al., 2002a). The patterns were Consciousness as Attribute or Trait subnormal and in the case of painful stim- of Psychological Processes uli did not activate the entire pain matrix. In addition, the activated areas’ functional Even when the organism is in a fully con- connectivity to the higher frontal and pari- scious and alert state, it is not capable of etal cortices seemed impaired in VS patients, being conscious of everything dealt with by but showed recovery when the patients its brain. Many brain processes are devoted 2004 recovered (Boly et al., ; Laureys et al., to the regulation of homeostasis, and at 2002 a). Nevertheless, these data do not pre- least in humans, hormone secretion, diges- clude the possibility that the VS patients still tion, breathing, immune defence, and the have some conscious experience even if it is like are in principle unconscious. We can not in context. learn about them, but have no direct access to them, and only feel the consequences of Conclusion their functions and dysfunctions. What the organism can be conscious of depends on The body of data on the three states of its sensory, cognitive, and behavioural fac- altered, reduced, dissociated, or lost con- ulties. Some of the brain processes involved sciousness demonstrates how difficult it is to in their mediation are potentially conscious, provide incontrovertible evidence for even in that we can consciously perceive, think, a transient absence of all consciousness. In feel, wish, and act upon them. However, at all instances, we need to ‘wake’ the person any point in time we are in fact only con- and require a (verbal or non-verbal) report scious of a small subset of what we can be to learn whether anything was experienced conscious of. Despite the richness of our before the ‘waking’. Even when this is pos- moment-to-moment experience, our con- sible – in sleep or anaesthesia, and to some sciousness is limited in its capacity for simul- extent after emergence from VS – the effects taneous representation. Broadbent’s bottle of the state change on memory make it dif- neck (1974) and Koestler’s administration ficult, if not impossible, to conclude any- (1968) metaphors both focus on this restric- thing from a negative report. Process fraction, illustrating that only a fraction of what tionation, discussed both in the context goes on in the brain reaches the top repre- of sleep behaviour disorders (Mahowald & sentative. Schenck, 2001) and anaesthesia (Cariani, 2000; Mashour, 2004) and referring to disso- Neuropsychological Approaches ciations within neurobiological patterns that to Conscious Representations regulate the normal stages of sleep and wakefulness, could be extended to the patholog- How can we learn in which ways brain ical state changes. Such a dissociation-based processes that can cause conscious experi- concept of state changes in general anaesthe- ences differ from those that cannot, and how sia, the vegetative, and the minimally con- those involved in mediating the presently scious state could accommodate conscious conscious differ from those that are poten- experiences whether or not they are acces- tially, but not now, consciously represented? sible to the subject after ‘waking’ and could One of the inroads to the first question is explain aspects of wakefulness manifest in the neuropsychological study of patients or vegetative state and sleep behaviour disor- animals who have suffered circumscribed der patients. brain lesions. If the lesion destroyed the That the thalamus and its thalamo- conscious representation of a particular cortico-thalamic loops that integrate specific modality or faculty, we can explore whether and non-specific neurons (Jones, 1998) play implicit (non-conscious) processes remain a special role in maintaining state conscious- and how the neuronal activity mediating ness seems likely in the light of both pathol- them and the performance they allow differ ogy and anaesthesiology. from their normal conscious counterparts. P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

hunting the ghost: toward a neuroscience of consciousness 717

Figure 25.4. The metabolically most impaired brain regions in vegetative state patients were found in prefrontal, premotor, and parietotemporal areas (Fig. 2A from Laureys et al., 2002b, reproduced with kind permission from Steven Laureys).

The best-known examples of this approach the hippocampal regions, complements pre- are probably the implicit functions that vious findings. He too presents with dense remain in amnesic and in cortically blind anterograde amnesia, but, unlike HM, is patients. incapable of recollecting episodic memories Claparede` (1911) was the first to describe from his life before the accident, although, telling evidence of implicit memory. His like HM, he can recall semantic information densely amnesic patient refused to shake his from that time. In addition to providing evi- hand after he had, on a previous occasion, dence for a dissociation between semantic pricked her with a needle he had hidden and episodic memory, he implicitly recov- between his fingers. When asked about her ered new associations between word pairs refusal, she might even explain that some- and took longer when responding to rear- times people hide needles in their hands, ranged or new pairs than to the original indicating that the experience had left a intact pairs, although he failed to recollect trace that could be incorporated into her the items on explicit testing (see Rosen- behaviour as well as her reasoning; however, baum et al., 2004, for a review). More she could not consciously recall the original evidence for implicit memory comes from painful episode. Results of extensive formal the work of Warrington and Weiskrantz testing rather than anecdotal evidence have (1968). They introduced the fragmented fig- been reported on patient HM whose dense ure test and found that their amnesic sub- anterograde amnesia resulted from bilateral jects recognized the figures at an earlier, removal of medial temporal cortex. A series more fragmented stage when having seen the of papers impressively demonstrates that set before. Other forms of implicit mem- HM, despite his superior intelligence, was ory include various forms of priming, word- unable to recall or recognize verbal as well as stem completion, and conditioning (Rovee- pictorial material if prevented from constant Collier et al., 2001; Schacter, Dobbins, & rehearsal between presentation and recall. Schnyer, 2004; Chapter 28). Nevertheless, HM showed motor learning Amnesic patients thus provide a prime and from trial to trial improved his ability to example of implicit access to informa- trace a line between the double outlines of tion that informs behaviour despite being a star that, like his pencil, were only visible unavailable to conscious recall. The infor- in a mirror. Despite never recalling having mation – the pinprick, the word pairs, performed these tasks before, HM showed the fragmented figures – is phenomenally similar improvements in solving the Tower represented as long as it is present, and of Hanoi puzzle (Cohen & Corkin, 1981; the patients’ responses to it show that they Milner, Corkin, & Teuber, 1968). Another are perfectly aware of it and able to deal thoroughly studied case, that of KC who suf- with it in multitudinous ways. The conscious fered extensive bilateral damage involving access that manifests itself in their behaviour, P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

718 the cambridge handbook of consciousness

however, is lost quickly when the informa- cal visual nuclei (Pasik & Pasik, 1982) and tion is no longer present. Implicit memory extrastriate visual cortex (Goebel et al., remains when a form of conscious acess – 2001). The blindness in blindsight can there- recall – is lost, and thus differs from the fore not be attributed to an absence of second of the widely known examples of all cortical involvement, a conclusion sup- implicit functions that survive the lesion- ported by evidence for cortical activation in induced loss of their explicit representation, a variety of other neuropsychological syn- that of blindsight (Weiskrantz,Warrington, dromes (e.g., Rees et al., 2000). More likely, Sanders, & Marshall, 1974), where cortical some functional or quantitative difference blindness prevents the conscious phenom- accounts for the loss of the conscious rep- enal respresentation of visual information. resentation. Such loss, albeit of a particu- Cortical blindness results from the destruc- lar visual feature, also characterizes cerebral tion or denervation of the primary visual cor- colour blindness that results from destruc- tex (V1). Patients with absolute field defects tion of the colour complex in the fusiform consistently claim that they do not see visual and lingual gyri (see Meadows, 1974; Zeki, stimuli that are confined to the blind field. 1990, for review) and cerebral motion blind- Nevertheless, they can exhibit non-reflexive ness that results from bilateral destruction visual functions when they are forced to of the motion complex (see Zeki, 1991, for guess whether, where, or what stimulus has review). been presented and may detect, localize, Zeki’s concept of micro-consciousnesses and discriminate visual stimuli at statisti- (Zeki, 2001) is based both on the selec- cally significant levels (Stoerig, 1999; see tive loss of visual qualia from circumscribed Fig. 25.5; Weiskrantz, 1986). Evidence for lesions of visual cortical areas and on the blindsight has also been reported in hemi- (re)appearance of residual qualia that Rid- anopic monkeys (Cowey & Stoerig, 1995) doch (1917) first observed in the cortically who behaved like the patients; they showed blind fields of some patients. The concerted excellent localization performance but nev- activity of the early visual cortical areas is ertheless treated the same stimuli as blanks necessary to provide the repertoire of our when given that response option (Stoerig, visual qualia – brightness, depth, colour, and Zontanou, & Cowey, 2002). Numerous motion – that are the phenomenal fabric of other examples of implicit visual functions conscious vision (Stoerig, 1996). Although include covert processing of colours in cere- the primary visual cortex plays first vio- bral achromatopsia (Heywood, Cowey, & lin in this concert, as its destruction abol- Newcombe, 1991), faces in prosopagnosia ishes all visual qualia in the vast major- (Bruyer et al., 1983), orientation and size in ity of cases, its precise role is still under agnosia (Milner & Goodale, 1995), as well debate. Certainly functional blindness also as deaf hearing (Mozaz Garde & Cowey, ensues when (V1) is disconnected from the 2000) and unfeeling touch (Paillard, Michel, higher visual cortical areas (Bodis-Wollner, & Stemach, 1983) following lesions of the Atkin, Raab, & Wolkstein, 1977; Horton auditory or somatosensory cortices, respec- & Hoyt, 1991). Whether this finding indi- tively. Probably it is fair to say that most cates that vision is only conscious when every loss that results from a cortical lesion higher visual, and possibly also non-visual, spares implicit functions that can be revealed areas receive retinal input via the retino- with appropriate testing. geniculo-striate–extrastriate cortical route, Studies of implicit processes and the or whether the normally massive backpro- pathways that mediate them may help jections from higher to earlier visual areas us get a better grasp of the neural basis are required for conscious vision, is still of explicit representations. Blindsight for uncertain (Hochstein & Ahissar, 2002; Sto- example seems to involve all the remaining erig & Cowey, 1993; Crick & Koch, 1995; projections from the retina onto subcorti- Kohler¨ & Moscovitch, 1997; Lamme, 2001). P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

hunting the ghost: toward a neuroscience of consciousness 719

100 100

75 P 80 ercentage a

50 60 w are

ercentage correct 25 P 40

0 0 9 20304661799199 % Contrast Figure 25.5. In a 2AFC localization task, the hemianopic patient HK performed signiﬁcantly above chance with a grating contrast of 46% or higher. When asked to indicate whether or not he was aware of the stimulus, which could appear with equal probability at either of the two target positions when he touched the start light shown beneath the ﬁxation cross in the display (left), he did not report awareness on a single trial, regardless of contrast.

Selective inactivation of feedforward and Binocular rivalry results from dichoptic feedback projections would help disentan- stimulation with incompatible images that gle this issue. cannot be fused into a meaningful percept. If one eye sees an upward moving grating while the other sees a downward moving Experiments on Normal Observers one, rather than perceiving no motion as How processes that are consciously repre- would happen if the two inputs cancelled sented at this moment differ from those that each other out, subjects see ﬁrst upward, may, but are not, presently so represented then downward, then upward moving grat- has been addressed in normal subjects. The ings in succession. As monkeys also report main paradigms manipulate the experimen- these changes when trained to indicate tal conditions so that the subjects cannot upward or downward motion (Logothetis become aware of parts of the physically pre- & Schall, 1989), brain processes during sented information; in addition, manipula- rivalrous stimulation have been recorded tion of the subjects by means of transcranial from both species in the hope that they may magnetic or electrical stimulation is used. help distinguish between processes that Note that unlike the patients who have lost are presently conscious and those that are the conscious representation of a sensory presently not conscious. Results of neural modality, the normal subjects consciously recordings in various areas of the occipito- perceive information in these manipulations temporal stream in awake behaving monkeys even if they miss whatever target they are who constantly indicate which stimulus they seeking. Binocular rivalry, change blindness, presently see have shown that the number of inattentional blindness, visual masking, rep- neurons that follow the percept (rather than etition blindness, and the attentional blink responding throughout to the always present are all examples drawn primarily from the stimulus direction they prefer) increase sub- visual domain. stantially from early to late visual cortical P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

720 the cambridge handbook of consciousness

areas (see Fig. 25.6). In the motion area switches agrees with observations in patients MT and in area V4, ∼40% of neurons with damage to the right frontal lobe who responded to their preferred stimulus only may fail to report perceptual alternations of when the monkey indicated seeing it, ambiguous figures (Ricci & Blundo, 1990). whereas in inferotemporal (IT) and superior Leopold and Logothetis (1999) use this find- temporal cortex, ∼90% of neurons showed ing to back their hypothesis that the (fron- this behaviour, indicating that binocular toparietal) sensorimotor systems punctuate rivalry is largely resolved at this advanced the visual areas and thereby actively shift the processing stage (Logothetis, 1998). perceptual interpretation by means of atten- How do these neurons, which were also tion poised for action. found in medial temporal lobe areas of Frontal regions have also been implicated humans (Kreiman et al., 2002), know which in change blindness. Here, two complex stimulus is (or should) just (be) perceived? visual stimuli, such as photographs of real- Do they determine it among themselves, in world scenes or arrangements of symbols, some kind of oscillating majority vote? If are presented in alternation. The two stim- so, do the patterns alternate because fatigue uli are similar overall, but differ in some causes the balance to shift to the suppressed feature; a person may wear a different pair population? Or does another brain proof trousers, or a mountain may shift posi- cess determine the perceptual and neuronal tion. In between stimulus presentations, a switches? Tononi, Srinivasan, Russell, and blank screen appears briefly to simulate an Edelman (1998) used magnetoencephalog- eye blink and mask the transient change, raphy and presented their subjects with rendering change detection difficult enough dichoptically presented gratings of different for subjects to require several cycles of pre- colour that flickered at different frequencies sentation (Rensink, 2000; Simons & Levin, to allow attribution of the recorded activ- 1997; see Chapter 9). Beck and colleagues ity to one or the other grating. In line with (2001) used this phenomenon in a neu- previous results (Brown & Norcia, 1997), roimaging study in which they substituted they found stronger amplitudes correlating one image of a face (or a place) for another. with the perceived stimulus. This finding is They compared activation patterns corre- consistent with more neurons firing in con- lating with stimulus presentations yielding cert, but does not elucidate how the per- change detection with those in which the ceptual switches are brought about. Evi- change went unreported and found signifi- dence in favour of a switch located outside of cant differences in frontoparietal (right mid- the visual areas comes from a neuroimaging dle frontal gyrus and bilateral superior pari- study in humans that compared brain acti- etal lobule) activation. Here, activation was vation patterns from rivalrous stimulation pronounced when the change was detected, with a control condition in which the stimuli whereas the visual areas that are responsive were physically alternating so as to produce to faces, or places, respectively, appeared not a rivalrous percept without interocular con- to differentiate between detected and unde- flict (Lumer et al., 1998). Such a fake rivalry tected changes. stimulus is useful for comparison with the Results based on a variety of paradigms real one, because it appears the same to the (e.g., detection of threshold stimuli [Pins and subject if well done and externally induces ffytche, 2003]; or detection of a moving tar- the perceptual switches that the subject’s get in a field of relative cortical blindness brain initiates on its own under conditions [Sahraie et al., 1997]) support the hypoth- of interocular conflict. The results of the esis that frontoparietal networks play some comparison of activation patterns evoked role in conscious vision. However, they do by real and fake rivalry indicated that a not address whether frontoparietal neurons frontoparietal network was more active in play an important part in the phenomenal real rivalry. That a frontoparietal network rendering of the information and, if they do, may be involved in initiating the perceptual which neurons are implicated. Conceivably, P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

hunting the ghost: toward a neuroscience of consciousness 721

Figure 25.6. In the visual cortical areas of the monkey (top), Logothetis and co-workers recorded neurons’ responses to dichoptically presented stimuli tuned to each area’s preferences. Orthogonal orientations were used for V1/V2 and V4, gratings moving in opposite directions for MT (V5), and complex images (face and starburst) for the temporal visual areas TPO and TE. The bars show the proportion of neurons that respond to their stimulus when the monkey indicated perceiving it (black) and when it was perceptually suppressed (grey). Like neurons that responded selectively only during rivalry (darker grey), these were encountered only at the intermediate processing stages V4 and MT. Note that the vast majority of neurons in the highest region investigated ﬁred only when their stimulus was also perceived. (With kind permission from Nicos Logothetis).

all or some of these activated neurons are impaired provided the second stimulus involved in attentional selection, and again follows the detected first target after a all or some of them may be required to pro- (distractor-filled) interval of ∼100–300 ms vide conscious access to that information. As (Chun & Potter, 1995). Marois, Yi, and Chun detection is inferred on the basis of report, (2004) found that activity in the parahip- and both attention and access are necessary pocampal place area (PPA) was enhanced for report, it is premature to ascribe the con- when the second of the two target stimuli (a scious detection of stimulus or change to the scene) embedded in a rapid series of scram- frontoparietal regions. bled versions of the same scenes that served Some light is shed on this issue by recent as distractors was consciously perceived; studies. An attentional blink paradigm was moreover, frontal cortex was activated only used in the first such study. In this ‘experi- when the target was also reported. Whereas mental blindness’ paradigm, different visual the first finding indicates that the PPA, stimuli are presented briefly and in rapid which responds preferentially to images alternation, and two targets are embed- of places and houses, is activated differ- ded in the series. Although the targets dif- entially when the appropriate stimulus is fer in some prominent feature from the consciously perceived, the latter result im- distractors – they may be white letters when plicates the frontal activation in report. the distractors are black letters – subjects’ This result gains support from a change detection of the second stimulus is severely detection study. Here, subjects had to P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

722 the cambridge handbook of consciousness

maintain a sample stimulus in short-term tion of backward masked targets when deliv- memory in order to decide whether the test ered 40 ms before target onset (Grosbras stimulus, which appeared 6 s later at the & Paus, 2003). As electrical microstimula- same central position, matched the sam- tion of the FEF in monkeys has been shown ple (Pessoa & Ungerleider, 2004). Impor- to enhance the visually evoked responses in tantly, the authors used a stimulus/response- area V4 (Moore & Armstrong, 2003), this contingent analysis of their imaging data, finding is in line with a frontal ‘prompting’ contrasting false alarms (reports of a change of visual areas as well as a gain effect (i.e., that had not occurred) with misses (miss- an enhancement of visual responsivity as has ing a change that had occurred), as well as been repeatedly demonstrated to result from hits (correctly reporting a change) with cor- attention). rect rejections (correctly denying a change). In difficult tasks, such as those used for In this fashion, they could show that, albeit studying processes pertaining to presently less strongly, false alarms activated the same invisible or inaccessible information in nor- frontoparietal network as hits. Missed tar- mal subjects, attention is certainly required. gets produced less (Marois et al., 2004)or In its different guises (focal, spatial, divided, none of this dorsal frontoparietal activa- voluntary, automatic), attention can mod- tion even if subjects indicated high confi- ulate the activity of sensory neurons dence in their responses (Pessoa & Ungerlei- (Desimone & Duncan, 1995) even in the der, 2004). Instead, focal activation increases absence of stimuli (Kastner et al., 1999), were seen in visual areas tuned to the stimuli affect the apparent contrast of stimuli that and in more ventrally situated frontal areas are presented (Carrasco, Ling, & Read, (inferior frontal gyrus: Beck et al., 2001; sup- 2004), and enhance spatial resolution and plementary eye field: Pessoa & Ungerleider, recognition. Depending on guise and con- 2004). text, attention invokes different subcortical That different brain regions are involved and cortical neural networks including dorsal in different aspects of processing, such as and ventral frontoparietal regions (Corbetta phenomenal rendering, report, and atten- & Shulman, 2002). It may thus account for tion, is thus more than a conceptual some aspects of the different stimulus- and possibility. Moreover, different neuronal response-contingent activation patterns in populations in the same area also seem to the change detection and attentional blink be differentially involved, as indicated by results. physiological recordings in the frontal eye Because the effects of attention perme- field (FEF) of monkeys performing a detec- ate the entire perceptual process from the tion task in which a proportion of targets phenomenal rendering through recognition was masked (Thompson & Schall, 2000). and working memory to report of stim- Reminiscent of Pessoa and Ungerleider’s uli, its neuronal correlates require care- results, the authors found FEF neurons that ful differentiation from those involved in responded when the monkeys reported a tar- consciousness. This is both important and get, regardless of whether the target had difficult. It is difficult because, as just been presented or not. However, visual FEF described, attention affects the structures neurons that do not project to oculomo- that are involved in perceptual processing. tor structures also showed a selective post- It is important because attention is distinct mask activity that was synchronized to target from consciousness: Like sensory informa- presentation. These neurons may send feed- tion processing, but unlike consciousness, back to extrastriate visual cortex, prompting attention comes in covert as well as explicit and questioning them as prefrontal cortex forms, being automatically deployed in the has been suggested to do in several ambigu- first, and intentionally focussed in the sec- ous and difficult visual tasks. Finally, tran- ond. That voluntary focal attention can be scranial magnetic stimulation over the FEF required for above-chance performance in in humans improved the subjects’ detec- tasks in which the stimuli are not consciously P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

hunting the ghost: toward a neuroscience of consciousness 723

perceived – masking in normally sighted sub- singleton, or presented simultaneously with jects is one example (Naccache, Blandin, distractors, or in a series of other stim- & Dehaene, 2002) – further underlines its uli – will also influence both the neuronal being different from the conscious represen- response it evokes and its detectability. The tation. subjects’ task bears on the level of processing that a stimulus achieves; for example, Conclusion detection of face stimuli among non-face distractors requires a level of processing differ- Whether information, which is processed ent from what is required for recognition to an extent that makes it behaviourally of individual faces. What the subject senses effective, becomes consciously represented or experiences depends on the magnitude is not determined by attention alone. In the and probably the kind of neuronal response; absence of sufficient visual information, as in the strength of the response, the size of blindsight or under stringent masking con- the activated population, and whether its ditions, attention cannot render the repre- members act in concert are all likely to sentation phenomenal. Both attention and play a role. What becomes consciously re- the conscious access to the information that presented may depend on what the subject is required for the subjects to be able to attends to, and attention invokes extravisual communicate that they have a conscious structures that in their turn affect visual representation of the information invoke neurons. What becomes consciously repre- frontoparietal cortices. Although their role sented may also depend on the subject’s is thus of great importance in establishing alertness; some events pass unnoticed when whether a representation is conscious, the one is daydreaming, whereas others, like process that renders it phenomenal requires slipping on a banana skin, reach conscious- the functionality of the systems that pro- ness if anything does. Whether anything vide the information. In the present vision- becomes consciously represented depends only examples, it is the destruction or den- on the subject being in a conscious state in ervation of the visual cortices, rather than the first place. Neuronal processes involved that of frontoparietal circuits, that abolishes in mediating the conscious state, whether (aspects of ) phenomenal vision or object via the reticular formation, or through non- recognition dependent on what parts are specific thalamo-cortico-thalamic loops, or affected. The jury that decides whether through some other type of process, pro- these regions are only necessary, as people vide the critical background for any con- who argue that it is only by virtue of access scious representation. It is therefore not sur- that conscious representations are formed prising that manipulations of non-specific contend (see below), or whether they are brain structures alter the stimulus-driven also sufficient is still out. activity of specific neurons (Jasper, 1949) as well as neuronal networks (Munk et al., 1996); Bachmann’s approach to the micro- Consciousness and Confounds genesis of conscious perception takes this non-specific thalamic-enabling effect into The neural correlates of conscious sensations account (Bachmann, 2000). and representations depend on many fac- If all of these factors, and the list is not tors. The stimulus itself is one; presentation exhaustive, determine what gets to be rep- of a brief beep causes a cascade of brain resented and also to what depth it does so, it events that is different from that caused by is a hard task to define the neural correlate a briefly presented red blob, and the neu- even of a single phenomenal sensation. To ronal response to a red blob depends on make matters worse, observers who want to stimulus size, position, contrast, and so on. study conscious processes require a report The experimental conditions under which from the subject that they can confidently the stimulus is presented – whether it is a use to assert that something was consciously P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

724 the cambridge handbook of consciousness

seen, heard, or felt. In view of the possibility This argument extends to the interpre- of implicitly guided motor behaviour, this tation of data from patients in pathological (verbal or non-verbal) report cannot be of states of consciousness. A marked decrease a kind that could also be elicited when no in frontoparietal network activity is con- conscious representation was formed. Giv- sistent with the patients’ inability to give ing an unequivocal report requires not only consistent reports about any experiences sufficient motor control on the part of the they may have, but does not preclude that subject but also conscious access to his or they have any. Functional disconnections her phenomenal percept. In Block’s defini- between sensory and frontoparietal areas tion (1995), ‘access-consciousness’ is avail- as inferred from neuroimaging of vegeta- ability for use in reasoning and rationally tive state patients agree with this cautious guiding speech and action. Conscious access interpretation. In fact, the functional neu- and reportability are the ones most likely to roanatomy of sleep in its various stages pro- involve additional neuronal processes, which vides experimental support for it, as both further encumbers the search for the neu- REM and non-REM sleep are characterized ronal correlates of trait consciousness, espe- by a decrease in frontoparietal activation, cially if neither access nor reportability is although extended periods of sleep brim necessary for having the experience in the with phenomenal experiences. Although first place. This point – whether or not access these experiences are not lacking in vivid- is necessary for the phenomenal representa- ness, they are lacking in immediate and gen- tion – is controversial (Block, 1995, 2005; eral reportability and other forms of access. Rosenthal, 1986; Stoerig & Barth, 2001). If access consciousness depends on phenom- Studies tackling the neural substrate of enal consciousness, there is no access if the conscious representations indicate that the phenomenal representation is lost; but if sensory systems are necessary for the phe- access is lost while phenomenality remains, nomenal rendering of stimuli. However, only the subject learns about it. their activation as such can be insufficient The multiplicity of conjoined factors is to generate conscious sensation, as demon- likely to contaminate the conclusions we strated by the finding that implicit and draw from the experimental data. To un- explicit perceptual processes are mediated confound them, we need a means to estab- by overlapping sensory regions, suggesting lish whether something is consciously repre- that a certain level, type, or pattern of activ- sented, a means that, like a brain potential or ity is necessary. To what extent non-sensory some other physiological marker, is indepen- cortical networks are required for phenom- dent of a behavioural response. As yet, we do enal perception is still unresolved (see dis- not know whether someone is conscious at cussion in Kohler¨ & Moscovitch, 1997). all unless he or she displays behaviour we My working hypothesis is that the sen- are able to identify as consciously initiated, sory neurons’ concerted activity is respon- and we do not know whether something sible for providing conscious phenomenal is consciously represented unless he or she content; reflecting on this content could informs us about it. Only people who equate well be enabled by non-sensory networks access and phenomenal consciousness need that also access the sensory ones during not mind what to others appears a barrier recall and imagery and prompt them con- large enough to warrant a Nobel prize to tinuously – challenging the consistency of reward its dissolution. their report and requesting different interpretations in the face of ambiguous input. The frontal cortex is indispensable for con- Acknowledgements scious report because voluntary behaviour is impossible without it, but any interfer- Thanks to the colleagues who kindly pro- ence with reportability is moot regarding the vided figures to illustrate their data – presence or absence of a phenomenal repre- Michael Alkire, David Leopold, Steven sentation. Laureys, and Nicos Logothetis – and to P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

hunting the ghost: toward a neuroscience of consciousness 725

Morris Moscovitch for his helpful comments Block, N. (1995). On a confusion about a function on the manuscript. of consciousness. Behavioral and Brain Sciences, 18, 227–287. Block, N. (2005). Two neural correlates of consciousness. Trends in Cognitive Sciences, 9, 46– References 52. Bodis-Wollner, I., Atkin, A., Raab, E., & Wolk- 1968 Allison. A. C., & Nunn, J. F. ( ). Effects of stein, M. (1977). Visual association cortex and general anesthetics on microtubules. A possi- vision in man: Pattern-evoked occipital poten- 1326 ble mechanism of anesthesia. Lancet, 2 , – tials in a blind boy. Science, 198, 629–631. 1329. Bogen, J. E. (1995). On the neurophysiology of Alkire, M. T., Haier, R. J., Barker, S. J., Shah, consciousness: I. An overview. Consciousness 1995 N. K., Wu, J. C., & Kao, Y. J. ( ) Cere- and Cognition, 4, 52–62. bral metabolism during propofol anesthesia in Boly, M., Faymonville, M.-E., Peigneux, P. et al. humans studied with positron emission tomog- (2004). Auditory processing in severely brain raphy. Anesthesiology, 82, 393–400. injured patients: differences between the min- 2000 Alkire, M. T., Haier, R. J., & Fallon, J. H. ( ). imally conscious state and the persistent vege- Toward a unified theory of narcosis: Brain imag- tative state. Archives of Neurology, 61, 233–238. ing evidence for a thalamocortical switch as the Bonhomme, V., Fiset, P., Meuret, P., Backman, neurophysiologic basis of anesthetic-induced S., Plourde, G., Paus, T., Bushnell, M. C., & unconsciousness. Consciousness and Cognition, Evans, A. C. (2001). Propofol anesthesia and 9, 387–395. cerebral blood flow changes elicited by vibro- 1991 Angel, A. ( ). The G. L. Brown lecture. tactile stimulation: A positron emission tomog- Adventures in anaesthesia. Experimental Physi- raphy study. Journal of Neurophysiology, , 1 38 85 ology, 76, – . 1299–1308. Angel, A. (1993). Central neuronal pathways and Bonnebakker, A. E., Bonke, B., Klein, J., Wolters, the process of anaesthesia. British Journal of G., Stijnen, T., Passchier, J., & Merikle, P. M. Anaesthesia, 71, 148–163. (1996). Information processing during general Anis, N. A., Berry, S. C., Burton, N. R., & anesthesia: Evidence for unconscious memory. 1983 Lodge, D. ( ). The dissociative anaesthetics, Memory and Cognition, 24, 766–776. ketamine and phencyclidine, selectively reduce Born, A. P., Law, I., Lund, T. E., Rostrup, E., excitation of central mammalian neurons by N- Hanson, L. G., Wildschiødtz, G., Lou, H. C., methyl-asparatate. British Journal of Pharma- & Paulson, O. B. (2002). Cortical deactivation cology, 79, 565–575. induced by visual stimulation in human slow- Antognini, J. F., Buonocore, M. H., Disbrow, wave sleep. Neuroimage, 17, 1325–1335. E. A., & Carstens, E. (1997). Isoflurane anes- Bosinelli, M. (1995). Mind and consciousness dur- thesia blunts cerebral responses to noxious and ing sleep. Behavioral Brain Research, 69, 195– innocuous stimuli: A fMRI study. Life Sciences, 201. 61, 349–354. Braun, A. R., Balkin, T. J., Wesensten, N. J., Car- Aserinsky, E., & Kleitman, N. (1953). Regularly son, R. E., Varga, M., Baldwin, P., Selbie, S., occurring periods of eye motility, and concomi- Belenky, G., & Herscovitch, P. (1997). Regional tant phenomena during sleep. Science, 118, 273– cerebral blood flow throughout the sleep-wake 274. cycle. Brain, 120, 1173–1197. Bachmann, T.(2000). Microgenetic approach to the Braun, A. R., Balkin, T. J., Wesensten, N. J., conscious mind. Amsterdam: John Benjamins. Gwadry, F., Carson, R. E., Varga, M., Bald- Bassetti, C., Vella, S., Donati, F., Wielapp, P., & win, P., Belenky, G., & Herscovitch, P. (1998). 2000 Weder, B. ( ). SPECT during sleepwalking. Dissociated pattern of activity in visual cortices 484 485 Lancet, 356, – . and their projections during human rapid eye Beck, D. M., Rees, G., Frith, C. D., & Lavie, N. movement sleep. Science, 279, 91–95. 2001 ( ). Neural correlates of change detection Broadbent, D. E. (1974). Division of function and change blindness. Nature Neuroscience, 4, and integration of behavior. In F. O. Schmitt 645 650 – . & F. G. Worden (Eds.), The neurosciences third Blake, R., & Logothetis, N. K. (2002). Visual com- study program (pp. 31–42). Cambridge, MA: petition. Nature Reviews Neuroscience, 3, 13–21. MIT Press. P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

726 the cambridge handbook of consciousness

Broughton, R. et al. (1994). Homicidal Electroencephalography and Clinical Neurophys- somnambulism: A case report. Sleep, 17, iology, 9, 673–690. 253–264. Desimone, R., & Duncan, J. (1995). Neural mech- Brown, R. J., & Norcia, A. M. (1997). A method anisms of selective attention. Annual Review of for investigating binocular rivalry in real-time Neuroscience, 18, 193–222. with the steady-state VEP. Vision Research, 37, Fiset, P., Paus, T., Daloze, T., Plourde, G., Meuret, 2401–2408. P., Bonhomme, V., Haji-Ali, N., Backman, Bruyer, R., Laterre, C., Seron, X., Feyereisen, S. B., & Evans, A. C. (1999). Brain mecha- P., Srypstein, E., Pierrard, E., & Rectern, D. nisms of propofol-induced loss of conscious- (1983). A case of prosopagnosia with some pre- ness in humans: A positron emission tomogra- served covert remembrance of familiar faces. phy study. Journal of Neuroscience, 19, 5506– Brain and Cognition, 2 , 257–284. 5513. Cariani, P. (2000). Anesthesia, neural informa- Flohr, H. (1995). An information processing the- tion processing, and conscious awareness. Con- ory of anaesthesia. Neuropsychologia, 33, 1169– sciousness and Cognition, 9, 387–395. 1180. Carrasco, M., Ling, S., & Read, S. (2004). Atten- Flohr, H., Glade, U., & Motzko, D., (1998). The tion alters appearance. Nature Neuroscience, 7, role of the NMDA synapse in general anesthe- 208–209. sia. Toxicology Letters, 100–101, 23–29. Cavallero, C., Cicogna, P., Natale, V., Occhio- Foulkes, D., & Schmidt, M. (1983). Tempo- nero, M., & Zito, A (1992). Slow wave sleep ral sequence and unit composition in dream dreaming. Sleep, 15, 562–566. reports from different stages of sleep. Sleep, 6, Chun, M. M., & Potter, M. C. (1995). A two-stage 265–80. model for multiple target detection in rapid Foulkes, W. D. (1962). Dream reports from dif- serial visual presentation. Journal of Experme- ferent stages of sleep. Journal of Abnormal and ntal Psychology: Human Perception and Perfor- Social Psychology, 65, 14–25. mance, 21, 109–127. Franks, N. P., & Lieb, W. R. (1994). Molecular Cicogna, P., Natale, V., Occhionero, M., & and cellular mechanisms of general anesthesia. Bosinelli, M. (2000). Slow wave and REM sleep Nature, 367, 607–614. mentation. Sleep Research Online, 3, 67–72. Franks, N. P., & Lieb, W.R. (1996). An anesthetic- Claparede,` E. (1911). Recognition et moite.´ sensitive superfamily of neurotransmitter- Archives de Psychologie, 11, 79–90. gated ion channels. Journal of Clinical Anesthe- Cohen, N. J., & Corkin, S. (1981). The amnesic sia, 8,S3–S7. patient H. M.: Learning and retention of a cog- Giacino, J. T., Ashwal, S., Childs, N., et al. nitive skill. Neuroscience Abstracts, 7, 235. (2002). The minimally conscious state. Deﬁni- Corbetta, M., Shulman G. L. (2002). Control of tion and diagnostic criteria. Neurology, 58, 349– goal-directed and stimulus-driven attention in 353. the brain. Nature Reviews Neuroscience, 3, 201– Goebel, R., Muckli, L., Zanella, F. E., Singer, 215. W., & Stoerig, P. (2001). Sustained extrastri- Cowey, A., & Stoerig, P. (1995). Blindsight in ate cortical activation without visual aware- monkeys. Nature, 373, 247–249. ness revealed by fMRI studies of hemianopic 1459 1474 Crick, F., & Koch, C. (1995). Are we aware of patients. Vision Research, 41, – . neural activity in primary visual cortex? Nature, Grosbras, H., & Paus, T.(2003). Transcranial mag- 375, 121–123. netic stimulation of the human frontal eye ﬁeld Czisch, M., Wetter, T. C., Kaufmann, C., facilitates visual awareness. European Journal of 3121 3126 Pollmacher,¨ T., Holsboer, F., & Auer, D. P. Neuroscience, 18, – . (2002). Altered processing of acoustic stim- Hameroff, S. (1998). Anesthesia, consciousness uli during sleep: Reduced auditory activation and hydrophobic pockets – a unitary quantum and visual deactivation detected by a com- hypothesis of anesthetic action. Toxicology Let- bined fMRI/EEG study. Neuroimage, 16, 251– ters, 100/101, 31–39. 258. Hartmann, E. (1983). Two case reports: Night ter- Dement, W., & Kleitman, N. (1957). Cyclic vari- rors with sleepwalking – a potentially lethal dis- ations in EEG during sleep and their relation to order. Journal of Nervous and Mental Disease, eye movements, body motility, and dreaming. 171, 503–505. P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

hunting the ghost: toward a neuroscience of consciousness 727

Heywood, C. A., Cowey, A., & Newcombe, F. Kajimura, N., Uchiyama, M., Takayama, Y., (1991). Chromatic discrimination in a cortically Uchida, S., Uema, T., Kato, M., et al. (1999). colour blind observer. European Journal of Neu- Activity of midbrain reticular formation and roscience, 3, 802–812. neocortex during the progression of human Hobson, J. A. (1989). Sleep. New York: Scientific non-rapid eye movement sleep. Journal of Neu- American Library. roscience, 19, 10065–10073. Hobson, J. A., Goldfrank, F., & Snyder, F. Kales, A., Hoedemaker, F. S., Jacobson, A., Kales, (1965). Respiration and mental activity during J. D., Paulson, M. J., & Wilson, T. E. (1967). sleep. Journal of Psychiatric Research, 3, 79– Mentation during sleep: REM and NREM 90. recall reports. Perceptual and Motor Skills, 24, 555–60. Hochstein, S., & Ahissar, M. (2002). View from the top: Hierarchies and reverse hierarchies in Kassubek, J., Juengling, F. D., Els, T., et al. 2003 the visual system. Neuron, 36, 791–804. ( ). Activation of residual cortical net- 1991 work during painful stimulation in long-term Horton, J. C., & Hoyt, W. F. ( ). Quadran- 15 postanoxic vegetative state: A 2 tic visual field defects. A hallmark of lesions in O-H O PET Journal of Neurological.Science, 85 extrastriate (V2/V3) cortex. Brain, , 1703– study. 212, – 114 91 1718. . Hubel, D. H., & Wiesel, T. N. (1968). Receptive Kastner, S., Pinsk, M. A., De Weerd, P., Des- 1999 fields and functional architecture of monkey imone, R., & Ungerleider, L. G. ( ). In- striate cortex. Journal of Physiology (London), creased activity in human visual cortex during directed attention in the absence of visual stim- 195, 215–243. ulation. Neuron, 22, 751–761. Jasper, H. H. (1949). Diffuse projection systems: 1963 The integrative action of the thalamic reticu- Kleitman, N. ( ). Sleep and wakefulness. lar system. Electroencephalography and Clinical Chicago: University of Chicago Press. Neurophysiology, 1, 405–420. Koestler, A. (1968). Das Gespenst in der Maschine. Jasper, H. H. (1998). Sensory information and Wien, Zurich,¨ Munchen:¨ Molden: conscious experience. In H. H. Jasper, L. Kohler,¨ S., & Moscovitch, M. (1997) Uncon- Descarries, V. F. Castellucci, & S. Rossignol scious visual processing in neuropsychological (Eds), Consciousness at the frontiers of neuro- syndromes: A survey of the literature and eval- science. Advances in neurology (Vol. 77,pp. uation of models of consciousness. In M. D. 33–48). Philadelphia, New York: Lippincott- Rugg (Ed.), Cognitive neuroscience (pp. 305– Raven Publishers. 373). Cambridge, MA: MIT Press. Jasper, H. H., & Droogleever-Fortuyn, J. Kreiman, G., Fried, I., & Koch, C. (2002) (1947). Experimental studies of the functional Single-neuron correlates of subjective vision in anatomy of petit mal epilepsy. Proceedings of the human temporal lobe. Proceedings of the the association for research in nervous & mental National Academy of Science USA, 99, 8378– disease, 26, 272–298. 8383. Jennett, B. (2002). The vegetative state. Journal Lamme, V. L. (2001). Blindsight: The role of of Neurology, Neurosurgery, and Psychiatry, 73, feedforward and feedback corticocortical con- 355–356. nections. Acta Psychologica (Amsterdam), 107, Jennett, B., & Plum, F. (1972, April 1). Persis- 209–228. tent vegetative state after brain damage. Lancet, Laureys, S., Faymonville, M.-E., Luxen, A., 734–737. Lamy, M., Franck, G., & Maquet, P. (2000a). Jouvet, M., & Delorme, F. (1965). Locus Restoration of thalamocortical connectivity coeruleus et sommeil paradoxical. Societéde after recovery from persistent vegetative state. 1790 1791 Biologie, 159, 895–899. Lancet, 355, – . Jones, E. G. (1998). A new view of specific and Laureys, S., Faymonville M.-E., Degueldre, C., nonspecific thalamocortical connections. In et al. (2000b). Auditory processing in the H. H. Jasper, L. Descarries, V. F. Castellucci, vegetative state. Brain, 123, 1589–1601. & S. Rossignol (Eds), Consciousness at the fron- Laureys, S., Faymonville, M.-E., Peigneux, P., tiers of neuroscience. Advances in neurology (Vol. et al. (2002a). Cortical processing of noxious 77,pp.49–71). Philadelphia: Lippincott-Raven somatosensory stimuli in the persistent vegeta- Publishers. tive state. Neuroimage, 17, 732–741. P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

728 the cambridge handbook of consciousness

Laureys, S., Antoine, S., Boly, M., et al. (2002b). Marois, R., Yi, D.-J., & Chun, M. M. (2004). Brain function in the vegetative state. Acta The neural fate of consciously perceived and Neurologica (Belgium), 102, 177–185. missed events in the attentional blink. Neuron, Laureys, S., Owenm A. M., & Schiff, N. D. 41, 465–472. (2004). Brain function in coma, vegetative Martin, E., Thiel, T., Joeri, P., Loenneker, T., Eka- state, and related disorders. Lancet Neurology, todramis, E., Huisman, T., Hennig, J., & Marcar, 3, 537–546. V. L. (2000). Effect of pentobarbital on visual Leopold, D. A., & Logothetis, N. K. (1999). Multi- processing in man. Human Brain Mapping, 10, stable phenomena: Changing views in percep- 132–139. tion. Trends in Cognitive Sciences, 3, 254–264. Mashour, G. A. (2004). Consciousness unbound. Leopold, D. A., Plettenberg, H. K., & Logo- Toward a paradigm of general anesthesia. Anes- thetis, N. K. (2003). Visual processing in the thesiology, 100, 428–433. ketamine-anesthetized monkey. Optokinetic Meadows, J. C. (1974). Disturbed perception and blood oxygenation-dependent responses. of colours associated with localized cerebral Experimental Brain Research, 143, 359–372. lesions. Brain, 97, 615–632. Levy, D. E., Sidtis, J. J., Rottenberg, D. A., Jar- Merikle, P. M., & Daneman, M. (1996). Memory den, J. O., et al. (1987). Differences in cerebral for unconsciously perceived events; evidence blood flow and glucose utilization in vegetative from anesthetized patients. Consciousness and versus locked-in patients. Annals of Neurology, Cognition, 5, 542–561. 22, 673–682. Milner, A. D., & Goodale, M. A. (1995). The Logothetis, N. K. (1998). Single units and con- visual brain in action. New York: Oxford Uni- scious vision. Philosophical Transactions of the versity Press. Royal Society, London B, 353, 1801–1818. Milner, B., Corkin, S., & Teuber, H.-L. (1968). Logothetis, N. K., Guggenberger, H., Peled, S., & Further analysis of the hippocampal amnesic Pauls, J. (1999). Functional imaging of the mon- syndrome: 14-year follow-up study of H. M. key brain. Nature Neuroscience, 2 , 555–562. Neuropsychologia, 6, 215–234. Logothetis, N. K., & Schall, J. D. (1989). Neuronal Moore, T., & Armstrong, K. M. (2003). Selective correlates of subjective visual perception. Sci- gating of visual signals by microstimulation of ence, 245, 761–763. frontal cortex. Nature, 421, 370–373. Lumer, E. D., Friston, K. J., & Rees, G. (1998) Morruzzi, G., & Magoun, W. (1949). Brain stem Neural correlates of perceptual rivalry in the reticular formation and activation of the EEG. human brain. Science, 280, 1930–1934. Electroencephalography and Clinical Neurophys- Madler, C., Keller, I., Schwender, D., & Poppel,¨ iology, 1, 455–473. E. (1991). Sensory information processing dur- Mozaz Garde, M. , & Cowey, A. (2000). ‘Deaf ing general anaesthesia: Effect of isoflurane on hearing’: Unacknowledged detection of audi- auditory evoked neuronal oscillations. British tory stimuli in a patient with cerebral deafness. Journal of Anaesthesiology, 66, 81–87. Cortex, 36, 71–80. Madler, C., & Poppel,¨ E. (1987). Auditory evoked Munk, M. H. J., Roelfsema, P. R., Konig,¨ P., potentials indicate the loss of neuronal oscil- Engel, A. K., & Singer, W. (1996). Role of lations during general anaesthesia. Naturwis- reticular activation in the modulation of intra- senschaften, 74, 42–43. cortical synchronization. Science, 272, 271– Mahowald, M. W., & Schenck, C. H. (2001). 274. Evolving concepts of human state dissociation. Nacchache, L., Blandin, E., & Dehaene, S. (2002). Archives Italienne de Biologie, 139, 269–300. Unconscious masked priming depends on tem- Maquet, P. (2000). Functional neuroimaging of poral attention. Psychological Science, 13, 416– normal human sleep by positron emission 424. tomography. Journal of Sleep Research, 9, 207– Nielsen, T. A. (2000) A review of mentation in 231. REM and NREM sleep: “covert” REM sleep Maquet, P., Peters,´ J.-M., Aerts, J., Delfiore, G., as a possible reconciliation of tow opposing Degueldre, G., Luxen, A., & Franck, G. (1996). models. Behavioral and Brain Sciences, 23, 793– Functional neuroanatomy of human rapid-eye- 1121. movement sleep and dreaming. Nature, 383, Nofzinger, E. A., Mintun, N. A., Wiseman, M., 163–166. Kupfer, D. J., & Moore, R. Y. (1997). Forebrain P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

hunting the ghost: toward a neuroscience of consciousness 729

activation in REM-sleep: An FDG PET study. reference to appreciation of movement. Brain, Brain Research, 770, 192–201. 40, 15–57. Paillard, J., Michel, F., & Stemach, G. (1983). Rosenbaum, R. S., Kohler,¨ S., Schacter, D. L., Localization without content: A tactile ana- Moscovitch, M., Westmacott, R., Black, S. E., logue of blindsight. Archives of Neurology, 40, Gao, F., & Tulving, E. (2004). The case of K. 548–551. C.: Contributions of a memory-impaired per- Pasik, P., & Pasik, T. (1982). Visual function in son to memory theory. Neuropsychologia, 43, monkeys after total removal of visual cerebral 989–1021. cortex. Contributions to Sensory Physiology, 7, Rosenthal, D. (1986). Two concepts of conscious- 147–200. ness. Philosophical Studies, 94, 329–359. Pessoa, L., & Ungerleider, L. G. (2004). Neu- Rovee-Collier, C., Hayne, H., & Colombo, M. ral correlates of change detection and change (2001). The development of implicit and explicit blindness in a working memory task. Cerebral memory. Advances in consciousness research 24. Cortex, 14, 511–520. Amsterdam: John Benjamins. Pins, D., & ffytche, D. (2003). The neural corre- Sahraie, A., Weiskrantz, L., Barbur, J. L., Sim- lates of conscious vision. Cerebral Cortex, 13, mons, A., Williams, S., & Brammer, M. J. 461–474. (1997). Pattern of neuronal activity associ- Plourde, G., Baribeau, J., & Bonhomme, V. ated with conscious and unconscious pro- (1997). Ketamine increases the amplitude of cessing of visual signals. Proceedings of the the 40-Hz auditory steady-state response in National Academy of Sciences USA, 94, 9406– humans. British Journal of Anaesthesia, 78, 9411. 524–529. Schacter, D. L., Dobbins, I. G., & Schnyer, D. M. Plum, F., & Posner, J. B. (1982). The diagno- (2004). Specificity of priming: A cognitive neu- sis of stupor and coma. Philadephia: F. A. roscience perspective. Nature Reviews Neuro- Davis. science, 5, 853–862. Plum, F., Schiff, N., Ribary, U., & Ll´ınas, R. Schenck, C. H., Bundlie, S. R., Ettinger, M. G., & (1998). Coordinated expression in chronically Mahowald, M. W. (1986). Chronic behavioral unconscious persons. Philosophical Transactions disorders of human REM sleep: A new category of the Royal Society, London B: Biological Science, of parasomnia. Sleep, 9, 293–308. 1929 1933 353, – . Schenck, C. H., & Mahowald, M. W. (1994). Poppel,¨ E. (1997). A hierarchical model of time Review of nocturnal sleep-related eating disor- perception. Trends in Cognitive Sciences, 1, 56– ders. International Journal of Eating Disorders, 61. 15, 343–356. Purpura., K. P., & Schiff, N. D. (1997). The thala- Schenck, C. H., Milner, D. M., Hurwitz, T. mic intralaminar nuclei: A role in visual aware- D., Bundlie, S. R., & Mahowald, M. W. ness. The Neuroscientist, 3, 8–15. (1989). Dissociative disorders presenting as Rechtschaffen, A., & Kales, A. A. (1968). A man- somnambulism: Polysomnographic, video and ual of standardized terminology, techniques and clinical documentation (8 cases). Dissociation, scoring system for sleep stages of human subjects. 2 , 194–204. Bethesda, MD: National Institute of Neurolog- Schiff, N. D., Ribary, U., Moreno, D. R., Beat- ical Diseases and Blindness. tie, B., et al. (2002). Residual cerebral activity Rees, G., Wojciulik, E., Clarke, K., Husain, M., and behavioural fragments can remain in the Frith, C., & Driver, J. (2000). Unconscious acti- persistently vegetative brain. Brain, 125, 1210– vation of visual cortex in the damaged right 1234. hemisphere of a parietal patient with extinc- Shulman, G., Fiez, J., Corbetta, M., Buckner, 1624 1633 tion. Brain, 123, – . R. L., Miezin, F. M., Raichle, M. E., & Petersen, Rensink, R. A. (2000). Seeing, sensing, scrutiniz- S. E. (1997). Common blood flow changes ing. Vision Research, 40, 1469–1487. across visual tasks: II. Decreases in cerebral cor- Ricci, C., & Blundo, C. (1990). Perception of tex. Journal of Cognitive Neuroscience, 9, 648– ambiguous figures after focal brain lesions. 663. Neuropsychologia, 28, 1163–1173. Simons, D. J., & Levin, D. T. (1997). Change Riddoch, G. (1917). Dissociations of visual per- blindness. Trends in Cognitive Sciences, 1, 261– ception due to occipital injuries, with especial 267. P1: KAE 0521857430c25 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:36

730 the cambridge handbook of consciousness

Steriade, M. (2000). Corticothalamic resonance, relates of conscious perception by frequency- states of vigilance and mentation. Neuroscience, tagged neuromagnetic responses. Proceedings of 101, 243–276. the National Academy of Science USA, 95, 3198– Stoerig, P. (1996). Varieties of vision: From blind 3203. processing to conscious recognition. Trends in Wagner, U., Gais, S., Haider, H., Verleger, R., & Neurosciences, 19, 401–406. Born, J. (2004). Sleep inspires insight. Nature, 352 355 Stoerig, P. (1999). Blindsight. In R. Wilson & F. 427, – . Keil (Eds.), The MIT encyclopedia of the cogni- Warrington, E. K., & Weiskrantz, L. (1968). New tive sciences (pp. 88–90). Cambridge, MA: MIT method of testing long-term retention with Press. special reference to amnesic patients. Nature, 972 972 Stoerig, P. (2002). Neural correlates of conscious- 217, – . ness as state and trait. In L. Nadel (Ed.), Ency- Weiskrantz, L. (1986). Blindsight: A case study clopedia of cognitive neuroscience (pp. 233–240). and implications. Oxford: Oxford University London: Macmillan Press. Press. Stoerig, P., & Barth, E. (2001). Low-level phe- Weiskrantz, L., Warrington, E. K., Sanders, nomenal vision despite unilateral destruction M. D., & Marshall, J. (1974). Visual capacity of primary visual cortex. Consciousness and in the hemianopic ﬁeld following a restricted Cognition, 10, 574–587. cortical ablation. Brain, 97, 709–728. Stoerig, P., & Cowey, A. (1993). Blindsight and White, N. S., & Alkire, M. T. (2003). Impaired perceptual consciousness: neuropsychological thalamocortical connectivity in humans dur- aspects of striate cortical function. In B. Gulyas, ing general-anesthetic-induced unconscious- D. Ottoson, & P. Roland (Eds.), Functional ness. Neuroimage, 19, 402–411. organisation of the human visual cortex (pp. 181– Yamada, T., Kameyama, S., Fuchigami, Y., 193). Oxford, New York: Pergamon Press. Nakazumi, Y., Dickins, Q. S., & Kimura, S. Stoerig, P., & Cowey, A. (1997). Blindsight in man (1988). Changes of short latency somatosen- and monkey. Brain, 120, 535–559. sory potential in sleep. Electroencephalography 126 136 Stoerig, P., Zontanou, A., & Cowey, A. (2002). and Clinical Neurophysiology, 70, – . Aware or unaware? Assessment of cortical Zeki, S. (1990). A century of cerebral achro- blindness in four men and a monkey. Cerebral matopsia. Brain, 113, 1721–1777. Cortex, 12, 565–574. Zeki, S. (1991). Cerebral akinetopsia. Brain, 114, Thompson, K. G., & Schall, J. D. (2000). 811–824. Antecedents and correlates of visual detection Zeki, S. (2001). Localization and globilization and awareness in macaque prefrontal cortex. in conscious vision. Annual Review of Neuro- Vision Research, 40, 1523–1538. science, 24, 57–86. Tononi, G., Srinivasan, R., Russell, D. P., & Edel- Zeman, A. (1997). Persistent vegetative state. man, G. M. (1998). Investigating neural cor- Lancet, 350, 795–798. P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

CHAPTER 26 Neurodynamical Approaches to Consciousness

Diego Cosmelli, Jean-Philippe Lachaux, and Evan Thompson

Abstract cal basis of consciousness. We continue by discussing the relation between spatiotem- One of the outstanding problems in the cog- poral patterns of brain activity and con- nitive sciences is to understand how ongo- sciousness, with particular attention to pro- ing conscious experience is related to the cesses in the gamma frequency band. We workings of the brain and nervous system. then adopt a critical perspective and high- Neurodynamics offers a powerful approach light a number of issues demanding further to this problem because it provides a coher- treatment. Finally, we close the chapter by ent framework for investigating change, vari- discussing how phenomenological data can ability, complex spatiotemporal patterns of relate to and ultimately constrain neurody- activity, and multiscale processes (among namical descriptions, with the long-term aim others). In this chapter, we advocate a neu- being to go beyond a purely correlational rodynamical approach to consciousness that strategy of research. integrates mathematical tools of analysis and modeling, sophisticated physiological data recordings, and detailed phenomenological The Intuition descriptions. We begin by stating the basic intuition: Consciousness is an intrinsically The central idea of this chapter is the notion dynamic phenomenon and must therefore of dynamics – the dynamics of neural activ- be studied within a framework that is capa- ity, the dynamics of conscious experience, ble of rendering its dynamics intelligible. We and the relation between them. then discuss some of the formal, analytical Dynamics is a multifarious concept. In a features of dynamical systems theory, with narrow sense, it refers to the change a cir- particular reference to neurodynamics. We cumscribed system undergoes in some time- then review several neuroscientiﬁc propos- dependent descriptive variable; for example, als that make use of dynamical systems the- a neuron’s membrane voltage (Abarbanel ory in characterizing the neurophysiologi- & Rabinovich, 2001). In a broad sense, 731 P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

732 the cambridge handbook of consciousness

dynamics indicates a field of research con- inefficient (think of what happens when you cerned with non-linear dynamical systems. try to catch a fish out of water; Garfinkel, Such systems range from mathematical Spano, Ditto, & Weiss, 1992; Schiff et al., models to experimental problems to actual 1994). In general terms, there will be a concrete world systems (Van Gelder, 1999). certain degree of dissociation between the Finally, in a more intimate sense, dynamics observed behavior of these systems and the refers to the temporal nature of our observa- patterns of external constraints that can be tions themselves and thus to our conscious imposed on them. In other words, these experience and how it is deployed in time systems exhibit a certain degree of auton- (Varela, 1999). The interplay of these dif- omy: When external perturbations cease, ferent senses of dynamics is at the heart of the system goes on; when external pertur- this chapter. bations become stationary, the system does What exactly are the properties of not. Somehow we intuitively recognize such dynamical systems, and why are they of systems as animated or alive, in contrast to interest in relation to consciousness? The simpler systems that respond in a linear and entirety of this chapter is concerned with predictable way to external control (e.g., this question, but to begin addressing it, we a stone that flies twice the distance when wish, in this introductory section, first to give thrown twice as strong). an overview of the basic intuition underlying Thus, such systems exhibit an intrinsic dynamical approaches to consciousness. variability that cannot be attributed to noise, Briefly stated, a complex, non-linear but appears to be constitutive of their func- dynamical system can be described, at any tioning. Moreover, in the case of certain time, by a position in a high-dimensional types of complex dynamical systems, one state space (Nicolis & Prigogine, 1989). The can reveal a characteristic spatiotemporal n coordinates of such a position are the val- balance of functional segregation and coop- ues of the set of n variables that define the erative integration.1 This balance depends system. This position changes in time and on the actual architecture of the system thus defines a trajectory, which will tend to (its internal connectivity, for example), and explore a subregion of the total state space. is revealed in the transient establishment One can then measure the distance between of distributed couplings among separated any two points of the trajectory and show subsystems that in themselves present local that under certain circumstances the tra- encapsulated dynamics. Finally, some of jectory can exhibit spontaneous recurrence: these systems display what has been termed Small portions of the state space will be ‘self-organization’; that is, the emergence of explored over and over, but never along the collective coherent behavior starting from exact same path. When perturbed by exter- random initial conditions. This last feature, nal events, such a system will change its tra- although not necessary for a system to jectory in a way that is never quite the same be considered dynamical, has proven espe- and that depends on its position in the state cially interesting when dealing with bio- space at the time of the perturbation. logical phenomena (Haken, 1983; Kelso, Given this feature, plus the system’s 1995). extreme sensitivity to initial conditions, the The brain is a major case in point. system’s response to perturbations will be The nervous system is a complex dynam- unpredictable in practice. It is therefore ical structure, in which individual neurons quite difficult to control such a system and have intrinsic activity patterns and cooper- constrain its movement along a predefined ate to produce coherent collective behavior trajectory. For example, in the case of chaotic (Llinas, 1988). The explosion of neuroimag- systems, such control involves applying a ing studies in the last 15–20 years, as well continuous succession of carefully chosen, as the substantial amount of data produced delicate inputs, brute force usually being by electrophysiological techniques since the P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

neurodynamical approaches to consciousness 733

beginning of the 20th century, has shown This intuitive convergence of complex that the brain is never silent, but always dynamical patterns in experience and in in a state of ongoing functioning (Wicker, brain activity is highly suggestive. It sug- Ruby, Royet, & Fonlupt, 2003). The ner- gests that the framework of dynamical sys- vous system has a domain of viability, of tems theory could offer a valuable way of allowed functioning, but within this domain bridging the two domains of brain activ- it explores a multiplicity of possible states ity and subjective experience. If we wish to in a recurrent, yet always changing manner study the neurobiological processes related (Palus, 1996). Incoming events are not suffi- to consciousness, then we must provide a cient to determine the system’s behavior, for description of these processes that is (some- any incoming event will change the system’s how) compatible with the dynamics of lived activity only as a result of how the system, experience. On the other hand, dynami- given its current activity, responds to that cal aspects of experience might serve as event (Engel, Fries, & Singer, 2001). a leading clue for uncovering and track- If we now follow the thread of dynam- ing the neurobiological processes crucial for ics back to our own conscious experi- consciousness. ence, we can immediately notice that our In the rest of this chapter, we explore this consciousness manifests subjectively as a guiding intuition through a discussion of the kind of continuously changing or flowing following topics: formal dynamical systems, process of awareness, famously called the neurobiological theories based on dynamical ‘stream of consciousness’ by William James system principles, and the attempt to dis- (1890/1981). Our experience is made up of tinguish dynamical structures within expe- recurring perceptions, thoughts, images, and rience that can constrain how we study the bodily sensations; yet, however similar these neurobiological basis of consciousness. events may be over time, there is always something new to each one, something ultimately unpredictable to every forthcoming Neurodynamics moment. We can try to plan our day as strictly as we want, but the wanderings of Dynamical Systems our minds and how we react to the encounters we have in the actual world are things we Dynamical cognitive science has been cannot fully control. There seems to be an defined as “a confederation of research endogenous, spontaneous, ongoing flow to efforts bound together by the idea that nat- experience that is quite refractory to exter- ural cognition is a dynamical phenomenon nal constraints (Hanna & Thompson, 2003). and best understood in dynamical terms” Indeed, this dissociation can easily be made (Van Gelder, 1999,p.243). Within this con- evident from the first-person perspective. federation, the job of the neurodynamicist is If you sit down and close the windows, to model the neural basis of cognition using turn off the lights, and close your eyes so the tools of dynamical systems theory. Thus, that external stimulation is greatly reduced the first thing we need to do is to define more for you, there is nevertheless still some- precisely the notion of a dynamical system. thing going on subjectively in you, with an A dynamical system is a collection of apparent temporal dynamics all of its own.2 interdependent variables that change in Furthermore, at any moment, conscious- time. The state of the system at any time ness appears diverse, complex, and rich with t is defined by the values of all the vari- multiple, synchronous, and local contents ables at that time; it can be represented by a (images, expectations, sounds, smells, kines- position in an abstract ‘state space’, whose thetic feelings, etc.), yet it seems to hold coordinates are the values of all the vari- together as a coherent and globally organized ables at t. The system’s behavior consists of experience. transitions between states and is described P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

734 the cambridge handbook of consciousness

geometrically by a trajectory in the state They can be considered as constant during space, which corresponds to the consecutive the given window of observation, but are positions the system occupies as time passes. potentially variable across different observa- At a first level of complexity, in the tion periods. Their dynamics thus contrasts context of neurodynamics, we can think with that of the external inputs, which can of the variables as being the membrane have a dynamics as fast as that of the sys- potentials of each individual neuron of the tem. The governing equations of the system nervous system.3 These membrane poten- are also a function of these inputs, but the tials are obviously interdependent. Thus, at temporal evolution of the external inputs this level, the state of the nervous system at cannot be predicted from those equations any time t would be defined by the value of (otherwise they could be considered as state all the membrane potentials at time t. variables of the system). Finally, all real sys- Although a dynamical system, in the most tems include a noise component, which also general terms, is any system that changes in counts as a factor in the governing equations time, dynamical systems theory gives spe- and thus affects the trajectory of the system. cial attention to non-linear dynamical systems. The behavior of such systems is gov- Neurodynamics and the Dynamical erned by non-linear equations; in other Approach in Neuroscience words, some of the mathematical functions used to derive the system’s present Neurodynamics emerged from the proposal, state from its previous states and possi- which can be traced back to Ashby in the ble external inputs are non-linear functions 1950s (Ashby, 1952), that the nervous sys- (for neurobiological examples, see Abar- tem can be described as a non-linear dynam- banel & Rabinovich, 2001; Faure & Korn, ical system. Although simple in appearance, 2001). Non-linearity can endow the sys- this proposal deserves some attention: What tem with certain interesting properties. For does the nervous system look like from a example, when convection cells in a horizon- dynamical point of view, and why is it non- tal water layer are submitted to a thermal linear? The majority of the dynamical mod- gradient above a critical value, the motions els of the nervous system describe the tem- of billions of molecules spontaneously orga- poral evolution of the membrane potentials nize into long-range correlated macroscopic of neurons (Arbib, 2002). The behavior of structures (Chandrasekhar 1961, cited in Le any neuron of the system is a function of Van Quyen, 2003,p.69; see also Kelso, both its own history of activation and the 1995). Such properties of non-linear systems history of activation of every other neuron, led in the 1970s to an increased interest in thanks to the intrinsic connectivity of the the mathematics of dynamical system the- nervous system. The precise influence of ory (Nicolis & Prigogine, 1977). a given neuron on a second one is deter- Several elements condition the behav- mined by the weight of the synapse that links ior of a dynamical system during a given them. Thus, the overall synaptic pattern in window of observation. First, the system’s the nervous system provides the main set of behavior is conditioned by the values of a set order parameters in such models (see Arbib, of so-called order parameters. By definition, 2002).4 To this desciption, it must be added these parameters determine the exact math- that the system is not isolated, but under the ematical equations that govern the system. constant influence of external sensory inputs This set of parameter values is a function that shape the behavior of peripheral sensory of the architecture of the system (e.g., the neurons. synaptic weights between neurons), factors There are many models available for the external to the system (e.g., outside temper- mathematical functions that link the mem- ature), and so on. These parameters cannot brane potentials of individual neurons to the necessarily be controlled, and their dynam- history of the larger system and to the exter- ics is slower than that of the system itself. nal inputs (Arbib, 2002). At this point, it P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

neurodynamical approaches to consciousness 735

is sufficient to state that these functions are is that their dynamics can organize around non-linear and that this is the reason why the presence of ‘strange’ attractors. A strange the nervous system is described as a (spa- attractor is a pattern of activity that captures tially extended) non-linear dynamical sys- nearby states (Arbib, Erdi,´ & Szentagothai,´ tem (for reviews, see Faure & Korn, 2001; 1997): It occupies a subregion of the state Korn & Faure, 2003).5 space, a manifold, and if the trajectory of the system comes into contact with this man- Chaos in the Brain ifold, the trajectory will stay on it subsequently, in the absence of external pertur- As Le Van Quyen (2003,p.69) notes, bations. The precise number and shapes of there was little echo to Ashby’s original pro- the attractors are determined by the param- posal to view the nervous system as a non- eters of the system, such as the intrinsic linear dynamical system, mostly because connectivity of the nervous system. Which the appropriate mathematical methods and particular attractor captures the system is computational tools to pursue this proposal determined by the system’s initial position. were lacking at that time. The real boost to When the parameters define several strange neurodynamics came later, in the 1980s and attractors, then there can be associations 1990s, with the widespread emergence in the between certain initial positions in the state scientific community of interest in the prop- space and certain attractors (Tsuda, 2001). erties of chaotic systems. This association is the basis for chaos-based Chaotic systems are simply non-linear perceptual systems: For example, in a com- systems, with their parameters set so that mon non-linear model of olfactory process- they possess an extreme sensitivity to ini- ing (as reviewed in Korn & Faure, 2003), tial conditions. Such sensitivity means that each odor is represented by a specific attrac- if one changes the initial position of the sys- tor, such that when confronted with slightly tem in its state space, however slightly, the different olfactory stimuli, the trajectory of subsequent positions on the modified trajec- the system will converge onto the same tory will diverge exponentially from what attractor, if the stimuli actually correspond they would have been otherwise. Given this to the same odorant.7 Thus, in this model, sensitive dependence on initial conditions, perception is based on several coexisting combined with the impossibility of deter- attractors in a multistable system. An addi- mining the present state of the system with tional and important feature is that external perfect precision, the future behavior of a perturbations to the system can make the chaotic system is unpredictable. The system system jump from one attractor to another; thus appears to have an inherent source of therefore, chaotic systems should not be variability, for it will never react twice in thought of as ‘static’ and unreactive to the the same way to identical external pertur- environment. Moreover, chaotic systems can bations, even in the absence of noise.6 be controlled; that is, they can be ‘forced’ The possible existence of ‘chaos in the to stay within specific portions of the state brain’ sparked much speculation and excite- space via external perturbations. It must ment. There were two related matters of immediately be added, however, that the debate: (i) whether the nervous system is term ‘forced’ is misleading, for the external actually chaotic (or whether there are sub- perturbations are nothing like brute force; systems in the brain that are; Faure & Korn, rather, they must be thought of as like a series 2001; Korn & Faure, 2003) and (ii) what of subtle touches, carefully chosen to adapt use the nervous system could make of such to the system’s dynamic properties. chaotic behaviors (Faure & Korn, 2001; Korn A sobering thought is that it is not clear & Faure, 2003; Skarda & Freeman, 1987; whether the activity of the nervous sys- Tsuda, 2001). tem, considered as a whole, is chaotic. One The second question proved to be the requirement for a system to be chaotic is easier one. One property of chaotic systems that its trajectory in the state space be P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

736 the cambridge handbook of consciousness

constrained within geometrical structures lation is no longer viewed as the simple trig- that have a lower dimension than the space gering of a systematic, prespecified, chain of itself (this requirement is needed mainly to neural operations that would unfold inde- distinguish between chaotic and stochastic pendently of the brain’s current activity, as processes; Wright & Liley, 1996). Unfortu- in a computer algorithm. For this reason, nately, the behavior of the nervous system the neurodynamical approach is often pre- cannot be observed directly in its actual sented as a sharp alternative to the computer state space, but only via limited sets of mea- metaphor of the brain (Freeman, 1999a; surements that are crude projections of its Kelso, 1995; Van Gelder, 1998). actual state (like the electroencephalogram or EEG, which retains only the average activ- Self-Organization and the Emergence ity of millions of neurons; Korn & Faure, of Spatiotemporal Patterns 2003).8 Nevertheless, the debate over the chaotic As Crutchfield and Kaneko (1987) note, nature of brain activity proved productive dynamical system theory has developed and brought out of the shadows some ideas largely through the study of low-dimen- crucial to neurodynamics. For instance, cen- sional systems, with no spatial extension. To tral to neurodynamical thought is the idea be useful for neuroscience, however, dynam- that the variability of neural activity may ical system theory needs to consider the be an integral part of the nervous sys- special properties conferred on the nervous tem’s dynamics. This notion is orthogonal system by its spatial extension. to a number of traditional (and still largely Fortunately, there has been a recent coin- dominant) approaches in neuroscience that cidence between, on the theoretical side, attribute this variability to ‘meaningless the development of a theory of large- noise’. As a case in point, most brain imag- scale non-linear systems and, on the exper- ing studies try to get rid of the variabil- imental side, the advent of multielectrode ity of the neural activity by averaging brain recordings and imaging techniques to map recordings over multiple repetitions of the precisely the electrical activity of entire pop- same process.9 These averaging procedures ulations of neurons. This coincidence has most likely give an oversimplified view of led to renewed interest, in the biological brain dynamics. In the near future, neurosci- community, in large-scale models of neural entists will undoubtedly have to go to the activity. trouble of making sense of the neural vari- The study of large, spatially extended ability by finding its experiential and behav- non-linear systems is a field in itself, in which ioral correlates. Fortunately, new approaches the interest in attractors shifts to the related along these lines are emerging, such as try- one of spatiotemporal patterns. (We rec- ing to understand the brain response to sen- ommend the reader spend a few minutes sory stimulation in the context of the brain’s looking for pictures of ‘cellular automata’ active state preceding the stimulation and with the Google image search engine to thus in relation to an active ‘baseline’ of neu- see some beautiful examples of spatiotem- ral activity that is far from neutral (Lutz, poral structures.) As a result, the neurody- Lachaux, Martinerie, & Varela, 2002; Engel namical community is now becoming less et al., 2001) focused on chaos and more focused on the This shift in the focus of brain imaging properties of self-organization in non-linear should not be underestimated: The brain’s systems, and particularly the formation of reaction is no longer viewed as a passive transient spatiotemporal structures in the ‘additive’ response to the perturbation, but brain.10 As noted by Freeman, in the pref- as an active ‘integration’ of the perturbation ace to the second (electronic) printing of his into the overall dynamics (Arieli, Sterkin, seminal book, Mass Action in the Nervous Grinvald, & Aertsen, 1996). In other words, System (Freeman, 1975): “The word ‘chaos’ the processing of an incoming sensory stimu- has lost its value as a prescriptive label and P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

neurodynamical approaches to consciousness 737

should be dropped in the dustbin of history, macroscale oscillations that can be picked but the phenomenon of organized disorder up by mesoscale recordings (such as local constantly changing with fluctuations across field potentials) or macroscale recordings the edge of stability is not to be discarded” (such as an EEG). For this reason, syn- (Freeman, 2004). chronous oscillations have been the easiest Spatiotemporal structures are ubiquitous form of spatiotemporal structure to measure in the brain. Apart from the obvious physical in the brain and not surprisingly, the first one construction of the system, they correspond to be observed (Bressler & Freeman, 1980; to the emergence of transient functional Gray, Konig, Engel, & Singer, 1989); see couplings between distributed neurons. For also the discussion of functional connectivity a given period of time, the activity of a below). set of neurons shows an increased level One reason why resonant cell assem- of statistical dependency, as quantified, for blies in particular, and spatiotemporal struc- example, by mutual information.11 In pio- tures in general, are so appealing to neuro- neering work, Freeman (reviewed in Free- science is because they provide a flexible and man, 2000a) observed spatiotemporal activ- reversible way to bind together distributed ity patterns in the olfactory bulb and neurons that may be primarily involved in interpreted them within the framework of very different functional processes. This type dynamical system theory. In an influen- of binding has three fundamental features: tial theoretical paper with Skarda, he pro- (i) the ability to integrate distributed neural posed that sensory information was encoded activities (integration); (ii) the ability to pro- in those patterns (Skarda & Freeman, mote, and virtually extract, one particular set 1987). of neural activities above the rest of the brain The classic example of spatiotemporal activity (segregation); and (iii) the capacity to structures in the brain is the Hebbian rever- evolve easily through a succession of flexi- berant cell assembly, which Hebb (1949) ble and adaptive patterns (metastability). For hypothesized to be the basis for short-term example, one resonant assembly could tran- memory (see Amit, 1994).12 (This notion is siently bind together the different popula- also closely related to Varela’s [1995] idea of tions of neurons involved in analyzing the resonant cell assemblies, described below.) shape, color, and motion of a visual object, Reverberant cell assemblies are labile sets and this temporary assembly would con- of neurons that transiently oscillate together stitute a neural substrate for the transient at the same frequency, at the level of their perception of a visual object. This idea is membrane potential. They are the best- the starting point of a very active stream of studied spatiotemporal structures in the research that we discuss later in this chap- brain. Indeed, the cortex has sometimes ter (for an overview, see Roskies, 1999). been modeled as a lattice of coupled oscil- Some authors have even proposed that every lators – in other words, as a juxtaposition cognitive act corresponds to the formation of reverberant cell assemblies. One advan- of such a transient spatiotemporal pattern tage of such models is that the behavior of (Varela, 1995). oscillator lattices has been abundantly inves- In summary, dynamical system theory tigated, mainly using numerical simulations proposes a precise framework to analyze (see Gutkin, Pinto, & Ermentrout, 2003; the spatiotemporal neural phenomena that Kuramoto, 1984; Nunez, 2000; Wright & occur at different levels of organization in Liley, 1996). the brain, such as the firing of individual The formation of spatiotemporal struc- neurons and the collective dynamics of syn- tures in such systems often takes the form of chronous oscillations within large networks. phase-synchronization patterns between the The future challenge is to relate these prop- oscillators (Le Van Quyen, 2003).13 In the erties of self-organization to various aspects brain, phase synchronization of large pop- of mental life. This endeavor is still in its ulations of neural oscillators can produce early phases, but the future looks promising. P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

738 the cambridge handbook of consciousness

For example, recent analysis methods mak- J. Freeman stands out as one of the most ing use of the brain’s dynamical properties elaborate and truly dynamical theories of have been proposed as a means to antic- brain function (Freeman, 1975, 1999a,b, ipate epileptic seizures (Martinerie et al., 2000a,b, 2004). His work is based mainly 1998). There is now a strong general sense on animal studies, in particular electrophys- that the properties of metastable, neural spa- iological recordings of the olfactory sys- tiotemporal patterns match crucial aspects tem of awake and behaving rabbits. This of conscious experience and that neurody- approach can be summarized as follows: namics may provide the tools and concepts The point of departure is the neuronal to understand how the neural activity crucial population. A neuronal population is an for consciousness temporally unfolds. This aggregate of neurons, in which, through pos- trend is patent in a set of influential neu- itive feedback, a state transition has occurred roscientific models of consciousness that we so that the ensemble presents steady-state, review in the next section of this chapter. non-zero activity. When negative feedback is established between populations, where one is excitatory and the other inhibitory, Examples of Neurodynamical oscillatory patterns of activity appear. This Approaches to Consciousness change implies a second state transition, where the resulting attractor is a limit cycle Introduction that reveals the steady-state oscillation of Although neurodynamics is quite popular the mixed (excitatory-inhibitory) popula- in the neurobiology of consciousness, it is tion. When three or more mixed popula- still not a widespread practice to formu- tions combine among themselves by further late theories about the relation between con- negative and positive feedback, the result- sciousness and the brain in purely dynamical ing background activity becomes chaotic. terms. Dynamical concepts are incorporated This chaotic activity now distributed among to varying degrees by different researchers the populations is the carrier of a spatial and used alongside concepts from informa- pattern of amplitude modulation that can tion theory or functionalist models of cogni- be described by the local heights of the tive processing. In this section, we review recorded waveform. some models that make use of dynamical When an input reaches the mixed pop- concepts in attempting to explain the phe- ulation, an increase in the non-linear feed- nomenon of consciousness. The list of mod- back gain will produce a given amplitude- els we cover is not meant to be exhaus- modulation pattern. The emergence of this tive, but rather a small sample of a large pattern is considered to be the first step in spectrum of dynamical approaches to brain perception: Meaning is embodied in these activity and consciousness. Furthermore, we amplitude-modulation patterns of neural do not intend to scrutinize these models in activity, whose structure is dependent on detail, but instead to highlight some com- synaptic changes caused by previous expe- mon aspects while providing an overview of rience. Thus, the whole history of the ani- their main proposals and hypotheses. The mal sets the context in which the emerg- reader is referred to the original sources for ing spatiotemporal pattern is meaningful. more details. Through the divergence and convergence of neural activity onto the entorhinal cortex, Neurodynamical Models the pulse patterns coming from the bulb of Consciousness are smoothed, thereby enhancing the macroscopic amplitude-modulation pattern while consciousness as order parameter attenuating the sensory-driven microscopic and dynamical operator activity. Thus, what the cortex ‘sees’ is a con- Among the different approaches to neural struction made by the bulb, not a mapping dynamics, the pioneering work of Walter of the stimulus. P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

neurodynamical approaches to consciousness 739

Hence, in Freeman’s view, perception dynamic large-scale integration is an essentially active process, closer to and radical embodiment hypothesis testing than to passive recovery Another proposal that falls squarely within of incoming information. This active stance the neurodynamical framework is one for- is embodied in the process of ‘preafference’ mulated initially by Francisco J. Varela by which the limbic system (including the (1995) and then developed with his collabo- entorhinal cortex and the hippocampus in rators, especially Evan Thompson (Thomp- mammals), through corollary discharges to son & Varela, 2001). Varela proposes to all sensory cortices, maintains an attentive address the question of how neural mech- expectancy of what is to come. The stimulus anisms bring about “the flow of adapted then confirms or disconfirms the hypothesis and unified cognitive moments” (Varela, through state transitions that generate the Lachaux, Rodriguez, & Martinerie, 2001,p. amplitude-modulation patterns described 229). The main working hypothesis is that previously.14 The multisensory convergence a specific neuronal assembly underlies the onto the entorhinal cortex becomes the basis operation of every unitary cognitive act. for the formation of Gestalts underlying the Here a neuronal assembly is understood as unitary character of perception. a distributed set of neurons in the brain Finally, through multiple feedback loops, that are linked through reciprocal and selec- global amplitude-modulation patterns of tive interactions, where the relevant vari- chaotic activity emerge throughout the able is no longer single-neuron activity, but entire hemisphere, directing its subsequent rather the dynamic nature of the links that activity. These loops comprise feedforward are established between them. Varela and flow from the sensory systems to the entorhi- collaborators propose that such dynamical nal cortex and the motor systems, and feed- links are mediated by the transient establish- back flow from the motor systems to the ment of phase relations (phase synchrony) entorhinal cortex, and from the entorhinal across multiple frequency bands, especially cortex to the sensory systems. Such global in the beta (15–30 Hz) and gamma (30– brain states “emerge, persist for a small frac- 80 Hz) range (Varela et al., 2001). More- tion of a second, then disappear and are over, the transient nature of such dynamical replaced by other states” (Freeman, 1999b, links (and therefore of the neural assemblies p. 153). themselves) is central to the idea of large- For Freeman, it is this level of emergent scale integration, for it brings to the fore and global cooperative activity that is crucial the notion that the system, rather than pre- for consciousness, as these remarks indicate: senting a series of well-defined states (attrac- “Consciousness ...is a state variable that tors), shows metastable (self-limiting and constrains the chaotic activities of the parts recurrent) patterns of activity: “In the brain, by quenching local fluctuations. It is an order there is no ‘settling down’ but an ongo- parameter and an operator that comes into ing change marked only by transient coor- play in the action-perception cycle as an dination among populations, as the attractor action is being concluded, and as the learn- itself changes owing to activity-dependent ing phase of perception begins” (Freeman, changes and modulations of synaptic con- 1999a, p. 132). Furthermore: “[T]he globally nections” (Varela et al., 2001,p.237). Large- coherent activity ...may be an objective cor- scale integration through phase relations relate of awareness. ...In this view, aware- becomes fundamental for understanding ness is basically akin to the intervening state brain dynamics as coordinated spatiotempo- variable in a homeostatic mechanism, which ral patterns and provides a plausible solu- is both a physical quantity, a dynamical oper- tion to the problem of how to relate the ator, and the carrier of influence from the local specificity of activity in specialized past into the future that supports the rela- cortical regions to the constraints imposed tion between a desired set point and an exist- by the connectivity established with other ing state” (Freeman, 1999b, p. 157). distributed areas. We see in the next two P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

740 the cambridge handbook of consciousness

sections below, as well as in the other dynam- munication. This last type of cycle depends ical approaches described in this section, on various levels of sensorimotor coupling, that this balance of segregation and integra- mediated in particular by the so-called mir- tion has been considered the hallmark of ror neuron systems that show similar pat- brain complexity and a plausible prerequi- terns of activation for both self-generated, site for consciousness. goal-directed actions and when one observes Thompson and Varela (2001) then qual- someone else performing the same action ify this view by placing it in a ‘radical (Rizzolatti & Craighero, 2004). embodiment’ framework. They propose As a final aspect of this proposal, Thomp- that, although the neural processes relevant son and Varela hypothesize that “conscious- to consciousness are best mapped at the level ness depends crucially on the manner in of large-scale, transient spatiotemporal pat- which brain dynamics are embedded in the terns, the processes crucial for consciousness somatic and environmental context of the are not brain-bound events, but comprise animal’s life, and therefore there may be no also the body embedded in the environment. such thing as a minimal internal neural corre- By taking into account the notion of emer- late whose intrinsic properties are sufficient gent processes as understood in complex sys- to produce conscious experience” (Thomp- tems theory (order parameters or collective son & Varela, 2001,p.425; see also Noe&¨ variables and the boundary conditions they Thompson, 2004a). impose on local activities), they propose that conscious awareness (as an order parameter cortical coordination dynamics or dynamical operator) is efficacious with Based on extensive work in human motor respect to local neural events (see also Free- coordination, J. A. Scott Kelso has developed man 1999a,b, and above) and that the pro- a detailed dynamical framework for under- cesses crucial for consciousness so understanding human cognition (Kelso, 1995). stood span at least three ‘cycles of operation’ His main focus is the appearance of self- that cut across brain-body-world divisions organized patterns caused by non-linear (Thompson & Varela, 2001,p.424). The first interactions between system components, at is the regulatory organismic cycle, in which both the neural and motor levels, as well as the maintenance of internal variables within their role in human behavior. Kelso views the a viable range is achieved “through sensors brain as fundamentally “a pattern forming and effectors to and from the body that link self-organized system governed by poten- neural activity to the basic homodynamic tially discoverable, non-linear dynamic laws” processes of internal organs.” This cycle is (Kelso, 1995,p.257). He proposes that cog- supposed to be the basis of the “inescapable nitive processes “arise as metastable spa- affective backdrop of every conscious state,” tiotemporal patterns of brain activity that also called ‘core consciousness’ (Damasio, themselves are produced by cooperative 1998, 1999) or ‘primary-process conscious- interactions among neural clusters” (257). ness’ (Panksepp, 1998). The second cycle He then goes one step further, proposing is sensorimotor coupling between organism that “an order parameter isomorphism con- and environment, whereby what the organ- nects mind and body, will and brain, mental ism senses is a function of how it moves and neural events. Mind itself is a spatiotem- and how it moves is a function of what poral pattern that molds the metastable it senses. Here, “transient neural assemblies dynamic patterns of the brain” (288). mediate the coordination of sensory and What are the specific neural mecha- motor surfaces, and sensorimotor coupling nisms underlying the establishment of such with the environment constrains and mod- self-organized patterns? Kelso, in collabo- ulates this neural dynamics.” The third is a ration with Steven Bressler, proposes that cycle of intersubjective interaction involving the answer lies in the notion of ‘coordi- the recognition of the intentional meaning nation dynamics’ (Bressler & Kelso, 2001). of actions and (in humans) linguistic com- Coordination dynamics is presented as an P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

neurodynamical approaches to consciousness 741

integrative framework, in which the main sense, etc., come together in a unique coher- issue is “to identify the key variables of coor- ent experience. Differentiation is the capac- dination (defined as a functional ordering ity to experience any of a vast number among interacting components) and their of different possible conscious states. This dynamics (rules that govern the stability capacity is intimately tied to what Edel- and change of coordination patterns and the man and Tononi call the informativeness of non-linear coupling among components that conscious experience, where each conscious give rise to them)” (Bressler & Kelso, 2001, state would be highly informative, given p. 26). Using this framework, Bressler and the reduction in uncertainty that is accom- Kelso address the question of how inter- plished by the selection of one among a acting, distributed cortical areas allow the potentially infinite number of possible states. emergence of ongoing cognitive functions. Edelman and Tononi stress that con- On the basis of previous studies of biman- sciousness is not a thing, but a process, and ual coordination, they propose more specif- therefore should be explained in terms of ically that the relevant collective variable neural processes and interactions and not is the relative phase (the continuous phase in terms of specific brain areas or local difference) among the given neural struc- activities. More specifically, they postulate tures, which are themselves considered to that to understand consciousness it is nec- be accurately described by non-linear oscilla- essary to pinpoint neural processes that tors. They argue that this coordination vari- are themselves integrated, yet highly differ- able is adequate because (i) it reveals the entiated. Their answer to this problem is spatiotemporal ordering between interact- what they call ‘the Dynamic Core hypoth- ing structures, (ii) changes in the relative esis’ (Edelman & Tononi, 2000; Tononi & phase occur more slowly than changes in Edelman, 1998). They describe this hypoth- the local component variables, and (iii) rel- esis as follows:, ative phase shows abrupt changes during phase transitions or bifurcations. When the 1) a group of neurons can contribute directly two coordinated local neuronal populations to conscious experience only if it is part of a have different intrinsic frequencies, the rel- distributed functional cluster that achieves high integration in hundreds of millisec- ative phase shows a metastable regime in onds. 2 ) To sustain conscious experience, the form of ‘attractiveness’ toward preferred it is essential that this functional cluster be modes of coordination, without settling into highly differentiated, as indicated by high any unique one. Accordingly, Bressler and values of complexity. We call such a cluster Kelso propose that “a crucial aspect of cog- of neuronal groups that are strongly inter- nitive function, which can both integrate acting among themselves and that have dis- and segregate the activities of multiple distinct functional borders with the rest of the tributed areas, is large-scale relative coor- brain at the time scale of fractions of a sec- dination governed by way of metastable ond a ‘dynamic core,’ to emphasize both dynamics” (Bressler & Kelso, 2001,p.30). its integration and its constantly changing composition. A dynamic core is therefore the ‘dynamic core’ hypothesis a process, not a thing or a place, and it is defined in terms of neural interactions, Gerald M. Edelman and Giulio Tononi have rather than in terms of specific neural loca- developed an account of the neural basis tions, connectivity or activity (Edelman & of consciousness that aims to explain two Tononi, 2000,p.144). fundamental properties of conscious experience, which they call ‘integration’ and In addition, they argue that “the dynamic ‘differentiation’ (Edelman & Tononi, 2000; core is a functional cluster: its participat- Tononi & Edelman, 1998). Integration refers ing neuronal groups are much more strongly to the unitary character of conscious expe- interactive among themselves than with the rience, whereby the multiplicity of aspects, rest of the brain. The dynamic core must such as color, taste, audition, kinesthetic also have high complexity: its global activity P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

742 the cambridge handbook of consciousness

patterns must be selected within less than ception, and consciousness. Therefore, we a second out of a very large repertoire” believe that it is important to keep these (Tononi & Edelman, 1998,p.1849). They models in mind as part and parcel of the hypothesize that the dynamic core achieves wider research program of neurodynamics. integration on the basis of reentrant interactions among distributed neuronal groups, the cortico-thalamic dialogue most likely mediated by the thalamocor- Rodolfo Llinas and his collaborators have tical system. Specifically, for primary con- proposed a model of how consciousness sciousness to arise, interactions are required is related to brain activity, in which the between sensory cortices in different modal- notion of emergent collective activity plays ities and value-category and memory sys- a central role (Llinas & Pare, 1991; Llinas tems in frontal, temporal, and parietal & Ribary, 2001; Llinas, Ribary, Contreras, & areas.15 Pedroarena, 1998). In particular, Llinas pos- Edelman and Tononi claim that the tulates that consciousness arises from the dynamic core provides a “neural reference ongoing dialogue between the cortex and space for conscious experience” (Edelman & the thalamus (Llinas & Pare, 1991). He calls Tononi, 2000,p.164). They depict this space attention to the fact that most of the input to as an n-dimensional neural space, where the thalamus comes from the cortex, rather the number of dimensions is given by the than from peripheral sensory systems. On number of neuronal groups that are part this basis, he proposes that the brain be con- of the dynamic core at that moment. Such sidered as a ‘closed system’ that can gen- neuronal groups would be segregated into erate and sustain its own activity thanks neural domains specialized for various func- to the intrinsic electrical properties of neu- tions, such as form, color, or orientation dis- rons (Llinas, 1988) and the connectivity they crimination, proprioceptive or somatosen- establish. The interplay of these two main sory inputs, and so on, and they would be characteristics underlies the establishment brought together through re-entrant inter- of “global resonant states which we know as actions. The local activities of these groups cognition” (Llinas & Ribary, 2001,p.167). would therefore need to be understood in A crucial feature of this proposal is the relation to the unified process constituted precise temporal relations established by by the functional cluster; that is, the entire neurons in the cortico-thalamic loop. This dynamic core: “The pure sensation of red temporal mapping is viewed as a ‘functional is a particular neural state identified by geometry’ and involves oscillatory activity at a point within the N-dimensional neural different spatial scales, ranging from individ- space defined by the integrated activity of ual neurons to the cortical mantle. In par- all the groups of neurons that constitute the ticular, 40-Hz oscillations that traverse the dynamic core. ...The conscious discrimina- cortex in a highly spatially structured man- tion corresponding to the quale of seeing red ner are considered as candidates for the pro- acquires its full meaning only when consid- duction of a “temporal conjunction of rhyth- ered in the appropriate, much larger, neural mic activity over large ensemble of neurons” reference space” (Edelman & Tononi, 2000, (Llinas & Ribary, 2001,p.168). Such gamma p. 167). oscillations are believed to be sustained by a thalamo-cortical resonant circuit involving pyramidal neurons in layer IV of the neo- Related Models cortex, relay thalamic neurons, and reticu- Several other authors have advanced mod- lar nucleus neurons. In particular, tempo- els in which dynamical system concepts are ral binding is supposed to be generated by present, yet appear less explicitly. Neverthe- the conjunction of a specific circuit involving less, these approaches also aim to describe specific sensory and motor nuclei projecting the formation of spatiotemporal patterns of to layer IV and the feedback via the reticular brain activity that are crucial for action, per- nucleus, and a non-specific circuit involving P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

neurodynamical approaches to consciousness 743

non-specific intralaminar nucei projecting to via dynamical cell assemblies. (4) Binding of the most superficial layer of the cortex and such assemblies is effected by transient syn- collaterals to the reticular and non-specific chrony that establishes a code for relatedness thalamic nuclei. Thus, the ‘specific’ system among features and facilitates downstream is supposed to supply the content that relates evaluation and impact. (5) Such assemblies to the external world, and the non-specific need desynchronized EEG16 (which corre- system is supposed to give rise to the tempo- lates with phenomenal awareness in the ral conjunction or the context (on the basis waking state and REM dream state and of a more interoceptive context concerned shows high-frequency beta and gamma oscil- with alertness). Together they would gener- latory activity), and are facilitated by atten- ate a single cognitive experience (Llinas & tion (Singer, 1998). Ribary, 2001,p.173). Engel, Fries, and Singer (Engel, Fries, & Singer, 2001; Engel & Singer, 2001) also timing and binding explicitly espouse a ‘dynamicist view’ of Wolf Singer and collaborators have exten- brain function. According to this view, sively investigated the issue of temporal cor- brain processes are not passive, stimulus- relations between cortical neurons and the driven, and hierarchical, but active, context- role this phenomenon could play in solv- dependent, endogenously driven, and dis- ing what has been called ‘the binding prob- tributed. In particular, “spatio-temporal pat- lem’ (Engel & Singer, 2001; Gray et al., 1989; terns of ongoing activity ...translate the Singer & Gray, 1995). This is the problem functional architecture of the system and its of how the signals from the separate neu- pre-stimulation history into dynamic states ronal populations concerned with distinct of anticipation” (Engel et al., 2001,p.705). object features (color, shape, motion, etc.) In this dynamicist account of top-down are bound together into a unified percep- influences, relevant patterns are generated tual representation. The main idea behind as a result of continuous large-scale interac- their approach is that there is a “tempo- tions, and these patterns can bias the saliency rary association of neurons into functionally of sensory signals by changes in their tempo- coherent assemblies that as a whole repre- ral correlations. Endogenous, self-generated sent a particular content whereby each indi- activity displays distinct spatiotemporal pat- vidual neuron is tuned to one of the ele- terns, and these patterns bias the self- mentary features of composite perceptual organizing process that leads to the tempo- objects” (Singer, 1998,p.1831). The specific ral coordination of input-triggered responses hypothesis is that neurons become mem- and their binding into functionally coher- bers of such coherent assemblies through the ent assemblies. This dynamicist approach precise synchronization of their discharges; thus stresses the importance of top-down in other words, such synchronization estab- influence in the form of large-scale dynam- lishes a “code for relatedness” (Singer, 1998, ics that express contextual influences and p. 1837). stored knowledge in the system and that Recently, Singer and colleagues (Engel can modulate local processing and hence the & Singer, 2001, Engel et al., 1999) have downstream effect of the impinging event extended this framework to address the issue (Engel et al., 2001). of phenomenal awareness. Their argument can be summarized as follows: (1) Brains the neural correlates of capable of phenomenal awareness should be consciousness able to generate metarepresentations of their Francis Crick and Christof Koch have own cognitive processes. (2) Metarepresen- employed dynamical concepts in a series tations are realized by an iterative process, of proposals regarding the relation between in which higher-order cortical areas read neural activity and conscious perception low-order (sensory) areas. (3) Combinatorial (Crick & Koch, 1990, 1998, 2003). In their flexibility of metarepresentations is obtained view, the best way for the neuroscience P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

744 the cambridge handbook of consciousness

of consciousness to proceed is first to snapshots are possibly related to alpha and uncover the neural correlates of conscious- theta rhythms and are the reflection of a cer- ness (NCCs), in particular the neural contain threshold that has been overcome (for tents of visual consciousness. They define a a certain amount of time) by neural activ- neural correlate of consciousness as a mini- ity, enabling it to become conscious. Con- mal set of neuronal events necessary and suf- scious coalitions would therefore be con- ficient for a given state of phenomenal continually “forming, growing or disappearing” sciousness (see also Chalmers, 2000). Here (Crick & Koch, 2003,p.122). Crick and we summarize the version of their theory Koch propose that attention is fundamental presented in one of their last joint articles in biasing the competition among coalitions on consciousness (Crick & Koch, 2003). that share critical nodes. Attention produces They begin with the notion of an ‘uncon- the effective binding of different attributes scious homunculus’, which is a system con- of the given conscious content by means of sisting of frontal regions of the brain “looking shared “membership in a particular coali- at the back, mostly sensory region” (Crick tion” (Crick & Koch, 2003,p.123). & Koch, 2003,p.120). Crick and Koch Although Crick and Koch recognize that propose that we are not conscious of our the mechanism to establish such coali- thoughts, but only of sensory representations probably involves synchronous fir- tions of them in imagination. The brain ing between distributed populations, they presents multiple unconscious processing explicitly state that they no longer believe modules, mostly feedforward, that act as that 40-Hz oscillatory activity is a suffi- ‘zombie’ modes. These modules present cient condition for consciousness. Finally, stereotyped responses in a sort of ‘cortical they propose that there is a set of neural pro- reflex’, whereas conscious modes are nec- cesses that, although not part of the NCC, essary only to deal with time-consuming, is affected by the NCC, both with respect less stereotyped situations that need plan- to its actual firing and with respect to synap- ning and decision making. The most impor- tic modifications caused by previous expe- tant point, however, with regard to the NCC rience. This ‘penumbra’ (Crick and Koch, issue, is the existence of dynamic coalitions 2003,p.124), could eventually become con- of neurons in the form of neural assemblies scious, if incorporated into the NCC. whose sustained activity embodies the contents of consciousness. Explicit representa- consciousness as global workspace tions of particular aspects of the (visual) Stanislas Dehaene, Jean-Pierre Changeux, scene are present in special brain regions and collaborators have explored an alterna- (‘critical nodes’), and these representations tive model of brain functioning that under- are bound together in the dynamic neural lies the accessibility to verbal report of coalitions. Additionally, Crick and Koch sug- conscious experience (Dehaene, Kerszberg, gest that higher levels of cortical processing & Changeux, 1998; Dehaene & Naccache, are first reached by the feedforward sensory 2001; Dehaene, Sergent, & Changeux, sweep and that only through backpropaga- 2003). The main proposal of their model tion of activity from higher to lower levels do is the existence of “two main computa- the lower levels gain access to this informational spaces within the brain” (Dehaene, tion. They distinguish ‘driving’ from ‘modu- Kerszberg, & Changeux, 1998,p.14529). lating’ connections and suggest that the feed- The first computational space consists of a forward sweep is mostly driving activity in series of functionally segregated and spe- the frontal regions, whereas the feedback cialized modules or processors that consti- return onto sensory cortices is mainly mod- tute a parallel distributed network (exam- ulating. In the specific case of conscious per- ples of modular processors would be primary ception, they propose that it is not a contin- visual cortex (V1) or the mirror neuron sys- uous phenomenon, but rather that it works tem in area F5 of the premotor cortex). The on the basis of a series of ‘snapshots’. Such second computational space is not confined P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

neurodynamical approaches to consciousness 745

to a series of brain areas, but rather is dis- distance connections to a vast set of defined tributed among multiple cortical regions. areas, thus creating a global and exclusive The main property of this second space is availability for a given stimulus, which is massive reciprocal connectivity on the basis then subjectively experienced as conscious” of horizontal projections (long-range cor- (Dehaene, Kerszberg, & Changeux, 2003). tico-cortical connections). Through descending connections, this ‘global workspace’ summary determines the contributions of the modular The majority of the approaches reviewed processors of the first computational space above stress the importance of a certain type by selecting a specific set while suppress- of distributed, spatiotemporal pattern of ing another. Through this selective mobiliza- neural activity that ‘demarcates’ itself from tion of the specialized processors into the the background activity of the brain. Such global workspace. a ‘brain scale’ state can be patterns are described as ongoing, transient, reached, in which a group of workspace neu- metastable coordination processes among rons are spontaneously coactivated while the separate neurons, and they are considered to rest are suppressed. As a result, an exclusive be crucial for the moment-to-moment emer- ‘representation’ invades the workspace and gence and formation of conscious experi- may remain active in an autonomous man- ence. Another related feature crucial to sev- ner and resist changes in peripheral activ- eral of the above approaches is that these ity. If it is negatively evaluated, or if atten- spatiotemporal patterns reveal the interplay tion fails, it may however be spontaneously of two apparently fundamental principles and randomly replaced by another discrete of brain organization and function; namely, combination of workspace neurons. Func- functional segregation and cooperative inter- tionally, this neural property implements action or integration. This interplay and an active ‘generator of diversity,’ which con- the dynamical properties of the brain’s spa- stantly projects and tests hypotheses (or tiotemporal activity patterns are the focus of pre-representations) on the outside world. the following sections. The dynamics of workspace neuron activity is thus characterized be a constant flow of individual coherent episodes of The Search for Meaningful variable duration (Dehaene, Kerszberg, & Spatiotemporal Patterns in the Brain Changeux, 1998,p.14530). This postulated workspace has access to the Introduction world through ‘perceptual circuits’; ‘motor Despite their significant differences, all the programming circuits’ enable action guida- above models agree that the constitution of nce’; ‘long-term memory circuits’ enable dynamic spatiotemporal patterns of neural access to past experiences; ‘evaluation activity plays a central role in the emergence circuits’ allow negative-positive judg- of consciousness. This section discusses the ments; and ‘attention circuits’ endow the practical aspects of the search for such pat- workspace with the capacity to alter its own terns. After a short review of the connec- activity separately from the influence of tivity of the brain, we discuss the detection external inputs. Through connections with of such patterns in real brain data. A short motor and language centers, the workspace mathematical presentation leads us to the makes its resident representation avail- concept of synchrony, which is the preferred able for verbal report by the subject. Thus, candidate to date for such patterns. Dehaene and Changeux see consciousness as a selective global pattern: “When a piece of Connectivity in the Brain information such as the identity of a stimulus accesses a sufficient subset of workspace The organization of the brain’s connectivity neurons, their activity becomes self- is what ultimately determines the form of sustained and can be broadcasted via long- the neural spatiotemporal patterns. For this P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

746 the cambridge handbook of consciousness

reason, it is useful to start with a review of rule clearly favors the establishment of recur- some basic facts about this architecture. sive loops. Nevertheless, some basal gan- The brain is probably one of the most glia nuclei present a slightly different con- complex biological systems we know (Edel- nectivity structure: Although they receive man & Tononi, 2000). Its complexity is cer- axons from cortical neurons, they project tainly due in great part to its histological only through the thalamus into frontal lobe and morphological structure, and one of the regions (Edelman & Tononi, 2000). most striking aspects of the brain as a sys- Interestingly, however, no one zone in the tem is the connectivity pattern it exhibits. brain can be distinguished as the ultimate This pattern is that of a compact but dis- highest level, at least in terms of the connec- tributed tissue, with local clusters of highly tivity patterns. Indeed, the massively interconnected neurons that establish long-range connected nature of the brain suggests that interactions. In general, two neurons in the dynamic relations between local and dis- brain are always in interaction either directly tant activities will necessarily be established or via a certain number of intermediate cells. whatever the observed origin of a given It is useful to distinguish two levels of con- activation is. On the other hand, it is true nectivity in the brain. that clusters of more strongly interconnected Local connections: Several types of neu- regions are evident. Stephan and collabora- rons coexist in the neocortex. Within a tors recognize at least three main clusters given portion of neocortex, a complex in the primate cortex: (i) visual (occipito- arrangement of pyramidal, spiny stellate, and temporal); (ii) somatomotor (mainly pre- smooth stellate cells can be found. This and post-central, but extending into pari- arrangement of collateral axons, dentritic etal regions); and (iii) orbito-temporopolar- trees, and cell bodies gives rise to clusters insular (Stephan et al., 2000). It is inter- of interconnected neurons that extend over esting that the overall structural connectiv- a fraction of a millimeter. Neurons tend to ity (Hilgetag & Kaiser, 2004) and functional organize into radial clusters that share func- connectivity (Stephan et al., 2000) show a tional characteristics, known as functional ‘small world’ architecture. Networks hav- columns. These structures are particularly ing such an architecture display remarkable evident in somatic sensory cortex and visual properties, such as reduced average length cortex and are believed to play a fundamen- path (reaching any node from any other tal role in basic discriminative capacities. node is accomplished in a minimal number Long-range connections: In addition to the of steps), high synchronizability, enhanced bodies and dentritic trees of the local neu- signal propagation speed, and stability (one rons, axons from deep structures and other can randomly eliminate links without affect- cortical regions terminate at different points ing substantially the network properties; in the six-layered structure of the neocor- Watts & Strogatz, 1998). tex. Likewise, pyramidal neurons in a given One of the most important conclusions region of the neocortex have axons that of the study of the brain as a system is that, extend into the white matter and reach both despite its massive interconnectedness, the deep structures and other cortical regions. brain shows a strong segregation into clus- At least four patterns of long-distance con- ters at both structural and functional lev- nectivity can be distinguished in the brain els. The interplay of these two characteristic (Abeles, 1991): (i) between cortical neu- features of the brain lies at the basis of one rons within one hemisphere, (ii) between of the most interesting issues in contempo- cortical neurons of different hemispheres, rary neuroscience – the large-scale integra- (iii) between cortical neurons and deep tion of brain activity and its role in the uni- nuclei, and (iv) between brainstem modu- ﬁed nature of experience (James, 1890/1981; latory systems and extended areas of the Varela, 1995; Von der Malsburg, 1981). cortex. In general, long-range connectivity The combination of extensive neuropsy- obeys a reciprocity rule (Varela, 1995): If chological studies since Broca and the explo- A projects to B, then B projects to A. This sive use of imaging techniques in the last P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

neurodynamical approaches to consciousness 747

15–20 years has highlighted two main prin- regions establish interdependent activation ciples of brain functioning (Edelman & to account adequately for the integration Tononi, 2000; Friston, 2002a,b, 2005). On of functionally separate activity (Bressler, the one hand, a functional encapsulation 1995). Thus, one task facing the neurody- is evident: Distinct regions in the brain namicist is to detect such interdependent contribute differentially to different aspects activation from the brain recordings avail- of adaptive behavior; for instance, bilateral able today. damage to the human homologue of V5/MT (in the middle temporal area) can lead to Detecting Interdependent Activations a restricted impairment in the capacity to from Real Brain Recordings discriminate movement (akinetopsia; Zihl, Von Cramon, & Mai, 1983). On the other the data hand, for a given cognitive task, it is rarely The activity of the brain can be recorded the case that only one isolated region shows at several different spatial and temporal significant activation. For example, directing scales. The neurodynamicist will be primar- attention to a particular location of the visual ily interested in those techniques that are field correlates with the concomitant activa- fast enough to follow the formation of spa- tion of several cortical regions, preferentially tiotemporal patterns in the time scale of right parietal, anterior cingulate, and occip- hundreds of milliseconds. Because the con- ital cortices (Mesulam, 1999). struction of such patterns often involves Indeed, as we saw above, the connectiv- activities at the millisecond time scale ity pattern of the mammalian brain reveals a (e.g., in the case of fast neural oscilla- complex structure of recursively connected tions), the desired temporal resolution is on distant areas (Hilgetag & Kaiser, 2004; this order of milliseconds. In practice, this Stephan et al., 2000). Although the cortico- excludes the neuroimaging techniques based cortical connectivity pattern is paradigmatic, on slow metabolic measures, such as func- the structure of recursive connections is tional magnetic resonance imaging (fMRI) reflected most prominently in the thalam- or positron emission topography (PET).17 ocortical matrix (Edelman & Tononi, 2000; Millisecond temporal resolution is accessi- Llinas & Ribary, 2001). For instance, the lat- ble through direct intracellular and extra- eral geniculate nucleus (LGN) of the tha- cellular measurements of individual neu- lamus receives only around 5–10%ofits rons and recordings of local field potentials inputs (not more than 20%) from the retina, (LFPs) or of the electromagnetic fields of whereas the remaining connections come large neural populations that produce the from local inhibitory networks, descend- electroencephalographic (EEG) and mag- ing inputs from layer VI of the visual cor- netoencephalographic (MEG) signals. LFPs tex, and ascending inputs from the brain- are the summation of the membrane poten- stem (Sherman & Guillery, 2002). Yet the tials of populations of neurons. The size of LGN is the major relay in the visual path- the populations depends on the site and pre- way from the retina to the cortex. Such a cision of the recordings: Local microelec- complex structure of recursive, re-entrant, trodes can record small populations, extend- and interconnected networks that pervade ing over less than a square millimeter of the mammalian brain (Edelman & Tononi, tissue, whereas scalp EEG electrodes or 2000) strongly suggests the existence of con- MEG sensors (and optical imaging) record stitutive cooperative interactions, and there- the average activity of several square cen- fore integrative activity, among different timeters of cortex. At an intermediate level, regions. intracranial recordings from human patients Nevertheless, the presence of anatomical can record from a couple of square millime- connectivity is not enough to explain effec- ters of cortex (Lachaux, Rudrauf, & Kahane, tive interactions among separate regions 2003). Except in those rare situations justi- (Friston, 2002a,b). Indeed, in addition to fied by therapeutical reasons, human record- being connected, it is necessary that such ings are almost exclusively non-invasive and P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

748 the cambridge handbook of consciousness

performed therefore at the centimeter-wide dimensional state space, and the evolution spatial resolution of MEG or EEG. of this position in time defines a trajectory. With this panel of recording techniques, If the measurements are independent from spatiotemporal patterns can in principle be each other, then the trajectory will progres- observed at three levels: (i) as interactions sively completely fill a hypercubic portion of between simultaneous recordings of multi- the state space, leaving no hole. In contrast, ple individual neurons, (ii) as interactions if there are interactions among the mea- between simultaneous recordings of multi- sured neuronal populations, then the trajec- ple individual LFPs, and (iii) in single LFP tory will fit into a restricted portion of the recordings. The third level is intermediate full space and be constrained onto a manifold between the first two: Because an individual with a (fractal) dimension less than that of LFP records from a single neural population, the state space.18 What this means in infor- the average activity of the LFP is sensitive mational terms is that, for at least one pair of to the spatiotemporal organization of activ- the measured neural populations, measuring ity within this population. For example, if the activity of the first population provides all the neurons are coactive periodically, the some information about the activity of the average activity in the LFP will be a massive second one. The probability distribution of oscillation, much stronger than if the neu- the activity of Population 2 (the probabil- rons are not synchronous. ity p(y) that this activity is y), given that the activity of Population 1 is x, is different from some simple mathematical what it would be if the activity of Popula- considerations tion 1 were x. Consider this metaphor: If As we have seen, the organization of the we know where John will spend the after- brain suggests that interactions among dis- noon, we can predict with some accuracy tributed neuronal groups are bound to occur, that his wife Ann will spend the afternoon in given their massive interconnectedness. We the same city, but we cannot predict where also mentioned that recordings of neuronal Jane, unrelated to John, will be. Certain mea- activity can be obtained by a diversity of sures, such as mutual information (David, approaches and at several levels of spatial Cosmelli, & Friston, 2004), quantify exactly resolution. With these points in mind, let us this sharpening of the probability distribu- return to the central question of this section: tion. How does one detect neural spatiotempo- The transient nature of neural inter- ral patterns from real brain data? The defi- actions, however, makes general measures nition of the ‘dynamic core’ by Tononi and based on such geometrical formulations dif- Edelman (1998) provides a useful starting ficult to apply. The main problem is that, point: “The dynamic core is a functional to know whether or not the trajectory fills cluster: its participating neuronal groups up the whole space, the experimenter needs are much more strongly interactive among to observe it during time windows that are themselves than with the rest of the brain.” typically orders of magnitude longer than The challenge for the neurodynamicist is the typical lifespan of cell assemblies. This therefore to find neurons or groups of neu- difficulty can be avoided if the researcher rons with particularly strong (but transient) assumes a priori what will be the shape interactions. of the manifold onto which the trajectory In keeping with the dynamical approach, is constrained. Because spatiotemporal pat- we can usefully consider this question in terns can potentially take an infinite num- geometrical terms. Consider the n simulta- ber of shapes in the state space, a possible neous measures of brain activity that one solution is to assume a specific shape and to can record in a typical electrophysiologi- build a special detector for this shape (surf- cal setting (e.g., 64 measures from 64 EEG ing on the advances of signal processing). It channels). At any time t, the n simultane- is easy to see that it takes fewer measure- ous measurements define a position in an n- ments to test whether the trajectory stays on P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

neurodynamical approaches to consciousness 749

a circle or whether it follows some general, tend to fire at periodic latencies. Numerous unknown, geometrical structure. animal studies conducted over the past 20 In the absence of noise, three succes- years have now established that synchrony is sive measurements are sufficient to know ubiquitous in virtually all sensory and motor whether the trajectory stays on a straight modalities. It has often been found to be line, which would correspond to a linear related to perception, memory, and motor relationship between the recorded activities. programming (see Roskies, 1999, for a group Synchrony between quasi-periodic oscilla- of excellent reviews that summarize these tors – that is, the transient phase-locking of results). their oscillations – is a good example of such Synchrony has also been studied in hu- an interaction. But linear relationships can mans as an instance of spatiotemporal pat- be extended to other trajectories constrained terns of interdependent neural activity, on simple manifolds of dimension 1,19 such although at a different level from animal as a circle. This is the case for two oscilla- studies. Here an important distinction needs tors rotating at the same frequency with a to be made between the local recordings constant phase lag. of individual neurons, almost always accessible only in animals, and the more global recordings of entire neural populations, Synchrony: Perhaps Not the Best accessible in humans through scalp EEG Candidate, but at Least the Simplest or MEG. EEG and MEG average across The practical reason that synchrony has so large neuronal assemblies, and hence oscil- far been the best-studied (if not the only) latory synchrony between neurons shows type of transient interaction between neu- up as changes of power in particular fre- ral populations is the ease with which it quency bands. The reason this happens is can be detected. Furthermore, since the that groups of synchronously firing neu- development of the EEG, it has been evi- ral oscillators can be modeled as oscillators dent that oscillations are ubiquitous in the themselves, with the amplitude of the oscil- brain. This fact, combined with the rela- lations depending on the number of individ- tion between coordinated oscillatory activ- ual oscillators in the group and on the pre- ity and several important cognitive functions cision of the synchrony between them. This (discussed below), has also contributed to point entails a further distinction: On the the development of approaches that seek to one hand, oscillatory activity as recorded by detect the occurrence of synchrony from real an individual EEG electrode or MEG sen- neurobiological signals. sor implies already a certain amount of local In its original neurophysiological formu- synchronous activity. On the other hand, lation, ‘synchrony’ refers to a positive cor- one can choose to consider synchroniza- relation between the spike timing of a set tion between oscillations produced by dis- of neurons. In other words, if we consider tant neuronal populations (separated by sev- two neurons within a synchronous popula- eral centimeters) to describe distributed spa- tion, the probability of the first neuron to tiotemporal patterns that occur at a more fire a spike is significantly higher at specific global level. In any case, when dealing with delays relative to the spikes of the other EEG and MEG recording one is always in neuron.20 In the simplest case, this delay is the presence of noisy data. Hence, any inter- zero, which means that the neurons have dependence measure must be understood in a high probability of firing simultaneously. a statistical sense throughout a given tempo- In general, this probability, as well as the ral window (Lachaux et al., 2002; Lachaux, eventual delay, is quantified by the cross- Rodriguez, Martinerie, & Varela, 1999;Le correlogram between the spike trains of the Van Quyen et al., 2001). two neurons (Perkel, Gerstein, & Moore, Synchrony at the more regional or 1967). One speaks of oscillatory synchrony, local level has been demonstrated repeat- or synchronous oscillations, if the neurons edly in humans in relation to integrative P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

750 the cambridge handbook of consciousness

mechanisms in language, memory, atten- cal limitations in the estimation of gener- tion, and motor tasks and in virtually all alized measures of synchronous activation. the sensory modalities. For example, the With this point in mind it is possible to say perception of coherent objects in humans that synchrony, as presented here, appears is specifically associated with synchronous as a simple measure of precise temporal oscillations in the gamma range (above 30 relations between neural processes that can Hz), the so-called induced gamma response enable one to follow the formation of spa- (for a review, see Tallon-Baudry & Bertrand, tiotemporal brain patterns relevant for con- 1999). This response, although not com- sciousness. Not surprisingly, this mechanism pletely time-locked to the stimulus presen- is referred to in several of the dynamical tation, typically starts in posterior brain areas models reviewed above. In the next sec- (over the occipital cortex) around 200 ms tion, we review more specifically a set of after the stimulus and then returns grad- results concerning the relation between con- ually to the prestimulus level when the sciousness and the current ‘crowd’s favorite’ stimulus does not require further analysis among neurodynamicists – synchrony in the (Lachaux et al., 2000, 2005; Tallon-Baudry gamma range. & Bertrand, 1999). As we mentioned above, oscillations produced by two neural populations can also The Crowd’s Favorite: be synchronous within larger cell assem- The Gamma Band blies. This synchrony can be detected by a transient phase-locking between the oscilla- Evidence for a Relation Between Gamma tions of the two local fields (Lachaux et al., Synchrony and Consciousness 1999, 2002; Rodriguez et al., 1999). Such long-range synchrony between distant neu- We have mentioned that synchrony among ral populations has been suggested as a plau- oscillating neural populations is a plausible sible candidate to mediate the integration candidate to mediate functional connectiv- of activity in functionally specialized and ity and therefore to allow the formation distinct brain regions (Bressler, 1995; Varela of spatiotemporal structures, such as those et al., 2001). For example, Tallon-Baudry and reviewed in the previous sections. In this sec- colleagues have shown in humans that dur- tion, we return to this hypothesis in more ing the maintenance of a complex shape detail, with a particular focus on gamma in visual short-term memory, two function- band oscillations, which have been repeat- ally distinct regions within the ventral visual edly associated with consciousness in the last pathway, the fusiform gyrus and the lat- 15 years. eral occipital sulcus, produce synchronous The putative role of gamma band oscil- oscillations around 20 Hz (Tallon-Baudry, lations in the formation of conscious experi- Bertrand, & Fischer, 2001). ence was proposed by Crick and Koch (Crick Synchronization is a complex concept & Koch, 1990), shortly after Singer and col- that can cover several possibly distinct types leagues (Gray et al., 1989) had completed a of temporal relations, such as coherence, fre- series of observations in the cat visual cor- quency synchronization, phase synchroniza- tex showing that neurons tend to synchro- tion, generalized synchronization, as well nize their spiking activity when stimulated as others (Brown & Kocarev, 2000; Friston, with parts of the same visual object, such as 1997; Pikovsky et al., 1997). Here we have a moving bar (whereas they do not synchro- focused on synchrony as either occurring nize when stimulated with features that can- between stochastic point-processes, such as not be part of the same object). Those obser- spike trains, or in terms of phase rela- vations matched theoretical predictions by tions between oscillatory processes (phase- Von der Marlsburg (1981) that synchrony locking synchrony). As mentioned above, we could be used to achieve figure/ground seg- chose this focus mainly because of techni- mentation during perception of the visual P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

neurodynamical approaches to consciousness 751

scene. Thus, synchrony was assumed to Several studies have also demonstrated provide a solution to the visual binding prob- that the presentation of sensory stimuli elic- lem (the problem, discussed above, of inte- its stronger gamma synchrony when atten- grating distinct visual features into a unified tion is focused on the stimulus than when and coherent perception; see Roskies, 1999). attention is diverted away. This finding In their 1990 paper, Crick and Koch was observed in monkeys for somatosensory pushed this idea further by stating that stimulations and found again recently for visual consciousness of the object occurs neurons in area V4 of monkeys presented only when its features are bound together as with small visual gratings (Fries, Reynolds, a result of this type of synchronous activity. Rorie, & Desimone, 2001). This hypothesis was in good agreement with There is also evidence for a more direct the feature-integration-theory, proposed by relation between gamma activity and con- Anne Treisman, suggesting that attention is sciousness. Lachaux and colleagues have necessary to bind together the features of recently shown that the perception of faces objects (Treisman & Gelade, 1980).21 Hence- is associated with strong gamma oscillations forth, a close relation between gamma syn- in face-specific regions along the ventral chrony and attention and consciousness has visual stream (Lachaux et al., 2005). Epilep- ensued (Fell et al., 2003; Varela, 1995). This tic patients with intracranial electrodes that association has been very appealing for neu- record directly from the fusiform face area rodynamicists addressing the consciousness (a region along the ventral visual path- issue, because it provides ground material way particularly associated with the percep- for the neural spatiotemporal patterns they tion of faces) were presented with high- associate with consciousness on the appro- contrast ‘Mooney figures’ representing faces. priate time scale. Indeed, several dynamic Because the figures were presented briefly, models among those reviewed above, specif- for 200 ms, they were consciously perceived ically those by Singer, Llinas, Varela, and as faces only half of the time. The authors their respective collaborators, consider syn- reported that the gamma band response to chronous activity in high-frequency ranges, the images was significantly stronger when most preferentially the gamma range, as cru- the figures were actually consciously per- cial for conscious experience. ceived as faces than when they were not. Fortunately, the association between the This high-resolution study followed a pre- gamma band and attention, vigilance, and vious one (Rodriguez et al., 1999), using the consciousness is not just based on its the- same protocol in normal subjects with non- oretical appeal but also on sound exper- invasive scalp EEG recordings; this study imental evidence. For instance, it is well showed that gamma oscillations tend to syn- known that the precise synchronization of chronize across widely separated brain areas neuronal discharges is more prevalent dur- (typically frontal versus occipital) only when ing states characterized by arousal and more- the figures are perceived as faces. over that gamma oscillations are particularly Fries and colleagues (Fries, Roelfsema, prominent during epochs of higher vigilance Engel, Konig, & Singer, 1997) have shown an (Herculano-Houzel, Munk, Neuenschwan- even more direct relation between gamma der, & Singer, 1999; Rodriguez, Kallenbach, synchrony and consciousness. They showed Singer, & Munk, 2004). In cats, for exam- that, during binocular rivalry in cats, the ple, gamma synchrony is stronger after the level of synchrony between visual neurons stimulation of the mesencephalic reticular follows in time the shift of perceptual dom- formation (Munk, Roelfsema, Konig, Engel, inance. Cats were presented with two visual & Singer, 1996). Furthermore, EEG/MEG patterns moving simultaneously in different gamma-band activity is present both during directions: One pattern was presented to the REM sleep and awake states, with a much left eye and the other to the right eye. Under stronger amplitude than during deep sleep such circumstances, the visual percept can- (reviewed in Engel et al., 1999).22 not encompass the two contradictory P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

752 the cambridge handbook of consciousness

patterns and instead alternates between Further hints about the role of gamma them (hence the term ‘binocular rivalry’). synchrony come, albeit indirectly, from The results of this study showed that neu- the experimental contributions of Benjamin rons stimulated by the perceived stimulus Libet (Gomes, 1998; Libet, 2002). In a were strongly synchronized, with strong series of classic experiments in patients gamma oscillations, whereas cells stimulated mixing direct intracranial electric stimula- by the suppressed visual pattern showed tions and peripheral somatosensory stimu- only weak synchrony. This experiment is lations, Libet revealed a number of inter- highly relevant to the study of visual con- esting properties of somatosensory aware- sciousness, because conscious perception ness: (1) An electrical cortical or thalamic is decoupled from the drive of the sensory stimulus requires a duration of more than inputs (the physical stimulus remains con- 250 ms to be felt, whereas a skin stimu- stant while perception does not) and gamma lus of 20 ms is sufficient. (2) If a direct synchrony is used as an indicator of which cortical (electrical) stimulus occurs within pattern is being consciously perceived by 250 ms after a skin stimulus, it can suppress the cat. or enhance the felt perception of the latter Gamma synchrony has been further asso- stimulus. (3) For a skin stimulus to be felt as ciated with consciousness in the context of synchronous with a non-overlapping cortical the attentional blink effect. The attentional stimulus, the skin stimulus must be delayed blink occurs when a subject must detect two about 250 ms relative to the latter stimu- targets in a series of rapidly presented pic- lus. Interestingly, all three properties match tures (at a rate of about 10 per second). Typi- quite closely the known temporal dynamics cally, the second target is detected (and con- of the cortical gamma response induced by sciously perceived) less frequently when it sensory stimuli. This match is particularly comes within 500 ms of the first target, as intriguing considering the fact that Libet if the subject had ‘blinked’. Fell and col- used rhythmic electrical stimulations in the leagues have argued that the blink could be gamma range (typically 60-Hz trains of due to the suppression of gamma synchro- electric pulses). nization shortly after the response to the If the induced gamma response is first target (Fell, Klaver, Elger, & Fernandez, involved in the conscious perception of a 2002). Once again, gamma synchrony would sensory stimulus, then one would indeed be necessary for conscious perception. expect that a rhythmic train of electri- This proposal is consistent with a recent cal stimulations in the gamma range could observation from Lachaux and colleagues, mimic the effect of the induced gamma in the face perception paradigm detailed response, if it possesses the same temporal above (Lachaux et al., 2005), that parts of properties; that is, if it starts roughly 250 ms the primary visual cortex shut down, with after the mimicked stimulus onset and lasts respect to gamma activity, after the presen- for at least 250 ms. Then it should be felt as tation of a Mooney figure: There is a drop of synchronous with a corresponding skin stim- energy in the gamma band, below the base- ulus and possibly interfere with perception line level, which lasts a couple of hundreds of that latter stimulus. In brief, Libet’s obser- of milliseconds, and is simultaneous with vations can readily be interpreted via the the induced gamma increase in the fusiform involvement of the sensory-induced gamma face area. This drop in gamma activity could response in sensory awareness, at least in the be the trace of a transient deactivation of the case of somatosensory awareness. primary visual cortex that could cause the In summary, the previous studies cer- transient attentional blink after a meaning- tainly build a strong case for the role of res- ful visual stimulus. The visual cortex would onant assemblies, oscillating in the gamma be transiently ‘unavailable’ while processing range, as neural correlates of sensory aware- particularly meaningful stimuli, as in a reflex ness. Nevertheless, this assessment is not the protective mode. end of the story, for a number of arguments P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

neurodynamical approaches to consciousness 753

make it difficult to equate gamma synchrony in the emergence and the stabilization of and consciousness. the conscious percept. The presence of such structures is especially the case for short- term memory, which has been proposed Problems Concerning the Link Between as a central component of consciousness Gamma Synchrony and Consciousness (Baars & Franklin, 2003). In visual short- The first problem to mention is that gamma term memory, when an individual has to synchrony can be observed in unconscious maintain a conscious representation of a anesthetized animals, although it is stronger complex visual shape, using mental imagery, when animals are awake (see Sewards & for a couple of seconds, synchrony occurs not Sewards, 2001, for arguments against the in the gamma range, but in the lower beta role of gamma synchrony in conscious- range (between 15 and 20 Hz), between dis- ness). Sewards and Sewards further argue tributed sites of the ventral visual pathway that gamma oscillatory activities have been (Tallon-Baudry et al., 2001). Therefore, res- detected in structures that most likely do onant cell assemblies in the beta range may not participate in the generation of sen- also subserve continuous visual perception sory awareness, such as the hippocampal (if only in its imagery aspect). formation: “Obviously hippocampal activi- The above studies emphasize the point ties could not contribute to sensory aware- that gamma synchrony may be necessary for ness since lesions to that structure do not the emergence of a conscious perception, result in purely sensory deficits of any kind” but perhaps only in this emergence. Once (Sewards & Sewards, 2001,p.492). This formed, the percept could then continue via argument, as well as others, leads them to other cell mechanisms, in the form of other conclude that “while synchronization and types of spatiotemporal structures. oscillatory patterning may be necessary con- Nevertheless, even at the initial level of ditions for activities to participate in gener- this emergence, the role of gamma syn- ating awareness, they are certainly not suffi- chrony needs to be clarified. As we have cient” (Sewards & Sewards, 2001,p.492).23 seen, gamma synchrony occurs in anes- This point echoes the conclusions from a thetized animals and is therefore not suffi- study by Revonsuo and colleauges (Revon- cient for consciousness (Sewards & Sewards, suo, Wilenius-Emet, Kuusela, & Lehto, 2001). One interesting possibility, in the 1997). In this study, they recorded the case of the visual system, is that gamma gamma band response of normal subjects synchrony could be involved in the forma- during the fusion of random-dot stere- tion of visual objects. Visual objects are ograms. They observed that, although 40-Hz the preferred targets of visual attention, synchronized oscillations seemed to partici- and yet they present themselves to us only pate in the construction of the unified per- via conscious perception. Furthermore, as cept, they were not maintained during the argued by the Feature Integration Theory continuous viewing (and conscious percep- (Treisman & Gelade, 1980), visual objects tion) of the same stimulus once it had seem to require visual attention to form. been constructed. Lachaux (unpublished The question thus arises of which comes findings) repeatedly confirmed this obser- first: objects or attention. One solution to vation with human intracranial recordings: this problem is that in the absence of atten- The gamma response induced by durable tion there are only ‘pre-objects’; that is, visual stimuli in the visual system often bundles of features that are object candi- stops before the end of the stimulus pre- dates and that are sufficient to attract atten- sentation, despite the fact that the sub- tion, which would then finish the construc- jects still fixate the images and consciously tion and remain grabbed by them (Wolfe & perceive them.24 Bennett, 1997). These considerations indicate that other Engel and Singer (Engel et al., 1999) pro- spatiotemporal structures may participate pose that gamma synchrony may mediate P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

754 the cambridge handbook of consciousness

this mechanism. According to this proposal, view is, after all, one of the main points proto-objects, based on their physical fea- of agreement among the different neurody- tures and Gestalt properties, assemble in the namical proposals we reviewed earlier. In form of nascent cell assemblies via gamma this section, we take a more critical stance synchrony. This synchrony corresponds to regarding this central issue and put forth the kind of ‘automatic’ synchrony observed some qualifications we believe are important in anesthetized animals. This nascent syn- to keep in mind. chrony is reinforced in awake animals, such In several of the neurodynamical theo- that there is a formation of the visual object. ries we have discussed, the notion of a dis- This process corresponds to the grabbing of tributed neuronal assembly, understood as attention by the object and is simultaneous some kind of synchronous pattern of acti- with the object’s actual formation for per- vation, is central to explaining the neuronal ception. In this model, attention and gamma basis of consciousness. As we saw in the pre- synchrony become two sides of the same ceding section, the gamma band has been a coin, as long as one is ready to extend the preferred region of the frequency domain, concept of attention (usually associated with in which such assemblies have been studied. conscious perception) to a general selec- Whether restricted to this frequency band or tion mechanism that includes an uncon- spanning multiple frequencies, an emergent scious preselection mechanism. This pres- and stabilized spatiotemporal pattern is seen election mechanism is the one observed in as a prerequisite for conscious experience to anesthetized animals. Attention, in its clas- happen. sic ‘conscious’ sense, is thus envisioned as the This viewpoint, however, raises at least tip of the selection iceberg. two related questions. On the one hand, if Can we therefore relate the full formation such patterns are necessary for conscious- of resonant gamma assemblies to the emer- ness, and if we can distinguish them as gence of consciousness? The answer would having a certain spatiotemporal unity, what seem to be yes, in a certain sense; namely, happens between patterns? Are we con- that the content that is correlated with the scious during such transitions? Or is con- formation of the resonant gamma assem- sciousness a sequence of snapshots, in which bly is accessible to verbal report, working the apparently seamless fusion of succes- memory, and so on. On this view, gamma sive moments into the ongoing flow of synchrony is necessary for any kind of sen- experience is achieved by some additional sory awareness. This view gains support from mechanism? Engel and Singer’s observation that syn- On the other hand, can we define a stable chrony is related to all of the four presumed conscious moment within the flow, and are component processes of awareness: arousal, we therefore entitled to suppose that during segmentation, selection, and working mem- such a moment, the assembly will ‘hold’ or ory (Engel et al., 1999). In the following, ‘contain’ a certain unity, even though during however, we examine certain problems with that moment one can distinguish a change this idea that lead us to qualify it. (or changes) in one’s experience? Recall that dynamic assemblies are supposed to last for several hundreds of milliseconds, but our Consciousness and Dynamical sensory experience can change within that duration. Suppose, for example, you are sit- Structures: Some Qualifications ting in a train, staring out of the window, and as you look out into the countryside, trees, Introduction electricity poles, and other objects swiftly Throughout this chapter we have explored cross your visual field, without your being the view that consciousness seems to require able to grasp them fully and stably. Yet you the formation of distinct, dynamic spa- know they are trees, electricity poles, and tiotemporal structures in the brain. This other objects. Does your rapid experience P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

neurodynamical approaches to consciousness 755

of each of these objects correspond to a dis- and what they mean in terms of the expe- tinct assembly? Or is it rather a matter of riencing subject is addressed less frequently. one global assembly, in which various local With regard to this issue, the proposals assemblies ‘ride’? In several neurodynamical of Varela and Kelso are the most explicit proposals, as we have seen, an experience of and developed.26 These authors stress the an object is supposed to depend on the for- metastable nature of such patterns, so that mation of distinct, coherent brain patterns. successive moments of distributed neural But a conscious moment can include full- coherence combine in a continuous and fledged objects as well as less definite visual ongoing fashion, in contrast to a sequence of patterns that, although conscious to a cer- clear-cut states.27 These approaches present tain extent, cannot be completely described attractive alternatives that seem to fit nicely as stable entities. with James’s intuitions. They also allow for These two interrelated features – ongoing a different interpretation of what counts as flow and fleeting experiences – need to be a meaningful dynamic pattern. Rather than addressed by any neurodynamical approach seeing these patterns as individual assem- to consciousness. In the remainder of this blies that arise, maintain themselves for a section, we discuss both features and pro- brief period, and then subside, they can be pose a simple distinction that may help clar- viewed as one itinerant trajectory, and thus ify the issues at hand. as one pattern (Friston, 1997, 2000; Varela, 1999) in which the rate of change is the only internal definition of the stability of a Ongoing Flow and Fleeting Experiences given moment. In any case, neurodynami- The issue of the ongoing, fluid nature of cal approaches must deal explicitly with this conscious experience is certainly not new.25 issue of the apparent unity of the flow of William James, in his famous chapter on consciousness,28 as opposed to the unity of “The Stream of Thought” (James, 1890/1891, moment-to-moment experience. Chapter IX), provides a detailed description The second question to which we wish of the structure of this flow. He distinguishes to draw attention is related to the stabil- at least two fundamental aspects – ‘substan- ity of actual perceived objects during a con- tive’ stable moments, in which one is actu- scious moment. As we mentioned above, the ally conscious of something, and ‘transitive’ notion of an assembly implicitly incorpo- fleeting moments, in which one passes from rates a notion of stability during the lifespan one content to another. He describes con- of the pattern in question. Our sensory envi- sciousness as like a bird’s life, for it seems to ronment, however, can be subject to rapid be made up of an alternation of flights and change in time windows lasting less than sev- perchings. James remarks that substantive eral hundred milliseconds, and yet we are, moments can be recognized as such, whereas to a certain extent, aware of the change as transitive moments are quite difficult to pin- taking place. This fact would seem to pose point accurately. They present themselves as a difficulty for any theory that postulates tendencies and changes between states, and a neural assembly, organized on a slower not as distinct contents immediately defin- time scale, as necessary for conscious expe- able in themselves, save by some retrospec- rience. On the other hand, not every object tive exercise. of the visual scene is perceived as stably as How do these phenomenological obser- one might naively think. This fact is espe- vations relate to the neurodynamical picture cially clear in inattentional blindness experi- of the brain and its relation to consciousness? ments (Simons, 2000). In such experiments, As we have seen, most neurodynamical subjects are asked to focus on a particular proposals stress that each conscious state task and set of stimuli in a visual scene. If depends on a specific neural assembly or an additional stimulus appears unexpectedly emerging dynamic pattern, but the issue of in that scene, the subjects are often unable how transitions between states take place to report it afterward. What is particularly P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

756 the cambridge handbook of consciousness

striking with such ‘inattentional blindness’ accordingly. In a certain sense, this dual- is that it can happen even for very distinc- ity mirrors the access/phenomenal distinctive and salient objects. In one famous exam- tion, but without assuming that there can ple (described in Simons, 2000), subjects be fleeting phenomenally conscious expe- watch people passing basketballs. Three peo- riences that are inaccessible in principle ple wearing white T-shirts pass a ball to to verbal report. In endorsing the need to each other, while three other people wearing make this stable/fleeting distinction, we also black T-shirts pass another ball to each other. stress the need to consider the possibility The subjects have to count the number of of the more ephemeral aspects of experi- passes between the white players, which ence as being accessible to verbal report, occur at a fast enough rate to require the if approached with the appropriate first- full attention of the viewer. After 45 softhe person and second-person phenomenologi- display, a man in a gorilla suit walks across cal methods (Petitmengin, in press; Varela & the scene, stops for a moment in between Shear, 1999). the players, waves his hands in the air, and Given this structural distinction between then exits through the other side 5 s later. stable and fleeting aspects of experience, it It is well documented that a high portion of would be interesting to see how a neuro- the viewers fail to report seeing this gorilla. dynamical theory that relates the formation In models like the one advocated by of well-defined spatiotemporal patterns in Singer and collaborators (see above) there brain activity to conscious experience would is a strong correspondence between a fig- deal with the intrinsic mobility of any given ure/ground distinction (and therefore an perceptual act. For example, the feedfor- object) and the formation of a synchronous ward stream (or sweep, FFS) is defined as the assembly. This correspondence would seem earliest activation of cells in successive areas to imply that only fully formed assemblies of the cortical hierarchy. In the visual modal- can ‘support’ some type of perceptual recog- ity, it starts with the retina, the LGN, V1, nition of the object in question. As dis- and then the extrastriate visual areas and the cussed above, however, both phenomeno- parietal and temporal cortex. Thorpe and logical observation of one’s own experi- colleagues (Thorpe, Fize, & Marlot, 1996) ence and experiments such as the unnoticed have shown that the FFS is sufficient to gorilla suggest that a great deal of experience carry out complex visual processing, such as may be unstable and fleeting. Where would detecting whether a natural scene presented such fleeting experiences of quasi-objects for 20 ms contains an animal. It is tempting fall in the framework of dynamic neural to relate the more stable aspect of experi- assemblies? Lamme (2003, 2004) has pro- ence to the formation of spatiotemporal pat- posed that such fleeting experiences belong terns, in the sense of dynamic neural assem- to ‘phenomenal consciousness’ (i.e., are sub- blies mediated by recurrent neural interac- jectively experienced, but not necessarily tions, whereas the fleeting, unstable aware- accessible to verbal report), whereas more ness could be embodied through the rapid stable experiences belong also to ‘access con- FFS that modulates and continuously affects sciousness’ (i.e., are available to verbal report the formation of such assemblies while not and rational action guidance; see Block, being fully excluded from a certain level 1997, 2001, for this distinction between of perceptual experience. This proposal is phenomenal consciousness and access con- highly speculative, but is intended simply sciousness).29 Neurodynamical models need as a way to highlight the necessity of deal- to be able to account for this evanescent ing with the stable/fleeting structure that aspect of conscious experience in a more appears to be inherent in each and every con- explicit way. scious moment. More precisely, we propose that the sta- To conclude this section on qualifica- ble/fleeting duality be considered a struc- tions to the dynamic approach, we would tural feature of consciousness experience like briefly to draw the reader’s attention (see also the next section) and dealt with to another aspect of consciousness that is P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

neurodynamical approaches to consciousness 757

significant in light of the preceding discus- sciousness. This crucial feature is often put sion and the overall topic of this chapter. to the side as something to deal with once This aspect is the subjectivity or subjec- the issue of the neural correlates of percep- tive character of consciousness. For exam- tual consciousness has been resolved (e.g., ple, Damasio (1999) has stressed that, in Crick & Koch, 2003). Our view, however, is addition to understanding the neurobiolog- that unless the subjectivity of consciousness ical basis for the stream of object-directed is adequately confronted and its biological conscious experiences, it is also necessary to basis understood, proposals about the neural understand the neurobiological basis for “the correlates of perceptual consciousness will sense of self in the act of knowing” (Parvizi & provide limited insight into consciousness Damasio, 2001; see also Panksepp, 1998, for overall. Thus, the issue of subjectivity is a a convergent argument, and Wicker et al., non-trival matter that any neurodynamical 2003). The sense of self with which Dama- approach must confront sooner or later if it is sio is concerned is a primitive kind of con- to become a cogent theory of consciousness. scious self-awareness that does not depend We briefly pick up this thread in the when on reflection, introspection, or possession of discussing how to relate phenomenological the concept of a self. In phenomenologi- descriptions to neurodynamical accounts. cal terms, it corresponds to the fundamental ‘ipseity’ (I-ness or selfhood, by contrast with otherness or alterity) belonging to subjective The Future: Beyond Correlation? experience (see Chapters 4 and 19). In a related line of argument, Searle Introduction (2000) has suggested that a major draw- So far we have dealt primarily with the issue back of current attempts to uncover the of meaningful spatiotemporal patterns in the neural correlates of consciousness in human brain and their relevance to the study of beings is that they begin with already con- conscious experience. It may have become scious subjects. He advocates a ‘field of increasingly evident to the reader, however, 30 consciousness’ viewpoint, in which the that the issue of how to relate such pat- perceptual experience of an object arises as terns to experience as a first-person phe- a modification of a pre-existing conscious nomenon has been left untouched. Indeed, ‘ground-state’ that is unified, subjective, and one of the major challenges facing the cogni- qualitative. In this context, the transition tive sciences is precisely how to relate these between conscious states need not be punc- two domains – the domain of third-person tuated by a radical gap in consciousness, but biobehavioral processes and the domain of can rather be a modulation of a more basic first-person subjective experience. What is state of background consciousness, which the right way to conceptualize this rela- accounts for the fact that even such transition, and what is the best way to approach tive moments are felt as belonging to one- it methodologically? These questions have self. Here dynamic patterns in the form not yet received anything near a satisfactory of transient and distributed coactive assem- answer from the neuroscientific community. blies would mainly reflect the nervous sys- We do not intend to propose an answer tem’s own homeodynamic activity; that is, to them here. Rather, we wish to highlight its maintenance of a range of internal regular- some conceptual and practical issues in the ities in the face of its ongoing compensation quest to understand the relation between for the systematic perturbations to which it these two domains while keeping in mind is exposed from both the sensory environ- the dynamical insights we have gained from ment and the internal bodily milieu (Dama- the previous exposition. sio, 1999; Maturana & Varela, 1980). Nevertheless, it remains difficult to see Correlation and Emergence how metastable assemblies of coactive neurons could by themselves account for this The first question that comes to mind is the crucial aspect of the subjectivity of con- extent to which the entire neurodynamical P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

758 the cambridge handbook of consciousness

approach rests on a merely correlational the notion of emergence or emergent phenom- strategy. In coarse terms, one isolates a given ena and thus can be considered as attempts target experience, say the perception of a to go beyond a purely correlational descrip- figure; one determines the neural patterns tion. Although ‘emergence’ is a complex that correlate with the moment the subject concept subject to multiple interpretations sees the figure; and one then concludes that (see Keslo, 1995; Thompson, 2007; Thomp- the conscious experience depends on such son & Varela, 2001), in simple terms it can neural patterns.31 In the last decade or so be defined as follows: A process is emergent this correlational approach, in the form of when (i) it belongs to an ensemble or net- the search for the neural correlates of con- work of elements, (ii) it does not belong sciousness, has undergone important devel- to any single element, and (iii) it happens opments and become more sophisticated spontaneously given both the way the ele- with regard to its conceptual formulation, ments interact locally and the way those methodological commitments, and empiri- interactions are globally constrained and reg- cal results (Block, 1996; Crick & Koch, 1990, ulated. Thus, an emergent process cannot be 1998; Rees, Kreiman, & Koch, 2002). Here understood at the level of local components the central idea is that, rather than formu- taken individually, but depends rather on the lating explanatory principles about the rela- relations established between them. Further- tion between neural activity and experience, more, an emergent process not only depends what has to be done first is to determine on the local components but also constrains those neural processes that can count as a their degrees of freedom, a two-way pro- “specific system in the brain whose activity cess that has been termed ‘circular caual- correlates directly with states of conscious- ity’ (Haken, 1983). Especially in Freeman’s ness” (according to the Association for the and Varela’s approaches, conscious experi- Scientific Study of Consciousness, cited by ence is considered to be an emergent process. Chalmers, 2000,pp.17–18). Once such pro- The difference between their views is that cesses have been found, then one can turn whereas Freeman (1999a,b) proposes that to the issue of how they are causally related consciousness is a global brain state, Varela to experience itself.32 proposes that consciousness may encompass Neurodynamics as a research program multiple cycles of organismic regulation that is devoted, at least methodologically, to are not fully restricted to the brain (Thomp- this correlational strategy and in this sense son & Varela, 2001). Neverthless, although remains closely linked to the NCC program. principles of emergence have been clearly Of course, this commitment is due to the formulated at the level of physical processes fact that, in the scientific tradition, establish- and molecular interactions (Nicolis & Pri- ing a relation between two target events or gogine, 1989), in the case of conscious expe- phenomena is mainly approached by estab- rience, such principles still need to be under- lishing a correlation in their occurrence. stood and formulated in a more rigorous Causal relations can then be assessed on the way. We believe that the study of complex basis of altering one of the target events systems offers a promising approach in this and observing whether and how the other direction (Le Van Quyen, 2003; Thompson, changes. This ‘interventionist’ strategy can 2007). be employed in the case of brain functioning and consciousness by using microstimula- Gaining Access to Experience tion during surgery or transcranial magentic stimulation (TMS). Nevertheless, by itself As mentioned earlier, any neurodynamical this strategy does not guarantee the eluci- approach to consciousness must eventually dation of the underlying causal mechanisms. deal with the issue of how to describe Several of the proposals reviewed above, experience itself. In the previous sections however, formulate explicit links between of this chapter we have discussed mainly the neural and the experiential in terms of spatiotemporal brain patterns in relation to P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

neurodynamical approaches to consciousness 759

consciousness, but we now turn to consider ments is to dissociate what is presented to the other side of the issue; namely, how to the subject from what the subject sees in gain access to experience itself and render it order to distinguish the neural patterns that accessible to scientific description. are specific for conscious perception. Among On the more operational side of this ques- these approaches, three stand out as the tion, one can ask how it is possible to set most well studied and influential. The first is up an experimental paradigm that addresses visual masking (Dehaene & Naccache, 2001; the issue of gaining access to experience in Hollender, 1986), in which short-lived visual a way that allows us to study the underly- stimuli flanked by meaningless masks are not ing neuronal processes. Not all that is going perceived consciously, yet can alter future on in the brain is necessarily related to what behavior (an index of non-conscious pro- the subject is consciously experiencing. It is cessing). The second is inattentional blind- known that, during our conscious engage- ness and change blindness (O’Regan & Noe,¨ ment with the world, a non-negligible part 2001; Simons, 2000), in which diverting the of our adapted behavior depends on non- subject’s attention can render major changes conscious processes that are carried on within the scene unnoticed. The third is binocu- out us being aware of their functioning. For lar rivalry (Blake, 1989; Blake & Logothetis, example, do you have any feeling what- 2002), in which the presentation to each eye soever of the oxygen level in your blood of a different image induces an alternation in right now? Yet this bodily state of affairs conscious perception between the two alter- can be crucial to your capacity to be here natives, despite the fact that both are always right now reading this text. Although this present. example is extreme, carefully crafted exper- This last experimental paradigm is par- iments reveal that even perceptual informa- ticularly relevant to the issue of gaining tion can be used to guide behavior in a non- access to experience because it provides conscious way. For example, when a subject an ongoing, slow phenomenon that can be is presented with a small circle surrounded described by the subject. In virtue of its by larger circles, the small circle appears alternating character, the experience lends smaller than if it is presented in isolation. itself to repetitive scrutiny, in order to bet- Yet if the subject is asked to reach for it, ter characterize ‘what it is like’ subjectively his fingers adopt a grip size that is consis- to undergo it. Finally, because both stim- tent with the true size of the circle and uli do not change, yet perception changes not with its illusory dimension (Milner & dramatically, binocular rivalry evidences the Goodale, 1995). Another classic example is endogenous and ongoing character of expe- known as blindsight (Danckert & Goodale, rience and therefore calls for attending to 2000; Weiskrantz, 1990). In this neurologi- those neural processes that share this funda- cal condition, conscious visual experience is mentally dynamical structure.33 impaired due to damage in primary visual These considerations suggest that, in cortex, yet subjects can produce quite accu- addition to using experimental paradigms rate motor actions, such as introducing an for dissociating unconscious and conscious envelope through a horizontal slot or point- processes, we need to be able to capture ing to a target they claim not to see. Thus, the the dynamics of experience itself. Hence it problem arises of how to determine those is necessary for the experimenter to take neural processes that show some kind of measurements of each phenomenon – the direct relation to the actual conscious expe- dynamics of the brain and the dynamics of rience of the subject, in contrast to those that experience. Measurements should provide sustain ongoing and non-conscious adaptive public data; that is, information that can be behavior in the world. shared with another observer. One recurrent Several experimental approaches have problem with consciousness is that the direct become paradigmatic in this endeavor. In observation of experience is accessible only general, the rationale behind these experi- to the subject, and such observation is not P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

760 the cambridge handbook of consciousness

a public measurement in itself. The experi- bility requires distance. Stability means that ence therefore needs to be transcribed into the phenomenon remains somewhat con- public data in a subsequent step to provide stant during a certain time interval; it fur- so-called first-person data. What the status ther means that the distance between con- of first-person data is and to what extent the secutive measures is shorter now than what subject can play an active role in describ- it was in earlier observation windows. The ing his or her experience are matters of notion of distance is also central to the con- active debate in the science of consciousness cept of recurrence: If we find a certain neu- (Jack & Roepstorff, 2003; Varela & Shear, ral pattern that correlates with a conscious 1999). We cannot review these debates here. experience, we expect this neural pattern to Rather, in the remainder of this section, we repeat when the same experience repeats. wish to explore two complementary lines of Because neither neural patterns nor experi- investigation that are relevant to the issue ences repeat in a perfectly reproducible way, of making experience more scientifically we also need a way to know whether a cer- accessible. tain neural pattern or experience looks like one that occurred in the past. This requires a quantification of resemblance between two A ‘Topological Approach’ measures; that is, a distance. to First-Person Data Note that this first definition is large In a general sense, one expects measures to enough to include many possible measures. be somehow organized in a universe, the In fact, a dance could be considered as a mea- measurement universe, that is the set of all sure or a series of measures if each succes- the possible ‘values’ that measure can take. sive body configuration constitutes by itself a The term ‘universe’ must be understood in measure. A drawing could also be a measure. its statistical sense and simply refers to the But to be actually useful, we insist that the set of possible values, states, or items that subject and the experimenter should agree can be valid measurements. For instance, a on a measure of distance, which enables single word is one particular item among all anybody to evaluate the degree of similar- the possible words. The universe may be dis- ity between two measurements. The Basic crete (as for words or sentences) or continu- Requirement (so called in the following) is ous (as for magnetic fields). In any case, a that the distance should be consistent with measurement will be the selection of one the experience of the subject (as only the particular value allowed in a given universe, subject can tell): If measure A is closer to based on the present state of the observed measure B than to measure C, then the ele- phenomenon. For a given subjective experiments of experience that led the subject to ence, this may correspond to the selection of select measure A should appear to him as one description, among all the possible writ- closer to the elements that led him to choose ten descriptions that can be produced (say) measure B than to the elements associated in a couple of minutes.34 with measure C.35 This requirement directly We mentioned that the measurement implies, for instance, that recurrences in the should be ‘organized’. This means that subject’s experience should translate into it should be provided with some sort of recurrences in the measure. topology: It should be possible to estimate Once provided with measures of (some a distance between two measures. Indeed, it elements of) the subjective experience and should be possible to say whether measure A with measures of neural phenomena, it is closer to measure B than it is to measure C should be possible to establish a relationship (see Fell 2004 for a convergent discussion). between the two phenomena by compar- Without any kind of topology, it would be ing the dynamics of those measures: Related difficult to compare the dynamics of the two phenomena should provide sets of mea- phenomena. For instance, the notion of sta- sures with compatible dynamics. That is, P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

neurodynamical approaches to consciousness 761

once again, stability in experience should be ditions are met, at any given point of the associated with stability (or stationarity) in visual image, only one of the images (or part the neural dynamics, whereas moments of of it) will be seen (will dominate) in an change should be correlated with changing exclusive way. In general, binocular rivalry is (or non-stationary) neural processes.36 considered a clear-cut alternation between two states, and average measures of the brain state during one or the other dominance A ‘Structural Invariants’ Approach period are contrasted. Most commonly, the to First-Person Data subject’s indication via a button press of As a complement to the fine-grained topo- the moment when the alternation takes logical description presented above, it seems place is used to fix a rigid temporal reference possible to adopt what can be termed a around which the average brain responses ‘structural invariant’ strategy. Here the main are defined. aim is to obtain, through descriptions of the We recently used this experimental pro- target experience, an account of that which tocol to investigate the underlying neural is invariant (or stable) as a feature of the patterns, but with the specific objective experience, regardless of whether it is one of describing their spatiotemporal evolu- or another subject that undergoes it. The tion throughout extended periods and with- roots of this approach go back to the method out presupposing a rigid two-state structure adopted in phenomenological philosophy (Cosmelli et al., 2004). To do so, we worked (see Chapter 4). Here, through several rep- with a group of subjects who were exten- etitions of the same experience in different sively exposed to the experience and pro- contexts, one can arrive first at a certain sub- duced free, ongoing descriptions of what jective invariant, and then, through contrast they were seeing and how they were expe- with other subjects, intersubjective invari- riencing it. As conflicting stimuli we used ants that are present in the original expe- a human face and a moving pattern with rience, no matter how many versions of it an intrinsic frequency (a frequency tag; see one tries and no matter how many different Brown & Norcia, 1997; Tononi & Edelman, subjects engage in it. A traditional example 1998). This intrinsic frequency was incor- is the structure of the visual field, in which porated in order to tag a neural evoked what one sees focally always appears as a response that could be followed by magne- relatively detailed center surrounded by an toencephalography (MEG). increasingly less detailed region, which, at The descriptions produced by the sub- the limit, fades into an ungraspable indeter- jects showed some interesting features: In minacy. In the particular context of the neu- addition to experiencing the well-known rodynamics of consciousness, the relevance alternation between both images, the sub- of this type of approach can be illustrated jects repeatedly described this alternation as by recent work on the experience of binoc- extremely variable in the way it occurred. ular rivalry (Cosmelli et al., 2004). Although sometimes the alternation from As we briefly described above, binocu- one image to the other started in the cen- lar rivalry occurs whenever one is presented ter of the field and progressed toward the with dissimilar images, one to each eye. The outer limits, in other occasions it began on subjective experience is that of an ongoing one side, from the top or the bottom, or even alternation between both possible images, from the external borders, and then progres- with only one of them consciously perceived sively invaded the pre-existing image. Most at a time. If the images are large, then dur- subjects claimed that it was difficult to give ing the transition from one to the other, one a stable description of how these transitions can distinguish a mosaic, patchwork pattern took place, because at each time they devel- composed of both images, but as a rule, if oped in a different way. Nevertheless, all the adequate contrast and luminance con- subjects invariantly stated that dominance P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

762 the cambridge handbook of consciousness

periods would alternate and recur, no mat- are more visual, or more auditory, for ins- ter what the subjects did or how much they tance, and eventually segment it along such tried to prevent it from happening. dimensions; and (iv) recurrent, in the sense At a first coarse level, these descriptions that we recognize objects, feelings, thoughts, already provide us with some crucial aspects memories, etc., as seen or felt before, even of the experience of rivalry: This experience though they are never experienced in the is one of an ongoing flow of recurrent domi- same way. These properties, although cer- nant periods, in which alternations are ex- tainly not exhaustive of our conscious lives, tremely variable in the way they develop. do suggest that methods that allow for pro- This feature is indeed a hallmark of binocular cesses of compatible dynamics should be rivalry that will be experienced by any nor- preferred if we want to advance in our under- mal observer and is thus a structural invari- standing of the neural underpinnings of con- ant in the sense described above. Although sciousness. this descriptive feature is not particularly In addition to this methodological con- novel, it nevertheless points toward a constraint, however, the structural invariants crete restriction in the methods we need approach can potentially make a further to choose to analyze the underlying neural contribution. As we mentioned above, one processes (and consequently what we under- of the most prominent structural invari- stand as the neural underpinnings of con- ants of consciousness is precisely its subjec- sciousness). If we wish to reveal neural pat- tive character, in the sense of its fundamen- terns that are meaningful in the context tal prereflective and preconceptual ‘ipsiety’ of this specific experience, then we cannot (see Zahavi, 2005, for an extended discus- impose a rigid temporal grid and suppose sion). This backdrop of consciousness per- that there is such a thing as an average transi- vades the occurrence of specific states of per- tion from one image to the other. This point, ceptual consciousness. It would appear to however, is rarely acknowledged. We there- call for an explanation not so much in terms fore developed a statistical framework that of the dynamic behavior of the system (e.g., considered significant any neural activity only in terms of the dynamical properties that is recurrent in time, without any restric- of the nervous system’s patterns of activ- tions on the temporal pattern of activation. ity), but rather in terms of how a certain The result was an original description of a self-referring perspective can emerge from a cer- network of distributed cortical regions that tain dynamical organization (Rudrauf et al., showed synchronous activation modulated 2003; Thompson, 2007). Whether this type in concert with conscious dominance peri- of account is beyond the domain of neurody- ods. Moreover, the dynamics of modulation namics as we have defined it here is an empir- of these brain patterns showed a striking sim- ical issue. The crucial point is that if, through ilarity to the bell-type pattern that William some enriched neurodynamical plus organis- James had predicted (more than a century mic plus biological approach (e.g., Damasio, ago) would underlie the occurrence of 1999; Varela, 1979), one could account for any given conscious moment (James, 1890/ the conditions of possibility of a minimally 1981). subjective system, then transcending a purely An important contribution of the struc- correlational strategy would become a real tural invariant approach is thus that it can possibility. serve as an effective constraint on how we study the dynamic brain patterns. Basic Can We Avoid the Pitfalls phenomenological observation shows that of Introspectionism? experience (or the stream of consciousness) is at least (i) dynamic and ongoing; One recurrent question, when discussing (ii) continuous;37 (iii) able to be parsed, the use of first-person data, is how so one can distinguish in a given subjec- to avoid the pitfalls of introspectionism. tive experience components or aspects that Introspectionism was an attempt to use P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

neurodynamical approaches to consciousness 763

introspection as a scientific method to elab- other extreme, first-person data about the orate psychological theories. It was the main precise dynamic of subtle variations of emo- scientific approach to mental phenomena at tions would probably be considered less reli- the beginning of psychology, but was later able (this means that they would not meet dismissed by the scientific community in the Basic Requirement – the same subtle favor of behaviorism (reviewed by Vermer- variations would not lead to the same first- sch, in Depraz, Varela, & Vermersch, 2003). person data, if repeated).38 So, in fact, the The main problem with introspection, as real question concerning first-person data used at that time, was that it provided con- is, Where shall we draw the line between flicting theories. The root of the problem what is acceptable, perfectly good data, and was in fact methodological: It was never what is not?39 A related question of equal possible to ascertain whether the introspec- importance is whether this line is the same tive reports met the Basic Requirement for all individuals and whether it is fixed mentioned above, and there were serious within a single individual or whether train- doubts about the correspondence between ing can move the line (see Chapter 19).40 the descriptions of the experiences and the We believe that this question should become experiences themselves. On the other hand, central in cognitive neuroscience in the near there was little explicit description of the future, especially in view of the advent introspective method by which to proceed of new fields, such as the neuroscience of to explore and describe experience, and emotions or the neuroscience of conscious- hence the actual testing and refinement of ness itself. Such emergent fields rely heav- the research method, as opposed to the con- ily on trustworthy measures of subjective tent of its descriptions, remained underde- experience. veloped (Varela, 1996). Consequently, an important part of the cognitive science community is generally reluctant to use Conclusion first-person data. Is it therefore possible to build a neurodynamics of conscious- In summary, the neurodynamics of con- ness, given that it must rely on first-person sciousness is an attempt to relate two data? dynamical phenomena that take place in a Our position is that the whole issue is subject – the formation of metastable pat- a technical one: If the measure providing terns in the subject’s neural activity and the first-person data meets the Basic Require- transient emergence of discernible elements ment described above, then the measure is or aspects of his or her conscious experience. useful. Alternatively, in the structural invari- To establish such a relation, cognitive neu- ant approach, if a given invariant is stable roscientists need to observe systematic sim- across all subjects for a given experimental ilarities between the dynamical properties paradigm, it should be considered valid. In of these two phenomena. In this sense, the fact, the real question is not whether cogni- neurodynamical approach works at the level tive scientists should ‘trust the subject’, but of correlations, albeit refined ones. On the in which conditions they can trust the sub- experiential side, this approach requires the ject and what they should ask. First-person subject to provide first-person descriptions data, defined as measures of the subjective that can serve as ‘public’ measures of expe- experience, are continuously being used in rience, with at least two objectives. The first psychophysics: When a subject presses a objective is to capture reliably the degree button to indicate that he saw a blue square, of similarity (or disparity) between differ- and not a red circle, he provides a mea- ent subjective phenomena and produce tim- sure of his immediate perceptual experience ings that can be compared to the timing of in its simplest form. In this extremely sim- neural measurements. The second objective ple form, first-person reports are considered is to produce descriptions of the structural as perfectly valid and trustworthy. At the invariants of the experience in question, in P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

764 the cambridge handbook of consciousness

order to constrain the methods that are cho- brane potentials; for instance, via long-term sen to determine which neural activity is to potentiation (LTP) mechanisms. Models that be considered significant. It is not yet clear include a changing connectivity can quickly how much of the complexity of conscious- become unmanageable, however, both math- ness can be revealed in this way, and this ematically and computationally (but see Ito 2002 question constitutes an important field of & Kaneko, ). 5 investigation for the future. . For further details on this subject, we strongly recommend one of the original and most influential sources in neurodynamics by Walter Freeman (Freeman, 1975). See also Notes Bressler & Kelso, 2001. 6. For this reason, it would not necessarily be 1. We define in more detail the notions of func- meaningful to repeat the same perturbation tional segregation and cooperative interac- over and over to study the average reaction tion in the section, “The Search for Mean- of a chaotic system, for there is no guaran- ingful Spatiotemporal Patterns in the Brain.” tee that this average reaction would have any Here we only say that they can be considered meaning. Yet, such averaging procedures are analogous to local specialization and collec- the basis of almost all the imaging studies of tive interaction, respectively. the nervous system. 2. This situation is, of course, only suggestive. 7. Note that the exact reaction of the system – A dream state might be a more rigorous case that is, the precise trajectory that the system of a true sensory filter (Llinas & Pare, 1991). will follow to converge on the attractor – can be very different from one olfactory stim- 3. One might rightly consider other variables, ulation to another, even though the target such as the local concentrations of certain attractor is the same for all. Therefore, the neurotransmitters. Quantitative measures of existence of attractors is compatible with the the glial system should probably also be intrinsic variability of chaotic systems. included. In fact, there is no one single way of choosing which variables to include in 8. A recent review notes that “incontrovert- the system. This choice is largely driven ible proof that EEG reflects any simple by our current knowledge of the nervous chaotic process is generally lacking. There are system, which unfortunately remains quite grounds for reservation concerning reports of limited. As a starting point, one needs to the dimensionality of EEG from direct mea- keep the following three elements in mind surement. Fundamental difficulties lie in the when choosing the variables of the system: applicability of estimation algorithms to EEG (1) The time scale: If one candidate vari- data because of limitation in the size of data able maintains a constant value during the sets, noise contamination, and lack of signal time of observation of the system, then it stationarity” (Wright & Liley, 1996). does not need to be counted as a variable, 9. In human electrophysiology, for instance, the but rather can be considered as a parame- dominant paradigm is recording the EEG of ter (see below). (2) The spatial scale: The human subjects while presenting them with nervous system can be modeled at several series of similar sensory stimulations. The sig- scales – molecules, neurons, neural popula- nal studied is the evoked potential: the mean tions, etc. The variables should be meaningful EEG response averaged over all the stim- at the spatial level of investigation. (3) The ulations. The intertrial variability is consid- interdependence within the system: If the ered as noise and disappears in the averaging value of one candidate variable is fully deter- procedure. mined by the values of the other variables, 10. As Le Van Quyen (2003,p.69) notes, “In then it does not need to be included in the physics, what is usually referred to as self- system. organization is the spontaneous formation of 4. Alternative definitions of the nervous sys- well organized structures, patterns, or behav- tem as a dynamical system can include the iors, from random initial conditions. Typ- synaptic weights themselves among the vari- ically, these systems possess a large num- ables. At certain time scales, the weights are ber of elements or variables interacting in a a function of the evolution of the mem- complex way, and thus have very large state P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

neurodynamical approaches to consciousness 765

spaces. However, when started with some accurately termed phase-locking and phase- initial conditions, they tend to converge to scattering respectively (see below). small areas of this space which can be inter- 17. Mathematical methods are available to de- preted as a form of emergent eigenbehavior.” tect correlations between localized meta- 11. Mutual information quantifies the ability to bolic activations as measured by fMRI and predict the behavior of one element in the PET. The advantage of these methods is that system from the behavior of one or sev- they provide maps of functional connectiv- eral other elements (David, Cosmelli, & Fris- ity with a high spatial resolution in normal ton, 2004). This measure is one of several human subjects. These methods, however, tools used to quantify statistical dependence measure interactions that occur on the time within a system. Note that what counts as scale of a couple of seconds at best. This a spatiotemporal structure will depend on temporal resolution may be sufficient when which measure of statistical dependence is studying the neural correlates of slow expe- used. riential patterns, such as the evolution of cer- 12. “It seems that short-term memory may be tain emotions (Buchel & Friston, 1997). a reverberation in the closed loops of the 18. The fractal dimension d of a trajectory can cell assembly and between cell assemblies” be envisioned as follows: Imagine that each (Hebb, as cited in Amit, 1994,p.621). point along the trajectory is in fact a small 13. A precise definition of synchronization be- ball of lead. Then, the total mass of lead con- tween chaotic systems can be found in tained in a sphere centered on the trajec- (Pikovsky, Zaks, Rosenblum, Osipov, & tory will increase as a function of the sphere Kurths, 1997,p.680): “The phase synchro- radius r proportionally to rd. If the trajectory nization of a chaotic system can be defined as occupies all the space, then d is equal to the the occurrence of a certain relation between dimension of the space. If the trajectory is a 1 the phases of interacting systems or between straight line or a circle, d equals . the phase of a system and that of an exter- 19. See the previous note. nal force, while the amplitudes can remain 20. Note that synchrony can occur between two chaotic and are, in general, uncorrelated (see neurons without an actual direct relation also Brown & Kocarev, 2000; Rosenblum, between them, if they are driven by a com- Pikovsky, & Kurths, 1996). mon driver. This fact reveals one of the limi- 14. In Freeman’s words, “preafference provides tations of the synchrony measure so far. The an order parameter that shapes the attrac- three-neuron system that includes the driver, tor landscapes, making it easier to capture however, can be seen as a larger spatiotem- expected or desired stimuli by enlarging or poral pattern revealed by the synchrony mea- deepening the basins of their attractors. [...] sure. corollary discharges do this by a macroscopic 21. This theory was based, among other things, bias that tilts sensory attractor landscapes” on the observation of false conjunctions in (Freeman, 1999a, p. 112). the absence of attention: When presented 15. Edelman and Tononi distinguish primary briefly with a red square and a blue circle consciousness from higher-order conscious- outside of the scope of attention, a subject ness. The former involves the capacity to would sometimes report having seen a red construct a mental scene to guide behav- circle and a blue square. Such perception is ior without the semantic, linguistic, and self- typically an incorrect binding of the color and reflective capacities unique to the latter. shape attributes. 16. Here we are referring to desynchronized 22. This has led to the suggestion that gamma- EEG in the classical sleep/wake cycle sense, band synchrony is the trace of similar in which desynchronized gamma and beta processes in the emergence of dreaming frequencies dominate the EEG of the wak- consciousness in REM sleep and waking ing state, by contrast with the synchronized consciousness (Engel et al., 1999). slow-wave delta frequency EEG of sleep. 23. We do not wish at this point to step into This notion of desynchronized and synchro- the debate about which brain areas actu- nized EEG as a whole should not be confused ally participate in the generation of sen- with the synchronization and desynchroniza- sory awareness (see Rees, Kreiman, & Koch, tion of particular EEG signals, which is more 2002). P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

766 the cambridge handbook of consciousness

24. Letter strings presented to the subject for 1 on this controversial and interesting question s, for instance, generate an induced gamma (Crick & Koch, 2003; Metzinger, 2000;Noe¨ response that lasts roughly only for the & Thompson, 2004a,b; Pessoa, Thompson, & first 500 ms (Lachaux, unpublished observa- Noe,¨ 1998). tions). 33. This feature pertains to multistable and 25. For an extensive presentation of questions ambiguous perception in general (see concerning the experience of time, we refer Leopold & Logothetis, 1999). the reader to the notable work by Charles 34. This way of defining measures of subjec- Sherover (1991). tive experience should be sufficiently gen- 26. Varela in particular proposed a neurodynam- eral to include all the measures used in ical account of Husserl’s phenomenological psychophysics: Choosing to press one but- account of time consciousness (see Varela, ton among two, or to press one at a particu- 1999, and for further extensive discussion, lar time, for example, fits into that definition. Thompson, 2007). Psychophysics is indeed partly about first- 27. Tononi and Edelman (1998) do mention that person data. For example, the experimenter their dynamic core is constantly changing, shows a shape to a subject and asks him to but they do not develop this point further. press button A if what he sees looks more like a circle, and button B if it looks more 28. The question of whether this unity is illusory like a square. The subject’s answer is based or real remains an unresolved problem (Van- on one particular element of his subjective Rullen & Koch, 2003). experience (he selects one particular action 29 . Lamme’s distinction, however, is not com- in the universe of allowed responses, based on pletely equivalent to Block’s initial pro- the observation of his conscious visual expe- 1996 posal (Block, ). In its original formula- rience). The button press can therefore be tion, phenomenal consciousness is subjective seen as a (very crude) description of a con- experience, in the sense that there is some- scious content. thing it is like for the subject to be in the state. 35. In other words, what is needed here is first Access consciousness, on the other hand, is an a possible one-to-one monotonic correspon- information-theoretical concept that is sup- dence between the phenomena under inves- posed to account for the availability of con- tigation and the measurements. ‘Monotonic’ scious information for further rational guid- is to be understood in its usual mathemat- ance of behavior, including reportability. The ical sense: For three phenomena pa, pb, conceptual and empirical validity of this dis- and pc and their corresponding measure- tinction are a matter of lively debate in the ments m(pa), m(pb), and m(pc), it would science of consciousness (see Block, 1997, be desirable that if D(pa,pb) > D(pa,pc)(D 2000; see also the discussion in Thompson, being a subjective distance between experi- Lutz, & Cosmelli, 2005). ential phenomena), then d(m(pa), m(pb)) > 30 . The notion that consciousness has a unified d(m(pa),m(pc)) (d being the distance defined field structure goes back to A. Gurwitsch by the experimenter and the subject in the 1964 ( ). universe of measures;for a convergent per- 31. We see below that this general characteriza- spective see also Fell, 2004). tion needs some important qualifications, in 36. This relation implies an additional require- particular at the level of determining what ment for the measures of the subjective counts as a valid conscious experience and experience: They should be timed. Indeed, how to contrast such a conscious experience the dynamics of experiential phenomena can with possibly unconscious processing in sim- only be accessed through series of consecu- ilar situations. tive timed measures (as simple as a series 32. The theoretical validity and empirical plau- of button presses, for instance, or the time sibility of this approach remain a matter of course of the pressure applied on a joy- extensive discussion. Rather than endorse or stick). Therefore, to establish a strong rela- reject this approach, we wish to highlight tion between the dynamics of an experience it as an influential approach that can serve and the formation of certain patterns of neu- as a reference for further discussion. The ral activity, one should be able to say that interested reader is referred to several inter- the experience started at time t = 2s and esting publications (and references therein) fully developed between t = 5s and t = 10s P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

neurodynamical approaches to consciousness 767

(this is easy to understand in the case of Amit, D. J. (1994). The Hebbian paradigm an emotional reaction to a sound, for exam- reintegrated: Local reverberations as internal ple). It does not follow, however, that the representations. Behavioral and Brain Sciences, time as experienced must correspond precisely 18, 617–626. to the timing of neural processes. The for- Arbib, M. A. (2002). Handbook of brain theory mer is a matter of the content of experience and neural networks (2nd ed.). Cambridge, MA: and the latter of the neural vehicles that (in MIT Press. ways we do not fully understand) embody or Arbib, M. A., Erdi,´ P., & Szentagothai,´ J. (1997). encode those contents. Within certain small Neural organization: Structure, function, and temporal windows, a given neural vehicle dynamics. Cambridge, MA: MIT Press. could encode one event as happening before Arieli, A., Sterkin, A., Grinvald, A., & Aert- another event, even though that neural vehi- sen, A..(1996). Dynamics of ongoing activity: cle occurs after the neural vehicle encoding Explanation of the large variability in evoked the second event (see Dennett & Kinsbourne, cortical responses. Science, 273(5283), 1868– 1992). 1871. 37. ‘Continuous’ here is not meant as the oppo- Ashby, R. (1952). Design for a brain. London.: site of discrete, but rather is used to mean that Chapman-Hall. consciousness does not jump around with 2003 no connection whatsoever from one sort of Baars, B. J., & Franklin, S. ( ). How con- experience to another. scious experience and working memory interact. Trends in Cognitive Sciences, 7(4), 166–172. 38. Consider, however, the possibility of working 1989 with individuals who can produce and stabi- Blake, R. A. ( ). A neural theory of binocular 145 167 lize mental states more reliably (see Chap- rivalry. Psychology Review, 96, – . ter 19). The issue of working with ‘experts’ Blake, R., & Logothetis, N. K. (2002). Visual com- or trained subjects is important and con- petition. Nature Reviews Neuroscience, 3(1), troversial (Jack & Roepstorff, 2003; Lutz 13–21. & Thompson, 2003; Varela & Shear, 1999; Block, N. (1996). How can we find the neural Chapter 19). correlate of consciousness? Trends in Neuro- 39. Cognitive psychologists sometimes ask sub- sciences, 19(11), 456–459. jects very difficult questions, so how can they Block, N. (1997). On a confusion about a trust their answers? Why shall we trust the function of consciousness. In N. Block, O. button presses of a subject during a binocu- Flanagan, & G. Guzeldere¨ (Eds.), The nature lar rivalry experiment? The subject is asked of consciousness: Philosophical debates (pp. 375– to press the button as soon as one pattern 416). Cambridge, MA: MIT Press. dominates completely, but how can one be Block, N. (2001). Paradox and cross purposes in sure that the subject can actually do this task recent work on consciousness. Cognition, 79, reliably or that he has this sort of fine capac- 197–219. ity to attend to his own visual experience and Bressler, S. L. (1995). Large-scale cortical net- its dynamics in time? works and cognition. Brain Research Review, 40. There is a similar problem with the measure 20(3), 288–304. of neural events. For instance, with EEG, the Bressler, S. L., & Freeman, W.J. (1980). Frequency noise level is sometimes simply so strong that analysis of olfactory system EEG in cat, rab- measures of gamma activity cannot be made: bit, and rat. Electroencephalography and Clinical The Basic Requirement is not met. Neurophysiology, 50(1–2), 19–24. Bressler, S. L., & Kelso, J. A. (2001). Cortical coordination dynamics and cognition. Trends in References Cognitive Sciences, 5(1), 26–36. Brown, R., & Kocarev, L. (2000). A unifying def- Abarbanel, H. D., & Rabinovich, M. I. (2001). inition of synchronization for dynamical sys- Neurodynamics: Nonlinear dynamics and neu- tems. Chaos, 10(2), 344–349. robiology. Current Opinion in Neurobiology, Brown, R. J., & Norcia, A. M. (1997). A method 11(4), 423–430. for investigating binocular rivalry in real-time Abeles, M. (1991). Corticonics. New York: Cam- with steady-state VEP. Vision Research, 37, bridge University Press. 2401–2408. P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

768 the cambridge handbook of consciousness

Buchel, C., & Friston, K. J. (1997). Modulation of Dehaene, S., Sergent, C., & Changeux, J. P. connectivity in visual pathways by attention: (2003). A neuronal network model linking sub- Cortical interactions evaluated with structural jective reports and objective physiological data equation modelling and fMRI. Cerebral Cortex, during conscious perception. Proceedings of the 7(8), 768–778. National Academy of Sciences USA, 100(14), Chalmers, D. J. (2000). What is a neural corre- 8520–8525. late of consciousness? In T. Metzinger (Ed.), Dennett, D. C., & Kinsbourne, M. (1992). Time Neural correlates of consciousness: Empirical and and the observer: The where and when of con- conceptual questions. Cambridge, MA: MIT sciousness in the brain. Behavioral and Brain Press. Sciences, 15, 183–247. Chandrasekhar, S. (1961). Hydrodynamic and Depraz, N., Varela, F., & Vermersch, P. (2003). hydromagnetic stability. Oxford: Oxford Uni- On becoming aware. A pragmatics of experienc- versity Press. ing. Amsterdam: Benjamins. Cosmelli, D., David, O., Lachaux, J. P., Mar- Edelman, G. M., & Tononi, G. (2000). A universe tinerie, J., Garnero, L., Renault, B., et al. of consciousness: How matter becomes imagina- (2004). Waves of consciousness: Ongoing cor- tion. New York: Basic Books. tical patterns during binocular rivalry. Neu- Engel, A. K., Fries, P., Konig, P., Brecht, M., & roimage, 23(1), 128–140. Singer, W. (1999). Temporal binding, binoc- Crick, F. C., & Koch, C. (1990). Towards a neu- ular rivalry, and consciousness. Consciousness robiological theory of consciousness. Seminars and Cognition, 8(2), 128–151. in Neuroscience, 2 , 263–275. Engel, A. K., Fries, P., & Singer, W. (2001). Crick, F. C., & Koch, C. (1998). Consciousness Dynamic predictions: Oscillations and syn- and neuroscience. Cerebral Cortex, 8(2), 97– chrony in top-down processing. Nature Reviews 107. Neuroscience, 2 (10), 704–716. Crick, F. C., & Koch, C. (2003). A framework for Engel, A. K., & Singer, W. (2001). Temporal bind- consciousness. Nature Neuroscience, 6(2), 119– ing and the neural correlates of sensory aware- 126. ness. Trends in Cognitive Sciences, 5(1), 16–25. Crutchfield, J. P., & Kaneko, K. (1987). Phe- Faure, P., & Korn, H. (2001). Is there chaos in the nomenology of spatiotemporal chaos. In Direc- brain? I. Concepts of nonlinear dynamics and tions in chaos (pp. 272–353). World Scientific. methods of investigation. Les Comptes rendus Damasio, A. R. (1998). Investigating the biology de l’Académie des sciences III, 324(9), 773–793. of consciousness. Philosophical Transactions of Fell, J. (2004). Identifying neural correlates of the Royal Society London B: Biological Science, consciousness: The state space approach. Con- 353(1377), 1879–1882. sciousness and Cognition, 13, 709–729. Damasio, A. (1999). The feeling of what happens: Fell, J., Fernandez, G., Klaver, P., Elger, C. E., Body and emotion in the making of consciousness. & Fries, P. (2003). Is synchronized neuronal New York: Harcourt. gamma activity relevant for selective attention? Danckert, J., & Goodale, M. A. (2000). A con- Brain Research: Brain Research Reviews, 42(3), scious route to unconscious vision. Current 265–272. Biology, 10(2), 64–67. Fell, J., Klaver, P., Elger, C. E., & Fernandez, G. David, O., Cosmelli, D., & Friston, K. J. (2004). (2002). Suppression of EEG gamma activity Evaluation of different measures of functional may cause the attentional blink. Consciousness connectivity using a neural mass model. Neu- and Cognition, 11(1), 114–122. roimage, 21(2), 659–673. Freeman, W. J. (1975). Mass action in the nervous Dehaene, S., Kerszberg, M., & Changeux, system. New York: Academic Press. J. P. (1998). A neuronal model of a global Freeman, W. J. (1999a). How brains make up their workspace in effortful cognitive tasks. Proceed- minds. London: Weidenfeld and Nicolson. ings of the National Academy of Sciences USA, Freeman, W. J. (1999b). Consciousness, inten- 95(24), 14529–14534. tionality and causality. Journal of Consciousness Dehaene, S., & Naccache, L. (2001). Towards a Studies, 6(11–12), 143–172. cognitive neuroscience of consciousness: Basic Freeman, W. J. (2000a). Neurodynamics: An evidence and a workspace framework. Cogni- exploration of mesoscopic brain dynamics. tion, 79(1–2), 1–37. London: Springer-Verlag. P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

neurodynamical approaches to consciousness 769

Freeman, W. J. (2000b). Mesoscopic neuro- organization in physics, chemistry, and biology. dynamics: From neuron to brain. Journal of New York: Springer-Verlag. Physiology Paris, 94(5–6), 303–322. Hanna, R., & Thompson, E. (2003). The spon- Freeman, W. J. (2004). Mass action in the ner- taneity of consciousness. Canadian Journal of vous system (2nd ed.). Retrieved from http:// Philosophy, 29, 133–162. sulcus.berkeley.edu. Hebb, D. O. (1949). The organization of behavior. Fries, P., Reynolds, J. H., Rorie, A. E., & Des- New York: Wiley. imone, R. (2001). Modulation of oscillatory Herculano-Houzel, S., Munk, M. H., Neuen- neuronal synchronization by selective visual schwander, S., & Singer, W. (1999). Precisely attention. Science, 291, 1560–1563. synchronized oscillatory firing patterns require Fries, P., Roelfsema, P. R., Engel, A. K., Konig, P., electroencephalographic activation. Journal of & Singer, W. (1997). Synchronization of oscil- Neuroscience, 19(10), 3992–4010. latory responses in visual cortex correlates with Hilgetag, C. C., & Kaiser, M. (2004). Clustered perception in interocular rivalry. Proceedings of organization of cortical connectivity. Neuroin- the National Academy of Sciences USA, 94(23), formatics, 2 (3), 353–360. 12699 12704 – . Hollender, D. (1986). Semantic activation with- Friston, K. J. (1997). Transients, metastability, out conscious identification in dichotic lis- and neuronal dynamics. Neuroimage, 5(2), 164– tening, parafoveal vision, and visual masking: 171. A survey and appraisal. Behavioral and Brain Friston, K. J. (2000). The labile brain. I. Neuronal Sciences, 9, 1–66. transients and nonlinear coupling. Philosophical Ito, J., & Kaneko, K. (2002). Spontaneous struc- Transactions of the Royal Society London B: Bio- ture formation in a network of chaotic units logical Science, 355(1394), 215–236. with variable connection strengths. Physical Friston, K. J. (2002a). Beyond phrenology: What Review Letters, 88(2). can neuroimaging tell us about distributed Jack, A., & Roepstorff, A. (2003). Trusting the circuitry? Annual Review of Neuroscience, 25, subject (Vol. 1). Journal of Consciousness Stud- 221 250 – . ies, 10(9–10). 2002 Friston, K. ( b). Functional integration and James, W. (1981). The principles of psychology. inference in the brain. Progress in Neurobiology, Cambridge, MA: Harvard University Press. 2 113 143 68( ), – . (Original work published 1890) 2005 Friston, K. J. ( ). Models of brain function in Kelso, J. A. S. (1995). Dynamic patterns: The self- neuroimaging. Annual Review of Psychology, 56, organization of brain and behavior. Cambridge, 57 87 – . MA: MIT Press. Garfinkel, A., Spano, M. L., Ditto, W.L., & Weiss, Korn, H., & Faure, P. (2003). Is there chaos in the 1992 J. N. ( ). Controlling cardiac chaos. Science, brain? II. Experimental evidence and related 1230 1235 257, – . models. Comptes rendus biologies, 326(9), 787– Gomes, G. (1998). The timing of conscious 840. experience: A critical review and reinterpre- Kuramoto, Y. (1984). Cooperative dynamics of tation of Libet’s research. Consciousness and oscillator community: A study based on lat- 4 559 595 Cognition, 7( ), – . tice of rings. Progress of Theoretical Physics,79 Gray, C. M., Konig, P., Engel, A. K., & Singer, (Suppl.), 223–240. 1989 W. ( ). Oscillatory responses in cat visual Lachaux, J.-P., George, N., Tallon-Baudry, C., cortex exhibit inter-columnar synchroniza- Martinerie, J., Hugueville, L., Minotti, L., et al. tion which reflects global stimulus properties. (2005). The many faces of the gamma band 6213 334 337 Nature, 338( ), – . response to complex visual stimuli. Neuroim- Gurwitsch, A. (1964). The field of consciousness. age, 25(2), 491–501. Pittsburgh: Duquesne University Press. Lachaux, J. P., Lutz, A., Rudrauf, D., Cos- Gutkin, B., Pinto, D., & Ermentrout, B. (2003). melli, D., Le Van Quyen, M., Martinerie, Mathematical neuroscience: From neurons to J., et al. (2002). Estimating the time- circuits to systems. Journal of Physiology Paris, course of coherence between single-trial brain 97(2–3), 209–219. signals: An introduction to wavelet coher- Haken, H. (1983). Synergetics: An introduc- ence. Neurophysiology Clinics, 32(3), 157– tion. Nonequilibrium phase transitions and self- 174. P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

770 the cambridge handbook of consciousness

Lachaux, J. P., Rodriguez, E., Martinerie, J., Naionatl Academy of Sciences USA, 99(3), Adam, C., Hasboun, D., & Varela, F. J. (2000). 1586–1591. A quantitative study of gamma-band activity Lutz, A., & Thompson, E. (2003). Neuro- in human intracranial recordings triggered by phenomenology: Integrating subjective experi- visual stimuli. European Journal of Neuroscience, ence and brain dynamics in the neuroscience of 12(7), 2608–2622. consciousness. Journal of Consciousness Studies, Lachaux, J. P., Rodriguez, E., Martinerie, J., & 10, 31–52. Varela, F. J. (1999). Measuring phase synchrony Martinerie, J., Adam, C., Le Van Quyen, M., in brain signals. Human Brain Mapping, 8(4), Baulac, M., Clemenceau, S., Renault, B., et al. 194–208. (1998). Epileptic seizures can be anticipated by Lachaux, J. P., Rudrauf, D., & Kahane, P. (2003). non-linear analysis. National Medicine, 4(10), Intracranial EEG and human brain mapping. 1173–1176. Journal of Physiology Paris, 97(4–6), 613–628. Maturana, H. R., & Varela, F. J. (1980). Autopoiesis Lamme, 2003, 2004 and cognition: The realization of the living. Boston Le Van Quyen, M. (2003). Disentangling the studies in the philosophy of science, Vol. 42. dynamic core: A research program for a neuro- Dordrecht: D. Reidel. dynamics at the large-scale. Biological Research, Mesulam, M. M. (1999). Spatial attention and 36(1), 67–88. neglect: Parietal, frontal and cingulate contri- Le Van Quyen, M., Foucher, J., Lachaux, J.-P., butions to the mental representation and atten- Rodriguez, E., Lutz, A., Martinerie, J., et al. tional targeting of salient extrapersonal events. (2001). Comparison of Hilbert transform and Philosophical Transactions of the Royal Society wavelet methods for the analysis of neuronal London B: Biological Science, 354(1387), 1325– synchrony. Journal of Neuroscience Methods, 1346. 111(2), 83–98. Metzinger, T. (Ed.). (2000). Neural correlates of Leopold, D. A., & Logothetis, N. K. (1999). Multi- consciousness: Empirical and conceptual ques- stable phenomena: Changing views in perceptions. Cambridge, MA: MIT Press. tion. Trends in Cognitive Sciences, 3(7), 254– Milner, D., & Goodale, M. (1995). The visual 264. brain in action. Oxford: Oxford University Libet, B. (2002). The timing of mental events: Press. Libet’s experimental ﬁndings and their impli- Munk, M. H., Roelfsema, P. R., Konig, P., Engel, cations. Consciousness and Cognition, 11(2), A. K., & Singer, W. (1996). Role of reticular 291–299; discussion 304–233. activation in the modulation of intracortical 5259 271 274 Llinas, R. (1988). The intrinsic electrophysiologi- synchronization. Science, 272( ), – . cal properties of mammalian neurons: Insights Nicolis, G., & Prigogine, G. (1977). Self- into central nervous system function. Science, organization in nonequilibrium systems. New- 242(4886), 1654–1664. York: Wiley. Llinas, R., & Pare, D. (1991). Of dreaming and Nicolis, G., & Prigogine, I. (1989). Exploring wakefulness. Neuroscience, 44(3), 521–535. complexity: An introduction. New York: W. H. Llinas, R., & Ribary, U. (2001). Consciousness Freeman. and the brain. The thalamocortical dialogue Noe,¨ A., & Thompson, E. (2004a). Are there neu- in health and disease. Annals of the New York ral correlates of consciousness? Journal of Con- Academy of Sciences, 929, 166–175. sciousness Studies, 11, 3–28. Llinas, R., Ribary, U., Contreras, D., & Noe,¨ A., & Thompson, E. (2004b). Sorting out Pedroarena, C. (1998). The neuronal basis for the neural basis of consciousness. Authors’ consciousness. Philosophical Transactions of reply to commentators. Journal of Conscious- the Royal Society London B: Biological Science, ness Studies, 11, 87–98. 353(1377), 1841–1849. Nunez, P. L. (2000). Toward a quantitative Lutz, A., Lachaux, J. P., Martinerie, J., & Varela, description of large scale neocortical dynamic F. J.. (2002). Guiding the study of brain dynam- function and EEG. Behavioral and Brain Sci- ics by using ﬁrst-person data: Synchrony pat- ences, 23(3). terns correlate with ongoing conscious states O’Regan, K., & Noe,¨ A. (2001). sensorimotor during a simple visual task. Proceedings of the account of vision and visual consciousness. P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

neurodynamical approaches to consciousness 771

Behavioral and Brain Sciences, 24(5), 939– Roskies, A. L. (1999). The binding problem 973. [Special issue]. Neuron, 24(7–9). Palus, M. (1996). Nonlinearity in normal human Rudrauf, D., Lutz, A., Cosmelli, D., Lachaux, EEG: Cycles, temporal asymmetry, nonstation- J.-P., & Le Van Quyen, M. (2003). From arity and randomness, not chaos. Biological autopoiesis to neurophenomenology: Francisco Cybernetics, 75, 389–396. Varela’s exploration of the biophysics of being. Panksepp, J. (1998). The periconscious substrates Biological Research, 36, 27–66. of consciousness: Affective states and the evo- Schiff, S. J., Jerger, K., Duong, D. H., Chang, T., lutionary origins of self. Journal of Conscious- & Spano, M. L. (1994). Controlling chaos in ness Studies, 5, 566–582. the brain. Nature, 8, 615–620. Parvizi, J., Damasio, A. (2001). Consciousness and Searle, J. (2000). Consciousness. Annual Review the brainstem. Cognition, 79, 135–159. of Neuroscience, 23. Perkel, D. H., Gerstein, G. L., & Moore, G. P. Sewards, T. V., & Sewards, M. A. (2001). On (1967). Neuronal spike trains and stochas- the correlation between synchronized oscilla- tic point processes. I. The single spike train. tory activities and consciousness. Consciousness Biophysical Journal, 7(4), 391–418. and Cognition, 10(4), 485–495. Pessoa, L., Thompson, E., & Noe,¨ A. (1998). Find- Sherman, S. M., & Guillery, R. W. (2002). The ing out about filling-in: A guide to perceptual role of the thalamus in the flow of informa- completion for visual science and the philos- tion to the cortex. Philosophical Transactions of ophy of perception. Behavioral and Brain Sci- the Royal Society London B: Biological Science, ence, 21(6), 723–748. 357(1428), 1695–1708. Petitmengin, C. (in press). Describing one’s sub- Sherover, C. (1991). The human experience of time. jective experience in the second-person. Phe- Evanston, IL: Northwestern University Press. nomenology and the Cognitive Sciences. Simons, D. J. (2000). Attentional capture and Pikovsky, A., Zaks, M., Rosenblum, M., Osipov, inattentional blindness. Trends in Cognitive Sci- G., & Kurths, J. (1997). Phase synchroniza- ence, 4(4), 147–155. tion of chaotic oscillations in terms of periodic Singer, W. (1998). Consciousness and the struc- 4 680 687 orbits. Chaos, 7( ), – . ture of neuronal representations. Philosophi- Rees, G., Kreiman, G., & Koch, C. (2002). Neural cal Transactions of the Royal Society London B: correlates of consciousness in humans. Nature Biological Science, 353(1377), 1829–1840. 4 261 270 Reviews Neuroscience, 3( ), – . Singer, W., & Gray, C. M. (1995). Visual fea- Revonsuo, A., Wilenius-Emet, M., Kuusela, J., & ture integration and the temporal correlation Lehto, M. (1997). The neural generation of a hypothesis. Annual Review of Neuroscience, 18, unified illusion in human vision. Neuroreport, 555–586. 18 3867 3870 8( ), – . Skarda, C. A., & Freeman, W. J. (1987). How Rizzolatti, G., & Craighero, L. (2004). The brains make chaos in order to make sense of mirror-neuron system. Annual Review of Neu- the world. Behavioral and Brain Sciences, 10(2), roscience, 27, 169–192. 161–195. Rodriguez, E., George, N., Lachaux, J.- P., Mar- Stephan, K. E., Hilgetag, C. C., Burns, G. A., tinerie, J., Renault, B., & Varela, F. J. (1999). O’Neill, M. A., Young, M. P., & Kotter, R. Perception’s shadow: Long-distance synchro- (2000). Computational analysis of functional nization of human brain activity. Nature, connectivity between areas of primate cerebral 397(6718), 430–433. cortex. Philosophical Transactions of the Royal Rodriguez, R., Kallenbach, U., Singer, W., & Society London B: Biological Science, 355(1393), Munk, M. H. (2004). Short- and long-term 111–126. effects of cholinergic modulation on gamma Tallon-Baudry, C., & Bertrand, O. (1999). Oscil- oscillations and response synchronization in latory gamma activity in humans and its role the visual cortex. Journal of Neuroscience, in object representation. Trends in Cognitive 24(46), 10369–10378. Sciences, 3(4), 151–162. Rosenblum, M., Pikovsky, A., & Kurths, J. (1996). Tallon-Baudry, C., Bertrand, O., & Fischer, C. Phase synchronization of chaotic oscillators. (2001). Oscillatory synchrony between human Physical Review Letters, 76(11), 1804–1807. extrastriate areas during visual short-term P1: KAE 0521857430c26 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 1:55

772 the cambridge handbook of consciousness

memory maintenance. Journal of Neuroscience, Journal of Consciousness Studies, 3(4), 330– 21(20), RC177. 349. Thompson, E. (2007). Mind in life: Biology, phe- Varela, F. (1999). The specious present: The nomenology, and the sciences of mind. Cam- neurophenomenology of time consciousness. bridge, MA: Harvard University Press. In J. Petitot, F. J. Varela, B. Pachoud, & Thompson, E., Lutz, A., & Cosmelli, D. (2005). J. M. Roy (Eds.), Naturalizing phenomenology Neurophenomenology: An introduction for (pp. 266–314). Stanford: Stanford University neurophilosophers. In A. Brook & K. Akins Press. (Eds.), Cognition and the brain: The philosophy Varela, F., Lachaux, J.-P., Rodriguez, E., & Mar- and neuroscience movement (pp. 40–97). New tinerie, J. (2001). The brainweb: Phase synchro- York: Cambridge University Press. nization and large-scale integration. Nature Thompson, E., & Varela, F. J. (2001). Radical Reviews Neuroscience, 2 (4), 229–239. embodiment: Neural dynamics and conscious- Varela, F. J., & Shear, J. (1999). First-person ness. Trends in Cognitive Sciences, 5(10), 418– methodologies: Why, when and how. Journal 425. of Consciousness Studies, Thorpe, S., Fize, D., & Marlot, C. (1996). Speed of Von der Malsburg, C. (1981). The correlation the- processing in the human visual system. Nature, ory of brain function.Gottingen:¨ Germany: Max 381(6582), 520–522. Planck Institute for Biophysical Chemistry. Tononi, G., & Edelman, G. M. (1998). Conscious- Watts, D. J., & Strogatz, S. H. (1998). Collective ness and complexity. Science, 282(5395), 1846– dynamics of ‘small-world’ networks. Nature, 1851. 393(6684), 440–442. Treisman, A. M., & Gelade, G. (1980). A feature- Weiskrantz, L. (1990). The Ferrier lecture, 1989. integration theory of attention. Cognitive Psy- Outlooks for blindsight: Explicit methodolo- chology, 12(1), 97–136. gies for implicit processes. Philosophical Trans- Tsuda, I. (2001). Toward an interpretation of actions of the Royal Society London B: Biological dynamic neural activity in terms of chaotic Science, 239(1296), 247–278. dynamical systems. Behavioral and Brain Wicker, B., Ruby, P., Royet, J. P., & Fonlupt, P. Sciences, 24(5), 793–810; discussion 810–748. (2003). A relation between rest and the self Van Gelder, T. J. (1998). The dynamical hypoth- in the brain? Brain Research: Brain Research esis in cognitive science. Behavioral and Brain Review, 43(2), 224–230. Sciences, 21, 1–14. Wolfe, J. M., & Bennett, S. C. (1997). Preattentive Van Van Gelder, T. J. (1999). Dynamic app- object ﬁles: Shapeless bundles of basic features. roaches to cognition. In R. Wilson & F. Keil Vision Research, 37(1), 25–43. (Eds.), The MIT encyclopedia of cognitive sciences Wright, J. J., & Liley, D. T. J. (1996). Dynamics (pp. 244–246). Cambridge MA: MIT Press. of the brain at global and microscopic scales: VanRullen, R., & Koch, C. (2003). Is percep- Neural networks and the EEG. Behavioral and tion discrete or continuous? Trends in Cognitive Brain Sciences, 19, 285–320. Sciences, 7(5), 207–213. Zahavi, D. (2005). Subjectivity and selfhood: Inves- Varela, F. (1995). Resonant cell assemblies: A tigating the ﬁrst-person perspective. Cambridge, new approach to cognitive functioning and MA: MIT Press. neuronal synchrony. Biological Research, 28(81– Zihl, J., Von Cramon, D., & Mai, N. (1983). 95). Selective disturbance of movement vision after Varela, F. J. (1996). Neurophenomenology: A bilateral brain damage. Brain, 106(2), 313– methodological remedy for the hard problem. 340. P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

B. Neuropsychological Aspects of Consciousness Disorders and Neuroimaging

773 P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

774 P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

CHAPTER 27 The Thalamic Intralaminar Nuclei and the Property of Consciousness

Joseph E. Bogen

Abstract (1853). Thalamic contributions were a principal emphasis of the landmark Laurentian An anatomico-physiologic approach to con- Symposium entitled “Brain Mechanisms and sciousness is facilitated by recognizing that Consciousness” held a century later (Adrian, the various meanings of the term “conscious- Bremer, & Jasper, 1954). However, as argued ness” have in common a crucial core var- by Walshe (1965) in his criticism of that iously called subjectivity, sentience, aware- emphasis, toward the end of the 19th century ness, conscious itself, consciousness-as-such, the cerebral cortex had become accepted by consciousness per se, primary consciousness, many neurologists as the “seat of conscious- or simply, by some earlier authors, con- ness,” as well as of intelligence, ideation, and sciousness. In this chapter it is called C. A memory. In the Laurentian Symposium it sharp distinction is made between the prop- was Lashley and Bremer who mainly dis- erty C and the manifold contents of con- sented from the thalamic emphasis. Bremer sciousness, the partial loss of which is typi- insisted that “it is the dynamic integration cal of circumscribed neocortical lesions. The of all cerebral processes at a single moment neuronal mechanism needed for C is here which makes consciousness” (Adrian et al., proposed to involve the intralaminar thala- 1954,p.497). This so-called global view mic nuclei. Subsidiary hypotheses are brieﬂy still appeals to many people although the considered concerning volition and episodic impressive extent to which “dynamic inte- memory. gration” has since been shown to occur without consciousness (as discussed further below) has somewhat lessened the connec- Introduction tion. Lashley reiterated his earlier opinion, possibly the strongest argument against a The belief that consciousness depends signif- thalamic “seat” of consciousness. He wrote icantly on thalamic function is at least as old that “the neurological activity in a psycho- as the “automatic apparatus” of Carpenter logical event cannot be simpler in structure 775 P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

776 the cambridge handbook of consciousness

or in number of elements than the psycho- the view that insisting on a single criterion logical activity which it constitutes. . . . The for a supposedly unitary property has in sci- complexity of mental processes is too great ence been “long since given up”; a standard to be represented by the permutations of example is length, which is measured in dif- a small number of cells” (Lashley, 1952). ferent ways depending on the context. And I This argument has perhaps been answered in follow, as described in some detail elsewhere part by writers who have made a distinction (Bogen, 1995b), those philosophers of sci- between the property of consciousness and ence who advise that formal definitions must the contents of consciousness, as discussed follow rather than precede understanding of further on, emphasizing that only an exceed- a subject. ingly small part of the potential contents can Although it seems desirable not to delay be conscious at any moment. Toward the end a neurobiological approach to conscious- of the 20th century, brain-oriented books on ness by getting mired down in discussions the subject commonly have supposed that of definition or of uncertain cases (Crick & mentation, conscious or not, depends upon Koch, 1990, 2000), it is nonetheless trou- thalamocortical interaction while offering blesome that the word “consciousness” is varying views as to when or how this activity used in many differing ways (Galin, 1992; is endowed with consciousness (Baars, 1988, Guzeldere,¨ 1997; Natsoulas, 1983). The vari- 1997; Cairns-Smith, 1996; Churchland, ety of usages contributed to the opinion 2002: Crick, 1994; Darnasio, 1999; Edel- of G. Ryle that “there would seem to be man, 1989, 1992; Koch, 2004; Llinas, 2001; no actual thing or process that our past Taylor, 1999). usages have been getting at” (Guzeldere,¨ This chapter is devoted to reviewing some 1997,p.473). Others have taken a less rad- aspects of thalamic structure and function as ical approach, recognizing that the many they may contribute to making small frac- usages have in common (as their intersections of mentation momentarily conscious, tion or commonality) a crucial, central core with particular emphasis on those thalamic for which have been suggested many semi- constituents sometimes called non-specific. synonyms. These semi-synonyms include Considerable emphasis is given to the effects first-person experience (Fessard, 1954), of certain thalamic lesions. Before presenting subjectivity (Chalmers, 2002; Flanagan, a simplified description of thalamic anatomy 1992), awareness (Goldman, 1993; Jennett, as a prelude to selected physiologic and 2002), sentience (Searle, 1993, 1998), inner pathologic observations, I offer a brief dis- experience (Metzinger, 1995), phenome- cussion of usages of the word “conscious- nal consciousness (Block, 1995), primary ness” and a few explicit assumptions about consciousness (Edelman, 1992; Revonsuo, neuronal activity patterns. 1995), crude consciousness (Cairns, 1952), consciousness-as-such (Baars, 1988, 1993), consciousness per se (Gray, 1995), con- The Concept of “C” or Elementary sciousness itself (Moscovitch, 1995), the cru- Consciousness cial core of consciousness (Bogen, 1995a), core consciousness (Damasio, 1999) or sim- Although there is often considerable agree- ply, consciousness (Grossman, 1980; Moore, ment as to who or what is conscious, the 1922). In this chapter I simply refer to basis on which we make this attribution consciousness as C. is not always the same – there seems to be no single universally applicable criterion (Allport, 1988). We often use accurate eye C Includes One (Minimal) Aspect tracking of a moving object as an indicator of the Complex Concept “Self” that someone is conscious, but this can be fallible as well as often inappropriate. I fol- The word “self” seems to be used even more low Hempel (1966) and Goldman (1993)in variously than the word “consciousness.” My P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

the thalamic intralaminar nuclei and the property of consciousness 777

Webster’s Collegiate Dictionary has about 350 activity pattern that represented an object, usages spread over three and a half pages. in his broad sense, as a “pre-percept”). As a possible example of a relatively narrow According to Hart and Whitlow (1995), usage, Dennett (1991) refers to the “self as the human sense of self has five facets: the center of narrative gravity.” This is a felic- None of these, even Hart’s “subjective self- itous phrase, but linguistic ability (good or awareness,” seems terribly close to what bad) is not essential to the C we share with Flanagan and Churchland call “me-ness.” so many other species. At another extreme, Hart’s other four facets – objective self- Galin (1992, 2003) proposed that self be awareness, memories of self, representations defined as “the overall organization that of self, and theories of self – are even less makes a person a unity.” A similar concept close to me-ness. That even the “subjective of self was discussed by P. S. Churchland in self-awareness” of Hart is not as low level a multiauthored discussion entitled “Reflec- as Flanagan and I have in mind is indicated tions on Self” in the April 12, 2002 issue of by maintaining that monkeys do not have Science. This “overall organization” view may even this aspect of self (Hart & Whitlow, not be the most inclusive usage possible, but 1995). I do believe monkeys know who they it is surely a contender, it involving such a are, which infants are theirs, etc. (Bergman multitude of unifying mechanisms, both hor- et al., 2003) even though they do not under- monal and neuronal. We need not attempt stand the function of a mirror a` la Gallup’s a catalogue of all that self might contain in (Gallup, 1995) chimps. order to point to one aspect that is typi- It needs to be emphasized that the me- cally implied when the word “consciousness” ness ascribed here to C is neither an aspect is used. of cognition nor does it require cognition; Flanagan (1991,p.352) quotes William it is, in Thomas Sullivan’s term, subconcep- James as saying, “Whatever I may be thinking tual. Taylor (2002) offers a nice summary of, I am always at the same time more or less of authors advocating related concepts of aware of myself, of my personal existence.” “primitive self” or “minimal self” or “pre- Flanagan goes on to observe, “This low level reflective self” or “ipseity.” This minimal, sense of ‘me-ness,’ of ‘something happen- subconceptual aspect of self is, of course, a ing here’ does seem to underlie all conscious far cry from what some philosophers have experience. All conscious experiences are, in mind. For example, “While a bat, like in addition to being experienced, experi- even the lowly lobster, has a biological self, enced as attached to the subject of these it has no selfy self to speak of – no Cen- very experiences.” As Churchland wrote, “A ter of Narrative Gravity . . . no regrets, no single thread of “me-ness” runs through the complex yearnings, no nostalgic reminis- entire fabric of one’s existence” (Church- cences, no grand schemes” (Dennett, 1991). land, 2002,p.61). There might be arguable According to P. S. Churchland (personal exceptions (e.g., certain meditative states) to communication, June 30, 2003), Dennett C having this quite minimal aspect of self, still holds firmly to his 1991 views, including but “me-ness” names what I believe to be an that brains do not have any “hardwiring” for essential characteristic of C. consciousness. That a connection with self is an indis- Identification of the neuronal substrates pensable aspect of any sensible usage of the for the richness of a maximally “selfy self” term “consciousness” has often been main- may be a desirable long-term goal; consid- tained. Damasio (1999) suggests repeat- erably less ambitious is the hope expressed edly that consciousness depends upon bring- in this chapter of identifying the neuronal ing together what he calls an “object” and requirements for the primitive, low-level the self. Damasio includes as objects not sense of me-ness and giving some indica- only places and persons but also pains tion of how this neuronal mechanism is and emotions; he sometimes uses the term transiently connected to the neuronal repre- “object proxy” (I would refer to the neuronal sentation of a pre-percept, thus embodying P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

778 the cambridge handbook of consciousness

what Damasio discussed as the connection point in common. . . . I will call this common of self and “object.” element ‘consciousness’. . . . We have then in every sensation two distinct terms, (1); consciousness; in respect of which all sen- There Is a Lot That C Does sations are alike, and (2) something else [the Not Include content], in respect of which one sensation differs from another.” I have discussed elsewhere (Bogen, 1995a) One fact motivating the approach of this further philosophical considerations con- chapter is that partial, specific loss of concerning the semantics of C, including tents is typical of cerebro-cortical lesions Carnap’s concept of explication and Naess’s that leave C otherwise intact. This chapter notion of precisation. In paraphrase, the con- concentrates on the search for a process or cept of Naess says the following: The formu- mechanism providing the property C, with- lation C is more “precise” than conscious- out trying to explain everything else that ness if there are properties of consciousness consciousness might include. that are not properties of C, but there are no properties of C that are not also properties of consciousness. C is a much narrower concept Neuronal Activity Patterns than many people’s usage of consciousness. C is intended to include the above described A persistent question about consciousness low-level sense of me-ness, accompanying is, Where does it come from? I take it some percept or affect, but C is intended for granted in this chapter that C depends to convey very little else. upon specific brain processes. I assume that Whatever else various usages of con- every sensation, thought, feeling, memory, sciousness sometimes include (e.g. planning, and expectation is represented in the brain control, time-binding, monitoring, supervi- by a neuronal activity pattern (NAP). This sion, language, “higher thoughts,” self-image, view, although explicitly mechanistic, is theory-of mind, sociality), none of these is not necessarily materialistic or completely involved in the concept C, whereas C does deterministic; this is discussed in detail else- appear to be involved in most of the more where (Bogen, 1995b; Bogen et al., 1998) expansive concepts of consciousness. for those who care about the metaphysics. The view urged here is that we are I only assume here that C requires specific well served by making a sharp distinc- neuronal activity that we hope to identify. tion between the property C and the This approach suggests the availability to contents, potentially enormous, of con- experimentation of various questions: First, sciousness. Most of the authors cited in what is the change in a NAP that occurs the previous paragraph using some semi- when it is associated with C? This implies synonym for C have made this distinction. that recording from brain cells (electrically Other examples include Panksepp’s distinc- or chemically or both) will provide infor- tion between “the essential foundations of mation that distinguishes, in the words of consciousness and the contents of conscious- Marcel, “processing without consciousness ness” (Panksepp, 1998,p.313) and the view from processing which is also describable as of Grossman (1980) who wrote, “We can also conscious” (Marcel, 1988,p.126). Second, introspectively discriminate between the is there an identifiable mechanism that pro- contents of consciousness . . . and the qual- duces the change in a NAP? Such a mech- ity of being conscious.” The earliest example anism, which associates NAPs with C, is I have found is the distinction made by G. named in this chapter Mc. A third question E. Moore. Landesman (1967), quotes Moore is, What determines which NAP will acquire (1922) as follows, “The sensation of blue dif- C at any particular time? fers from that of green. But it is plain that Regarding the first question, it is some- if both are sensations they also have some times suggested that the change from NAP P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

the thalamic intralaminar nuclei and the property of consciousness 779

to CNAP is simply an increase in activity Similarly with C, we commonly detect its as measured by neuronal firing rates (Fuster, presence by its effects. The effects of C, 2002). As Moutoussis and Zeki (2002) put when it is generated by Mc, depend on the it, “When certain levels of activation are specific NAPs involved. Determining which reached and probably in combination with effects might be considered most typical of the activation of other areas as well, there consciousness is important, but is not the is the generation of a conscious visual per- immediate problem, which is to find the cept.” On the other hand, recordings from mechanisms necessary to achieve any of the neurons known to respond to a specific stim- various effects. ulus sometimes show a decrease in firing rate when the subject indicates it is aware of that stimulus (Koch, 2004; Logothetis, 2002). C Requires a Time-Dependent Process In answer to the second question, it has been suggested that the change in NAPs Most discussants agree that NAPs that are involves increased synchrony (Singer, 1998). potentially conscious make up a restricted Of particular interest are synchronous dis- fraction of the NAPs possible in a brain. charges in the gamma range (20–70 Hz), And of those NAPs that can acquire C at most often around 40 Hz. Although origi- some time or another, a very small fraction nally described in anesthetized animals, the are endowed with C at any one moment. gamma synchrony (at least for vision) is There is considerable debate as to how more evident in the alert state, is enhanced long a “moment” is, probably one to sev- by arousal, and is related to the presence eral hundred milliseconds and possibly vary- and properties of experimental stimuli (Fries ing with the NAP involved. The exact time et al., 2001; Gray & Viana Di Prisco, 1997; is not essential for most hypotheses about Munk et al., 1996). consciousness. Most views currently advo- As to the third question, what factors cated seem compatible with a metaphor of determine which NAP will acquire C, the Pavlov: “If the place of optimal excitabil- current answers are usually expressed in ity were luminous, then we should see [it] such psychological terms as “salience” or playing over the cerebral surface” (Pavlov, “attention”; in this chapter, I briefly men- 1928,p.222). I believe that the “hot spot” tion some thalamic contributions to these can involve the depths, as well as the sur- two important aspects. Mainly, however, I face, and to some extent can be multiple. discuss the second question: how to go about As mentioned in the previous paragraph, finding Mc, a mechanism that can endow a “optimal excitability” does not necessarily NAP with C. mean “increased excitability.” Evidence is presented that specific, wired-in thalamic structures are needed for Mc; in addition, We Are Looking for Mc, there needs to be some active process that Rather Than C transiently involves each NAP for the brief time it is endowed with C. This implies an C is provided by some cerebral mechanism, affirmative answer to the question, “Does Mc. It is this mechanism that we hope to time help to understand consciousness?” identify and ultimately analyze. Trying to (Engel et al., 1999). point to C may turn out to be like pointing to the wind. We can point to the effects of the wind, and we can usually give a good C Is an Emergent Property account of what causes the wind, the causes often being quite distant (in miles) from the A question not always explicitly addressed effects. We do not actually see the wind when looking for some Mc is whether C itself, only leaves fluttering, or material such is an aggregative or an emergent prop- as dust or smoke being carried by the wind. erty. The term “emergent” is used variously; P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

780 the cambridge handbook of consciousness

sometimes it simply means “arising from” computers; Bunin & Wightman, 1999; Fuxe as a bather emerging from the water. & Agnati, 1991; Marder & Bucher, 2001). An Somewhat idiosyncratic is the definition in extensive review of the possible role of glia is Rumbaugh, Beran, and Pate (2003,p.53): found in Fields and Stevens-Graham (2002). “Emergent behaviors are new patterns of That C is aggregative has had many responding with no antecedent in previously advocates; a good example is the suggestion learned behavior [and] are applied appropri- in Greenfield (1998) that the intensity of ately to novel situations.” A more common subjective experience may correlate with the meaning among philosophers, not necessar- size of a non-specialized neuronal network. ily involving novelty, was adopted by Dudai In line with the idea of a critical mass of con- (2002,p.218), who noted that in a com- nections is the view of Cairns-Smith (1996) plex system there may be properties that are who suggested that consciousness arises by “emergent, i.e., appear only at a higher level virtue of the extensive white matter making of organization.” up so much of the cerebrum; this seems to This chapter employs a philosophi- echo R. Sperry’s (1980, 1991) discussions of cally technical usage, following Wimsatt widespread connections leading to an “emer- (1976, 2000), who contrasts “emergent” with gent” consciousness (see also Bogen, 1998). “aggregative.” A property possessed by a Different, but also suggesting aggregation is combination of parts is said to be aggregative the belief of R. Llinas´ (2001) that each indi- if it results from adding up smaller amounts vidual neuron has what he called “a mod- of that property already possessed by the icum of qualia” so that when many are parts; common examples are mass or electri- added together there is a substantial amount. cal charge. Sometimes an aggregate reaches (Llinas` also believes that the circuit “archi- what is called “critical mass,” as when a forest tecture” is important.) A similar view was fire is large enough to create a wind that then expressed earlier by Lashley who spoke of fans the flames that created it; this sort of “consciousness quanta” possessed even by property is sometimes called “emergent.” By simple reflexes (Lashley, 1954,p.437). contrast, according to Wimsatt’s usage and in Claims that an entire brain must be this chapter, an emergent property arises not involved are sometimes called “global” theo- from an accumulation of smaller amounts ries. Lycan seemed to express such a view possessed by each part, but from the particu- when he wrote, “The central nervous sys- lar way in which the parts are put together. A tem is as central as it gets” (Lycan, 1997, common example of an emergent property p. 762). More restricted, but non-localized is the ability of a wheel to roll – whether it hypotheses include the complexity theory is made of wood, steel, plaster-of-paris, or of Tonini and Edelman (1998), which pro- clay fired in a kiln, it rolls. The property of poses a highly mobile “dynamic core” of “rollness” is largely independent of the prop- activity whose composition “can transcend erties of the parts, although not entirely; one traditional anatomic boundaries.” When the cannot very well make a wheel out of jello emphasis is on aggregation rather than archi- or oatmeal. tecture, theories are often called “global” If, as I urge here, C is an emergent prop- even when they do not intend that an entire erty of some Mc in the Wimsatt sense, there brain is involved. For example, Kinsbourne may well be restrictions on the sort of parts in his “integrated field theory” maintained that can be assembled for this purpose. At that awareness is a property of a sufficiently present, many of us believe that neurons sizable neural network, “not of any particu- are the essential constituents. An open mind lar locus in the brain” (Kinsbourne & Smith, on this issue requires remembering that real 1974). brains also have an abundance of glial cells Until we have a more widely agreed- and lots of different juices surrounding the upon understanding of how C is produced, neurons. (The wetness of brains is one of the this question remains open. The view in big differences from all known man-made this chapter is that Mc depends upon P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

the thalamic intralaminar nuclei and the property of consciousness 781

an anatomically specifiable arrangement of normally available to awareness (Milner & nerve cells, probably including and inducing Goodale, 1996,p.200). specific chemical environments. A theory emphasizing the role of specialized cells was offered by Jones (1998a). He pointed out that in each thalamus there are cells, staining for the presence of parvalbu- C Depends upon an Anatomically min, that project focally to cortex. Others Localizable Mechanism that he calls matrix cells and that stain for calbindin project more diffusely. He suggests Once we agree that C depends upon brain, that it is the diffusely projecting matrix cells and we want to decide whether to look that contribute to “the binding of all aspects for a global or a local mechanism, a rel- of sensory experience into a single frame- evant question is the following: Are some work of consciousness” (Jones, 1998a, p. 69). brain parts more important than others? Jones’s proposal is discussed in more detail One can approach a physiologic understand- later in the chapter, after a brief discussion ing at levels from the microscopic to the of thalamic anatomy. most encompassing. That is, C might be sup- Some describe consciousness in terms of posed to appear at the subcellular level (e.g., much larger frameworks; one of these is the the microtubules as in Hameroff & Penrose, Extended Reticular Thalamocortical Acti- 1996). Another approach is to attribute C to vating System (ERTAS) of Newman and specialized neurons in cortical clusters; this Baars (1993), which can act as a “global seems implied by the view that processing workplace” or “theatre of consciousness” of information is best described as depend- (Baars, 1997). Baars (1988) emphasized early ing upon a large number of nodes and each on that consciousness of particular data of these nodes is capable of generating its is associated with what might be called own “microconsciousness” (Zeki & Bartels, a “broadcasting” of those particular data. 1999). The idea that consciousness is gener- This aspect of consciousness has now been ated independently in different places is sup- confirmed by fMRI studies showing that ported by the finding that color and motion whereas a stimulus can produce cerebral occurring at the same time can sometimes be activity without the subject’s awareness, perceived at different times (Zeki & Bartels, when the subject does become aware, then 1999). It is hoped that the reader will find cerebral activity appears in many more loca- that this phenomenon is also explicable by tions. This fact about consciousness might thalamocortical interaction as described in suggest that it is the broadcasting of data this chapter. that makes them conscious. By contrast, Special cells or circuits seem implied by the hypothesis advanced in this chapter is the idea that C can accompany the ven- that the broadcasting is not the cause of tral stream but not the dorsal stream. This C, but rather is a consequence of data hav- last suggestion was proposed by Milner and ing acquired C. It may be that deciding Goodale (1996) and reflects the proposal between these two views will require com- that visual information flows forward from bining fMRI with some other procedure hav- the primary visual area in two directions: the ing finer time resolution. ventral stream into the temporal lobe, giving In contrast to much of the foregoing, I rise to knowledge of what an object is, and present below evidence for the view that C the dorsal stream into the parietal lobe, giv- depends upon centrally located, hard-wired ing rise to knowing where it is (Ungerleider circuits that embody what I have called Mc & Mishkin, 1982). Milner and Goodale pre- (Bogen, 1993, 1995a,b, 1997c). That is, Mc is fer to regard the ventral and dorsal streams as supposed to depend upon a specific arrange- serving perception and action, respectively ment of neurons so organized, so centrally (Milner &Goodale, 1996,p.178) and suggest located, and so widely connected to other that processing in the dorsal stream is not brain parts that it can transiently endow a P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

782 the cambridge handbook of consciousness

NAP elsewhere in brain with the property typically do not (Benson, 1994: Damasio & C. Kinsbourne (1995) has criticized this con- Damasio, 1989). cept, facetiously referring to a postulated Mc as “the subjectivity pump”; this term describes precisely what we are trying to Neocortex Provides Content, But Is find. Insufficient to Provide the Property C NAPs that are potentially conscious undoubtedly number in the billions. How- That C is not produced by cerebral cortex ever, only a few can have C at any one was particularly urged by Penfield and Jasper moment. Probably relevant here is the pro- (1954). Their views derived largely from cess of “sparsening” by which complex, dis- observations of epilepsy, including that contributed material can be rapidly and specif- sciousness could be absent during complex ically retrieved using only a fraction of the behavior (requiring neocortex). Conversely, material to be retrieved (Kanerva, 1988), severe disturbances of function either from as concisely explained by Laurent (2002). cortical removals or cortical hyperactivity (as Because Mc, as conceived here, needs to in focal seizures) need not be accompanied reflect only a very small set of NAP at any by loss of consciousness. one moment, Mc need not be very large as As early as 1937, Penfield expressed one would be required in a so-called Cartesian theme of this chapter: “All parts of the brain theatre; that is, “a place where it all comes may well be involved in normal conscious together” (Dennett & Kinsbourne, 1992). processes but the indispensable substratum of consciousness lies outside of the cerebral cortex, probably in the diencephalon.” (The Evidence for Localization of Mc diencephalon includes both thalamus and hypothalamus). It is an interesting fact that The usual localizationist argument involves this view has often been ignored by recent two findings. First, a large deficit in some theorizers on consciousness, many of whom function (f ) is produced by a small lesion are fixated almost exclusively on neocorti- in the “center” or “node” for that f (this cal function. We would do well to keep in does not imply that the representation of mind the point made by Steriade and col- f is wholly contained within some sharp leagues that for cortical information pro- boundary; Bogen, 1976; Bogen & Berker, cessing, “at each step of this information 2002; Bogen & Bogen, 1976). Second, a flow, corticothalamic volleys engage synapti- large lesion elsewhere results in very little cally the intrathalamic networks” (Steriade, if any disturbance of f. A familiar example is Jones, & McCormick, 1997,p.13). These the profound disturbance of speech begin- authors go so far as to refer to the thalamus ning with “a complete linguistic suppres- as a “switchboard” for the cerebrum. More sion” (Lecours, Lhermitte, & Bryans, 1983) recently, Sherman and Guillery (2002)have and often leading to what is called agram- suggested the likelihood that “the thalamus matism, which stems from relatively small sits at an indispensable position for the mod- lesions in and around Broca’s area of the left ulation of messages involved in corticocorti- hemisphere (Henderson, 1990). By contrast, cal processing.” a very large lesion in the right hemisphere or even its complete removal rarely affects the syntactic (grammatic) competence of a Conscious Content Includes Primitive, right-hander. Non-Cognitive Components With respect to C, quite small bithalamic lesions involving both sets of intralaminar To Penfield and Jasper’s reasons that C is nuclei (ILN) typically impair Mc (detailed produced outside the cerebral cortex can be below), whereas very large bicortical lesions added the observation that some important P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

the thalamic intralaminar nuclei and the property of consciousness 783

contents of consciousness are quite pri- cephalon.” They originally described it as mitive; that is, they are unneedful of neocor- “the neurone systems which are symmetri- tical discrimination, association, or learning. cally connected with both hemispheres and Examples are nausea, fatigue, unelabo- which serve to co-ordinate their functions” rated pain (for example, the “electric” (Penfield & Jasper, 1954,p.27). This original jabs of trigeminal neuralgia), thirst, and the description takes in a great deal of anatomy. like. It may be that Mc evolved to give these including the 200 million fiber corpus callo- percepts greater potency, either to facilitate sum. However, in subsequent discussions in learning or to provide another layer of con- that book and in later descriptions (Adrian trol over the stopping of ongoing action. et al., 1954; Eccles, 1966) they made it clear In either case, it would appear that Mc that they had in mind structures in the dien- was only subsequently recruited to serve cephalon, and they particularly stressed the so-called higher functions and more elab- role of the ILN. Why was this concept largely orate responses (Seager, 2002). We under- abandoned? At least three reasons can be stand that Mc of humans routinely endows seen: with C patterns of complex cortical activ- 1. The centrencephalon was supposed to ity describable as “representations of rep- be not only a mechanism for conscious- resentations” or “higher order thoughts” ness but also a source of seizures that were (Rolls, 2000; Rosenthal, 1990). But these “generalized from the start.” The concept are special contents, not the crucial core of of centrencephalic seizures has been largely consciousness. abandoned by epileptologists. However, that Conversely, complex representations can Penfield and Jasper combined these two sometimes be effective without conscious- ideas does not require that we do so; argu- ness. Earlier experiments that attempted ments for a thalamic role in C can be to show that meaningful contents requir- made quite independently of theories about ing neocortical activity can influence be- seizure origin and spread. havior without being conscious have been 2. The centrencephalon was intended subject to a variety of methodological criti- to explain the unification of conscious- cisms (Hollender, 1986; Perruchet & Vinter, ness. However, cerebral commissurotomy 2002). However, considerable evidence has (the split-brain, discussed further below) for most of us established the existence renewed interest in the corpus callosum as of nonconscious perception and cognition an integrating structure and raised doubts (Bar & Biederman, 1998; Berti & Rizzolatti, about the usefulness of the centrencephalon 1992; Castiello, Paulignan, & Jeannerod, concept for explaining consciousness (Doty, 1991; Henke et al., 1993; Jeannerod, 1992; 1975). On the basis of our current knowl- Kihlstrom, 1987; Libet, Gleason, Wright, edge, the problem of localizing Mc can be & Pearl, 1983; Milner & Goodale, 1996; approached in terms of a single hemisphere Ro & Rafal, 1996; Schacter, McAndrews, (which we know can have C), postponing & Moscovitch, 1988; Taylor & McCloskey, to the future the problem of integrating two 1990; Wexler, Warrenburg, Schwartz, & Mc and the implications of this problem for Janer, 1992; Zaidel, Hugdahl, & Johnsen, the unity of consciousness. 1995; Zaidel et al., 1999; see Chapter 9). 3. When Penfield and Jasper emphasized the function of thalamic nuclei, especially the intralaminar nuclei (ILN), objections An Historical Antecedent: to their concept arose because of consid- The Centrencephalon erable doubt concerning the existence of ILN projections to cortex. This doubt arose When Penfield and Jasper (1954) argued because unilateral decortication did not pro- against cerebral cortex as the source of C, duce rapid degeneration in the ILN as it they introduced the concept of a “centren- did in the principal nuclei. However, more P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

784 the cambridge handbook of consciousness

Figure 27.1. This is a stylized drawing of the right, human thalamus as seen from above and behind (the left thalamus is an anatomical mirror image). A thin slice has been taken out about two-thirds of the way back from the anterior (front) end in order to show some of the internal structure. This figure shows what modern anatomists call the “dorsal thalamus” because of its embryonic origin. In reality, the dorsal thalamus is enshrouded by the reticular nucleus (R), a thin shell of cells that are of ventral embryonic origin. The R is not shown in this figure because it would hide the dorsal thalamus from view. Anatomists of the Renaissance were unaware of both the embryology and the microscopic detail; they applied the term “thalamus” to the entire assembly including R, and this is what neurologists usually mean when they use the word thalamus. (The word “thalamus” is Latin for bridal couch; the older anatomists enjoyed a rather randy vocabulary.) The internal lamina is shown in white. (There are other laminae (layers of fibers) outside of the thalamus, but are not relevant here.) The lamina consists of millions of nerve fibers going to or from nearby neurons. To the right of the lamina are the lateral and ventral parts; they are largely concerned with specific sensory or motor functions and are not discussed further in this chapter. Toward the anterior end the lamina bifurcates, with the lateral branch continuing on toward the cortex. The medial branch dives under the collections of neurons to which most of its fibers are headed. Each thalamus (the right thalamus is shown here) contains many subparts called nuclei. There are, in each thalamus, about 40 nuclei according to Jones (and about 80 according to Hassler). We do not need to bother with most of these nuclei (they are unrelated to the text), so we do not show the subdivisions of the lateral and anterior parts; and we do not label most of the nuclei, with four exceptions: MD, PL, LD, and CM. The medial dorsal nucleus (MD) takes up most of the medial thalamus. The MD is almost certainly important for emotion because of its large input from the amygdala of the same hemisphere. And MD is likely important for thoughts about the future, as its output is to the prefrontal cortex; more specifically the large-celled part of MD projects to the orbitofrontal cortex, which is an important part of the so-called emotional brain. On top of MD is the rather thin nucleus called lateral dorsal (LD); it is included by Jones (1985) in his discussion of the Anterior nuclei. These have been implicated in both memory and emotion by various authors and may be involved in other processes because they project to the anterior cingulate gyrus, P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

the thalamic intralaminar nuclei and the property of consciousness 785

recent tracer techniques have shown that findings, some introductory anatomy may ILN do indeed project to cortex and do so be helpful. widely. The centrencephalon concept tried to Simplified Image of a Thalamus explain too much. But it contained a germ of truth, which now needs to be nourished in Within each cerebral hemisphere, close to terms of thalamic contributions to the mech- the midline, sits the thalamus of that hemi- anisms that provide us, as well as creatures sphere. Each human thalamus is about the like us, with consciousness. size and shape of a pecan in the shell, with a bulbous swelling at the back end called the pulvinar. In other words, each thalamus is an ellipsoid about 4 cm long and about 2 cm Different Kinds of Evidence wide (see Fig. 27.1). That Mc Requires the ILN The neurons in each thalamus are in clusters called nuclei. According to Netter Evidence for the importance of thalamic (1953), long a favorite of medical students, nuclei, especially the ILN, is presented here each thalamus contains 11 nuclei. Accord- from clinical neurology, neuroanatomy, elec- ing to Hassler (in Schaltenbrand & Wahren, trophysiology, and other studies, includ- 1977), long favored by neurosurgeons who ing experiments with magnetic resonance operate on thalami, there are about 80 dif- imaging (MRI). Before getting to these ferent nuclei. According to E. G. Jones ← Figure 27.1 (cont.) which itself has been implicated in a wide variety of cognitive, volitional, and emotional processes. Note that, in the human, the LD is neither lateral nor dorsal; it acquired its name from its position in more primitive brains, such as rabbit, cat, or the most primitive primates like the bushbaby. At the posterior end of the dorsal thalamus there is a quite large nucleus, having several subdivisions; it is called the pulvinar (PL). (The word “pulvinar” means pillow.) The pulvinar is connected very widely to almost all of the cortex known to respond to external stimuli of whatever nature; it is particularly large in primates where it is thought to help in determining salience of information, and it therefore plays a role in directing attention. The ILN are shown in black, surrounding the MD; in reality the nerve cells of the ILN are not all in a discrete band, as shown here, because many of them are dispersed within the lamina. The subdivisions of the ILN have been given a variety of names, which are unnecessary in this chapter. Because of their arrangement in a shell around the MD, they were lumped together by Grunthal¨ (1934) as the “nucleus circularis.” This nicely descriptive usage was followed by a half-dozen anatomists subsequently, but is no longer current. The entromedian nucleus (CM), more commonly called centrum medianum in Latin or centre mediane in French, is visible where the slice was taken out. The CM is much larger in monkeys than in most mammals and is even bigger in humans. It projects to the motor cortex, which is one of the reasons for thinking that it is probably essential to volition. It has long been customary to consider CM to be part of the ILN because it is located in the lamina (as shown on the cut surface). Moreover, it has a major output to the striatum, which is a defining characteristic of the ILN. At present, in simple terms, we can think of CM as the motor part of ILN and the rest of ILN (the veil surrounding MD) as sensory; in other words, roughly speaking, this reflects the difference between the intentional and perceptual aspects of C. [This stylized drawing was done by Michael Stern, using Photoshop, with advice from JEB, referring to the coronal sections in Fig. 6 of Jones (1998a) and the coronal sections of Figs. 1–8 in Munkle et al. (2000).] P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

786 the cambridge handbook of consciousness

(1985, 1998a,b), whose views reflect mod- and left MD, and the clusters are merged ern staining techniques, each thalamus has together so they are given a separate name, about 40 nuclei. Thalamic nuclei have been the midline nuclei. In humans in whom the delineated and named in conflicting ways, midline nuclei are largely absent, the two often subject to vociferous argument (Jones thalami are for the most part clearly sepa- & Macchi, 1997; Percheron, 1997). For rated. In addition to the “blanket,” it is cus- some, what mattered was whether groups tomary to include as one of the ILN the of cells seemed circumscribed by fiber bun- centrum medianum (CM), which resembles dles. Also, anatomists gave individual names a small bean up against the blanket just to groups of cells if they seemed to have below the back half of MD. The CM seems the same appearance and were clustered to have a different function as judged by the together more closely. Modern terminol- fact that it stains differently from the rest ogy also gives major importance to connec- of the ILN. It is notably bigger in primates tions, as well as to chemical constituents than in most mammals, and even more so in as shown by various staining techniques. humans, a fact to which I return. Another source of complexity is that even when anatomists are largely agreed on which The Definition of the ILN Has Changed thalamic neurons belong together, they use different names. I follow here the usage of Modern thalamic terminology was greatly Jones (1985, 1998a,b), influenced by LeGros Clark (1932). He For the purposes of this chapter I am not wrote, “Broadly speaking, the thalamic concerned with all the nuclei; most of them nuclei may be divided into two groups: prin- contribute to specific sensory or motor func- cipal nuclei and intralaminar nuclei.” Fur- tions, like vision or hearing or coordination. ther on he wrote, “Among the principal In the older literature these are sometimes nuclei and outlining them are fiber tracts called “relay nuclei”; more recently the terms which form the medullary laminae. These “specific nuclei” or “principal nuclei’ are pre- laminae are strewn with cells [which] form ferred. Of greater relevance for this chap- the intralaminar nuclei” (LeGros Clark, ter are the intralaminar nuclei (ILN), often 1932,p.418). That is, the ILN originally called non-specific; it is the case that com- got their name from their association with pared to the specific nuclei, inputs to the the laminae, whether actually inside them ILN are dispersed more widely, and much or close by outside them. Anatomists have of their output is distributed more diffusely. argued about what nuclei should be so One specific nucleus important for present characterized; most nowadays accept the purposes is the medial dorsal nucleus (MD), view that the ILN are defined mainly by their whose connections from amygdala and with projection to the striatum (Jones, 1985, 1989, prefrontal cortex give it a special role in emo- 1998b). The considerable importance of this tion. Moreover the location of MD helps in projection comes up later with respect to the description of the ILN, as well as sug- volition and procedural memory. First, I con- gesting their hypothetical role in the experi- sider evidence alluded to earlier that Mc encing of emotion. depends on the ILN. The medial dorsal nucleus (MD) is an ellipsoid about the size and shape of a sin- Some Neurologic Evidence: gle Virginia peanut (out of its shell) located Three Clinical Facts inside the pecan. MD is nestled up against the roof of the thalamus against the medial 1. There are two levels in the CNS where wall. The ILN are collections of nerve cells very small lesions (less than 1 g) placed bilat- draped like a blanket all around the MD, top erally can abruptly abolish responsiveness: and bottom, front and back, and all along the These are either in the brainstem reticu- side. In rats, cats, and monkeys similar cells lar formation (BSRF) or in the thalamic also cover the medial surfaces of both right intralaminar nuclei (ILN). These facts are P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

the thalamic intralaminar nuclei and the property of consciousness 787

discussed in more detail, with appropriate for very large damage or removal of an entire references, further on. hemisphere, in excess of 400 g. Corollary: The effect on consciousness of Corollary: The mechanism for conscious- brain damage depends crucially on its loca- ness is paired, existing in duplicate. This con- tion. One of the most widely used textbooks clusion is consonant with the bilateralism of on human neuropsychology states, “There the anatomy and is supported experimen- are no reports that individuals, who have lost tally by the results of hemispherectomy and a certain restricted portion of the brain have the splitbrain. lost their consciousness. The idea that consciousness is a property of a single system or The ILN Subserve Mc: The Widespread brain area receives no support from clinical Connections studies of people who have suffered brain damage” (Kolb & Wishaw, 2000,p.484). Experiments using anatomic tracers have People who want to understand con- shown that with small lesions in almost every sciousness can be led astray by these two cortical region these regions project to ILN. sentences. Perhaps the authors had in mind However, Koch (1995) pointed out that both “cortical system or brain area”; if so they primary visual cortex (V1) and inferotempo- were undoubtedly correct. ral cortex (IT) in the monkey have not been 2. The larger the thalamic lesion, the shown to connect with ILN. There is consid- more widespread the diaschisis (the shock erable controversy over V1-ILN connections; effect that depresses nerve cell function if they are indeed absent, this would be con- elsewhere; Szelies et al., 1991). Also, the sistent with the evidence that processing in larger the lesion, the more long-lasting V1 is inaccessible to consciousness (Koch, the deficit. This follows a basic prin- 2004). A systematic search for ILN-IT con- ciple of behavioral neurology (and of nections or their absence is needed; if they neuropsychology); that is, the location of the are not present, our attention will be increas- lesion commonly determines the nature of ingly focused on the pulvinar because it is the behavioral deficit, whereas the size of widely connected to all of postrolandic cor- the lesion is usually the main determinant tex and is thought to be essential to the of the extent of recovery (Kertesz, 1993). determination of the salience of visual stim- Corollary: The effects on consciousness uli (Robinson & Cowie, 1997). of appropriately placed focal lesions follow Groenewegen and Berendse recently the same rules as the effects produced by reviewed evidence that ILN has effects focal lesions on other brain functions. The more specific than the traditional term “non- fact that responsiveness returns after small specific” might suggest. They concluded that ILN lesions is not evidence against the neces- “the major role of the midline-intralaminar sity of ILN for C (Petit, Rousseaux, Clarisse, nuclei presumably lies in the regulation of & Delafosse, 1981; Smythies, 1997) any more the activity and the ultimate functioning than rapid recovery from a small lesion in of individual basal-ganglia-thalamocortical Broca’s area (Dronkers, 2000; Mohr, 1973, circuits” (Groenewegen & Berendse, 1994, 1976; Moutier, 1908) means that this cortical p. 56). Moreover, “the midline-intralaminar area is irrelevant for aphasia, as once claimed nuclei are positioned in the forebrain circuits by P. Marie (1906). The ability of brain func- like a spider in its web” (Groenewegen & tion to recover significantly (though rarely Berendse, 1994,p.57). ever completely) from structural damage is, The emphases placed on the ubiqui- along with wetness, an important property tous role of thalamus in neocortical func- distinguishing brains from known artifacts. tion by Steriade et al. (1997) and by Sher- 3. So long as the lesion is unilateral, there man and Guillery (2002) were made with will be no loss of consciousness. This is true respect to the processing of specific infor- not only for a small lesion weighing a few mation. A different but related point is that grams in one of the two thalami but even essentially an entire thalamus is engaged P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

788 the cambridge handbook of consciousness

in switching between more general states, pain.” When recording, they found neurons such as arousal, spindling (as in drowsi- responding to visual and auditory as well ness), slow-wave sleep, or seizure activity as somesthetic stimuli. It was emphasized (Contreras et al., 1997). How large num- by Jasper (1966) that such responses only bers of thalamic cells, located in differ- appeared when the stimuli were novel and ent nuclei, can be activated collectively for were less evident with repetition, unlike the such overall state changes depends upon responses in the sensory (ventrobasal) por- several mechanisms. Jones (1998a) includes tion of thalamus that “continue to respond (1) the action on underlying thalamic nuclei faithfully . . . whether the patient is aware of of the reticular nucleus whose cells are the stimuli or not” (Jasper, 1966,p.260). widely interconnected; (2) interconnections When Albe-Fessard and Besson published (which are controversial) among the ILN; their monumental review in 1978, they (3) cortical projections to ILN, which in turn emphasized the polysensory functions of the project diffusely to cortex; and (4) the more ILN although admitting they had “a role recent proposal (Jones, 1998b) that the dif- to play” in the appreciation of pain. By fusion of activity is effected via the diffusely 1980, McGuinness and Krauthamer had con- projecting matrix cells present in all thalamic cluded that the ILN acted not only as a tha- nuclei, principal as well as ILN. The relative lamic pacemaker and as a relay for cortical importance of these mechanisms remains arousal but was characterized also by the unsettled. presence of cells responding to visual, audi- Ascending input to ILN can help explain tory, and somesthetic stimuli. They found C of primitive percepts. The input to ILN further that the suggestion that at least part includes a large fraction of the ascending of the ILN were active in central pain mech- output of the brainstem reticular forma- anisms and in addition acted as a mod- tion, which not only subserves arousal but ulator of both striatal output and input, also contains information about other bod- hence could be considered part of the motor ily states. Other input comes from a phy- apparatus. Similar conclusions were reached logenetically old spinothalamic system and by Schlag and Schlag-Rey (Schlag, 1984) trigeminal complex (both conveying tem- after nearly twenty years of experiments on perature and nociceptive information) and thalamic function. In addition, they estab- from the dentate nuclei in the cerebel- lished a major role for ILN in the control of lum conveying proprioceptive signals. There eye movements and visual attention (Orem, are also ascending inputs to ILN from 1973). deep layers of the periacqueductal gray, Purpura and Schiff (Purpura, 1997) con- substantia nigra, and amygdala with affec- cluded that a wide range of data, “sug- tive information and from the vestibular gest that in the state of wakefulness, the nuclei with information about body position ILN neurons promote the formation of an (Kaufman & Rosenquist, 1985; McGuinness “event-holding” function in the cortex”. An & Krauthamer, 1980; Royce, Bromley, & abundance of neurological, anatomical and Gracco, 1991). functional evidence was subsequently reviewed in Schiff (1999) who wrote, “cortical and subcortical innervations of the ILN place The ILN Subserve Mc: Multiple them in a central position to inﬂuence dis- Functional Roles tributed networks underlying arousal, atten- In addition to observations from clinical tion, intention, working memory, and senso- cases and the use of anatomical tracers, infor- rimotor integration, including gaze control.” mation on the ILN has come from recording and stimulation. By 1966, Ervin and Mark were stimulating the ILN in pain Bilaterality of the Mechanism patients, obtaining diffuse dysphoria and making lesions that (at least for a time) A striking example of loss of cerebral tis- produced “a striking loss of the clinical sue is provided by a total hemispherectomy P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

the thalamic intralaminar nuclei and the property of consciousness 789

also called “hemicerebrectomy” (including the most prominent neurologists of the late cortex, underlying white matter, and basal 19th century to speak of a double brain. ganglia). Of the four patients reported by The rise and subsidence of these views are Austin and Grant (1958), one was stu- thoroughly described by Harrington (1987); porous preoperatively. The other three con- her book was stimulated by the appearance tinued speaking and were “acutely aware” of new evidence for cerebral duality (the of their surroundings throughout the oper- split-brain) as it became well known in the ation, which was done under local anes- 1970s through the work of Roger Sperry and thesia. Whatever the anatomical basis for colleagues. producing C, the anatomy exists in dupli- Because consciousness requires only one cate. That only one of the pair suffices for hemisphere, many 19th-century neurologists C is clear from total hemispherectomy in inferred that there can be a doubling of humans (Austin & Grant, 1958; Bogen et al., C with two hemispheres. This inference 1998; Peacock, 1996; Smith & Burklund, was empirically confirmed in the testing of 1966; Smith & Sugar, 1975). The same con- split-brain cats and monkeys (Sperry, 1961) clusion follows from hemicerebrectomy in as well as humans (Sperry, Gazzaniga, & monkeys and cats (Bogen, 1974; Bogen & Bogen, 1969). How this duality is negoti- Campbell, 1962; Koskoff, Patton, Migler, & ated by an assortment of integrating mecha- Kruper, 1959; Kruper, Patton, & Koskoff, nisms in the intact cerebrum is a fascinating 1971; White et al., 1959). It also follows and important question (Bogen, 1990, 1993, from temporarily anesthetizing a human 1997a; Bogen & Bogen, 1969; Gazzaniga, hemisphere rather than removing it (Lesser, 1970; Iacoboni & Zaidel, 1996; Kinsbourne & Dinner, Luders, & Morris, 1986). Smith, 1974; Landis & Regard, 1988; Regard, Often the term “hemispherectomy” has Cook, Wieser, & Landis, 1994; Sperry, 1968; been applied to operations in humans in Tramo et al., 1995; Trevarthen, 1974; Zaidel which there has been only a hernidecor- & Iacoboni, 2002). But this problem of inter- tication (Kossoff et al., 2003). In addi- hemispheric integration is not the concern of tion, some operations are “functional hemi- this chapter. It is pointed out here to empha- spherectomies’ in which the hemisphere has size the following conclusion: Because an been disconnected from the rest as fully individual needs only one hemisphere to be as possible, but left in situ (Arzimanoglou conscious, our immediate problem of under- et al., 2000). In such cases it is always pos- standing the physiology of consciousness can sible that some function is attributable to be much simplified by restricting our atten- parts left behind, rather than solely to the tion to how C is engendered in someone with residual hemisphere. However, in the cat a single cerebral hemisphere. and monkey experiments described above It is worth emphasizing at this point that the sections were precisely in the midline, it is a mistake to argue against the possibil- so no parts of the extirpated hemisphere ity of an anatomically specifiable mechanism remained. In the human cases of “total hemi- (an Mc) necessary for consciousness on the spherectomy” cited above, the basal ganglia ground that no suitable anatomic structure is were removed with all the cortex, as well present in the middle of the head (Dennett as all the intervening white matter (Austin & Kinsbourne, 1992; Flanagan, 1991). What is & Grant, 1958; Bogen et al., 1998; Smith & evident on any horizontal or frontal section Burklund, 1966; Smith & Sugar, 1975). With of the cerebrum is the duality of any and all no targets remaining for their axons, the tha- candidates; they are all present in pairs. The lamic cells soon atrophy and are resorbed; hemispherectomy data have established that half of the supratentorial space is thus left either member of a pair is, in general, suffi- empty, containing only cerebrospinal fluid. cient for C. If one is looking for a structure Even before hemispherectomies had been located “in the middle” of the hemisphere, performed, there had been observed natu- one obvious place is the thalamus of that rally occurring cases with resorption of an hemisphere. Many people are so accustomed entire hemisphere; these cases led some of to speak of “the” thalamus or “the” amygdala P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

790 the cambridge handbook of consciousness

or “the” limbic system that they easily forget bilateral damage was accompanied by the that there are two, one in each hemisphere. abrupt loss of responsiveness. We should not let our thinking about the Emergence from unresponsiveness after physiology be misdirected by long-standing, incomplete bithalamic lesions depends on convenient figures of speech. the size of the lesion, taking only a few hours if the lesions are tiny, as in case 4 of Graff-Radford et al. (1990). When the The ILN Subserve Mc: Results of lesions are larger, it takes longer for Bithalamic Paramedian Strokes the patient to arouse. With larger lesions, Support of the proposal that ILN sub- the recovery is commonly accompanied by serve Mc includes the results of thalamic mental impairments variously described as obstructive strokes involving the thalam- confusion, dementia, amnesia, and/or hyper- operforating (paramedian) arteries (Tatu, somnia. Which of these impairments dom- Moulin, Bogousslavsky, & Duvernoy, 1998). inates depends on the precise lesion site Simultaneous bimedial thalamic damage can and size (Abe, Yokoyama, & Yorifuji, 1993; occur because the medial parts of both Bassetti et al., 1996; Bewermeyer et al., 1985; thalami are occasionally supplied by a sin- Bogousslavsky & Caplan, 1995; Bogous- gle arterial trunk that branches, with one slavsky et al., 1991; Gentilini, De Renzi, branch to each thalamus. If the trunk is & Crisi, 1987; Graff-Radford et al., 1990; occluded before it branches, both thalami Guberman & Stuss, 1983; Malamut et al., will be affected. When there is simultaneous, 1992; Markowitsch, Cramon, & Schuri, partial damage of the two sets of ILN, unre- 1993; Meissner, Sapir, Kokmen, & Stein, sponsiveness typically ensues (see Table 2 1987; Michel et al., 1982; Mills & Swanson, of Guberman & Stuss, 1983). Sudden onset 1978; Plum & Posner, 1985; Schott et al., of coma can occur even when the lesions 1980; Yasuda et al., 1990). are only a few cubic centimeters in vol- With quite large bilateral lesions the ume, as in case 4 of Graff-Radford, Tranel, patients usually remain in coma until they Van Hoesen, and Brandt (1990). This is in die. However, a patient with extensive ILN contrast to retention of responsiveness with lesions may survive for many years, if med- very large infarctions elsewhere. Even a quite ically supported, in a persistent vegetative large lesion involving one (and only one) tha- state. That is, such a patient remains unre- lamus rarely if ever causes coma (Plum & sponsive in spite of relatively intact brain- Posner, 1985). stem and cerebral cortex, as in the case Correlating thalamic damage with behav- (widely publicized at the time) of Karen Ann ioral deficit is more precise with strokes from Quinlan (Kinney et al., 1994). occlusion of blood supply than from bleed- ing or from tumors, so that emphasis is given here to cases of occlusion. Shortness of Breath Classical autopsy examples of bithalamic occlusion often had extension of the dam- Among the most primitive of subjective sen- age downward into subthalamus (Castaigne sations is the experience of shortness of et al., 1981;Façon, Steriade, & Wertheim, breath (SOB), sometimes called breathless- 1958). Many cases carried uncertainties ness or air hunger. The usual cause of SOB inherent in now outdated imaging tech- is increased CO2 in the blood, whether from niques. More recently, published cases typ- exertion, inadequate ventilation, lung dis- ically show crisp MRI images of small ease, or direct introduction of CO2 into the lesions whose extent is well demonstrated subject’s inspired air. Extensive investiga- (Tatemichi et al., 1992; van Domburg, tions have made it clear that SOB does not ten Donkelaar, & Noterman, 1996; Wiest, depend on feedback from respiratory mus- Mallek, & Baumgartner, 2000); in these cases cles or the rib cage; it is directly related to it has again been reported that onset of the “respiratory drive”; that is, the frequency of P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

the thalamic intralaminar nuclei and the property of consciousness 791

neuronal firing in the respiratory centers of This change has subsequently come to the medulla (Eldridge & Chen, 1996). The be called arousal of the EEG because of amount of respiratory drive can be mea- its common coexistence with behavioural sured by recording from medulla, but is mea- arousal. When Bremer (1935, 1936) tran- sured more conveniently by recording from sected the neuraxis at C1, the top end of the the phrenic nerve. Because the sensation of spinal cord, he found that the isolated brain SOB does not stem from bodily sensations (encephale´ isole)´ alternated between HVLF (SOB occurs even when these sensations and LVHF and the cat’s pupils altemated are interrupted as in spinal cord injury), its between small, as in sleep, and dilated as in likely source is a corollary discharge travel- the aroused animal. In other words, the cat’s ing upward from the medulla. A search of brain alternated between sleep and wakeful- the midbrain and of thalamus has disclosed ness in spite of being disconnected from its neurons whose activity is closely correlated body below the head and in spite of the head with phrenic nerve activity, as CO2 levels being fixed in a frame so the EEG could be are manipulated. In the thalamus, the neu- recorded. From my own personal experience rons that become active with increased CO2 with the encephale´ isole,´ I can also add that appear to be in the ILN (Eldridge & Chen, when the cat is awake, as judged by EEG cri- 1996). teria and pupil size, it follows with its eyes When CO2 is elevated, the pulse rate and objects or persons moving around the room. respiratory rate rise along with increased res- By contrast, Bremer found that after piratory effort. These compensatory reac- midbrain transection the isolated cerebrum tions occur automatically. So what advan- (cerveau isole)´ stayed in the HVLF state. tage is there in also having a subjective That is, without the brainstem the cere- experience? The answer is not altogether brum was apparently unarousable. Subse- clear, but likely has to do with encourag- quently, Moruzzi and Magoun (1949) found ing the individual to change whatever else that stimulation within the brainstem retic- is going on. ular formation produced the EEG arousal. Other questions remain, as the experi- As was pointed out above, small (bilat- ments of Eldridge and Chen lacked histo- eral) lesions in the BSRF can cause imme- logical controls and have not been repli- diate unresponsiveness. Together with rele- cated to my knowledge. When they are vant autopsy studies of coma in humans, this repeated, it would be of interest to know if line of research led to the widely accepted the ILN neurons correlated with SOB are notion of an ascending reticular activating those already known to be responsive to system (ARAS; Magoun, 1952; O’Leary & pain, visual, auditory, and somesthetic stim- Coben, 1958). uli, especially when these stimuli are novel By 1960, it was generally accepted that (Jasper, 1966). (1) for consciousness, the cerebrum must remain attached to the upper end of the brainstem and that (2) this fact depended Ascending Activation upon ascending activating influences from the BSRF. For several decades, in the 1970s It is important to distinguish Mc from the and 1980s, the concept of an ARAS was var- ascending activating influences arising from iously criticized (Brodal, 1969; Jouvet, 1967) several sources in the brainstem reticular for- and thought by some to be passe.´ How- mation (BSRF). First, some background. ever, newer evidence in the 1980s and 1990s, It has been known since work conducted much of it from Steriade and colleagues, by Berger (1930) that transitions from resting enabled Steriade (1996) to say that the con- to behavioral alertness are reflected in the cept of ascending activation from the BSRF electroencephalogram (EEG) as transitions “has been rescued from oblivion.” Further- from high-voltage low frequencies (HVLF) more, it now appears that the BSRF stim- to lower-voltage higher frequencies (LVHF). ulation that arouses the EEG can facilitate P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

792 the cambridge handbook of consciousness

synchrony in the gamma range (20–70 the striatum, which suggests a role of ILN spikes/s: Munk et al., 1996). This is impor- in the control of motor output, as discussed tant because synchrony in the gamma range further below. has been thought to underlie aspects of cog- Fourth, activation of cortex occurs inde- nition, so this ﬁnding ﬁts the idea that arousal pendently of ILN as evidenced by rare cases facilitates thinking. In addition, an imag- of nearly complete destruction of the ILN ing study found that procedures that called but relative sparing of cortex; the result- for focusing attention caused increased acti- ing vegetative state includes cycling between vation of both BSRF and the thalamic “sleep” and “wakefulness” (i.e., states with ILN (Kinomura, Larsson, Gulyas, & Roland, and without cortical arousal; Kinney et al., 1996). 1994; Schiff & Plum, 1999).

The ILN Are Not Simply Relays Thalamic Contribution to Attention for Activation Consideration of the relation of conscious- When the concept of ascending activation ness and attention is made difficult by the was first popular, the arrival at cortex of acti- various usages of both words. Attention vating influences appeared attributable to an can be synonymous with an orientation “unspecific thalamocortical system” includ- of the head and eyes to a brief stimulus, ing the ILN (Jasper, 1954). Since then, it has attributable to midbrain function and occur- often been supposed that the function of the ring well before awareness of the stimulus. ILN is best understood as relaying to cortex What we see may or may not correspond to the activation originating in the brainstem. the “spotlight of attention” (He, Cavanagh, This appeared similar to the “relay” function & Intriligator, 1997; Jung, 1954). Or atten- attributed to the specific nuclei. Perhaps the tion may refer to an individual’s keeping most explicit expression of this is the ERTAS some location “in mind” while fixating vision advocated by Newman and Baars (1993). elsewhere (Treue & Maunsell, 1996) as evi- However, considering the ILN to act princi- denced, for example, by a faster response pally as relays for activation seems incorrect to that location than to others. It is not for several reasons. surprising then, that multiple mechanisms First, the view that thalamic nuclei act have been implicated in attention, including mainly as relays (and are often labeled “relay an assortment of cortical regions (Posner & nuclei” in introductory texts) is no longer Rothbart, 1992). tenable (Guillery & Sherman, 2002). In this Almost completely surrounding each tha- connection, I note that only 10 to 20%of lamus is a thin layer of cells called the LGN synapses are from the eyes (LGN is reticular nucleus (which is called R, RE, the visual “relay” nucleus of thalamus). This Ret, Rt, nRt, RTN, or TRN depending on means that information arriving from the the author). It has long been suggested eyes is subjected in LGN to modification that R provides part of the physiologic from a variety of sources. This includes fibers basis for selective attention (Crick, 1984; from visual cortex back to LGN that are Guillery, Feig, & Lozsadi, 1998; Jones, 2002; about ten times as numerous as the fibers Mitrofanis & Guillery, 1993; Scheibel & carrying information from LGN up to visual Scheibel, 1966; Yingling & Skinner, 1976, cortex (Sherman & Koch, 1986). 1977). Briefly, a physiologic mechanism for Second, the internal organization of the attention is ascribed to R for several reasons: ILN implies selective processing, a conclu- (1) Each R envelops, in a thin layer, most sion consonant with their diverse inputs of the thalamic nuclei. The thalamocortical from all sensory modalities. fibers, as well as the fibers returning from Third, the output of the ILN is not only cortex to thalamus, when passing through to cortex, but even more prominently is to R, give off collaterals as they pass through; P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

the thalamic intralaminar nuclei and the property of consciousness 793

(2) R efferents terminate in the immedi- of neocortex; it is this diffuse projection that, ately underlying thalamic nuclei; and (3) I argue here, has to do with C. One can R efferents are GABAergic (the neurons in understand how ILN could directly influ- R are exclusively inhibitory, using GABA ence ideation, as ideation is a function of as their transmitter). The likelihood exists, cortex. This implies that awareness of con- therefore, that thalamocortical communica- tent depends upon some as yet unspecified tion can be simultaneously inhibited over- “appropriate interaction” between ILN and all with highly selective non-inhibition. Such the neural representation of that content. localized gating could provide a mechanism The possibility has been raised that what for selective attention in cognition. How R I call appropriate interaction involves syn- selectively gates is clearly affected by the chronization of neuronal activity at 40 Hz: fibers from cortex ending in R, which may Stimulus-dependent synchronization of cor- help explain top-down control of atten- tical cells with ILN cells at 40 Hz could be an tion. Bottom-up effects on attention may example of C plus content (Steriade, Curro involve several mechanisms, but one prob- Dossi, & Contreras, 1993). ably depends upon the collaterals, ending As an example of appropriate interaction in R, of thalamic projection to cortex from between ILN and a specific cortical area, the sensory nuclei of the thalamus. In addi- we can consider awareness of the direction tion to the focal gating, R can generally of motion of a stimulus. It is now widely inhibit all the other thalamic nuclei dur- understood that motion direction informa- ing slow-wave sleep. When the ascending tion (MDI) is represented in cortex of the activation from BSRF increases, it has a superior temporal sulcus (STS), especially widespread inhibitory effect throughout R; area V5, also called MT (Allman & Kass, this suppresses the inhibition exerted by 1971; Maunsell & Newsome, 1987; Rodman, R and momentarily opens up widespread Gross, & Albright, 1990; Zeki, 1974, 1993). (ungated) communication from thalamus According to the present hypothesis, for the to cortex. As Guillery and Harting (2003) MDI to have a subjective aspect (i.e., to wrote, “Whereas [brainstem and basal acquire C) there must occur the appropri- forebrain] afferents can be expected to have ate interaction between STS neurons and global actions on thalamocortical transmis- Mc. Keep in mind that the MDI in STS sion, relevant for overall attentive state, might well be available for adaptive behavior the [topographically focal corticothalamic whether or not it acquires C. input] will have local actions, modulating In the neurally intact individual, the transmission through the thalamus to cortex appropriate interaction can be on, or off, with highly specific local effects.” or in between and is quickly adjustable. However, when V1 (striate cortex) has been ablated, the appropriate interaction for Appropriate Interaction vision typically does not occur. That is, the MDI in STS is not available to verbal output There are three main types of thalamic (the individual denies seeing the stimulus). projections: the specific (as for vision, audi- At the same time, the MDI may be available tion), the diffuse (characteristic of the for some other behavior. This is an example matrix cells according to Jones as men- of blindsight (Grusser¨ & Landis, 1991; Ptito, tioned above), and the projection to stria- Lepore, Ptito, & Lassonde, 1991; Stoerig & tum (essentially all from ILN). The striatal Cowey, 1995, 1997; Weiskrantz, 1986, 1997). projection is discussed below with respect When we accept availability to verbal output to volition. The first type, specific projec- as the index of C (which we commonly do) tion, conforms more to the concept of the it appears that the appropriate interaction sensory relay of information to cortex. As between STS and ILN cannot readily occur to the second type, diffuse ILN efferents are without an essential influence from striate widely though sparsely distributed to most cortex (V1). An explanation for the fact that P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

794 the cambridge handbook of consciousness

C usually does not accompany the process- flow) done by Laureys et al. (2000) showed ing of visual information in the absence of thalamocortical disconnection as the cause V1 is not yet established (Weiskrantz, 1997). of a vegetative state. These authors con- However, the excitatory drive of the extras- cluded that “restoration of consciousness triate regions by striate cortex may be the seems to be paralleled by the resumption of essential ingredient; more feeble input from the functional relation between thalami and subcortical sources (such as pulvinar) can associative cortices.” The next sections dis- enable some low-level processing by MT cuss interference with the ILN-cortex con- without awareness (Koch, 2004). nections that are physiologic (functional) When visual processing in the absence of rather than anatomic. V1 is accompanied by C (Stoerig & Barth, 2001), this may indicate stronger than usual Loss of Function from Unbalanced input from pulvinar. Alternatively, Zeki and Inhibition Bartels (1999) consider visual C without participation of V1 to be evidence for their the- An important principle of neuropsychology ory, mentioned above, of multiple microcon- is that loss of some ability is not neces- sciousnesses. sarily the result of destruction of a brain part essential to that ability. A fairly common example is the loss of a stroke patient’s Interruption of Access to C ability to look toward the side of a cortical lesion; this deficit sometimes subsides in a This chapter’s hypothesis suggests that for few hours, but may take much longer. The some content represented by NAPs in cortex problem is not primarily in the midbrain cir- to be endowed with C, there must be a con- cuits controlling eye movement; the prob- nection between the ILN and the region(s) lem is that these circuits are receiving from of cortex containing those NAPs. Years ago one hemisphere influences that are no longer it was suggested that unawareness of one’s balanced by those from the damaged hemi- bodily derangement (called anosognosia) sphere. There are multiple influences, both could be caused by an interruption between inhibitory and facilitatory. thalamus and parietal cortex (Sandifer, As some of these influences subside, the 1946). The disconnection of ILN from cor- ability to gaze in either direction returns. tex should result in the unavailability to con- Recovery from a brain injury results from sciousness of some content; the more exten- a variety of processes; one of these is a sive the disconnection, the greater the loss of rebalancing of the facilitatory and inhibitory content. An extreme example would be the influences. When performance has been lost vegetative state in which there is eye opening because the competence has lost some facil- to minimal stimuli, cycling between sleep- itation, re-emergence of the performance ing and waking states, and preservation of can result simply from the subsidence of many brainstem functions but little if any inhibition (Sherrington, 1932). Much the evidence of mental function (Plum & Pos- same suggestion was made by von Monakow ner, 1985). The undetectability of mental (1911). function is accompanied by loss of metabolic The main point is that a loss of perfor- activity and decreased blood flow in the cere- mance is not necessarily the result of dam- brum. Although the vegetative state may age to the competence for that performance; follow widespread cerebral damage it can it may result from unbalanced or excessive also appear either from damage solely to the inhibition of the competence. thalami (Szirmai, Guseo, & Molnar, 1977) The concept of an appropriate interaction or from interruption of thalamocortical con- between ILN and a NAP elsewhere implies nections with little if any damage to the cor- that C of some representations can be lost tex (Jennett, 2002; Kinney et al., 1994; Schiff not only when the representations are dam- et al., 2002). A PET study (measuring blood aged or when anatomical connections are P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

the thalamic intralaminar nuclei and the property of consciousness 795

severed. In addition, contents (represented also of any awareness at all. We understand by various NAPs) that are ordinarily avail- that most bodily adaptations (including pos- able to consciousness can be unavailable tural adjustments, sensorimotor coordina- because the interaction is inhibited. An tion, phoneme generation during speech, as example may be found in some cases of pari- well as most autonomic regulation) not only etal hemi-neglect (Bogen, 2003). can proceed independently of C but, more often than not, are also unavailable to C. The foregoing contrasts with what I call The Conviction of Volition “deliberate” actions. And some activities are That the basal ganglia (BG) are crucial in only intermittently volitional; probably the controlling motor function is well known. commonest example of incessant activity for Many of the relevant facts come from which we occasionally do take responsibility studies of Parkinsonism: “Damage to the is breathing. What is the physiological differ- striatum results in major defects in vol- ence among these actions? untary movement” (Graybiel, Aosaki, Fla- Velmans (1991) and Gray (1995)have herty, & Kimura, 1994). The BG are par- described a range of processes in which ticularly involved in self-initiated movement awareness follows rather than precedes the (Taniwaki et al., 2003). information processing. A famous experi- The prominent connections to ILN from ment (Libet, 1993, 2003; Libet et al., 1983) globus pallidus suggests a monitoring of showed that an intent to act develops, as motor systems, as do the cortical projec- evidenced by development of the readiness tions to ILN from sensorimotor and pre- potential (RP) several hundred milliseconds motor cortex. A role for ILN in the con- before the subject decides to act, as noted trol of motor output can be inferred from by the subject who is watching an indica- the very substantial projections from ILN to tor sweeping a large clock face marked in striatum (Sadikot, Parent, Smith, & Bolam, milliseconds. This experiment of Libet has 1992; Sidibe´ & Smith, 1996). We can ask, been furiously debated for more than 20 Is the ILN projection to striatum a pathway years, sometimes regarding the method but for the inhibition (or release from inhibition) mainly with respect to the interpretation. of motor plans that have been developing Much of the argumentation has stemmed for several hundred milliseconds? Is this the from suspicions that the experiment has, or basis for a “volitional” decision? has not, some significance for certain meta- Closely related to or synonymous with physical issues; these debates do not relate “volitional” are such adjectives as voluntary, to my suggestion here that the experiment discretionary, conative, spontaneous, inten- tells us something about the conviction of tional, deliberate, and the like. We need not volition. The methodological criticisms have consider here how these concepts differ, nor probably been laid to rest by replication of mire down in the centuries-old problem of the original findings (Keller & Heckhausen, free will versus either theological or materi- 1990; Trevena & Miller, 2002). Trevena and alistic determinism. We need recognize only Miller (2002) did point out that, although the fact that each of us has a conviction of the RP started earlier, what is called the lat- volition that we attach to some of our acts eralized readiness potential (LRP) appeared (those for which we feel responsible), but at the time indicated by the subjects, prob- not to some other acts, and that our con- ably concordant with activation of motor viction of volition has a neurophysiological cortex (their article is part of a thorough explanation. It helps to consider first acts discussion in Consciousness and Cognition that are non-volitional. A common example 11(2)). The discussion largely concluded with of action for which people do not consider the statement by Pockett (2002) who was themselves responsible is the pupillary con- in many respects critical of Libet; however, striction to light. Many activities carry on in she wrote, “Thus by no means all of the the absence not only of voluntary control but time difference between the start of the P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

796 the cambridge handbook of consciousness

RP and the consciousness of initiating an sonant with the common dictionary defini- action can be accounted for by measure- tion of affect as the conscious aspect of emo- ment errors. Clearly there is at least some tion. Attributing Mc to the ILN provides a form of unconscious preparation going on tangible mechanism for the conscientiation well before a subject consciously initiates a of emotion (i.e., turning a non-conscious voluntary action” (Pockett, 2002,p.323). emotion into an affect). Before considering The foregoing does not mean that intent the anatomy, it helps to comment briefly on plays no role in what happens because, the point that being conscious of emotions although conscious intent appears well after may be more fundamental than being con- the intent to move has been developing, scious of ideas (Damasio, 1999; Panksepp, it appears some 150 ms before the action 1998; Seager, 2002; Watt, 2000). (pushing of a button). There exists, there- The C of nausea, thirst, fatigue, and the fore, time for the subject either to stop like is at least as typical of being conscious the process or allow it to continue. More- (having qualia) as having C of time, place, over, motor plans can be voluntarily for- person, meanings, memories, and expecta- mulated and stored, the storage probably tions. In all likelihood C of emotions ante- involving the motor areas in medial frontal dated C of cognition in both phylogeny cortex (Libet, 1993; Passingham, 1993; Tanji and ontogeny. An excessive emphasis on C & Shima, 1994). Stored motor plans can attending cognition by many current stu- be subsequently released by a triggering dents of consciousness may derive, as do stimulus that acquires C after the action so many attitudes, from Descartes who has been initiated. Indeed, a stored action claimed, “I think, therefore I am.” A neuro- may be triggered by stimuli that never are biological understanding of C might come “perceived”; that is, never acquire C (Taylor more quickly if we were to aver, I feel, & McCloskey, 1990). therefore I am (Bogen, 1997b; Prof. T. Sul- The proposal here is that a motor livan has recently brought to my atten- plan develops over time and that partway tion that when Descartes said “thinking,” he through this development or while it is meant to include feelings as well as thoughts. held in readiness but not yet executed, we However, the idea that Descartes meant to become aware of it (Libet, 1993). We can exclude emotions seems to be a venera- readily suppose that appropriate interaction ble misunderstanding; see Descartes, 1953, between premotor cortex and ILN occurs p. 374, where Principle 9 is entitled, “Ce que early enough (i.e., 150 ms) before action so c’est penser”). that the self is associated with the develop- Looking for the mechanism endowing ing motor plan. There would thus be time for emotions with C means finding connections ILN to exert an inhibition that would stop between structures subserving C and struc- the action or to overcome any tonic inhibi- tures subserving emotion. The latter are, tion and thus allow the action. If the motor roughly speaking, the constituents of the plan is permitted to run to completion, self limbic system; these include the amygdala, or me-ness would be associated with the orbital cortex, and, connected to both of action, and the individual would feel respon- these, the medial dorsal nucleus of the tha- sibility precisely because there had been an lamus. As pointed out above, the ILN are opportunity to abort the plan. draped over and around the MD like a blanket. This relationship is particularly well demonstrated by staining for the compound Endowing Emotions with C calretinin, which is largely restricted to the ILN (Cicchetti, Lacroix, Beach, & Parent, The present hypothesis, including that a 1998; Fortin, Asselin, Gould, & Parent, 1998; NAP can be converted to CNAP by Mc, Munkle, Waldfogel, & Faull, 1999, 2000). So implies that the NAPs representing emo- striking is this relation that one is tempted tions can exist (and cause behavioral signs of to think of it as the anatomical basis for the emotion) while non-conscious. This is con- statement, “I feel, therefore I am.” P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

the thalamic intralaminar nuclei and the property of consciousness 797

Importance of C for Episodic Memory 2. Our search for that basis will be facilitated by making a sharp distinction between Understanding the relations between mem- the property of consciousness (here called ory and consciousness can help us under- C) and the contents of consciousness, stand better both of these (Schacter, 1989; which vary widely from time to time, Tulving, 1985). One function of C is to facil- from one individual to the next, and from itate certain kinds of learning. We can con- one species to another. sider as examples two forms of learning: 3. The property C depends upon a specific trace conditioning (TC) of the blink reflex mechanism (here called Mc) that evolved to a puff of air and the delayed non-match- to enable the transient association of to-sample (DNMS). The former is relatively self with (a small fraction at any one simple and has been studied extensively in time) of the thoughts and feelings gener- rabbits as well as humans (Clark & Squire, ated in mammalian (and possibly other) 1998). The DNMS is more demanding; it brains. has mainly been studied in rats as well 4. Proper function of Mc requires a state of as monkeys (Dudai, 2002; Eichenbaum & arousal ordinarily dependent upon activa- Cohen, 2001). Both TC and DNMS are tion ascending from the brainstem retic- dependent on neuronal circuits that include ular formation. the entorhinal cortex (ento), which is the 5. In mammals the Mc exists in duplicate. principal bottleneck for information to and 6. Mc crucially includes the intralaminar from the hippocampus. C is maintained by nuclei of the thalami. structures other than the entorhinal cortex because C persists after removal of entorhinal cortex. If it is correct that both TC and DNMS require consciousness, then they evi- References dently depend not only on ento but also on a connection between ento and other Abe, K., Yokoyama, R., Yorifuji, S. (1993). Repet- structure(s) important for C. A likely can- itive speech disorder resulting from infarcts in didate is the “blanket” of ILN. This follows the paramedian thalami and midbrain. Journal not only because of the evidence presented of Neurology, Neurosurgery, and Psychiatry, 56, 1024 1026 in this chapter that the ILN are essential – . to C but also because it is known that Adrian, E. D., Bremer, F., & Jasper, H. H. (Eds.). 1954 damage to the ILN impairs DNMS (Mair, ( ). Brain mechanisms and consciousness. 1994). So far as I am aware, no one has yet Springfield, IL: C. C. Thomas. examined whether lesions of the ILN would Albe-Fessard, D., & Besson, J. M. (1978). Conver- prevent trace conditioning. These facts also gent thalamic and cortical projections, the non- suggest that both TC and DNMS could specific system. In A. Iggo (Ed.), Handbook of sensory physiology: Somatosensory system. New be experimentally impaired by lesions of York: Springer. the tracts connecting the entorhinal cor- Allman, J. M., & Kass, J. H. (1971). A represen- tex with ILN while sparing both termini 2001 tation of the visual field in the caudal third of (Bogen, ). the middle temporal gyrus of the owl monkey (Aotus trivirgatus). Brain Research, 31, 85–105. Allport, A. (1988). What concept of con- Summary of the Thalamic ILN sciousness? In A. J. Marcel & E. Bisiach (Eds.), Consciousness in contemporary science. Oxford: and the Property C Clarendon Press. Arzimanoglou, A. A., Andermann, F., Aicardi, This chapter includes a series of sugges- J., Sainte-Rose, C., Beaulieu, M. A., Ville- tions: mure, J. G., Olivier, A., & Rasmussen, T. (2000). Sturge-Weber syndrome: Indications 1. An essential aspect for the understanding and results of surgery in 20 patients. Neurol- of consciousness is its physiologic basis. ogy, 55, 1472–1479. P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

798 the cambridge handbook of consciousness

Austin, G. M., & Grant, F. C. (1958). Physiologic Bogen, J. E. (1990). Partial hemispheric indepen- observations following total hemispherectomy dence with the neocommissures intact. In C. in man. Surgery, 38, 239–258. Trevarthen (Ed.), Brain circuits and functions Baars, B. J. (1988). A cognitive theory of conscious- of the mind. Cambridge: Cambridge University ness. Cambridge: Cambridge University Press. Press. Baars, B. J. (1993). How does a serial, inte- Bogen, J. E. (1993). The callosal syndromes. In grated and very limited stream of consciousness K. M. Heilman & E. Valenstein (Eds.), Clini- emerge from a nervous system that is mostly cal neuropsychology (3rd ed., pp 337–381). New unconscious, distributed, parallel and of enor- York: Oxford University Press. mous capacity? In G. Broch & J. F. Marsh (Eds.), Bogen, J. E. (1995a). On the neurophysiology Experimental and theoretical studies of conscious- of consciousness: II. Constraining the semantic ness. New York: John Wiley and Sons. problem. Consciousness and Cognition, 4, 137– Baars, B. J. (1997). In the theatre of consciousness. 158. New York: Oxford University Press. Bogen, J. E. (1995b). On the neurophysiology of Bar, M., & Biederman, I. (1998). Subliminal visual consciousness: I. An overview. Consciousness priming. Psychological Science, 9, 464–469. and Cognition, 4, 52–62. Bar, M., & Biederman, I. (1999). Localizing Bogen, J. E. (1997a). Does cognition in the discon- the cortical region mediating visual awareness nected right hemisphere require right hemi- of object identity. Proceedings of the National sphere possession of language? Brain and Lan- Academy of Sciences USA, 96, 1790–1793. guage, 57, 12–21. Bassetti, C., Mathis, J., Gugger, M., Lovblad, K. Bogen, J. E. (1997b). The crucial central core O., & Hess, C. W. (1996). Hypersomnia follow- of consciousness does not require cognition. ing paramedian thalamic stroke: A report of 12 Proceedings of the Society for Neuroscience, 23, patients. Annals of Neurology, 39, 471–480. 1926. Benson, D. F. (1994). The neurology of thinking. Bogen, J. E. (1997c). Some neurophysiologic New York: Oxford University Press. aspects of consciousness. Seminars in Neurol- 95 103 Berger, H. (1930). Uber¨ das Elektrenkephalogram ogy, 17, – . des Menschen, II. Archives of Psychology and Bogen, J. E. (1998). My developing understand- Neurology (Lpz), 40, 160–169. ing of Roger Wolcott Sperry’s philosophy. Neu- 1089 1096 Bergman, T. J., Beehner, J. C., Cheney, D. L., & ropsychologia, 36, – . Seyfarth, R. M. (2003). Hierarchical classiﬁca- Bogen, J. E. (2001). An experimental disconnection by rank and kinship in baboons. Science, tion approach to a function of consciousness. 302, 1234–1236. International Journal of Neuroscience, 111, 135– 136 Berti, A., & Rizzolatti, G. (1992). Visual process- . ing without awareness: Evidence from unilat- Bogen, J. E. (2003). Is TPO hemineglect the result eral neglect. Journal of Cognitive Neuroscience, of unbalanced inhibition? International Journal 345–351. of Neuroscience, 114(6), 655–670. Bewermeyer, H., Dreesbach, H. A., Rackl, A., Bogen, J. E., & Berker, E. (2002). Face module, Neveling, M., & Heiss, W. D. (1985) Presen- face network. Neurology, 59, 652–653. tation of bilateral thalamic infarction on CT, Bogen, J. E., Berker, E., van Lancker, D., Sudia, MRI and PET. Neuroradiology, 27, 414–419. S., Lipstad, B., Sadun, A., & Weekes, N. (1998). Block, N. (1995). On a confusion about a function Left hemicerebrectomy: Vision, olfaction and of consciousness. Behavioral and Brain Sciences, mentation 45 years later. Proceedings of the Soci- 173 18, 227–247. ety for Neuroscience, 24, . Bogen, J. E. (1974). Hemispherectomy and the Bogen, J. E., & Bogen, G. M. (1969). The other placing reaction in cats. In M. Kinsbourne & side of the brain. 3. The corpus callosum and W. L. Smith (Eds.), Hemispheric disconnection creativity. Bulletin of the Los Angeles Neurology and cerebral function. Springﬁeld, IL: C. C. Society, 34, 191–220. Thomas. Bogen, J. E., & Bogen, G. M. (1976). Wernicke’s Bogen, J. E. (1976). Hughlings Jackson’s hetero- region. Where is it? Annals of the New York gram. In D. Walter, L. Rogers, & J. M. Finzi- Academy of Sciences, 280, 834–843. Fried (Eds.), Cerebral dominance. Los Angeles: Bogen, J. E., & Campbell, B. (1962). Recov- UCLA BIS Conference Report. ery of foreleg placing after ipsilateral frontal P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

the thalamic intralaminar nuclei and the property of consciousness 799

lobectomy in the hemicerebrectomized cat. of spindle oscillations in cortex and thalamus. Science, 135. Journal of Neuroscience, 17, 1179–1196. Bogousslavsky, J., & Caplan, L. (1995). Stroke Crick, F. (1984). Function of the thalamic reticu- syndromes. Cambridge: Cambridge University lar complex: The search light hypothesis. Pro- Press. ceedings of the National Academy of Sciences Bogousslavsky, J., Regli, F., Delaloye, B., USA, 81, 4586–4590. Delaloye-Bischof, A., Assal, G., & Uske, A. Crick, F. (1994). The astonishing hypothesis: (1991). Loss of psychic self-activation with The scientific search for the soul. New York: bithalamic infarction. Neurobehavioural, CT, Scribner’s. MRI and SPECT correlates. Acta Neurologica Crick, F., & Koch, C. (1990). Towards a neurobio- 309 316 Scandinavica, 83, – . logical theory of consciousness. Seminars in the Bremer, F. (1935). Cerveau isole´ et physiologie du Neurosciences, 2 , 263–275. sommeil. Comptes Rendus Societe de la Societe Crick, F., & Koch, C. (2000). The problem of con- 1235 1242 de Biologie, 118, – . sciousness. Scientific American, 12, 10–17. 1936 Bremer, F. ( ). Nouvelles recerches sur le Damasio, A. (1999). The feeling of what happens: mecanisme du sommeil. Comptes Rendus Soci- Body and emotion in the making of consciousness. 460 464 ete de la Societe de Biologie, 122, – . New York: Harcourt Brace. 1969 2 Brodal, A. ( ). Neurological anatomy ( nd ed). Damasio, H., & Damasio, A. R. (1989). Lesion London: Oxford University Press. analysis in neuropsychology. New York: Oxford Bunin, M. A., & Wightman, R. M. (1999). University Press. Paracrine neurotransmission in the CNS: Dennett, D. C. (1991). Consciousness explained. 5 Involvement of -HT. Trends in Neuroscience, Boston: Little, Brown. 22, 377–382. Dennett, D. C., & Kinsbourne, M. (1992). Time 1952 Cairns, H. ( ). Disturbances of conscious- and the observer: The where and when of con- ness with lesions of the brain-stem and dien- sciousness in the brain. Behavioral and Brain 109 146 cephalon. Brain, 75, – . Sciences, 15, 183–201. 1996 Cairns-Smith, A. G. ( ). Evolving the mind. Descartes, R. (1953) Traite de l’homme. In O Cambridge: Cambridge University Press. Euvres et letters (A. Bridoux A, Ed.). Paris: Carpenter, W. B. (1853). Principles of human phys- Gallimard. iology. London: Churchill. Doty, R. W., Sr. (1975). Consciousness from Castaigne, P., Lhermitte, F., Buge, A., Escourolle, neurons. Acta Neurobiologiae Experimentalis R., Hauw, J. J., &Lyon-Caen, O. (1981). Para- (Warsz), 35, 791–804. median thalamic and midbrain infarct: Clinical Dronkers, N. F. (2000). The gratuitous relation- and neuropathological study. Annals of Neurol- ship between Broca’s aphasia and Broca’s area. 127 148 ogy, 10, – . Behavioral and Brain Sciences, 23, 30–31. Castiello, U., Paulignan, Y., & Jeannerod, M. Dudai, Y. (2002). Memory from A to Z. Oxford: (1991). Temporal dissociation of motor res- Oxford University Press. ponses and subjective awareness. A study in Eccles, J. C. (1966). Conscious experience and 2639 2655 normal subjects. Brain, 114, – . memory. In J. C. Eccles (Ed.), Brain and con- Chalmers, D. J. (2002). The puzzle of conscious scious experience. New York: Springer. 90 98 experience. Scientific American, 12, – . Edelman, G. M. (1989). The remembered present: Churchland, P. S. (2002). Brain-wise. Cambridge, A biological theory of consciousness. New York: MA: MIT Press. Basic Books. Cicchetti, F., Lacroix, S., Beach, T. G., & Par- Edelman, G. M. (1992). Bright air, brilliant fire. ent, A. (1998). Calretinin gene expression in New York: Basic Books. the human thalamus. Brain Research: Molecu- Eichenbaum, H., & Cohen, N. J. (2001). From lar Brain Research, 54, 1–12. conditioning to conscious recollection. New York: Clark, R. E., & Squire, L. R. (1998). Classical con- Oxford University Press. ditioning and brain systems: The role of aware- Eldridge, F. L., & Chen, Z. (1996). Respiratory ness. Science, 280, 77–81. sensation: A neurophysiologic perspective. In Contreras, D., Destexhe, A., Sejnowski, T. J., & L. Adams & A. Guz (Eds.), Respiratory sensa- Steriade, M. (1997). Spatiotemporal patterns tion. New York: Marcel Dekker. P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

800 the cambridge handbook of consciousness

Engel, A. K., Fries, P., Konig, P., Brecht, M., & Gazzaniga, M. S. (1970). The bisected brain. New Singer, W. (1999). Does time help to under- York: Appleton-Century-Crofts. stand consciousness? Consciousness and Cogni- Gentilini, M., De Renzi, E. , & Crisi, G. (1987). tion, 8, 260–268. Bilateral paramedian thalamic artery infarcts: Ervin, F. R., & Mark, V. H. (1966). Studies of Report of eight cases. Journal of Neurology, Neu- the human thalamus: IV. Evoked responses. rosurgery, and Psychiatry, 50, 900–909. Annals of the New York Academy of Sciences, Goldman, A. I. (1993). Consciousness, folk psy- 112, 81–92. chology and cognitive science. Consciousness Façon, E., Steriade, M., & Wertheim, N. (1958). and Cognition, 2 , 364–382. Hypersomnie prolongee´ engendree´ par des Graff-Radford, N. R., Tranel, D., Van Hoesen, G. lesions´ bilaterales´ du systeme´ activateur medial´ W., & Brandt, J. P. (1990). Diencephalic amne- lesyndrome thrombotique de la bifurcation du sia. Brain, 113, 1–25. tronc basilaire. Revue of Neurologique (Paris), Gray, C. M., & Viana Di Prisco, G. (1997). 117 133 98, – . Stimulus-dependent neuronal oscillations and Fessard, A. E. (1954). Mechanisms of nervous local synchronization in striate cortex of the integration and conscious experience. In E. D. alert cat. Journal of Neuroscience, 17, 3239– Adrian & F. Bremer, Eds.), Brain mechanisms 3253. 200 236 and consciousness (pp – ). Springfield, Gray, J. A. (1995). The contents of consciousness: IL: C. C. Thomas. A neuropsychological conjecture. Behavioral Fields, R. D., & Stevens-Graham, B. (2002). and Brain Sciences, 18, 659–722. New insights into neuron-glia communication. Graybiel, A. M., Aosaki, T., Flaherty, A. W., & Science, 298, 556–562. Kimura, M. (1994). The basal ganglia and adap- Flanagan, O. (1991). The modularity of conscious- tive motor control. Science, 265, 1826–1831. ness. Behavioral and Brain Sciences, 14, 446– Greenfield, S. A. (1998). A rosetta stone for mind 447. and brain? In S. R. Hameroff, A. W. Kasz- Flanagan, O. (1992). Consciousness reconsidered. niak, & A. C. Scott (Eds.), Toward a science Cambridge: MIT Press. of consciousness: II. The second Tucson discus- Fortin, M., Asselin, M. C., Gould, P. V., & Parent, sions and debates (pp 231–236). Cambridge, A. (1998). Calretinin-immunoreactive neurons MA: MIT Press. in the human thalamus. Neuroscience, 84, 537– Groenewegen, H. J., & Berendse, H. W. (1994). 548. The specificity of the ‘nonspecific’ midline and Fries, P., Reynolds, J. H., Rorie, A. E., & Des- intralaminar thalamic nuclei. Trends in Neuro- imone, R. (2001). Modulation of oscillatory science, 17, 52–57. neuronal synchronization by selective visual Grossman, R. G. (1980). Are current concepts attention. Science, 291, 1560–1563. and methods in neuroscience adequate for Fuster, J. (2002). Cortex and mind: Unifying cog- studying the neural basis of consciousness and nition. New York: Oxford University Press. mental activity? In H. M. Pinsker & W. D. Fuxe, K., & Agnati, L. F. (Eds.). (1991). Volume Willis, Jr. (Eds.), Information processing in the transmission in the brain: Novel mechanisms for nervous system. New York: Raven Press. neural transmission. New York: Raven Press. Grunthal,¨ E. (1934). Der Zellbau im Thalamus Galin, D. (1992). Theoretical reflections on der Sauger¨ und des Menschen. Journal of Psy- 41 112 awareness, monitoring, and self in relation to chology and Neurology, 46, – . anosognosia. Consciousness and Cognition, 1, Grusser,¨ O. J., & Landis, T. (1991). Visual agnosias 152–162. and other disturbances of visual perception and Galin, D. (2003). The concepts “self,” “person” cognition. London: Macmillan & Co. and “I” in Buddhism and in western psychology. Guberman, A., & Stuss, D. (1983). The syn- In B. A. Wallace (Ed.), Breaking new ground: drome of bilateral paramedian thalamic infarc- Essays on Tibetan Buddhism and modern sci- tion. Neurology, 33, 540–546. ence (p. 432). New York: Columbia University Guillery, R. W., Feig, S. L., & Lozsadi, D. A. Press. (1998). Paying attention to the thalamic retic- Gallup, G. G. (1995). Mirrors minds, and ular nucleus. Trends in Neuroscience, 21, 28–32. cetaceans. Consciousness and Cognition, 4, 226– Guillery, R. W., & Harting, J. K. (2003). Struc- 228. ture and connections of the thalamic reticular P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

the thalamic intralaminar nuclei and the property of consciousness 801

nucleus: Advancing views over half a century. Jeannerod, M. (1992). The where in the brain Journal of Comparative Neurology, 463, 360– determines the when in the mind. Behavioral 371. and Brain Sciences, 15, 212–213. Guillery, R. W., & Sherman, S. M. (2002). Tha- Jennett, B. (2002). The vegetative state. Journal lamic relay functions and their role in cortico- of Neurology, Neurosurgery, and Psychiatry, 73, cortical communication: Generalizations from 355–357. the visual system. Neuron, 33, 163–175. Jones, E. G. (1985). The thalamus. New York: Guzeldere,¨ G. (1997). The many faces of Plenum Press. consciousness: A field guide. In N. Block, Jones, E. G. (1989). Defining the thalamic O. Flanagan, & G. Guzeldere¨ (Eds.), The intralaminar nuclei in primates. In G. Gain- nature of consciousness. Cambridge, MA: MIT otti, M. Bentivoglio, P. Bergonzi, & F. M. Ferro Press. (Eds.), Neurologia e scienze de base: Scritti in Hameroff, S. R., & Penrose, R. (1996). Orches- onore di Giorgio Macchi. Milan: Universita` Cat- trated reduction of quantum coherence in brain tolica del Sacro Cuore. microtubules: A model for consciousness. In Jones, E. G. (1998a). The thalamus of primates. In S. R. Hameroff, A. W. Kaszniak, & A. C. F. E. Bloom, A. Bjorklund,¨ & T. Hokfelt¨ (Eds.), Scott (Eds.), Toward a science of consciousness Handbook of chemical neuroanatomy (pp. 1– (pp. 507–540). Cambridge, MA: MIT Press. 298). Amsterdam: Elsevier. Harrington, A. H. (1987). Medicine, mind, Jones, E. G. (1998b). A new view of specific and and the doublebrain; a study in 19th cen- nonspecific thalamocortical connections. In H. tury thought. Princeton: Princeton University H. Jasper, L. Descarries, V. F. Castellucci, & Press. S. Rossignol (Eds.), Consciousness: At the fron- Hart, D., & Whitlow, J. W.(1995). The experience tiers of neuroscience. Philadelphia: Lippincott- of self in the bottlenose dolphin. Consciousness Raven. and Cognition, 4, 244–247. Jones, E. G. (2002). Thalamic circuitry and tha- He, S., Cavanagh, P., & Intriligator, J. (1997). lamocortical synchrony. Philosophical Transac- Attentional resolution. Trends in Neuroscience, tions of the Royal Society of London. Series B: 1, 115–121. Biological Sciences (London), 357, 1659–1673. Hempel, C. G. (1966). Philosophy of natural sci- Jones, E. G., & Macchi, G. (1997). Terminology ence. Englewood Cliffs, NJ: Prentice Hall. of the motor thalamus. Journal of Neurosurgery, Henderson, V. W. (1990). Alalia, aphemia, and 87, 981–982. aphasia. Archives of Neurology, 47, 85–88. Jouvet, M. (1967). Neurophysiology of the states Henke, K., Landis, T., & Markowitsch, H. J. of sleep. Physiological Reviews, 47, 117–177. (1993). Subliminal perception of pictures in Jung, R. (1954). Correlation of bioelectrical and the right hemisphere. Consciousness and Cog- autonomic phenomena with alterations of con- nition, 2 , 225–236. sciousness and arousal in man. In E. D. Adrian, Hollender, D. (1986). Semantic activation with- R. Bremer, & H. H. Jasper (Eds.), Brain mech- out conscious identification in dichotic listen- anisms and consciousness. Springfield, IL: C. C. ing, parafoveal vision, and visual masking: A Thomas. survey and appraisal. Behavioral and Brain Sci- Kanerva, P. (1988). Sparse distributed memory. ences, 9, 1–66. Cambridge, MA: MIT Press. Iacoboni, M., & Zaidel, E. (1996). Hemispheric Kaufman, E. F., & Rosenquist, A. C. (1985). Affer- independence in word recognition: Evidence ent connections of the thalamic intralaminar from unilateral and bilateral presentations. nuclei in the cat. Brain Research, 335, 281– Brain and Language, 53, 121–140. 296. Jasper, H. H. (1954). Functional properties of Keller, I., & Heckhausen, H. (1990). Readiness the thalamic reticular system. In J. F. Delofres- potentials preceding spontaneous motor acts: naye (Ed.), Brain mechanisms and conscious- Voluntary vs. involuntary control. Electroen- ness. Oxford: Blackwell. cephalography and Clinical Neurophysiology, 76, Jasper, H. H. (1966). Pathophysiological studies 351–361. of brain mechanisms in different states of con- Kertesz, A. (1993). Clinical forms of aphasia. Acta sciousness. In J. C. Eccles (Ed.), Brain and con- Neurochirurgica Supplementum (Wien), 56, 52– scious experience. Berlin: Springer-Verlag. 58. P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

802 the cambridge handbook of consciousness

Kihlstrom, J. F. (1987). The cognitive uncon- mechanisms in behavior (pp. 112–136). New scious. Science, 237, 1445–1452. York: Wiley. Kinney, H. C., Korein, J., Panigrahy, A., Dikkes, Lashley, K. S. (1954). Dynamic processes in per- P., & Goode, R. (1994). Neuropathological ception. In E. D. Adrian, F. Bremer, & H. H. findings in the brain of Karen Ann Quinlan. Jasper (Eds.), Brain mechanisms and conscious- The role of the thalamus in the persistent veg- ness. Springfield, IL: C. C. Thomas. etative state. New England Journal of Medicine, Laurent, G. (2002). Olfactory network dynam- 330, 1469–1475. ics and the coding of multidimensional signals. Kinomura, S., Larsson, J., Gulyas, B., & Roland, P. Nature Reviews Neuroscience, 3, 884–895. E. (1966) Activation by attention of the human Laureys, S., Faymonville, M. E., Luxenm A., reticular formation and thalamic intralaminar Lamy, M., Franck, G., & Maquet, P. (2000). nuclei. Science, 271, 512–515. Restoration of thalamocortical connectivity Kinsbourne, M. (1995). The intralaminar thala- after recovery from persistent vegetative state. mic nuclei: Subjectivity pumps or attention Lancet, 355, 1790–1791. action coordinators? Consciousness and Cogni- Lecours, A. R., Lhermitte, F., & Bryans, B. (1983). tion, 4, 167–171. Aphasiology. London: Balliere Tindall. Kinsbourne, M., & Smith, W. L. (1974). Hemi- LeGros Clark, W. E. (1932). The structure and spheric disconnection and cerebral function. connections of the thalamus. Brain, 55, 406– Springfield, IL: C. C. Thomas. 470. Koch, C. (1995). Visual awareness and the tha- Lesser, R. P., Dinner, D. S., Luders, H., & Morris, lamic intralaminar nuclei. Consciousness and H. H. (1986). Memory for objects presented Cognition, 4, 163–166. soon after intracarotid amobarb¨ıtal sodium Koch, C. (2004). The quest for consciousness: injections in patients with medically intractable A neurobiological approach. Englewood, CO: complex partial seizures. Neurology, 36, 889– Roberts and Company. 899. Kolb, B., & Wishaw, I. Q. (2000). Fundamentals Libet, B. (1993). Neurophysiology of consciousness. of human neuropsychology (4th ed). New York: Boston: Burkhauser.¨ W. H. Freeman. Libet, B. (2003). Timing of conscious experience: Koskoff, Y. D., Patton, R., Migler, B., & Kru- Reply to the 2002 commentaries on Libet’s per,D.(1959, Sept-Oct). Hemicerebrectomy findings. Consciousness and Cognition, 12, 321– in the Rhesus monkey: Surgical technique and 331. preliminary behavioral observations. Cerebral Libet, B., Gleason, C. A., Wright, E. W., & Pearl, Palsy. D. K. (1983). Time of conscious intention to Kossoff, E. H., Vining, E. P., Pillas, D. J., Pyzik, P. act in relation to onset of cerebral activities L., Avellino, A. M., Carson, B. S., & Freeman. (readiness potential): The unconscious initia- J. M. (2003). Hemispherectomy for intractable tion of a freely voluntary act. Brain, 106, 623– unihemispheric epilepsy: Etiology vs outcome. 642. Neurology, 61, 887–890. Llinas,´ R. R. (2001). I of the vortex: From neurons Kruper, D. C., Patton, R. A., & Koskoff, Y. D. to self. Cambridge, MA: MIT Press. (1971). Visual discrimination in hemicerebrec- Logothetis, N. K. (2002). Vision: A window tomized monkeys. Physiology and Behavior, 7, on consciousness. The hidden mind. Scientific 173–179. American [Special issue], 12, 18–25. Landesman, C. (1967). Consciousness. In P. Lycan, W. G. (1997). Consciousness as internal Edwards (Ed.), Encyclopedia of philosophy monitoring. In N. Block, O. Flanagan, & G. (pp. 191–195). New York: Macmillan. Guzeldere (Eds.), The nature of consciousness: Landis, T., & Regard, M. (1988). The right Philosophical debates. Cambridge, MA: MIT hemisphere’s access to lexical meaning: A Press. function of its release from left hemisphere Magoun, H. W. (1952). An ascending reticular control? In C. Chiarello (Ed.), Right hemisphere activating system in the brainstem. Archives of contributions to lexical semantics. New York: Neurology and Psychiatry, 57, 145–154. Springer. Mair, R. G. (1994). On the role of thalamic Lashley, K. S. (1952). The problem of serial order pathology in diencephalic amnesia. Reviews in in behavior. In L. A. Jeffress (Ed.), Cerebral the Neurosciences, 5, 105–140. P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

the thalamic intralaminar nuclei and the property of consciousness 803

Malamut, B. L., Graff-Radford, N., Chawluk, J., Studies in neurolinguistics. New York: Academic Grossman, R. I., & Gur, R. C. (1992). Memory Press. in a case of bilateral thalamic infarction. Neu- Moore, G. E. (1922). Philosophical studies. rology, 42, 163–169. London: Routledge & Kegan Paul. Marcel, A. J. (1988). Phenomenal experience and Moruzzi, G., & Magoun, H. W.(1949). Brain stem functionalism. In A. J. Marcel & E. Bisiach reticular formation and activation of the EEG. (Eds.), Consciousness in contemporary science Electroencephalography and Clinical Neurophys- (pp, 121–158). Oxford: Clarendon Press. iology, 1, 455–473. Marder, E., & Bucher, D. (2001). Central pattern Moscovitch, M. (1995). Models of consciousness generators and the control of rhythmic move- and memory. In M. S. Gazzaniga (Ed.), The ments. Current Biology, 11,R986–996. cognitive neurosciences. Cambridge, MA: MIT Marie, P. (1906). The third left frontal convolu- Press. tion plays no special role in the function of lan- Moutier, F. (1908). L’aphasie de Broca. Paris: guage. Semaine Medicale, 26, 241–247. Steinheil. Markowitsch, H. J., von Cramon, D. Y., & Schuri, Moutoussis, K., & Zeki, S. (2002). The relation- U. (1993). Mnestic performance proﬁle of a ship between cortical activation and perception bilateral diencephalic infarct patient with pre- investigated with invisible stimuli. Proceedings served intelligence and severe amnesic distur- of the National Academy of Sciences USA, 99, bances. Journal of Clinical and Experimental 9527–9532. 627 652 Neuropsychology, 15, – . Munk, M. H., Roelfsema, P. R., Konig, P., Engel, Maunsell, J. H., & Newsome, W. T. (1987). Visual A. K., & Singer, W. (1996). Role of reticular processing in monkey extrastriate cortex. activation in the modulation of intracortical Annual Review of Neuroscience, 10, 363–401. synchronization. Science, 272, 271–274. McGuinness, C. M., & Krauthamer, G. M. (1980). Munkle, M. C., Waldvogel, H. J., & Faull, R. L. The afferent projections to the centrum medi- (1999). Calcium-binding protein immunoreac- anum of the cat as demonstrated by ret- tivity delineates the intralaminar nuclei of the rograde transport of horseradish peroxidase. thalamus in the human brain. Neuroscience, 90, Brain Research, 184, 255–269. 485–491. Meissner, I., Sapir, S., Kokmen, E., & Stein, Munkle, M. C., Waldvogel, H. J., & Faull, R. L. S. D. (1987). The paramedian diencephalic (2000). The distribution of calbindin, calre- syndrome: A dynamic phenomenon. Stroke, 18, tinin and parvalbumin immunoreactivity in the 380–385. human thalamus. Journal of Chemical Neu- Metzinger, T.(1995). The problem of consciousness. roanatomy, 19, 155–173. Thorverton, UK: Imprint Academic. Natsoulas, T. (1983). Concepts of consciousness. Michel, D., Laurent, B., Foyaher, N., Blanc, A., Journal of Mind and Behavior, 4, 13–59. & Portafaix, M. (1982). Infarctus thalamique Netter, F. H. (1953). Nervous system. Summit, NJ: paramedian` gauche: tude de la memoire´ et Ciba. du langage. Revue Neurologique, 138, 533– Newman, J., & Baars, B. J. (1993). A neural atten- 550. tional model for access to consciousness: A Mills, R. P., & Swanson, P. D. (1978). Vertical ocu- global workspace perspective. Concepts in Neu- lomotor apraxia and memory loss. Annals of roscience, 4, 255–290. Neurology, 4, 149–153. O’Leary, J. L., & Coben, L. A. (1958). The retic- Milner, A. D., & Goodale, M. A. (1996). The ular core. 1957. Physiological Reviews, 38, 243– visual brain in action. Oxford: Oxford Univer- 276. sity Press. Orem, J., Schlag-Rey, M., & Schlag, J. (1973). Mitrofanis, J., & Guillery, R. W. (1993). New Unilateral visual neglect and thalamic intralam- views of the thalamicreticular nucleus in the inar lesions in the cat. Experimental Neurology, adult and the developing brain. Trends in Neu- 40, 784–797. roscience, 16, 240–245. Panksepp, J. (1998). Affective neuroscience. New Mohr, J. P. (1973). Rapid amelioration of motor York: Oxford University Press. aphasia. Archives of Neurology, 28, 77–82. Passingham, R. (1993). The frontal lobes and Mohr, J. P. (1976). Broca’s area and Broca’s apha- voluntary action. Oxford: Oxford University sia. In H. Whitaker & H. A. Whitaker (Eds.), Press. P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

804 the cambridge handbook of consciousness

Pavlov, I. P. (1928). Lectures on conditioned behavioural components of visual salience. In reflexes. New York: Liveright Publishing. E. G. Jones, M. Steriade, & D. A. McCormick Peacock, W. J. (1996). Hemispherectomy for (Eds.), The thalamus. Amstedam: Elsevier. intractable seizures in children: A report of 58 Rodman, H. R., Gross, C. G., & Albright, T. D. cases. Child’s Nervous System, 12, 376–384. (1990). Afferent basis of visual response prop- Penfield, W. (1937). The cerebral cortex and con- erties in area MT of the macaque. II. Effects of sciousness. In R. H. Wilkins (Ed.), The Harvey superior colliculus removal. Journal of Neuro- lectures. New York: Johnson Reprint Corp. science, 10, 1154–1164. Penfield, W., & Jasper, H. (1954). Epilepsy and the Rolls, E. T. (2000). The brain and emotion. functional anatomy of the human brain. Boston: Behavioral and Brain Sciences, 23, 177–234. Little & Brown. Rosenthal, D. M. (1990). Two concepts of con- Percheron, G. (1997). The motor thalamus. Jour- sciousness. In D. M. Rosenthal (Ed.), The nature nal of Neurosurgery, 87, 981. of mind. Oxford: Oxford University Press. Perruchet, P., & Vinter, A. (2002). The self- Royce, G. J., Bromley, S., & Gracco, C. (1991). organizing consciousness. Behavioral and Brain Subcortical projections to the centromedian Sciences, 25, 297–330. and parafascicular thalamic nuclei in the cat. Petit, H., Rousseaux, M., Clarisse, J., & Journal of Comparative Neurology, 306, 129– Delafosse, A. (1981). Oculocephalomotor 155. disorders and bilateral thalamo-subthalamic Rumbaugh, D. M., Beran, M. J., & Pate, J. L. infarction. Revue Neurologique (Paris), 137, (2003). Uncertainty monitoring may promote 709–722. emergents. Behavioral and Brain Sciences, 26, Plum, F., & Posner, J. B. (1985). The diagnosis of 353. stupor and coma (3rd ed). Philadelphia: F. A. Sadikot, A. F., Parent, A., Smith, Y., & Bolam, Davis. J. P. (1992). Efferent connections of the cen- Pockett, S. (2002). Backward referral, flash-lags, tromedian and parafascicular thalamic nuclei and quantum free will: A response to commen- in the squirrel monkey: A light and electron taries on articles by Pockett, Klein, Gomes, and microscopic study of the thalamostriatal pro- Trevena and Miller. Consciousness and Cogni- jection in relation to striatal heterogeneity. tion, 11, 314–425. Journal of Comparative Neurology, 320, 228– Posner, M. I., & Rothbart, M. K. (1992). Atten- 242. tional mechanisms and conscious experience. Sandifer, P. H. (1946). Anosognosia and disorders In A. D. Milner & M. D. Rugg (Eds.), The neu- of body scheme. Brain, 69, 122. ropsychology of consciousness. San Diego: Aca- Schacter, D. L. (1989). On the relation between demic Press. memory and consciousness: Dissociable Ptito, A., Lepore, F., Ptito, M., & Lassonde, M. interactions and conscious experience. In H.L (1991). Target detection and movement dis- Roediger & F. I. M. Craik (Eds.), Varieties of crimination in the blind field of hemispherec- memory and consciousness: Essays in honour of tomized patients. Brain, 114, 497–512. Endel Tulving. Hillsdale, NJ: Erlbaum. Purpura, K. P., & Schiff, N. D. (1997). The thala- Schacter, D. L., McAndrews, M. P., Moscov- mic intralaminar nuclei: A role in visual aware- itch, M. (1988). Access to conscousness: Dis- ness. The Neuroscientist, 3, 8–15. sociation between implicit and explicit knowl- Regard, M., Cook, N. D., Wieser, H. G., & Landis, edge in neuropsychological syndrome. In L. T. (1994). The dynamics of cerebral dominance Weiskrantz (Ed.), Thought without language. during unilateral limbic seizures. Brain, 117, 91– Oxford: Clarendon Press. 104. Schaltenbrand, G., & Wahren, W.(1977). Atlas for Revonsuo, A. (1995). Conscious and noncon- stereotaxy of the human brain, 2nd ed. Stuttgart: scious control of action. Behavioral and Brain Thieme. Sciences, 18, 265–266. Scheibel, M. E., & Scheibel, A. B. (1966). The Ro, T., & Rafal, R. D. (1996). Perception of geo- organization of the nucleus reticularis thalami: metric illusions in hemispatial neglect. Neu- A Golgi study. Brain Research, 1, 43–62. ropsychologia, 34, 973–978. Schiff, N. D., & Plum, F. (1999). Web forum: The Robinson, D. L., & Cowie, R. J. (1997). The neurology of impaired consciousness: Global primate pulvinar: Structural, functional and disorders and implied models. http://athena. P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

the thalamic intralaminar nuclei and the property of consciousness 805

english.vt.edu/cgi-bin/netforum/nic/a/34-- years after left hemispherectomy. Neurology, 1.2.1 25, 813–818. Schiff, N. D., Ribary, U., Moreno, D. R., Beat- Smythies, J. (1997). The functional neu- tie, B., Kronberg, E., Blasberg, R., Gicino, roanatomy of awareness: With a focus on the J., McCagg, C., Fins, J. J., Llinas, R., & role of various anatomical systems in the con- Plum, F. (2002). Residual cerebral activity and trol of intermodal attention. Consciousness and behavioural fragments can remain in the persis- Cognition, 6, 455–481. tently vegetative brain. Brain, 125, 1210–1234. Sperry, R. W. (1961). Cerebral organization and Schlag, J., & Schlag-Rey, M. (1984). Visuomo- behavior. Science, 133, 749–1757. tor functions of central thalamus in monkey. Sperry, R. W. (1968). Mental unity following sur- II. Unit activity related to visual events, target- gical disconnection of the cerebral hemispheres. ing, and fixation. Journal of Neurophysiology, 51, New York: Academic Press. 1175–1195. Sperry, R. W. (1980). Mind-brain interaction: Schott, B., Mauguiere,` F., Laurent, B., Serclerat, Mentalism, yes; dualism, no. Neuroscience, 5, O., & Fischer, C. (1980). L’amnesie´ thalamique. 195–206. Revue Neurologique (Paris), 136, 117–130. Sperry, R. W. (1991). In defense of mentalism Seager, W. (2002). Emotional introspection. Con- and emergent interaction. Journal of Mind and sciousness and Cognition, 11, 666–687. Behavior, 12, 221–245. Searle, J. R. (1993). The problem of conscious- Sperry, R. W., Gazzaniga, M. S., & Bogen, ness. Ciba Foundation Symposium, 174, 61–69; J. E. (1969). Interhemispheric relationships: discussion 70–80. The neocortical commissures; syndromes of Searle, J. R. (1998). How to study consciousness hemisphere disconnection. Handbook of Clin- scientifically. Philosophical Transactions of the ical Neurology, 4, 273–290. Royal Society of London. Series B: Biological Sci- Steriade, M. (1996). Arousal: Revisiting the retic- ences (London), 353, 1935–1942. ular activating system. Science, 272, 225–226. Sherman, S. M., & Guillery, R. W. (2002). The Steriade, M., Curro´ Dossi, R., & Contreras, D. role of the thalamus in the flow of informa- (1993). Electrophysiological properties of tion to the cortex. Philosophical Transactions of intralaminar thalamocortical cells discharging the Royal Society of London. Series B: Biological rhythmic (approximately 40 HZ) spike-bursts Sciences (London), 357, 1695–1708. at approximately 1000 HZ during waking and Sherman, S. M., & Koch, C. (1986). The control rapid eye movement sleep. Neuroscience, 56, 1 9 of retinogeniculate transmission in the mam- – . malian lateral geniculate nucleus. Experimental Steriade, M., Jones, E. G., & McCormick, D. A. Brain Research, 63, 1–20. (1997). Thalamus I: Organization and function. Sherrington, C. S. (1932). Inhibition as a coordi- New York: Elsevier. native factor. In Amsterdam: Elsevier. Stoerig, P., & Barth, E. (2001). Low-level phe- Sidibe,´ M., & Smith, Y. (1996), Differential nomenal vision despite unilateral destruction synaptic innervation of striatofugal neurones of primary visual cortex. Consciousness and 574 587 projecting to the internal or external segments Cognition, 10, – . of the globus pallidus by thalamic afferents in Stoerig, P., & Cowey, A. (1995). Visual percep- the squirrel monkey. Journal of Comparative tion and phenomenal consciousness. Behavioral Neurology, 365, 445–465. Brain Research, 71, 147–156. Singer, W. (1998). Consciousness and the struc- Stoerig, P., & Cowey, A. (1997). Blindsight in man ture of neuronal representations. Philosophi- and monkey. Brain, 120, 535–559. cal Transactions of the Royal Society of Lon- Szelies, B., Herholz, K., Pawlik, G., Karbe, H., don. Series B: Biological Sciences (London), 353, Hebold, I., & Heiss, W. D. (1991). Widespread 1829–1840. functional effects of discrete thalamic infarc- Smith, A., & Burklund, C. W. (1966). Dominant tion. Archives of Neurology, 48, 178–182. hemispherectomy: Preliminary report on neu- Szirmai, I., Guseo, A., & Molnar, M. (1977). Bilat- ropsychological sequelae. Science, 153, 1280– eral symmetrical softening of the thalamus. 1282. Journal of Neurology, 217, 57–65. Smith, A., & Sugar, O. (1975). Development Taniwaki, T., Okayama, A., Yoshiura, T., Naka- of above normal language and intelligence 21 mura, Y., Goto, Y., Kira, J., & Tobimatsu, S. J. P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

806 the cambridge handbook of consciousness

(2003). Reappraisal of the motor role of Goodale, & R. J. W. Mansfield (Eds.), Analysis basal ganglia: A functional magnetic resonance of visual behavior (pp. 549–546). Cambridge, image study. Journal of Neuroscience, 23, 3432– MA: MIT Press. 3438. van Domburg, P. H. M. F., ten Donkelaar, H. J., Tanji, J., & Shima, K. (1994). Role for supplemen- & Noterman, S. L. H. (1996). Akinetic mutism tary motor area cellsin planning several move- with bithalamic infarction neurophysiologic ments ahead. Nature, 371, 413–416. correlates. Journal of Neuroscience, 139, 58–64. Tatemichi, T. K., Steinke, W., Duncan, C., Bello, Velmans, M. (1991). Is human information pro- J. A., Odel, J. G., Behrens, M. M., Hilal, S. K., & cessing conscious? Behavioral and Brain Sci- Mohr, J. P. (1992). Paramedian thalamopedun- ences, 14, 651–726. cular infarction: Clinical syndromes and mag- von Monakow, C. (Ed.). (1911). Localization of netic resonance imaging. Annals of Neurology, brain functions. Springfield, IL: C. C. Thomas. 32, 162–171. Walshe, F. M. R. (1965). Further critical studies in Tatu, L., Moulin, T., Bogousslavsky, J., & Duver- neurology. London:E&S.Livingstone. noy, H. (1998). Arterial teritories of the human Watt, D. F. (2000). Emotion and consciousness. brain. Neurology, 50, 1699–1708. Journal of Consciousness Studies, 7, 72–84. Taylor, J. G. (1999). The race for consciousness. Weiskrantz, L. (1986). Blindsight: A case study and Cambridge, MA: MIT Press. implications. Oxford: Clarendon Press. Taylor, J. G. (2002). Paying attention to con- Weiskrantz, L. (1997). Consciousness lost and sciousness. Trends in Cognitive Science, 6, 206– found. Oxford: Oxford University Press. 210 . Wexler, B. E., Warrenburg, S., Schwartz, G. E., & Taylor, J. L., & McCloskey, D. I. (1990). Trigger- Janer, L. D. (1992). EEG and EMG responses ing of preprogrammed movements as reactions to emotion-evoking stimuli processed with- to masked stimuli. Journal of Neurophysiology, out conscious awareness. Neuropsychologia, 30, 63, 439–446. 1065–1079. Tononi, G., & Edelman, G. M. (1998). Conscious- White, R. J., Schreiner, L. H., Hughes, R. A., ness and complexity. Science, 282, 1846–1851. MacCarty, C. S., & Grindlay, J. H. (1959). Phys- Tramo, M. J., Baynes, K., Fendrich, R., Mangun, iologic consequences of total hemispherec- G. R., Phelps, E. A., Reuter-Lorenz, P. A., tomy in the monkey. Neurology, 9, 149–159. & Gazzaniga, M. S. (1995). Hemispheric spe- Wiest, G., Mallek, R., & Baumgartner, C. (2000). cialization and interhemispheric integration: Selective loss of vergence control secondary to Insights from experiments with commissuro- bilateral paramedian thalamic infarction. Neu- tomy patients. In A. G. Reeves & D. W. Roberts rology, 1997–1999. (Eds.), Epilepsy and the corpus callosum. New Wimsatt, W. C. (1976). Reductionism, levels of York: Plenum Press. organization, and the mind-body problem. In Treue, S., & Maunsell, J. H. R. (1996). Attentional G. G. Globus, G. Maxwell, & I. Savodnik modulation of visualmotion processing in cor- (Eds.), Consciousness and the brain. New York: tical areas MT and MST. Nature, 382, 539– Plenum. 541. Wimsatt, W. C. (2000). Emergence as non- Trevarthen, C. (1974). Functional relations of dis- aggregativity and the biases of reductionisms. connected hemispheres with the brain stem, Foundations of Science, 5, 269–297. and with each other: Monkey and man. In M. Yasuda, Y., Akiguchi, I., Ino, M., Nabatabe, H., Kinsbourne & W. L. Smith (Eds.), Hemispheric & Kameyama, M. (1990). Paramedian thalamic disconnection and cerebral function. Springfield, and midbrain infarcts associated with palilalia. IL: C. C. Thomas. Journal of Neurology, Neurosurgery, and Psychi- Trevena, J. T., & Miller, J. (2002). Cortical move- atry, 53, 797–799. ment preparation before and after a conscious Yingling, C. D., & Skinner, J. E. (1976). Selective decision to move. Consciousness and Cognition, regulation of thalamic sensory relay nuclei by 11, 162–190. nucleus reticularis thalami. Electroencephalog- Tulving, E. (1985). Memory and consciousness. raphy and Clinical Neurophysiology, 41, 476– Canadian Psychology, 26, 1–12. 482. Ungerleider, L. G., & Mishkin, M. (1982). Two Yingling, C. D., & Skinner, J. E. (1977). Gat- cortical visual systems. In D. J. Ingle, M. A. ing of thalamic input to cerebralcortex by P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

the thalamic intralaminar nuclei and the property of consciousness 807

nucleus reticularis thalami. In J. E. Desmedt Zaidel, E., & Iacoboni, M. (Eds.). (2002). The par- (Ed.), Attention, voluntary contraction and allel brain: The cognitive neuroscience of the cor- event-related cerebral potentials. Basel: Krager. pus callosum. Cambridge, MA: MIT Press. Zaidel, D. W., Hugdahl, K., & Johnsen, B. H. Zeki, S. (1974). Functional organization of a (1995). Physiological responses to verbally visual area in the superiortemporal sulcus of inaccessible pictorial information in the left the rhesus monkey. Journal of Physiology, 236, and right hemispheres. Neuropsychology, 9, 549. 1–6. Zeki, S. (1993). A vision of the brain. Oxford: Zaidel, E., Zaidel, D. W., & Bogen, J. E. (1999). Blackwell. The split brain. In G. Adelman & B. Smith Zeki, S., & Bartels, A. (1999). Toward a theory of (Eds.), Encyclopedia of neuroscience (2nd ed.; pp. visual consciousness. Consciousness and Cogni- 1027–1032). Amsterdam: Elsevier. tion, 8, 225–259. P1: JzG 0521857430c27 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 5:49

808 P1: JzG 0521857430c28 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 2:50

CHAPTER 28 The Cognitive Neuroscience of Memory and Consciousness

Scott D. Slotnick and Daniel L. Schacter

Abstract gesting that low retrieval content memories are associated with greater post-retrieval In this chapter, we delineate the neural activ- monitoring (although this activity does not ity associated with conscious memories char- necessarily reﬂect differential retrieval con- acterized by different degrees of ‘retrieval tent per se). Finally, memories with high content’ (i.e., sensory/contextual detail). retrieval content, to a greater degree than Based primarily on neuroimaging evidence, memories with low retrieval content, are we identify the neural regions that are asso- associated with activity in the parietal cor- ciated most consistently with the following tex and sensory cortex (along with the conscious memory processes: retrieval suc- medial temporal lobe for retrieval success > cess versus retrieval attempt, remembering attempt and remembering > knowing). versus knowing, and true recognition ver- This increased activity in sensory cortex sus false recognition. A number of patterns (and medial temporal lobe) for memo- emerge from the comparison of memories ries with high retrieval content indicates with high retrieval content (i.e., retrieval that conscious memories are constructed success, remembering, and true recognition) by reactivation of encoded item features at and memories with low retrieval content retrieval. (i.e., retrieval attempt, knowing, and false recognition). Memories with both high and low retrieval content are associated with Introduction activity in the prefrontal and parietal cortex, indicating that these regions are gener- In 1985, Endel Tulving lamented the lack ally associated with retrieval. There is also of interest in the topic of memory and evidence that memories with low retrieval consciousness shown by past and present content are associated with activity in the memory researchers: “One can read article prefrontal cortex to a greater degree than after article on memory, or consult book memories with high retrieval content, sug- after book, without encountering the term 809 P1: JzG 0521857430c28 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 2:50

810 the cambridge handbook of consciousness

‘consciousness.’ Such a state of affairs must During the 20 years that have elapsed be regarded as rather curious. One might since the publication of these papers, a think that memory should have something vast amount of research has been pub- to do with remembering, and remembering lished on the distinctions that they intro- is a conscious experience (Tulving, 1985b, duced. For example, many cognitive stud- p. 11).” ies have used the techniques introduced Though Tulving provided an accurate by Tulving (1985b) to delineate the func- assessment of the field at the time, the year tional and phenomenological characteristics in which he voiced his complaint proved of remembering and knowing (for reviews to be a kind of turning point in research and contrasting perspectives, see Dunn, on memory and consciousness. Tulving’s 2004; Gardiner, Ramponi, & Richardson- (1985b) own article focused on the impor- Klavehn, 2002). Likewise, cognitive studies tant distinction between remembering, which have also explored numerous aspects of the involves specific recollections of past experi- relation between implicit and explicit forms ences, and knowing, which involves a general of memory (for reviews, see Roediger & sense of familiarity without specific recollec- McDermott, 1993; Schacter, 1987; Schacter tion, and introduced seminal techniques for & Curran, 2000). experimentally assessing these two forms of Although purely cognitive studies have memory. played a significant role in advancing our In a different paper published that same understanding of memory and conscious- year, Tulving (1985a) argued that each ness, cognitive neuroscience studies – which of three dissociable memory systems is attempt to elucidate the nature of, and uniquely associated with a particular type relations between, the brain systems and of consciousness. Specifically, he contended processes that support various forms of that procedural memory (learning of motor, memory – have also been critically impor- perceptual, and cognitive skills) is associated tant. Indeed, much of the impetus for the with anoetic or “nonknowing” consciousness, distinction between implicit and explicit which entails simple awareness of exter- memory was provided initially by neu- nal stimuli; semantic memory (general fac- ropsychological studies of amnesic patients, tual knowledge) is associated with noetic or who exhibit severe impairment of explicit “knowing” consciousness; and episodic mem- memory for previous experiences as a ory (recollection of personal experiences) is result of damage to the hippocampus and associated with autonoetic or “self-knowing” related structures in the medial temporal consciousness. lobe (MTL; e.g., Moscovitch, Vriezen, & Finally, in 1985 Graf and Schacter Goshen-Gottstein, 1993; Nadel & Moscov- introduced the related distinction between itch, 1997; Squire, 1992; Squire, Stark, & implicit and explicit memory. According to Clark, 2004). Nonetheless, it has been Graf and Schacter (1985), explicit memory demonstrated repeatedly that conditions refers to the conscious recollection of pre- exist in which amnesics can exhibit robust vious experiences, as revealed by standard and sometimes normal implicit memory for tests of recall and recognition that require aspects of prior experiences, as exempli- intentional retrieval of previously acquired fied by such phenomena as preserved prim- information. Implicit memory, by contrast, ing and skill learning (for recent reviews, refers to non-conscious effects of past expe- see Gooding, Mayes, & van Eijk, 2000; riences on subsequent behavior and perfor- Schacter, Dobbins & Schnyer, 2004). Stud- mance, such as priming or skill learning, that ies of other neuropsychological syndromes are revealed by tests that do not require have likewise revealed dissociations between conscious recollection of previous experi- implicit and explicit forms of percep- ences (for precursors, see also Cermak, 1982; tion, language, and related cognitive and Moscovitch, 1984). motor processes (e.g., Goodale & Westwood, P1: JzG 0521857430c28 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 2:50

the cognitive neuroscience of memory and consciousness 811

2004;Guzeldere,¨ Flanagan, & Hardcastle, post-retrieval monitoring (for a review, see 2000;Kohler¨ & Moscovitch, 1997; Schacter, Schacter & Slotnick, 2004). Parietal cortex, McAndrews, & Moscovitch, 1988; Warring- particularly in Brodmann Area (BA) 39/40, ton & Weiskrantz, 1974; Young, 1994). has recently been associated with the ten- Although neuropsychological studies dency to make “old” responses (Velanova have been crucial to advancing our under- et al., 2003; Wheeler & Buckner, 2003). standing of the relation between memory The MTL, as mentioned previously, is nec- and consciousness, during the past decade essary for explicit memory, with the hip- cognitive neuroscience analyses have pocampus proper possibly serving the role focused increasingly on research using of binding together information from dis- functional neuroimaging techniques, such parate cortical regions (Squire, 1992). That as positron emission tomography (PET) is, the hippocampus may serve a central and functional magnetic resonance imaging role in combining disparate features to con- (fMRI). A vast amount of neuroimaging struct a unitary memory (e.g., Moscovitch, research has been published, and much of 1994; Schacter, Norman, & Koutstaal, 1998a; it is well beyond the scope of this chapter. Squire, 1992). However, we believe that several lines Providing support for this constructive of research concerned with elucidating view of memory, recent neuroimaging evi- the neural correlates of explicit memory dence indicates that explicit memory evokes processes do provide useful insights into activity in the appropriate domain-specific the cognitive neuroscience of memory and processing regions (i.e., retrieval-related consciousness (for reviews of neuroimaging reactivation of processing regions associ- studies concerning the neural substrates ated with memorial encoding). Specifically, associated with implicit memory, see memory for actions activates motor pro- Henson, 2003; Schacter & Buckner, 1998; cessing regions (Nyberg, Petersson, Nilsson, ´ Schacter et al., 2004; Wiggs & Martin, 1998). Sandblom, Åberg, & Ingvar, 2001), mem- Many neuroimaging studies of explicit ory for sounds activates auditory process- memory have used a recognition paradigm, ing regions (Nyberg, Habib, McIntosh, & where items such as words or objects are Tulving, 2000; Wheeler & Buckner, 2003; studied, and then on a subsequent test, Wheeler, Petersen, & Buckner, 2003), mem- these old items are randomly intermixed ory for odors activates olfactory process- with new items, and participants decide ing regions (Gottfried, Smith, Rugg, & whether each item is “old” or “new.” Item Dolan, 2004), and memory for visual stimuli recognition has been associated most consis- (e.g., shapes or objects) activates occipital- tently with activity in three neural regions: temporal regions in the ventral visual pro- (1) prefrontal cortex (anterior and dorso- cessing stream (Moscovitch, Kapur, Kohler,¨ lateral), (2) parietal cortex, and (3) the & Houle, 1995; Slotnick et al., 2003; Vaidya, MTL (for reviews, see Buckner & Schacter, Zhao, Desmond, & Gabrieli, 2002; Wheeler 2004; Buckner & Wheeler, 2001; Slotnick, & Buckner, 2003; Wheeler et al., 2000). Moo, Segal, & Hart, 2003; Tulving, Kapur, Such domain-specific sensory reactivation Craik, Moscovitch, & Houle, 1994). The is typically taken as evidence for the con- functional role(s) subserved by each of these scious re-experiencing of sensory attributes regions is currently an active area of inves- of items from the study episode. tigation, as we discuss later in this chap- This chapter considers three lines of ter. At a very general level, prefrontal cor- research that have examined aspects of this tex has been associated with the control memory-related sensory/contextual activity of retrieval (e.g., increases in activity that and the associated subjective experience correlate with retrieval demands; Velanova (or phenomenal consciousness; see Block, et al., 2003: Wheeler & Buckner, 2003) 1995; also referred to as the ‘contents of and in addition has been associated with consciousness’ or more simply ‘retrieval P1: JzG 0521857430c28 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 2:50

812 the cambridge handbook of consciousness

content’; Wheeler & Buckner, 2003). We explicit retrieval is accompanied by activa- believe that each of these lines of research tion in specific brain regions – most promi- has provided new information regarding the nently, regions within prefrontal cortex, but neural underpinnings of conscious experi- also within the MTL (e.g., Schacter, Alpert, ences of remembering. First, we consider Savage, Rauch, & Albert, 1996a; Squire et al., attempts to separate explicit retrieval into 1992; Tulving et al., 1994) – it became impor- separate components that can be grouped tant to specify further the nature of the broadly into two categories: retrieval suc- observed activity. Early PET studies adopted cess, which involves the recovery of informa- two main experimental approaches to this tion presented during a prior study episode, issue: (1) producing high and low levels of and retrieval attempt, which refers to strate- successful retrieval by manipulating study gic processes involved in explicit retrieval conditions and (2) manipulating the num- that operate even when recovery is not suc- ber of previously studied items that appear cessful. Evidence from recent neuroimag- during a particular test. We briefly sum- ing studies points toward different neu- marize studies that have used each type of ral substrates subserving these two broad approach. classes of conscious memory processes. Sec- In a PET study by Schacter et al. (1996a), ond, we discuss imaging experiments con- subjects studied some words four times cerned with the distinction between remem- and judged the number of meanings associ- bering and knowing (Tulving, 1985b) that ated with each item (high-recall condition); examine how neural activity correlates with they studied other words once and judged differing degrees or types of conscious expe- the number of t-junctions in each item riences. Third, we consider recent work con- (low-recall condition). Subjects were then cerned with delineating the neural substrates scanned during an explicit retrieval task of true versus false memories, where the (stem-cued recall, e.g., tab for table), with role of sensory reactivation in the conscious separate scans for high-recall words and low- experience of remembering has been exam- recall words. The logic underlying the exper- ined in the context of questions concerning iment is that regions that are selectively the accuracy of explicit retrieval. Although activated during the high-recall condition, we focus on neuroimaging studies in each when subjects correctly recall a large proof the three lines of research, we also dis- portion of the study list words, are pref- cuss, when relevant, complementary data erentially associated with successful con- from neuropsychological studies of brain- scious recollection; by contrast, regions that damaged patients. are activated during the low-recall condition, when subjects retrieve only a few study lists words, are preferentially asso- Neural Substrates of Retrieval Success ciated with retrieval attempt. Analysis of Versus Attempt PET data revealed blood flow increases in the hippocampal formation during the When a brain region shows changes in activ- high-recall but not the low-recall condi- ity during explicit retrieval, the changes are tion, and a significant difference between not necessarily associated with the conscious the two conditions, thereby suggesting that experience of successfully recovering pre- hippocampal activation is associated with viously studied information. Such changes some aspect of the successful conscious could instead reflect, entirely or in part, recall of a previously studied word, rather conscious processes involved in the deploy- than retrieval attempt (see also, Nyberg, ment of attention or effort when individ- McIntosh, Houle, Nilsson, & Tulving, 1996; uals attempt to retrieve the target mate- Rugg, Fletcher, Frith, Frackowiak, & Dolan, rial, independent of whether retrieval is 1997). Schacter et al. (1996a) also found that successful. Once neuroimaging studies of anterior/dorsolateral areas within prefrontal episodic memory had demonstrated that cortex were preferentially activated in the P1: JzG 0521857430c28 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 2:50

the cognitive neuroscience of memory and consciousness 813

low-recall condition, thus raising the possi- retrieval-related activation observed in right bility that blood flow increases in anterior anterior prefrontal cortex. However, results prefrontal cortex during stem-cued recall are from these studies were inconclusive, with associated with retrieval orientation effects some evidence linking right prefrontal acti- (cf., Nyberg et al., 1995). Such findings vation with retrieval attempt (e.g., Kapur, accord well with theoretical proposals that Craik, Jones, Brown, Houle, & Tulving, 1995) have linked the MTL/hippocampal region and others reporting evidence for retrieval with the automatic recovery of stored infor- success effects (e.g., Rugg, Fletcher, Frith, mation and regions within prefrontal cor- Frackowiak, & Dolan, 1996; for an attempt tex with strategic aspects of retrieval (e.g., to reconcile some of these conflicting early Moscovitch, 1994). results, see Wagner, Desmond, Glover, & In a related PET study by Rugg et al. Gabrieli, 1998). (1997), subjects studied word lists and The development of event-related fMRI either generated sentences for each word in the late 1990s provided a more direct (deep encoding) or made judgments about means of examining brain activations asso- the letters in each word (shallow encod- ciated with retrieval success and retrieval ing). Following each type of encoding task, attempt. The PET studies reviewed above they were given either an old-new recog- used blocked designs in which items from nition test (intentional retrieval) or an ani- different conditions were presented in sepa- mate/inanimate decision task (unintentional rate blocks, and data concerning brain activ- retrieval). Deep encoding produced more ity were collapsed across subjects’ behavioral accurate memory on the intentional retrieval responses. Taking advantage of the superior task. Performance was at ceiling levels on temporal resolution of fMRI compared with the unintentional task, but subjects reported PET, event-related fMRI allows intermix- spontaneously noticing that test words came ing of items from different conditions and, from the study list more often after deep more importantly, permits analysis of brain than shallow encoding, perhaps providing a activity conditional on subjects’ responses rough index of unintentional conscious rec- (Dale & Buckner, 1997). Thus, for exam- ollection. There was greater activation in ple, in a recognition memory task, “old” and left MTL areas after deep encoding than “new” responses can be analyzed separately after shallow encoding during both inten- for old and new items. Thus, retrieval suc- tional and unintentional retrieval. Thus, cess should be maximal when subjects make these data suggest that hippocampal activity “old” responses to old items (hits) and mini- during retrieval is observed with high lev- mized when subjects make “new” responses els of conscious recollection, regardless of to new items (correct rejections). whether subjects voluntarily try to remem- A number of studies have used event- ber the study list items. By contrast, there related fMRI to examine retrieval success was greater right prefrontal activation dur- versus retrieval attempt with an old-new ing intentional retrieval than during uninten- recognition test for previously studied items tional retrieval after both deep and shallow intermixed with new, non-studied items. encoding. The critical comparison involves a contrast Several PET studies have attempted of brain activity during hits and correct to separate retrieval success and retrieval rejections. A number of early studies using attempt by manipulating the proportion of these procedures failed to reveal clear evi- old items presented to subjects during a par- dence of brain activation differences dur- ticular scan. The reasoning here is that pre- ing hit versus correct rejection trials (e.g., senting large numbers of old items during Buckner, Koutstaal, Schacter, Dale, Rotte, & a particular scan will produce more suc- Rosen, 1998; Schacter, Buckner, Koutstaal, cessful retrieval than presenting only a few Dale, & Rosen, 1997a). However, as dis- old items. In general, these studies focused cussed by Konishi, Wheeler, Donaldson, and on issues concerning the characterization of Buckner (2000), these failures to observe P1: JzG 0521857430c28 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 2:50

814 the cambridge handbook of consciousness

evidence for retrieval success effects likely ticular memory test (for a contrasting view, reflected technical limitations of early event- see Slotnick & Dodson, 2005). Moreover, related fMRI procedures, such as low statis- each of the two processes are potentially sep- tical power resulting from the use of long arable into the two components on which we intertrial intervals and a correspondingly low have focused in this section of the chapter, number of items per experimental condi- retrieval success and retrieval attempt (Dob- tion. Consistent with this possibility, studies bins et al. used the closely related phrase, using more powerful event-related meth- retrieval orientation to refer to the extent to ods revealed evidence for greater activa- which subjects recruit each process during tion during hits than during correct rejec- particular retrieval tasks). With respect to tions in a number of cortical regions, most the issues raised in the preceding paragraph, consistently in prefrontal and parietal cor- Dobbins et al. (2003) noted that, when tices (e.g., Konishi et al., 2000; McDermott, presented with new items, subjects could Jones, Petersen, Lageman, & Rodeiger, 2000; rely entirely on familiarity-based processes, Nolde, Johnson, & D’Esposito, 1998) but also rejecting new items when they are not famil- in the MTL (as is discussed below). iar, and might not even attempt to engage in The results of the foregoing studies are recollection-based retrieval. Thus, it is con- consistent with the conclusion that regions ceivable that previous findings of prefrontal within prefrontal and parietal cortices are and parietal activations associated with hits specifically related to successful conscious greater than correct rejections might reflect recollection of some aspects of a pre- attempted recollective retrieval, rather than vious experience. However, this conclu- successful conscious recollection. sion depends critically on the assumption To address this issue, Dobbins et al. that comparing hits with correct rejections (2003) used a different type of experimen- isolates successful retrieval. Although the tal design in which all items had been pre- assumption appears straightforward enough, sented previously, and task demands were the comparison between hits and correct varied to require differential reliance on rejections necessarily confounds subjects’ recollection and familiarity. Prior to scan- responses (“old” or “new”) and item type ning, subjects were presented visually with (old and new). It is conceivable, therefore, a long list of nouns, and then they alter- that hit greater than correct rejection-related nated between two semantic encoding tasks brain activations do not exclusively reflect (pleasant/unpleasant and concrete/abstract differences in conscious experience related judgments). Subjects were then scanned to subjects’ responses (e.g., calling an item during two different two-alternative forced- “old” versus calling it “new”), but instead choice tests: a source memory test and a reflect differences in responses to old versus recency memory test. During source mem- new items, irrespective of subjects’ experi- ory, subjects selected the member of the ences. For example, differential responses to pair previously associated with a particu- old and new items might reflect the occur- lar encoding task; that is, they had to rec- rence of priming or related processes that ollect some type of detail associated with can occur independently of conscious mem- the particular encoding judgment performed ory (Schacter & Buckner, 1998; Schacter earlier. In contrast, the recency judgment et al., 2004; Wiggs & Martin, 1998). required subjects to select the most recently Dobbins, Rice, Wagner, and Schacter encountered item of the pair, regardless (2003) approached this issue within the con- of how it had been encoded. The source text of the theoretical distinction between memory test is assumed to rely on recol- recollection (i.e., memory for the contextual lection, whereas recency decisions can rely details of a prior encounter) and familiar- on a familiarity signal. Furthermore, suc- ity (i.e., recognition without recollection of cessful and unsuccessful trials within each contextual details). Both recollection and retrieval task were contrasted to determine familiarity can, in principle, operate on a par- whether retrieval success effects occurred in P1: JzG 0521857430c28 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 2:50

the cognitive neuroscience of memory and consciousness 815

Table 28.1. Neural regions associated most consistently with conscious memory processes Region Prefrontal Cortex Parietal Cortex MTL Sensory Cortex Retrieval Success & Attempt X X Retrieval Success > Attempt X Retrieval Attempt > Success X Remembering & Knowing X X Remembering > Knowing X X X Knowing > Remembering X True & False Recognition X X X True > False Recognition X X False > True Recognition X

Note: Regions of common (&) and differential (>) activity were identiﬁed via review of the neuroimaging literature.

overlapping or dissimilar brain regions com- Kahn, Davachi, and Wagner (2004)have pared to those associated with each retrieval provided converging evidence on the fore- orientation. going conclusions using an old-new recog- Results revealed left lateral prefrontal nition test for previously presented words and parietal activations that distinguished and new words in which subjects also made attempted source recollection from judg- a source memory judgment (whether they ments of relative recency; these retrieval had read a word at study or imagined attempt or orientation effects were largely a scene related to the word). They con- independent of retrieval success. Impor- cluded that left prefrontal/parietal regions tantly, these activations occurred largely in are related to attempted recollection of the same left prefrontal and parietal regions source information, but not to success- that had been previously identified with ful recollection of that information; by retrieval success. Because these regions were contrast, MTL activation (in the parahip- not associated with successful retrieval in the pocampal region) was related to successful Dobbins et al. (2003) design, which con- source recollection. Importantly, Kahn et al. trolled for old-new item differences present (2004) also provided evidence indicating in previous studies, it is plausible that the that left frontal/parietal activity is related to prefrontal and parietal activations in ear- familiarity-based retrieval success. Thus, the lier studies reflect attempted, rather than general distinction between retrieval success successful, conscious recollection (for fur- and retrieval attempt (or orientation) may ther relevant analyses, see Dobbins, Foley, be too coarse to prove useful theoretically. Schacter, & Wagner, 2002). In contrast, Instead, it may be necessary to specify a Dobbins et al. (2003) found that MTL struc- particular form of retrieval to make sense tures (hippocampus and parahippocampal of neuroimaging data concerning the neu- gyrus) were differentially more active dur- ral correlates of successful and attempted ing successful recollection, showing similarly retrieval (e.g., recollection versus familiarity, reduced responses during failed source rec- or remembering versus knowing, which are ollection and judgments of recency. These considered in detail below). findings complement previous data link- The results summarized in this sec- ing MTL regions with successful conscious tion indicate that neuroimaging studies recollection (e.g., Maril, Simons, Schwartz, are beginning to dissociate components of Mitchell, & Schacter, 2003; Rugg et al., 1997; conscious retrieval that are related to activ- Schacter et al., 1996a; but see, Buckner et al., ity in particular brain regions. In particu- 1998; Rugg, Henson, & Robb, 2003), as well lar, three patterns of results can be observed as other results from related paradigms con- (Table 28.1). First, both retrieval success sidered later in the chapter. and retrieval attempt were associated with P1: JzG 0521857430c28 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 2:50

816 the cambridge handbook of consciousness

activity in the prefrontal cortex and parietal ber responses and correct know responses, cortex. Second, retrieval success to a greater relative to new-correct rejections, were asso- degree than retrieval attempt was associated ciated with activity in prefrontal cortex (dor- with activity in the MTL. Third, there is solateral and medial) and medial parietal some evidence that retrieval attempt may cortex (precuneus). Relative to new-correct be associated with greater prefrontal cor- rejections, remember judgments (but not tex activity than with retrieval success. As know judgments) were also associated with research in this area progresses, increasingly additional activity in parietal cortex (supe- finer distinctions will be made regarding the rior parietal lobule and inferior parietal lob- neural substrates associated with particular ule) and the MTL (parahippocampal gyrus). aspects of conscious memorial experience. The direct contrast between remember and Of note, the fact that retrieval success – know responses complemented these results which can be assumed to reflect greater by showing activity in the parietal cortex retrieval content than retrieval attempt – is (superior parietal lobule and inferior pari- preferentially associated with the MTL sug- etal lobule). Although the MTL activation gests this region plays a role during conscious did not survive this direct contrast, it should remembering. be noted that only remember responses (versus new-correct rejections) evoked activity in the MTL, providing some indication that Neural Activity Associated with this region is preferentially associated with Remembering and Knowing remembering. The reverse contrast between know and remember was associated with We reviewed research in the preceding sec- activity in the prefrontal cortex (dorsolat- tion that attempts to dissociate recollec- eral and medial), albeit to a less extensive tion and familiarity by manipulating task degree than that associated with remem- demands. However, as noted earlier, rec- ber or know responses (versus new-correct ollection and familiarity can be assessed rejections), and medial parietal cortex directly by asking participants about their (precuneus). subjective experiences during a memory A subsequent event-related fMRI recog- task; that is, to classify “old” responses based nition memory study, also using words as on the associated memorial experience study and test materials, replicated and of remembering or knowing. Remember extended the previous pattern of results responses indicate recollection of specific by focusing on differential neural activ- contextual detail associated with a previous ity associated with remember and know experience, whereas know responses refer responses (Eldrige, Knowlton, Furmanski, to a sense of familiarity without contextual Bookheimer, & Engel, 2000). The contrast detail (Tulving, 1985b). Comparing the neu- between remember and know responses was ral activity associated with remember and associated with activity in the prefrontal cor- know responses is thus expected to pro- tex (dorsolateral), the parietal cortex (infe- vide additional insight into the substrates of rior parietal lobule), the MTL (both hip- specific types of conscious experiences con- pocampus and parahippocampal gyrus), and sidered under the general rubric of explicit the fusiform gyrus. This fusiform gyrus activ- memory. ity (coupled with the MTL activity) likely In an event-related fMRI study of reflects a greater degree of sensory reacti- remembering and knowing (Henson, Rugg, vation associated with remember as com- Shallice, Josephs, & Dolan, 1999), subjects pared to know responses. The know greater first studied a list of words. For each item than remember contrast was associated with on a subsequent recognition test, partici- a distinct region in the (anterior) prefrontal pants responded “remember” or “know” to cortex. items they judged to be “old” and other- In an event-related fMRI remember- wise responded “new.” Both correct remem- know paradigm conducted by Wheeler and P1: JzG 0521857430c28 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 2:50

the cognitive neuroscience of memory and consciousness 817

Buckner (2004; adapted from a paradigm Fabiani, & Snodgrass, 1999; see also Wilding originally designed to investigate memory- & Rugg, 1996). related domain specific sensory reactivation; Neuropsychological evidence converges see Wheeler & Buckner, 2003; Wheeler to some extent with the neuroimaging (and et al., 2000), words were paired with either ERP) findings. In a study by Knowlton and sounds or pictures at study. On the sub- Squire (1995), amnesic patients with MTL sequent recognition test, old words (those damage studied a list of unrelated words and previously paired with pictures, the only then made “remember,” “know,” or “new” type of old items considered in the analysis) judgments on a subsequent recognition test. and new words were presented. Participants Amnesic patients showed a large decrement responded “remember,” “know,” or “new.” in remember responses as compared to con- Correct remember and know responses were trol participants and a more modest but still associated with the same degree of activ- significant decline in know responses (at a ity in one subregion of the parietal cor- 10-minute delay between study and test). tex, whereas another subregion was asso- Subsequent studies showed a similar pattern ciated with greater activity for remember of results, where amnesic patients showed than know responses (both regions were in a severe impairment in remembering along the inferior parietal lobule). The contrast with more modest trends for impair- of remember versus know was also asso- ments in knowing (Schacter, Verfaellie, & ciated with activity in the prefrontal cor- Anes, 1997b; Schacter, Verfaellie, & Pradere, tex (medial), the MTL (hippocampus), and 1996c; Yonelinas, Kroll, Dobbins, Lazzara, & the fusiform cortex. Because subjects were Knight, 1998) and unilateral temporal lobec- remembering previously studied pictures, tomy patients have been shown to only be the fusiform cortex activity in this study impaired in remembering (Moscovitch & likely reflects memory-related sensory reac- McAndrews, 2002). tivation. Know versus remember responses Although these group studies include were associated with activity in the (dorso- patients with damage to a variety of lateral) prefrontal cortex. MTL structures, more recent studies have Across the studies reviewed, a num- attempted to distinguish between patients ber of patterns emerge (Table 28.1). First, with damage restricted to the hippocam- remembering and knowing, as compared pal formation and those with more exten- to new-correct rejections, were associated sive MTL damage. Yonelinas et al. (2002) with activity in prefrontal cortex and pari- found deficits in both remembering and etal cortex. Second, remembering evoked knowing in patients with damage to both greater activity than knowing most consis- the hippocampus and surrounding parahip- tently in parietal cortex, the MTL, and sen- pocampal gyrus. By contrast, they found sory cortex. Third, knowing evoked greater impairments of remembering – but not activity than remembering in the prefrontal knowing – in patients who developed mem- cortex. These findings are largely consis- ory deficits as a result of hypoxia, which tent with evidence from remember-know is known to produce damage restricted ERP studies, which have shown greater to the hippocampal formation in patients remember than know activity at parietal whose deficits are restricted to memory scalp electrodes (approximately 400–800 (see Yonelinas et al., 2002). Note, how- ms from stimulus onset) in addition to ever, that anatomical information was not similar remember and know activity (both provided concerning the precise lesion sites greater than new) at frontal scalp electrodes of the hypoxic patients included in the (approximately 1000–1600 ms from stimu- Yonelinas et al. (2002) study, so the anatom- lus onset; Curran, 2004; Duarte, Ranganath, ical implications of these findings are uncer- Winward, Hayward, & Knight, 2004;Duzel,¨ tain. Manns, Hopkins, Reed, Kitchener, Yonelinas, Mangun, Heinze, & Tulving, and Squire (2003) reported significant and 1997; Smith, 1993; Trott, Friedman, Ritter, comparable deficits of remembering and P1: JzG 0521857430c28 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 2:50

818 the cambridge handbook of consciousness

knowing in amnesics with restricted hip- Roediger and McDermott (1995), extend- pocampal damage, compared with controls. ing earlier work by Deese (1959), reported By contrast, a recent case study of patient a paradigm that produces extremely high B.E., who has selective bilateral hippocam- levels of false memories (now commonly pal damage, suggests that damage to the hip- referred to as the DRM paradigm). In the pocampal region alone can result in a specific DRM paradigm, participants are presented deficit in remembering with relative spar- with lists of associated words (e.g., fly, bug, ing of knowing (Holdstock, Mayes, Gong, insect, web, and other related words) that Roberts, & Kapur, 2005; see also Holdstock are related to a non-studied lure item (e.g., et al., 2002). spider). Roediger and McDermott’s (1995) In summary, although all neuropsycho- study showed that subjects falsely recog- logical studies of amnesic patients with MTL nized a high proportion of these related or restricted hippocampal damage reveal lure items and often claimed to specifically severe deficits of remembering, the evi- “remember” (versus “know”) that the lure dence is mixed concerning the role of the items appeared on the study list. A simi- MTL generally, and of hippocampus specif- lar paradigm has been used to study false ically, in knowing. Given current controver- memory for visual shapes, in which subjects sies in the interpretation of remember/know study physically related shapes and later pro- data (cf., Dunn, 2004; Gardiner et al., duce high levels of false alarms to perceptu- 2002; Rotello, Macmillan, & Reeder, 2004; ally similar shapes that had not been pre- Wixted & Stretch, 2004), it is perhaps not viously seen (Koutstaal, Schacter, Verfaellie, entirely surprising that clarification of the Brenner, & Jackson, 1999). From the per- relative status of remembering and know- spective of the present chapter, the developing in amnesic patients will require further ment of such paradigms allows us to exam- study. Nonetheless, these neuropsychologi- ine the similarities and differences in the cal studies complement imaging data by pro- neural correlates of conscious experiences viding evidence that the MTL is critically associated with accurate and inaccurate involved in remembering, which reflects a memories. rich form of conscious recollective experi- In the first neuroimaging study to com- ence (see discussion in Moscovitch, 1995, pare true and false memory (Schacter, 2000). Reiman, Curran, Yun, Bandy, McDermott, & Roediger, 1996b), participants heard DRM- associated lists followed by a recognition test Sensory Reactivation in True (consisting of studied/old words, lures/non- and False Memory studied related words, and non-studied unrelated/new words). Each item type at test In the type of recognition memory para- (old, related, and new), in addition to a digms we have considered thus far, analy- baseline passive fixation condition, was pre- ses of cognitive and brain activity typically sented in a separate PET scanning block. focus on accurate responses: “old” responses Both true and false recognition, compared to studied items (old-hits) or “new” respon- to baseline fixation, were associated with ses to non-studied items (new correct rejec- activity that included anterior/dorsolateral tions). False alarms to new items in such prefrontal cortex (BA 10/46), precuneus paradigms are usually too few to allow mean- (medial parietal cortex), and parahippocam- ingful analysis. However, cognitive psycho- pal gyrus (within the MTL). The direct logists have developed a number of para- contrast between true and false recognition digms that yield much larger numbers of was associated with activity in a left tempo- false alarms, thus allowing comparison of the ral parietal cortex, a region linked to audi- cognitive and neural properties of true mem- tory processing. This latter finding can be ories and false memories (for a recent review, taken as evidence for greater sensory reac- see Schacter & Slotnick, 2004). For example, tivation (i.e., auditory cortex activation dur- P1: JzG 0521857430c28 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 2:50

the cognitive neuroscience of memory and consciousness 819

ing memory for previously spoken words) with greater activity in the parietal cor- during true memory as compared to false tex (inferior parietal lobule) and another memory. region of the MTL, the parahippocampal A similar experiment was conducted gyrus. This parahippocampal gyrus activity using event-related fMRI (Schacter et al. (which has also been reported in a true/false 1997a), where event types during the recog- recognition paradigm by Okado and Stark, nition test were intermixed, and it showed 2003) may reflect greater true than false that both true and false recognition (com- recognition-related contextual reactivation pared to baseline fixation) were associated (possibly reflecting memory for the video- with similar patterns of activity including taped speakers), because this region has been the prefrontal cortex, parietal cortex, and associated with processing visual context the MTL. However, unlike the previous (Bar & Aminoff, 2003; Epstein & Kanwisher, study, the true greater than false recog- 1998). nition contrast did not reveal activity in Slotnick and Schacter (2004) used event- any region. At the same time, an ERP related fMRI to investigate the neural sub- experiment suggested that true greater than strates of true and false recognition for false recognition-related activity could be abstract visual shapes. During the study attributed to differences in blocked ver- phase, participants viewed sets of exemplar sus event-related designs (Johnson, Nolde, shapes that were similar to a non-presented Mather, Kounios, Schacter, & Curran, 1997). prototype shape (analogous to DRM word Although this latter finding suggests com- lists). At test, old shapes, related shapes mon neural substrates underlying true and (e.g., non-studied but similar shapes), or false recognition, subsequent fMRI studies new shapes were presented, and participants have shown more convincing evidence of made an old-new recognition judgment. true/false differences in brain activity and True recognition and false recognition, as have begun to elucidate the nature of that compared to new-correct rejections (i.e., activity. responding “new” to unrelated new items), In an event-related fMRI study conducted were associated with activity in the ante- by Cabeza, Rao, Wagner, Mayer, and Schac- rior/dorsolateral and medial prefrontal cor- ter (2001), a male or female (on videotape, tex, parietal cortex (superior parietal lobule, a relatively rich contextual environment) inferior parietal lobule, and precuneus), the spoke words from DRM lists of seman- MTL (hippocampus), and ventral occipital- tic associates or similar categorized lists temporal visual processing regions (BA (e.g., onion, cucumber, and pea are exem- 17/18/19/37). Although the true greater than plars of the category ‘vegetable’). Partici- false recognition contrast and the reverse pants were instructed to remember each contrast were each associated with activity word and whether it was spoken by the in different regions of the dorsolateral pre- male or female. At test, old words, related frontal cortex and parietal cortex (includ- words (non-presented associates and cate- ing precuneus and inferior parietal lobule), gories), or new words were presented, and only the true greater than false recogni- participants made an old-new recognition tion contrast was associated with activity in decision. True recognition and false recog- visual processing regions, specifically in BA nition, as compared to new items, were 17 and BA 18. These latter regions may reflect associated with activity in the dorsolateral greater visual sensory reactivation associ- prefrontal cortex, parietal cortex (medial ated with true recognition as compared to and inferior parietal lobule), and the MTL, false recognition. The results of ERP stud- specifically the hippocampus. The contrast ies investigating the neural basis of true of false recognition versus true recognition and false visual spatial memory are consis- was associated with greater activity in ven- tent with these findings (Fabiani, Stadler, & tromedial prefrontal cortex. True recogni- Wessels, 2000; Gratton, Corballis, & Jain, tion versus false recognition was associated 1997). P1: JzG 0521857430c28 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 2:50

820 the cambridge handbook of consciousness

Thus, across a number of studies, cor- familiar items, such as words and pictures of tical activity that is likely associated with common objects (for reviews, see Henson, sensory/contextual processing is greater for 2003; Schacter & Buckner, 1998; Schacter true than false recognition. Such differen- et al., 2004; Wiggs & Martin, 1998). How- tial activity might be taken as reflecting ever, it has also been found that repetition of conscious recollection of sensory/contextual novel (or masked) faces, objects, or shapes details that are remembered during true elicits increases in regional brain activity but not false recognition, an idea that (e.g., Henson, Shallice, & Dolan, 2000; has received some support from behavioral James, Humphrey, Gati, Menon, & Goodale, studies (e.g., Mather, Henkel, & Johnson, 2000; Schacter, Reiman, Uecker, Polster, 1997; Norman & Schacter, 1997). Slotnick Yun, & Cooper, 1995; Uecker, Reiman, and Schacter (2004) attempted to iden- Schacter, Polster, Cooper, Yun, & Chen, tify visual processing regions that reflect 1997). Because novel abstract shapes served conscious memory by contrasting old-hits as materials in the Slotnick and Schac- (responding “old” to old items) and old- ter study, the priming hypothesis remains misses (responding “new” to old items; see viable. Note also that Slotnick and Schac- also, Wheeler & Buckner, 2003, 2004). If ter used the identical shapes at study and activity within such regions reflects contest, thus allowing for repetition priming scious memory, then brain activity should be to occur (see also, Slotnick et al., 2003), greater for old-hits than for old-misses. Con- whereas other visual memory studies that versely, regions that reflect non-conscious failed to observe memory-related activity memory should respond equivalently dur- in early visual regions (but found memory- ing old-hits and old-misses, but in both cases related activity in late visual regions BA to a greater degree than during new cor- 19/37) did not use the identical stimuli at rect rejections (Rugg, Mark, Walla, Schlo- study and test, thus reducing the possibility erscheidt, Birch, & Allan, 1998). Slotnick of repetition priming (Vaidya et al., 2002; and Schacter found that conscious mem- Wheeler & Buckner, 2003; Wheeler et al., ory, as identified by the old-hits greater 2000). than old-misses contrast, was associated with The overall pattern of results thus sug- activity in later visual processing regions gests that memory-related activity in BA (BA 19/37), whereas non-conscious mem- 17/18 may be non-conscious. This observa- ory – identified by contrasting both old- tion has ramifications for interpreting activ- hits and old-misses each with new-correct ity associated with performance on explicit rejections – was associated with activity in memory tests. Typically, activation associ- earlier visual processing regions (BA 17/18). ated with explicit memory tests such as The same functional-anatomic dichotomy old-new recognition is attributed to con- was also observed in a follow-up experiment. scious processing; however, the present anal- Both the true greater than false recogni- ysis indicates this is not always the case. tion activity and the old-hits and old-misses Rather, additional analyses (such as the old- greater than new-correct rejections results hits versus old-misses contrast) appear nec- provide convergent evidence that activity in essary to investigate and characterize the BA17/18 reflects nonconscious memory, at nature of activity associated with explicit least in the paradigm used by Slotnick and memory. Schacter. The idea that activity in early visual We have proposed that this early visual regions that distinguishes between true area activity may reflect the influence of and false recognition reflects non-conscious priming, which as noted earlier is a non- memory processes may also help explain conscious form of memory (Slotnick & why false recognition occurs at high lev- Schacter, 2004). One possible problem with els, even though brain activity can distin- this idea is that neuroimaging studies have guish between true and false memories. If often shown that priming is associated with the activity in early visual regions that dis- decreases in activity following repetition of tinguished between true and false memories P1: JzG 0521857430c28 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 2:50

the cognitive neuroscience of memory and consciousness 821

had been consciously accessible, participants nition as compared to control participants should have used this activity to avoid mak- (Balota, Watson, Duchek, & Ferraro, 1999; ing false alarms to the related shapes. The Budson, Daffner, Desikan, & Schacter, 2000; fact that there was nonetheless a high rate of Budson, Desikan, Daffner, & Schacter, false recognition makes sense if the activity 2001; Budson, Sullivan, Daffner, & Schacter, within these regions reflects a non-conscious 2003). These neuropsychological studies form of memory. indicate that the MTL is critically involved Across the true and false recognition stud- in both true and false recognition. ies reviewed, a number of patterns can be observed (Table 28.1). First, consistent with regions previously associated with explicit Concluding Comments retrieval, true and false recognition (versus new-correct rejections) were both asso- In this chapter we have reviewed cog- ciated with activity in prefrontal cortex, nitive neuroscience evidence concerning parietal cortex, and the MTL (most con- three distinctions that illuminate different sistently within the hippocampus). Second, aspects of the relation between memory true greater than false recognition was asso- and consciousness: retrieval success versus ciated with activity in the parietal cortex attempt, remembering versus knowing, and sensory/contextual processing regions. and true versus false recognition. Retrieval Third, false greater than true recognition success involves memory of a previously was associated with activity in the prefrontal experienced item or event, whereas retrieval cortex (distinct from the commonly active attempt refers to the effort associated with regions). remembering (without success). As such, Neuropsychological studies have pro- successful retrieval (based on the asso- vided convergent evidence, particularly ciated memorial experience/details) can regarding the role of the MTL in both be said to reflect high retrieval content. true and false recognition. In a study by whereas retrieval attempt can be said to Schacter et al. (1996c), amnesic patients reflect low retrieval content. By definition, (with MTL damage) took part in a recog- remember-know studies are used to study nition memory paradigm that used associa- distinctions between contextual differences tive word lists. As expected, these patients in explicit memory: Remember responses showed lower levels of true recognition are associated with greater sensory/ (and higher levels of false alarms to new contextual detail (i.e., high retrieval con- words) as compared to control participants; tent), whereas know responses are not in addition, the patients had lower levels associated with sensory/contextual detail of false recognition (i.e., a reduced rate of (i.e., low retrieval content). True recog- false alarms to semantically related words; nition has been associated with access to see also, Melo, Winocur, & Moscovitch, greater sensory/contextual detail as com- 1999; Schacter, Verfaellie, Anes, & Racine, pared to false recognition (Mather et al., 1998b). Similarly, reduced levels of both true 1997; Norman & Schacter, 1997; Schooler, and false recognition in amnesic patients Gerhard, & Loftus, 1986). Accordingly, have also been shown in recognition mem- retrieval content can be considered greater ory paradigms that have employed concep- during true as compared to false memory tually related words (e.g., “twister,” “fun- (although not to such a degree as to preclude nel”) and perceptually related words (e.g., the occurrence of false memories). Thus, “hate,” “mate”; Schacter et al., 1997b), or although both true and false recognition are abstract visual patterns (Koutstaal et al., forms of explicit memory, where common 1999; similar to those used by Slotnick & neural substrates likely reflect mechanisms Schacter, 2004). Furthermore, Alzheimer’s of general retrieval, regions differentially disease patients (with neuropathology that associated with true and false recognition includes, but is not limited to, the MTL can be assumed to reflect high and low regions) also have lower levels of false recog- retrieval content, respectively. P1: JzG 0521857430c28 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 2:50

822 the cambridge handbook of consciousness

As reflected in our summaries at the low retrieval content items (which can be conclusion of each section of the chapter, considered access-consciousness; see Block, the patterns of results for retrieval success 1995), and is perhaps attributable to greater and attempt, for remembering and know- task difficulty, this is typically not the cen- ing, and for true and false recognition show tral focus in discussions of consciousness striking parallels (Table 28.1). The patterns and memory. Rather, high retrieval content of results for retrieval success and attempt and low retrieval content refer to the sen- differed from the patterns for remember- sory/contextual experience associated with ing and knowing only in that remembering retrieval of episodic memories. Relevant to greater than knowing (and not retrieval suc- this point, high retrieval content memories, cess greater than attempt) was associated to a greater degree than low retrieval con- with activity in parietal and sensory cor- tent memories, were associated with activ- tex (which may simply reflect general dif- ity in the parietal cortex (most consistently ferences in the use of stimulus materials; the inferior parietal lobule) and sensory e.g., pictures versus words). The patterns of processing regions (at least for remember- results for true versus false recognition were ing and true recognition, with a null result largely identical to the patterns of results for for retrieval success). The parietal activ- remembering versus knowing, except that ity may reflect a greater degree of atten- true and false recognition were both asso- tion during retrieval of memories with high ciated with MTL activity, whereas some retrieval content as compared to those with data indicate remembering but not know- low retrieval content (Corbetta & Shul- ing were associated with MTL activity (as man, 2002; Hopfinger, Buonocore, & Man- noted earlier, however, the neuropsycholog- gun, 2000). Critically, however, the greater ical evidence for this conclusion is uncertain, degree of sensory activity associated with with some data indicating a link between memories with high versus low retrieval con- knowing and MTL structures). That the tent provides evidence that memories are MTL is associated with false recognition constructed by reactivation of features that may provide some explanation why partic- comprised a previous item or event (Squire, ipants respond “old” despite the fact there 1992; Schacter et al., 1998a). may be less contextual detail associated with The present chapter shows that a cogni- these items. tive neuroscience approach can illuminate We now consider the common neural the relation between memory and con- activity associated with high retrieval consciousness, highlighting how explicit mem- tent (i.e., retrieval success, remembering, ories with different degrees of retrieval and true recognition) and low retrieval content can be linked to distinct neural sub- content (i.e., retrieval attempt, knowing, strates. Although we would be remiss not and false recognition). Memories with both to point out that this area of research is in high and low retrieval content were associ- its infancy, we also believe that the field has ated with activity in the prefrontal cortex advanced significantly since the publication and parietal cortex, which indicates these of Tulving’s (1985b) lament concerning the regions are generally associated with explicit lack of interest in memory and conscious- retrieval. There was also some evidence ness. We suspect that advances during the that memories with low retrieval content, next 20 years will be even more impressive to a greater degree than those with high than those of the past two decades. retrieval content, may be associated with increased prefrontal cortex activity; however, this activity has been attributed to Acknowledgments greater low retrieval content-related post- retrieval monitoring (Schacter & Slotnick, Preparation of this chapter was sup- 2004). That is, although there may be ported by grants NIA AG08441 and NIMH more effortful conscious processing with MH060951. P1: JzG 0521857430c28 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 2:50

the cognitive neuroscience of memory and consciousness 823

References in the brain. Nature Reviews Neuroscience, 3, 201–215. Balota, D. A., Watson, J. M., Duchek, J. M., & Curran, T. (2004). Effects of attention and con- Ferraro, F. R. (1999). Cross-modal semantic and ﬁdence on the hypothesized ERP correlates of homograph priming in healthy young, healthy recollection and familiarity. Neuropsychologia, old, and Alzheimer’s disease individuals. Jour- 42, 1088–1106. nal of International Neuropsychological Society, Dale, A. M., & Buckner, R. L. (1997). Selective 5, 626–640. averaging of rapidly presented individual trials Bar, M., & Aminoff, E. (2003). Cortical analysis using fMRI. Human Brain Mapping, 5, 329– 340 of visual context. Neuron, 38, 347–358. . Block, N. (1995). On a confusion about a function Deese, J. (1959). On the prediction of occurrence of consciousness. Behavioral and Brain Sciences, of particular verbal intrusions in immediate 18, 227–287. recall. Journal of Experimental Psychology, 58, 17 22 Buckner, R. L., Koutstaal, W., Schacter, D. L., – . Dale, A. M., Rotte, M., & Rosen, B. R. Dobbins, I. G., Foley, H., Schacter, D. L., & Wag- (1998). Functional-anatomic study of episodic ner,A.D.(2002). Executive control during retrieval: Selective averaging of event-related episodic retrieval: Multiple prenfrontal pro- fMRI trials to test the retrieval success hypoth- cesses subserve source memory. Neuron, 35, esis. NeuroImage, 7, 163–175. 989–996. Buckner, R. L., & Schacter, D. L. (2004). Neural Dobbins, I. G., Rice, H. J., Wagner, A. D., & correlates of memory’s successes and sins. In M. Schacter, D. L. (2003). Memory orientation S. Gazzaniga (Ed.), The cognitive neurosciences and success: Separable neurocognitive com- (3rd ed., pp. 739–752). Cambridge, MA: MIT ponents underlying episodic recognition. Neu- Press. ropsychologia, 41, 318–333. Buckner, R. L., & Wheeler, M. E. (2001). Dobbins, I. G., Schnyer, D. M., Verfaellie, M., & The cognitive neuroscience of remember- Schacter, D. L. (2004). Cortical activity reducing. Nature Reviews Neuroscience, 2 , 624– tions during repetition priming can result from 634. rapid response learning. Nature, 428, 316–319. Budson, A. E., Daffner, K. R., Desikan, R., & Duarte, A., Ranganath, C., Winward, L., Hay- Schacter, D. L. (2000). When false recogni- ward, D., & Knight, R. T. (2004). Dissociable tion is unopposed by true recognition: Gist- neural correlates for familiarity and recollec- based memory distortion in Alzheimer’s dis- tion during the encoding and retrieval of pic- ease. Neuropsychology, 14, 277–287. tures. Cognitive Brain Research, 18, 255–272. Budson, A. E., Desikan, R., Daffner, K. R., & Dunn, J. C. (2004). Remember-know: A matter Schacter, D. L. (2001). Perceptual false recog- of conﬁdence. Psychological Review, 111, 524– nition in Alzheimer’s disease. Neuropsychology, 542. 15, 230–243. Duzel,¨ E., Yonelinas, A. P., Mangun, G. R., Budson, A. E., Sullivan, A. L., Daffner, K. R., Heinze, H., & Tulving, E. (1997). Event-related & Schacter, D. L. (2003). Semantic versus brain potential correlates of two states of con- phonological false recognition in aging and scious awareness in memory. Proceedings of the Alzheimer’s disease. Brain and Cognition, 51, National Academy of Sciences USA, 94, 5973– 251–261. 5978. Cabeza, R., Rao, S. M., Wagner, A. D., Mayer, Eldridge, L. L., Knowlton, B. J., Furmanski, C. S., A. R., & Schacter, D. L. (2001). Can medial Bookheimer, S. Y., & Engel, S. A. (2000). temporal lobe regions distinguish true from Remembering episodes: A selective role for the false? An event-related functional MRI study hippocampus during retrieval. Nature Reviews of veridical and illusory recognition memory. Neuroscience, 3, 1149–1152. Proceedings of the National Academy of Sciences Epstein, R., & Kanwisher, N. (1998). A cortical USA, 98, 4805–4810. representation of the local visual environment. Cermak, L. S. (1982). Human memory and amne- Nature, 392, 598–601. sia. Hillsdale, NJ: Erlbaum. Fabiani, M., Stadler, M. A., & Wessels, P. M. Corbetta, M., & Shulman, G. L. (2002). Control (2000). True but not false memories produce of goal-directed and stimulus-driven attention a sensory signature in human lateralized brain P1: JzG 0521857430c28 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 2:50

824 the cambridge handbook of consciousness

potentials. Journal of Cognitive Neuroscience, 12, nition spared relative to recall after selective 941–949. hippocampal damage in humans? Hippocam- Gardiner, J. M., Ramponi, C., & Richardson- pus, 12, 341–351. Klavehn, A. (2002). Recognition memory and Hopﬁnger, J. B., Buonocore, M. H., & Mangun, decision processes: A meta-analysis of remem- G. R. (2000). The neural mechanisms of top- ber, know, and guess responses. Memory, 10, down attentional control. Nature Neuroscience, 83–98. 3, 284–291. Goodale, M. A., & Westwood, D. A. (2004). James, T. W., Humphrey, G. K., Gati, J. S., An evolving view of duplex vision: Separate Menon, R. S., & Goodale, M. A. (2000). The but interacting cortical pathways for percep- effects of visual object priming on brain activation and action. Current Opinion in Neurobiol- tion before and after recognition. Current Biol- ogy, 14, 203–211. ogy, 10, 1017–1024. Gooding, P. A., Mayes, A. R., & van Eijk, R. Johnson, M. K., Nolde, S. F., Mather, M., (2000). A meta-analysis of indirect memory Kounios, J., Schacter, D. L., & Curran, T. tests for novel material in organic amnesics. (1997). The similarity of brain activity associ- Neuropsychologia, 38, 666–676. ated with true and false recognition memory Gottfried, J. A., Smith, A. P. R., Rugg, M. D., depends on test format. Psychological Science, & Dolan, R. J. (2004). Remembrance of odors 8, 250–257. past: Human olfactory cortex in cross-modal Kahn, I., Davachi, L., & Wagner, A. D. recognition memory. Neuron, 42, 687–695. (2004). Functional-neuroanatomic correlates Graf, P., & Schacter, D. L. (1985). Implicit and of recollection: Implications for models of explicit memory for new associations in normal recognition memory. Journal of Neuroscience, subjects and amnesic patients. Journal of Exper- 24, 4172–4180. imental Psychology: Learning, Memory, and Cog- Kapur, S., Craik, F. I., Jones, C., Brown, G. M., nition, 11, 501–518. Houle, S., & Tulving, E. (1995). Functional Gratton, G., Corballis, P. M., & Jain, S. (1997). role of the prefrontal cortex in retrieval of Hemispheric organization of visual memories. memories: A PET study. NeuroReport, 6, 1880– Journal of Cognitive Neuroscience, 9, 92–104. 1884. Guzeldere, G., Flanagan, O., & Hardcastle, Knowlton, B. J., & Squire, L. R. (1995). Remem- V. G. (2000). The nature and function of bering and knowing: Two different expressions consciousness: Lessons from blindsight. In M. of declarative memory. Journal of Experimental S. Gazzaniga (Ed.), The cognitive neurosciences Psychology: Learning, Memory, & Cognition, 21, (2nd ed., pp. 1277–1284). Cambridge, MA: 699–710. MIT Press. Kohler,¨ S., & Moscovitch, M. (1997). Uncon- Henson, R. N. (2003). Neuroimaging studies of scious visual processing in neuropsychological priming. Progress in Neurobiology, 70, 53–81. syndromes: A survey of the literature and eval- Henson, R. N. A., Rugg, M. D., Shallice, T., uation of models of consciousness. In M. D. Josephs, O., & Dolan, R. J. (1999). Recollec- Rugg (Ed.), Cognitive neuroscience (pp. 305– tion and familiarity in recognition memory: 373). Cambridge, MA: MIT Press. An event-related functional magnetic reso- Konishi, S., Wheeler, M. E., Donaldson, D. I., nance imaging study. Journal of Neuroscience, & Buckner, R. L. (2000). Neural correlates of 19, 3962–3972. episodic retrieval success. NeuroImage, 12, 276– Henson, R., Shallice, T., & Dolan, R. (2000). 286. Neuroimaging evidence for dissociable forms Koutstaal, W., Schacter, D. L., Verfaellie, M., of repetition priming. Science, 287, 1269–1272. Brenner, C., & Jackson, E. M. (1999). Perceptu- Holdstock, J. S., Mayes, A. R., Gong, Q. Y., ally based false recognition of novel objects in Roberts, N., & Kapur, N. (2005). Item recog- amnesia: Effects of category size and similarity nition is less impaired than recall and associa- to category prototypes. Cognitive Neuropsychol- tive recognition in a patient with selective hip- ogy, 16, 317–341. pocampal damage. Hippocampus, 15, 203–215. Manns, J. R., Hopkins, R. O., Reed, J. M., Kitch- Holdstock, J. S., Mayes, A. R., Roberts, N., Ceza- ener, E. G., & Squire, L. R. (2003). Recognition yirli, E., Isaac, C. L., O’Reilly, R. C., & Norman, memory and the human hippocampus. Neu- K. A. (2002). Under what conditions is recog- ron, 37, 171–180. P1: JzG 0521857430c28 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 2:50

the cognitive neuroscience of memory and consciousness 825

Maril, A., Simons, J. S., Mitchell, J. P., Schwartz, Moscovitch, M., Vriezen, E., & Goshen- B. L., & Schacter, D. L. (2003). Feeling-of- Gottstein, Y. (1993). Implicit tests of memory knowing in episodic memory: An event-related in patients with focal lesions or degenerative fMRI study. NeuroImage, 18, 827–836. brain disorders. In F. Boller & J. Grafman (Eds.), Mather, M., Henkel, L. A., & Johnson, M. K. Handbook of neuropsychology (Vol. 8,pp.133– (1997). Evaluating characteristics of false 173). Amsterdam: Elsevier. memories: Remember/know judgments and Nadel, L., & Moscovitch, M. (1997). Memory memory characteristics questionnaire com- consolidation, retrograde amnesia and the hip- pared. Memory & Cognition, 25, 826–837. pocampal complex. Current Opinion in Neuro- McDermott, K. B., Jones, T. C., Petersen, S. E., biology, 7, 217–227. Lageman, S. K., & Roediger, H. L. III. (2000). Nolde, S. F., Johnson, M. K., & D’Esposito, Retrieval success is accompanied by enhanced M. (1998). Left prefrontal activation during activation in anterior prefrontal cortex during episodic remembering: An event-related fMRI recognition memory: An event-related fMRI study. NeuroReport, 9, 3509–3514. study. Journal of Cognitive Neuroscience, 12, Norman, K. A., & Schacter, D. L. (1997). 965–976. False recognition in younger and older adults: Melo, B., Winocur, G., & Moscovitch, M. Exploring the characteristics of illusory mem- (1999). False recall and false recognition: An ories. Memory & Cognition, 25, 838–848. examination of the effects of selective and Nyberg, L., Habib, R., McIntosh, A. R., & Tul- combined lesions to the medial temporal ving, E. (2000). Reactivation of encoding- lobe/diencephalon and frontal lobe structures. related brain activity during memory retrieval. 343 359 Cognitive Neuropsychology, 16, – . Proceedings of the National Academy of Sciences Moscovitch, M. (1984). The sufficient conditions USA, 97, 11120–11124. for demonstrating preserved amnesia: A task Nyberg, L., McIntosh, A. R., Houle, S., Nils- analysis. In L. R. Squire & N. Butters (Eds.), son, L. G., & Tulving, E. (1996). Activation Neuropsychology of memory (pp. 104–114). New of medial temporal structures during episodic York: Guilford Press. memory retrieval. Nature, 380, 715–717. Moscovitch, M. (1994). Memory and working Nyberg, L., Petersson, K. M., Nilsson, L., Sand- ´ with memory: Evaluation of a component pro- blom, J., Åberg, C., & Ingvar, M. (2001). Reac- cess model and comparisons with other models. tivation of motor brain areas during explicit In D. L. Schacter & E. Tulving (Eds.) Memory memory for actions. NeuroImage, 14, 521–528. 269 331 systems 1994 (pp. – ). Cambridge, MA: Nyberg, L., Tulving, E., Habib, R., Nilsson, L. G., MIT Press. Kapur, S., Houle, S., Cabeza, R., & McIntosh, Moscovitch, M. (1995). Recovered conscious- A. R. (1995). Functional brain maps of retrieval ness: A hypothesis concerning modularity and mode and recovery of episodic information. episodic memory. Journal of Clinical and Exper- NeuroReport, 7, 249–252. 276 290 imental Neuropsychology, 17, – . Okado, Y., & Stark, C. (2003). Neural process- Moscovitch, M. (2000). Theories of memory and ing associated with true and false memory consciousness. In E. Tulving & F. I. M. Craik retrieval. Cognitive, Affective, & Behavioral Neu- (Eds.), Oxford handbook of memory (pp. 609– roscience, 3, 323–334. 625 ). London: Oxford University Press. Roediger, H. L. III, & McDermott, K. B. (1993). Moscovitch, M., Kapur, S., Kohler,¨ S., & Houle, Implicit memory in normal human subjects. In S. (1995). Distinct neural correlates of visual H. Spinnler & F. Boller (Eds.), Handbook of neu- long-term memory for spatial location and ropsychology (Vol. 8,pp.63–131). Amsterdam: object identity: A positron emission tomog- Elsevier. raphy study in humans. Proceedings of the Roediger, H. L.III, & McDermott, K. B. (1995). 3721 National Academy of Sciences USA, 92, – Creating false memories: Remembering words 3725 . not presented in lists. Journal of Experimental Moscovitch, M., & McAndrews, M. P. (2002). Psychology: Learning, Memory, and Cognition, Material-specific deficits in “remembering” in 21, 803–814. patients with unilateral temporal lobe epilepsy Rotello, C. M., Macmillan N. A., & Reeder, and excisions. Neuropsychologia, 40, 1335– J. A. (2004). Sum-difference theory of remem- 1342. bering and knowing: A two-dimensional P1: JzG 0521857430c28 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 2:50

826 the cambridge handbook of consciousness

signal-detection model. Psychological Review, out language (pp. 242–278). New York: Oxford 111, 588–616. University Press. Rugg, M. D., Fletcher, P. C., Frith, C. D., Frack- Schacter, D. L., Norman, K. A., & Koutstaal, W. owiak, R. S., & Dolan, R. J. (1996). Differential (1998a). The cognitive neuroscience of con- activation of the prefrontal cortex in successful structive memory. Annual Review of Psychology, and unsuccessful memory retrieval. Brain, 119, 49, 289–318. 2073–2083. Schacter, D. L., Reiman, E., Curran, T., Yun, Rugg, M. D., Fletcher, P. C., Frith, C. D., Frack- L. S., Bandy, D., McDermott, K. B., & Roediger, owiak, R. S., & Dolan, R. J. (1997). Brain H. L. III. (1996b). Neuroanatomical correlates regions supporting intentional and incidental of veridical and illusory recognition memory: memory: A PET study. NeuroReport, 24, 1283– Evidence from positron emission tomography. 1287. Neuron, 17, 267–274. Rugg, M. D., Henson, R. N. A., & Robb, W. G. K. Schacter, D. L., Reiman, E., Uecker, A., Polster, (2003). Neural correlates of retrieval process- M. R., Yun, L. S., & Cooper, L. A. (1995). Brain ing in the prefrontal cortex during recognition regions associated with retrieval of structurally and exclusion tasks. Neuropsychologia, 41, 40– coherent visual information. Nature, 376, 587– 52. 590. Rugg, M. D., Mark, R. E., Walla, P., Schloer- Schacter, D. L., & Slotnick, S. D. (2004). The scheidt, A. M., Birch, C. S., & Allan, K. (1998). cognitive neuroscience of memory distortion. Dissociation of the neural correlates of implicit Neuron, 44, 149–160. and explicit memory. Nature, 392, 595– Schacter, D. L., Verfaellie, M., & Anes, M. D. 598. (1997b). Illusory memories in amnesic patients: Schacter, D. L. (1987). Implicit memory: His- Conceptual and perceptual false recognition. tory and current status. Journal of Experimental Neuropsychology, 11, 331–342. Psychology: Learning, Memory, and Cognition, Schacter, D. L., Verfaellie, M., Anes, M. D., 13, 501–518. & Racine, C. (1998b). When true recognition Schacter, D. L., Alpert, N. M., Savage, C. R., suppresses false recognition: Evidence from Rauch, S. L., & Albert, M. S. (1996a). Con- amnesic patients. Journal of Cognitive Neuro- scious recollection and the human hippocam- science, 10, 668–679. pal formation: Evidence from positron emis- Schacter, D. L., Verfaellie, M., & Pradere, sion tomography. Proceedings of the National D. (1996c). The neuropsychology of mem- Academy of Sciences USA, 93, 321–325. ory illusions: False recall and recognition in Schacter, D. L., & Buckner, R. L. (1998). Priming amnesic patients. Journal of Memory and Lan- and the brain. Neuron, 20, 185–195. guage, 35, 319–334. Schacter, D. L., Buckner, R. L., Koutstaal, W., Schooler, J. W., Gerhard, D., & Loftus, E. F. Dale, A. M., & Rosen, B. R. (1997a). Late onset (1986). Qualities of the unreal. Journal of Exper- of anterior prefrontal activity during true and imental Psychology: Learning, Memory, and Cog- false recognition: An event-related fMRI study. nition, 12, 171–181. NeuroImage, 6, 259–269. Slotnick, S. D., & Dodson, C. S. (2005). Sup- Schacter, D. L., & Curran, T. (2000). Memory port for a continuous (single-process) model of without remembering and remembering with- recognition memory and source memory. Mem- out memory: Implicit and false memories. In ory & Cognition, 33, 151–170. M. S. Gazzaniga (Ed.), The cognitive neuro- Slotnick, S. D., Moo, L. R., Segal, J. B., & Hart, sciences (2nd ed., pp. 829–840). Cambridge, J. Jr. (2003). Distinct prefrontal cortex activity MA: MIT Press. associated with item memory and source mem- Schacter, D. L., Dobbins, I. G., & Schnyer, ory for visual shapes. Cognitive Brain Research, D. M. (2004). Speciﬁcity of priming: A cogni- 17, 75–82. tive neuroscience perspective. Nature Reviews Slotnick, S. D., & Schacter, D. L. (2004). A Neuroscience, 5, 853–862. sensory signature that distinguishes true from Schacter, D. L., McAndrews, M. P., & Moscov- false memories. Nature Neuroscience, 7, 664– itch, M. (1988). Access to consciousness: Dis- 672. sociations between implicit and explicit knowl- Smith, M. E. (1993). Neurophysiological man- edge. In L. Weiskrantz (Ed.), Thought with- ifestations of recollective experience during P1: JzG 0521857430c28 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 2:50

the cognitive neuroscience of memory and consciousness 827

recognition memory judgments. Journal of Cog- tex and recognition memory: Functional-MRI nitive Neuroscience, 5, 1–13. evidence for context-dependent retrieval pro- Squire, L. R. (1992). Memory and the hippo- cesses. Brain, 121, 1985–2002. campus: A synthesis from findings with rats, Warrington, E. K., & Weiskrantz, L. (1974). The monkeys, and humans. Psychological Review, effect of prior learning on subsequent retention 99, 195–231. in amnesic patients. Neuropsychologia, 12, 419– Squire, L. R., Ojemann, J. G., Miezin, F. M., 428. Petersen, S. E., Videen, T. O., & Raichle, M. E. Wheeler, M. E., & Buckner, R. L. (2003). (1992). Activation of the hippocampus in nor- Functional dissociation among components of mal humans: A functional anatomical study of remembering: Control, perceived oldness, and memory. Proceedings of the National Academy content. Journal of Neuroscience, 23, 3869– of Sciences USA, 89, 1837–1841. 3880. Squire, L. R., Stark, C. E. L., & Clark, R. E. Wheeler, M. E., & Buckner, R. L. (2004). (2004). The medial temporal lobe. Annual Functional-anatomic correlates of remember- Review of Neuroscience, 27, 279–306. ing and knowing. NeuroImage, 21, 1337– Trott, C. T., Friedman, D., Ritter, W., Fabiani, 1349. M., & Snodgrass, J. G. (1999). Episodic prim- Wheeler, M. E., Petersen, S. E., & Buckner, R. L. ing and memory for temporal source: Event- (2000). Memory’s echo: Vivid remembering related potentials reveal age-related differences reactivates sensory-specific cortex. Proceedings in prefrontal functioning. Psychology and Aging, of the National Academy of Sciences USA, 97, 14, 390–413. 11125–11129. Tulving, E. (1985a). How many memory systems Wiggs, C. L., & Martin, A. (1998). Properties are there? American Psychologist, 40, 385–398. and mechanisms of perceptual priming. Cur- Tulving, E. (1985b). Memory and consciousness. rent Opinion in Neurobiology, 8, 227–233. Canadian Psychology, 26, 1–12. Wilding, E. L., & Rugg, M. D. (1996). An event- Tulving, E., Kapur, S., Craik, F. I. M., Moscovitch, related potential study of recognition memory M., & Houle, S. (1994). Hemispheric encod- with and without retrieval of source. Brain, 119, ing/retrieval asymmetry in episodic memory: 889–905. Positron emission tomography findings. Pro- Wixted, J. T., & Stretch, V. (2004). In defense of ceedings of the National Academy of Sciences the signal detection interpretation of remem- USA, 91, 2016–2020. ber/know judgments. Psychonomic Bulletin & Uecker, A., Reiman, E. M., Schacter, D. L., Pol- Review, 11, 616–641. ster, M. R., Cooper, L. A., Yun, L. S., & Yonelinas, A. P., Kroll, N. E. A., Dobbins, I., Chen, K. (1997). Neuroanatomical correlates Lazzara, M., & Knight, R. T. (1998). Recollec- of implicit and explicit memory for structurally tion and familiarity deficits in amnesia: Con- possible and impossible visual objects. Learning vergence of remember-know, process dissocia- & Memory, 4, 337–355. tion, and receiver operating characteristic data. Vaidya, C. J., Zhao, M., Desmond, J. E., & Neuropsychology 12, 323–339. 2002 Gabrieli, J. D. E. ( ). Evidence for corti- Yonelinas, A. P., Kroll, N. E., Quamme, J. R., Laz- cal encoding specificity in episodic memory: zara, M. M., Sauve, M. J., Widarman, K. F., & Memory-induced re-activation of picture pro- Knight, R. T. (2002). Effects of extensive tem- 2136 2143 cessing areas. Neuropsychologia, 40, – . poral lobe damage or mild hypoxia on recol- Velanova, K., Jacoby, L. L., Wheeler, M. E., lection and familiarity. Nature Neuroscience, 5, McAvoy, M. P., Petersen, S. E., & Buckner, 1236–1241. 2003 R. L. ( ). Functional-anatomic correlates of Young, A. W. (1994). Conscious and nonconsious sustained and transient processing components recognition of familiar faces. In C. Umilta&M.` engaged during controlled retrieval. Journal of Moscovitch (Eds.), Attention and performance 8460 8470 Neuroscience, 23, – . 15: Conscious and nonconscious information pro- Wagner, A. D., Desmond, J. E., Glover, G. H., cessing (pp. 153–178). Cambridge, MA: MIT & Gabrieli, J. D. E. (1998). Prefrontal cor- Press. P1: JzG 0521857430c28 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 2:50

828 P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

C. Affective Neuroscience of Consciousness

829 P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

830 P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

CHAPTER 29 The Affective Neuroscience of Consciousness: Higher-Order Syntactic Thoughts, Dual Routes to Emotion and Action, and Consciousness

Edmund T. Rolls

Abstract particular functions (Rolls, 1999a, 2005a). A reward is anything for which an ani- In this chapter, a theory of the nature mal (which includes humans) will work. of emotion and the functions of emotions A punisher is anything that an animal are described. A Higher-Order Syntactic will escape from or avoid. An example of Thought theory of consciousness is then an emotion might thus be happiness pro- developed. It is argued that the adaptive duced by being given a reward, such as a value of higher-order thoughts is to solve pleasant touch, praise, or winning a large the credit assignment problem that arises if sum of money. Another example of an a multistep syntactic plan needs to be cor- emotion might be fear produced by the rected. It is then suggested that it feels like sound of a rapidly approaching bus or the something to be an organism that can think sight of an angry expression on someone’s about its own linguistic and semantically face. We will work to avoid such stim- based thoughts. It is suggested that qualia uli, which are punishing. Another exam- and raw sensory and emotional feels arise ple would be frustration, anger, or sadness secondary to having evolved such a higher- produced by the omission of an expected order thought system and that sensory and reward. such as a prize, or the termination emotional processing feels like something of a reward, such as the death of a loved because it would be unparsimonious for it one. Another example would be relief, pro- to enter the planning, higher-order thought duced by the omission or termination of a system and not feel like something. punishing stimulus, such as the removal of a painful stimulus, or sailing out of danger. These examples indicate how emotions Emotions as States can be produced by the delivery, omission, or termination of rewarding or pun- Emotions can usefully be deﬁned as states ishing stimuli and go some way to indicate elicited by rewards and punishers that have how different emotions could be produced 831 P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

832 the cambridge handbook of consciousness

and classified in terms of the rewards and ing “secondary reinforcers”. This type of punishments received, omitted, or termi- learning may thus be called “stimulus- nated. A diagram summarizing some of the reinforcement association” and occurs via a emotions associated with the delivery of re- process like classical conditioning. If a rein- ward or punishment or a stimulus associated forcer increases the probability of emission with them, or with the omission of a reward of a response on which it is contingent, or punishment, is shown in Figure 29.1. it is said to be a “positive reinforcer” or Before accepting this approach, we “reward”; if it decreases the probability of should consider whether there are any such a response it is a “negative reinforcer” exceptions to the proposed rule. Are any or “punisher”. For example, fear is an emo- emotions caused by stimuli, events, or tional state that might be produced by a remembered events that are not rewarding sound (the conditioned stimulus) that has or punishing? Do any rewarding or punish- previously been associated with an electrical ing stimuli not cause emotions? We con- shock (the primary reinforcer). sider these questions in more detail below. The converse reinforcement contingen- The point is that if there are no major cies produce the opposite effects on behav- exceptions, or if any exceptions can be ior. The omission or termination of a clearly encapsulated, then we may have positive reinforcer (“extinction” and “time a good working definition at least of what out”, respectively, sometimes described as causes emotions. Moreover, it is worth point- “punishing”) decreases the probability of ing out that many approaches to or the- responses. Responses followed by the omis- ories of emotion (see Strongman, 1996) sion or termination of a negative reinforcer have in common that part of the pro- increase in probability, this pair of nega- cess involves “appraisal” (e.g., Frijda, 1986; tive reinforcement operations being termed Lazarus, 1991; Oatley & Jenkins, 1996). In all “active avoidance” and “escape” respec- these theories the concept of appraisal pre- tively (see further Gray, 1975; Mackintosh, sumably involves assessing whether someth- 1983). ing is rewarding or punishing. The descrip- This foundation has been developed (see tion in terms of reward or punishment Rolls, 1986a, b, 1990, 1999a, 2000a, 2005a,) adopted here seems more tightly and oper- to show how a very wide range of emotions ationally specified. I next consider a slightly can be accounted for by the operation of six more formal definition than rewards or pun- factors: ishments, in which the concept of reinforcers is introduced, and show how there has been 1. The reinforcement contingency (e.g., whe- a considerable history in the development of ther reward or punishment is given or ideas along this line. withheld; see Fig. 29.1). The proposal that emotions can be use- 2. The intensity of the reinforcer (see Fig. fully seen as states produced by instru- 29.1). mental reinforcing stimuli (Rolls, 2005a) 3 follows earlier work by Millenson (1967), . Any environmental stimulus might have Weiskrantz (1968), Gray (1975, 1987), and a number of different reinforcement associa- Rolls (1986a, b, 1990, 1999a, 2000a). (Instru- tions. (For example, a stimulus might be mental reinforcers are stimuli that, if their associated both with the presentation of occurrence, termination, or omission is a reward and of a punisher, allowing such made contingent upon the making of a res- states as conflict and guilt to arise.) ponse, alter the probability of the future 4. Emotions elicited by stimuli associated emission of that response.) Some stimuli with different primary reinforcers will be are unlearned reinforcers (e.g., the taste different. of food if the animal is hungry, or pain), 5. Emotions elicited by different secondary whereas others may become reinforcing by reinforcing stimuli will be different from learning, because of their association with each other (even if the primary reinforcer such primary reinforcers, thereby becom- is similar). P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

the affective neuroscience of consciousness 833

S+

Ecstasy

Elation

Pleasure Rage Anger Frustration Relief S+ or S+ ! S- or S- !

Grief Sadness Apprehension

Fear

Terror

S- Figure 29.1. Some of the emotions associated with different reinforcement contingencies are indicated. Intensity increases away from the centre of the diagram, on a continuous scale. The classiﬁcation scheme created by the different reinforcement contingencies consists of (1) the presentation of a positive reinforcer (S+), (2) the presentation of a negative reinforcer (S –), (3) the omission of a positive reinforcer (S+) or the termination of a positive reinforcer (S+!), and (4) the omission of a negative reinforcer (S–)orthe termination of a negative reinforcer (S–!).

6. The emotion elicited can depend on negative. In that an emotion is produced whether an active or passive behavioral by a stimulus, philosophers say that emo- response is possible. (For example, if tions have an object in the world and that an active behavioral response can occur emotional states are intentional, in that they to the omission of a positive rein- are about something. We note that a mood forcer, then anger might be produced, or affective state may occur in the absence but if only passive behavior is possible, of an external stimulus, as in some types then sadness, depression, or grief might of depression, but that normally the mood occur.) or affective state is produced by an external stimulus, with the whole process of It is also worth noting that emotions can stimulus representation, evaluation in terms be produced just as much by the recall of reward or punishment, and the result- of reinforcing events as by external rein- ing mood or affect being referred to as forcing stimuli and that cognitive process- emotion. ing (whether conscious or not) is important It is worth raising the issue that philoso- in many emotions, for very complex cog- phers usually categorize fear as an emotion, nitive processing may be required to deter- but not pain. The distinction they make may mine whether or not environmental events be that primary (unlearned) reinforcers do are reinforcing. Indeed, emotions normally not produce emotions, whereas secondary consist of (1) cognitive processing that anal- reinforcers (stimuli associated by stimulus- yses the stimulus and then determines its reinforcement learning with primary rein- reinforcing valence and then (2) an elicited forcers) do. They describe the pain as a sen- mood change if the valence is positive or sation. But neutral stimuli (such as a table) P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

834 the cambridge handbook of consciousness

can produce sensations when touched. It bility of the stimulus and the response, accordingly seems to be much more useful to and which can be seen as a very limited categorise stimuli according to whether they type of brain solution to the elicitation of are reinforcing (in which case they produce behavior. The emotional route to action emotions) or are not reinforcing (in which is flexible not only because any action case they do not produce emotions). Clearly can be performed to obtain the reward there is a difference between primary rein- or avoid the punishment but also be- forcers and learned reinforcers; but this is cause the animal can learn in as little as most precisely caught by noting that this is one trial that a reward or punishment is the difference, and that it is whether a stim- associated with a particular stimulus, in ulus is reinforcing that determines whether what is termed “stimulus-reinforcer asso- it is related to emotion. ciation learning”. To summarize and formalize, two processes are involved in the actions The Functions of Emotion being described. The first is stimulus- reinforcer association learning, and the The functions of emotion also provide in- second is instrumental learning of an sight into the nature of emotion. These func- operant response made to approach and tions, described more fully elsewhere (Rolls obtain the reward or to avoid or escape 1990, 1999a, 2000a, 2005a), can be summa- from the punisher. Emotion is an integral rized as follows: part of this, for it is the state elicited in the first stage by stimuli that are decoded 1. The elicitation of autonomic responses (e.g., as rewards or punishers, and this state a change in heart rate) and endocrine has the property that it is motivating. responses (e.g., the release of adrenaline). The motivation is to obtain the reward These prepare the body for action. or avoid the punisher, and animals must 2. Flexibility of behavioral responses to rein- be built to obtain certain rewards and forcing stimuli. Emotional (and motiva- avoid certain punishers. Indeed, primary tional) states allow a simple interface or unlearned rewards and punishers are between sensory inputs and action sys- specified by genes that effectively specify tems. The essence of this idea is that goals the goals for action. This is the solution for behavior are specified by reward and that natural selection has found for how punishment evaluation. When an envi- genes can influence behavior to promote ronmental stimulus has been decoded their fitness (as measured by reproductive as a primary reward or punishment or success) and for how the brain could inter- (after previous stimulus-reinforcer associ- face sensory systems to action systems. ation learning) as a secondary rewarding Selecting among available rewards or punishing stimulus, then it becomes a with their associated costs, and avoiding goal for action. The animal can then per- punishers with their associated costs, is a form any action (instrumental response) process that can take place both implic- to obtain the reward or to avoid the pun- itly (unconsciously) and explicitly using isher. Thus there is flexibility of action, a language system to enable long-term and this is in contrast with stimulus- plans to be made (Rolls, 1999a, 2005a). response, or habit, learning in which a These many different brain systems, some particular response to a particular stim- involving implicit evaluation of rewards, ulus is learned. It also contrasts with the and others explicit, verbal, conscious, elicitation of species-typical behavioral evaluation of rewards and planned long- responses by sign-releasing stimuli (such term goals, must all enter into the selec- as pecking at a spot on the beak of the tor of behavior (see Fig. 29.2). This selec- parent herring gull in order to be fed; see tor is poorly understood, but it might Tinbergen, 1951), where there is inflexi- include a process of competition between P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

the affective neuroscience of consciousness 835

Cortical Language Motor and Explicit Cortex Planning Actions areas

Association Amygdala Cortex and Orbitofrontal Cortex

Striatum; cingulate cortex Secondary Primary Reinforcers Cortex e.g. taste, touch, pain Premotor Implicit Thalamus Cortex etc Actions Primary Cortex Ventral Learned Striatum Autonomic brainstem Responses

INPUT spinal cord Reflexes Figure 29.2 . Dual routes to the initiation of action in response to rewarding and punishing stimuli. The inputs from different sensory systems to brain structures, such as the orbitofrontal cortex and amygdala, allow these brain structures to evaluate the reward- or punishment-related value of incoming stimuli or of remembered stimuli. The different sensory inputs enable evaluations within the orbitofrontal cortex and amygdala based mainly on the primary (unlearned) reinforcement value for taste, touch, and olfactory stimuli and on the secondary (learned) reinforcement value for visual and auditory stimuli. In the case of vision, the ‘association cortex’ that outputs representations of objects to the amygdala and orbitofrontal cortex is the inferior temporal visual cortex. One route for the outputs from these evaluative brain structures is via projections directly to such structures as the basal ganglia (including the striatum and ventral striatum) to enable implicit, direct behavioural responses based on the reward- or punishment-related evaluation of the stimuli to be made. The second route is via the language systems of the brain, which allow explicit (verbalizable) decisions involving multistep syntactic planning to be implemented.

all the competing calls on output and into the categories happy, sad, fearful, might involve the basal ganglia in the angry, surprised, and disgusted and that brain (see Fig. 29.2 and Rolls, 2005a). this categorization may operate similarly 3. Emotion is motivating, as just described. in different cultures. He also describes For example, fear learned by stimulus- how the facial muscles produce differ- reinforcement association provides the ent expressions. Further investigations of motivation for actions performed to avoid the degree of cross-cultural universality noxious stimuli. of facial expression, its development in 4. Communication. Monkeys, for example, infancy, and its role in social behavior 1991 may communicate their emotional state are described by Izard ( ) and Frid- 1994 to others by making an open-mouth lund ( ). As shown elsewhere (Rolls, 2000 2005 2002 threat to indicate the extent to which c, d, a, Rolls and Deco, ), they are willing to compete for resources, there are neural systems in the amygdala and this may inﬂuence the behavior and overlying temporal cortical visual of other animals. This aspect of emo- areas that are specialized for the face- tion was emphasized by Darwin and has related aspects of this processing. been studied more recently by Ekman 5. Social bonding. Examples of this func- (1982, 1993). He reviews evidence that tion are the emotions associated with the humans can categorize facial expressions attachment of the parents to their young P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

836 the cambridge handbook of consciousness

and the attachment of the young to their tex could perform this function are dis- parents. cussed by Rolls and Treves (1998). 8 6. The current mood state can affect the cog- . Another function of emotion is that, by nitive evaluation of events or memories (see enduring for minutes or longer after a Oatley & Jenkins, 1996). This may facili- reinforcing stimulus has occurred, it may tate continuity in the interpretation of the help produce persistent and continuing reinforcing value of events in the envi- motivation and direction of behavior to help ronment. A hypothesis that backprojec- achieve a goal or goals. tions from parts of the brain involved in 9. Emotion may trigger the recall of memo- emotion such as the orbitofrontal cortex ries stored in neocortical representations. and amygdala implement this evaluation Amygdala backprojections to the cortex of events or memories is described in Emo- could perform this function for emotion Explained (Rolls, 2005a). tion in a way analogous to how the hippocampus implements the retrieval in the 7 . Emotion may facilitate the storage of mem- neocortex of recent (episodic) memories ories. One way this occurs is that episodic (Rolls & Treves 1998; Rolls & Stringer, memory (i.e., one’s memory of particu- 2001). lar episodes) is facilitated by emotional states. This may be advantageous in that storing many details of the prevailing situation when a strong reinforcer is delivered Reward, Punishment, and Emotion may be useful in generating appropriate in Brain Design: An Evolutionary behavior in situations with some simi- Approach larities in the future. This function may be implemented by the relatively non- The theory of the functions of emotion is speciﬁc projecting systems to the cere- further developed in Chapter 3 of Emotion bral cortex and hippocampus, including Explained (Rolls, 2005a), and some of the the cholinergic pathways in the basal fore- points made there help elaborate greatly brain and medial septum and the ascend- on the above section. Rolls (1999a) and ing noradrenergic pathways (see Rolls, (2005a) considers the fundamental question 2005 1998 a; Rolls & Treves, ). A second of why we and other animals are built to way in which emotion may affect the stor- use rewards and punishments to guide or age of memories is that the current emo- determine our behavior. Why are we built tional state may be stored with episodic to have emotions, as well as motivational memories, providing a mechanism for the states? Is there any reasonable alternative current emotional state to affect which around which evolution could have built memories are recalled. A third way that complex animals? In this section I outline emotion may affect the storage of mem- several types of brain design, with differing ories is by guiding the cerebral cortex in degrees of complexity, and suggest that evo- the representations of the world that are lution can operate to inﬂuence action with set up. For example, in the visual system only some of these types of design. it may be useful for perceptual representations or analyzers to be built that are Taxes different from each other if they are associated with different reinforcers, and for A simple design principle is to incorpo- these to be less likely to be built if they rate mechanisms for taxes into the design of have no association with reinforcement. organisms. At their simplest, taxes consist Ways in which backprojections from parts of orientation toward stimuli in the envi- of the brain important in emotion (such as ronment; for example, the bending of a plant the amygdala) to parts of the cerebral cor- toward light, which results in maximum P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

the affective neuroscience of consciousness 837

light collection by its photosynthetic sur- be built by such natural selection to make faces. When just turning rather than loco- responses that will enable them to obtain motion is possible, such responses are called more rewards; that is, to work to obtain tropisms. When locomotion is possible, as stimuli that will increase their fitness. Cor- in animals, taxes include movements toward respondingly, animals must be built to make sources of nutrient and movements away responses that will enable them to escape from hazards, such as very high tempera- from, or learn to avoid, stimuli that will tures. The design principle here is that ani- reduce their fitness. There are likely to be mals have through a process of natural selec- many dimensions of environmental stim- tion built receptors for certain dimensions of uli along which actions can alter fitness. the wide range of stimuli in the environment Each of these dimensions may be a separate and have linked these receptors to mecha- reward-punishment dimension. An example nisms for particular responses in such a way of one of these dimensions might be food that the stimuli are approached or avoided. reward. It increases fitness to be able to sense nutrient need, to have sensors that respond to the taste of food, and to perform behav- Reward and Punishment ioral responses to obtain such reward stimuli As soon as we have an approach toward stim- when in that need or motivational state. Sim- uli at one end of a dimension (e.g., a source of ilarly, another dimension is water reward, in nutrient) and away from stimuli at the other which the taste of water becomes reward- end of the dimension (in this case lack of ing when there is body fluid depletion (see nutrient), we can start to wonder when it is Chapter 6 of Emotion Explained). appropriate to introduce the terms “rewards” With many reward/punishment dimen- and “punishers” for the stimuli at the differ- sions for which actions may be performed ent ends of the dimension. By convention, if (see Table 2.1 of Emotion Explained for a the response consists of a fixed reaction to non-exhaustive list!), a selection mechanism obtain the stimulus (e.g., locomotion up a for actions performed is needed. In this chemical gradient), we shall call this a taxis, sense, rewards and punishers provide a com- not a reward. On the other hand, if an arbi- mon currency for inputs to response selection trary operant response can be performed by mechanisms. Evolution must set the magni- the animal in order to approach the stim- tudes of each of the different reward sys- ulus, then we call this rewarded behavior, tems so that each will be chosen for action and the stimulus the animal works to obtain in such a way as to maximize overall fitness. is a reward. (The operant response can be Food reward must be chosen as the aim for thought of as any arbitrary action the animal action if a nutrient is depleted, but water will perform to obtain the stimulus.) This reward as a target for action must be selected criterion of an arbitrary operant response is if current water depletion poses a greater often tested by bidirectionality. For exam- threat to fitness than the current food deple- ple, if a rat can be trained either to raise or tion. This indicates that each reward must be lower its tail to obtain a piece of food, then carefully calibrated by evolution to have the we can be sure that there is no fixed relation right value in the common currency for the between the stimulus (e.g., the sight of food) competitive selection process. Other types and the response, as there is in a taxis. of behavior, such as sexual behavior, must The role of natural selection in this probe selected sometimes, but probably less fre- cess is to guide animals to build sensory quently, to maximise fitness (as measured by systems that will respond to dimensions of gene transmission into the next generation). stimuli in the natural environment along Many processes contribute to increasing the which actions can lead to a better ability to chances that a wide set of different environ- pass genes on to the next generation; that mental rewards will be chosen over a period is, to increased fitness. The animals must of time, including not only need-related P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

838 the cambridge handbook of consciousness

satiety mechanisms that decrease the lution to operate using reward and pun- rewards within a dimension but also sensory- ishment systems tuned to fitness-related specific satiety mechanisms, which facili- dimensions of the environment, if arbitrary tate switching to another reward stimulus responses are to be made by the animals, (sometimes within and sometimes outside rather than just preprogrammed movements the same main dimension) and attraction to such as tropisms and taxes. Is there any alter- novel stimuli. Finding novel stimuli reward- native to such a reward/punishment-based ing is one way that organisms are encour- system in this evolution by natural selection aged to explore the multidimensional space situation? It is not clear that there is, if the in which their genes are operating. genes are to control behavior. The argument The above mechanisms can be contrasted is that genes can specify actions that will with typical engineering design. In the latter, increase their fitness if they specify the goals the engineer defines the requisite function for action. It would be very difficult for them and then produces special-purpose design in general to specify in advance the particu- features that enable the task to be per- lar responses to be made to each of a myr- formed. In the case of the animal, there iad of different stimuli. This may be why we is a multidimensional space within which are built to work for rewards, avoid punish- many optimisations to increase fitness must ers, and to have emotions and needs (moti- be performed. The solution is to evolve vational states). This view of brain design reward/punishment systems tuned to each in terms of reward and punishment systems dimension in the environment that can built by genes that gain their adaptive value increase fitness if the animal performs the by being tuned to a goal for action offers, appropriate actions. Natural selection guides I believe, a deep insight into how natural evolution to find these dimensions. In con- selection has shaped many brain systems, trast, in the engineering design of a robot and is a fascinating outcome of Darwinian arm, the robot does not need to tune itself thought. to find the goal to be performed. The contrast is between design by evolution, which is ‘blind’ to the purpose of the ani- To What Extent Is Consciousness mal, and design by a designer who speci- Involved in the Different Types fies the job to be performed (cf. Dawkins, of Processing Initiated by 1986). Another contrast is that for the ani- Emotional States? mal the space will be high-dimensional, so that the most appropriate reward for current It might be possible to build a computer behavior (taking into account the costs of that would perform the functions of emo- obtaining each reward) needs to be selected, tions described above and in more detail whereas for the robot arm, the function to by Rolls (1999a, 2000a, 2005a), and yet we perform at any one time is specified by the might not want to ascribe emotional feelings designer. Another contrast is that the behav- to the computer. We might even build the ior (the operant response) most appropriate computer with some of the main process- to obtain the reward must be selected by the ing stages present in the brain and imple- animal, whereas the movement to be made mented using neural networks that simulate by the robot arm is specified by the design the operation of the real neural networks in engineer. the brain (see Rolls & Deco, 2002; Rolls & The implication of this comparison is that Treves, 1998), yet we might not still wish to operation by animals of reward and punish- ascribe emotional feelings to this computer. ment systems tuned to dimensions of the This point often arises in discussions with environment that increase fitness provides a undergraduates, who may say that they fol- mode of operation that can work in organ- low the types of point made above about isms that evolve by natural selection. It is emotion, yet believe that almost the most clearly a natural outcome of Darwinian evo- important aspect of emotions, the feelings, P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

the affective neuroscience of consciousness 839

have not been accounted for nor their neu- we “pay attention” to events in the world, ral basis described. In a sense, the functions we can process those events in some better of reward and punishment in emotional way – that is process or access as opposed behaviour are described by the above types to phenomenal aspects of consciousness – of process and their underlying brain mecha- may be easier to analyse (Allport, 1988; nisms in such structures as the amygdala and Block, 1995; Chalmers, 1996). The puzzle orbitofrontal cortex as described by Rolls of qualia – that is of the phenomenal aspect (1999a, 2000b, c, 2002), but what about the of consciousness – seems to be rather dif- subjective aspects of emotion, what about ferent from normal investigations in science, the pleasure? A similar point arises when in that there is no agreement on criteria we consider the parts of the taste, olfac- by which to assess whether we have made tory, and visual systems in which the reward progress. So, although the aim of this chap- value of the taste, smell, and sight of food ter is to address the issue of consciousness, is represented. One such brain region is especially of qualia, in relation to emotional the orbitofrontal cortex (Rolls, 1997, 1999a, feelings and actions, what is written can- 2000b, 2002, 2005a). Although the neu- not be regarded as being establishable by ronal representation in the orbitofrontal cor- the normal methods of scientiﬁc enquiry. tex is clearly related to the reward value of Accordingly, I emphasize that the view on food, is this where the pleasantness (the sub- consciousness that I describe is only pre- jective hedonic aspect) of the taste, smell, liminary, and theories of consciousness are and sight of food is represented? Again, we likely to develop considerably. Partly for could (in principle at least) build a com- these reasons, this theory of consciousness, puter with neural networks to simulate each at least, should not be taken to have practical of the processing stages for the taste, smell, implications. and sight of food that are described by Rolls (2005a, and more formally in terms of neural networks by Rolls & Deco, 2002; Rolls A Theory of Consciousness & Treves, 1998), and yet would probably not wish to ascribe feelings of pleasantness A starting point is that many actions can be to the system we have simulated on the performed relatively automatically, without computer. apparent conscious intervention. An exam- What is it about neural processing that ple sometimes given is driving a car. Such makes it feel like something when some actions could involve control of behaviour types of information processing are taking by brain systems that are old in evolution- place? It is clearly not a general property ary terms, such as the basal ganglia. It is of processing in neural networks, for there of interest that the basal ganglia (and cere- is much processing – for example that con- bellum) do not have backprojection systems cerned with the control of our blood pres- to most of the parts of the cerebral cortex sure and heart rate – of which we are not from which they receive inputs (see e.g., aware. Is it then that awareness arises when Rolls, 1994; Rolls & Johnstone, 1992; Rolls a certain type of information processing is & Treves, 1998). In contrast, parts of the being performed? If so, what type of infor- brain, such as the hippocampus and amyg- mation processing? And how do emotional dala, which are involved in such functions as feelings and sensory events come to feel episodic memory and emotion, respectively, like anything? These feels are called qualia. about which we can make (verbal) decla- These are great mysteries that have puz- rations (hence declarative memory, Squire, zled philosophers for centuries. They are at 1992) do have major backprojection sys- the heart of the problem of consciousness, tems to the high parts of the cerebral cortex for why it should feel like something at all from which they receive forward projections is the great mystery. Other aspects of con- (Rolls, 1996a, 2000e; Rolls & Deco, 2002; sciousness, such as the fact that often when Rolls & Treves, 1998; Treves & Rolls, 1994). It P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

840 the cambridge handbook of consciousness

may be that evolutionarily newer parts of the discrimination reversal (Hornak et al., 2004; brain, such as the language areas and parts Rolls et al., 1994). In these types of patients, of the prefrontal cortex, are involved in an confabulation may occur, in that a verbal alternative type of control of behaviour, in account of why the action was performed which actions can be planned with the use may be given, which may not be related of a (language) system that allows relatively at all to the environmental event that actu- arbitrary (syntactic) manipulation of seman- ally triggered the action (Gazzaniga, 1988, tic entities (symbols). 1995; Gazzaniga & LeDoux, 1978; Rolls The general view that there are many et al., 1994). routes to behavioural output is supported It is possible that sometimes in normal by the evidence that there are many input humans when actions are initiated as a systems to the basal ganglia (from almost all result of processing in a specialized brain areas of the cerebral cortex) and that neu- region, such as those involved in some types ronal activity in each part of the striatum of rewarded behaviour, the language sys- reﬂects the activity in the overlying corti- tem may subsequently elaborate a coherent cal area (Rolls, 1994, 2005a; Rolls & John- account of why that action was performed stone, 1992; Rolls & Treves, 1998). The evi- (i.e., confabulate). This would be consistent dence is consistent with the possibility that with a general view of brain evolution in different cortical areas, each specialised for which, as areas of the cortex evolve, they are a different type of computation, have their laid on top of existing circuitry connecting outputs directed to the basal ganglia, which inputs to outputs, and in which each level then select the strongest input and map it in this hierarchy of separate input-output into action (via outputs directed for example pathways may control behaviour according to the premotor cortex; Rolls & Johnstone, to the specialised function it can perform 1992; Rolls & Treves, 1998). Within this (see schematic in Fig. 29.2). (It is of interest scheme, the language areas would offer one that mathematicians may get a hunch that of many routes to action, but one that is par- something is correct, yet not be able to ver- ticularly suited to planning actions because balise why. They may then resort to formal, of the syntactic manipulation of semantic more serial, and language-like theorems to entities that may make long-term planning prove the case, and these seem to require possible. A schematic diagram of this sug- conscious processing. This is a further indica- gestion is provided in Fig. 29.2. Consistent tion of a close association between linguistic with the hypothesis of multiple routes to processing and consciousness. The linguistic action, only some of which utilise language, processing need not, as in reading, involve is the evidence that split-brain patients may an inner articulatory loop.) not be aware of actions being performed We may next examine some of the advan- by the “non-dominant” hemisphere (Gaz- tages and behavioural functions that lan- zaniga, 1988, 1995; Gazzaniga & LeDoux, guage, present as the most recently added 1978). Also consistent with multiple routes layer to the above system, would confer. One to action, including non-verbal routes, is major advantage would be the ability to plan the ﬁnding that patients with focal brain actions through many potential stages and to damage – for example to the prefrontal evaluate the consequences of those actions cortex – may perform actions, yet com- without having to perform the actions. For ment verbally that they should not be per- these functions, the ability to form propo- forming those actions (e.g., Hornak et al., sitional statements and to perform syntactic 2003, 2004; Rolls, 1999b; Rolls, Hornak, operations on the semantic representations Wade, & McGrath, 1994). The actions that of states in the world would be important. appear to be performed implicitly, with sur- Also important in this system would be the prise expressed later by the explicit system, ability to have second-order thoughts about include making behavioral responses to a no- the type of thought that I have just described longer rewarded visual stimulus in a visual (e.g., I think that he thinks that . . . ), as this P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

the affective neuroscience of consciousness 841

would allow much better modelling and pretive consciousness (or reflexive conscious- diction of others’ behaviour, and therefore ness or self-consciousness) is the attentive, of planning, particularly planning when it deliberately focused consciousness of one’s involves others1. This capability for higher- mental states. It is noted that not all of order thoughts would also enable reflection the higher-order thoughts need themselves on past events, which would also be useful be conscious (many mental states are not). in planning. In contrast, non-linguistic be- However, according to the analysis, hav- haviour would be driven by learned reinfor- ing a higher-order thought about a lower- cement associations, learned rules, etc., but order thought is necessary for the lower- not by flexible planning for many steps ah- order thought to be conscious. A slightly ead that involves a model of the world inc- weaker position than Rosenthal’s (and mine) luding others’ behaviour. (For an earlier view on this is that a conscious state corresponds that is close to this part of the argument to a first-order thought that has the capacity see Humphrey, 1980.) (The examples of be- to cause a second-order thought or judge- haviour from non-humans that may reflect ment about it (Carruthers, 1996). Another planning may reflect much more limited and position that is close in some respects to that inflexible planning. For example, the dance of Carruthers and the present position is that of the honeybee to signal to other bees the of Chalmers (1996) that awareness is some- location of food may be said to reflect plan- thing that has direct availability for behav- ning, but the symbol manipulation is not ar- ioral control, which amounts effectively for bitrary. There are likely to be interesting him in humans to saying that conscious- examples of non-human primate behaviour, ness is what we can report (verbally) about. perhaps in the great apes, that reflect the This analysis is consistent with the points evolution of an arbitrary symbol-manipu- made above that the brain systems that are lation system that could be useful for flexi- required for consciousness and language are ble planning (cf. Cheney & Seyfarth, 1990.) similar. In particular, a system that can have It is important to state that the language abil- second- or higher-order thoughts about its ity referred to here is not necessarily human own operation, including its planning and verbal language (though this would be an linguistic operation, must itself be a language example). What it is suggested is important processor in that it must be able to bind to planning is the syntactic manipulation of correctly to the symbols and syntax in the symbols, and it is this syntactic manipula- first-order system. According to this expla- tion of symbols that is the sense in which nation, the feeling of anything is the state language is defined and used here. that is present when linguistic process- It is next suggested that this arbitrary sym- ing that involves second- or higher-order bol manipulation using important aspects of thoughts is being performed. language processing and used for planning It might be objected that this captures but not in initiating all types of behaviour some of the process aspects of conscious- is close to what consciousness is about. In ness, what it is good for in an information- particular, consciousness may be the state processing system, but does not capture that arises in a system that can think about the phenomenal aspect of consciousness. (or reflect on) its own (or other peo- (Chalmers, following points made in his ples’) thoughts; that is, in a system capa- 1996 book, might make this point.) I agree ble of second- or higher-order thoughts that there is an element of “mystery” that (Carruthers, 2000; Dennett, 1991; Gennaro, is invoked at this step of the argument, 2004; Rolls, 2004b, 2005a; Rosenthal, 1986, when I say that it feels like something for 1990, 1993, 2004, 2005). On this account, a a machine with higher-order thoughts to be mental state is non-introspectively (i.e., non- thinking about its own first- or lower-order reflectively) conscious if one has a roughly thoughts. But the return point (discussed simultaneous thought that one is in that further below) is the following: If a human mental state. Following from this, introspec- with second-order thoughts is thinking about P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

842 the cambridge handbook of consciousness

its own first-order thoughts, surely it is very dif- connectionist modelling) and implemented ficult for us to conceive that this would NOT a first-order linguistic system that would feel like something? (Perhaps the higher-order not be conscious. [Possible examples of lan- thoughts in thinking about the first-order guage processing that might be performed thoughts would need to have some sense non-consciously include computer programs of continuity or self, so that the first-order implementing aspects of language or ritual- thoughts would be related to the same sys- ized human conversations (e.g., about the tem that had thought of something else a few weather). These might require syntax and minutes ago. But even this continuity aspect correctly grounded semantics and yet be per- may not be a requirement for conscious- formed non-consciously. A more complex ness. Humans with anterograde amnesia can- example, illustrating that syntax could be not remember what they felt a few min- used, might be, “If A does X, then B will utes ago, yet their current state does feel like probably do Y, and then C would be able to something.) do Z.” A first-order language system could It is suggested that part of the evolu- process this statement. Moreover, the first- tionary adaptive significance of this type of order language system could apply the rule higher-order thought is that it enables cor- usefully in the world, provided that the sym- rection of errors made in first-order linguis- bols in the language system (A, B, X, Y etc.) tic or in non-linguistic processing. Indeed, are grounded (have meaning) in the world.] the ability to reflect on previous events In line with the argument on the adaptive is extremely important for learning from value of higher-order thoughts and thus con- them, including setting up new long-term sciousness given above, that they are useful semantic structures. It was shown above that for correcting lower-order thoughts, I now the hippocampus may be a system for such suggest that correction using higher-order “declarative” recall of recent memories. Its thoughts of lower-order thoughts would close relation to “conscious” processing in have adaptive value primarily if the lower- humans (Squire, 1992, has classified it as a order thoughts are sufficiently complex to declarative memory system) may be simply benefit from correction in this way. The that it enables the recall of recent mem- nature of the complexity is specific: It should ories, which can then be reflected upon involve syntactic manipulation of symbols, in conscious, higher-order processing (Rolls, probably with several steps in the chain, and 1996a). Another part of the adaptive value of the chain of steps should be a one-off (or in a higher-order thought system may be that, American, “one-time”, meaning used once) by thinking about its own thoughts in a given set of steps, as in a sentence or in a partic- situation, it may be able to better understand ular plan used just once, rather than a set the thoughts of another individual in a simi- of well-learned rules. The first- or lower- lar situation and therefore predict that indi- order thoughts might involve a linked chain vidual’s behaviour better (cf. Barlow, 1997; of “if” . . . “then” statements that would be Humphrey, 1980, 1986). involved in planning, an example of which As a point of clarification, I note has been given above. It is partly because that according to this theory, a language- complex lower-order thoughts, such as these processing system is not sufficient for con- that involve syntax and language, would sciousness. What defines a conscious system benefit from correction by higher-order according to this analysis is the ability to thoughts that I suggest that there is a close have higher-order thoughts, and a first-order link between this reflective consciousness language processor (that might be perfectly and language. The hypothesis is that by competent at language) would not be con- thinking about lower-order thoughts, the scious in that it could not think about its higher-order thoughts can discover what own or others’ thoughts. One can perfectly may be weak links in the chain of rea- well conceive of a system that obeyed the soning at the lower-order level and, having rules of language (which is the aim of much detected the weak link, might alter the plan P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

the affective neuroscience of consciousness 843

to see if doing so gives better success. In our control or even overseeing program check- example above, if it transpired that C could ing the operation of the first program. We not do Z, how might the plan have failed? would want to say that in such a situation it Instead of having to go through endless ran- would feel like something to be running the dom changes to the plan to see if by trial and higher-level control program only if the first- error some combination does happen to pro- order program was symbolically performing duce results, what I am suggesting is that by operations on the world and receiving input thinking about the previous plan, one might, about the results of those operations, and for example using knowledge of the situation if the higher-order system understood what and the probabilities that operate in it, guess the first-order system was trying to do in that the step where the plan failed was that the world. The issue of symbol grounding is B did not in fact do Y. So by thinking about considered further by Rolls (2005a, Chapter the plan (the first- or lower-order thought), 10; 2000a). The symbols (or symbolic rep- one might correct the original plan in such a resentations) are symbols in the sense that way that the weak link in that chain, that “B they can take part in syntactic processing. will probably do Y”, is circumvented. The symbolic representations are grounded To draw a parallel with neural networks, in the world in that they refer to events there is a “credit assignment” problem in in the world. The symbolic representations such multistep syntactic plans: If the whole must have a great deal of information about plan fails, how does the system assign credit what is referred to in the world, including or blame to particular steps of the plan? the quality and intensity of sensory events, (In multilayer neural networks, the credit emotional states, etc. The need for this is assignment problem is that if errors are being that the reasoning in the symbolic system specified at the output layer, the problem must be about stimuli, events, and states, and arises about how to propagate back the error remembered stimuli, events, and states; for to earlier, hidden layers of the network to the reasoning to be correct, all the informa- assign credit or blame to individual synaptic tion that can affect the reasoning must be connection; see Rolls & Deco, 2002; Rumel- represented in the symbolic system, includ- hart, Hinton, & Williams, 1986). The sug- ing for example just how light or strong the gestion is that this is the function of higher- touch was, etc. order thoughts and is why systems with Indeed, it is pointed out in Emotion higher-order thoughts evolved. The sugges- Explained that it is no accident that the tion I then make is that if a system were shape of the multidimensional phenomenal doing this type of processing (thinking about (sensory etc.) space maps so clearly onto its own thoughts), it would then be very the space defined by neuronal activity in plausible that it should feel like something sensory systems, for if this were not the to be doing this. I even suggest to the reader case, reasoning about the state of affairs in that it is not plausible to suggest that it the world would not map onto the world would not feel like anything to a system if and would not be useful. Good exam- it were doing this. ples of this close correspondence are found Two other points in the argument should in the taste system, in which subjective be emphasized for clarity. One is that the space maps simply onto the multidimen- system that is having syntactic thoughts sional space represented by neuronal firing about its own syntactic thoughts (higher- in primate cortical taste areas. In particu- order syntactic thoughts or HOSTs) would lar, if a three-dimensional space reflecting have to have its symbols grounded in the the distances between the representations of real world for it to feel like something to different tastes provided by macaque neu- be having higher-order thoughts. The inten- rons in the cortical taste areas is constructed, tion of this clarification is to exclude sys- then the distances between the subjective tems, such as a computer running a pro- ratings by humans of different tastes is very gram when there is in addition some sort of similar (Plata-Salaman, Smith-Swintowsky, P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

844 the cambridge handbook of consciousness

& Scott, 1996; Smith-Swintowsky, Plata- ation is very different from the slow learning Salaman, & Scott, 1991; Yaxley, Rolls, of fixed rules by trial and error or the appli- & Sienkiewicz, 1990; Kadohisa, Rolls, & cation of fixed rules by a supervisory part of Verhagen, 2005). Similarly, the changes in a computer program. human subjective ratings of the pleasantness This analysis does not yet give an account of the taste, smell, and sight of food paral- for sensory qualia (“raw sensory feels”; for lel very closely the responses of neurons in example, why “red” feels red), for emotional the macaque orbitofrontal cortex (see Emo- qualia (e.g., why a rewarding touch pro- tion Explained, Chapter 5). The represen- duces an emotional feeling of pleasure), or tations in the first-order linguistic proces- for motivational qualia (e.g., why food depri- sor that the HOSTs process include beliefs vation makes us feel hungry). The view I sug- (for example “Food is available”, or at least gest on such qualia is as follows. Information representations of this), and the HOST sys- processing in and from our sensory systems tem would then have available to it the con- (e.g., the sight of the colour red) may be rel- cept of a thought (so that it could represent evant to planning actions using language and “I believe [or there is a belief] that food is the conscious processing thereby implied. available”). Given that these inputs must be represented However, as summarised by Rolls (2000a, in the system that plans, we may ask whether 2005a), representations of sensory processes it is more likely that we would be conscious and emotional states must be processed by of them or that we would not. I suggest the first-order linguistic system, and HOSTs that it would be a very special-purpose sys- may be about these representations of sen- tem that would allow such sensory inputs, sory processes and emotional states capable and emotional and motivational states, to be of taking part in the syntactic operations of part of (linguistically based) planning and the first-order linguistic processor. Such sen- yet remain unconscious. It seems to be much sory and emotional information may reach more parsimonious to hold that we would the first-order linguistic system from many be conscious of such sensory, emotional, and parts of the brain, including those parts, such motivational qualia because they would be as the orbitofrontal cortex and amygdala, being used (or are available to be used) in that are implicated in emotional states (see this type of (linguistically based) higher- Emotion Expalined, Fig. 10.3 and p. 408). order thought processing, and this is what When the sensory information is about the I propose. identity of the taste, the inputs to the first- The explanation for emotional and moti- order linguistic system must come from the vational subjective feelings or qualia that primary taste cortex, in that the identity this discussion has led toward is thus that of taste, independent of its pleasantness (in they should be felt as conscious because they that the representation is independent of enter into a specialised linguistic symbol- hunger), must come from the primary taste manipulation system that is part of a higher- cortex. In contrast, when the information order thought system that is capable of that reaches the first-order linguistic sys- reflecting on and correcting its lower-order tem is about the pleasantness of taste, it thoughts involved, for example, in the flex- must come from the secondary taste cor- ible planning of actions. It would require a tex, in that there the representation of taste very special machine to enable this higher- depends on hunger. order linguistically based thought process- The second clarification is that the plan ing, which is conscious by its nature, to would have to be a unique string of steps, in occur without the sensory, emotional, and much the same way as a sentence can be a motivational states (which must be taken unique and one-off (or one-time) string of into account by the higher-order thought words. The point here is that it is helpful system) becoming felt qualia. The qualia to be able to think about particular one-off are thus accounted for by the evolution of plans and to correct them; this type of oper- the linguistic system that can reflect on and P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

the affective neuroscience of consciousness 845

correct its own lower-order processes and are not affected by hunger, but the pleasant- thus has adaptive value. ness of its taste is decreased to zero (neu- This account implies that it may be espe- tral) or even becomes unpleasant after we cially those animals with a higher-order have eaten it to satiety. The implication of belief and thought system and with linguis- this is that for quality and intensity informa- tic symbol manipulation that have qualia. tion about taste, we must be conscious of It may be that much non-human animal what is represented in the primary taste cor- behaviour, provided that it does not require tex (or perhaps in another area connected flexible linguistic planning and correction to it that bypasses the secondary taste corby reflection, could take place accord- tex), and not of what is represented in the ing to reinforcement guidance (using e.g., secondary taste cortex. In contrast, for the stimulus-reinforcement association learn- pleasantness of a taste, consciousness of this ing in the amygdala and orbitofrontal could not reflect what is represented in the cortex; Rolls, 1990, 1996b, 1999a, 2000b, c, primary taste cortex, but instead what is rep- 2002, 2004a, 2005a), and rule following resented in the secondary taste cortex (or (implemented, for example, using habit or in an area beyond it). The same argument stimulus-response learning in the basal gan- arises for reward in general, and therefore glia, Rolls, 1994; Rolls & Johnstone, 1992). for emotion, which in primates is not rep- Such behaviours might appear very similar resented early on in processing in the sen- to human behaviour performed in similar sory pathways (nor in or before the infe- circumstances, but would not imply qualia. rior temporal cortex for vision), but in the It would be primarily by virtue of a sys- areas to which these object analysis system for reflecting on flexible, linguistic, plan- tems project, such as the orbitofrontal cor- ning behaviour that humans (and animals tex, where the reward value of visual stim- close to humans, with demonstrable syntac- uli is reflected in the responses of neurons tic manipulation of symbols and the abil- to visual stimuli (see Rolls, 1990, 1995a, b, ity to think about these linguistic processes) 2005a). It is also of interest that reward sig- would be different from other animals and nals (e.g., the taste of food when we are would have evolved qualia. hungry) are associated with subjective feel- For processing in a part of our brain to ings of pleasure (see Rolls, 1990, 1995a,b, be able to reach consciousness, appropri- 1997, 2005a). I suggest that this correspon- ate pathways must be present. Certain con- dence arises because pleasure is the subjec- straints arise here. For example, in the tive state that represents in the conscious sensory pathways, the nature of the repre- system a signal that is positively reinforcing sentation may change as it passes through (rewarding) and that inconsistent behaviour a hierarchy of processing levels, and to be would result if the representations did not conscious of the information in the form in correspond to a signal for positive reinforce- which it is represented in early processing ment in both the conscious and the non- stages, the early processing stages must have conscious processing systems. access to the part of the brain necessary for Do these arguments mean that the con- consciousness. An example is provided by scious sensation of, for example, taste quality processing in the taste system. In the primate (i.e., identity and intensity) is represented or primary taste cortex, neurons respond to occurs in the primary taste cortex, and of the taste independently of hunger, yet in the sec- pleasantness of taste in the secondary taste ondary taste cortex, food-related taste neu- cortex, and that activity in these areas is suf- rons (e.g., responding to sweet taste) only ficient for conscious sensations (qualia) to respond to food if hunger is present and grad- occur? I do not suggest this at all. Instead the ually stop responding to that taste during arguments I have put forward above suggest feeding to satiety (see Rolls, 1997, 2005a; that we are only conscious of representations Rolls & Scott, 2003). Now the quality of when we have higher-order thoughts about the tastant (sweet, salt etc.) and its intensity them. The implication then is that pathways P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

846 the cambridge handbook of consciousness

must connect from each of the brain areas no immediate action is possible (see Rolls, in which information is represented about 1990, 1995b, 2005a). Grief may be adap- which we can be conscious, to the system tive by leading to a cessation of the for- that has the higher-order thoughts, which merly rewarded behaviour and thus facil- as I have argued above, requires language. itating the possible identification of other Thus, in the example given, there must positive reinforcers in the environment. In be connections to the language areas from humans, grief may be particularly potent the primary taste cortex, which need not because it becomes represented in a system be direct, but which must bypass the sec- that can plan ahead and can understand the ondary taste cortex, in which the informa- enduring implications of the loss. (Think- tion is represented differently (see Rolls, ing about or verbally discussing emotional 1995a, 2005a). There must also be pathways states may also in these circumstances help, from the secondary taste cortex, not neces- because doing so can lead toward the iden- sarily direct, to the language areas so that tification of new or alternative reinforcers we can have higher-order thoughts about and the realization that for example negative the pleasantness of the representation in the consequences may not be as bad as feared.) secondary taste cortex. There would also This account of consciousness also leads need to be pathways from the hippocampus, to a suggestion about the processing that implicated in the recall of declarative mem- underlies the feeling of free will. Free will ories, back to the language areas of the cere- would in this scheme involve the use of lan- bral cortex (at least via the cortical areas that guage to check many moves ahead on a num- receive backprojections from the amygdala, ber of possible series of actions and their orbitofrontal cortex, and hippocampus, see outcomes, and then with this information to Fig. 29.2), which would in turn need con- make a choice from the likely outcomes of nections to the language areas). different possible series of actions. (If in con- One question that has been discussed is trast choices were made only on the basis whether there is a causal role for conscious- of the reinforcement value of immediately ness (e.g., Armstrong & Malcolm, 1984). available stimuli, without the arbitrary syn- The position to which the above arguments tactic symbol manipulation made possible lead is that indeed conscious processing by language, then the choice strategy would does have a causal role in the elicitation of be much more limited, and we might not behaviour, but only under the set of circum- want to use the term “free will”, as all the stances when higher-order thoughts play a consequences of those actions would not role in correcting or influencing lower-order have been computed.) It is suggested that thoughts and that it is a property of the when this type of reflective, conscious infor- higher-order thought system that it feels mation processing is occurring and leading like something when it is operating. As we to action, the system performing this pro- have seen, some behavioural responses can cessing and producing the action would have be elicited when there is not this type of to believe that it could cause the action, for reflective control of lower-order processing otherwise inconsistencies would arise, and nor indeed any contribution of language (see the system might no longer try to initiate further Rolls, 2003, for relations between action. This belief held by the system may implicit and explicit processing). There are partly underlie the feeling of free will. At many brain processing routes to output other times, when other brain modules are regions, and only one of these involves con- initiating actions (in the implicit systems), scious, verbally represented processing that the conscious processor (the explicit system) can later be recalled (see Fig. 29.2). may confabulate and believe that it caused It is of interest to comment on how the the action or at least give an account (possi- evolution of a system for flexible planning bly wrong) of why the action was initiated. might affect emotions. Consider grief, which The fact that the conscious processor may may occur when a reward is terminated and have the belief even in these circumstances P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

the affective neuroscience of consciousness 847

that it initiated the action may be a property ple inconsistent goals from day to day or a of it being inconsistent for a system that can poor recall memory. It may be noted that the take overall control using conscious verbal ability to recall previous steps in a plan and processing to believe that it was overridden bring them into the conscious, higher-order by another system. This may be the reason thought system is an important prerequisite why confabulation occurs. for long-term planning that involves check- In the operation of such a free will system, ing each step in a multistep process. the uncertainties introduced by the limited These are my initial thoughts on why we information possible about the likely out- have consciousness and are conscious of sen- comes of series of actions, and the inabil- sory, emotional, and motivational qualia, as ity to use optimal algorithms when combin- well as qualia associated with first-order lin- ing conditional probabilities, would be much guistic thoughts. However, as stated above, more important factors than whether the one does not feel that there are straight- brain operates deterministically or not. (The forward criteria in this philosophical field operation of brain machinery must be rel- of enquiry for knowing whether the sug- atively deterministic, for it has evolved to gested theory is correct. Therefore, it is likely provide reliable outputs for given inputs.) that theories of consciousness will continue Before leaving these thoughts, it may be to undergo rapid development, and current worth commenting on the feeling of con- theories should not be taken to have practi- tinuing self-identity that is characteristic of cal implications. humans. Why might this arise? One suggestion is that if one is an organism that can think about its own long-term multi- Dual Routes to Action step plans, then for those plans to be executed consistently and thus adaptively, the According to the present formulation, there goals of the plans would need to remain sta- are two types of route to action performed in ble, as would memories of how far one had relation to reward or punishment in humans proceeded along the execution path of each (see also Rolls, 2003, 2005a). Examples of plan. If one felt each time one came to exe- such actions include emotional and motiva- cute, perhaps on another day, the next step tional behaviour. of a plan, that the goals were different; or The first route is via the brain systems that if one did not remember which steps had have been present in non-human primates, already been taken in a multi step plan, the such as monkeys, and to some extent in other plan would never be usefully executed. So, mammals for millions of years. These sys- given that it does feel like something to be tems include the amygdala and, particularly doing this type of planning using higher- well developed in primates, the orbitofrontal order thoughts, it would have to feel as if cortex. These systems control behaviour in one were the same agent, acting toward the relation to previous associations of stimuli same goals, from day to day. Thus it is sug- with reinforcement. The computation that gested that the feeling of continuing self- controls the action thus involves assessment identity falls out of a situation in which there of the reinforcement-related value of a stim- is an actor with consistent long-term goals ulus. This assessment may be based on a and long-term recall. If it feels like anything number of different factors. One is the pre- to be the actor, according to the suggestions vious reinforcement history, which involves of the higher-order thought theory, then it stimulus-reinforcement association learning should feel like the same thing from occasion using the amygdala and its rapid updating to occasion to be the actor, and no special especially in primates using the orbitofrontal further construct is needed to account for cortex. This stimulus-reinforcement associ- self-identity. Humans without such a feel- ation learning may involve quite specific ing of being the same person from day to information about a stimulus; for example, day might be expected to have for exam- of the energy associated with each type of P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

848 the cambridge handbook of consciousness

food, by the process of conditioned appetite and depending on the investment in time and satiety (Booth, 1985). A second is that the partner is willing to put into making the current motivational state; for example, the touch pleasurable, a sign that indicates whether hunger is present, whether other the commitment and value for the partner needs are satisfied, etc. of the relationship). After the reward value A third factor that affects the computed of the stimulus has been assessed in these reward value of the stimulus is whether that ways, behaviour is then initiated based on reward has been received recently. If it has approach toward or withdrawal from the been received recently but in small quantity, stimulus. A critical aspect of the behaviour this may increase the reward value of the produced by this type of system is that it is stimulus. This is known as incentive moti- aimed directly toward obtaining a sensed or vation or the “salted peanut” phenomenon. expected reward, by virtue of connections The adaptive value of such a process is that to brain systems, such as the basal ganglia, this positive feedback of reward value in which are concerned with the initiation of the early stages of working for a particu- actions (see Fig. 29.2). The expectation may lar reward tends to lock the organism onto of course involve behaviour to obtain stimuli behaviour being performed for that reward. associated with reward, which might even be This means that animals that are for exam- present in a chain. ple almost equally hungry and thirsty will Now part of the way in which the show hysteresis in their choice of action, behaviour is controlled with this first route rather than continually switching from eat- is according to the reward value of the out- ing to drinking and back with each mouth- come. At the same time, the animal may ful of water or food. This introduction of only work for the reward if the cost is not hysteresis into the reward evaluation system too high. Indeed, in the field of behavioural makes action selection a much more efficient ecology, animals are often thought of as process in a natural environment, for con- performing optimally on some cost-benefit stantly switching between different types of curve (see e.g., Krebs & Kacelnik, 1991). behaviour would be very costly if all the dif- This does not at all mean that the animal ferent rewards were not available in the same thinks about the rewards and performs a place at the same time. (For example, walk- cost-benefit analysis using a lot of thoughts ing a half-mile between a site where water about the costs, other rewards available, and was available and a site where food was avail- their costs, etc. Instead, it should be taken able after every mouthful would be very to mean that in evolution, the system has inefficient.) The amygdala is one structure evolved so that the way in which the reward that may be involved in this increase in the varies with the different energy densities or reward value of stimuli early on in a series amounts of food and the delay before it is of presentations, in that lesions of the amyg- received can be used as part of the input to a dala (in rats) abolish the expression of this mechanism that has also been built to track reward-incrementing process, which is nor- the costs of obtaining the food (e.g., energy mally evident in the increasing rate of work- loss in obtaining it, risk of predation, etc.), ing for a food reward early on in a meal (Rolls and to then select, given many such types & Rolls, 1982). of reward and the associated cost, the cur- A fourth factor is the computed absolute rent behaviour that provides the most “net value of the reward or punishment expected reward”. Part of the value of having the com- or being obtained from a stimulus; for exam- putation expressed in this reward-minus- ple, the sweetness of the stimulus (set by cost form is that there is then a suitable “cur- evolution so that sweet stimuli will tend rency”, or net reward value, to enable the to be rewarding because they are generally animal to select the behaviour with currently associated with energy sources) or the pleas- the most net reward gain (or minimal aver- antness of touch (set by evolution to be sive outcome). pleasant according to the extent to which it The second route in humans invol- brings animals of the opposite sex together, ves a computation with many “if . . . then” P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

the affective neuroscience of consciousness 849

statements used to implement a plan to the one in which explicit, declarative pro- obtain a reward. In this case, the reward cessing occurs. Processing in this system is may actually be deferred as part of the plan, frequently associated with reason and ratio- which might involve working first to obtain nality, in that many of the consequences of one reward and only then to work for a sec- possible actions can be taken into account. ond more highly valued reward, if this was The actual computation of how rewarding thought to be overall an optimal strategy in a particular stimulus or situation is or will terms of resource usage (e.g., time). In this be probably still depends on activity in the case, syntax is required, because the many orbitofrontal cortex and amygdala, as the symbols (e.g., names of people) that are part reward value of stimuli is computed and of the plan must be linked or bound cor- represented in these regions and verbalised rectly. Such linking might be of the form: expressions of the reward (or punishment) “If A does this, then B is likely to do this, value of stimuli are dampened by damage to and this will cause C to do this . . . ”. The these systems. (For example, damage to the requirement of syntax for this type of plan- orbitofrontal cortex renders painful input ning implies that an output to language sys- still identifiable as pain, but without the tems in the brain is required for this type of strong affective, “unpleasant”, reaction to it.) planning (see Fig. 29.2). Thus, the explicit This language system that enables long-term language system in humans may allow work- planning may be contrasted with the first sys- ing for deferred rewards by enabling use of a tem in which behaviour is directed at obtain- one-off, individual plan appropriate for each ing the stimulus (including the remembered situation. Another building block for such stimulus) that is currently most reward- planning operations in the brain may be the ing, as computed by brain structures that type of short-term memory in which the pre- include the orbitofrontal cortex and amyg- frontal cortex is involved. For example in dala. There are outputs from this system, non-human primates, this short-term mem- perhaps those directed at the basal ganglia, ory may be of where in space a response which do not pass through the language sys- has just been made. A development of this tem, and behaviour produced in this way is type of short-term response memory sys- described as implicit, and verbal declarations tem in humans to enable multiple short- cannot be made directly about the reasons term memories to be held in place correctly, for the choice made. When verbal declara- preferably with the temporal order of the tions are made about decisions made in this different items in the short-term memory first system, those verbal declarations may coded correctly, may be another building be confabulations, reasonable explanations, block for the multiple-step “if . . . then” type or fabrications of reasons why the choice was of computation required to form a multi- made. These reasonable explanations would ple step plan. Such short-term memories be generated to be consistent with the sense are implemented in the (dorsolateral and of continuity and self that is a characteristic inferior convexity) prefrontal cortex of non- of reasoning in the language system. human primates and humans (see Goldman- The question then arises of how deci- Rakic, 1996; Petrides, 1996), and may be sions are made in animals such as humans part of the reason why prefrontal cortex that have both the implicit, direct reward- damage impairs planning (see Shallice & based and the explicit, rational, planning Burgess, 1996). systems (see Fig. 29.2). One particular sit- Of these two routes (see Fig. 29.2), it is uation in which the first, implicit, system the second that I have suggested above is re- may be especially important is when rapid lated to consciousness. The hypothesis is that reactions to stimuli with reward or punish- consciousness is the state that arises by virtue ment value must be made, for then the direct of having the ability to think about one’s connections from structures such as the own thoughts, which has the adaptive value orbitofrontal cortex to the basal ganglia may of enabling one to correct long multistep allow rapid actions (e.g., Rolls et al., 1994). syntactic plans. This latter system is thus The implicit system may be used to guide P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

850 the cambridge handbook of consciousness

action when there may be too many factors Alexander, 1975, 1979; Trivers, 1976, 1985; to be taken into account easily by the and the review by Nesse & Lloyd, 1992). explicit, rational, planning, system. In con- Another example might be that the explicit trast, when the implicit system continually system might because of its long-term plans makes errors, it would then be beneficial influence the implicit system to increase its for the organism to switch from automatic, response to, for example, a positive rein- direct action based on obtaining what the forcer. One way in which the explicit sys- orbitofrontal cortex system decodes as being tem might influence the implicit system is the most positively reinforcing choice cur- by setting up the conditions in which, for rently available, to the explicit conscious example when a given stimulus (e.g., person) control system that can evaluate with its is present, positive reinforcers are given to long-term planning algorithms what action facilitate stimulus-reinforcement association should be performed next. Indeed, it would learning by the implicit system of the per- be adaptive for the explicit system to regu- son receiving the positive reinforcers. Con- larly be assessing performance by the more versely, the implicit system may influence automatic system and to switch itself in the explicit system, for example by high- to control behaviour quite frequently, as lighting certain stimuli in the environment otherwise the adaptive value of having the that are currently associated with reward, to explicit system would be less than optimal. guide the attention of the explicit system to Another factor that may influence the bal- such stimuli. ance between control by the implicit and However, it may be expected that there explicit systems is the presence of phar- is often a conflict between these systems, in macological agents such as alcohol, which that the first, implicit system is able to guide may alter the balance toward control by behaviour particularly to obtain the great- the implicit system, may allow the implicit est immediate reinforcement, whereas the system to influence more the explanations explicit system can potentially enable imme- made by the explicit system, and may within diate rewards to be deferred and longer- the explicit system alter the relative value it term, multistep, plans to be formed. This places on caution and restraint versus com- type of conflict will occur in animals with a mitment to a risky action or plan. syntactic planning ability; that is, in humans There may also be a flow of influence and any other animals that have the ability from the explicit, verbal system to the to process a series of “if . . . then” stages of implicit system, in that the explicit system planning. This is a property of the human may decide on a plan of action or strategy language system, and the extent to which it and exert an influence on the implicit sys- is a property of non-human primates is not tem that will alter the reinforcement eval- yet fully clear. In any case, such conflict may uations made by and the signals produced be an important aspect of the operation of at by the implicit system. An example of this least the human mind, because it is so essen- might be that if a pregnant woman feels tial for humans to correctly decide, at every that she would like to escape a cruel mate, moment, whether to invest in a relationship but is aware that she may not survive in or a group that may offer long-term bene- the jungle, then it would be adaptive if the fits or whether to directly pursue immediate explicit system could suppress some aspects benefits (Nesse & Lloyd, 1992). of her implicit behaviour toward her mate, As Nesse and Lloyd (1992) describe, ana- so that she does not give signals that she is lysts have come to a somewhat similar posi- displeased with her situation. (In the litera- tion, for they hold that intrapsychic con- ture on self-deception, it has been suggested flicts usually seem to have two sides, with that unconscious desires may not be made impulses on one side and inhibitions on the explicit in consciousness (or may actually other. Analysts describe the source of the be repressed), so as not to compromise the impulses as the id, and the modules that explicit system in what it produces; see e.g., inhibit the expression of impulses, because P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

the affective neuroscience of consciousness 851

of external and internal constraints, the ego I will be punished”; see Dawkins, 1993). and superego, respectively (Leak & Christo- Nevertheless, the possible advantage of such pher, 1982; Trivers, 1985; see also Nesse Machiavellian planning could be one of the & Lloyd, 1992,p.613). The superego can adaptive guiding factors in evolution that be thought of as the conscience, whereas provided advantage to a multistep, syntactic the ego is the locus of executive functions system that enables long-term planning, the that balance satisfaction of impulses with best example of such a system being human anticipated internal and external costs. The language. present position differs because it is based However, another, not necessarily exclu- on the identification of dual routes to action sive, advantage for the evolution of a linguis- implemented by different systems in the tic multistep planning system could well be brain, each with its own selective advantage. not Machiavellian planning, but planning for Some investigations in non-human pri- social cooperation and advantage. Perhaps mates on deception have been interpreted in general an “if . . . then” multistep syntactic as showing that animals can plan to deceive planning ability is useful primarily in evolu- others (see e.g., Griffin, 1992); that is, to uti- tion in social situations of the type: “if X does lize “Machiavellian intelligence”. For exam- this, then Y does that; then I would/should ple, a baboon may “deliberately” mislead do that, and the outcome would be . . . ”. It another animal in order to obtain a resource, is not yet at all clear whether such plan- such as food (e.g., by screaming to sum- ning is required in order to explain the social mon assistance in order to have a compet- behaviour of social animals, such as hunting animal chased from a food patch) or ing dogs or socialising monkeys (Dawkins, sex (e.g., a female baboon who moved very 1993). However, in humans there is evidence gradually into a position from which the that members of “primitive” hunting tribes dominant male could not see her groom- spend hours recounting tales of recent events ing a subadult baboon; see Dawkins, 1993). (perhaps who did what, when, who then did The attraction of the Machiavellian argu- what, etc.), perhaps to help learn from expe- ment is that the behaviour for which it rience about good strategies, which is neces- accounts seems to imply that there is a sary for example when physically weak men concept of another animal’s mind and that take on large animals (see Pinker & Bloom, one animal is trying occasionally to mislead 1992). Thus, social cooperation may be as another, which implies some planning. How- powerful a driving force in the evolution of ever, such observations tend by their nature syntactical planning systems as Machiavel- to be field-based and may have an anecdo- lian intelligence. What is common to both tal character, in that the previous experience is that they involve social situations. How- of the animals in this type of behaviour, and ever, such a syntactic planning system would the reinforcements obtained, are not known have advantages not only in social systems, (Dawkins, 1993). It is possible for exam- for such planning may be useful in obtaining ple that some behavioural responses that resources purely in a physical (non-social) appear to be Machiavellian may have been world. An example might be planning how the result of previous instrumental learn- to cross terrain given current environmen- ing in which reinforcement was obtained for tal constraints in order to reach a particular particular types of response or of observa- place. tional learning, with again learning from the The thrust of this argument thus is that outcome observed. However, in any case, much complex animal, including human, most examples of Machiavellian intelligence behaviour can take place using the implicit, in non-human primates do not involve mul- non-conscious route to action. We should tiple stages of “if . . . then” planning requiring be very careful not to postulate intentional syntax to keep the symbols apart (but may states (i.e., states with intentions, beliefs, involve learning of the type “if the domi- and desires) unless the evidence for them is nant male sees me grooming a subadult male, strong, and it seems to me that a flexible, P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

852 the cambridge handbook of consciousness

one-off linguistic processing system that can “on the fly” reasoning procedure that errors handle propositions is needed for intentional that cannot be corrected easily by reward states. What the explicit, linguistic system or punishment received at the end of the does allow is exactly this flexible, one-off, reasoning, need ‘thoughts about thoughts’– multistep planning-ahead type of computa- that is, some type of supervisory and mon- tion, which allows us to defer immediate itoring process – to detect where errors in rewards based on such a plan. the reasoning have occurred. This suggestion This discussion of dual routes to action on the adaptive value in evolution of such a has been with respect to the behaviour pro- higher-order linguistic thought process for duced. There is of course, in addition, a multistep planning ahead, and correcting third output of brain regions, such as the such plans, may also be different from ear- orbitofrontal cortex and amygdala involved lier work. Put another way, this point is that in emotion, which is directed to produc- credit assignment when reward or punishing autonomic and endocrine responses. ment is received is straightforward in a one- Although it has been argued by Rolls (2005a, layer network (in which the reinforcement Chapter 2) that the autonomic system is can be used directly to correct nodes in error not normally in a circuit through which or responses). However, credit assignment is behavioural responses are produced (i.e., very difficult in a multistep linguistic process against the James-Lange and related somatic executed once “on the fly”. Very complex theories), there may be some influence from mappings in a multilayer network can be effects produced through the endocrine sys- learned if hundreds of learning trials are pro- tem (and possibly the autonomic system, vided. But once these complex mappings are through which some endocrine responses learned, their success or failure in a new situ- are controlled) on behaviour, or on the ation on a given trial cannot be evaluated and dual systems just discussed that control corrected by the network. Indeed, the com- behaviour. For example, during female plex mappings achieved by such networks orgasm the hormone oxytocin may be (e.g., backpropagation nets) mean that after released, and this may influence the implicit training they operate according to fixed rules system to help develop positive reinforce- and are often quite impenetrable and inflexi- ment associations and thus attachment. ble. In contrast, to correct a multistep, single occasion, linguistically based plan or procedure, recall of the steps just made in the rea- Discussion soning or planning and perhaps of related episodic material needs to occur, so that Some ways in which the current theory the link in the chain that is most likely to may be different from other related theo- be in error can be identified. This may be ries follow. The current theory holds that it part of the reason why there is a close rela- is higher-order syntactic thoughts (HOSTs) tion between declarative memory systems, that are closely associated with conscious- which can explicitly recall memories, and ness, and this may differ from Rosen- consciousness. thal’s higher-order thoughts (HOTs) the- Some computer programs may have ory (Rosenthal, 1986, 1990, 1993, 2005) supervisory processes. Should these count in the emphasis on language. Language in as higher-order linguistic thought processes? the current theory is defined by syntactic My current response to this question is that manipulation of symbols and does not nec- they should not, to the extent that they oper- essarily imply verbal language. The type of ate with fixed rules to correct the opera- language required in the theory described tion of a system that does not itself involve here is sometimes termed “mentalese” by linguistic thoughts about symbols grounded philosophers (Fodor, 1994). The reason that semantically in the external world. If on the strong emphasis is placed on language is that other hand it were possible to implement it is as a result of having a multistep flexible on a computer such a higher-order linguistic P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

the affective neuroscience of consciousness 853

thought supervisory correction process to 30–50 ms, and cortical neurons need fire correct first-order one-off linguistic thoughts for only this long during the identification with symbols grounded in the real world, of objects (Rolls, 2003; Rolls & Deco, 2002; then this process would prima facie be con- Rolls & Tovee, 1994; Rolls, Tovee, & Panz- scious. If it were possible in a thought exper- eri, 1999; Rolls & Treves, 1998; Rolls, Tovee, iment to reproduce the neural connectivity Purcell, Stewart, & Azzopardi, 1999; Tovee and operation of a human brain on a com- & Rolls, 1995; Tovee, Rolls, Treves, & Bel- puter, then prima facie it would also have lis, 1993): These are rather short time win- the attributes of consciousness. It might con- dows for the expression of multiple separate tinue to have those attributes for as long as populations of synchronized neurons). In power was applied to the system. addition, oscillations, at least, are not an Another possible difference from earlier obvious property of neuronal firing in theories is that raw sensory feels are sug- the primate temporal cortical visual areas gested to arise as a consequence of hav- involved in the representation of faces ing a system that can think about its own and objects (Tovee & Rolls, 1992; see fur- thoughts. Raw sensory feels, and subjective ther Rolls & Treves, 1998).] However, the states associated with emotional and moti- hypothesis that syntactic binding is neces- vational states, may not necessarily arise first sary for consciousness is one of the postu- in evolution. lates of the theory I am describing (for the A property often attributed to conscious- system I describe must be capable of cor- ness is that it is unitary. The current the- recting its own syntactic thoughts), and the ory would account for this property by the fact that the binding must be implemented limited syntactic capability of neuronal net- in neuronal networks may well place limita- works in the brain, which render it diffi- tions on consciousness, which lead to some cult to implement more than a few syntac- of its properties, such as its unitary nature. tic bindings of symbols simultaneously (see The postulate of Crick and Koch (1990) that McLeod, Plunkett, & Rolls, 1998 and Rolls & oscillations and synchronization are neces- Treves, 1998). This limitation makes it dif- sary bases of consciousness could thus be ficult to run several “streams of conscious- related to the present theory if it turns ness” simultaneously. In addition, given that out that oscillations or neuronal synchro- a linguistic system can control behavioural nization is the way the brain implements output, several parallel streams might pro- syntactic binding. However, the fact that duce maladaptive behaviour (apparent as, oscillations and neuronal synchronization for example, indecision), and might be are especially evident in anaesthetized cats selected against. The close relation between does not impress as strong evidence that and the limited capacity of both the stream oscillations and synchronization are criti- of consciousness and auditory-verbal short cal features of consciousness, for most peo- term memory may be that both imple- ple would hold that anaesthetized cats are ment the capacity for syntax in neural net- not conscious. The fact that oscillations works. Whether syntax in real neuronal and synchronization are much more diffi- networks is implemented by temporal bind- cult to demonstrate in the temporal corti- ing (see von der Malsburg, 1990) is still cal visual areas of awake behaving monkeys very much an unresolved issue (Rolls & (Aggelopoulos, Franco, & Rolls, 2005) might Deco, 2002; Rolls & Treves, 1998; Rolls, just mean that during evolution to primates 2007a). [For example, the code about which the cortex has become better able to avoid visual stimulus has been shown can be parasitic oscillations, as a result of developing read off from the end of the visual sys- better feedforward and feedback inhibitory tem without taking the temporal aspects circuits (see Rolls & Deco, 2002 and Rolls & of the neuronal firing into account; much Treves, 1998). of the information about which stimulus The current theory holds that conscious- is shown is available in short periods of ness arises by virtue of a system that can P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

854 the cambridge handbook of consciousness

think linguistically about its own linguistic erty that occurs when a linguistic system is thoughts. The advantages for a system of thinking about its lower-order thoughts. The being able to do this have been described, hypothesis that it does feel like something and this has been suggested as the reason when this processing is taking place is at least why consciousness evolved. The evidence to some extent testable: Humans performing that consciousness arises by virtue of hav- this type of higher-order linguistic processing a system that can perform higher-order ing, for example recalling episodic memories linguistic processing is however, and I think and comparing them with current circum- may remain, circumstantial. (Why must it stances, who denied being conscious, would feel like something when we are perform- prima facie constitute evidence against the ing a certain type of information processing? theory. Most humans would find it very The evidence described here suggests that it implausible, though, to posit that they could does feel like something when we are per- be thinking about their own thoughts, and forming a certain type of information pro- reflecting on their own thoughts, without cessing, but does not produce a strong rea- being conscious. This type of processing does son for why it has to feel like something. appear to be for most humans to be neces- It just does, when we are using this lin- sarily conscious. guistic processing system capable of higher- Finally, I provide a short specification order thoughts.) The evidence, summarized of what might have to be implemented in above, includes the points that we think a neural network to implement conscious of ourselves as conscious when for exam- processing. First, a linguistic system, not ple we recall earlier events, compare them necessarily verbal, but implementing syntax with current events, and plan many steps between symbols implemented in the envi- ahead. Evidence also comes from neuro- ronment would be needed (i.e., a mentalese logical cases, from for example split-brain language system). Then a higher- order patients who may confabulate conscious sto- thought system also implementing syntax, ries about what is happening in their other, able to think about the representations in non-language hemisphere; and from frontal the first-order language system, and able to lobe patients who can tell one consciously correct the reasoning in the first-order lin- what they should be doing, but nevertheless guistic system in a flexible manner would may be doing the opposite. (The force of this be needed. So my view is that conscious- type of case is that much of our behaviour ness can be implemented in neural networks, may normally be produced by routes about (and that this is a topic worth discussing), which we cannot verbalize and are not con- but that the neural networks would have to scious about.) implement the type of higher-order linguis- This raises the issue of the causal role of tic processing described in this chapter. consciousness. Does consciousness cause our behaviour?2 The view that I currently hold is that the information processing that is Conclusions and Comparisons related to consciousness (activity in a linguistic system capable of higher-order thoughts It is suggested that it feels like something to and used for planning and correcting the be an organism or machine that can think operation of lower-order linguistic systems) about its own (linguistic and semantically can play a causal role in producing our based) thoughts. It is suggested that qualia, behaviour (see Fig. 29.2). It is, I postulate, raw sensory, and emotional feels arise sec- a property of processing in this system (capa- ondary to having evolved such a higher- ble of higher-order thoughts) that it feels like order thought system and that sensory and something to be performing that type of pro- emotional processing feels like something cessing. It is in this sense that I suggest that because it would be unparsimonious for it consciousness can act causally to influence to enter the planning, higher-order thought our behaviour: Consciousness is the prop- system and not feel like something. The P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

the affective neuroscience of consciousness 855

adaptive value of having sensory and emo- Damasio’s theory of emotion (1994, tional feelings, or qualia, is thus suggested 2003) is similar to the James-Lange the- to be that such inputs are important to the ory (and is therefore subject to some of long-term planning, explicit processing sys- the same objections), but holds that the tem. Raw sensory feels and subjective states peripheral feedback is used in decision mak- associated with emotional and motivational ing rather than in consciousness. He does states may not necessarily arise ﬁrst in evolu- not formally deﬁne emotions, but holds that tion. Some issues that arise in relation to this body maps and representations are the basis theory are discussed by Rolls (2000a, 2004b, of emotions. When considering conscious- 2005a); reasons why the ventral visual sys- ness, he assumes that all consciousness is self- tem is more closely related to explicit than consciousness (Damasio, 2003,p.184) and implicit processing are considered by Rolls that the foundational images in the stream (2003) and by Rolls and Deco (2002); and of the mind are images of some kind of body reasons why explicit, conscious processing event, whether the event happens in the may have a higher threshold in sensory pro- depth of the body or in some specialized cessing than implicit processing are consid- sensory device near its periphery (Damasio, ered by Rolls (2003, 2005b). 2003,p.197). His theory does not appear The theory is different from some other to be a fully testable theory, in that he sus- theories of consciousness (Carruthers, 2000; pects that “the ultimate quality of feelings, a Gennaro, 2004; Rosenthal, 1990, 1993, part of why feelings feel the way they feel, 2004, 2005) in that it provides an account of is conferred by the neural medium” (Dama- the evolutionary, adaptive value of a higher- sio 2003,p.131). Thus, presumably if pro- order thought system in helping solve a cesses he discusses (Damasio, 1994, 2003) credit assignment problem that arises in a were implemented in a computer, then the multistep syntactic plan, links this type of computer would not have all the same prop- processing to consciousness, and therefore erties with respect to consciousness as the emphasises a role for syntactic processing real brain. In this sense he appears to be argu- in consciousness. The type of syntactic pro- ing for a non-functionalist position and for cessing need not be at the natural language something crucial about consciousness being level (which implies a universal grammar), related to the particular biological machin- but could be at the level of mentalese (Rolls, ery from which the system is made. It is in 1999a, 2004b, 2005a; cf Fodor, 1994). this respect that the theory seems somewhat The theory described here is also differ- intangible. ent from other theories of affect, which in LeDoux’s (1996) approach to emotion is some cases consider the issue of conscious- largely (to quote him) one of automatic- ness. James (1884) and Lange (1885/1922) ity, with emphasis on brain mechanisms held that emotional feelings arise when feed- involved in the rapid, subcortical mech- back from the periphery (about for exam- anisms involved in fear. LeDoux, in line ple heart rate) reach the brain, but had no with Johnson-Laird (1988) and Baars (1988), theory of why some stimuli and not others emphasises the role of working memory produced the peripheral changes, and thus in consciousness, where he views working of why some but not other events produce memory as a limited-capacity serial proces- emotion. Moreover, the evidence that feed- sor that creates and manipulates symbolic back from peripheral autonomic and pro- representations (p. 280). He thus holds that prioceptive systems is essential for emotions much emotional processing is unconscious is very weak; for example, blocking periph- and that when it becomes conscious it is eral feedback does not eliminate emotions, because emotional information is entered and producing peripheral (e.g., autonomic) into a working memory system. However, changes does not elicit emotion (Reisen- LeDoux (1996) concedes that conscious- zein, 1983; Rolls, 1999a, 2005a; Schachter ness, especially its phenomenal or subjective & Singer, 1962). nature, is not completely explained by the P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

856 the cambridge handbook of consciousness

computational processes that underlie work- Alexander, R. D. (1975). The search for a general ing memory (p. 281; see Rolls, 2007b). theory of behavior. Behavioral Sciences, 20, 77– Panksepp’s (1998) approach to emotion 100. has its origins in neuroethological investiga- Alexander, R. D. (1979). Darwinism and human tions of brainstem systems that when acti- affairs. Seattle: University of Washington Press. vated lead to behaviours like fixed action Allport, A. (1988). What concept of consci- patterns, including escape, flight, and fear ousness? In A. J. Marcel & E. Bisiach (Eds.), behaviour. His views about consciousness Consciousness in contemporary science (pp. 159– 182 include the postulate that feelings may ). Oxford: Oxford University Press. emerge when endogenous sensory and emo- Armstrong, D. M., & Malcolm, N. (1984). Con- tional systems within the brain that receive sciousness and causality. Oxford: Blackwell. direct inputs from the outside world, as well Baars, B. J. (1988). A cognitive theory of conscious- as the neurodynamics of the SELF (a Simple ness. New York: Cambridge University Press. Ego-type Life Form), begin to reverberate Barlow, H. B. (1997). Single neurons, commu- with each other’s changing neuronal firing nal goals, and consciousness. In M. Ito, Y. rhythms (Panksepp, 1998,p.309). Miyashita, & E. T. Rolls (Eds.), Cognition, computation, and consciousness (pp. 121–136). Oxford: Oxford University Press. Block, N. (1995). On a confusion about a function Notes of consciousness. Behavioral and Brain Sciences, 18, 227–247. 1. Second order thoughts are thoughts about Booth, D. A. (1985). Food-conditioned eat- thoughts. Higher order thoughts refer to sec- ing preferences and aversions with interocep- ond order, third order etc. thoughts about tive elements: Learned appetites and satieties. thoughts... Annals of the New York Academy of Sciences, 2. This raises the issue of the causal relation 443, 22–37. between mental events and neurophysiolog- Carruthers, P. (1996). Language, thought and con- ical events, part of the mind-body problem. sciousness. Cambridge: Cambridge University My view is that the relation between mental Press. events and neurophysiological events is similar Carruthers, P. (2000). Phenomenal consciousness. (apart from the problem of consciousness) to Cambridge: Cambridge University Press. the relation between the program running in a Chalmers, D. J. (1996). The conscious mind. computer and the hardware on the computer. Oxford: Oxford University Press. In a sense, the program causes the logic gates to 1990 move to the next state. This move causes the Cheney, D. L., & Seyfarth, R. M. ( ). How program to move to its next state. Effectively, monkeys see the world. Chicago: University of we are looking at different levels of what is Chicago Press. overall the operation of a system, and causality Crick, F. H. C., & Koch, C. (1990). Towards a neu- can usefully be understood as operating both robiological theory of consciousness. Seminars within levels (causing one step of the program in the Neurosciences, 2 , 263–275. to move to the next), as well as between lev- Damasio, A. R. (1994). Decartes’ error. New York: els (e.g., software to hardware and vice versa). Putnam. This is the solution I propose to this aspect of Damasio, A. R. (2003). Looking for Spinoza. the mind-body (or mind-brain) problem. London: Heinemann. Dawkins, M. S. (1993). Through our eyes only? The search for animal consciousness. Oxford: Free- References man. Dawkins, R. (1986). The blind watchmaker. Harlow: Longman. Aggelopoulos, N. C., Franco, L., & Rolls, E. T. 1991 (2005). Object perception in natural scenes: Dennett, D. C. ( ). Consciousness explained. Encoding by inferior temporal cortex simul- London: Penguin. taneously recorded neurons. Journal of Neuro- Ekman, P. (1982). Emotion in the human face (2nd physiology, 93, 1342–1357, ed.). Cambridge: Cambridge University Press. P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

the affective neuroscience of consciousness 857

Ekman, P. (1993). Facial expression and emotion. Izard, C. E. (1991). The psychology of emotions. American Psychologist, 48, 384–392. New York: Plenum. Fodor, J. A. (1994). The elm and the expert: Men- James, W. (1884). What is an emotion? Mind, 9, talese and its semantics. Cambridge, MA: MIT 188–205. Press. Johnson-Laird, P. N. (1988). The computer and the Fridlund, A. J. (1994). Human facial expression: mind: An introduction to cognitive science. Cam- An evolutionary view. San Diego: Academic bridge, MA: Harvard University Press. Press. Kadohisa, M., Rolls, E. T., & Verhagen, J. V. Frijda, N. H. (1986). The emotions. Cambridge: (2005). Neuronal representations of stimuli in Cambridge University Press. the mouth: the primate insular taste cortex, Gazzaniga, M. S. (1988). Brain modularity: orbitofrontal cortex, and amygdala. Chemical Towards a philosophy of conscious experience. Senses, 30, 401–419. In A. J. Marcel & E. Bisiach (Eds.), Conscious- Krebs, J. R., & Kacelnik, A. (1991). Decision ness in contemporary science (pp. 218–238). making. In J. R. Krebs & N. B. Davies (Eds.), Oxford: Oxford University Press. Behavioural ecology (3rd ed., pp. 105–136). Gazzaniga, M. S. (1995). Consciousness and the Oxford: Blackwell. cerebral hemispheres. In M. S. Gazzaniga (Ed.), Lange, C. (1922). The emotions. In E. Dunlap The cognitive neurosciences (pp. 1392–1400). (Ed.), The emotions. Baltimore: Williams & Cambridge, MA: MIT Press. Wilkins. (Original work published 1885) 1978 Gazzaniga, M. S., & LeDoux, J. ( ). The inte- Lazarus, R. S. (1991). Emotion and adaptation. grated mind. New York: Plenum. New York: Oxford University Press. 2004 Gennaro, R. J. (Ed.) ( ). Higher order theories Leak, G. K., & Christopher, S. B. (1982). Freudian of consciousness. Amsterdam: Benjamins. psychoanalysis and sociobiology: A synthesis. Goldman-Rakic, P. S. (1996). The prefrontal American Psychologist, 37, 313–322. landscape: Implications of functional archi- LeDoux, J. (1996). The emotional brain. New tecture for understanding human mentation York: Simon and Shuster. and the central executive. Philosophical Trans- Mackintosh, N. J. (1983). Conditioning and asso- actions of the Royal Society B, 351, 1445– ciative learning. Oxford: Oxford University 1453. Press. Gray, J. A. (1975). Elements of a two-process theory McLeod, P., Plunkett, K., & Rolls, E. T. (1998). of learning. London: Academic Press. Introduction to connectionist modelling of cogni- Gray, J. A. (1987). The psychology of fear and stress tive processes. Oxford: Oxford University Press. (2nd ed.). Cambridge: Cambridge University 1967 Principles of behavioral Press. Millenson, J. R. ( ). analysis. New York: MacMillan. Grifﬁn, D. R. (1992). Animal minds. Chicago: 1992 University of Chicago Press. Nesse, R. M., & Lloyd, A. T.( ). The evolution of psychodynamic mechanisms. In J. H Barkow, Hornak, J., Bramham, J., Rolls, E. T., Morris, L. Cosmides, & J. Tooby (Eds.), The adapted R. G., O’Doherty, J., Bullock, P. R., et al. mind (pp. 601–624). New York: Oxford Uni- (2003). Changes in emotion after circum- versity Press. scribed surgical lesions of the orbitofrontal and 1996 cingulate cortices. Brain, 126, 1691–1712. Oatley, K., & Jenkins, J. M. ( ). Understanding emotions. Oxford: Blackwell. Hornak, J., O’Doherty, J., Bramham, J., Rolls, E. T., Morris, R. G., Bullock, P. R., et al. (2004). Panksepp, J. (1998). Affective neuroscience: The Reward-related reversal learning after surgical foundations of human and animal emotions. excisions in orbitofrontal and dorsolateral pre- New York: Oxford University Press. frontal cortex in humans. Journal of Cognitive Petrides, M. (1996). Specialized systems for the Neuroscience, 16, 463–478. processing of mnemonic information within Humphrey, N. K. (1980). Nature’s psychologists. the primate frontal cortex. Philosophical Trans- In B. D. Josephson & V. S. Ramachandran actions of the Royal Society B, 351, 1455– (Eds.), Consciousness and the physical world 1462. (pp. 57–80). Oxford: Pergamon. Pinker, S., & Bloom, P. (1992). Natural language Humphrey, N. K. (1986). The inner eye. London: and natural selection. In J. H. Barkow, L. Cos- Faber. mides, & J. Tooby (Eds.), The adapted mind P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

858 the cambridge handbook of consciousness

(pp. 451–493). New York: Oxford University Rolls, E. T. (2000c). Neurophysiology and func- Press. tions of the primate amygdala, and the neural Plata-Salaman, C. R., Smith-Swintosky, V. L., & basis of emotion. In J. P. Aggleton (Ed.), The Scott, T. R. (1996). Gustatory neural coding in amygdala: A functional analysis (2nd ed., pp. the monkey cortex: Mixtures. Journal of Neu- 447–478). Oxford: Oxford University Press. rophysiology, 75, 2369–2379. Rolls, E. T. (2000d). Functions of the primate Reisenzein, R. (1983). The Schachter theory of temporal lobe cortical visual areas in invariant emotion: Two decades later. Psychology Bul- visual object and face recognition. Neuron, 27, 205 218 letin, 94, 239–264. – . 2000 Rolls, B. J., & Rolls, E. T. (1982). Thirst. Rolls, E. T. ( e). Hippocampo-cortical and Cambridge: Cambridge University Press. cortico-cortical backprojections. Hippocampus, 10, 380–388. Rolls, E. T. (1986a). Neural systems involved in 2002 emotion in primates. In R. Plutchik & H. Keller- Rolls, E. T. ( ). The functions of the man (Eds.), Emotion: Theory, research, and orbitofrontal cortex. In D. T. Stuss & R. T. experience: Vol. 3. Biological foundations of emo- Knight (Eds.) Principles of frontal lobe function 354 375 tion (pp. 125–143). New York: Academic Press. (pp. – ). New York: Oxford University Press. Rolls, E. T. (1986b). A theory of emotion, and its 2003 application to understanding the neural basis Rolls, E. T. ( ). Consciousness absent and of emotion. In Y. Oomura (Ed.), Emotions. neu- present: A neurophysiological exploration. 95 106 ral and chemical control (pp. 325–344). Tokyo: Progress in Brain Research, 144, – . Japan Scientiﬁc Societies Press. Rolls, E. T. (2004a). The functions of the Brain and Cognition, Rolls, E. T. (1990). A theory of emotion, and its orbitofrontal cortex. 55, 11 29 application to understanding the neural basis – . 2004 of emotion. Cognition and Emotion, 4, 161–190. Rolls, E. T. ( b). A higher order syntactic thought (HOST) theory of consciousness. In Rolls, E. T. (1994). Neurophysiology and cog- R. J. Gennaro (Ed.), Higher order theories of nitive functions of the striatum. Revue Neu- consciousness (pp. 137–172). Amsterdam: John rologique, 150, 648–660. Benjamins. Rolls, E. T. (1995a). Central taste anatomy and Rolls, E. T. (2005a). Emotion explained. Oxford: neurophysiology. In R. L. Doty (Ed.), Hand- Oxford University Press. book of olfaction and gustation (pp. 549–573). 2005 New York: Dekker. Rolls, E. T. ( b). Consciousness absent and present: A neurophysiological exploration of Rolls, E. T. (1995b). A theory of emotion and masking. In H. Ogmen & B. G. Breitmeyer consciousness, and its application to under- (Eds.), The ﬁrst half second: The microgenesis and standing the neural basis of emotion. In M. S. temporal dynamics of unconscious and conscious Gazzaniga (Ed.), The cognitive neurosciences visual processes (pp 89–108). Cambridge, MA: (pp. 1091–1106). Cambridge, MA: MIT Press. MIT Press. Rolls, E. T. (1996a). A theory of hippocampal Rolls, E. T. (2007a). Memory, attention, and function in memory. Hippocampus, 6, 601–620. decision-making. Oxford: Oxford University 1996 Rolls, E. T. ( b). The orbitofrontal cortex. press. Philosophical Transactions of the Royal Society Rolls, E. T. (2007b). Emotion, higher-order syn- B, 351, 1433–1444. tactic thoughts, and consciousness. In M. 1997 Rolls, E. T. ( ). Taste and olfactory processing Davies & L. Weiskrantz (Eds.), Frontiers of con- in the brain. Critical Reviews in Neurobiology, sciousness. Oxford: Oxford University Press. 11, 263–287. Rolls, E. T., & Deco, G. (2002). Computational 1999 Rolls, E. T. ( a). The brain and emotion. neuroscience of vision. Oxford: Oxford Univer- Oxford: Oxford University Press. sity Press. 1999 Rolls, E. T. ( b). The functions of the Rolls, E. T., Hornak, J., Wade, D., & McGrath, 301 312 orbitofrontal cortex. Neurocase, 5, – . J. (1994). Emotion-related learning in patients Rolls, E. T. (2000a). Precis´ of The Brain and Emo- with social and emotional changes associated tion. Behavioral and Brain Sciences, 23, 177– with frontal lobe damage. Journal of Neurology, 233. Neurosurgery and Psychiatry, 57, 1518–1524. Rolls, E. T. (2000b). The orbitofrontal cortex and Rolls, E. T., & Johnstone, S. (1992). Neurophysio- reward. Cerebral Cortex, 10, 284–294. logical analysis of striatal function. In G. Vallar, P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

the affective neuroscience of consciousness 859

S. F. Cappa, & C. Wallesch (Eds.), Neuropsy- tional state. Psychological Review, 69, 379– chological disorders associated with subcortical 99 lesions (pp. 61–97). Oxford: Oxford University Shallice, T., & Burgess, P. (1996). The domain of Press. supervisory processes and temporal organiza- Rolls, E. T., & Scott, T. R. (2003). Central tion of behaviour. Philosophical Transactions of taste anatomy and neurophysiology. In R. L. the Royal Society of London B, 351, 1405–1411. Doty (Ed.), Handbook of olfaction and gustation Smith-Swintosky, V. L., Plata-Salaman, C. R., & (2nd ed., pp. 679–705). New York: Dekker. Scott, T. R. (1991). Gustatory neural coding in Rolls, E. T., & Stringer, S. M. (2001). A model the monkey cortex: Stimulus quality. Journal of the interaction between mood and memory. of Neurophysiology, 66, 1156–1165. Network: Computation in Neural Systems, 12, Squire, L. R. (1992). Memory and the hippo- 89–109. campus: A synthesis from findings with rats, Rolls, E. T., & Tovee, M. J. (1994). Processing monkeys and humans. Psychological Review, 99, speed in the cerebral cortex and the neuro- 195–231. physiology of visual masking. Proceedings of the Strongman, K. T.(1996). The psychology of emotion Royal Society of London B, 257, 9–15. (4th ed.). Chichester: Wiley. Rolls, E. T., Tovee, M. J., & Panzeri, S. (1999). The Tinbergen, N. (1951). The study of instinct. neurophysiology of backward visual masking: Oxford: Clarendon Press. Information analysis. Journal of Cognitive Neu- Tovee, M. J., & Rolls, E. T. (1992). Oscillatory 335 346 roscience, 11, – . activity is not evident in the primate temporal Rolls, E. T., Tovee, M. J., Purcell, D. G., Stewart, visual cortex with static stimuli. Neuroreport, A. L., & Azzopardi, P. (1994). The responses of 3, 369–372. neurons in the temporal cortex of primates, and Tovee, M. J., & Rolls, E. T. (1995). Information face identification and detection. Experimental encoding in short firing rate epochs by single 474 484 Brain Research, 101, – . neurons in the primate temporal visual cortex. Rolls, E. T., & Treves, A. (1998). Neural networks Visual Cognition, 2 , 35–58. and brain function. Oxford: Oxford University Tovee, M. J., Rolls, E. T., Treves, A., & Bellis, Press. R. P. (1993). Information encoding and the Rosenthal, D. M. (1990). A theory of conscious- responses of single neurons in the primate tem- ness. ZIF Report No. 40. Bielefeld: Germany: poral visual cortex. Journal of Neurophysiology, Zentrum fur Interdisziplinaire Forschung. 70, 640–654. Rosenthal, D. M. (1986). Two concepts of con- Treves, A., & Rolls, E. T. (1994). A computa- sciousness. Philosophical Studies, 49, 329–359. tional analysis of the role of the hippocampus Rosenthal, D. M. (1993). Thinking that one in memory. Hippocampus, 4, 374–391. thinks. In M. Davies & G. W. Humphreys Trivers, R. L. (1976). Foreword. In R. Dawkins (Eds.), Consciousness (pp. 197–223). Oxford: (Ed.), The selfish gene. Oxford: Oxford Univer- Blackwell. sity Press. Rosenthal, D. M. (2004). Varieties of higher order Trivers, R. L. (1985). Social evolution. California: theory. In R. J. Gennaro (Ed.), Higher order the- Benjamin/Cummings. 17 44 ories of consciousness (pp. – ). Amsterdam: von der Malsburg, C. (1990). A neural archi- John Benjamins. tecture for the representation of scenes. In Rosenthal, D. M. (2005) Consciousness and mind. J. L. McGaugh, N. M. Weinberger, & G. Lynch Oxford: Oxford University Press. (Eds.), Brain organization and memory: Cells, Rumelhart, D. E., Hinton, G. E., & Williams, systems and circuits (pp. 356–372). New York: R. J. (1986). Learning internal representations Oxford University Press. by error backpropagation. In D. E. Rumelhart Weiskrantz, L. (1968). Emotion. In L. Weiskrantz & J. L. McClelland, et al. (Eds.), Parallel (Ed.), Analysis of behavioural change (pp. 50– distributed processing: Explorations in the micro- 90). London: Harper and Row. 1 structure of cognition,Vol . Cambridge, MA: Yaxley, S., Rolls, E. T., & Sienkiewicz, Z. J. MIT Press. (1990). Gustatory responses of single neurons Schacter, S., & Singer, J. E. (1962). Cognitive, in the insula of the macaque monkey. Journal of social and physiological determinants of emo- Neurophysiology, 63, 689–700. P1: JzG 0521857430c29 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 11:50

860 P1: KAE 0521857430c30 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 3:10

D. Social Neuroscience of Consciousness

861 P1: KAE 0521857430c30 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 3:10

862 P1: KAE 0521857430c30 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 3:10

CHAPTER 30 Consciousness: Situated and Social

Ralph Adolphs

Abstract Introduction: The Problem of Conscious Experience Conscious experience is usually presumed to depend in a direct way only on events in No other topic has stirred so much debate the brain. For instance, thought experimen- as has the question of conscious experience. ts in which events in the brain are held cons- Approaches have ranged from denying that tant, but events outside the brain are chan- we are conscious at all to claiming that we ged, are assumed to leave the nature of the already know how we are conscious. One conscious experience also unchanged. Here reason for the diversity of approaches is I argue to the contrary that conscious expe- that different thinkers take up fairly differ- rience depends in a direct way on a wider set ent questions, a situation resulting at least of events, including events outside the brain in part because there is no single problem and in the past. Nonetheless, not all events of consciousness, but rather quite a hetero- matter equally. For events outside the brain, geneous collection of problems. To simplify social relations appear especially important. the discussion in this chapter, I will discuss A complex social environment has shaped consciousness in the restricted sense of con- much of the architecture of the brain and can scious sensory experience, such as the expe- be partially internalized by the brain via sim- riences of seeing a certain color, hearing a cer- ulation. This view suggests that understand- tain sound, smelling a rose, and the various ing conscious experience requires under- other examples usually cited by those who standing the phylogenetic, ontogenetic, and want to make a convincing case for so-called current social relations in which an individ- qualia. Furthermore, I restrict the aspect of ual brain is embedded. conscious experience under scrutiny to what

863 P1: KAE 0521857430c30 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 3:10

864 the cambridge handbook of consciousness

has been the focus in some recent philosoph- ism’s own point of view at the time that the ical discussions. Many readers are familiar sunset is seen by the organism. Although it with the views of the philosopher Thomas is important to distinguish this meaning of Nagel, who has defended the view that our subjectivity from its more common every- difficulty in providing a scientific explana- day usage, it is also critical not to attach to it tion of conscious experience arises from an properties that would prevent any possibility essential feature of conscious experience: its of an epistemologically objective account of subjectivity (Nagel, 1974, 1993). Of course, the phenomenon and that would bar reduc- there are other reasons why explaining con- tive explanation. It is easy to jump from the scious experience is not easy, but subjectiv- realization that conscious experience is sub- ity looms as the paramount obstacle, because jective in Nagel’s sense to the conclusion that we cannot at present even conceive of how therefore it is not material. The “hard prob- this feature could be made commensurate lem” of consciousness arises because what it with our scientific, objective understanding is like to have a conscious experience seems of the world. so clearly contingent on all possible objec- By subjectivity, Nagel means that fea- tive accounts of the physical or functional ture of conscious experience that generates machinery on which the conscious experi- a point of view on the world by the owner ence supervenes. Let us consider one such of the experience. In Nagel’s words, there apparent example. is “something it is like to be” an organ- David Chalmers has championed thought ism undergoing a conscious experience. This experiments about the conceivability of ontological notion of subjectivity needs to “zombies.” These are conceivably possible be distinguished from epistemic subjectivity (logically possible, rather than nomologically (the way the word is used colloquially) – that possible) creatures identical to a human in is, the meaning that something is a matter of all physical and functional respects, minus opinion. Ontological subjectivity, the mean- the conscious experience (Chalmers, 1997). ing of the term that concerns us here, is a par- All objective accounts of consciousness we ticular mode of existence of a phenomenon have at present, all accounts that describe (existence as seen “from the inside”). It is consciousness from “the outside” as it were, important also to acknowledge that an expe- fail to provide an understanding of how con- rience is rendered phenomenal in virtue of a sciousness fits into the rest of the world as a collection of properties, the instantiation of natural feature of its constituents. They fail more than one of which is necessary for actu- to do so precisely because in every case we ally having a conscious experience. Thus, can imagine a system meeting all the crite- phenomenality consists of quite a few dif- ria of the objective theory, yet not be con- ferent properties or features, and it is possi- scious. The converse is also conceivable: that ble to distinguish conceptually among these the conscious experience remains the same (cf. Metzinger, 2003). For our purposes it is while its physical substrate is very different important to note that subjectivity, in the (or even non-existent, as in the conceivabil- sense in which Nagel is using the term and ity of disembodied spirits). in which I intend to use it here, is separate It is instructive to compare what we from ownership or from self-consciousness. might want to say about conscious expe- It may well be that subjectivity itself, upon rience to what we would say about how further analysis, fragments into a collection mental states refer to events in the world. of disparate properties. One much-defended answer about proposi- For present purposes it suffices to use our tional attitudes, such as beliefs, is to say that pretheoretical concept that there is some- they indeed are not specified or identified by thing it is like for a conscious organism in the pointing to events occurring in the brain, but act of becoming aware of a sensory object, rather by pointing to events outside of the such as opening one’s eyes and seeing a red brain (in fact, outside of the person and the sunset; something it is like from the organ- current physical events). The philosopher P1: KAE 0521857430c30 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 3:10

consciousness: situated and social 865

Hilary Putnam (1979) has defended this example I mentioned above. If Chalmers view on the basis of his famous “twin earth” by some unexplained and mysterious acci- examples, and a similar thought experiment dent were to suddenly lose all ability to has been provided by the late Donald David- have conscious experiences, but nonethe- son with the invention of “swampman”; let less remained physically identical, his behav- us consider the latter (Davidson, 1987). ior would, we suppose, remain unchanged Swampman is a molecular duplicate of and so would his beliefs. (We could further Davidson that was created by a spontaneous suppose that this would be so even if his assembly of the requisite atoms when light- physical constituency changed but his func- ning struck a swamp. If physical determin- tional architecture remained invariant, but ism is true, Swampman does and says exactly for present purposes this is irrelevant.) What the things that Davidson would do and he tells us and believes that was true before say; their behavior is indistinguishable. But is still true in the absence of the conscious Davidson tells us that his swamp duplicate experience. He would, like Swampman, be could nonetheless not have any of the beliefs physically indistinguishable from his earlier that he, the real Davidson, has, because the incarnation, and he would say all the things swamp being does not share any of David- the prior, conscious person would say. But, son’s history that is necessary for having such unlike Swampman, he would still be right beliefs. That is, although Swampman would, about the beliefs he avows (perhaps with like Davidson, say that it believes it was one exception: he might still, and now incor- born on such and such a date, is married rectly, believe that he is conscious). to such and such a person, etc., all of these These examples seem to show that the beliefs could not refer to the said events, representational properties of brain events because they depend on the correct histori- can come apart despite invariance in their cal embeddedness, which Swampman lacks physical composition. Moreover, they can and Davidson has. So, Swampman’s physical come apart in directions opposite to the way constituency would cause him to utter the in which conscious experience might come words, “I remember playing in the sand when apart: Swampman arguably still has con- I was a little boy,” a statement he believes scious experiences indistinguishable from to be true and one that would be true if those Davidson would have, despite the vac- Davidson uttered it (let us suppose). But in uous representations of his past; the zombie Swampman’s case it could not possibly refer on the other hand has (by stipulation) no to the stated event: Swampman didn’t play conscious experiences although he still has in the sand as a little boy; instead, he was the same representational powers. spontaneously created only a few minutes It is not hard (I think) to see how these ago in the swamp (and it would certainly examples work in the case of intentional sound bizarre to suppose that Swampman’s states like beliefs, because it is not hard to utterance somehow refers to Davidson’s acknowledge that beliefs might not be intrin- early life). The same goes for all the other sic states. Beliefs, like desires, wants, and things that Swampman might honestly say other such attitudes, are, as the name “atti- about his history preceding his actual exis- tude” already tells us, relational states: They tence. The way in which mental events refer depend on the historical relations between to states of affairs in the world depends not the believer and events in the world. But just on how we are constituted, on what goes what about conscious experiences? Are they, on in our brains and bodies; it matters also like beliefs, relational states speciﬁed in part how we are in fact situated with respect to by the historical context? the environment and our history of interac- The argument sounds implausible for tion within it. conscious experiences. Quite the opposite: The opposite conclusion can now be We seem to be able to imagine that Swamp- reached in the case of David Chalmers man has conscious experiences indistin- turning himself into a zombie, the ﬁrst guishable from Davidson’s and that zombie P1: KAE 0521857430c30 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 3:10

866 the cambridge handbook of consciousness

Chalmers has none although he believes he experience and (2) their difficulty (and their does. The reason we give these answers is persuasiveness) are a good warning signal. that our concept of conscious experience These difficulties tell us that we need to put ties it to intrinsic properties of the person on hold many of the preconceptions that we (and specifically the brain, at least in mod- may have about explanations of conscious ern day), rather than to relational proper- experience. The problem cannot be solved ties as in the case of beliefs. This ability simply by thinking about it, because the con- of ours to imagine the above scenarios, of cepts we would currently employ in its anal- course, in no way shows that they are actu- ysis are misguided. ally possible. What it shows is that we can What kinds of concepts would we need? imagine them, that they are conceivable, and To make the relationship between objective that therefore our current concept of con- physical events and the subjectivity of conscious experience fails to be explained by scious experience transparent, Nagel tells us our current concepts of the objective cri- in a more recent paper that we would need teria that determine it. And this is where to see how conscious experience is a neces- the problem lies: If conscious experiences sary accompaniment of the objective events, were like beliefs, we could explain them in rather than a contingent one (Nagel, 2001). virtue of the relations between organisms We would need to be able to understand not and their interactions with the environment. just that, but why, certain physical events Given that we are unwilling to take this step or information-processing events also have a with conscious experiences, that we insist subjective nature to them. This would mean on an intrinsic account, we run into prob- that the identification of conscious experi- lems. On the one hand, we want to insist ence with such physical or functional events that conscious experience depends only on would be a necessary truth, but one that we its proximal neural substrate, only on events could only appreciate a posteriori. That is, in the brain, and hence both Swampman once we would have the conceptual appara- and brains in vats have conscious experi- tus to see how conscious experience is real- ences identical to those of normal humans. ized physically or functionally, we would see On the other hand, the possibility of zom- the relationship to be necessary, not contin- bies seems to show that conscious expe- gent. But we do not see this now, and it can- rience is contingent on all physical events not be seen using the concepts we have cur- occurring inside the person who has the rently available – hence it is not an a priori experience, and hence the two cannot be truth, but only one we could see to be true identified with one another. We are faced if we change the way in which are able to with a dilemma: Consciousness seems com- think about the problem. pletely unaccounted for by any of the intrinsic physical facts about a person despite our firm belief that it must arise from or be Does the Mind Supervene identical with such physical facts. What to on the Brain Alone? do? One might postulate by fiat conscious experience as a brute addition to our objec- Situated Cognition tive understanding of the world. But this, of course, does not count as an explanation – The preceding section served to sharpen the it does not make transparent why con- kind of difficulty with which we are dealing. scious experience should accompany certain The problem stems in large part from what events that can also be described physically we can currently conceive or not. In this or functionally. chapter, I work my way toward proposing an I have dwelled on these difficulties for account of conscious experience that, a pri- two reasons: I think that (1) an externalist ori, may seem rather counter-intuitive, per- view of mental content may provide clues haps even inconceivable. The task, therefore, toward a similar treatment also of conscious is to make it seem, a posteriori, conceivable P1: KAE 0521857430c30 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 3:10

consciousness: situated and social 867

that what I write might be right. By way of to glean lots of information about a com- introducing the proposal, let us review what plex visual scene. For instance, you can iden- has been written recently, not specifically tify targets in it, count objects, and report regarding conscious experience, but regard- on innumerable other features in an image. ing the mind and cognition in general. But you can’t do it by first inspecting the Picking up some of the strands we dis- visual scene and then simply closing your cussed in the preceding section, a major eyes and reading off all the information from issue is whether we should think of men- an internal representation of the scene. The tal events as intrinsic to the brain or as information is not in your brain; it is in relations between brains and their environ- the environment to be obtained by inves- ment. Recent proposals have argued that we tigation (in this example, by moving your should think of them as relational. These eyes and your visual attention to partic- proposals have been motivated by a variety ular locations in the image about which of findings that demonstrate that, at least at you want to obtain information). A popular the level of information processing, much of demonstration of this fact comes from the such information processing doesn’t happen phenomenon of “change blindness.” People inside the skulls of people alone, but draws often fail to notice rather large changes in a on information that is distributed through- visual scene, despite the fact that they are out the environment with which the person looking right at those changes and nothing is interacts. Let’s consider some examples. preventing them from noticing the changes You are able, presumably, to calculate the (Rensink, O’Regan, & Clark, 1997). square of a five-digit number. The way in If we reflect carefully on our conscious which you do so (presumably) draws (for visual experience, we can realize that it arises example) on pencil and paper: You manipu- rather directly from how we probe the exter- late objects in your environment to take up nal world, rather than from what is hap- some of the computational load (specifically, pening in our brains alone. The visual world working memory load, in this example) with appears very rich and full of information – which your brain in isolation is unable to but much of that information is not in our deal. So calculating the square of a five-digit head; it is available for inspection in the number involves physical events outside of world. The sense we have of a rich visual your brain and indeed outside of your body. experience is not as of an image in our mind Yet it is your brain that is guiding these that offers this richness, but rather as of the external computations. Your brain is able world out there that provides it if we choose to structure events occurring in the body to look here and there (Noe, 2004; O’Regan and environment in such a way as to use & Noe, 2001). And this is all as it should be; if them in information processing. The cogni- we didn’t rely on the external world as a rich tive scientist Andy Clark has championed source of information, our brains would be a comprehensive account of how the rela- overwhelmed. The idea then is that we don’t tional interactions between a brain, a body, have detailed, rich, internal representations and the external environment are all neces- of the features of the world, but rather that sary for cognition (Clark, 1997). And one can our representations consist (at least in good go further: Not only are all these relational part) of prescriptions of how to obtain such interactions necessary for cognition but they information by exploring our environment. are also necessary for conscious experience. This is not to say that we don’t represent You as a subject of conscious experience are anything in our brains. It is just to say that not identical with your brain, but rather arise what we represent is not the sole source of from the relational interactions of that brain information of which we can become aware with its environment (Clark, 1995). and that the format of the representation is Similar comments apply to cases where often sensorimotor in nature, rather than just we do not actively manipulate the exter- sensory. The motor aspect emphasizes the nal environment, such as when you are able idea that we are situated (Suchman, 1987), P1: KAE 0521857430c30 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 3:10

868 the cambridge handbook of consciousness

embedded in our environments: Our rela- provided some of the first examples of tionship with the environment is bidirec- this idea: tional – it affects us and we affect it, and the way in which we affect it in turn influences Upon analysis, we find that we begin not with a sensory stimulus, but with a sensori- the way in which it affects us. So, clever motor coördination, the optical-ocular, and manipulation of the way in which we change that in a certain sense it is the movement our bodies and the environment will permit which is primary, and the sensation which us to use physical events outside our brains is secondary, the movement of body, head to do information processing for us. and eye muscles determining the quality of There is still a weak version of this view, what is experienced. In other words, the real according to which the environment influ- beginning is with the act of seeing; it is look- ences the mind only insofar as it influences ing, and not a sensation of light. The sen- the brain. That is, events external to us can sory quale gives the value of the act, just certainly contribute to cognition, but they as the movement furnishes its mechanism do so only through the proximal substrate and control, but both sensation and movement lie inside, not outside the act (Dewey, of cognition, namely the brain. However, 1896). there is also a stronger version of the view, according to which the entire, coupled sys- The idea that the mind is embodied tem of brain and environment is the proxi- and situated in the world has been devel- mal substrate of cognition. The “active exter- oped in considerable detail recently (Varela, nalism” proposed by Clark and Chalmers Thompson, & Rosch, 1991) and has been (1998), suggests this second view. If this is combined with findings in cognitive neuro- right, then it would appear that cognition science. In particular, the idea is that com- does not supervene solely on neural events, plex dynamical patterns of brain activity but depends on events in our bodies, sense arise from the way in which the brain and organs, and the environment with which body interact with the external environ- they interact as well. ment and that it is such dynamical brain patterns that generate conscious experience (Thompson & Varela, 2001). Extended Conscious Experience? One can agree with all these points as pro- You have probably guessed where this viding a useful analysis of how the contents discussion is going. If cognitive abilities of our conscious experience depend on rela- depend on resources external to our brains, tional facts about brains, bodies, and the rest might conscious experiences have a simi- of the world, and yet one can still adhere to lar dependency? Might our minds literally the belief that it is in fact only the events extend outside of our brains and bodies, happening in the brain that directly deter- not just with respect to their information- mine conscious experience. That is, one can processing capacities, but with respect to agree that interaction with the world shapes conscious experience? the inputs the brain receives, but that if we Of course, the above discussion already were to strip away the world and keep the indicates that this is so in a certain sense, events in the brain unchanged, we would but perhaps only an indirect sense. Situ- keep the conscious experience unchanged. ated cognition, the view espoused in some That, after all, is the idea that conscious of the writings quoted earlier (Clark 1997; experience supervenes solely on the brain, O’Regan & Noe, 2001), has a long history even though what happens in the brain will of acknowledging that sensation requires of course be influenced by events outside it action onto the environment and that sen- (under normal circumstances). sation and action are not temporally and We have noted that at least some are functionally separate phenomena, but con- willing to view beliefs as literally and current aspects of how an organism interacts directly supervenient on external and his- with its world. The philosopher John Dewey torical events, as we saw in the case of P1: KAE 0521857430c30 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 3:10

consciousness: situated and social 869

Swampman earlier. And as reviewed above, in the brain now seems misguided, because many are willing to view cognition and their very subjectivity presupposes a distinc- information processing that contribute to tion between subject and object; that is, a our conscious experience as similarly exter- relation. What it is like to have a conscious nalized. But it sounds very odd indeed experience of a red sunset depends not only to suppose that conscious experience itself on the subject undergoing the experience should depend on events external to the but also on the red sunset toward which brain, although some proposals have indeed the subject’s point of view can be directed. been made along exactly those lines (e.g., In this respect, the subjectivity of conscious Rowlands, 2003). experience is no different from the directed- What I propose is to bite the bullet and ness of intentional states like beliefs or from literally consider conscious experience as the situatedness of the mind’s information supervening on events occurring outside the processing. None of these supervenes solely brain. That is, brains in vats and perhaps on events intrinsic to the brain, but literally Swampman also could not in fact be con- arises from the brain’s relational interaction scious in the way that we imagine them to with its environment. The extended system be. They lack some of the necessary envi- as a whole – brain, body, and environment ronmental and historical relations. Note that and their interactions (now as well as his- this is a radical claim – we are not claiming torically) – is what makes the subject con- merely that such relations can affect consciously aware of objects and events. The scious experience via their effects on events subvenience base of conscious experience in the brain, but that they are constitutive has thus been radically broadened: Even if of conscious experience. Of course, the idea everything in the brain is held constant, a seems nonsensical at ﬁrst, a priori. It seems change in the web of relations that are linked an impossibility because we are wedded to to the brain would change the conscious the view that the subjectivity of conscious experience. Despite his claims to the con- experience is an intrinsic property of the trary, Swampman would not have the con- brain. But we saw earlier that this view faces scious experiences that he avows, for the serious problems. same reason that he does not have the beliefs Under the new proposal, conscious he avows. awareness is best thought of as arising not from events in the brain, but rather from the act of becoming conscious of an external Social Cognition event or object in the environment through our brain’s and body’s interaction with it. The Social Brain Hypothesis If we thus attach phenomenal properties to the act of apprehending that of which we We have covered a lot of ground so far, become consciously aware, we can see that mostly philosophical, and of necessity I’ve conscious awareness is situated in the same sketched only the roughest kind of picture. way in which cognition in general was pro- The question that a scientist might now like posed to be situated in our earlier discussion. to ask is, How are we to think of this envi- Like situated cognition, the subjectivity of ronment with which a subject interacts and conscious awareness depends not only on the that is doing so much work according to the subject’s brain but also on the relation that arguments above? At a minimum, some par- the subject has to the object of conscious titions of the environment must be more awareness in the environment and indeed to important to conscious experience than oth- his or her history of interaction with objects ers. If there is an explosion on alpha centauri, in the environment. it cannot be that my conscious experience Several perspectives follow naturally suddenly undergoes a large change. Presum- from this idea. In retrospect, thinking of sub- ably, some set of the relations between the jectivity as an intrinsic property of events brain and the rest of the world matter more P1: KAE 0521857430c30 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 3:10

870 the cambridge handbook of consciousness

than others to conscious experience. An ini- skills may be traced to evolution in an envi- tial visual metaphor would be that subjec- ronment in which there was a premium on tivity is centered in the head, as it were, but social skills. In support of this idea, there is a leaks out into the body and past the bounds correlation between mean group size among of the body, becoming fainter the more dis- various primate species and their neocor- tally that events are located. Of course, it is tex volume (Dunbar, 1998); to be exact, it not spatial location as such that matters here, is the ratio of their neocortex volume to but the strength of the relational interaction their total brain volume, but the details are within the web of relations that involves the not important here. Such a correlation has brain: Breaking some of these relations intro- been found also for several other mammals duces a large perturbation into the entire that all feature a complex social structure relational network, whereas breaking others (e.g., bats, carnivores, and toothed whales): has a negligible effect. The larger the social groups, the larger the One aspect of the external environment brains. Although brain size has been argued is of course just the non-biological environ- to correlate with a number of other fac- ment of mountains and tables and chairs that tors, including dietary foraging strategy, tool surrounds us. Another critical aspect of the use, and longevity (Allman, 1999), it may external environment, however, is the social be that large brain size is at least a par- environment: the other people, the other tial consequence of the fact that primates minds, with whom we interact. The social have a complex ecological niche with respect environment has several features that distin- to social structure (including its effect on guish it: It is more complex than the non- food and mate availability). This hypothesis, social environment, it is reciprocally inter- variously dubbed the Machiavellian Intelli- active with us, and it includes other minds gence Hypothesis (Whiten & Byrne, 1997)or and other subjects of conscious awareness. the Social Brain Hypothesis (Dunbar, 1998), How important is this distinction? argues that the complexity of primate social According to some views, very impor- structure, together with certain of its unique tant. One influential line of thinking has pro- features, such as cooperativity and decep- posed that the cognitive abilities that make tion, led to an advantage for larger brains. us distinctively human in fact arose pre- Primates are clearly highly skilled at precisely from the need to function in a socially dicting other individuals’ behavior, but there complex environment, with other minds and is presently vigorous debate regarding how the reciprocal interactions they provide us. to interpret such an ability. Research into As a source of competition for resources how we represent other minds began with a like shared food and mates, the social envi- question about whether or not chimpanzees ronment would arguably have driven the might possess a theory of mind (Premack evolution of our abilities to predict other & Woodruff, 1978), a question that is still people’s behavior and to guide our behav- the subject of debate (Povinelli & Vonk, ior accordingly. Given that everyone would 2003; Tomasello, Call, & Hare, 2003). In have been driven in the same fashion, some- humans, the question was posed concretely thing like a “mental arms race” might thus in terms of the ability to attribute beliefs, have provided the impetus for the evolution specifically false beliefs, to other individu- of many of our cognitive abilities, abilities als, an ability that begins to emerge around that collectively allow us to outsmart others. age 4 or possibly earlier (Wimmer & Perner, I next briefly review this issue in somewhat 1983). The abilities that constitute a the- more detail. ory of mind have been fractionated into sev- It is clear that primates are exceedingly eral distinct components, such as the abil- adept at negotiating the social environment. ity to attribute desires, to recognize objects This ability is most striking in the most of shared attention, and to monitor oth- social primate, Homo sapiens, suggesting the ers’ direction of gaze (Baron-Cohen, 1995). hypothesis that our exceptional cognitive All these different components appear at P1: KAE 0521857430c30 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 3:10

consciousness: situated and social 871

distinct developmental stages in humans, have started at the input end, as it were: by and there is evidence that some of them may showing subjects (often under passive view- be disproportionately impaired in subjects ing conditions) pictures of social relevance with autism, a disorder that exhibits marked and associating differences in the social con- difficulties in social behavior. tent of stimuli with differences in the neural There is thus considerable support structures engaged in their processing. This (although by no means unequivocal sup- work has found covariances between stimu- port) for the idea that social interactions gave lus dimensions and brain structures, primar- rise to the kinds of complex cognitive abil- ily from functional imaging studies. ities seen in humans. If the human mind, Investigations have focused on the visual with respect to cognition, arose out of the modality in primates. Social visual signals need to function in a complex social envi- include information about the face, such as ronment, might the same be true of human its expression and direction of gaze, as well consciousness? Returning to our metaphor as about body posture and movement. There of the mind as literally leaking out beyond is a wealth of signals available in the face, the bounds of our skull to encompass the reviewed further below, that illustrates the web of relations in which our brains partic- tight reciprocal link between action and per- ipate, social relations may be some of the ception. We look at someone’s face to pick strongest links in this web. up salient information, but doing so may I have argued that relational facts about well result in the other person also noticing a brain’s interactions with its environment that he is being scrutinized and changing his are just as important to conscious experi- face in response. ence as they have been claimed to be impor- Human viewers are surprisingly adept tant for propositional attitudes like beliefs also at making reliable judgments about and for cognitive abilities in general. There is social information contained in rather now the further argument that a very impor- impoverished stimuli, such as very faint tant, and perhaps the most important, com- changes in facial expression or only a few ponent of those relations concern relations seconds of videos of full-body interper- between an individual and others of its kind. sonal interactions. For instance, very brief At least at the gross level of evolution of the and sparse depictions of biological motion brain, the nature of social relations appears depicting people can already trigger a rich to have been a driving force in that evolu- social interpretation (Ambady & Rosenthal, tion. The critically social nature of the envi- 1992). Not only are we exceedingly sensi- ronment in which our brains are embedded tive to the social signal itself but we also also has not escaped the notice of neurosci- are sensitive to the details of the context entists, and there is now a burgeoning litera- in which it occurs. A good example of this ture on how socially relevant stimuli are pro- is direction of eye gaze – direct gaze is a cessed by the brain. This literature is briefly potent social signal, but it can mean different reviewed next. things, such as threat or flirtation, depending on who it is that is doing the gazing. More- over, there are clear differences in the brain Social Cognitive Neuroscience when social signals are embedded in differ- Social cognitive neuroscience, the study of ent contexts (Pelphrey, Singerman, Allison, the neurological underpinnings of social cog- & McCarthy, 2003), and the influence of nition, has recently exploded as a field. Data such social context on brain function appears are pouring in from cross-disciplinary stud- to be impaired in people with autism. ies using techniques that range from fMRI, Regions of non-primary sensory cortices lesion, and behavioral studies in humans already appear to be relatively specialized to cellular and molecular studies in rats to process certain socially relevant attributes and flies (Adolphs, 2003; Cacioppo et al., of stimuli. The evidence is best in regard to 2001). Many recent neuroscience studies faces, for which higher-order visual cortices P1: KAE 0521857430c30 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 3:10

872 the cambridge handbook of consciousness

can be regarded as an assembly of mod- and generate exceptionally robust shape- ules that process distinct attributes, as borne from-motion cues that permit recognition of out by a variety of lesion studies, scalp and identity, gender, emotions, and personality intracranial recordings, and a rapidly growing traits. In line with the role of superior tempo- list of functional imaging studies. The data ral cortices in processing dynamic aspects of point to the fusiform gyrus in processing the faces, this region is also activated by viewing structural, static properties of faces that are biological motion in whole bodies or their reliable indicators of personal identity, and point-light displays (Grossman & Blake, to regions more anterior and dorsal in the 2002) and by more abstract movements of temporal lobe (such as the superior temporal geometric shapes, likely reflecting its role in gyrus and sulcus) in processing information processing biological motion information on about the changeable configurations of faces, the basis of which we make social attribu- such as facial expressions, eye movements, tions. In these examples, relatively subtle or and mouth movements (Haxby, Hoffman, & impoverished trigger stimuli in the environ- Gobbini, 2000). Activation along the supe- ment are sufficient to result in a large neural rior temporal sulcus and gyrus has been modulation. In this case it is the brain that found when subjects view stimuli depicting is providing most of the information gener- biological motion, such as eye gaze shifts and ated, rather than the stimulus – but of course mouth movements. The fusiform gyrus, the the brain does so on the basis of its history superior temporal gyrus, and other as yet less of interaction within a social environment. well-specified regions of occipitotemporal Several brain regions are activated not cortex could thus be thought of as an inter- only as a function of properties inherent connected system of regions that together to the stimuli but also as a function of construct a spatially distributed perceptual the psychological judgments we make about representation of different aspects of faces them. In a sense, the influence of such judg- (Haxby et al., 2001). There is good evi- ments reflects a progressive decoupling from dence that activation in all of these regions responses dictated by the stimulus itself to can be modulated by attention (Vuilleu- information that is generated by the brain via mier, Armony, Driver, & Dolan, 2001) and associations and inferences. The amygdala by the context in which the visual social sig- is one structure that is anatomically posi- nal appears. All of these findings certainly tioned to participate in such post-perceptual indicate that the brain is exquisitely tuned processing, as it receives highly processed to respond to socially relevant signals with visual information (from anterior temporal which it can interact. cortices) and stores codes for subsequent A growing body of work has used visual processing of such perceptual information in stimuli that signal biological motion. Social other brain regions. In this way, it can influ- psychologists first demonstrated our propen- ence memory, attention, decision making, sity to make social inferences from visual and other cognitive functions on the basis motion of abstract shapes in the 1940s of the social significance of the stimuli being (Heider & Simmel, 1944), and recent stud- processed. ies suggest that specific movement cues gen- The bulk of research on the human amyg- erate attributions of animacy, intentional- dala has used emotional facial expressions ity, and agency (Scholl & Tremoulet, 2000). as stimuli and has pointed most consis- Visual motion stimuli elicit attributions of tently to processing of fear and related emo- intentionality and animacy in infants, and tions (Adolphs, Tranel, Damasio, & Damasio, robustly elicit intentional, emotional, and 1994; Calder et al., 1996), although recent personality attributions in adults, even when evidence suggests that its role is probably only static depictions of their trajectories considerably broader than that. Functional are shown. Point-light displays offer more imaging studies demonstrate processing at specific information about the movements multiple stages: rapid, automatic evaluation made by a human body (Johansson, 1973) and tagging of stimuli for further processing, P1: KAE 0521857430c30 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 3:10

consciousness: situated and social 873

feedback modulation of attentional process- with other visual regions of the brain but ing in visual cortices, and modes of pro- that it does so also by directing gaze to cessing subject to self-regulation and voli- specific features of stimuli. Damage to the tional guidance. The first and last of these amygdala results in an inability to recognize in fact illustrate complementary roles for certain emotions from facial expressions, but the amygdala, probably operating at comple- it appears to do so, at least in good part, via an mentary time scales. On the one hand, some impairment in exploring the social environ- amygdala activation is seen early, regard- ment with one’s eyes. In one case, a patient less of conscious perception of the stim- with bilateral damage to the amygdala failed ulus and regardless of attention allocation to look normally at other people’s faces and in some tasks. On the other hand, effort- consequently failed to pick up and use visual ful self-regulation of the emotions induced information normally in judging the emo- by stimuli, reappraisal of the emotional sig- tion that was shown on the face (Adolphs nificance of the stimuli, and difficult atten- et al., 2005). This example brings us back tional tasks all modulate amygdala activa- to our discussion of situated cognition and tion. These findings urge caution in the rigid examples like change blindness: The visual assignment of cognitive processes to neural world we experience is not entirely in our structures, because it is likely that a given heads. Rather, our brains possess strategies structure participates in multiple processes, for actively and efficiently probing the exter- depending on the point in time at which nal environment in order to seek out rel- its activity is sampled and on the details of evant visual information – for instance, by the task and context in which the subject deciding where to direct our gaze in the first is engaged. It is plausible that the amygdala place. participates both in the initial, rapid evalua- Abilities that have been dubbed “theory tion of the emotional significance of stimuli of mind” enable the attribution of men- and in their later assessment within a given tal states to other people (Siegal & Varley, context and goals. 2002). Particularly studied have been attri- Beyond its role in the recognition of basic butions of beliefs, specifically false beliefs, emotions, the amygdala is involved in more to other individuals. As we discussed earlier complex social judgments. It shows differ- in this section, such abilities emerge around ential habituation of activation to faces of age 4, may be unique to humans, and may be people of another race, and amygdala acti- assembled out of a collection of more basic vation has been found to correlate with skills by which we assign animacy, actions, racial stereotypes of which the viewer may goals, and intentions to stimuli, an issue that be unaware (Phelps et al., 2000). However, has seen intense recent investigation using the amygdala’s role in processing informa- visual motion stimuli (Blakemore & Decety, tion about race is still unclear: Other brain 2001). In addition to the reliable activation of regions, in extrastriate visual cortex, are also superior temporal gyrus already mentioned activated differentially as a function of race, earlier, a number of functional imaging stud- and lesions of the amygdala do not appear to ies have demonstrated activation of cortex in impair race judgments. the medial frontal lobe, as well as the infe- All of this demonstrates that the neural rior parietal lobule, when people view visual events that generate knowledge, behavior, motion or eye gaze stimuli that signal such and awareness of the social world unfold directed mental states. over multiple iterations of processing, iter- Many of the same stimuli that engage ations that include multiple interactions the superior temporal gyrus and that lead with the environment. There is a recent viewers to attribute actions, intentions, particularly clear example of such action- and goals also activate regions of neocor- perception coupling in regard to the amyg- tex involved in representing actions. These dala. It turns out that the amygdala not only regions include premotor-related cortices as processes visual information by interacting well as somatosensory-related cortices, the P1: KAE 0521857430c30 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 3:10

874 the cambridge handbook of consciousness

efferent and afferent sides of actions. A series dynamics of the brain processes that were of recent studies have investigated the role of just outlined. Well, there is certainly some right somatosensory-associated cortices and consequence of the more elaborated social left premotor cortex in making emotional information processing that primate brains and personality attributions also from point- afford. But, following the metaphor outlined light displays and movements of geometric earlier, the consequence of more elaborate shapes; damage in both regions impairs the neural processing is that it brings certain rel- ability to make such attributions. evant relations closer to the nexus of the web There is a rapidly growing literature in which they are embedded. supporting the idea that we understand Let us return briefly to the example of other people’s behavior in part by simula- calculating the square of a five-digit num- tion (Rizzolatti, Fogassi, & Gallese, 2001). ber. With some training, one could do so in Observing another’s actions results in desyn- the absence of pencil and paper. Perhaps one chronization in motor cortex as measured strategy of doing so might even be to con- with magnetoencephalography (Hari et al., centrate very hard to create a detailed visual 1998); imitating another’s actions via obser- image as if one were doing it with pencil vation activates premotor cortex in func- and paper. What has been achieved by this tional imaging studies (Iacoboni et al., 1999), is to bring what used to require relations and such activation is somatotopic with with events in the external environment into respect to the body part that is observed a simulation of those relations within the to perform the action, even in the absence brain. But that in no way shows that the cog- of any overt action on the part of the sub- nitive ability to square a five-digit number ject (Buccino et al., 2001). In fact, in both now depends only on the brain. It depends humans (Hutchison, Davis, Lozano, Tasker, on events happening in the brain, as well as & Dostrovsky, 1999) and monkeys (Gallese on historical relations between the brain and & Goldman, 1999), so-called mirror neu- the environment. Without that history of rons have been discovered that respond both interaction, there would have been nothing when the subject is doing something specific to internalize. This issue, of how the brain and when he or she observes another per- might come to internalize relevant aspects of son doing the same thing. Damage restricted its relational interaction with body and envi- to somatosensory cortex impairs the abil- ronment, requires some further treatment. ity to recognize complex blends of emotions in facial expressions (Adolphs, Dama- Simulation sio, Tranel, Cooper, & Damasio, 2000), and there is an association between impaired Like all cognition, social knowledge goes somatic sensation of one’s own body and beyond the mere information present in the impaired ability to judge other people’s evidence on which it is based; it is inferen- emotions (Adolphs et al., 2000). Functional tial and creative in nature. But do the mech- imaging studies also support a role for right anisms that subserve it literally extend out- somatosensory-related cortices in represent- side the brain, as I’ve been suggesting? One ing the actions we observe others to perform, starting point is to view brains as containing as distinct from those we perform ourselves emulators that are in the business of con- (Ruby & Decety, 2001). structing models of the world, including the All of the above review might seem to social world. The mechanisms required by suggest that, after all, richer minds arise from the theory-of-mind abilities reviewed above larger and more complex brains; richer con- can be considered to use emulators of sorts. scious experience depends on richer brains; The simulation theory of mind, for instance, and perhaps we should jettison everything I might consist in an “articulated” emulator: argued for in the first half of this chapter and a model that mimics aspects of what it is return to the idea that conscious experience modeling to achieve its predictive power arises entirely out of the complex internal (Grush, 2003). P1: KAE 0521857430c30 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 3:10

consciousness: situated and social 875

The notion of an articulated emulator ies, and although there are indeed somatic has substantial support from ﬁndings in responses during dreaming, these responses cognitive neuroscience. As reviewed above, are actively inhibited from full expression. there is considerable empirical evidence that It would seem odd to have evolved such humans obtain knowledge about other peo- efferent processing and inhibition if the ple’s emotional states, for example, at least body itself did not play an important role in part, via some kind of articulated emu- in building the models that help us predict lation. Premotor cortices are engaged when the world. we observe others behaving emotionally, as We can now imagine extending the mod- are somatosensory cortices and insula. These eling outside the bounds of the body. To components would correspond to the initial obtain social information, we may query not goal command for emotional behaviors and only our own bodies but also those of other to representation of the feedback signal from people. Clearly, this is the case in a general such behavior, respectively, raising the ques- sense: We probe other people’s reactions to tion of what happens between the two to initial and often subtle behaviors on our own connect them: Where is the emulator itself? part and use their response as feedback in One possibility is that we engage some constructing a more accurate model of the of the same machinery during emulation as social world. during actual emotions: the body outside the brain. There is good evidence for this idea as well: Observing other people express emo- Conclusions tions results in some mirroring of the physiological emotional state in the viewer. In this The picture that emerges from the above dis- case it seems that the emulator is the same cussion is more subtle and complex than the as the system in normal operation, although one we previously had in mind. The relations it may engage only a subset of a hierarchi- between brain, body, and external environ- cally structured system. The possibility of ment are still the base out of which the mind using the body itself as the emulator when is generated, but the brain has some control we model another person’s emotion would over the relative importance of these rela- be not only economical but also suggests tions insofar as it can internalize some of an interesting way in which actual, analog them and externalize others. Thus, depend- physical processes – state changes in various ing on the particular cognitive and behav- parameters of the body that normally com- ioral requirements, relations can be modeled prise an emotional response – can be used in internally to provide greater processing information processing. The body might be speed or can be externalized where very thought of as a “somatic scratchpad” that we large or detailed computations are required can probe with efferent signals in order to that go beyond what the brain in isolation reconstruct knowledge about the details of could compute. But nowhere is there an an emotional state. Given the complexity of independence of brain from environment – interaction among multiple somatic param- the only way in which the environment eters, in action as well as emotion, it may not can, by certain brains under certain circum- be feasible to emulate this entirely neurally. stances, be partially modeled is through a Typically, of course, our emulation should history of interaction with that environment be less than the real thing. Thus, emotion in the ﬁrst place. The mind, including con- emulation involves faint somatic changes scious experience, arises from a web of rela- that are a subset of having the emotion tions that involve the brain, but the brain oneself and that involve active inhibition remains the hub in this web, in the sense of expression of some of the components. that neural events contribute substantially Much the same happens when we dream: to mental events and the brain has a role in As in waking emulation, we construct mod- orchestrating the relational web as its orga- els that include responses in our bod- nizing center. P1: KAE 0521857430c30 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 3:10

876 the cambridge handbook of consciousness

Needless to say, and despite the length of ness are often correlated with changes in the chapter here, I’ve only begun to sketch, social cognition, this would be preliminary in the barest outline, how precisely the fact evidence in support of the idea that con- that we are embedded in a web of social sciousness is related to our social nature. interactions would inform our understand- There is already some evidence for such ing of consciousness. My aim has been to a correlation from experiments in patients focus the philosophical problem on one par- who have damage to the frontal lobes: Not ticular view of the problem of consciousness, only is their social behavior severely com- to suggest reasons for doubting that we can promised but also the nature of their con- understand consciousness through intrinsic scious experience may well be altered. It is properties of the brain, and to argue that likely, for instance, that their ability to expe- we need to look instead at relational proper- rience certain emotions, perhaps especially ties – between the brain and the social envi- social ones, is severely diminished and, at the ronment (i.e., between multiple brains). My same time, their social relations with others review of social neuroscience has, of course, are severely disturbed (Anderson et al., 1999; not directly addressed the deeper issue at Damasio, 1994). hand. But I hope it has provided a brief More direct evidence for the idea that overview of what we do know about the consciousness literally depends on relational neural underpinnings of social behavior and social events would require other kinds social cognition and, in so doing, empha- of data. One could imagine developmen- sized the fact that, as in the case of situ- tal studies in which the social environment ated cognition more generally, our represen- is systematically manipulated that might tation of the social world depends both on investigate the issue. One could also imag- events in our brains and events external to us. ine comparative studies between different We perceive other people’s behavior, they species that are more or less social. Of in turn perceive ours, and mutual interac- course, all of these would run up against tion permits one to construct a model of the another, separate practical difficulty, one other. Moreover, I hope that even this very that I have not mentioned here: how to brief sketch of an overview of social neuro- decide on convincing observable criteria for science suggested that our brains are indeed conscious awareness in the first place. highly adapted for a socially interactive environment. The evolutionary evidence I mentioned supports the idea that much of what References has driven the evolution of the brain was the nature of the complex social environ- Adolphs, R. (2003). Cognitive neuroscience of ment in which brains were situated. The human social behavior. Nature Reviews Neuro- 165 178 neurobiological evidence reinforced the idea science, 4, – . that the brain is replete with mechanisms Adolphs, R., Damasio, H., Tranel, D., Cooper, G., 2000 for processing socially relevant information, & Damasio, A. R. ( ). A role for somatosen- is exceedingly sensitive to such information sory cortices in the visual recognition of emotion as revealed by 3-D lesion mapping. Journal and its context, and has evolved clever tools of Neuroscience, 20, 2683–2690. for actively probing the social environment. Adolphs, R., Gosselin, F., Buchanan, T. W., At this stage, my feeling is that we have Tranel, D., Schyns, P., & Damasio, A. R. (2005). enough ammunition available to begin for- A mechanism for impaired fear recognition mulating experiments that would test some following amygdala damage. Nature, 433, of the ideas put forth in this chapter. There is 68–72. good evidence, as reviewed above, that social Adolphs, R., Tranel, D., Damasio, H., & Damasio, information processing draws more on some A. (1994). Impaired recognition of emotion in structures than on others. What happens facial expressions following bilateral damage to when those structures are lesioned? What the human amygdala. Nature, 372, 669–672. happens if they are artificially stimulated? Allman, J. M. (1999). Evolving brains. New York: If we find that changes in conscious aware- Scientific American Library. P1: KAE 0521857430c30 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 3:10

consciousness: situated and social 877

Ambady, N., & Rosenthal, R. (1992). Thin slices mind-reading. Trends in Cognitive Sciences, 2 , of expressive behavior as predictors of inter- 493–500. personal consequences: A meta-analysis. Psy- Grossman, E., & Blake, R. (2002). Brain areas chological Bulletin, 111, 256–274. active during visual perception of biological Anderson, S. W., Bechara, A., Damasio, H., stimuli. Neuron, 35, 1167–1175. Tranel, D., & Damasio A. R. (1999). Impair- Grush, R. (2003). The emulation theory of ment of social and moral behavior related representation: Motor control, imagery and to early damage in human prefrontal cortex. perception. Behavioral and Brain Sciences, 27, Nature Neuroscience, 2 , 1032–1037. 377–442. Baron-Cohen, S. (1995). Mindblindness: An essay Hari, R., Forss, N., Avikainen, S., Kirveskari, E., on autism and theory of mind. Cambridge, MA: Salenius, S., & Rizzolatti, G. (1998). Activation MIT Press. of human primary motor cortex during action Blakemore, S.-J., & Decety, J. (2001). From the observation: A neuromagnetic study. Proceed- perception of action to the understanding of ings of the National Academy of Sciences, 95, intention. Nature Reviews Neuroscience, 2 , 561– 15061–15065. 568. Haxby, J. V., Gobbini, M. I., Furey, M. L., Buccino, G., Binkofski, F., Fink, G. R., Fadiga, L., Ishai, A., Schouten, J. L., & Pietrini, P. (2001). Fogassi, L., Gallese, V. V., et al. (2001). Action Distributed and overlapping representation of observation activates premotor and parietal faces and objects in ventral temporal cortex. areas in a somatotopic manner: An fMRI study. Science, 293, 2425–2429. European Journal of Neuroscience, 13, 400–404. Haxby, J. V., Hoffman, E. A., & Gobbini, M. I. Cacioppo, J. T., Berntson, G. G., Adolphs, R., (2000). The distributed human neural system Carter, C. S., Davidson, R. J., McClintock, for face perception. Trends in Cognitive Sciences, M. K., et al. (Eds.). (2001). Foundations in social 4, 223–233. neuroscience. Cambridge, MA: MIT Press. Heider, F., & Simmel, M. (1944). An experimen- Calder, A. J., Young, A. W., Rowland, D., Per- tal study of apparent behavior. American Jour- rett, D. I., Hodges, J. R., & Etcoff, N. L. (1996). nal of Psychology, 57, 243–259. Facial emotion recognition after bilateral amyg- Hutchison, W. D., Davis, K. D., Lozano, A. M., dala damage: Differentially severe impairment Tasker, R. R., & Dostrovsky, J. O. (1999). Pain- of fear. Cognitive Neuropsychology, 13, 699– related neurons in the human cingulate cortex. 745. Nature Neuroscience, 2 , 403–405. Chalmers, D. J. (1997). The conscious mind. Iacoboni, M., Woods, R. P., Brass, M., Bekkering, Oxford: Oxford University Press. H., Mazziotta, J. C., & Rizzolatti, G. (1999). Clark, A. (1997). Being there: Putting brain, body Cortical mechanisms of human imitation. Sci- and world together again. Cambridge, MA: MIT ence, 286, 2526–2528. Press. Johansson, G. (1973). Visual perception of bio- Clark, A. (1995). I am John’s brain. Journal of Con- logical motion and a model of its analysis. Per- sciousness Studies, 2 , 144–148. ception and Psychophysics, 14, 202–211. Clark, A., & Chalmers, D. J. (1998). The extended Metzinger, T (2003). Being no one: The self mind. Analysis, 58, 10–23. model of subjectivity. Cambridge, MA: The MIT Damasio, A. R. (1994). Descartes’ error: Emo- Press. tion, reason, and the human brain. New York: Nagel, T. (1974). What is it like to be a bat? Philo- Putnam. sophical Review, 83, 435–450. Davidson, D. (1987). Knowing one’s own mind. Nagel, T. (1993). What is the mind-body In D. Davidson (Ed.), Subjective, intersubjective, problem? In Experimental and theoretical stud- objective: Philosophical essays of Donald David- ies of consciousness (Vol. 174, Ciba Founda- son (pp. 15–38). New York: Oxford University tion Symposium, pp. 1–13). New York: John Press. Wiley. Dewey, J. (1896). The reﬂex arc concept in psy- Nagel, T. (2001). The psychophysical nexus. In P. chology. Psychological Review, 3, 357–370. Boghossian & C. Peacocke (Eds.), New essays on Dunbar, R. (1998). The social brain hypothesis. the a priori (pp. 433–471). New York: Oxford Evolutionary Anthropology, 6, 178–190. University Press. Gallese, V., & Goldman, A. (1999). Mirror Noe, A. (2002). Is the visual world a grand illu- neurons and the simulation theory of sion? Journal of Consciousness Studies, 9, 1–12. P1: KAE 0521857430c30 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 13, 2007 3:10

878 the cambridge handbook of consciousness

Noe, A. (2004). Action in perception. Cambridge, Ruby, P., & Decety, J. (2001). Effect of subjective MA: MIT Press. perspective taking during simulation of action: O’Regan, J. K., & Noe, A. (2001). A sensorimo- A PET investigation of agency. Nature Neuro- tor account of vision and visual consciousness. science, 4, 546–550. Behavioral and Brain Sciences, 24, 939–1031. Scholl, B. J., & Tremoulet, P. D. (2000). Percep- Pelphrey, K. A., Singerman, J. D., Allison, T., & tual causality and animacy. Trends in Cognitive McCarthy, G. (2003). Brain activation evoked Sciences, 4, 299–308. by the perception of gaze shifts: The inﬂuence Siegal, M., & Varley, R. (2002). Neural systems of timing and context. Neuropsychologia, 41, involved in ‘theory of mind’. Nature Reviews 156–170. Neuroscience, 3, 463–471. Phelps, E. A., O’Connor, K. J., Cunningham, Suchman, L. (1987). Plans and situated actions. W. A., Funayama, E. S., Gatenby, J. C., Gore, Cambridge: Cambridge University Press. J. C., et al. (2000). Performance on indirect Thompson, E., & Varela, F. J. (2001). Radical measures of race evaluation predicts amygdala embodiment: neural dynamics and conscious- activation. Journal of Cognitive Neuroscience, 12, ness. Trends in Cognitive Sciences, 5, 418–425. 729 738 – . Tomasello, M., Call, J., & Hare, B. (2003). Chim- Povinelli, D. J., & Vonk, J. (2003). Chimpanzee panzees understand psychological states – the minds: Suspiciously human? Trends in Cogni- question is which ones and to what extent. tive Sciences, 7, 157–160. Trends in Cognitive Sciences, 7, 153–156. Premack, D., & Woodruff, G. (1978). Does the Varela, J. F., Thompson, E., & Rosch, E. chimpanzee have a theory of mind? Behavioral (1991). The embodied mind: Cognitive science and Brain Sciences, 1, 515–526. and human experience. Cambridge, MA: MIT Putnam, H. (1979). The meaning of ‘meaning’. In Press. Mind, language and reality: Philosophical papers Vuilleumier, P., Armony, J. L., Driver, J., & (Vol. 2,pp.215–272). Cambridge: Cambridge Dolan, R. J. (2001). Effects of attention and University Press emotion on face processing in the human brain. Rensink, R. A., O’Regan, J. K., & Clark, J. J. An event-related fMRI study. Neuron, 30, (1997). To see or not to see: The need for atten- 829. tion to perceive changes in scenes. Psychological Whiten, A., & Byrne, R. W. (Eds.). (1997). Science, 8, 368–373. Machiavellian intelligence II: Extensions and Rizzolatti, G., Fogassi, L., & Gallese, V. (2001). evaluations. Cambridge: Cambridge University Neurophysiological mechanisms underlying Press. the understanding and imitation of action. Wimmer, H., & Perner, J. (1983). Beliefs about Nature Reviews Neuroscience, 2 , 661–670. beliefs: Representation and constraining func- Rowlands, M. (2003). Externalism: Putting mind tion of wrong beliefs in young children’s under- and world back together again. Chesham: standing of deception. Cognition, 13, 103– Acumen. 128. P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

Part III QUANTUM APPROACHES TO CONSCIOUSNESS

879 P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

880 P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

CHAPTER 31 Quantum Approaches to Consciousness

Henry Stapp

Abstract brain. Classical physics is an approximation to quantum mechanics that eliminates this The concepts of classical physics and the effect. That makes the physical efficacy of implied doctrine of the causal closure of our conscious choices a strictly quantum the physical description fail to accommodate effect. The Penrose, Bohm, and Eccles-Beck observed behaviors of macroscopic systems quantum approaches to the mind-brain sys- when those behaviors depend sensitively tem are discussed, along with the orthodox upon the properties of atomic-sized par- von Neumann approach. ticles. To adequately describe such behaviors physicists were forced to replace, in the mathematics, the numbers that describe the Introduction intrinsic properties of a system by actions that correspond to the probing actions per- Quantum approaches to consciousness are formed upon that system by an observ- sometimes said to be motivated simply by ing system. According to orthodox quan- the idea that quantum theory is a mystery tum theory these probing actions are “freely” and consciousness is a mystery, so perhaps chosen by human agents, where “freely” the two are related. That opinion betrays a means that these choices are not fixed by profound misunderstanding of the nature of the known laws of physics or laws of any quantum mechanics, which consists funda- yet-known kind. But these choices influ- mentally of a pragmatic scientific solution ence the course of physical events. Von to the problem of the connection between Neumann’s orthodox formulation of quan- mind and matter. tum mechanics makes this influence of our The key philosophical and scientific conscious choices into influences, described achievement of the founders of quantum by specified causal laws, of a person’s con- theory was to forge a rationally coherent and scious choices on the state of his or her practically useful linkage between the two

881 P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

882 the cambridge handbook of consciousness

kinds of descriptions that jointly comprise nor any of the other entities, nor combi- the foundation of science. Descriptions of nations of the entities that populate the the first kind are accounts of psychologically world make choices on the basis of ideas. The experienced empirical findings, expressed in world described by the concepts of classical a language that allows us to communicate to physics has been systematically stripped of, our colleagues what we have done and what and is consequently bereft of, the concept we have learned. Descriptions of the second of choices based on consciously experienced kind are specifications of physical proper- ideas. Thus, the stubborn fact that idea-like ties, which are expressed by assigning mathe- realities do exist enforces an awkward depar- matical properties to space-time points, and ture of science from a purely naturalistic formulating laws that determine how these stance. Non-physical features, such as con- properties evolve over the course of time. scious thoughts, ideas, and feelings, must be Bohr, Heisenberg, Pauli, and the other inven- added, for no apparent naturalistic, physi- tors of quantum theory discovered a useful cal, or rational reason, to the features that way to connect these two kinds of descrip- enter into the putative laws of nature. There tions by causal laws; their seminal discov- is thus a conceptual mismatch between the ery was extended by John von Neumann world described by the basic laws of classi- from the domain of atomic science to the cal physics and the world we inhabit and do realm of neuroscience and in particular to science in. the problem of understanding and describing These difficulties have been much dis- the causal connections between the minds cussed by many philosophers, who have and the brains of human beings. proposed many different approaches. But The magnitude of the difference between in view of the known failure of classical the quantum and classical conceptions of physics to be able to describe the macro- the connection between mind and brain can scopic properties of systems whose behaviors scarcely be exaggerated. All approaches to can depend sensitively on the behaviors of this problem based on the precepts of clas- their atomic constituents, and the further sical physics founder first on the problem of fact that orthodox contemporary physical the lack of any need within classical mechan- theory brings conscious choices by human ics for consciousness to exist at all, and sec- agents into physical theory in an essential ond on a conceptual gap that blocks any way, the question must be asked whether rational understanding of how the experien- these philosophical efforts accord with 20th- tial realities that form our streams of con- century science or are, instead, clever ways sciousness could ever be produced by, or of trying to justifying the use of approx- naturally come to be associated with, the imately valid but fundamentally incorrect motions of the things that classical physics 19th-century physics in a domain where that claims the physical world to be made of. The approximation is inadequate. first problem is that, according to precepts of Both of the above-mentioned difficulties classical physics, the causal properties that are resolved in a rationally coherent and it explicitly mentions suffice, by themselves, practically useful way by quantum mechan- with no acknowledgment of the existence ics. On the one hand, a key basic pre- of consciousness, to completely specify all cept of the quantum approach, as it is physical properties of the universe, includ- both practiced and taught, is that choices ing the activities of our bodies and brains. made by human beings play a key and According to the conceptual structure of irreducible role in the dynamics. On the classical physics, everything physical would other hand, the great disparity within clas- go on just the same if nothing existed but sical physics between the experiential and the physical properties explicitly mentioned physical aspects of nature is resolved in the in the theory. The second problem is that quantum approach by altering the assump- within that conceptual framework of clas- tions about the nature of the physical uni- sical physics neither planets nor electrons, verse. The physical world, as it appears in P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

quantum approaches to consciousness 883

the theory, is transformed from a structure basic laws of nature, as they are now under- based on substance or matter to one based stood, not only fail to determine how we on events, each of which has both expe- will act, but, moreover, inject our choices riential aspects and physical aspects: Each about how to act directly into the dynamical such event injects information, or “knowl- equations. edge,” into an information-bearing mathe- This altered role of conscious agents matically described physical state. An impor- is poetically expressed by Bohr’s famous tant feature of this radical revamping of dictum: the conceptual foundations is that it leaves In the great drama of existence we ourselves unchanged, at the practical level, most of are both actors and spectators. (Bohr, 1958, classical physics. Apart from making room p. 81; 1963,p.15) for, and a strict need for, efficacious conscious choices, the radical changes introduced at It is expressed more concretely in statements the foundational level by quantum mechan- such as the following: ics preserve at the pragmatic level almost all The freedom of experimentation, presup- of classical physics. posed in classical physics, is of course In the remainder of this introductory sec- retained and corresponds to the free choice tion I sketch out the transition from the of experimental arrangement for which the classical physics conception of reality to mathematical structure of the quantum von Neumann’s application of the princi- mechanical formalism offers the appropri- ples of quantum physics to our conscious ate latitude. (Bohr, 1958,p.73) brains. In succeeding sections I describe the The most important innovation of quantum most prominent of the many efforts now theory, from a philosophical perspective, is being made by physicists to apply von Neu- the fact that it is formulated in terms of an mann’s theory to recent developments in interaction between the physically described neuroscience. world and conscious agents who are, within The quantum conception of the connec- the causal structure defined by the known phys- tion between the psychologically and physical laws, free to choose which aspect of nature ically described components of scientific they will probe. This crack, or gap, in the practice was achieved by abandoning the mechanistic world view leads to profound classical picture of the physical world that changes in our conception of nature and the had ruled science since the time of New- place of human beings within it. ton, Galileo, and Descartes. The building Another key innovation pertains to blocks of science were shifted from descrip- the nature of the stuff of the phys- tions of the behaviors of tiny bits of mind- ically/mathematically described universe. less matter to accounts of the actions that we The switch is succinctly summarized in take to acquire knowledge and of the knowl- Heisenberg’s famous assertion: edge that we thereby acquire. Science was thereby transformed from its 17th-century The conception of the objective reality of form, which effectively excluded our con- the elementary particles has thus evapo- scious thoughts from any causal role in the rated not into the cloud of some obscure mechanical workings of Nature, to its 20th- new reality concept, but into the transpar- century form, which focuses on our active ent clarity of a mathematics that repre- engagement with Nature and on what we sents no longer the behavior of the particle but rather our knowledge of this behavior. can learn by taking appropriate actions. (Heisenberg, 1958a) Twentieth-century developments have thus highlighted the fact that science is a What the quantum mathematics describes is human activity that involves us not as passive not the locations of tiny bits of matter. What witnesses of a mechanically controlled uni- is described by the mathematics is a causal verse, but as agents who can freely choose structure embedded in space-time that car- to perform causally efficacious actions. The ries or contains information or knowledge, P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

884 the cambridge handbook of consciousness

but no material substance. This structure is, ton considered this non-local feature of his on certain occasions, abruptly altered by dis- theory to be unsatisfactory, but proposed crete events that inject new information into no alternative. Eventually, Albert Einstein, it. But this carrier structure is not purely building on ideas of James Clerk Maxwell, passive. It has an active quality. It acts as a constructed a local classical theory in which bearer of “objective tendencies” or “poten- all dynamical effects are generated by con- tia” or “propensities” for new events to occur tact interactions between mathematically (Heisenberg, 1958b, p. 53). described properties localized at space-time To appreciate this new conception of points and in which no effect is transmitted the connection between the psychologically faster than the speed of light. described empirical part and the mathemat- All classical physics models of Nature ically described physical part of the new sci- are deterministic: The state of any isolated entific description of physical phenomena, system at any time is completely fixed by one needs to contrast it with what came the state of that system at any earlier time. before. The Einstein-Maxwell theory is deterministic in this sense; it is also “local” in the just- mentioned sense that all interactions are via The Classical-Physics Approach contact interactions between neighboring Classical physics arose from the theoreti- localized mathematically describable prop- cal effort of Isaac Newton to account for erties, and no influence propagates faster the findings of Johannes Kepler and Galileo than the speed of light. Galilei. Kepler discovered that the planets By the end of the 19th century certain dif- move in orbits that depend on the loca- ficulties with the general principles of clas- tion of other physical objects – such as the sical physical theory had been uncovered. sun – but not on the manner or the tim- One such difficulty was with “black-body ings of our observations: Minute-by-minute radiation.” If one analyzes the electromag- viewings have no more influence on a plan- netic radiation emitted from a tiny hole in etary orbit than daily, monthly, or annual a big hollow heated sphere, then it is found observations. The nature and timings of our that the manner in which the emitted energy observational acts have no effect at all on is distributed over the various frequencies the orbital motions described by Kepler. depends on the temperature of the sphere, Galileo observed that certain falling terres- but not upon the chemical or physical char- trial objects have similar properties. Newton acter of the interior surface of the sphere: then discovered that he could explain simul- The spectral distribution depends neither taneously the celestial findings of Kepler and on whether the interior surface is smooth the terrestrial findings of Galileo by postu- or rough nor on whether it is metallic or lating, in effect, that all objects in our solar ceramic. This universality is predicted by system are composed of tiny planet-like par- classical theory, but the specific form of the ticles whose motions are controlled by laws predicted distribution differs greatly from that refer to the relative locations of the var- what is empirically observed. ious particles and that make no reference In 1900 Max Planck discovered a universal to any conscious acts of experiencing. These law of black-body radiation that matches the acts are taken to be simply passive witness- empirical facts. This new law is incompatible ings of macroscopic properties of large con- with the basic principles of classical phys- glomerations (such as tables and chairs and ical theory and involves a new constant of measuring devices) of the tiny individually Nature, which was identified and measured invisible particles. by Planck and is called “Planck’s Constant.” Newton’s laws involve instantaneous By now a huge number of empirical effects action at a distance: Each particle has an have been found that depend upon this con- instantaneous effect on the motion of every stant and that conflict with the predictions other particle, no matter how distant. New- of classical physical theory. P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

quantum approaches to consciousness 885

During the 20th century a theory was under the control of the local deterministic devised that accounts for all of the successful Schroedinger equation, then the location of predictions of classical physical theory and the center of the moon would be represented also for all of the departures of the predic- in the theory by a structure spread out over tions of classical theory from the empirical a large part of the sky, in direct contradiction facts. This theory is called quantum theory. to normal human experience. No confirmed violation of its principles has This smeared-out character of the posi- ever been found. tion of (the center-point of) a macroscopic object is a consequence of the famous Heisenberg Uncertainty Principle, com- The Quantum Approach bined with the fact that tiny uncertainties at The core idea of the quantum approach is the microscopic level usually get magnified the seminal discovery by Werner Heisenberg over the course of time, by the Schroedinger that the classical model of a physical system equation acting alone, to large uncertainties can be considered to be an approximation in macroscopic properties, such as location. to a quantum version of that model. This Thus a mathematical equation – the quantum version is constructed by replac- Schroedinger equation – that is a direct ing each numerical quantity of the classical mathematical generalization of the laws of model by an action: by an entity that acts motion of classical physical theory and that on other such entities and for which the yields many predictions of incomparable order in which the actions are performed accuracy – strongly conflicts with many facts matters. The effect of this replacement is to of everyday experience (e.g., with the fact convert each point-like particle of the clas- that the apparent location of the center of sical conceptualization – such as an electron the moon is well defined to within, say 10 – to a smeared-out cloudlike structure that degrees, as observed from a location on the evolves, almost always, in accordance with surface of the earth). Contradictions of this a quantum mechanical law of motion called kind must be eliminated by a satisfactory for- the Schroedinger equation. This law, like its mulation of quantum theory. classical analog, is local and deterministic: To put the accurate predictions of the The evolution in time is controlled by con- quantum mathematics into the framework tact interactions between localized parts, of a rationally coherent and practically useful and the physical state of any isolated sys- physical theory, the whole concept of what tem at any time is completely determined physical science is was transformed from its from its physical state at any earlier time 19th-century form – as a theory of the prop- by these contact interactions. The cloudlike erties of a mechanical model of Nature in structure that represents an individual “par- which we ourselves are mechanical parts ticle,” such as an electron or proton, tends, – to a theory of the connection between under the control of the Schroedinger equa- the physically and psychologically described tion, to spread out over an ever-growing aspects of actual scientific practice. In actual region of space, whereas according to the practice we are agents that probe nature in ideas of classical physics an electron always ways of our own choosing in order to acquire stays localized in a very tiny region. knowledge that we can use. I now describe in The local deterministic quantum law more detail how this pragmatic conception of motion is, in certain ways, incredibly of science works in quantum theory. accurate: It correctly fixes to one part in a hundred million the values of some measur- “The Observer” and “The Observed able properties that classical physics can- System” in Copenhagen Quantum Theory not predict. However, it does not correlate directly to human experience. For exam- The original formulation of quantum the- ple, if the state of the universe were to ory is called the Copenhagen Interpretation have developed from the big bang solely because it was created by the physicists P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

886 the cambridge handbook of consciousness

that Niels Bohr had gathered around him ever, the detailed forms of the laws that gov- in Copenhagen. A central precept of this ern the evolution in time of this mathemat- approach is that, in any particular appli- ical structure, and of the rules that specify cation of quantum theory, Nature is to the connection of this mathematical struc- be considered divided into two parts: “the ture to the empirical facts, are very different observer” and “the observed system.” The in the two theories. observer consists of the stream of conscious- I am endeavoring here to avoid math- ness of a human agent, together with the ematical technicalities. But the essential brain and body of that person, and also the conceptual difference between the two measuring devices that he or she uses to approaches rests squarely on a certain tech- probe the observed system. nical difference. This difference can be illus- Each observer describes himself and his trated by a simple two-dimensional picture. knowledge in a language that allows him to communicate to colleagues two kinds of The Paradigmatic Example information: how he has acted in order to prepare himself – his mind, his body, and Consider an experiment in which an exper- his devices – to receive recognizable and imenter puts a Geiger counter at some reportable data and what he has learned from location with the intention of finding out the data he thereby acquires. This descrip- whether or not this device will “fire” dur- tion is in terms of the conscious experiences ing some specified time interval. The exper- of the agent himself. It is a description of iment is designed to give one of two possible his intentional probing actions and of the answers: ‘Yes’, the counter will fire during experiential feedbacks that he subsequently the specified interval, or ‘No’, the counter receives. will not fire during this specified interval. In actual scientific practice the experi- This is the paradigmatic quantum measure- menters are free to choose which experiment process. ments they perform: The empirical proce- This experiment has two alternative dures are determined by the protocols and mutually exclusive possible responses, ‘Yes’ aims of the experimenters. This element of or ‘No.’ Consequently, the key mathematical freedom is emphasized by Bohr in state- connections can be pictured in a two- ments such as the following: dimensional space, such as the top of your desk. To my mind there is no other alternative Consider two distinct points on the top than to admit in this field of experience, of your desk called zero and p. The displace- we are dealing with individual phenomena ment that would move a point placed on zero and that our possibilities of handling the to the point p is called a vector. Let it be called measuring instruments allow us to make a choice between the different complementary V. Suppose V has unit length in some units, types of phenomena that we want to study. say meters. Consider any two other displace- (Bohr, 1958,p.51) ments V1 and V2 on the desk top that start from zero, have unit length, and are perpen- This freedom to choose is achieved in dicular to each other. The displacement V the Copenhagen formulation of quantum can be formed in a unique way by making a theory by placing the empirically/psycho- (positive or negative) displacement along V1 logically described observer outside the followed by a (positive or negative) displace- observed system that is being probed and ment along V2 . Let the lengths of these two then subjecting only the observed system to displacements be called X1 and X2 , respec- the rigorously enforced mathematical laws. tively. The theorem of Pythagoras says that The observed system is, according to X1 squared plus X2 squared is one (unity). both classical theory and quantum theory, Quantum theory is based on the idea that describable in terms of mathematical prop- the various experiencable outcomes have erties assigned to points in space-time. How- “images” in a vector space. The vector V1 P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

quantum approaches to consciousness 887

mentioned above is the image, or represen- The orientation of the set of basis vectors tation, in the vector space of the possible is thus, from a mathematical standpoint, a outcome ‘Yes,’ whereas V2 represents ‘No.’ I variable that can be, and is, specified inde- do not try to describe here how this mapping pendently of the state V of the system being of possible experiencable outcomes into cor- probed. responding vectors is achieved. But the basic This entry into the dynamics of choices presumption in quantum theory is that such made by the experimenters is not at all sur- a mapping exists. prising. If the experimenters are considered The vector V represents the state of the to stand outside, and apart from, the system to-be-observed system, which has been being observed, as specified by the Copen- prepared at some earlier time and has been hagen approach, then it is completely rea- evolving in accordance with the Schroed- sonable and natural that the choices made inger equation. The vector V1 represents the by the experimenters (about how to probe state that this observed system would be the observed system) should be treated as known to be in if the observed outcome variables that are independent of the vari- of the measurement were ‘Yes.’ The vector ables that specify the physical state of the V2 represents the state that the observed system they are probing. system would be known to be in if the Bohr (1958,p.100) argued that quantum observed result of the measurement were theory should not be applied to living sys- ‘No.’ Of course, the directions of the two tems. He also argued that the classical con- perpendicular vectors V1 and V2 depend cepts were inadequate for that purpose. So upon the exact details of the experiment: the strict Copenhagen approach is simply to on exactly where the experimenters have renounce the applicability of contemporary placed the Geiger counter and on other physical theories, both classical and quan- details controlled by the experimenters. tum, to neurobiology. The outcome of the probing measurement will be either V1 (Yes) or V2 (No). Von Neumann’s Formulation The predicted probability for the outcome to be ‘Yes’ is X1 squared, and the predicted The great mathematician and logician John probability for the outcome to be ‘No’ is von Neumann (1955/1932) rigorized and X2 squared. These two probabilities sum to extended quantum theory to the point of unity, by virtue of the theorem of Pythago- being able to corporate the devices, the body, ras. The sudden jump of the state from V and the brain of the observers into the phys- to either V1 or V2 is called a “quantum ically described part of the theory, leaving, jump.” The general theory is expressed in in the psychologically described part, only terms of a many-dimensional generalization the stream of conscious experiences of the of your desktop. This generalization is called agents. The part of the physically described a Hilbert space, and every observable state system being directly acted upon by a psy- of a physical system is a represented by a chologically described “observer” is, accord- “vector” in such a space. ing to von Neumann’s formulation, the brain The crucial, though trivial, logical point of that observer (von Neumann, 1955,p.421). can now be stated: The two alternative pos- The quantum jump of the state of the brain sible outcomes, ‘Yes’ or ‘No’ of the chosen- of an observer to the ‘Yes’ basis state (vector) by-the-experimenter experiment, are asso- then becomes the representation, in the state ciated with a pair of perpendicular unit- of that brain, of the conscious acquisition of length vectors called “basis vectors.” The the knowledge associated with that answer orientation (i.e., directions) of the set of basis ‘Yes.’ Thus the physical features of the brain vectors, V1 and V2 , enters into the dynamics state actualized by the quantum jump to as a free variable controlled by the exper- the state V1 associated with the answer ‘Yes’ imental conditions, which are specified in constitute the neural correlate of that per- practice by choices made by experimenters. son’s conscious experience of the feedback P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

888 the cambridge handbook of consciousness

‘Yes.’ This fixes the essential quantum link Von Neumann shifted the boundary between consciousness and neuroscience. between the observer and the observed sys- This is the key point! Quantum physics is tem, in a series of steps, until the bodies built around “events” that have both phys- and brains of all observers, and everything ical and phenomenal aspects. The events else that classical physics would describe are physical because they are represented in as “physical,” were included as part of the the physical/mathematical description by a observed system, and he showed that this quantum jump to one or another of the basis form of the theory is essentially equiva- state vectors defined by the agent/observer’s lent, in practice, to the Copenhagen inter- choice of what question to ask. If the result- pretation. But it evades an unnatural fea- ing event is such that the ‘Yes’ feedback ture imposed by Bohr: It bypasses the ad experience occurs, then this event “col- hoc separation of the dynamically unified lapses” the prior physical state to a new phys- physical world into two differently described ical state compatible with that phenome- parts. Von Neumann’s final placement of the nal experience. Mind and matter thereby boundary allows the psychological descrip- become dynamically linked in a way that is tion to be – as is natural – the description of causally tied to the agent’s free choice of how a stream of conscious experiences that are he or she will act. Thus, a causal dynamical the experiential sides of a sequence of events connection is established between (1) a per- whose physical sides actualize the neural cor- son’s conscious choices of how to act, (2) relates of those experiences. that person’s consciously experienced incre- It is important that von Neumann’s sys- ments in knowledge, and (3) the physical tematic enlargement of the physical system actualizations of the neural correlates of the to include eventually the bodies and brains experienced increments in knowledge. of the observers does not disrupt the basic This conceptualization of the structure mathematical structure of the theory. In par- of basic physical theory is radically differ- ticular, it does not alter the critical need to ent from what it was in classical physics. specify the orientation of the set of basis vec- Classical physics was based on a guess that tors (e.g., V1 and V2) to make the theory worked very well for two centuries; namely, work. The specification of the basis states con- the notion that the concepts that provided tinues to be undetermined by anything in con- an “understanding” of our observations of temporary physical theory, even when the phys- planets and falling apples would continue to ical description is extended to include the entire work all the way down to the elementary physical world, including the bodies and brains particle level. That conjecture worked well of all human observers. until science became able to explore what This leap by von Neumann from the was happening at the elementary particle or realm of atomic physics to the realm of neu- atomic level. Then it was found that that roscience was way ahead of its time. Neu- simple “planetary” idea could not be right. roscience was then in a relatively primitive Hence scientists turned to a more sophisti- state compared to what it is today. It had cated approach that was based less on sim- a long way to go before mainstream inter- plistic ontological presuppositions and more est turned to the question of the connection on the empirical realities of actual scientific between brains and conscious experiences. practice. But 70 years of brain science has brought This new conceptual structure is not the empirical side up to the level where the some wild philosophical speculation. It details of the mind-brain connections are rationally yields – when combined with the being actively probed, and intricate results statistical rule associated with the theorem are being obtained that can be compared to of Pythagoras described above – all the the predictions of the psychophysical theory pragmatic results of quantum theory, which prepared long ago by John von Neumann. include, as special cases, all the valid predic- It is evident that a scientific approach tions of classical physics! to brain dynamics must in principle use P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

quantum approaches to consciousness 889

quantum theory in order to deal properly experiences, associated with a newly cho- with brain processes that depend heavily sen course of action, into the stream of con- on chemical and ionic processes. For exam- sciousness of the human agent and that, on ple, the release of neurotransmitter from a the physical side, actualize brain states that nerve terminal is controlled by the motions contain the neural correlates of those expe- of calcium ions, and these ions are small riences. This is the basic dynamical process enough so that the deterministic laws of that underlies the quantum approach to con- classical physics necessarily fail. Quantum sciousness. theory must in principle be used to describe the ion dynamics. But once one goes over Summary to a quantum-mechanical description of the ionic components of the brain, one must fol- The essential difference at the basic concep- low through and treat also the conglomera- tual level between the quantum and classi- tion of these cloudlike entities and the other cal approaches to consciousness is that the similar components of the brain by the quan- classical principles make no mention of con- tum rules in order to recover the connection sciousness. The causal structure is in prin- of that mathematical structure to our con- ciple completely “bottom up.” Everything scious experiences! is, in principle, fully determined by what According to this quantum description, goes on at the microscopic atomic level, the state of the brain is itself an expand- and any dependence of microscopic proper- ing cloudlike structure in a high-dimensional ties upon macroscopic properties, or on con- space. Just as the various points in the cloud sciousness, is, in the end, a round-about con- that describes a single particle represent dif- sequence of laws expressible exclusively in ferent classically conceived possibilities for terms of properties of atomic particles and of that single particle (where it is and how it the physical fields that they produce. But in is moving), so do the various points in the quantum theory the local-deterministic (i.e., cloud that describe the brain represent dif- bottom-up) physical process is in principle ferent classically conceived possible states of causally incomplete. It fixes, by itself, neither that brain. This cloudlike structure can, and our actions nor our experiences, nor even generally will, encompass, with appreciable any statistical prediction about how we will weights, many conflicting classical possibili- act or what we will experience. The bottom- ties. up process alone is unable to make statisti- The job of the brain is to accept clues cal predictions, because the statistical pre- from the environment and then to con- dictions depend upon the choice of a set struct a “Template for Action” (an “execu- of basis vectors, and the bottom-up local- tive” pattern of neurological activity) that, if deterministic quantum process does not fix it endures, will issue the sequence of neural this choice. signals that will cause some specific (and, it is This reorganization of the dynamical hoped, appropriate) action to unfold. In the structure leads to an altered perspective on cloudlike structure that represents a brain, the entire scientific enterprise. The psycho- many alternative conflicting Templates for logically described empirical side of scien- Action can arise. The generation, within the tific practice is elevated from its formerly quantum state of the brain, of important subservient status – as something that should components representing conflicting clas- be deduced from, or constructed from, the sical possibilities should occur particularly already dynamically complete physical side – when the low-level (essentially mechanical) to the new status of coequal dynami- processes cannot come to agreement on the cal partner. Science becomes the endeavor best course of action. In this circumstance, to describe the two-way interplay between the quantum mechanical rules allow choices the psychologically and physically described to be made that produce quantum jumps aspects of nature, rather than an attempt that, on the psychological side, inject new to deduce the existence and properties of P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

890 the cambridge handbook of consciousness

our streams of conscious experiences from a conclusion of the von Neumann formula- presumed-to-be-dynamically complete local tion of quantum theory; namely, that a con- mechanical model. scious human being can behave in ways that Within the von Neumann framework our a classical mechanical model cannot. Pen- conscious choices fix the orientations of the rose’s argument, if valid, would yield this basis vectors. These choices can strongly same conclusion, but within a framework influence our actions. Thus these influences that relies not on quantum concepts, which need not be illusions. The theory provides, are generally unknown to cognitive scien- as we see in the sectionon the von Neu- tists, but rather on Godel-type¨ arguments, mann/Stapp approach, a specific mechanism which are familiar to some of them. that allows our conscious “free” choices to The general idea of Penrose’s argument significantly influence our physical actions. is to note that, because of the mathematically deterministic character of the laws of Pragmatic Neuroscience classical physics, the output at any specified finite time of any computer behaving Von Neumann, in his 1932 book, followed in accordance with the classical laws should the Copenhagen tack of focusing on scien- in principle be deducible, to arbitrarily good tific practice, rather than ontological issues. accuracy, from a finite-step procedure based Indeed, it can be argued that science is on a finite set of mutually consistent rules intrinsically pragmatic, rather than ontolog- that encompass the laws of arithmetic. But ical. The true nature of things, other than then a human being who can be adequately our experiences themselves, can never be modeled as a classical computer should be provably ascertained by the methods of sci- able to know, at any finite time, the truth ence. Thus von Neumann’s formulation of only of those statements that can be deduced quantum theory provides the foundations from a finite-step computation based on the of a pragmatic neuro-psycho-dynamics that finite set of rules that govern that com- is built on contemporary physical theory, puter. Yet, Godel-theorem-type¨ arguments rather than an inadequate classical physics. allow real mathematicians to know, given All quantum approaches to consciousness any finite set of consistent logical rules that build upon this foundation laid by von Neu- encompass the laws of arithmetic, the truth mann, but various physicists have proposed of mathematical statements that cannot be different ways of developing that core struc- deduced by any finite-step proof based on ture. We now turn to the descriptions of a those rules. This seems to imply that a real number of these proposals. mathematician can know things that no classical physics model of himself could ever know; namely, the truth of statements that The Penrose-Hameroff Approach his classical computer simulation could not establish in a finite time. Perhaps the most ambitious attempt to cre- Filling in the details of this argument is ate a quantum theory of consciousness is the not an easy task. Penrose spends the bet- one of Roger Penrose and Stuart Hameroff. ter part of five chapters in The Emperor’s Their proposal has three parts: the Godel¨ New Mind (Penrose, 1986) and some 200 Part, the Gravity Part, and the Microtubule pages in Shadows of the Mind (Penrose, 1994) Part. explaining and defending this thesis. How- The Godel¨ Part, which is due to Penrose, ever, the Harvard philosopher Hilary Put- is an effort to use the famous Godel¨ Incom- nam challenged Penrose’s conclusion in a pleteness Theorem to prove that human debate appearing in the New York Times beings have intellectual powers that they Review of Books (Putnam, 1994), and numer- could not have if they functioned in accor- ous logicians have since weighed in, all, to dance with the principles of classical phys- my knowledge, challenging the validity of ical theory. Proving this would reaffirm a Penrose’s argument. Thus, the Godel¨ Part of P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

quantum approaches to consciousness 891

the Penrose-Hameroff approach cannot now it are made within a classical physics ide- be regarded as having been established suc- alization. But serious problems arise when cessfully. the quantum character of matter is consid- The Gravity Part of the Penrose-Hameroff ered. For, according to orthodox quantum approach addresses a key question pertain- theory, a particle, such as an electron or an ing to quantum dynamics: Exactly when do ion, has no well-defined location: Its loca- the sudden quantum jumps occur? In von tion is specified by a smeared-out “probabil- Neumann’s theory these jumps should pre- ity cloud.” But if the locations of the material sumably occur when the neural correlates of particles are not well defined then, accord- conscious thoughts become sufficiently well ing to General Relativity, neither is the form formed. But von Neumann gives no precise of the space-time structure in which the par- rule for when this happens. ticle structures are embedded. The lack of specificity on this issue of pre- Penrose conjectures that Nature abhors cisely “when” is a serious liability of the von uncertainty in the structure of space-time Neumann theory, insofar as it is construed as and that when too much ambiguity arises a description of the ontological mind-matter in the space-time structure a quantum jump reality itself. That difficulty is the basic rea- to some less ambiguous structure will occur. son why both the original Copenhagen for- This “principle” allows him to tie quantum mulation and von Neumann’s extension of jumps to the amount of uncertainty in the it eschew traditional ontological commit- structure of space-time. ments. They hew rather to the pragmatic There is no compelling reason why position that the job of science is to establish Nature should be any more perturbed by an useful practical connections between empir- uncertainty in the structure of space-time ical findings and theoretical concepts, rather than by an uncertainty in the distribution than advancing shaky speculations about the of matter. However, by adopting the princi- ultimate nature of reality. The pragmatic ple that Nature finds intolerable an excessive position is that theoretical ideas that opti- ambiguity in the structure of space-time Pen- mally provide reliable practical connections rose is able to propose a specific rule about between human experiences constitute, when the quantum jumps occur. themselves, our best scientific understanding Penrose’s rule depends on the fact of “reality.” Added ontological superstruc- that Planck’s constant gives a connection tures are viewed as not true science, because between energy and time: This constant additions that go beyond optimal theoretical divided by any quantity of energy gives a descriptions of connections between human corresponding interval of time. Thus if an experiences cannot be tested empirically. energy associated with a possible quantum Penrose wants to provide an ontology that jump can be defined, then a time interval has “real quantum jumps.” Hence he must associated with that potential jump becomes face the question: When do these jumps specified. occur? He seeks to solve this problem by To identify the pertinent energy, consider linking it to a problem that arises when one a simple case in which, say, a small object attempts to combine quantum theory with is represented quantum mechanically by a Einstein’s theory of gravity. small cloud that divides into two similar Einstein’s theory of gravity, namely Gen- parts, one moving off to the right and the eral Relativity, is based on the idea that other moving off to the left. Both parts space-time is not a rigid flat structure, as of the cloud are simultaneously present, had previously been thought, but is rather and each part produces a different distortion a deformable medium, and that the way it is of the underlying space-time structure, deformed is connected to the way that mat- because matter is distributed differently in ter is distributed within it. This idea was the two cases. One can compute the amount developed within the framework of classi- of energy that it would take to pull apart, cal physical theory, and most applications of against their gravitational attraction, two P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

892 the cambridge handbook of consciousness

copies of the object, if each copy is located ity of these anesthetics to diminish the abil- at the position specified by one of the two ity of the special electron to move from one clouds. If one divides Planck’s constant stable location to the other. This suggests by this “gravitational energy,” then a time a possible close connection between con- interval associated with this distortion of sciousness and the configurational activity of space-time into these two disparate struc- microtubules. tures becomes defined. Penrose proposes This putative linkage allows an empirical that this time interval is the duration of test of Penrose’s rule to be made. time for which Nature will endure this Suppose, in keeping with the case consid- bifurcation of its space-time structure into ered by Penrose, you are in a situation where the two incompatible parts before jumping one of two possible experiences will proba- to one or the other of these two forms. bly occur. For example, you might be star- This conjectured rule is based on two ing at a Necker Cube, or walking in a dark very general features of Nature: Planck’s uni- woods when a shadowy form jumps out and versal constant of action and the Newton- you must choose “fight” or “flight,” or per- Einstein universal law of gravitation. This haps you are checking your ability to freely universality makes the rule attractive, but choose to raise or not raise your arm. Thus no reason is given why Nature must com- one of two alternative possible experiences ply with this rule. is likely to occur. Various experiments sug- Does this rule have any empirical gest that it takes about a half-second for an support? experience to arise. Given this time interval, An affirmative answer can be provided Penrose’s formula specifies a certain corre- by linking Penrose’s rule to Hameroff’s sponding energy. Then Hameroff can com- belief that consciousness is closely linked to pute, on the basis of available information the microtubular substructure of the neuron concerning the two configurational states of (Hameroff & Penrose, 1996). the tubulin molecule, how many tubulin- It was once thought that the interiors of molecule configurational shifts are needed to neurons were basically structureless fluids. give this energy. That conclusion arose from direct micro- The answer is about 1% of the estimated scopic examinations. But it turns out that number of tubulin molecules in the human in those early studies the internal substruc- brain. This result seems reasonable. Its rea- ture was wiped out by the fixing agent. It is sonableness is deemed significant because now known that neurons are filled with an the computed fraction could have come out intricate structure of microtubules. to be perhaps billions of times smaller than Each microtubule is a cylindrical struc- or billions of times greater than 100%. The ture that can extend over many millimeters. fact that the computed value is “in the ball- The surface of the cylinder is formed by a park” supports the idea that consciousness spiral chain of tubulin molecules, with each may indeed be connected via gravity to tubu- circuit formed by 13 of these molecules. The lin configurational activity. tubulin molecule has a molecular weight of Given this rather radical idea – it was pre- about 110,000, and it exists in two slightly viously thought that gravity was not essen- different configurational forms. Each tubu- tial to consciousness (orbiting astronauts lin molecule has a single special electron that can think) and that the microtubules were can be in one of two relatively stable loca- merely a construction scaffolding for the tions. The molecule’s configurational state building and maintenance of the physical depends on which of these two locations this structure of the neurons – many other exotic special electron is occupying. possibilities arise. The two configurational Hameroff is an anesthesiologist, and he forms of the tubulin molecule mean that it noted that there is close correspondence be- can hold a “bit” of information, so maybe tween, on the one hand, the measured the microtubular structure forms the sub- effects of various anesthetics upon con- strate of a complex computer located within sciousness and, on the other hand, the capac- each neuron, thus greatly expanding the P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

quantum approaches to consciousness 893

computational power of the brain. And ening or disappearance of the interference maybe each such computer is in fact a “quan- effects). tum computer.” And maybe these quantum If a system interacts with its environment, computers are all linked together to form it is difficult to prevent a “perceptible influ- one giant brain-wide quantum computer. ence” of the system on the environment. If And maybe these hollow microtubes form even a single one of the thousands of par- wave guides for quantum waves. ticles in the environment is displaced by a These exotic possibilities are exciting and discernible amount, then the coherence is heady. They go far beyond what conservative lost, and the quantum interference effect physicists are ready to accept and far beyond will disappear. what the 1% number derived from Penrose’s Because the medium in which the puta- rule actually supports. What is supported is tive quantum information waves are moving merely a connection between consciousness involves different conformational states of and microtubular activity, without the pres- huge tubulin molecules of molecular weight ence of the further stringent coherence condi- ∼110,000, it would seemingly be exceed- tions required for the functioning of a quan- ingly hard to ensure that the passage of these tum computer. waves will not disturb even one particle of Coherence means preservation of the the environment by a discernible amount. “phase” relationships that allow waves that Max Tegmark wrote an influential paper have traveled via different paths to come in Physical Review E (Tegmark, 2000) that back together so that, for example, crest mathematically buttressed the intuition meets crest and trough meets trough to build of most physicists that the macroscopic an enhanced effect. Quantum computation coherence required by Penrose-Hameroff – requires an effective isolation of the quan- namely that the microtubular conformal tum informational waves from the surround- states can form the substrate of a quantum ing environment, because any interaction computer that extends over a large part between these waves and the environment of the brain – could not be realized in a tends to destroy coherence and the required living human brain. Tegmark concluded isolation is difficult to maintain in a warm, that the coherence required for macro- wet, noisy brain. scopic quantum computation would be lost The simplest system that exhibits a in a ten-trillionth of a second and hence behavior that depends strongly on quan- should play no role in consciousness. This tum interference effects, and for which the paper was widely heralded. However, maintenance of coherence is essential, is the Hagan, Hameroff, and Tuszynski (2002) famous double-slit experiment. When pho- wrote a rejoinder in a later issue of the same tons of a single wave length are allowed journal. They argued that some of Tegmark’s to pass, one at a time, through a pair of assumptions departed significantly from closely spaced narrow slits, and each pho- those of the Penrose-Hameroff model. The ton is later detected by some small detec- associated corrections lengthened the coher- tion device that is embedded in a large ence time by 8 or 9 orders of magnitude, thus array of such devices, one finds that if the bringing the situation into a regime where photonic system is not allowed to perceptibly the non-equilibrium conditions in a living influence any environmental degree of free- brain might become important: Energetic dom on its way to the detection device, biological processes might conceivably inter- then the pattern of detected events depends vene in a way that would make up the still- on an interference between the parts of the needed factor of ten thousand. However, the beam passing through the two different details of how this might happen were not slits. This pattern is very different from supplied. Hence the issue is, I believe, still what it is if the photon is allowed to up in the air, with no detailed explanation perceptibly disturb the surrounding envi- available to show how the needed macro- ronment. Disturbing the environment pro- scopic quantum coherence could be main- duces a “decoherence” effect (i.e., a weak- tained in a living human brain. P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

894 the cambridge handbook of consciousness

It must be stressed, however, that these The question thus arises: What determines exotic “quantum computer” effects are not these choices? necessary for the emergence of strong quan- One possibility is that these choices arise tum effects within the general framework in some yet-to-be-specified way from what supplied by the combination of Penrose’s we conceive to be the idealike aspect of rule pertaining to gravity and Hameroff’s reality. That option was pursued by Pen- claim concerning the importance of micro- rose, with his suggestion that our thoughts tubules. According to von Neumann’s gen- are linked to Plato’s world of ideal forms. eral formulation, the state of the brain – or Another – seemingly different – possibility of the microtubular part of the brain – is is that a physical description exists that is adequately represented by what physicists more detailed than the smeared-out cloud- call the “reduced density matrix” of that sub- like structures of the orthodox formulations system. This representation depends only and that this more detailed physical descrip- on the variables of that subsystem itself tion determines all features left undeter- (i.e., the brain, or microtubular array), but mined in the orthodox formulations. nevertheless takes adequate account of the This second approach was developed by interactions of that system with the envi- David Bohm (1952: Bohn & Hiley, 1993). ronment. It keeps track of the quantum His formulation of quantum theory postu- coherence or lack thereof. Penrose’s rule can lates, in effect, the existence of the old- be stated directly in terms of the reduced fashioned world of classical physical theory. density matrix, which displays, ever more This classical-type world is supposed to exist clearly as the interaction with the environ- in addition to the cloudlike wave function of ment grows, the two alternative states of the orthodox quantum theory and is supposed to brain – or of the microtubular array – that evolve in a way completely determined by Nature must choose between. This reduced- what precedes it in time. Bohm specifies density-matrix representation shows that new laws of motion that are able to reinstate the powerful decoherence effect produced determinism in a way compatible with the by strong interactions with the environ- predictions of quantum theory, but at the ment actually aids the implementation of expense of a very explicit abandonment of Penrose’s rule, which is designed to spec- locality: Bohm’s theory entails very strong, ify when the quantum jump occurs (and and very long-range, instantaneous action- perhaps to which states the jump occurs). at-a-distance. The capacity of the brain to be or not to One serious failing of Bohm’s approach is be a quantum computer is a very different that it was originally formulated in a non-re- question, involving enormously more strin- lativistic context, and it has not yet – after gent conditions. It thus is important, for a half-century and great effort – been ex- logical clarity, to separate these two issues tended to cover deterministically the most of the requirements for quantum computa- important domain in physics; namely, the tion and for quantum jumps, even though realm of quantum electrodynamics. This is they happen to be interlocked in the par- the theory that covers the atoms that make ticular scenario described by Penrose and up our bodies and brains, along with the Hameroff. tables, chairs, automobiles, and computers that populate our daily lives. This deficiency means that Bohm’s theory is, at present, The Bohm Approach primarily a philosophically interesting curiosity, not a viable deterministic physical The Copenhagen and von Neumann for- theory. mulations of quantum theory are non- Also, Bohm’s theory, at least in its origi- deterministic. Both specify that human nal form, is not really germane to the issue choices enter into the dynamics, but neither of consciousness. For Bohm’s theory success- specifies the causal origins of these choices. fully achieved its aim, which was precisely P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

quantum approaches to consciousness 895

to get rid of consciousness (i.e., to elimi- controllable, and knowable input boundary nate consciousness from the basic dynamical conditions. equations, just as classical physics had done). In Bohm’s theory these choices are not Bohm recognized, later on, that some actually free: Freedom is an illusion. The understanding of consciousness was needed, apparently free choice is, at a deeper dynam- but he was led instead to the notion of an ical level, completely determined by phys- infinite tower of mechanical levels, each con- ical conditions, just as it was in classical trolling the one below, with consciousness physics. However, the putative existence of somehow tied to the mystery of the infi- this deeper dynamical underpinning does nite limit (Bohm, 1986, 1990). This infinite- not upset scientific practice. It does not dis- tower idea tends to diminish the great place, within science, the orthodox quantum achievement of the original theory, which dynamics. The analysis by Heisenberg shows was to reinstate physical determinism in a that, even within the context of a deter- simple way. ministic Bohmian mechanics, the human Perhaps the most important use of observers can never determine, or know,to Bohm’s model is to provide an understand- which of the conceivable logically possible ing of the consequences of assuming, as some classical Bohmian worlds their experiences philosophers do, that we live in a world in belong. The Heisenberg Uncertainty Princi- which everything is deterministically fixed ple is a limitation upon human knowledge by purely physical processes and our con- that is not evaded by Bohm’s deterministic scious choices are epiphenomenal. dynamics. The most that “experiencers” can Bohm’s model is an example of such a ever know about the Bohmian classical world theory, which moreover agrees with the of which they are a putative part is repre- predictions of quantum theory, at least in sented by a quantum mechanical cloudlike the non-relativistic regime. It is thus instruc- wave function. tive to examine Bohm’s model and see how, This limitation in human knowledge is within that deterministic framework in acknowledged by Bohm. Indeed, Bohm’s which consciousness plays no fundamental theory leaves actual scientific practice the causal role, consciousness nevertheless same as it is in the Copenhagen approach. enters, at the level of scientific practice,in This equivalence at the practical level of just the way specified by the orthodox Bohm’s model to the Copenhagen formu- formulations. lation means that the unavoidable gap in As explained in the introductory sec- human knowledge mandated by the uncer- tion, actual scientific practice involves set- tainty principle forces a return, at the ting up experimental conditions that pro- level of scientific practice, to Copenhagen mote consciously conceived objectives. In quantum theory. The theoretically speci- von Neumann’s theory these consciously fied, but in principle unknowable and uncon- chosen actions influence the subsequent trollable information about the supposedly course of events in the observed system, deterministic microscopic realities are replaced which, according to von Neumann’s ver- in actual practice by knowable and con- sion of quantum theory, is primarily the trollable realities; namely, our human con- brain of the human participant. A key point scious choices about which actions we will is that these choices, made by the exper- take and their consciously experienced feed- imenter about how he or she will act, backs. But this means that the structure are treated in von Neumann’s theory, and added by Bohm lies beyond the scope of also by Copenhagen quantum theory, as science, in the sense that it adds nothing input data, to be fixed by the experimenter: testable. It is pure speculation. No matter what these choices actually are, An important feature of quantum the- or where they come from, or what they ory, as it is understood by physicists, is the actually do, these conscious choices are never rebutted argument of the founders treated in orthodox quantum theory as free, that any added superstructure of the general P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

896 the cambridge handbook of consciousness

kind proposed by Bohm can never add any- via the quantum Process 2 normally has thing testable, and hence lies outside sci- the effect of expanding the microscopic ence. Quantum physicists tend to be skep- uncertainties beyond what is demanded by tical of any claim that ascribes reality to the Heisenberg Uncertainty Principle: The properties that are unknowable in princi- cloud of microscopic possibilities spreads ple and give nothing testable. They identify out. This growth in the microscopic regime, science with actual scientific practice, not if unchecked by any other process, spreads untestable philosophical speculations. This into the macroscopic domain and causes attitude may be as beneficial in neuroscience even the centers of large objects to tend to as it is in atomic physics. become diffused over large regions. The disparity between this Process-2-generated theoretical indefiniteness of the locations of the The von Neumann/Stapp Approach centers of large objects and the consciously experienced definiteness of the positions of John von Neumann converted Copenhagen visible objects is resolved by Process 3. quantum theory, in a series of steps, into a Process 3 is sometimes called the Dirac form in which the entire physical universe, Choice. Dirac called it a “choice on the part including the brain of each agent, is repre- of Nature.” It can be regarded as Nature’s sented in one basic quantum state, which is answer to the question posed by Process called the state of the universe. The state of 1. This posed question might be, Will the any subsystem, such as a brain, is formed by detecting device be found to be in the averaging (tracing) this basic state over all state that signifies “Yes, a detection has variables other than those that describe the occurred”? Or, “Will the Geiger counter be state of that subsystem. The dynamics con- observed to ‘fire’ in accordance with the sists of three processes. experiential conditions that define a ‘Yes’ Process 1 is the choice on the part of the response?” Each Process 3 reply must be experimenter about how to act. This choice preceded by a Process 1 question. This is is sometimes called the Heisenberg Choice because Process 2 generates a continuous because Heisenberg strongly emphasized its infinity of possible questions that cannot crucial role in quantum dynamics. At the all be answered consistently within the pragmatic level it is a “free choice” because mathematical framework provided by it is controlled in practice by the conscious quantum theory. Process 1 specifies a set of intentions of the experimenter/participant, distinct allowed possible answers such that and neither the Copenhagen nor von Neu- the Pythagoras Rule for probabilities yields mann formulations provide any description the conclusion that the probabilities for the of the causal origins of this choice, apart from allowed possible answers sum to unity. the thoughts, ideas, and feelings of the agent. Process 1 brings the conscious choices Each intentional action involves an effort to made by the observer/participant directly produce a conceived experiential feedback, into the dynamics. On the other hand, there which, if it occurs, will be an experiential is a tendency for the effect of the Process confirmation of the success of that effort. 1 choices (of the questions) on the state of Process 2 is the quantum analog of the observed system to be washed out, in the equations of motion of classical physics. long run, by the averaging over the two pos- As in classical physics, these equations sible answers, ‘Yes’ and ‘No.’ However, it has of motion are local: All interactions are been stressed by Stapp (1999) that if will- between immediate neighbors. They are also ful effort can control the rate at which a deterministic. They are obtained from the sequence of similar Process 1 events occur classical equations by a certain quantization then the course of brain events could be procedure and are reduced to the classical strongly affected by mental effort. The tim- equations by taking the classical approxima- ing of the Process 1 events is, within the tion of setting to zero the value of Planck’s orthodox Copenhagen/von Neumann quan- constant everywhere it appears. Evolution tum theoretical framework, governed by the P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

quantum approaches to consciousness 897

choice made by the experimenter/agent, and Still later, James says, this choice is not specified by any known law of physics. But a rapid sequence of Everywhere, then, the function of effort is the same: to keep affirming and adopting pairs of questions and answers (Process 1/ 3 the thought which, if left to itself, would slip Process events) can, by virtue of the away. quantum laws themselves, hold a particular pattern of neurological activity in place, The conclusion here is that the apparent against the physical forces that would, both capacity of our conscious efforts/choices to in the absence of such pairs, and also in influence our physical actions, which seems classical physics, tend quickly to disrupt so puzzling and necessarily illusory within it. If this pattern of neurological activity classical physics, has a straightforward expla- were to be a Template for Action (i.e., nation within quantum theory. This causal an executive pattern of neurological activ- connection follows directly from the ortho- ity that tends to produce a specific action) dox quantum laws of motion. Moreover, then the prolongation of the activation of the details of how the process works is in this executive pattern of brain activity can close accord with William James’s account tend to cause the intended bodily action to of how willful effort brings about intended occur, in accordance with William James’s actions. Unlike the situation in classical “ideo-motor” theory of action (James, 1890, physics, these willful choices themselves are p. 522). (According to that theory, it is the not controlled by the known laws of physics. holding in place of the idea of an action that There is, therefore, no warrant in contem- tends to make that action happen.) porary physical theory for the assumption This fact that a sufficiently rapid sequence that our human choices are strict conse- of consciously selected probing events can quences of local mechanical processes akin hold the associated pattern of physical activ- to, or analogous to, those appearing in the ity in place longer than what would be spec- classical physics approximation. The clas- ified either by the classical laws of motion or sical approximation completely wipes out its quantum analog, Process 2,isan auto- the uncertainties within which the free matic consequence of the quantum laws of choices are allowed to act. This approx- motion. It has been extensively studied by imation contracts the spreading cloudlike quantum physicists, both empirically (Itano, structures of quantum theory into narrow Heinzen, Bollinger, & Wineland, 1990) and pencil-like beams, thus eliminating the free- theoretically (Misra & Sudarshan, 1977), dom provided by quantum theory that Bohr under the title “Quantum Zeno Effect.” strongly emphasized. In contrast to the Pro- This quantum process can provide a cess 3 choices on the part of Nature, which physics-based account of the causal efficacy are subject to statistical laws, and hence of conscious willful effort. This account cor- are forced to be “random,” the Process 1 responds closely to the ideas of William choices on the part of agents are not sub- James, as is made evident by the following ject to any known law, statistical or other- quotations: wise, and hence need not be ruled by pure chance. This is important, because it is often Thus we find that we reach the heart of claimed by the ill-informed that all of the our inquiry into volition when we ask by what process is it that the thought of any indeterminateness introduced by quantum given action comes to prevail stably in the theory is controlled by statistical laws, and 1 mind. (James, 1890,p.564) is hence random. But the crucial Process choices on the part of the agents are not sub- and later ject to any known statistical or deterministic The essential achievement of the will, in conditions. short, when it is most ‘voluntary,’ is to In quantum theory the connection bet- attend to a difficult object and hold it fast ween mental effort and physical action can before the mind....Effort of attention is be explained as a causal consequence of the thus the essential phenomenon of will. laws of quantum physics, combined with an P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

898 the cambridge handbook of consciousness

assumption that an agent’s conscious effort Pashler organizes his discussion by sep- to produce some experientially character- arating perceptual processing from post- ized effect increases the rapidity of a set of perceptual processing. The former covers Process 1 probing actions that focus attention processing that, first of all, identifies such on the intended experience. The experien- basic properties of stimuli as location, color, tial side of each such Process 1 action/event loudness, and pitch, and, secondly, identi- is specified by an intended (projected) expe- fies stimuli in terms of categories of mean- riential state. The physical side collapses ing. The post-perceptual process covers the the prior physical state of the brain to a tasks of producing motor and cognitive sum of two parts. The first part is the actions beyond mere categorical identifica- part of the prior state in which the neural tion. Pashler emphasizes (p. 33) that “the correlate (Template for Action) of the con- empirical findings of attention studies specif- scious intention is definitely present. The ically argue for a distinction between per- second part is the part of the prior state ceptual limitations and more central limita- in which the neural correlate of the con- tions involved in thought and the planning scious intention is definitely not present. In of action.” The existence of these two dif- quantum theory there are generally parts of ferent processes, with different characteris- the prior state that are not compatible with tics, is a principal theme of Pashler’s book either of those possibilities. Those parts are (pp. 33, 263, 293, 317, 404). He argues that eliminated by Process 1, which is thus associ- the former processes are carried out in paral- ated with asking a question. Process 3 gives lel, but that the latter processes, which seem Nature’s immediate answer: It collapses the to require effortful choosing, operate in state to the ‘Yes’ part or to the ‘No’ part. series and have a capacity that, although lim- These pairs of abrupt events can be regarded ited, can often be enlarged by willful effort. as the “posing by agents” and the “answering Pashler’s conclusion is based on the analy- by Nature” of specific experientially formu- sis of a huge array of recent experiments. But lated questions with ‘Yes’ or ‘No’ answers. the central finding is succinctly illustrated These events with their associated experien- in a finding dating from the 19th century; tial and physical sides are the basic building namely, that mental exertion reduces the blocks of quantum mechanics. Between such amount of physical force that a person can event-pairs the state evolves via the local apply. He notes, “This puzzling phenomena mechanical Process 2 that is analogous to the remains unexplained” (p. 387). However, if local deterministic law of motion in classical we take the sequence of Process 1 events physics. associated with an agent to have a limited This tripartite quantum dynamics involv- “capacity” in terms of events per second, ing Choice, Causation, and Chance (Pro- then this effect is a natural consequence of cesses 1, 2,&3, respectively) and the quantum theory. Creating a physical force implementation of Will (Volition) via the by muscle contraction requires a conscious conscious control of the rapidity of Pro- effort that prolongs the existence of the neu- cess 1 events provides the mathematical and ral template for action, in opposition to the logical foundation of a pragmatic quantum Process-2-generated tendency of the brain to approach to neuropsychology. How well evolve toward a more relaxed state. This pro- does this pragmatic quantum approach work longation is produced by the Quantum Zeno in actual practice? Effect, and its effect is roughly proportional to the number of bits per second of central processing capacity that is devoted to the Pashler’s Analysis task. So if part of this processing capacity A great deal of experimental work in the is directed to another task, then the applied field of the psychology of attention is sum- force will diminish. marized in Harold Pashler’s recent book of This example is just one simple case, that title (Pashler, 1998). but it illustrates the general principle. The P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

quantum approaches to consciousness 899

identification of Pashler’s limited central se- famous Libet experiments pertaining to will- rial “capacity” with the rate of occurrence of ful action (Libet, 2003; see Chapter 12). Process 1 events, assumed to be increasable The empirical fact established by the by willful effort, up to a limit, appears to Libet data is that when an action is explain the general features of all of the ‘willed’ – such as ‘willing’ a finger to rise – many diverse empirical results cited by Pash- a readiness potential (RP) appears before the ler in support of his thesis (Schwartz, Stapp, conscious experience of willing appears. The & Beauregard, 2003; Stapp, 2001). most straightforward conclusion is that the The apparent success of this quantum causal efficacy of free will is an illusion. psychophysical theory in accounting for The motion of the finger seems clearly to Pashler’s data does not mean that classical be caused by neural activity that began well physics could not be supplemented in some before the conscious act of willing occurs. ad hoc way that would enable it to match Thus, consciousness is seemingly a conse- that performance. However, the von Neu- quence of neural activity, not a cause of it. mann theory allows the data to be explained The quantum mechanical analysis of directly in terms of the already existing explic- this experiment leads to a more subtle itly described tripartite process that constitutes conclusion. the core of contemporary basic physical the- In the Libet experiment the original com- ory, whereas an explanation based on clas- mitment by the subject to, say, “raise my sical physics is predicated on the untenable finger within the next minute” will condi- idea that the classical concepts of causation tion his brain to tend to produce a sequence can be extrapolated from the motions of of potential RPs distributed over the next planets and falling apples to the motions of minute. That is, the cloud of quantum pos- ions inside nerve terminals. It rests on a the- sibilities will begin to generate a sequence of ory that is not only demonstrably false but possible RPs, each one beginning at a dif- that also claims to be dynamically and log- ferent time. Each such RP will be associ- ically complete without entailing the exis- ated with the ‘Yes’ answer to the question, tence of a part of reality that we know does “Shall I choose (make an effort) to raise my exist, namely human consciousness. In con- finger now?” If the answer is ‘No’ then the trast, von Neumann’s equations specify def- ‘template for the action of making an effort inite dynamical connections between con- to raise the finger at that moment’ will not sciousness and brain activity, and they do be actualized, and the brain state associated so in a theoretical framework that auto- with the answer ‘No’ will then evolve until matically entails all of the valid predictions the possibility of actualizing the template of classical physics. So what is the ratio- and RP corresponding to a later moment of nale, in neuropsychology, for rejecting the choice arrives. When the brain activity asso- fundamental equations and ideas of con- ciated with any one of these RPs reaches temporary physics, which can mathemat- a certain triggering condition the Process 1 ically account for the “directly observed” action associated with that particular RP will causal efficacy of consciousness and also occur. Because the original commitment is explain all of the valid classical features spread over a minute the probability, for any of phenomena, in favor of an extrapola- individual RP in this sequence, for Nature’s tion of “planetary” concepts into a micro- answer to be ‘Yes’ will be small. Hence most scopic regime where they are known to of the possible RPs up to the one correspond- fail? ing to some particular moment will not be actualized: They will be eliminated by the ‘No’ answer on the part of Nature. But for The Libet Experiment one of these Process 1 events the associated Perhaps the best way to understand the Process 3 will deliver the answer ‘Yes,’ and essence of the quantum approach to con- the associated experience E will occur. Up sciousness is to see how it applies to the to this point the conscious will has entered P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

900 the cambridge handbook of consciousness

only via the original commitment to raise the gies. For example, the subjects are trained finger at some time within the next minute. how to reduce their emotional reaction to But to be efficacious the later experi- a violent or sexual visual scene by cogni- ence E must contain an element of effort, tively re-evaluating the content; for exam- which will cause the Process 1 associated ple, by interpreting or contextualizing it with this experience (or a very similar one) in a different way. Their reactions to such to occur quickly again, and then again and stimuli are then studied using fMRI under again, thereby activating the Quantum Zeno differing choices of mental set. The brain Effect. This will cause the finger-raising tem- scans reveal profoundly different patterns of plate for action to be held in place, and the response to the stimuli according to whether effect of this will be the issuing of the neu- the subject does or does not apply the cog- ral messages to the muscles that will cause nitive re-evaluation. Without cognitive re- the finger to rise. Without this willful effort, evaluation the brain reaction is focused in which occurs in conjunction with the answer the limbic system, whereas when cognitive ‘Yes’, the sustained activation of the tem- re-evaluation is employed the focus shifts plate for action will not occur and the fin- to prefrontal regions. This demonstrates the ger will not rise. The willful effort causes powerful effect of cognitive choices upon the rapid repetition of the Process 1 action brain functioning. to occur. This holds the template in place, This effect is not surprising. Within the which causes the finger to rise. Thus the ris- pragmatic framework this effect appears as a ing of the finger is caused, in the quantum causal connection between a psychologically formulation, by the willful effort, in con- described input variable, namely a knowable cordance with the idea expressed by James and controllable free conscious choice, and (1892,p.227): an ensuing brain behavior. Quantum theory contains a mechanism that can explain this I have spoken as if our attention were empirically observed apparent causal effect wholly determined by neural conditions. of the conscious choice upon the ensuing I believe that the array of things we can brain activity, but can provide no explana- attend to is so determined. No object can catch our attention except by the neural tion of any causal effect of brain process machinery. But the amount of the attention upon our conscious choices. The contend- which an object receives after it has caught ing classical approach asserts that the causal our attention is another question. It often connections are wholly in the opposite direc- takes effort to keep mind upon it. We feel tion. It claims that both the conscious choice that we can make more or less of the effort and the subsequent brain behavior are both as we choose. If this feeling be not deceptive, consequences of prior physical processes. if our effort be a spiritual force, and an inde- Superficially, it might seem that there is terminate one, then of course it contributes no way to decide between these contending coequally with the cerebral conditions to the theories, and hence that a scientist is free result. Though it introduces no new idea, it here to choose between classical physics and will deepen and prolong the stay in consciousness of innumerable ideas which else quantum physics. It might be argued that the would fade more quickly away. dispute is all words, with no empirical or theoretical content. However, the claim that the sufficient cause of the subject’s subsequent Applications in Neuropsychology brain state does not include the controllable This theory has been applied to neu- variable that empirically controls it is prima ropsychology (Oschner, Bunge, Gross, & facie unreasonable. It would become rea- Gabrieli, 2002; Schwartz, Stapp, & Beaure- sonable only if supported by strong theo- gard, 2003). In these studies human subjects retical arguments or empirical evidence. But are first instructed how to alter their men- there is no strong theoretical support. The tal reactions to emotionally charged visual only theoretical support for this prima facie stimuli by adopting certain mental strate- implausible claim comes from a physical P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

quantum approaches to consciousness 901

theory, classical physics, whose domain of the intended experience and the other of applicability fails in principle to cover the which is the neural correlate of the nega- phenomena in question. The more accurate tion of that possibility. Contemporary phys- physical theory that should cover these phe- ical theory gives no statistical or determinis- nomena provides no support at all for an tic conditions on the choice of the intended epiphenomenal explanation of this data. action. But Nature’s feedback is required to This argument does not prove that a conform to the Pythagoras statistical rule classical-type causal explanation that leaves described above. consciousness out is rationally untenable. How is the necessary connection between However, if one were to search for evidence the experiential and physical regimes to support the classical idea then most sci- established? The answer is by trial-and- entists would probably agree that one must, error empirical testing of the correspon- at this point in time, expect the data to be dence between “the feeling of the conscious at least compatible with the predictions of effort” and “the feeling of the experien- quantum theory. But then the example pro- tial feedback.” Every healthy alert infant vided by Bohm’s model becomes instructive. is incessantly engaged in mapping out the That model is deterministic at the purely correspondences between efforts and feed- physical level, without reference to con- backs, and he or she builds up over the sciousness. To be sure, it involves, in a direct course of time a repertoire of correspon- way, instantaneous long-range action at a dis- dences between the feel of the effort and the tance, and it has not been made compatible feel of the feedback. This is possible because with the special theory of relativity. But the different effortful choices have, according model does give an idea of how conscious- to the quantum equations, different physi- ness might in principle be rendered impotent cal consequences, which produce different in a model compatible with the predictions of experiential consequences. This whole pro- quantum theory. However, as discussed in cess of learning depends crucially upon the the section on Bohm’s model, the Heisen- causal efficacy of chosen willful efforts: If berg uncertainty principle excludes the pos- efforts have no actual consequences then sibility of knowing anything about, or test- how can learning occur and the fruits of ing for the presence of, the proposed causal learning be obtained by appropriate effort? substructure. This circumstance argues for The focus here has been on the the- the adoption in neuroscience of the atti- oretical foundations of pragmatic neuro- tude adopted in atomic physics: Avoid the science. However, von Neumann’s formu- introduction of speculative concepts that are lation makes the theory more amenable to unknowable and untestable in principle, and ontological interpretation. instead treat conscious choices in the way The essential difference between quan- that they actually enter scientific practice; tum theory and classical physics, both onto- namely, as empirically knowable and con- logically construed, is that the classical trollable input parameters. state of the universe represents a purported The basic elements of von Neumann’s material reality, whereas the von Neumann theory are the experiences of conscious quantum state of the universe represents agents and the neural correlates of those a purported informational reality. This lat- experiences, the NCCs. The fundamen- ter reality has certain matter-like features. tal building blocks of quantum theory are It can be expressed in terms of micro-local “information/action events.” On the psycho- entities (local quantum fields) that evolve logical level, each Process 1 event focuses by direct interactions with their neighbors, attention and effort on an intended experi- except when certain abrupt “reductions” or ential/informational feedback. On the phys- “collapses” occur. These sudden changes are ical level this event reduces (collapses) the sometimes called quantum jumps. In ortho- state of the brain to a sum of two terms, dox pragmatic quantum theory what the one of which is the neural correlate of state represents is the collective knowledge P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

902 the cambridge handbook of consciousness

of all agents. Hence it abruptly changes work for studying the neuroscience of con- whenever the knowledge of any agent sciousness. We now turn to some efforts to changes. This behavior is concordant with tie this structure to the detailed structure of the term “our knowledge” used by the the brain. founders of quantum theory. Thus the quantum state has a form that manifests certain of the mechanical properties of matter, but a content that is basically idea-like: It rep- The Eccles-Beck Approach resents an objective kind of knowledge that changes when someone acquires knowledge. Sir John Eccles suggested in 1990, in the Pro- This radical revamping of physics was ceedings of the Royal Society (Eccles, 1990), not something lightly entered into by the that quantum theory plays a key role in the founders of quantum theory or docilely workings of the conscious brain. Based in accepted by their colleagues. A blithe disre- part on his discussions with Henry Marge- gard, in the study of the mind-brain connec- nau, Eccles noted that the statistical element tion, of this profound and profoundly rele- in quantum theory allows an escape from the vant revision by scientists of the idea of the rigid determinism of classical physics that interplay between the experiential and phys- has plagued philosophy since the time of ical aspects of nature is not easy to justify. Isaac Newton. In his later book, How the If one shifts over to an explicitly ontologi- Self Controls Its Brain, Eccles (1994) notes, cal interpretation, the question arises, What “There is of course an entrenched materialist systems besides human beings are agents? orthodoxy, both philosophic and scientific, There is currently a lack of replicable empir- that rises to defend its dogmas with a self- ical data that bear on this question, and I righteousness scarcely equaled in the ancient shall therefore not enter into philosophical days of religious dogmatism.” He says at the speculation. outset,”Following Popper (1968) I can say: I It needs to be emphasized that every- wish to confess, however, at the very begin- thing said about the von Neumann theory ning, that I am a realist: I suggest somewhat is completely compatible with there being like a na¨ıve realist that there is a physical very strong interactions between the brain world and a world of states of consciousness, and its environment. The state S(t) of the and that these two interact.” brain is what is known as the statistical oper- Eccles gives “two most weighty reasons” ator (reduced density matrix) corresponding for rejecting the classical-physics-based con- to the brain. It is formed by averaging (trac- cept of materialism (Eccles 1994,p,9). First, ing) over all non-brain degrees of freedom, classical physics does not entail the exis- and it automatically incorporates all of the tence or emergence of the defining charac- decoherence effects arising from interactions teristic of consciousness, namely “feelings,” with the environment. The key point is that and hence entails no theory of conscious- strong environmental decoherence does not ness. Second, because the nature of the map- block the Quantum Zeno Effect. ping between brain states and states of con- There is an approach to quantum theory sciousness never enters into the behavior of that tries to ignore von Neumann’s Process an organism, there is no evolutionary reason 1 (and Process 3 as well). This approach is for consciousness to be closely connected to called the “many-minds” or “many-worlds” behavior, which it clearly is. approach. No demonstration has been given Eccles’ approach to the mind-brain prob- that such a radical break with orthodox lem has three main points. The first is quantum theory is mathematically possible. that consciousness is composed of elemental Hence I do not include it among the pos- mental units called psychons and that each sible quantum approaches to consciousness psychon is associated with the activation described here. of a corresponding macroscopic physical Von Neumann’s theory provides a gen- structure in the cerebral cortex that Eccles eral physics-based psycho-physical frame- calls a dendron. It is anatomically defined and P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

quantum approaches to consciousness 903

is connected to the rest of the brain via a large is both. Until the brain process reaches the number of synapses. level of organization corresponding to the The second point is the claim that quan- occurrence of a Process 1 action, one must in tum theory enters brain dynamics in connec- principle retain all of the possibilities gen- tion with exocytosis, which is the release of erated by the Schroedinger equation, Pro- the contents of a “vesicle” – filled with neu- cess 2. In particular, one must retain both the rotransmitter – from a nerve terminal into a possibility that the ion activates the trigger, synaptic cleft. and exocytosis occurs, and also the possibil- The third point is a model developed by ity that the ion misses the trigger site, and the physicist Friedrich Beck that describes exocytosis does not occur. the quantum mechanical details of the pro- For cortical nerve terminals the observed cess of exocytosis. fraction of action potential pulses that result The first claim, that psychological pro- in exocytosis is considerably less than 100%. cesses have elemental units associated with This can be modeled classically (Fogelson dendrons, places Eccles’ theory somewhat & Zucker, 1985). But the large Heisenberg apart from the currently widespread belief uncertainty in the locations of the trigger- that the neural correlates of our conscious ing calcium ions entails that the classical experiences are spead out over larger regions uncertainties will carry over to similar quan- of the brain, which may be related more by tum uncertainties, and the two possibilities phase synchronization in time than by very at each synapse, exocytosis and no exocyto- restictive localization in space. sis, will, prior to the occurrence of the Pro- More germane to our topic is the sec- cess 3 action, both be present in the quantum ond component of Eccles’ proposal; namely, state S(t). If N such synaptic events occur that quantum effects are important in brain in the brain during some interval of time in dynamics in connection with cerebral exocy- which no Process 3 events occur, then the tosis. This conclusion is plausible and indeed state S(t) of the brain will evolve during that inescapable. Exocytosis is instigated by an interval into a form that contains (at least) action potential pulse that triggers an influx 2 N contributions, one for each of the alterna- of calcium ions through ion channels into a tive possible combinations of the exocytosis nerve terminal. These calcium ions migrate and no exocytosis options at each of the N from the ion channel exits to sites on or near synapse events. the vesicles, where they trigger the release There is a lot of parallel processing and of the contents of the vesicle into the synap- redundancy in brain dynamics, and many of tic cleft. The diameter of the ion channel these possible contributions may correspond through which the calcium ion enters the to exactly the same possible experience E. nerve terminal is very small, less than a But in real-life situations where there could nanometer, and this creates, in accordance be several different reasonable actions, one with the Heisenberg uncertainty principle, cannot expect that every one of the 2 N alter- a correspondingly large uncertainly in the native possible brain states will be a neural direction of the motion of the ion. That correlate of exactly the same possible E. If means that the quantum wave packet that the agent is conscious then the von Neu- describes the location of the ion spreads out, mann Processes 1 and 3 must enter to deter- during its travel from ion channel to trigger mine which of the various alternative possi- site, to a size much larger than the trigger ble experiences E actually occurs. site (Stapp, 1993/2004). That means that the The analysis just given assumes, in accor- issue of whether or not the calcium ion (in dance with the model of Fogelson and combination with other calcium ions) pro- Zucker, that the condition that triggers exo- duces an exocytosis is a quantum question cytosis is the presence of a specified num- basically similar to the question of whether ber of calcium ions on a trigger site. Beck or not a quantum particle passes through one and Eccles (2003) consider another possibil- or the other slit of a double-slit experiment. ity. They say that the “low exocytosis prob- According to quantum theory the answer ability per excitatory impulse . . . means that P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

904 the cambridge handbook of consciousness

there is an activation barrier against open- quantum probability rules, would in princi- ing an ion channel in the PVG (presynaptic ple allow such anomalies to occur. vesicular grid).” It is often emphasized, correctly, in con- They propose that “an incoming nerve nection with quantum approaches to brain pulse excites some electronic configuration dynamics, that “the environment” will be to a metastable level, separated energetically affected differently by interactions with the by a potential barrier V(q) from the state brain states in which an exocytosis has or has that leads to the unidirectional process of not occurred, and that this difference will exocytosis.” In this scenario the state in destroy, almost immediately, all (practically which the exocytosis does occur can be achievable) interference effects between considered to be connected by a quantum these macroscopically distinct states. tunneling process to the state where it does This environmental decoherence effect is not occur. automatically included in the formulas used Beck’s tunneling mechanism would achieve here, which refer explicitly to the brain state the same result as the mechanism, described S(t), which is the brain-state statistical oper- above, which is based simply on the spread- ator obtained by averaging (tracing) over all ing of the wave packets of the calcium ions non-brain variables. due to Heisenberg’s uncertainty principle. It is then sometimes concluded, incor- Both mechanisms lead to the result that the rectly, that one can immediately replace the brain state S(t) will contain 2 N states, defined brain state S(t) by just one of these 2 N com- by the independent exocytosis or no exocy- ponents. That conclusion might follow if one tosis option at each of the N synapses. Hence were to ignore Process 1, which is part of the Eccles-Beck model does not lead to any the brain process that defines which of our essential difference, as regards this key point, alternative possible thoughts occurs next. from the model that emphasizes the spread- Because Process 1 is part of the process that ing of the calcium ions inside the nerve ter- determines which thought occurs next, it minal. (Of course, it should not be thought should depend upon the state S(t) of the that these explicitly considered effects are brain before the thought occurs, not on the the only places where quantum effects enter part of that state that will eventually be actu- into brain dynamics. These explicitly treated alized. Hence all of the 2 N components of processes are just special cases where enough S(t) should be retained prior to the Process empirical evidence is available to make a cal- 3 collapse, whether they interfere or not: culation and where the alternative possibili- The model of the brain used above, with ties should feed into the generation of non- N identical brain states.) its 2 well defined distinct components is, of course, highly idealized. A more realis- The Eccles-Beck proposal does, however, tic model would exhibit the general smear- differ significantly from the von Neumann/ ing out of all properties that follows from Stapp proposal in regard to its third point. the quantum smearing out of the positions The von Neumann/Stapp theory attributes and velocities of all the particles. Thus the the efficacy of will to the assumed power of state S(t) prior to the collapse cannot be mental effort to increase the rate of Process expected ever to be rigorously divided, solely 1 actions, whereas the Eccles-Beck proposal by Process 2 action, including interaction attributes the efficacy of will to the assumed with the environment, into strictly orthogo- power of mental effort to modify the prob- nal non-interfering components correspond- abilities associated with the Process 3 action, ing to distinct experiences. It is Process 1 the collapse of the quantum state. that makes this crucial separation, not Pro- cess 2 . The recognition of the need to bring The von Neumann/Stapp proposal stays in a separate process to define the ques- rigorously within the framework of relativis- tion is the critical element of the Copen- tic quantum field theory and hence produces hagen approach, and it was formalized by no causal anomalies, such as the possibility von Neumann as Process 1. Any attempt of sending messages backward in time. The to leave out Process 1 faces daunting chal- Eccles-Beck proposal, by violating the basic lenges. P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

quantum approaches to consciousness 905

photons in the electromagnetic ﬁelds of, for The Jibu-Yasue Approach example, the dendritic network. They say:

The preceding sections are conservative and With the help of quantum field theory, we incomplete. They are conservative because have found that the creation and annihi- they (1) build on the orthodox philosophy lation dynamics of corticons and photons of quantum theory, which recognizes that in the QBD system in the sub-microscopic science, like every human endeavor, arises world of the brain to be the entity we call consciousness or mind. from the fact that human beings choose their actions and experience the feedbacks; and However, they have not made clear 2 they ( ) exploit the quantum laws that relate why “the creation and annihilation dynam- these choices to those feedback. ics of corticon and photons” should pos- The preceeding sections are incomplete sess the defining characteristics of conscious because they say very little about the actual processes; namely, the fact that they are brain mechanisms. “feelings.” Conscious experiences have a In regard to this second point there is quality of “feelingness” about them that is a related question of how memories are not contained in, or entailed by, the phys- 1966 1991 stored. Karl Pribram ( , ) has sug- ical concepts of corticons and photons, or gested that consciousness operates on prin- of the dynamics of these entities that they ciples similar to that of a hologram, in which claim “to be the entity we call consciousness or tiny variations of a myriad of physical vari- mind.” Thus their work does not address the ables, dispersed over a large region, combine basic question of how rationally to get the to modulate a carrier wave. These physical concepts that characterize the experiential variables might be the strengths of the syna aspects of reality out of the concepts of the ptic junctions. Pribram identifies the den- physical theory. That question is the one that dritic network (a dense set of neural fibers) was answered by the work of von Neumann, as the likely substrate of such a brain process. and that has been the primary focus of this This holographic model would appear chapter. to be implementable within quantum electrodynamics, which is the physical theory that would normally be expected to control Glossary brain dynamics. However, Umezawa and co- workers (Riccardi & Umezawa, 1967; Stuart, Quantum Jumps. The mystery of quantum Takahashi, & Umezawa, 1978, 1979)have theory is concentrated wholly in the pecu- suggested that an exotic physical process is liarities of the quantum jumps, which involved; namely, a process similar to what seem to have both subjective experi- appears in the theory of superconductively. ential aspects associated with “increases That theory is characterized by the existence in knowledge” and also objective phys- of a continuum of states of the same (low- ical aspects associated with changes in est) energy, and Umezawa has suggested that the expectations or potentialities for long-term memory is associated with break- future quantum jumps. In the chapter ing the symmetry of these ground states, “The Copenhagen Interpretation” in his instead of, for example, enduring changes in book Physics and Philosophy, Heisenberg the physical structures of nerve cells. (1958b, p. 54) says, “When the old Jibu and Yasue (1995) have attempted to adage ‘natura non facit saltus’ (nature weave these ideas of Pribram and Umezawa makes no jumps), is used as a basis into a unified quantum theory of brain for criticism of quantum theory, we can dynamics (QBD). Their theory takes the reply that certainly our knowledge can substrate associated with Umezawa’s ideas change suddenly and that this fact jus- to be the water that pervades the brain. Exci- tifies the use of the term ‘quantum tations of certain states of the water system jump.’” This explanation stresses the sub- are called corticons, and they interact with jective/experiential knowledge-increasing P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

906 the cambridge handbook of consciousness

aspect of the quantum jumps and is very part of the human agent. Its psychologi- much in line with the words of Bohr. cally described aspect is experienced and But Heisenberg then begins a further dis- described as a focusing of attention and course with the words, “If we want to effort on some intention, and the physi- know what happens in an atomic event” cally described aspect consists of the asso- and then speaks of “the transition from ciated choice of the basis vectors and of ‘possible’ to ‘actual’ that takes place as the timings of the action. Then there is soon as the interaction of the object with a feedback quantum jump whose psy- the measuring device, and hence with the chologically described aspect is experi- rest of the world, comes into play.” He says enced and described as an increment in that this transition is “is not connected to knowledge and whose physical aspect is the registration of the result in the mind of a “quantum jump” to a new physical the observer,” but that “the discontinuous state that is compatible with that incre- change in the probability function, however, ment in knowledge. Thus the two aspects takes place with the act of registration, of the conceptual foundation of science, because it is the discontinuous change of the objective/mathematical and the sub- our knowledge in the instant of registra- jective/experiential, are linked together, tion that has its image in the discontinuous within the theory, in the conceptual struc- change in the probability function.” In this ture of the quantum jump. account there are two different kinds of Planck’s Constant. This number, discovered jumps, one purely physical and occurring by Max Planck in 1900, is a number that at the device, and one purely psychologi- characterizes quantum phenomena. It cal/experiential and occurring in the mind specifies a definite connection between of the observer. The idea is that the proba- energy and frequency. The energy carried bility function is a mathematical construct by a quantum entity, such as an electron or that lives in the minds of human scien- photon (a quantum of “light”), divided by tists and represents “our knowledge” and Planck’s constant, defines a “frequency” that it consequently “jumps” when our (the number of oscillations per second) knowledge increases. There are also phys- that is associated with this entity. Thus ical jumps that occur at the devices. Planck’s constant links together a discrete But what happens in a person’s brain “lump” or “quantum” of energy with an when an abrupt increase in knowledge oscillatory motion of specified frequency. occurs? A brain is similar in many ways This constant thus forms the basis for to a measuring device: There is energy linking a property normally associated available to magnify quickly the effects with a “particle” – namely, a discrete of some small triggering event. And how amount of energy – with a property that does one explain the fantastic accuracy is associated in quantum theory with a of quantum calculations if the quantum wave motion. In classically describable mathematics that exists in our minds phenomena the products of energy and is not closely related to what is really frequency of the individual quantum happening? The von Neumann formula- entities that combine to produce the tion answers these questions by shifting macroscopic phenomena are so small on the boundary between the observed sys- the scale of observable parameters as to tem and the observer so that the phys- be individually undetectable, even though ically described state includes the brain the macroscopically measurable proper- of the observer, and the two kinds of ties of materials depend strongly upon the “jumps” described by Heisenberg become quantum properties of their constituent two aspects of a single kind of quan- elements. tum jump that occurs at the mind-brain Quantum Zeno Effect. The equations of interface. Von Neumann’s Process 1 is motion of a quantum system differ in the physical aspect of the choice on the many ways from the classical equations. P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

quantum approaches to consciousness 907

One such way is this: The motion of a or incomplete: it entails either statements planet, as described by Kepler’s empirical that can be proved to be both true and laws and by Newton’s theoretical ones, are false, or statements that cannot be proved such that the motion does not depend on to be either true or false.” how often we look at it. But the evolution Heisenberg’s Uncertainty Principle. This in time of an observed quantum system principle asserts the mathematical fact depends on Process 1, which is specified in that within the logical framework pro- the theory by observational choices made vided by quantum theory the position by the experimenter/observer/agent. In and momentum (velocity times mass) of a the case of a planet the effect of our obser- quantum entity cannot be simultaneously vations is negligible. But experiments defined to arbitrary accuracy. The product on atoms (Itano, Heinzen, Bollinger, & of the uncertainties in these two quantities Wineland, 1990) show that the behav- is given by Planck’s constant. This prin- ior of atoms can be strongly affected by ciple renders the classical concept of the the rapidity at which Process 1 events deterministic motions of the planets in the are occurring. If these events occur suf- solar system inapplicable to the motions of ficiently rapidly then the normal behavior the ions in a human brain. of the atom, such as a transition from an excited state to a state of lower energy, is slowed down. This was predicted by quantum theory, and the empirical results are References in good agreement with theoretical predictions. This slowing effect is called the Beck, F., & Eccles, J. (2003). Quantum processes Quantum Zeno Effect (Misra & Sudar- in the brain: A scientific basis of consciousness. shan, 1977). If the observed system is the In N. Osaka (Ed.), Neural basis of consciousness 141 166 brain of an agent then, as a consequence of (pp. – ). Amsterdam: Benjamins. evolutionary and educational processing, Bohm, D. (1952). A suggested interpretation of this brain could be expected to become quantum theory in terms of hidden variables. 166 179 very sensitive to variations in the charac- Physical Review, 85, – . 1986 ter of the Process 1 events. This is because Bohm, D. J. ( ). A new theory of the relation- Templates for Action could be held in ship of mind to matter. Journal of the Amer- ican Society for Psychical Research, 80, 113– place longer than normal by “free” choices 135. made by agents, and this could strongly Bohm, D. J. (1990). A new theory of the relation- influence behavior. Such an effect would ship of mind to matter. Philosophical Psychology, elevate the choices made by human agents 3, 271–286. from their status in classical physics as Bohm, D., & Hiley, D. J. (1993). The undivided mechanically determined epiphenomenal universe. London: Routledge. side effects to causes that, in principle, are Bohr, N. (1958). Atomic physics and human knowl- not fully traceable within contemporary edge. New York: Wiley. physical theory to purely physical pro- Bohr, N. (1963). Essays 1958/1962 on atomic cesses, and hence are properly treatable as physics and human knowledge. New York: empirical inputs. Wiley. Template for Action. This is an “executive” Eccles, J. C. (1990). A unitary hypothesis of mind- patterns of neurological activity that if brain interaction in the cerebral cortex. Pro- held in place for an extended period will ceedings of the Royal Society of London, 240, issue the sequence of neural signals that 433–451. will initiate and monitor a particular phys- Eccles, J. C. (1994). How the self controls its brain. ical (or perhaps mental) action. Berlin: Springer. Godel’s¨ Incompleteness Theorem. “Any Fogelson, A., & Zucker, R., (1985). Presynaptic finite set of rules that encompass the calcium diffusion from various arrays of single rules of arithmetic is either inconsistent channels. Biophysical Journal, 48, 1003–1017. P1: KAE 0521857430c31 CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 27, 2007 19:28

908 the cambridge handbook of consciousness

Hagen, S., Hameroff, S., & Tuszynski, J. (2002). Penrose, R. (1994). Shadows of the mind. New Quantum computation in brain microtubules: York: Oxford University Press. Decoherence and biological feasibility. Physical Pribram, K. H. (1966). Some dimensions of Review, E65, 061901–1 – 061901–11. remembering: Steps towards a neurophysio- Hameroff, S., & Penrose, R. (1996). Orches- logical theory of memory. In J. Gaito (Ed.), trated reduction of quantum coherence in Macromolecules and behavior (pp. 165–187). brain microtubules: A model for conscious- New York: Academic Press. ness. Journal of Consciousness Studies, 3, 36– Pribram, K. H. (1991). Brain and perception. Hills- 53. dale, NJ: Erlbaum. Heisenberg, W. (1958a). The representation of Putnam, H. (1994, November 20). Review of Nature in contemporary physics. Daedalus, 87, Roger Penrose, Shadows of the Mind. New 95–108. York Times Book Review,p.7. Retrieved from Heisenberg, W. (1958b). Physics and philosophy. www.ams.org/journals/bull/pre-1996data/ New York: Harper, 199507/199507015.tex.html. 1967 Itano, A., Heinzen, D., Bollinger, J., & Wineland, Riccardi, L. M., & Umezawa, H. ( ). Brain and D. (1990). Quantum Zeno effect. Physical physics of many-body problems. Kybernetik, 4, 44 48 Review, A41, 2295–2300. – . Schwartz, J., Stapp, H., & Beauregard, M. (2003). James, W. (1890). The principles of psychology, Vol. The volitional inﬂuence of the mind on the II. New York: Dover. brain, with special reference to emotional self 1892 James, W. ( ). Psychology: The briefer course.In regulation. In M. Beauregard (Ed.), Conscious- William James: Writings 1879–1899. New York: ness, emotional self-regulation and the brain. Library of America. Amsterdam: Benjamins. 1995 Jibu, M., & Yasue, K. ( ). Quantum brain Stapp, H. (1999). Attention, intention, and will dynamics and consciousness. Amsterdam: Ben- in quantum physics. Journal of Consciousness jamins. Studies, 6, 143–164. 1985 Libet, B. ( ). Unconscious cerebral initiative Stapp, H. (2001). Quantum theory and the role and the role of conscious will in voluntary of mind in Nature. Foundations of Physics, 31, 529 action. Behavioral and Brain Sciences, 8, – 1465–1499. 566. Stapp, H. (1993/2004). Mind, matter, and quan- Libet, B. (2003). Cerebral physiology of con- tum mechanics. Heidelberg, Berlin, New York: scious experience: Experimental studies. In Springer. Sections 5.7 and 6.4. N. Osaka (Ed.), Neural basis of consciousness. Stuart, C. I. J. M., Takahashi, Y., & Umezawa, H. Amsterdam: Benjamins. (1978). On the stability and nonlocal properties Misra, B., & Sudarshan, E. C. G. (1977). of memory. Journal of Theoretical Biology, 71, The Zeno’s paradox in quantum theory. 605–618. 756 Journal of Mathematical Physics, 18, – Stuart, C. I. J. M., Takahashi, Y., & Umezawa, H. 763 . (1979). Mixed-system brain dynamics: Neural Ochsner, K. N., Bunge, S. J., Gross, J. J., & memory as a macroscopic ordered state. Foun- Gabrieli, J. D. (2002). Rethinking feelings: An dations of Physics, 9, 301–327. fMRI study of the cognitive regulation of emo- Tegmark, M. (2000). Importance of quantum 1215 tion. Journal of Cognitive Neuroscience, 14, – decoherence in brain process. Physical Review, 1229 . E61, 4194–4206. 1998 Pashler, H. ( ). The psychology of attention. von Neumann, J. (1955). Mathematical founda- Cambridge, MA: MIT Press. tions of quantum mechanics (R. T.Rever, Trans.). Penrose, R. (1986). The emperor’s new mind. New Princeton: Princeton University Press. (Origi- York: Oxford University Press. nal work published 1932.) P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

Author Index

Abarbanel, H. D., 731, 734 Akhtar, N., 619 Abe, K., 790 Akiguchi, I., 790 Abeles, M., 746 Alavi, A., 541 Abelson, R. P., 558 Albe-Fessard, D., 788 Åberg, C., 811 Albert, M. S., 231, 812, 815 Aberle, D., 648 Albright, T. D., 793 Abramowitz, A., 692 Alcock, K. J., 587 Abrams, M. H., 396 Alexander, M., 240 Abrams, R. L., 216, 218, 219, 220, 234 Alexander, R. D., 605, 850 Achilli, A., 584 Alkire, M. T., 711 Ackerman, B. P., 295, 296 Allan, K., 820 Ackerman, S. J., 681 Allison, A. C., 711 Adams, C., 738 Allison, T., 228, 230, 871 Adolphs, R. A., 229, 261, 871, 872, 873, 874 Allman, J. M., 793, 870 Adovasio, J. S., 584 Allport, A., 239, 776, 839 Adrian, E. D., 775, 783 Almagor, M., 689 Adriance, W., 535 Alpert, L. H., 454 Aerts, J., 709, 710 Alpert, N. M., 231, 450, 465, 466, 812, 815 Aertsen, A., 736 Amador, X., 489, 490, 493 Aftanas, L. I., 534 Ambady, N., 871 Agenta, B., 607 Ambrose, S. H., 583 Aggarwal, R., 681 Aminoff, E., 819 Aggelopoulos, N. C., 853 Amirkhanov, N. D., 584 Aglioti, S., 334 Amit, D. J., 737 Agnati, L. F., 780 Amsterdam, B., 420, 421 Agnetta, B., 578 Anand, B., 536, 537 Ahissar, M., 718 Anand, K. J., 406, 417 Ahlfors, S. P., 536 Anand, S., 333 Aicardi, J., 789 Andermann, F., 789 Aiello, L. C., 388, 605 Anderson, C. A., 417, 562 Aikhenvald, A. Y., 370 Anderson, J. R., 152, 153, 154, 155, 461, 466–467, 648 Aitken, K. J., 406 Anderson, M. C., 312 909 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

910 author index

Anderson, S. J., 260, 265 Avihour, N., 686 Anderson, S. W., 876 Avikainen, S., 874 Andrade, M. C., 618 Azzopardi, P., 208, 238, 242 Andresen, J., 533 Anes, M. D., 817, 821 Baars, B. J., 36, 56, 118, 124, 152, 155, 158, 162, 166, 168, Anis, N. A., 711 169, 182, 186, 192, 194, 195, 196, 201, 229, 633, 753, Anokhin, A. P., 427 776, 781, 792, 855 Ansay, C., 297, 308–309 Bachmann, T., 723 Anscombe, G. E., 60 Backman, S. B., 711 Anthony, J. L., 361 Bacon, E., 310, 484, 494 Antognini, J. F., 712 Baddeley, A. D., 187, 188, 393, 484, 563 Antoine, S., 715, 717 Bagchi, B. K., 533, 537 Anton, H. P., 584 Bahan, B., 587 Anwar, A., 164 Baillargeon, R., 618 Aosaki, T., 795 Baime, M., 541 Apanovitch, A. M., 676, 690, 694 Baird, A. A., 420 Apple, J., Baird, J., 290 Aquili, E., 541 Bakan, P., 463 Arad, D., 686 Baker, L., 681 Arbib, M. A., 388, 392, 563, 587, 734, 735 Bakhtin, M., 396 Arenander, A., 535 Bal, M., 376 Ariela, A., 736 Baldeweg, T., 588 Arkowitz, J., 690, 692 Baldwin, D. A., 617 Arlow, J. A., 680 Baldwin, J. M., 407, 408, 414 Armony, J. L., 229, 872 Baldwin, M. W., 685 Armstrong, D. F., 587 Baldwin, P., 441, 709, 710 Armstrong, D. G., 587 Balkin, T. J., 441, 709 Armstrong, D. M., 47, 57, 414, 418, 846 Balleine, B., 383 Armstrong, K. M., 722 Balota, D. A., 821 Arnold, M. B., 461 Balthazard, C., 155 Aron, L., 681 Banaji, M. R., 208 Artiges, E., 484 Bandura, A., 296, 557 Arzimanoglou, A. A., 789 Bandy, D., 818 As, A., 446 Banister, H., 467 Asch, S. E., 556 Banks, C., 556 Aschenbrenner-Scheibe, R., 533 Banquet, J. P., 537 Aschersleben, G., 329, 330, 331, 332, 339, 340, 341, Banyas, C. A., 614 560 Bar, M., 215, 335, 783, 819 Aserinsky, E., 448 Bar, S. K., 302, 304 Asfaw, B., 583, 585 Barber, T. X., 447, 449, 452, 454, 459 Ashburner, J., 522, 587 Barbour, J. S., 457 Ashby, R., 734 Barbur, J. L., 720 Ashwal, S., 715 Bard, K. A., 607 Aslin, R. N., 522 Bargh, J. A., 315, 556, 557, 558, 560, 561, 562, 563, Assal, G., 790 565 Asselin, M. C., 796 Baribeau, J., 712 Astington, J. W., 384, 391, 423, 424 Barinaga, M., 530 Aston-Jones, G., 439–440 Barkin, A., 524 Atance, C. M., 577 Barkow, J. H., 590, 599, 600, 609, 653 Athwal, B. S., 451 Barlngay, S. S., 113 Atkin, A., 718 Barlow, H. B., 842 Atkinson, G., 446 Barndollar, K., 562 Atkinson, R. C., 187, 259, 563 Barnes, A. E., 292, 308–309 Atran, S., 622 Barnes, J., 13, 465 Auer, D. P., 710 Barnhardt, T. M., 451, 455 Auerbach, E., 377, 394 Barnier, A. J., 452, 453, 462, 463, 464 Auerbach, J. S., 681 Barnouw, V., 648 Austin, G. M., 789 Baron-Cohen, S., 611, 870 Austin, J. H., 500, 536, 539, 540 Barresi, J., 393, 408 Avellino, A. M., 789 Barrett, H. C., 634 Avidan, S., 686 Barrett, L., 590 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

author index 911

Barsalou, L. W., 163, 397 Berke, R. L., 686 Bartels, A., 543, 781, 794 Berkeley, G., 23 Barth, E., 724, 794 Berker, E., 778, 782, 789 Bartlett, F. C., 367, 641 Berkowitz, L., 562 Basmajian, J., 182 Berman, T. R., 396 Bassetti, C., 710, 790 Bernard, L. L., 583 Bastian, A., 635 Bernat, E., 229, 233 Bates, B. L., 446 Bernet, R., 74 Bates, E., 582, 619 Berntson, G. G., 871 Battersby, W. S., 237 Berridge, K. C., 293 Bauer, L. O., 427 Berry, D. C., 156, 453 Bauer, R. M., 455 Berry, J. W., 680 Baulac, M., 738 Berry, S. C., 711 Baumeister, R. F., 557, 558, 618, 682 Berti, A., 239, 783 Baumgartner, C., 790 Bertrand, L. D., 452 Bayer, M., 306 Bertrand, O., 750, 753 Baynes, K., 789 Berukhim, E., 55 Beach, T. G., 796 Besner, D., 216 Beattie, B., 715, 794 Bessenoff, G. R., 262 Beaulieu, M. A., 789 Besson, J. M., 788 Beauregard, M., 899, 900 Bethel, W. M., 618 Bechara, A., 876 Betsch, T., 306 Beck, B. B., 577 Bettelheim, B., 679 Beck, D. M., 720, 722 Bewermeyer, H., 790 Becker, D. E., 530, 532, 537 Beyenne, Y., 585 Becker, E., 615 Beyer, S., 505 Beechner, J. C., 777 Bharati, A., 648 Begg, I., 296, 298, 301 Bhattacharya, R. S., 112 Begleiter, H., 427 Bibb, B. C., 459 Behe, M., 25 Bibbig, A., 532 Behne, T., 388 Biblarz, A., 618 Behrendt, R. P., 494 Bickerton, D., 573, 581, 583, 585, 590 Behrens, M. M., 790 Biederman, I., 215, 783 Bejamin, A. S., 295 Bieri, P., 424 Bekkering, H., 874 Bierschwale, D. T., 608 Belenky, G., 441, 709, 710 Bilik, M., 484, 490 Bell, M. A., 420 Binkofski, F., 563, 874 Bellezza, F. S., 255 Birch, C. S., 820 Bellis, R. P., 853 Birnbaum, G., 685, 686 Bello, J. A., 790 Birnbaumer, N., 427 Ben-Porath, Y. S., 689 Bjork, R. A., 291, 295, 298, 301, 304, 305, 466–467 Ben-Zur, H., 293, 304 Bjorklund, D. F., 291, 604, 605, 606, 607, 609, 612 Bender, M. B., 237 Bjorkman,¨ M., 306 Benjamin, A. S., 291, 298, 301, 305 Black, S. E., 717 Bennett, J., 575 Blagov, P., 690, 691, 692 Bennett, K., 334 Blake, R. A., 532, 540, 713, 759, 872 Bennett, S. C., 753 Blakemore, S. J., 339, 343, 344, 345, 560, 564, 873 Benotsch, E. G., 454 Blanc, A., 790 Benson, D. F., 782 Blanck, G., 680 Benson, D. M., 464 Blanck, R., 680 Benson, H., 525, 542 Blandin, E., 216 Berabum, K. S., 454 Blank, A., 558 Beran, M. J., 780 Blasberg, R., 794 Berbaum, M. L., 454 Blatt, S. J., 681 Berendse, H. W., 787 Blaxton, T. A., 253, 267, 268, 271, 272, 273 Berger, H., 531, 791 Bleske, A. L., 602 Berger, R. J., 535 Bless, H., 289, 296, 297, 314 Bergman, T. J., 777 Block, E., 217, 219 Berhe, S. M., 584 Block, N., 43, 45, 54, 55, 58, 61, 136, 202, 342, 707, Bering, J. M., 577, 578, 580, 606, 607, 608, 609, 612, 724, 756, 758, 776, 811–812, 822, 839 615, 617, 622 Blomﬁeld, L., 356 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

912 author index

Bloom, F. E., 439–440 Brandt, J. P., 790 Bloom, P., 12, 588, 619, 851 Brass, M., 874 Blum, G. S., 457, 458, 678 Bratslavsky, E., 558 Blumberg, S. J., 688 Braud, W., 648 Blumbo, C., 720 Braun, A. R., 441, 709, 710 Blumenthal, A. L., 194 Brazdil, M., 233 Blumenthal, J., 414 Brazier, M. A., 439 Boas, F., 636 Brebion,´ G., 489, 490, 493 Boccia, M. L., 607 Brecht, M., 751, 753, 754, 779 Boddy, M., 135, 139 Brefczynski-Lewis, J. A., 542 Bodhi, B., 95, 99, 113 Breinlinger, K., 618 Bodis-Wollner, I., 718 Breiter, H. C., 228, 229 Boehm, C., 607 Brekke, N., 561 Boesch, C., 606, 607, 612, 622 Bremer, F., 775, 783, 791 Bogen, G. M., 789 Bremmer, J. D., 613 Bogen, J. E., 712, 776, 778, 780, 781, 782, 789, 795, Brenman, M., 447 796, 797 Brenneman, H. A., 451 Bogin, B., 605 Brenner, C., 680, 690, 818, 821 Bogousslavsky, J., 790 Brentano, Franz, 26, 50, 414 Bohm, D. J., 894, 895 Bressler, S. L., 737, 740, 741, 747, 750 Bohr, N., 883, 886, 887 Brewer, M. B., 557, 560 Bois, J., 363 Bridgeman, B., 333, 335 Bolam, J. P., 795 Brinkmann, B., 306 Bollinger, J., 897, 907 Briquet, P., 450 Bolton, A., 256, 269 Briscoe, S., 453 Boly, M., 715, 717 Brisson, M. A., 456 Bonhomme, V., 711, 712 Broad, C. D., 30 Bonke, B., 714 Broadbent, D. E., 156, 716 Bonnebakker, A. E., 714 Broca, Paul, 26 Bonner, J. T., 605 Brockmeier, J., 376, 377 Bonus, K., 521, 522, 524 Brodal, A., 791 Booker, J., 274 Bromﬁeld, E. B., 533 Bookheimer, S. Y., 816 Bromley, S., 788 Booth, D. A., 848 Bronkhorst, J., 504 Booth, W. C., 376 Brook, A., 61 Born, A. P., 711 Brooks, A. S., 584 Born, J., 710 Brooks, L. R., 297, 313 Bornstein, R. F., 682–683 Brooks-Gunn, J., 420, 606 Boseovski, J., 423 Broughton, R., 709 Bosinelli, M., 709 Browman, C., 588 Botero, J., 84 Brown, A. L., 290, 291, 295, 297 Botscharow, L., 633 Brown, A. S., 360, 366, 467 Boucouvalas, M., 648 Brown, C., 232 Bourdieu, P., 644 Brown, D., 460 Bourgeois, P., 454 Brown, F., 582 Bourguignon, E., 648 Brown, G. M., 465, 813 Bower, G., 154, 155, 161 Brown, K. W., 427 Bowers, J. S., 273–274 Brown, M., 216, 454 Bowers, K. S., 447, 455, 460, 461, 462, 464, Brown, R., 261, 270, 453, 750 465 Brown, R. J., 720, 761 Bowers, P., 449 Brown, R. M., 618 Bowlby, J., 681, 685 Bruant, A., 494 Boyer, P., 622 Brueckner, A., 55 Bradley, F., 24 Bruggemann, J. H., 584 Bradley, R., 682 Bruner, E. M., 633 Brady, I. A., 652 Bruner, J. S., 376, 396, 561, 641 Brady, J. P., 456 Bruyer, R., 718 Braid, J., 466 Bryans, B., 782 Brainerd, C. J., 297, 310 Bryant, R. A., 452, 453, 456, 457, 458, 459 Bramham, J., 840 Buccino, G., 341, 563, 874 Brammer, M. J., 720 Buchanan, T. W., 261, 873 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

author index 913

Bucher, D., 780 Canessa, N., 341 Buchser, E., 454 Cant, J. G., 609 Buckner, R. L., 228, 229, 231, 709, 811, 813, 814, 815, Cantor, N., 557 817, 819, 820 Caplan, L., 790 Budson, A. E., 821 Caramazza, A., 299, 366 Buffalo, E. A., 262 Carbaugh, D., 376 Bufﬂer, R. T., 584 Carey, D. P., 236, 336, 563 Bullemer, P., 158 Carey, S., 579 Bullmore, E. T., 465 Cargile, J., 575 Bullock, P. R., 840 Carlsen, V., 456 Bunce, S., 229 Carlson, S., 423, 425 Bundlie, S. R., 708 Carlston, D. E., 561 Bunge, J., 900 Carpenter, C. J., 489, 490, 493 Bunge, S. A., 412 Carpenter, M., 388, 607, 619 Bunin, M. A., 780 Carpenter, W. B., 775 Buonocore, M. H., 712, 822 Carr, T. H., 240 Burckhardt, J., 396 Carrasco, M., 722 Burgess, A. P., 536 Carrier, B., 465 Burgess, J. A., 406, 407 Carroll, D., 305 Burgess, N., 522, 533 Carroll, M., 296 Burgess, P., 849 Carruthers, P., 49, 60, 61, 407, 419, 841, 855 Burghy, C. A., 522 Carruthers, P. K., 406 Burglen, F., 485 Carsen, R. E., 441, 710 Burke, D. M., 295, 330 Carson, B. S., 789 Burke, R. E., 533 Carson, R. E., 709 Burklund, C. W., 789 Carstens, E., 712 Burley, N., 614 Carter, C. S., 871 Burns, G. A., 746, 747 Carter, O. L., 528 Burrows, L., 562 Cartwright, J., 609 Burton, N. R., 711 Carver, C. S., 557, 688, 694 Burton, P., 216, 219 Cassidy, J., 681, 694 Busey, T. A., 302 Castaigne, P., 790 Bush, G., 542 Castaneda,˜ H., 60 Bushman, B. J., 562 Castellanos, F. X., 414 Bushnell, M. C., 450, 463, 464, 465 Castiello, U., 334, 783 Busnel, M. C., 414 Castles, A., 361 Buss, D. M., 599, 600, 602, 609, 619 Caston, V., 15, 50, 61 Buss, R., 153–154, 158 Cavanaugh, P., 792 Bussey, T. J., 579 Cermack, L. S., 810 Buswell, G. T., 359 Cezayirli, E., 818 Butterworth, G., 617 Chabris, C. F., 213 Bygott, J. D., 617 Chafe, W., 359, 365, 367, 369, 370, 371 Byrne, A., 49, 58 Chaiken, S., 158, 313, 557, 558, 559, 677 Byrne, R. W., 578, 605, 608, 870 Challis, B. H., 272, 273 Chalmers, D. J., 36, 39, 40, 57, 59, 118, 123, 598, 744, Cabanis, P., 23 758, 776, 839, 841, 864, 868 Cabeza, R., 813, 819 Chandler, C. C., 302 Cacioppo, J. T., 558, 677, 871 Chandrasekhar, S., 734 Cadusch, P. J., 534 Chang, T., 533, 732 Cahan, E., 407 Changeux, J. P., 199, 200, 203, 406, 651, 744, 745 Cairns, H., 776 Chanowitz, B., 558 Cairns-Smith, A. G., 776, 780 Chapman-Waldrop, A., 523 Calder, A. J., 228, 229, 872 Charman, T., 619 Call, J., 388, 578, 606, 607, 608, 870 Charney, D. S., 613 Calverley, D. S., 452 Chartrand, T. L., 557, 561, 562, 563, 565 Calvert, G. A., 588 Chauvet, J., 391 Campbell, B., 789 Chaves, J. F., 448 Campbell, D. T., 449, 633 Chavez, M., 533 Campbell, J., 128 Chawluk, J., 790 Campbell, R., 588, 648 Cheeseman, J., 216 Campion, J., 238 Chekaluk, E., 335 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

914 author index

Chen, M., 562 Coleman, J. W., 508 Chen, Z., 791 Coles, M. G., 217 Cheney, D. L., 777, 841 Collins, A., 153 Cherry, E. C., 209 Collinson, S., 451 Chiang, C., 533 Colon, M. R., 579 Chiefﬁ, S., 334 Conklin, H., 643 Childs, N., 715 Connelly, A., 587, 588 China, G., 536 Connor, K. J., 873 Chiu, C. Y., 269 Contreras, D., 201, 527, 742, 788, 793 Chomsky, N., 366, 573, 575, 576, 581, 586 Conway, M. A., 256, 260, 262, 265, 270, 487, 493 Chrisley, R., 128, 134, 153, 160 Cook, N. D., 789 Christianson, S., 261 Cooke, R. W., 417 Christopher, S. B., 851 Cooney, J. W., 202 Chrosniak, L. D., 274 Cooper, G., 874 Chuengsatiansup, K., 657 Cooper, J., 466–467 Chugani, H. T., 420 Cooper, L. A., 820 Chun, M. M., 228, 230, 721, 722 Cooper, L. M., 451, 453 Churchland, P. M., 14, 38, 55 Cooper, N. R., 531, 536 Churchland, P. S., 132, 378, 776, 777 Cooper, W., 28 Cicchetti, F., 796 Copeland, S. M., 277 Clancy, W. J., 145 Corballis, M. C., 577, 587, 588, 589 Claparede,` E., 383, 717 Corballis, P. M., 819 Clarac, F., 328, 329, 343 Corbetta, M., 709, 722, 822 Clare, L., 484 Corby, J. C., Clarisse, J., 787 Corkin, S., 261, 717 Clark, A., 157, 867, 868 Cornelissen, E., 584 Clark, D. D., 529 Cornoldi, C., 296, 304, 308 Clark, J. J., 867 Corr, P. J., 694 Clark, R. E., 269, 797, 810 Corteen, R. S., 210 Clark, S., 330, 331, 332, 340 Corwin, J., 272 Clarke K., 718 Cosmelli, D., 71, 527, 748, 761, 762 Clarke, D., 584 Cosmides, L., 158, 583, 590, 599, 600, 609, 653 Clarke, J. D., 585 Costantini-Ferrando, M. F., 450, 465, 466 Clarke, K., 228, 229, 230 Costermans, J., 297, 308–309 Classen, C. C., 633, 657, 681 Cotterill, R. M., 170, 598, 609 Clayton, N. S., 579 Coultheart, M., 361 Clech, G., 217, 219, 231 Count, E. W., 652 Cleeremans, A., 170, 407 Courage, M., 421 Cleghorn, J. M., 465 Cove, J., 648 Clemence, A. J., 681 Coward, L. A., 152, 160, 170 Clemenceau, S., 738 Cowey, A., 208, 237, 238, 242, 338, 718, 723, 793 Clifford, E., 213 Cowie, R. J., 787 Clifford, W., 25 Cozort, D., 506 Cloninger, C. R., 689 Craig, A. D., 543 Clore, G. L., 293, 296, 315 Craighero, L., 740 Cobb, P. C., 461 Craik, F. I. M., 259, 276, 292, 295, 296, 410, 811, 812, Coben, L. A., 791 813 Coe, W. C., 447, 461 Cramon, D. Y., 747, 790 Cohen, A., 241 Crawford, H. J., 455, 463, 464 Cohen, G. M., 260, 265 Crelin, E. S., 586 Cohen, J. D., 563, 685 Crick, F. C., 36, 152, 169, 718, 743, 744, 750, 757, 758, Cohen, L., 230, 231 776, 792, 853 Cohen, L. B., 414 Crisi, G., 790 Cohen, L. H., 685 Croft, R. J., 536 Cohen, N. J., 215, 227, 234, 235, 717, 797 Cronsniak, L. D., 272 Cohen, R. L., 294, 304 Crook, J. M., 601 Cohen, S., 524 Cross, D., 409, 419, 611, 620 Cohn, D., 369 Crowley, C., 686 Cole, M., 637, 638, 653 Crutchﬁeld, J. P., 736 Colebatch, J. G., 330 Csikszentmihalyi, M., 446 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

author index 915

Csordas, T. J., 657, 658, 659 Dawkins, M. S., 838, 851 Cubelli, R., 240, 241 Dawkins, R., 25, 125 Cuervo, C., 494 Daza, M. T., 216 Cummins, R., 62 De Catanazaro, D., 618 Cunningham, P. V., 457 De La Mettrie, J., 22 Cunningham, W. A., 873 De la Vallee´ Poussin, L., 113 Curran, T., 810, 817, 818, 819 De Marcas, S., 299 Currie, K., 118 De Renzi, E., 790 Curro,´ D., 793 De Rios, D., 648 Curtis, G. H., 584 De Waal, F., 389, 390, 578, 604, 613, 614, 615 Custance, D. M., 607 De Weerd, P., 722 Czisch, M., 710 Deacon, D., 232 Deacon, T., 581 D’Amato, T., 484 Deak,´ G. O., 426 D’Andrade, R., 633, 634, 649, 650 Deakin, J., 693 D’Aquili, E. G., 633, 634, 652 Dean, T., 135, 139 D’Errico, F., 584 DeBarbara, K., 425 D’Esposito, M., 814 DeBellis, M., 45 D’Holbach, P-H, 23 Debner, J., 224 D’Mello, D., 165 DeCasper, A., 414 Daffner, K. R., 821 Decety, J., 563, 873, 874 Dagenbach, D., 240 Deco, G., 838, 839, 843, 853, 855 Dahaene, S., 196, 198, 199 DeCorte, E., 291 Dalai Lama XIV, 505, 506, 508, 518, 519 Deese, J., 263, 277, 818 Dale, A. M., 813, 815, 819 Defoe, D., 396 Dalery, J., 484 DeGroot, H. P., 459 Dallas, M., 253, 259, 272 DeGuelder, C., 450, 454, 464, 465 Daloze, T., 711 Degueldre, G., 709, 710 Dalton, K. M., 524 DeGusta, D., 583, 585 Daly, M., 599, 615, 621 Dehaene, S., 199, 200, 201, 202, 203, 216, 217, 218, 219, Damaser, E., 462 226, 230, 231, 328, 342, 406, 484, 527, 744, 745, Damasio, A. R., 76, 152, 169, 201, 301, 406, 526, 543, 759 620, 740, 757, 762, 776, 782, 796, 855, 872, 873, Dehaene-Lambertz, G., 217, 231 874, 876 Dehlen, E., 618 Damasio, H., 229, 776, 777, 782, 872, 874, 876 Deikman, A. J., 537 Damian, M., 216, 219 Delafosse, A., 787 Danckert, J., 759 Delaloye, B., 790 Dancy, J., 81 Delaloye-Bischof, A., 790 Daneman, M., 714 DeLeonardis, D. M., 262 Daniel, P., 233 DelFiore, G., 450, 454, 464, 465, 709, 710 Daniel, V., 656 Dell’Acqua, R., 216 Danion, J. M., 310, 484, 485, 487, 490, 494 Della Sala, S., 343 Daprati, E., 334, 339, 345 Delmonte, M. M., 500, 533, 534, 535, 536 Dardenne, B., 291, 454 Delorme, F., 708 Darlington, R., 603, 604, 606 Dement, W. C., 438 Darlley, J. M., 556 Demiralp, T., 233 Darwin, C., 24, 25 Denes-Raj, V., 301 Das, N. N., 530, 533, 537, 538 Denhiere, G., 308 Davachi, L., 815 Dennett, D. C., 38, 56, 60, 62, 124, 127–128, 134, 137, David, A. S., 453, 465 201, 202, 203, 330, 378, 379, 380, 382, 573, 618, David, D., 453 777, 782, 789, 841 David, O., 527, 748, 761 Dennis, M., 414 Davidson, D., 83, 144, 147, 865 Depraz, N., 71, 763 Davidson, I., 582 Descartes, R., 10, 17, 18, 378, 556, 573, 796 Davidson, J. E., 290 Deschamb, A., 456 Davidson, R. J., 521, 522, 524, 530, 531, 533, 538, 542, Deschamps, E., 391 543, 688, 694, 871 Desi, E. L., 557 Davis, D. R., 617 Desikan, R., 821 Davison, M. C., 572 Desimone, R., 533, 537, 722, 751, 779 Dawes, R. M., 693 Desjarlais, R., 657, 658 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

916 author index

Desmond, A., 27 Dubreuil, D. L., 452, 457 Desmond, J. E., 811, 813, 820 Dubrovinskaya, N. V., 427 Desmurget, M., 334 Duchek, J. M., 821 DeSouza, J. F., 334 Dudai, Y., 780, 797 Destexhe, A., 788 Dufek, M., 233 Deutsch, J. A, 208 Duft, S., 296, 298, 301 Deutsch, R., 313 DuHamel, K. N., 454 Devine, P. G., 557, 560 Dulany, D., 209, 237 Devrim, M., 233 Dumais, S. T., 565 Dewey, J., 414, 868 Dumas, R. A., 463 Dewhurst, S. A., 262, 487 Dunbar, K., 563 Diamond, J. M., 590, 591, 617 Dunbar, R. I., 388, 392, 590, 605, 609, 616, 870 Dickinson, A., 383, 579 Duncan, C., 790 Dickman, S., 693 Duncan, G. H., 450, 465 Dienes, Z., 239, 407, 418, 419, 453 Duncan, J., 722 DiGirolamo, G., 564 Dunlosky, J., 290, 292, 294, 295, 296, 298, 303, 304, Dijksterhuis, A., 560, 562 308–309, 312 Dikkes, P., 790, 792, 794 Dunn, D., 210 DiLillo, D. K., 681 Dunn, J. C., 256, 810, 818 Dinner, D. S., 789 Dunning, D., 296, 309, 690 Dipietro, M., 417 Dunton, B. C., 560 Disbrow, E. A., 712 Duong, D. H., 732 Dissanayake, E., 633 Duranti, A., 633 Ditto, P. H., 676, 690, 694 Durkeim, E., 653 Ditto, W. L., 732 Duvernoy, H., 790 Dixon, M., 266–267, 270 Duzel,¨ E., 817 Dixon, N. F., 683 Dywan, J., 460 Dixon, R. M., 370 Dobbins, I. G., 269, 810, 814, 815, 817, 820 Eagleman, D. M., 331, 340 Dobbins, S., 717 Eagly, A. H., 559, 677 Dolan, P. O., 263 Eastwood, J., 209, 224 Dolan, R., 229, 543, 811, 812, 813, 815, 816, 820, Ebbinghaus, H., 252, 271, 278 872 Eccles, J. C., 783, 902 Dolski, I., 524 Eddy, T. J., 608 Dominey, S. J., 536 Edelman, G. M., 152, 155, 158, 163, 191, 192, 196, 201, Donald, M., 390, 394, 565, 601, 633 406, 414, 436, 527, 532, 533, 720, 741, 742, 746, Donaldson, D. I., 813, 814 747, 748, 761, 776, 780 Donaldson, W., 255, 269 Edelson, M., 377 Donati, F., 710 Edmonston, W. E., 464 Donchin, E., 217 Edry, E., 299 Donkelaar, H. J., 790 Edwards, G., 462, 463 Dorfman, J., 452, 454 Edwards, P., 13 Dosher, B. A., 220, 226 Edwards, W., 688 Doty, R. W., 783 Egan, J., 686 Douglas, G., 127 Ehrlinger, J., 296, 309 Douglas, R. N., 291 Eich, E., 452, 460 Downes, J. J., 487 Eichenbaum, H., 797 Draine, S. C., 216, 218, 220 Eijk, R., 810 Dreesbach, H. A., 790 Eimer, M., 216, 217, 223, 228, 330, 331, 332 Dretske, F. I., 47, 57, 58, 61, 62 Einstein, G. O., 463 Drey, E. A., 406 Ekatodramis, E., 712 Dreyfus, G., 113 Ekman, P., 835 Dreyfus, H., 83 Eldridge, F. L., 791, 816 Driesch, H., 24 Eldridge, N., 25 Driver, J., 228, 229, 230, 239, 718, 872 Elger, C. E., 751, 752 Dronkers, N. F., 587, 787 Ellenberger, H. F., 458 Druch, A., 304 Elliot, A. J., 694 Drummond, J., 72 Elliot, J. M., 260, 272 Duarte, A., 817 Elman, J., 582 Dubin, L. L., 534, 535 Els, T., 715 DuBois, D., 535 Elston, G. N., 237 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

author index 917

Enard, W., 587 Fernandez-Duque, D., 290 Engel, A. K., 527, 537, 723, 733, 737, 743, 751, 753, Ferrari, M., 408 754, 779 Ferraro, F. R., 821 Engel, S. A., 816 Ferster, C., 572 English, E., 506 Fessard, A. E., 776 Engquist, G., 363 Feyereisen, P., 718 Epstein, M. D., 525 Ffytche, D. H., 465, 720 Epstein, R., 230, 819 Fick, L. J., 454 Epstein, S., 301, 313, 314 Fiedler, F., 306 Epstein, W., 304, 457 Fields, R. D., 780 Erber, R., 558 Fiez, J., 709 Erdelyi, M., 675, 684 Fillmore, C. J., 388, 392 Erdi,´ P., 735 Fink, G. R., 563, 874 Ericksen, C. W., 449, 684 Finke, R. A., 262 Erickson, M. H., 456, 457, 462 Finlay, B., 603, 604, 606 Ermentrout, B., 737 Fins, J. J., 794 Ervin, F. R., 788 Finucane, M., 293, 296 Eshkoli, N., 686 Fireman, G. D., 377 Eskenazi, J., 208 Fischer, C., 750, 753, 790 Esterman, M., 241 Fischer, L. E., 467 Etcoff, N. L., 228, 229, 234, 872 Fischer, R., 500, 531 Ettinger, M. G., 708 Fischhoff, B., 291, 297, 302, 303, 306, 309 Evans, A. C., 414, 711 Fiset, P., 711 Evans, F. J., 452, 453, 454, 463, 467 Fisher, S. E., 463, 587 Evans, S., 579, 606–607 Fishman, D. L., 466–467 Ey, Henry, 483 Fiske, D. W., 449 Fiske, S. T., 556, 557, 558 Fabiani, M., 817, 819 Fitch, W. T., 576 Fabrega,´ H., 615 Fitzsimons, G. M., 557, 562 Fac¸on, E., 790 Fivush, R., 393 Fadian, D. G., 588 Fize, D., 756 Fadiga, L., 80, 341, 563, 874 Flaherty, A. W., 795 Falk, D., 653 Flanagan, O. J., 55, 377, 397, 776, 777, 789, 810–811 Fallon, J. H., 711 Flavell, E. R., 419, 422, 424 Fan, J., 457 Flavell, J. H., 290, 295, 419, 422, 424, 454 Farah, M. J., 240, 336 Fletcher, P. C., 587, 812, 813, 815 Farber, D. A., 427 Flinn, M. V., 605 Farrer, C., 484 Flocco, R., 685 Farrer, M., 339, 345 Flohr, H., 711 Farrow, T. F., 693 Fodor, J. A., 57, 134, 137, 145, 158, 197, 244, 342, 583, Farvolden, P., 464 855 Faull, R. L., 785, 796 Fogassi, L., 80, 341, 388, 563, 587, 874 Faure, P., 734, 735, 736 Fogelson, A., 903 Faust, D., 693 Foley, M. A., 262, 263 Faymonville, M. E., 450, 454, 464, 465, 715, 794 Fonagy, P., 681 Fazendeiro, T. A., 291 Fonlupt, P., 733, 757 Fazio, R. H., 559, 560, 561 Foote, S. L., 439–440 Fechner, G., 26 Forgas, J. P., 289, 314 Feenan, K., 272 Forss, N., 874 Feig, S. L., 792 Forster, A., 454 Feigel, H., 31 Fortin, M., 796 Feigenbaum, E. A., 142 Fosse, M. J., 438 Feit, A., 690, 692 Fosse, R., 438 Feld, S., 656 Foster, D. P., 313 Feldman, S., 557 Foster, J., 55 Fell, J., 751, 752 Foster, M., 633 Fendrich, R., 237, 789 Foucher, J., 749 Fenwick, P. B., 500, 533, 534 Foulkes, W. D., 709 Ferguson, M. L., 556 Fourneret, P., 339, 345 Ferguson, S. A., 262, 272, 274 Fouts, R., 606 Fernandez, G., 751, 752 Fowler, J. C., 681 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

918 author index

Fox, E., 216 Gallo, D. A., 277, 278 Fox, N. A., 420 Gallup, G. G., 579, 606, 777 Foxe, J. J., 536 Gandolfo, R. L., 463 Foyaher, N., 790 Gao, F., 717 Frackowiak, R. S., 451, 522, 812, 813, 815 Gao, H. H., 425, 426 Frake, C. O., 643 Gardiner, J. M., 252, 253, 256, 257, 258, 259, 265, Franck, G., 450, 454, 464, 465, 709, 710, 794 266–267, 268, 270, 271, 274, 275, 293, 482, 483, Franck, N., 339, 345, 484 484, 486, 810, 818 Franco, F., 617 Gardner, B. T., 573, 586 Franco, L., 853 Gardner, R. A., 573, 586 Franco, S., 465 Garﬁnkel, A., 732 Franconeri, S. L., 210–211 Garner, W. R., 449 Frankfurt, H. G., 564 Garnerno, L., 527, 761 Franklin, N., 262 Garnett, E. S., 465 Franklin, S., 155, 162, 164, 165, 166, 167, 168, 197, 753 Garrett, J. B., 467 Frankovich, L., 296 Garro, L., 633 Franks, N. P., 711 Garry, M., 300 Franks, R. H., 620 Gastaut, H. C., 530, 533, 537, 538 Freeman, J. M., 789 Gatenby, J. C., 873 Freeman, W. J., 76, 152, 193, 194, 196, 201, 532, 735, Gati, J. S., 820 736, 737, 738, 739, 740, 758 Gauld, A., 447, 463 Freidman, W. J., 421 Gaulin, S. J., 609 French, J. A., 600, 619 Gawlick, B., 259 Freud, A., 678 Gazzaniga, M., 3, 202, 237, 239, 277, 379, 789, 840 Freud, S., 27, 208, 396, 408, 675, 676, 677, 679 Geary, D. C., 604, 605, 612 Freund, H. J., 563 Geertz, C., 643 Freytag, P., 306 Gehrke, J., 329, 330, 331, 332, 340, 341 Friberg, L., 541 Geiselman, R. E., 466–467 Fricchione, G. L., 542 Gelade, G., 751, 753 Fridlund, A. J., 835 Gell, A., 656 Fried, I., 332 Gelman, R., 600 Friedman, D., 817 Gemperle, M., 454 Fries, P., 533, 537, 733, 743, 751, 753, 754, 779 Gennaro, R. J., 51, 61, 841, 855 Frijda, N. H., 832 Gentilini, M., 790 Friston, K. J., 747, 748, 750, 755 Gentilucci, M., 334 Frith, C. D., 228, 229, 230, 339, 343, 344, 345, 522, George, M., 648 543, 560, 564, 718, 720, 722, 812, 813, 815 George, N., 749, 750, 751, 752 Fromm, E., 447 Georgieff, N., 339, 345, 484 Frost, D., 237, 337 Georgopoulos, A. P., 572 Frye, D., 410, 412, 418, 422, 423, 425 Geraci, L., 260, 261, 262, 266, 273, 274 Funayama, E. S., 873 Gergeley, G., 681 Funnell, M., 314 Gerhard, D., 821 Furey, M. L., 872 Gerstein, G. L., 749 Furmanski, C. S., 816 Gerton, M. I., 466 Furst, M. L., 534, 535 Gethin, R., 113, 503, 504, 505 Furst, P., 648 Geurts, K., 657 Fuster, J., 779 Geurtz, C., 633 Fuxe, K., 780 Ghetti, S., 297 Giacino, J. T., 715 Gabbard, G. O., 690 Giambrone, S., 577, 615, 617 Gabor, D., 190 Gibson, J. J., 243 Gabrieli, J. D., 268, 269, 466–467, 811, 813, 820, 900 Gibson, K. R., 586 Gabrieli, S., 466–467 Gicino, J., 794 Gadian, D. G., 522 Giedd, J. N., 414, 426 Gaillard, R. C., 454 Gigerenzer, G., 297, 306, 688 Gais, S., 710 Gilbert, D. T., 297, 556, 557–558, 559, 688 Galileo, G., 16 Gilbert, H., 583, 585 Galin, D., 776, 777 Gilbert, P., 618 Gall, F., 25 Gill, M. M., 447, 678, 680 Gallagher, S., 78, 415 Gillath, O., 685–686 Gallese, V., 341, 388, 563, 587, 874 Gilovich, T., 289 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

author index 919

Giovanello, K. S., 269 Graffin, N. F., 463 Givon,´ T., 587 Graham, K. R., 455 Glade, U., 711 Grainger, J., 216 Glaser, D., 522 Grammer, K., 614 Glaser, J., 690 Grande, L., 239, 241 Glaser, R., 394, 524 Grange, D., 484, 490 Gleason, C. A., 330, 331, 783, 795 Granier-Deferre, C., 414 Gleberman, L., 458 Grant, F. C., 789 Glenberg, A. M., 304 Gras-Vincendon, A., 484, 490 Glisky, M. L., 447, 465 Gratton, G., 217, 819 Glover, G. H., 466–467, 813 Gravenstein, S., 524 Glover, S., 333 Gravitz, M. A., 466 Glucksberg, S., 299 Gray, A. L., 456 Goad, H., 587 Gray, C. M., 737, 743 Gobbini, M. I., 872 Gray, J. A., 688, 694, 776, 779, 795, 832 Goebel, R., 718 Graybiel, A. M., 795 Gokalsing, E., 310, 484, 485, 494 Grazzani-Gavazzi, I., 385 Gold, J. M., 489, 490, 493 Green, C., 466–467 Goldberg, T. E., 484, 489, 490, 493 Green, D. M., 211 Goldie, P., 376 Green, F. L., 419, 422, 424 Goldman, A. I., 49, 54, 81, 776, 874 Green, M. F., 493 Goldman, R. F., 525 Greenberg, J. R., 680 Goldman-Rakic, P. S., 849 Greenberg, R. P., 681 Goldsmith, M., 289–290, 292, 293, 300, 307, 309, 310, Greenfield, P., 607 311, 312, 313 Greenfield, S. A., 780 Goldstein, A. P., 454 Greenwald, A. G., 207, 208, 209, 216, 218, 219, 220, Goldstein, D. G., 688 234 Goldstein, L., Gregg, V. H., 256, 258, 266–267, 453 Gollub, R. L., 542 Greischar, L. L., 530, 531, 538 Gollwitzer, P. M., 562, 565 Grether, W. F., 457 Golocheikine, S. A., 534 Grew, N., 24 Gomes, G., 752 Grezes,` J., 563 Gomez,¨ J. C., 607 Griffin, D., 29, 289, 304, 306, 851 Gomez, L., 113 Grindal, B. T., 648 Gong, Q. Y., 818 Grindlay, J. H., 789 Good, C. D., 522 Grinvald, A., 736 Goodale, M. A., 80, 236, 244, 333, 334, 335, 336, 338, Groden, M., 376 563, 718, 759, 781, 783, 810–811, 820 Groenewegen, H. J., 787 Goodall, J., 388, 586, 590, 606, 612, 616 Groome, J. R., 417 Goode, R., 790, 792, 794 Grosbars, H., 722 Goodenough, W., 643 Gross, C. G., 793 Gooding, P. A., 810 Gross, J., 532, 900 Goodman, G. S., 415 Grossman, E., 872 Goodwin, C., 657 Grossman, P., 523 Gopnik, A., 81, 423, 424, 618, 619 Grossman, R. G., 776, 778 Gopnik, M., 587 Grossman, R. I., 790 Gorassini, D. R., 446, 457, 461 Grosz, H. J., 456 Gore, J. C., 228, 230, 873 Grun,¨ R., 585 Gorman, J. M., 489, 490, 493 Gruneberg, M. M., 306, 308–309 Goshen-Gottstein, Y., 253, 269, 810 Grunthal,¨ E., 785 Gosselin, F., 873 Grush, R., 874 Goto, Y., 795 Grusser,¨ O. J., 793 Gottfried, J. A., 811 Gruzelier, J. H., 463, 536 Gould, P. V., 25, 796 Guberman, A., 790 Gould, S. J., 25, 602, 603, 604 Guenther, H., 113 Goulet, J., 648 Guez, J. R., 456 Gowlett, J. A., 581 Guggenberger, H., 712 Gracco, C., 788 Gugger, M., 790 Graesser, A. C., 162, 197, 290 Guillary, R. W., 747 Graf, P., 252–253, 259, 268, 466, 810 Guillaume, M. M., 584 Graff-Radford, N. R., 790 Guillery, R. W., 782, 787, 792, 793 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

920 author index

Gumperz, J. J., 640, 641 Harnishfeger, K. K., 604, 605, 609, 612 Gunaratana, H., 508, 509 Harre,´ R., 377 Guntrip, H., 680 Harriman, P. L., 457 Gur, R. C., 456, 464, 790 Harrington, A. H., 521, 522, 524, 789 Gur, R. E., 464 Harris, J., 582 Gurwitsch, A., Harris, P. L., 384, 391, 393, 397, 419 Guseo, A., 794 Harsch, N., 261, 262 Gusnard, D. A., 529 Hart, D., 461, 777 Gutkin, B., 737 Hart, J. T., 290, 293, 294, 295, 306, 314, 534, 811, 820 Guttentag, R., 295, 296, 298, 305 Hart, W. K., 585 Guzeldere,¨ G., 55, 61, 810–811 Harting, J. K., 793 Guzman, A. E., 299 Hartley, T., 522 Gwynn, M. I., 455, 459 Hartmann, E., 709 Gyllensten, U., 583, 584 Hartmann, H., 680 Harvey, M. A., 457 Ha, Y., 465 Hasboun, D., 750 Habib, R., 811, 813 Haselton, M. G., 602 Hacker, D. J., 290 Hasher, L., 557 Haddon, A., 637 Hashtroudi, S., 262, 272, 274, 311, 467, 483 Hadley, R., 156, 157–158 Haslam, N., 613 Haffenden, A., 334 Hassin, R. R., 557, 563 Hagan, S., 893 Hastie, R., 300, 396 Haggard, P., 329, 330, 331, 332, 340, 341 Haugeland, J., 143 Hagoort, P., 232 Hauser, M. D., 576, 579 Hahn, J., 584 Hawley, K. J., 272 Haider, H., 710 Haxby, J. V., 872 Haier, R. J., 711 Hayes, C., 586 Hains, S. M., 414 Hayes, P., 147 Haith, M. M., 415 Hayes-Roth, B., 118, 195 Haji-Ali, N., 711 Haynal, A., 454 Hake, H. W., 449 Hayward, D., 817 Haken, H., 732, 758 Hazan, C., 415 Halevy, V., 686 Hazeltine, E., 228, 340 Halifax, J., 648 He, S., 792 Halligan, P. W., 451 He,Z.J.,528 Hallowell, A., 634, 641 Hebold, I., 787 Hallowell, I., 631, 633 Heckhausen, H., 795 Hamada, T., 525 Heelas, P., 633 Hamann, S., 261 Hegel, G., 24 Hameroff, S. R., 711, 781, 892, 893 Heidegger, M., 71, 73, 82 Hamill, J., 633 Heider, F., 872 Hamilton, D., 454 Heinecke, A., 335 Hamilton, M., 256, 265, 269, 277 Heinze, H., 817 Hamilton, W. D., 599, 618 Heinzen, D., 897, 907 Hammer, A. G., 452 Heisenberg, W., 883, 905 Hammond, D. C., 460 Heiss, W. D., 787, 790 Hancock, T., 278 Heit, F., 335 Haney, C., 556 Held, R., 237, 337 Hanna, R., 733 Helmholtz, H., 26 Hannula, D. E., 215, 227, 234, 235 Hempel, C. G., 776 Hanson, L. G., 711 Henderson, V. W., 782 Harbluk, J. L., 453 Hendry, D., 335 Hardaway, R. A., 685 Henke, K., 783 Hardcastle, V. G., 810–811 Henkel, L. A., 262, 263, 820, 821 Hardin, C. L., 657 Hennessy, R. T., 456 Hare, B., 388, 578, 607, 608, 870 Hennig, J., 712 Hargadon, R., 455 Henshilwood, C. S., 584 Hari, R., 874 Henson, R. N., 811, 815, 816, 820 Harkness, S., 636 Herculano-Houzel, S., 751 Harman, G., 44, 45 Herholz, K., 787 Harnad, S., 130, 144 Herman, L. M., 573 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

author index 921

Herndon, F., 307 Holsboer, F., 710 Herr, P. M., 561 Hommel, B., 339, 340 Herscovitch, P., 441, 709, 710 Honeycutt, H., 600 Herscovits, M. J., 633 Hopﬁnger, J. B., 822 Hertzog, C., 294, 298, 303, 308, 312 Hopkins, J., 525 Hesketh, B., 154 Hopkins, R. O., 817 Hess, C. W., 790 Horgan, T., 54 Hewes, G. W., 587 Hornak, J., 840 Hewitt, E. C., 455 Horton, J. C., 718 Hewitt, S., 232 Hossack, K., 51, 61 Heyes, C. M., 574, 619 Houle, S., 811, 812, 813 Heyl, B., 460 Howard, R. J., 465 Heywood, C. A., 293, 718 Howe, M., 421 Hickey, P. R., 417 Hoyt, I. P., 451, 463 Hicks, J. L., 255, 278 Hoyt, W. F., 718 Higgins, E. T., 560, 561 Hrbek, A., 417 Hilal, S. K., 790 Hrdy, S. B., 621 Hilgard, E. R., 454, 455, 456, 458, 459, 462, 465 Huang, C-T., 619 Hilgard, J. R., 446, 447, 451 Huang, H., 414 Hilgetag, C. C., 746, 747 Hubel, D. H., 712 Hill, C., 584 Huffman, K., 605 Hill, K., 605 Hugdahl, K., 463, 783 Hillaire, C., 391 Hughes, R. A., 789 Hillyard, S. A., 232 Hugueville, L., 749, 751, 752 Hilsenroth, M. J., 681 Huisman, T., 712 Hinchley, J. L., 295 Hull, C. L., 453, 466 Hinkley, L., 216 Humphrey, G. K., 367, 820 Hinton, A., 633, 653 Humphrey, N. K., 378, 379, 386, 598, 601, 605, 610, Hirai, T., 536, 537 614, 841, 842 Hirschfeld, L., Hunt, E., 153, 162 Hirshman, E., 255 Hunt, G. R., 573 Hirstein, W., 345 Hunt, H. T., 633, 660 Hitch, G. J., 187, 563 Hunt, R. R., 260, 261 Hobbes, T., 20 Huprich, S. K., 681 Hobson, J. A., 436, 438, 439, 440, 464, 710 Hurley, S., 80 Hochberg, J., 457 Huron, C., 485 Hochstein, S., 718 Hurtado, A. M., 605 Hodes, R., 438 Hurwitz, T. D., 708 Hodges, J. R., 872 Husain, M., 228, 229, 230 Hoerl, C., 423, 424 Husserl, E., 27, 67, 68, 71, 72, 73, 74, 75, 77, 79, 82, Hoesen, G. W., 790 85, 634, 649 Hofbauer, R. K., 450, 463, 464 Hutchison, W. D., 874 Hoffman, E. A., 872 Hoffman, J., 335 Iacoboni, M., 789, 874 Hoffman-Chemi, A., 689 Illingworth, H. A., 584 Hoffrage, U., 297, 306 Ingman, M., 583, 584 Hoﬂing, C. K., 460 Ingvar, M., 811 Hofstadter, D. R., 121, 164 Ino, M., 790 Hogan, P. C., 376, 396 Inoue, C., 255 Hogg, D., 380 Insel, T. R., 603 Holcombe, A. O., 331, 340 Intriligator, J., 792 Holdstock, J. S., 818 Isaac, C. L., 818 Holender, D., 207, 209, 210, 211, 213, 215, 220, 223, 232 Ishai, A., 872 Holgate, B., 296 Itakura, S., 617 Hollan, D., 654, 655 Itano, A., 897, 907 Holland, O., 170 Ito, J., Hollender, D., 759, 783 Ivry, R. B., 241 Holm, S., 541 Izard, C. E., 835 Holmes, D. S., 500, 535 Holmes, J. C., 618 Jack, A., 2, 760 Holsanov´ a,´ J., 359 Jackendoff, R., 152, 357, 358, 361, 365, 366, 370 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

922 author index

Jackson, D. C., 524, 688, 694 Jolly, A., 606 Jackson, E. M., 818, 821 Jones, B., 456 Jackson, F., 39 Jones, C., 813 Jackson, J. R., 560 Jones, E. G., 712, 716, 781, 782, 784, 785, 786, 787, Jackson, M., 633, 658 788, 792 Jacobson, K., 618 Jones, T. C., 263, 814 Jacoby, L. L., 223, 224, 252, 253, 256, 257, 258, 259, Jonsson,¨ F. U., 293 260, 263, 267, 268, 272, 274, 277, 289, 293, 295, Joordens, S., 207, 225, 257 297, 301, 310, 313, 314, 315, 453, 811 Josephs, O., 816 Jacques, S., 409, 425 Jost, J. T., 289, 291, 690 Jahoda, M., 679 Jouvet, M., 438, 708, 791 Jain, S., 819 Joyce, J. S., 37, 454 Jakobson, L. S., 236, 336, 563 Juengling, F. D., 715 James, T. W., 820 Jung, R., 792 James, W., 27, 28, 77, 94, 95, 186, 295, 315, 359, Juola, J. F., 259 360, 362, 381, 461, 565, 673, 733, 762, 855, 897, Jurak, P., 233 900 Jurist, E. L., 681 Jameson, K. A., 301 Juslin, P., 289, 291, 306 Janer, L. D., 783 Janet, P., 450 Kabat-Zinn, J., 508, 509, 521, 522, 524 Janis, C., 582 Kacelnik, A., 848 Jansen, R. D., 457 Kaessmann, H., 583, 584 Jasiukaitis, P., 463 Kagan, J., 406, 420 Jaskowski, P., 216 Kahana, M. J., 533 Jasper, H. H., 723, 775, 782, 783, 788, 791, 792 Kahane, P., 747 Java, R. I., 252, 253, 256, 258, 259, 266–267, 268, 270, Kahn, I., 815 274, 275, 483, 484 Kahneman, D., 289, 295, 298, 313, 688 Jayaraman, K. S., 501 Kaiser, M., 746, 747 Jaynes, J., 125, 394 Kajimura, N., 710 Jeannerod, M., 334, 335, 336, 337, 339, 345, 560, Kajiyama, Y., 113 598, 609, 783 Kales, A. A., 709 Jeffries, N. O., 414 Kalin, N. H., 524, 688, 694 Jenike, M. A., 229, 234 Kalitzin, S., 537 Jenkins, J. M., 832, 836 Kallio, S., 447 Jennett, B., 714, 715, 776, 794 Kalogeras, J., 330, 331, 332, 340 Jennings, J. M., 256, 257, 267 Kameyama, M., 790 Jerger, K., 732 Kamil, A. C., 600, 619 Jerrison, H. J., 603 Kaminska, Z., 266–267, 270 Jessee, S., 586 Kamiya, J., 449 Jibu, M., 905 Kaneko, K., 736 Jimenez,´ L., 407 Kanerva, P., 164, 782 Jimenez, R., 213 Kant, I., 22 Joaquim, S. G., 298, 299 Kanwisher, N., 201, 228, 230, 819 Joeri, P., 712 Kaplan, H., 605 Joffe, T. H., 605 Kaplan, N., 681, 694 Johansson, G., 872 Kapur, N., 818 Johansson, M., 232, 233, 234, 235 Kapur, S., 811, 812, 813 John, E. R., 202 Karbe, H., 787 John, O. P., 447 Karlberg, P., 417 Johnson, B. H., 783 Karma Chagme,´ 510, 513, 515, 517, 519 Johnson, J. T., 298, 307 Karmiloff-Smith, A., 157, 158, 409, 612 Johnson, K., 296 Kasamatsu, A., 536, 537 Johnson, L. S., 455 Kass, J. H., 793 Johnson, M. K., 262, 263, 293, 311, 483, 557, 814, 819, Kassubek, J., 715 820, 821 Kastner, S., 722 Johnson, R. F., 459, 467 Katz, A., 332 Johnson-Laird, P. N., 135, 169, 184, 855 Katz, R., 648 Johnsrude, I. S., 522 Kaube, H., 543 Johnston, W. A., 272 Kauffmann-Muller, F., 485, 494 Johnstone, S., 839, 840, 845 Kaufman, E. F., 788 Johnstone, T., 522 Kaufmann C., 710 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

author index 923

Kavanaugh, R. D., 419 Klein, R. M., 342 Kay, P., 641 Kleinbolting,¨ H., 297, 306 Keane, M. M., 269 Kleitman, N., 438, 448, 709 Kegl, J., 587 Klimach, V. J., 417 Keglevich, L., 294, 304 Klimesch, W., 530, 534, 535, 536 Keim, C. S., 456 Klin, C. M., 299 Kelemen, A., 155, 162 Klinger, M. R., 216, 218, 219 Keller, I., 795 Klotz, W., 217, 335 Kelley, C. M., 257, 258, 268, 289, 293, 295, 297, 298, Klumpp, G., 296 300, 309, 310, 313, 314, 315 Knecht, H., 584 Kelley, H. H., 556 Knight, R. T., 269, 817 Kelley, M., 692 Knippenberg, A., 562 Kelso, J. A., 732, 734, 736, 740, 741, 758 Knowlton, B. J., 269, 816, 817 Kempton, W., 641 Knox, V. J., 454, 455 Kennedy, W. A., 228, 229 Knuf, L., 560 Kensinger, E. A., 261 Kobes, B. W., 51, 60, 61 Kentridge, R. W., 293 Kocarev, L., 750 Kern, I., 74 Koch, C., 36, 169, 230, 718, 743, 744, 750, 757, 758, Kernberg, O., 681, 682 776, 779, 787, 792, 794, 853 Kerszberg, M., 199, 744, 745 Koechlin, E., 217, 219, 231 Kertesz, A., 787 Koepping, K., 635 Keysers, C., 341, 388 Koestler, A., 716 Khalsa, G., 542 Koestner, R., 467 Kidder, D. P., 298, 303 Koh, S. D., 489, 493 Kiecolt-Glaser, J. K., 394, 524 Kohler, E., 341 Kiefer, M., 232 Kohler,¨ S., 717, 810–811 Kierkegaard, S., 613 Kohler,¨ W., 572 Kiesel, A., 335 Kohn, C., 241 Kihlstrom, J. F., 447, 449, 450, 451, 452, 453, 454, Kohut, H., 682 455, 458, 459, 460, 461, 462, 463, 464, 465, 466, Kokmen, E., 790 467, 674, 783 Kolb, B., 787 Kilduff, P. T., 240, 241 Kolers, P., 272 Kim, J., 42 Konig, P., 533, 723, 737, 743, 751, 753, 754, 779 Kimball, D. R., 305 Konishi, S., 813, 814 Kimura, M., 795 Konner, M., 588 King, G. A., 560 Kopell, B. S., King, J. F., 304 Korein, J., 790, 792, 794 King, P., 16 Koren, D., 310 Kinney, H. C., 790, 792, 794 Korfmacher, J. E., 422 Kinoshita, S., 268, 276 Koriat, A., 270, 289–290, 291, 292, 293, 295, 296, 297, Kinsbourne, M., 782, 789 298, 299, 300, 301, 302, 303, 304, 305, 306, 307, Kipp, K., 604, 609 309, 310, 311, 312, 313, 314, 315 Kira, J., 795 Korn, H., 734, 735, 736 Kircher, T. T., 345 Kornell, N., 292, 312 Kirk, R., 147 Kosaka, H., 525 Kirkpatrick, J., 633 Koskoff, Y. D., 789 Kirsch, I., 447, 461 Kosslyn, S. M., 450, 465, 466 Kirveskari, E., 874 Kossoff, E. H., 789 Kirwan, A., 296 Kotler-Cope, S., 153–154, 158 Kisilevsky, B. S., 414, 417 Kotter, R., 746, 747 Kitano, T., 587 Kounios, J., 819 Kitchener, E. G., 817 Kourch, M., 458 Kiuchi, M., 525 Koutstaal, W., 811, 813, 815, 818, 819, 821, 822 Kjaer, T. W., 541 Kozin, M., 261 Klapp, S., 216 Kraft, P. M., 446 Klaver, P., 751, 752 Krauss, M., 494 Klein, D., 26 Kraut, M. A., 534 Klein, E., 310 Krauthamer, G. M., 788 Klein, G. S., 679 Krebs, J. R., 848 Klein, J., 714 Kreigel, U., 46, 49, 54, 55, 58, 59, 60, 61 Klein, R. G., 584, 585 Kreiman, G., 230, 758 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

924 author index

Kreiswirth, M., 376 LaPiere, R. T., 556, 559 Krinsky, R., 299 Larocque, L., 385 Kripke, S., 41 Larson, J., 112 Kris, E., 680 Lashley, K. S., 776, 780 Krogman, W. M., 582 Lassonde, M., 793 Kroll, N. E., 269, 817 Latane, B., 556 Kronberg, E., 794 Laterre, C., 718 Krovitz, G., 586 Latham, G. P., 557 Kruger, A. C., 607 Latto, R., 238 Kruger, J., 296 Laughlin, C. D., 633, 634, 648, 652, 660 Kruglanski, A. W., 289, 291, 690 Laurence, J. R., 455, 461 Kruper, D., 789 Laurent, B., 790 Kuhl, B., 466–467 Laurent, G., 782 Kuhn, T., 447 Laureys, S., 195, 454, 465, 715, 717, 794 Kuklick, H., 637 Laverriere, M., 454 Kulik, J., 261 Lavie, N., 720, 722 Kummer, H., 616 Law, I., 711 Kunda, Z., 690 Lazar, S. W., 542 Kunde, W., 335, 340 Lazarus, R. S., 832 Kunst-Wilson, W. R., 221 Lazzara, M. M., 269, 817 Kuperman, S., 427 LeDoux, J. E., 244 Kupfer, D. J., 441 Le Ny, J. F., 308 Kuramoto, Y., 737 Le Van Quyen, M., 533, 734, 735, 737, 738, 749, 758, Kurt, A., 233 762 Kurths, J., 750 Le, B., 230, 231 Kurzweil, Raymond, 120 Leach, K., 262 Kutas, M., 232 Leak, G. K., 851 Kuusela, J., 753 Leakey, R., 582 Kuzendorf, R., 465 Lebiere, C., 152 Lecanuet, J. P., 414 LaBerge, D., 460–461 Lecours, A. R., 782 Lachaux, J. P., 71, 202, 527, 528, 533, 736, 739, 747, Lederman, C., 648 749, 750, 751, 752, 761, 762 Ledger, D. W., 600, 619 Lachmann, F. M., 684 LeDoux, J. E., 152, 229, 239, 378, 684, 840, 855 Lack, L. C., 528 Lee, K. K., 414, 454 Lacroix, S., 796 Lee, M. B., 229, 234 Ladavas, E., 240, 241 Lee, R. B., 391 Lagerman, S. K., 814 Lee, R. G., 587 Lagravinese, G., 341 Lee, S. J., 406, 648 Lai, C. S., 587 Lee, V. A., 301 Laird, J., 155 Lee-Chai, A. Y., 562 Lalonde, P., 296, 298, 301 LeGros Clark, W. E., 786 Lambert, A. J., 310 Lehmann, J. W., 525 Lamme, V. L., 718 Lehrer, K., 50 Lampinen, J. M., 277 Lehto, M., 753 Lamy, M., 450, 454, 464, 465, 794 Leiblich, I., 298 Lancaster, J., 605 Leibniz, G., 17, 21 Lancker, D., 778, 789 Leibowitz, H. W., 456 Landesman, C., 778 Lenox, J. R., 455 Landis, T., 783, 789, 793 Leonesio, R. J., 292, 293, 296, 308, 312 Landsman, R., 300 Leopold, D. A., 712, 714, 720 Landzberg, K. S., 458 Lepore, F., 793 Lane, R. D., 230 Leslie, A. M., 579 Lang, A., 647 Lesser, R. P., 534, 789 Lang, B., 291 Lester, D., 618 Lang, E. V., 454 Leube, D. T., 345 Langacker, R. W., 356 Levelt, W. J., 230 Lange, A. F., 455 Levi-Strauss,´ C., 642 Lange, C., 855 Levin, D. T., 720 Langer, E. J., 558 Levine, B., 414 Lansman, M., 153, 162 Levine, J., 36, 46, 61, 62 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

author index 925

Levine, W. H., 299 Loose, C., 723 Levinson, S. C., 640, 641 Lopes da Silva, F. H., 534, 537 Levy, B. L., 466–467 Lories, G., 291, 297, 308–309, 454 Levy, K. N., 681 Lou, H. C., 540, 711 Levy, R. I., 654, 655 Lourenco, S. F., 425 Levy-Bruhl,´ L., 639 Lovblad, K. O., 790 Levy-Sadot, R., 291, 295, 297, 298, 299, 300, 310, 313, Low, J. A., 417 314, 315 Lowie, R. H., 635 Lewes, George, 29 Lozsadi, D. A., 792 Lewis, D. K., 56 Lubbe, R. H., 216 Lewis, M., 393, 406, 407, 420, 421, 606 Lucy, J., 641 Lewis, S., 335 Luders, H., 789 Lewis, V. J., 297 Ludwig, A. M., 448 Lhermitte, F., 343, 564, 782 Lui, F., 341 Liberman, A., 306 Lund, T. E., 711 Liberman, V., 303, 307 Lundy, R. M., 456, 463 Libet, B., 54, 125, 330, 331, 332, 752, 783, 795, 796, Luria, A. R., 408, 614 899 Lurz, R., 48, 58 Lichtenstein, S., 291, 297, 302, 303, 306, 309 Lutgendorf, S., 454 Lickliter, R., 600 Lutz, A., 71, 528, 530, 531, 533, 538, 542, 736, 749, Lieb, W. R., 711 762 Lieberman, D. E., 586 Lutz, C., 633 Lieberman, P., 586 Lutzenberger, W., 427 Lieblich, I., 299 Luxen, A., 450, 454, 464, 465, 709, 710, 794 Liegeois,´ F., 588 Lycan, W. G., 47, 48, 780 Liley, D. T., 736, 737 Lycett, J., 590 Lillard, A. S., 620 Lydon, J., 685 Lilleen, P. R., 464 Lynn, S. J., 447, 456, 461 Lincoln, J., 633, 648 Lyons, C., 371 Lind, D. I., 456 Lindgren, M., 232, 233, 234, 235 Ma’ayan, H., 298, 303, 304, 305, 315 Lindsay, D. S., 257, 262, 298, 300, 307, 311, 467, 483 Macaulay, V., 584 Ling, S., 722 MacCarty, C. S., 789 Lipsitt, L. P., 417 Macchi, G., 786 Lipstad, B., 778, 789 MacDonald, G. F., 648 Llewelynn, K., 335 Macdonald, H., 455 Llinas, R., 201, 527, 732, 742, 747, 776, 780, 794 MacDonald, J. L., 588, 648 Lloyd, A. T., 613, 850 MacGregor, D. G., 293, 296 Lloyd, D., 71 Macguire, E. A., 522 Loar, B., 40 Mack, A., 199, 213, 224 Lobaugh, N. J., 191 MacKay, D. G., 295 Lock, A., 633 MacKenzie, C. L., 334 Locke, E. A., 557 Mackie, J. L., 58 Lockhart, L. K., 676, 690, 694 Mackintosh, N. J., 832 Lockhart, R., 259, 292, 410 MacKuen, M., 688 Lockwood, M., 28 MacLaughlin, D., 587 Lodge, D., 397, 711 MacLeod, C. M., 460, 683 Loeb, G. E., 328, 329, 343 MacLeod-Morgan, C., 463 Loenneker, T., 712 Macmillan, N. A., 211, 213, 257, 818 Loewenstein, R. M., 680 Macomber, J., 618 Loftus, E. F., 261, 300, 302, 821 Madigan, S. A., 268, 294, 314 Loftus, G. R., 302 Madler, C., 712 Loftus, J., 153 Madsen, J. R., 533 Logan, G., 154 Maes, P., 166 Logan, H., 454 Mafﬁ, L., 657 Logothetis, N. K., 572, 712, 714, 719, 720, 759, 779 Magno, E., 330, 331, 332, 340 Lohr, N. E., 681 Magoun, H. W., 438, 791 Long, J. K., 648 Magoun, W., 708, 712 Long, P. J., 681 Mahalingam, I., 112 Lonigan, C. J., 361 Maharishi Mahesh Yogi, 534, 535 Loomis, J. M., 457 Maher, B. A., 459 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

926 author index

Mahoney, M. J., 690 Mathews, R., 153–154, 158 Mahowald, M. W., 708, 710 Mathis, D., 155 Mai, N., 747 Mathis, J., 790 Main, M., 681, 694 Matilal, B. K., 108, 113 Mair, R. G., 797 Matson, W., 14, 16 Maki, R. H., 293 Mattingley, J. B., 239 Malamut, B. L., 790 Mattler, U., 335 Malarkey, W. B., 524 Maturana, H. R., 529, 757 Malcolm, N., 438, 846 Matvey, G., 295, 296, 298 Malfara, A., 456 Mauguiere,` F., 790 Malhotra, M. S., 525 Maunsell, J. H., 792, 793 Mallard, D., 452, 457, 458 Mauss, M., 638 Mallek, R., 790 Mavin, G. H., 560 Maloney, J. C., 58 Mayeda, S., 113 Mandler, G. A., 186, 252–253, 259, 342 Mayer, A. R., 819 Mandler, J. M., 418 Mayer, D., 454 Mangin, J., 230, 231 Mayes, A. R., 487, 810, 818 Mangun, G. R., 789, 817, 822 Maynard, A., 607 Manning, C. G., 300 Mayr, E., 616 Manns, J. R., 817 Mazziotta, J. C., 874 Mantyla, T., 260, 266 Mazzoni, G., 292, 296, 298, 303, 304, 308–309 Maquet, P., 441, 450, 454, 464, 465, 709, 710, 794 McAllister, M. K., 618 Mar, R. A., 390 McAndrews, M. P., 269, 783, 810–811, 817 Marbach, E., 74 McAvoy, M. P., 811 Marcar, V. L., 712 McBratney, B. M., 586 Marcel, A., 169, 202, 210, 211, 213, 238, 239, 328, 329, McBrearty, S., 584 332, 342, 343, 344, 455, 778 McBurney, D. H., 609 Marchetti, C., 343 McCarley, R. W., 439, 440 Marchitelli, G., 296, 304 McCarthy, D. C., 572 Marcovitch, S., 410, 420, 422 McCarthy, G., 332, 871 Marcus, G. E., 688 McCarthy, J. C., 123, 124, 143, 147, 228, 230 Marcus, S., 396 McCauley, L., 155, 162 Marder, E., 780 McCauley, R. N., 618 Marean, C. W., 584 McClachlan, D. R., 453 Marie, P., 787 McClelland, D. C., 467 Marie-Cardine, M., 484 McClelland, J. L., 154, 160, 161, 170, 522, 563, Maril, A., 815 650 Marino, L., 607 McClintock, M. K., 871 Mark, R. E., 820 McCloskey, D. I., 330, 335, 783, 796 Mark, V. H., 788 McCloskey, M., 299 Markowitsch, H. J., 783, 790 McConkey, K. M., 447, 449, 456, 457, 458, 459, 460, Marlot, C., 756 462, 463, 464 Marois, R., 721, 722 McCormack, T., 421, 423, 424 Marsh, E. J., 266, 277 McCormick, D. A., 782, 787 Marsh, R. L., 255 McCormick, P., 225 Marshall, G., 455 McCullough, L., 678 Marshall, J., 237, 336, 418 McDaniel, M. A., 261, 463 Martelli, C., 484 McDermott, D., 131, 132, 134, 140, 144, 145, 147 Marteniuk, R. G., 334 McDermott, J., 228, 230 Martin, A., 811, 814, 820 McDermott, K. B., 253, 263, 267, 273, 277, 810, 814, Martin, E., 712 818 Martin, J. S., 534, 535 McDonough, L., 418 Martinerie, J., 202, 527, 528, 533, 736, 738, 739, 749, McDougall, W., 583, 637 750, 751, 752, 761 McDowd, J. M., 276 Mashour, G. A., 716 McGagg, C., 794 Masling, J. M., 682–683 McGinn, Colin, 36 Maslow, A., 160 McGlashan, T. H., 454, 613 Massin-Krauss, M., 310, 484 McGlinchey-Berroth, R., 239, 240, 241 Masters, S., 255 McGlynn, S. M., 451, 452 Mather, M., 263, 819, 820, 821 McGrew, W. C., 387, 590, 606, 611, 612 Mathews, A., 683 McGuinness, C. M., 788 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

author index 927

McGuire, M. J., 293 Mills, R. P., 790 McGurk, H., 588 Milner, A. D., 80, 236, 244, 333, 336, 338, 563, 717, McInerney, S. C., 229, 234 718, 759, 781, 783 McIntosh, A. R., 191, 811, 812, 813 Milner, B., 186 McKee, R. D., 418 Milo, R. G., 584 McKeefry, D., 465 Mils, C., 618 McKenna, P. J., 484 Minotti, L., 749, 751, 752 McLaughlin, B., 30 Minsky, M., 122, 123, 134, 163 McLeod, P., 853 Mintun, M. A., 441 McManus, J., 633, 634, 648, 652 Miozzo, M., 299, 366 McNaughton, B., 160, 161 Mischel, W., 557 McNeill, D., 290, 366, 587 Mishkin, M., 337, 588 McVay, T. E., 377 Misra, B., 897, 907 Mead, G. H., 384 Mitchell, D. B., 260 Meade, M. L., 266 Mitchell, K. J., 293 Meade, S. D., 191 Mitchell, M., 164 Meadows, J. C., 718 Mitchell, R. W., 607 Meaney, M. J., 522 Mitchell, S. A., 655, 680 Meehan, W., 584 Mithen, S., 391, 394, 611 Meehl, P. E., 693 Mitrofanis, J., 792 Mehlman, M. J., 584 Mitroff, S. R., 210–211 Meissner, I., 790 Moga, N., 310 Mellars, P. A., 583 Mohanty, J., 84 Mellers, B., 688 Mohr, J. P., 787, 790 Mellor, D. H., 60 Mojardin, A. H., 297, 310 Melnick, R., 296, 298, 301 Molnar, M., 794 Melo, B., 821 Monaco, A. P., 587 Meltzoff, A. N., 415, 418, 619 Monheit, M. A., 240 Melzack, R., 465 Monk, A. F., 297 Menon, R. S., 820 Monks, J., 306, 308–309 Menzel, C. R., 579 Monteith, M. J., 560 Menzel, E., 661 Montgomery, G. H., 454 Merikle, P. M., 207, 209, 210, 211, 214, 216, 220, 222, Moo, L. R., 534, 811, 820 224, 225, 227, 235, 257, 714 Moody, T., 123 Merikle, W. P., 244 Mooibroek, J., 534 Merleau-Ponty, M., 71, 72, 76, 82, 526, 658, 659 Moore, C., 408 Merrill, E., 157, 160, 161 Moore, G. E., 44, 121, 776, 778 Mesulam, M. M., 564, 747 Moore, G. P., 749 Metcalfe, J., 289, 290, 291, 292, 293, 295, 298, 299, Moore, J., 27 305, 306, 308, 312, 314, 454 Moore, M. K., 418 Metzinger, T., 449, 776, 864 Moore, R. Y., 441 Meuret, P., 711 Moore, T., 722 Meyer, D. E., 295 Moran, D., 68 Meyer-Lindenberg, A., 427 Moravec, H. P., 120 Michalski, R., 159 Moray, N., 209 Michel, D., 790 Moreno, D. R., 715, 794 Michel, F., 438, 718 Moret, V., 454 Michel, L., 484 Morgan, A. H., 454, 455 Miezin, F. M., 231, 709, 812 Morgan, C., 30 Migler, B., 789 Morgan, R., 415 Mikulincer, M., 681, 685–686 Morin, A., 407 Milberg, W. P., 239, 240, 241 Morrell, L. K., 537 Miles, H. L., 573 Morris, B., 633 Milgram, S., 556 Morris, C. C., 298, 304 Milich, R., 684 Morris, H. H., 789 Mill, J., 28, 29 Morris, J. S., 228, 229 Millenson, J. R., 832 Morris, R. G., 840 Miller, G. E., 524, 650 Morruzzi, G., 708, 712 Miller, J., 795 Morse, D. R., 534, 535 Miller, M. B., 277 Mortimer, A. M., 484 Miller, M. E., 455, 456 Moruzzi, G., 438, 791 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

928 author index

Moscovitch, M. M., 1, 253, 269, 276, 328, 342, 406, Nelson, D. L., 300 717, 776, 783, 810, 811, 812, 813, 817, 818, 821 Nelson, J., 15 Moses, L. J., 423, 425 Nelson, K., 392, 393 Moskowitz, G. B., 561 Nelson, T. O., 289, 290, 291, 292, 293, 296, 298, 299, Most, S. B., 213 300, 303, 304, 305, 308–309, 312, 316, 454 Motzko, D., 711 Nemirow, L., 56 Moule, S., 812 Nesse, R. M., 613, 850 Moulin, T., 790 Netter, F. H., 785 Mourelatos, A., 13 Neuberg, S. L., 558 Moutier, F., 787 Neuenschwander, S., 751 Moutoussis, K., 779 Neumann, O., 217, 335 Mozaz, G., 718 Neuschatz, J. S., 277 Mozer, M., 155 Neverling, M., 790 Muckli, L., 718 Newberg, A. B., 541, 652 Mueller, C. J., 522 Newcombe, F., 718 Mueller, M., 217, 231 Newell, A., 155, 195 Muesseler, J., 339 Newman, J. R., 140, 196, 688, 781, 792 Muir, D. W., 417 Newman, L. S., 561 Muller, D., 521, 522, 524 Newman, W. R., 688 Muller,¨ U., 410, 422, 425 Newport, E. L., 522, 582 Mulvaney, S., 458 Newsome, W. T., 793 Munk, M. H., 723, 751, 779, 792 Newtson, D., 363 Munkle, M. C., 785, 796 Nhouyvanisvong, A., 299 Munro, G. D., 676, 690, 694 Nicastro, N., 603, 604, 606 Munzel,¨ S., 584 Nichols, J., 370 Murata, T., 525 Nichols, S., 422 Muraven, M., 558 Nickerson, R. S., 306 Murphy, K. M., 644, 658 Nicolis, G., 732, 734, 758 Murray, D. M., 307 Niedeggen, M., 723 Niemann, L., 523 Nabatabe, H., 790 Nigg, J., 681 Naccache, L., 196, 198, 199, 201, 202, 216, 217, 218, 219, Nikolaev, A., 427 226, 230, 231, 328, 342, 484, 527, 744, 759 Nilsson, L. G., 811, 812, 813 Nace, E. P., 452, 462 Nisbett, R. E., 527, 557, 560 Nachmias, O., 685, 686 Nishida, T., 590, 606, 612 Nadel, L., 230, 810 Nissen, M., 158 Nadon, R., 463 Noble, W., 582 Nagata, M., 232 Noe,¨ A., 80, 740, 759, 867, 868 Nagel, E., 140 Nofzinger, E. A., 441 Nagel, T., 10, 12, 39, 74, 406, 864, 866 Nogrady, H., 460 Nagle, M., 335 Nolan, R. P., 454 Nahmias, C., 465 Nolde, S. F., 814, 819 Nakamura, Y., 795 Norcia, A. M., 720, 761 Nakash-Dura, M., 307 Nordlie, J. W., 456 Nakayama, K., 230 Norman, D. A., 186, 341, 342, 343 Nanez, J., 215 Norman, J., 563 Napper, E., 113 Norman, K. A., 278, 811, 818, 820, 821, 822 Narayan, M., 613 Noseworthy, M. D., 414 Narens, L., 289, 290, 291, 292, 293, 298, 299, 300, 301, Noterman, S. L., 790 305, 308–309, 454 Nouriani, B., 463 Nash, J. K., 457 Nowak, M., 541 Nash, M. R., 456, 460, 464 Nunez, P. L., 532, 534, 737 Natsoulas, T., 61, 342, 356, 776 Nunn, J. F., 711 Nattkemper, D., 339, 340 Nussinson, R., 315 Naveh, I., 160 Nyberg, L., 811, 812 Neal, A., 154 Neander, K., 61 O’Brien, G., 146, 155 Negassi, B., 584 O’Connell, D. N., 446 Negatu, A., 166 O’Conner, S. J., 427 Neidle, C., 587 O’Connor, L. E., 680 Neisser, U., 261, 262, 311 O’Connor, R. J., 618 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

author index 929

O’Daly, G., 16 Panksepp, J., 76, 381, 389, 526, 688, 740, 778, 796, 856 O’Doherty, J., 543, 840 Pansky, A., 289–290, 310, 311, 313 O’Donnell, S., 414 Panzeri, S., 853 O’Keefe, J., 533 Pare, D., 742 O’Leary, J. L., 791 Parent, A., 795, 796 O’Neill, D. K., 387, 577, 608 Paris, S. G., 291 O’Neill, M. A., 746, 747 Parker, S. T., 155, 607 O’Nell, C., 633 Parkin, A. J., 257, 259, 266, 268 O’Regan, J. K., 80, 867, 868 Parnas, J., 61, 481, 526 O’Regan, K., 759 Partridge, J. C., 406 O’Reilly, R. C., 160, 161, 818 Parvizi, J., 76, 757 Oakley, D. A., 451 Pashler, H., 898 Oatley, K., 377, 380, 384, 385, 386, 390, 396, 397, Pasik, P., 718 832, 836 Pasik, T., 718 Obstoj, I., 459 Passchier, J., 714 Ochsner, K., 261, 269, 466–467 Passingham, R., 587, 796 Odel, J. G., 790 Pate, J. L., 780 Ojemann, J. G., 231, 812 Patteri, I., 341 Okado, Y., 819 Patterson, F., 573 Okayama, A., 795 Pattie, F. A., 456 Olds, D. D., 690 Patton, R., 789 Oliver, W., 159 Paulhan, F., 390 Olivier, A., 789 Paulignan, Y., 334, 783 Olivier, C., 485 Pauls, J., 712 Olson, D. R., 384, 391 Paus, T., 414, 463, 464, 711, 722 Olsson, A., 232, 233, 234, 235 Pavlov, I. P., 779 Olsson, M. J., 293 Pawlik, G., 787 Olsson, T., 417 Payne, B. K., 310 Omori, M., 525 Payne, D. G., 277 Ooi, T. L., 528 Peacock, J. L., 633, 645–646 Opie, J., 146, 155 Peacock, W. J., 789 Oppenheimer, D. M., 297 Peacocke, C., 45 Oppenheimer, S., 584, 585 Pearl, D. K., 330, 331, 783, 795 Orem, J., 788 Pearlman-Avnion, S., 310 Orne, M. T., 446, 448, 449, 452, 453, 454, 459, 462 Pedroarena, C., 201, 527, 742 Ortells, J. J., 216 Peery, S., 333 Oschner, K., 900 Peigneux, P., 715 Osipov, G., 750 Peled, S., 712 Ossi, R., 793 Pelisson,´ D., 334, 335 Otto-Salaj, L. L., 463 Pellegrini, A. D., 604, 606 Overton, W. F., 425 Pellizzer, G., 572 Owen, A. M., 715 Pelphrey, K. A., 871 Oxbury, J., 239 Pembrey, M, E., 587 Oxbury, S., 239 Penfield, W., 191, 782, 783 Pennebaker, J. W., 394 Paäbo,¨ S., 583, 584 Pennington, N., 396 Pace-Schott, E., 464 Penrose, R., 118, 140, 152, 781, 890, 892 Pachoud, B., 68, 84 Percer, J. M., 277 Pacini, R., 313, 314 Percheron, G., 786 Pagano, R. R., 500, 533, 535 Pereg, D., 681 Paillard, J., 718 Perenin, M., 336, 337 Paladini, R., 240, 241 Peretti, C. S., 494 Palafi, T., 412 Perfect, T. J., 260, 265, 289–290, 296, 307, Palesh, O. G., 681 487 Palfai, T., 410, 423, 425 Perkel, D. H., 749 Palmer, C. T., 615 Perlis, D., 128 Palmer, F. R., 370 Perner, J., 291, 407, 418, 419, 421, 870 Palmer, J. A., 609 Perrett, D. I., 228, 229, 872 Palmer, L. K., 609 Perriloux, H. K., 608 Palus, M., 733 Perruchet, P., 414, 783 Panigraphy, A., 790, 792, 794 Perry, C., 455, 460 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

930 author index

Pessoa, L., 722 Posada, A., 339, 345 Peters, E., 293, 296 Posner, J. B., 715, 790, 794 Peters,´ J., 709, 710 Posner, M. I., 199, 290, 293, 342, 457, 460–461, 557, Peters, L. G., 648, 654 564, 792 Petersen, S. E., 231, 709, 811, 812, 814, 817, 820 Post, R. B., 456 Peterson, J., 383 Potter, E. K., 330 Peterson, M. A., 457 Potter, M. C., 721 Peterson, R. A., 489, 493 Poulin, R., 618 Peterson, T., 157, 158, 159, 160, 161 Pourdehnad, M., 541 Petersson, K. M., 811 Povinelli, D. J., 387, 421, 422, 572, 577, 578, 580, 606, Petit, H., 787 608, 609, 615, 617, 622, 870 Petitmengin, C., 756 Powell-Moman, A., 303 Petitot, J., 68, 84 Power, R., 385 Petrides, M., 849 Poyurovsky, M., 310 Petrusic, W., 457 Prablanc, C., 334, 335 Petty, R. E., 558, 677 Pradere, D., 269, 817, 821 Pfurtscheller, G., 537 Praslov, N. D., 584 Phelps, E. A., 789, 873 Pratkanis, A. R., 208 Phelps, M. E., 420 Premack, D., 573, 574, 577, 870 Phillips, L. D., 303, 306 Pressley, M., 289, 295 Piaget, J., 355, 407, 408, 414, 421 Pressman, S., 524 Piazza, M., 231 Preuss, T. M., 605 Pickering, T., 582 Pribram, K. H., 190, 191, 678, 905 Pierce, T., 685 Price, C. J., 587 Pierrard, E., 718 Price, D. D., 450, 463, 464, 465 Pietrini, P., 872 Price, M., 335 Pike-Tay, A., 584 Price-Williams, D., 648 Pikovsky, A., 750 Priestly, J., 23 Pillas, D. J., 789 Prigogine, I., 732, 734, 758 Pimm-Smith, M., 188–189 Prinz, W., 329, 330, 331, 332, 339, 340, 341, 560, 562, Pinard, A., 408 563 Pinel, E. C., 688 Prochazka, A., 328, 329, 343 Pinel, P., 231 Proudfoot, W., 500 Pinker, S., 582, 583, 587, 588, 609, 615, 851 Przeworski, M., 587 Pins, D., 720 Przybeck, T. R., 689 Pinsk, M. A., 722 Ptito, A., 793 Pinto, D., 737 Ptito, M., 793 Pisella, L., 333, 338 Puce, A., 228, 230 Pistole, D. D., 451 Purpura, K. P., 712, 788 Pitts, G., 216, 219 Putnam, H., 56, 59, 890 Plata-Salaman, C. R., 843–844 Pylyshyn, Z., 145 Plettenberg, H. K., 712, 714 Pyzik, P. L., 789 Pleydell-Pearce, C. W., 493 Ploog, D., 586 Quamme, J. R., 269, 817 Plooij, F. X., 617 Quartz, S., 650 Plourde, G., 711, 712 Quillian, M. R., 159 Plum, F., 714, 715, 788, 790, 792, 794 Quinn, N., 650 Plunkett, K., 853 Plutchik, R., 688 Raab, E., 718 Pockett, S., 795 Rabbitt, P. M., 309 Poeppel, E., 337 Rabin, B. S., 524 Pojoga, C., 453 Rabinovich, M. I., 731, 734 Poldrack, R. A., 228 Rabinowitz, J. C., 295, 296 Polich, J., 427 Rabten, G., 113 Poline, J., 230, 231 Racine, C., 817, 821 Pollmacher,¨ T., 710 Rackl, A., 790 Polster, M. R., 820 Radtke, H. L., 452 Poltrock, S. E., 360 Rafal, R. D., 228, 241, 783 Poppel,¨ E., 113, 237, 362, 712 Rahula, W., 113 Porjesz, B., 427 Raichle, M. E., 231, 529, 709, 812 Portafaix, M., 790 Rainville, P., 71, 450, 463, 464, 465 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

author index 931

Rajah, M. N., 191 Reyna, V. F., 297, 310 Rajaram, S., 253, 254, 256, 258, 259, 260, 261, 264, Reynolds, J. H., 533, 537, 751, 779 265, 266, 267, 268, 269, 270, 273, 274, 276, 277, Reynolds, V., 590, 606, 612, 621 278, 279 Reznick, J. S., 415, 422 Rakoczy, H., 392, 579 Rhodewalt, F., 682 Ralston, H. J., 406 Ribary, U., 201, 527, 715, 742, 747, 794 Ramachandran, V. S., 345 Ricard, M., 530, 531, 538 Ramamurthy, U., 165 Riccardi, L. M., 905 Ramponi, C., 266, 270, 810, 818 Ricci, C., 720 Ramsoy, T., 195 Rice, H. J., 814, 815 Randolph, C., 489, 490, 493 Richards, D. R., 299 Ranganath, C., 817 Richards, G. D., 583, 585, 637 Rao, S. M., 819 Richardson, S., 652 Rapaport, D., 675 Richardson-Klavehn, A., 252, 259, 266–267, 270, 271, Rapoport, J. L., 414 274, 275, 293, 483, 810, 818 Rappaport, R., 651 Rick, J., 618 Rappold, V. A., 272, 274 Riegler, G. L., 274 Rasmussen, S., 657 Rilling, J. K., 603 Rasmussen, T., 789 Rink, W. J., 584 Ratner, H., 607 Rittenauer-Schatka, H., 296 Rauch, S. L., 228, 229, 231, 234, 812, 815 Ritter, F. E., 298, 299 Raudsepp, J., 266 Ritter, W., 817 Rawlings, N. B., 530, 531, 538 Riutort, M., 494 Rawson, K. A., 304 Rizzolatti, G., 80, 341, 388, 392, 563, 587, 740, 783, Ray, W. J., 463, 464 874 Raye, C. L., 262, 263, 467 Rizzuto, D. S., 533 Raymont, P., 61 Ro, T., 783 Raz, A., 457 Robb, W. G., 815 Read, D. J., 307 Robert, P., 310, 484, 485, 494 Read, S., 722 Roberts, N., 818 Reason, J. T., 343 Robertson, D. A., 307 Reaux, J. E., 608 Robertson, E., 466–467 Reber, A. S., 156, 160, 453 Robinson, A. E., 298 Reber, R., 291 Robinson, D. L., 787 Rechtscaffen, A., 709 Robinson, M. D., 298, 307 Rectern, D., 718 Rochat, P., 415 Redd, W., 454 Roche, S. M., 447 Reddy, R., 195 Rock, I., 199, 213, 224 Reder, L. M., 289, 293, 298, 299, 309, 314, 454 Rodeiger, H. L., 814 Reed, J. M., 817 Rodemer, C. S., 456 Reeder, J. A., 257, 818 Rodman, H. R., 793 Rees, G., 202, 228, 230, 718, 720, 722, 758 Rodriguez, E., 202, 533, 739, 749, 750, 751 Regard, M., 789 Rodriguez, I. A., 579 Regehr, G., 155 Roebers, C. M., 310 Register, P. A., 460, 463 Roediger, H. L., 168–169, 253, 256, 259, 260, 263, 266, Regli, F., 790 267, 268, 270, 271, 272, 273, 274, 276, 277, 810, 818 Reich, T., 427 Roelfsema, P. R., 723, 751, 779 Reiman, E., 818, 820 Roelofs, C., 333 Reingold, E. M., 214, 220, 222, 235 Roepstorff, A., 2, 760 Reisenzein, R., 855 Rogers, L., 193, 194 Reiss, D., 607 Rogers, T. T., 522 Rektor, I., 233 Rohrbaugh, J. W., 427 Renault, B., 527, 738, 750, 751, 761 Roitberg, B. D., 618 Rengo, C., 584 Rolls, B. J., 848 Renne, P. R., 585 Rolls, E. T., 783, 831, 832, 834, 835, 836, 838, 839, Rensink, R. A., 720, 867 840, 841, 842, 843, 845, 846, 847, 848, 849, 852, Reuter-Lorenz, P. A., 789 853, 855 Revonsuo, A., 169, 447, 753, 776 Rorie, A. E., 533, 537, 751, 779 Rey, G., 61, 136, 290, 316 Rosa, M. G., 237 Reyher, J., 458, 462 Rosch, E., 113, 500 Reyna, S., 653 Rosen, B. R., 813, 815, 819 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

932 author index

Rosen, I., 232, 233, 234, 235 Sainte-Rose, C., 789 Rosen, M. A., 406 Salame, P., 494 Rosenbaum, R. S., 717 Salenius, S., 874 Rosenberg, J. S., 607 Sampson, H., 680 Rosenblatt, B., 681 Sanchez, I. C., 579 Rosenbloom, P., 155 Sandblom, J., 811 Rosenblum, M., 750 Sanders, M. D., 237, 336, 418, 718 Rosenkranz, M. A., 521, 522, 524 Sandifer, P. H., 794 Rosenquist, A. C., 788 Sandler, J., 681 Rosenthal, D. M., 46, 48, 54, 59, 60, 292, 418, 419, Sandler, S. P., 294, 304 724, 783, 841, 852, 855 Sandner, G., 484 Rosenthal, R., 461, 685, 871 Sanocki, T., 304 Roskies, A. L., 737, 749, 751 Santanna, J., 541 Ross, L., 557 Santorelli, S. F., 521, 522, 524 Rossetti, Y., 333, 334, 337, 338 Sanvito, J., 296, 298, 301 Rostrup, E., 711 Sapir, E., 640 Rotello, C. M., 257, 818 Sapir, S., 790 Roth, W. T., Sarauw, D., 606–607 Rothbart, M. K., 792 Sarbin, T. R., 447, 461, 465 Rothwell, J. C., 328, 329, 343 Sargent, C., 200, 203 Rotte, M., 813, 815 Sartre, J. P., 73 Rousseaux, M., 787 Sasaki, Y., 215 Rowland, D., 228, 229, 872 Sass, K. J., 332 Rowlands, M., 869 Sass, L. A., 481 Roy, J., 68, 84 Sauer, T., 533 Royce, G. J., 788 Saunders, S., 127 Royet, J. P., 733, 757 Sauve, M. J., 269, 817 Rozin, P., 271 Savage, C. R., 231, 812, 815 Rubin, D. B., 461 Savage-Rumbaugh, S., 573, 585, 586, 607 Rubin, D. C., 261, 262 Scepansky, J. A., 676, 690, 694 Rubinstein, R. A., 652 Schacter, D. L., 168, 169, 185, 221, 231, 252, 253, 267, Ruby, P., 733, 757, 874 269, 272, 273–274, 278, 406, 450, 451, 452, 453, Rudrauf, D., 533, 747, 762 466, 717, 783, 797, 810, 811, 812, 813, 814, 815, 817, Ruffman, T., 421 818, 819, 820, 821, 822 Rugg, M. D., 186, 811, 812, 813, 815, 816, 817, 820 Schacter, S., 855 Rumbaugh, D. M., 780 Schafer, R., 679 Rumelhart, D. E., 157, 159, 376, 650 Schall, J. D., 713, 719, 722 Rumsey, A., 641 Schaltenbrand, G., 785 Runions, K., 408 Schank, R. C., 376, 396 Russ, E., 682 Scharf, D., 680 Russell, B., 28, 355, 574 Scharf, J. S., 680 Russell, D. P., 720 Scheﬂin, A. W., 460 Russell, L. M., 681 Scheibe, K. E., 456 Russo, R., 268 Scheibel, A. B., 792 Ryan, R. M., 427, 557 Scheibel, M. E., 792 Ryff, C. D., 524 Scheid, K., 291 Ryle, G., 643 Scheier, M. F., 557 Schenck, C. H., 708, 710 Saad, C. L., 457 Scheutz, M., 129 Sabourin, M., 456 Schiff, N. D., 712, 715, 788, 792, 794 Sackeim, H. A., 456, 463 Schiff, S. J., 533, 732 Sadikot, A. F., 795 Schimek, J. G., 677, 680 Sadler, P., 447, 461, 464 Schlag, J., 788 Sadun, A., 778, 789 Schlag-Rey, M., 788 Saetti, M. C., 334 Schlaghecken, F., 216, 217 Saffran, J. R., 522 Schloerscheidt, A. M., 820 Sagiv, N., 228 Schlotterbeck, E., 216 Sahakyan, L., 309, 310 Schmidt, S. R., 260, 523 Sahlins, M., 637 Schmidt, T., 335 Sahraie, A., 237, 720 Schmidtling, E. Y., 607 Saint-Cyr, J. A., 484 Schmolck, H., 262 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

author index 933

Schneider, G. E., 337 Shallice, T., 169, 186, 187, 341, 342, 343, 816, 820, Schneider, W., 159, 188–189, 289, 295, 307, 310, 849 460–461, 557 Shanker, S. G., 573, 585, 586 Schnitzler, A., 532 Shannon, Benny, 3 Schnyer, D. M., 810, 820 Shapiro, D. H., 500, 533, 534, 537 Scholl, B. J., 213, 872 Shapiro, T., 457 Schooler, J. W., 406, 821 Shapiro, W., 686 Schopenhauer, A., 24 Sharf, R. H., 500 Schott, B., 790 Sharma, K., 307 Schotterbeck, E., 216 Shaughnessy, J. J., 296, 301, 304 Schouten, J. L., 872 Shaver, P. R., 681, 685–686 Schreiber, V., 306 Shaw, R. J., 296 Schreibner, T. A., 300 Shear, J., 71, 756, 760 Schreiner, L. H., 789 Shearer, J., 648, 652 Schubotz, R. I., 341 Shedler, J., 681, 682 Schuh, E. S., 218, 234 Sheehan, P. W., 449, 459, 460, 462, 463, 465 Schulze-Bonhage, A., 533 Sheffer, D., 293 Schumacher, J., 521, 522, 524 Sheffer, L., 302, 303, 304 Schuman, M., 533 Sheline, Y. I., 522 Schunn, C. D., 298, 299, 314 Shelley-Tremblay, J., 232 Schurhoff, F., 484 Sheridan, J. F., 521, 522, 524 Schuri, U., 790 Sherman, J. W., 262 Schuster, B. G., 501 Sherman, S. J., 300, 561 Schwarcz, H. P., 584 Sherman, S. M., 747, 782, 787, 792 Schwartz, B. L., 291, 298, 299, 301, 306, 308–309, 579, Sherover, C., 815 Sherrington, C. S., 794 Schwartz, G. E., 783 Shevrin, H., 229, 233, 693 Schwartz, J., 899, 900 Shields, W. E., 160 Schwarz, L. M., 455 Shiffrin, R. M., 187, 460–461, 557, 563, 565 Schwarz, N., 291, 293, 296, 297, 313, 314, 315 Shima, K., 796 Schwarzbach, J., 335 Shimamura, A. P., 272–273, 289, 453, 454 Schweinberger, S., 240 Shoben, E. J., 273 Schweizer, H. R., 458 Shoemaker, S., 58, 59 Schweizer, P., 113 Shor, R. E., 446, 447, 451, 462 Schyns, P., 873 Shore, B., 651 Scott, T. R., 843–844, 845 Shulman, G. L., 709, 722, 822 Seagal, J. D., 394 Shweder, R., 653, 654 Seager, W., 16, 61, 783, 796 Shyi, G. C., 262 Searle, J., 60, 61, 62, 72, 131, 141, 142, 146, 147, 152, 243, Sichel, J. P., 484, 490 386, 598, 757, 776 Sidebe,´ M., 795 Sebald, W. G., 397 Sidis, B., 208, 213 Seelig, D., 533 Siegel, M., 873 Segal, J. B., 811, 820 Siegel, P., 684 Segall, M. H., 633 Sienkiewicz, Z. J., 843–844 Seger, C., 153–158, 160, 453 Siewert, C. P., 54 Seidman, L. J., 310 Sigel, I., 410 Seitz, R. J., 563 Silananda, U., 504, 505 Sejnowski, T. J., 650, 788 Silberstein, R. B., 532, 534 Selbie, S., 441, 709, 710 Silk, K. R., 681 Sellars, W., 107 Silverman, L. H., 684 Selten, R., 688 Simmel, M., 872 Serclerat, O., 790 Simmons, A., 465, 720 Sergent, C., 744 Simon, D. A., 301 Seron, X., 718 Simon, H. A., 142, 688 Servan-Schreiber, E., 155 Simons, A., 296 Sevier, M., 680 Simons, D. J., 210–211, 213, 215, 227, 234, 235, 720, Sewards, M. A., 753 755, 759 Sewards, T. V., 753 Simons, J. S., 815 Seyfarth, R. M., 841 Simpson, G. V., 536 Seymour, B., 543 Singeman, J. D., 871 Shackelford, T. K., 602 Singer, A. K., 754 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

934 author index

Singer, B. H., 524 Spellman, B. A., 305 Singer, J. E., 855 Spence, S. A., 693 Singer, T., 527, 532, 543 Spencer, D. D., 332 Singer, W., 718, 723, 733, 737, 743, 751, 753, 779 Spencer, H., 639 Singh, B., 536 Spencer, S. S., 332 Sipprelle, C. N., 457 Sperber, D., 653 Sirevaag, E. J., 427 Sperry, R. W., 780, 789 Skarda, C. A., 735, 737 Spiegel, D., 450, 454, 463, 465, 466 Skinner, B. F., 572, 574 Spiers, H. J., 522 Skinner, J. E., 792 Spinnler, H., 343 Skolnick, B., 464 Spinoza, B., 20–21 Skowronski, J. J., 561 Spiro, M., 655 Slagle, R. W., 465 Spitzer, M., 232 Slamecka, N. J., 259, 268 Squire, L. R., 231, 252–253, 262, 269, 272–273, 418, Sletten, I., 454 453, 797, 810, 811, 812, 817, 822, 839, 842 Sloman, A., 128, 129, 130, 134, 160 Srinivas, K., 268 Sloman, S. A., 313 Srinivasan, R., 532, 534, 720 Slotnick, S. D., 534, 811, 814, 818, 819, 820, 821, 822 Srull, T. K., 561 Slovic, P., 293, 296, 302, 309 Srypstein, E., 718 Smart, J. J., 31, 41 Stadler, M. A., 819 Smilek, D., 209, 224 Stadler, M. L., 274 Smirnov, Y., 582 Staib, L. H., 613 Smith, A. P., 789, 811 Stam, H. J., 455 Smith, B., 129, 130, 272 Stamenov, M. I., 356 Smith, D. W., 46, 50, 51, 61 Standing, L., 183 Smith, J. D., 160 Stanger, C., 393 Smith, L. B., 424 Staniford, P., 648 Smith, M. E., 817 Stanley, S., 456 Smith, M. J., 489, 490, 493 Stanley, W., 153–154, 158 Smith, S. M., 299 Stanovich, K. E., 313 Smith, W. L., 789 Stapp, H., 896, 899, 900 Smith, Y., 795 Star, C. E., 269 Smith, Y. M., 238 Stark, C. E., 810, 819 Smith-Swintowsky, V. L., 843–844 Stark, H. A., 272 Smolensky, P., 157 Stark, L., 335 Smyth, L., 462 Stein, E., 82 Smythies, J., 787 Stein, S. D., 790 Snell, B., 395 Steinbock, A., 69 Snellgrove, D. L., 506 Steinke, W., 790 Snodgrass, J. G., 272, 817 Stemach, G., 334, 718 Snodgrass, M., 233 Stenberg, G., 232, 233, 234, 235 Snyder, C. R., 293, 460–461, 557 Stephan, K. E., 746, 747 Snyder, M., 461 Stephens, C. D., 648 So, P., 533 Steriade, M., 710, 782, 787, 790, 791, 793 Sobo, E. J., 657 Sterkin, A., 736 Soffer, O. J., 584 Stern, D. B., 425, 655 Sokoloff, L., 529 Stern, J. A., 454 Sokolowski, R., 70 Stern, W., 406 Solecki, R. S., 582 Sternberg, R. J., 290 Solms, M., 441 Stevens-Graham, B., 780 Sommerville, J. A., 391, 416, 422 Stewart, K., 584 Son, L. K., 292, 308 Stickgold, R., 438, 464 Sonntag, P., 485 Stief, V., 240 Sorrow, D. L., 682 Stijnen, T., 714 Southwick, S. M., 613 Stoerig, P., 237, 338, 707, 718, 723, 724, 793, 794 Spangenberg, E. R., 208 Stoet, G., 340 Spano, M. L., 732 Stokoe, W. C., 587 Spanos, N. P., 446, 447, 449, 452, 454, 455, 456, 457, Stoller, P., 633, 656, 657 459, 461, 465 Stout, D., 611 Spehn, M. K., 293 Stoyva, J., 449 Spelke, E. S., 618 Strack, F., 295, 296, 297, 301, 313, 314 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

author index 935

Strauss, C., 650 Tardif, T., 620 Strawson, Galen, 379 Target, M., 681 Street, M., 584 Tarrier, N., 693 Stretch, V., 818 Tataryn, D. J., 447, 451, 455, 459, 465, 674 Striano, T., 415 Tate, A., 118 Stringer, C. B., 583, 584, 585 Tatemichi, T. K., 790 Stringer, S. M., 836 Tattersall, I., 581, 585, 589, 602 Stroffolino, P. J., 299 Tatu, L., 790 Strogatz, S. H., 746 Tawia, S. A., 406, 407 Stroker,¨ E., 661 Taylor, J. G., 154, 169, 776, 777 Strongman, K. T., 832 Taylor, J. L., 783, 796 Stroop, J. R., 460 Taylor, S. E., 556, 558 Stuart, C., 905 Taylor, T. J., 573, 585, 586 Studdert-Kennedy, M., Taylor, T. L., 335 Stuss, D. T., 254, 577, 790 Tebecis, A. K., 535 Suchman, L., 867 Tegmark, M., 893 Sudarshan, E., 897, 907 Teixidor, P., 607 Suddendorf, T., 577, 606–607 Tellegen, A., 446, 689 Sudia, S., 778, 789 TenHouten, W., 653 Suengas, A. G., 262, 263 Tenpenny, P., 273 Suffczynski, P., 537 Terao, T., 451 Sugar, O., 789 Terriault, D., 312 Sugiyama, Y., 590, 606, 612 Tetlock, P., 688 Sullivan, A. L., 821 Teuber, H. L., 237, 717 Sullivan, G. D., 380 Thagard, P., 674, 690, 691, 692, 693 Sullivan, M. W., 393 Thau, M., 58 Sulloway, F. J., 690 Theall, L. A., 608 Sun, R., 152, 154, 155, 156, 157, 158, 159, 160, 161, 162, Theide, K. W., 304 170 Thein, R. D., 560 Suprijo, A., 584 Thiede, K. W., 308, 312 Sussman, G., 382 Thiel, T., 712 Sutcliffe, J. P., 448, 456 Thomasson, A. L., 61 Sutherland, A. E., 425 Thompson, D. M., 277, 561 Sutich, A., 648 Thompson, E., 46, 71, 77, 78, 80, 82, 84, 113, 500, 528, Suwa, G., 583, 585 532, 733, 739, 740, 758, 762, 868 Svrakic, D. M., 689 Thompson, K. G., 722 Swann, W. B., 461 Thompson, W. L., 450, 465, 466 Swanson, P. D., 790 Thorne, W. A., 453, 467 Swartz, K. B., 606–607 Thornhill, R., 615 Swets, J. A., 211 Thorpe, S., 756 Swick, D., 228 Thrangu, 502, 507, 508, 511, 512, 513, 515, 517, 519 Swisher III, C. C., 584 Thrash, T. M., 694 Sykes, R. N., 308–309 Throop, C., 633, 638, 639, 642, 643, 644, 645–646, Symons, D., 588 647, 654, 656, 657, 658, 660 Synnott, A., 657 Thurow, M. E., 522 Szechtman, H., 464, 465 Thurstone, L. L., 556 Szelies, B., 787 Tibbits, R., 452 Szentagothai,´ J., 735 Tice, D. M., 558 Szirmai, I., 794 Tienson, J., 54 Timm, H. W., 460 Taillanter, D., 308 Tinbergen, N., 834 Takahashi, T., 525 Titchener, E., 637, 638 Takahashi, Y., 905 Toates, F., 159 Talarico, J. M., 262 Tobias, B. A., 447, 458 Tallon-Baudry, C., 537, 749, 750, 751, 752, Tobias, P. V., 603 753 Tobimatsu, S. J., 795 Tambiah, S., 633 Tobis, I. P., 458 Tangney, J. P., 618 Toglia, M. P., 277 Taniwaki, T., 795 Toland, John, 23 Tanji, J., 796 Tomasello, M., 387, 388, 392, 572, 578, 579, 580, 590, Tanner, R., 621 606, 607, 608, 619, 870 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

936 author index

Tomkins, S. S., 688 Usher, M., 300 Tong, F., 230 Uske, A., 790 Toni, I., 334 Tononi, G., 155, 191, 192, 196, 201, 406, 527, 532, 533, Vaidya, C. J., 811, 820 720, 741, 742, 746, 747, 748, 761, 780 Vaillant, G. E., 678, 690 Tooby, J., 158, 590, 599, 600, 609, 653 Van den Broek, P., 387 Torres, F., 417 Van der Lubbe, R. H., 216 Toth, J. P., 260, 272, 274 Van Domburg, P. H., 790 Toth, N., 582 Van Eijk, R., 487 Tovee, M. J., 840, 849, 853 Van Gelder, T. J., 732, 733, 736 Towbes, L. C., 685 Van Gulick, R., 51, 60, 61 Trabasso, T., 387 Van Hoesen, G. W., 790 Tramo, M. J., 789 Van Reekum, C. M., 522 Tranel, D., 229, 790, 872, 873, 874, 876 Van Rullen, R., Trapnell, P. D., 447 Varela, F. J., 68, 71, 78, 84, 113, 202, 500, 527, 529, Traub, R. D., 532 532, 533, 732, 736, 737, 739, 740, 746, 749, 750, Travis, F., 535 751, 755, 756, 757, 758, 760, 763, 868 Treanor, J. J., 524 Varga, M., 441, 709, 710 Treisman, A. M., 209, 751, 753 Vargha-Khadem, F., 587, 588 Tremoulet, P. D., 872 Varley, R., 873 Treue, S., 792 Vaughan, J. T., 230 Trevarthen, C. B., 337, 406, 789 Vedeniapin, A. B., 427 Trevena, J. T., 795 Veith, I., 450 Treves, A., 836, 838, 839, 840, 853 Vella, S., 710 Trevethan, C. T., 237 Velleman, J. D., 383 Trivers, R. L., 580, 613, 850, 851 Velleman, R., 380 Troetschel, R., 562 Velmans, M., 54, 795 Trope, Y., 158, 306, 313, 557 Velonova, K., 811 Trott, C. T., 817 Verfaellie, M., 239, 240, 241, 268, 269, 817, 818, Tsapkini, K., 314 821 Tsongkhapa, 506, 507, 508, 511, 512, 513 Vergeler, R., 216 Tsuda, I., 735 Verleger, R., 216, 710 Tucker, D. M., 464, 534 Vermersch, P., 71, 763 Tulving, E., 252, 253, 254, 263, 267, 271, 272, 276, Vernon, D., 300 277, 294, 314, 415, 421, 452, 481, 482, 488, 561, Viana Di Prisco, G., 779 576, 577, 598, 717, 797, 810, 811, 812, 813, 816, 817, Viard, A., 484 822 Videen, T. O., 231, 812 Tunnicliff, J., 302 Vighetto, A., 336 Turing, A., 121, 140 Viksman, P., 310 Turner, E., 633, 648 Villemure, J. G., 789 Turner, S. R., 384 Vining, E. P., 789 Turner, V., 646, 647, 653 Vinski, E., 296 Tuszynski, J., 893 Vinter, A., 414, 783 Tutin, C. E., 606, 612 Visser, T., 222, 225, 227, 244 Tuvia, R., 298 Vogel, F., 427 Tversky, A., 304, 306, 688 Vohs, K. D., 682 Tweedale, R., 237 Volpe, B. T., 239 Tye, M., 40, 43, 45 Von Cramon, D. Y., 337, 341 Tylor, E., 639 Von der Malsburg, C., 746, 750 Von Helmholtz, H., 380, 673 Uecker, A., 820 Von Monakow, C., 794 Uleman, J. S., 561 Von Neumann, J., 887 Ulett, G. A., 454 Von Restorff, H., 261 Umezawa, H., 905 Von Stein, A., 533 Umilta,` C., 1 Vonk, J., 870 Umilta,` M. A., 328, 341, 342 Vorberg, D., 335 Underwood, B. J., 296 Vos, J. E., 534 Ungerleider, L. G., 337, 722 Vriezen, E., 253, 269, 810 Unkelbach, C., 306 Vuilleumier, P., 228, 229, 872 Urbanowski, F., 521, 522, 524 Vygotsky, L. S., 382, 393, 407, 408, 426, 617, Urry, H. L., 522 653 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

author index 937

Wada, Y., 525 Wessinger, C. M., 237 Wade, E., 295 West, M. A., 500, 533, 534 Wadlington, W. L., 456 West, R. F., 313 Wagner, A. D., 268, 269, 813, 814, 815, 819 Westdrop, A. F., 534 Wagner, U., 710 Westen, D., 674, 678, 680, 681, 682, 683, 686, 688, Wahren, W., 785 690, 692, 693, 694 Wakeﬁeld, J. C., 602 Westmacott, R., 717 Walach, H., 523 Westwood, D. A., 810–811 Waldron, W., 113 Wetter, T. C., 710 Waldvogel, H. J., 785, 796 Wetzel, R. D., 689 Walker, A., 582 Wexler, B. E., 783 Walla, P., 820 Whalen, D. H., Wallace, B., 465, 467, 500 Whalen, P. J., 228, 229, 234 Wallace, R. K., 504, 534, 535, 537 Wharton, E., 369 Walsh, R. N., 500, 533, 534 Wheatley, T. P., 688 Walshe, F. M., 775 Wheatstone, C., 230 Walter, R. C., 584 Wheeler, M. A., 254, 259, 406–407, 421, 577 Wang, Z., 414 Wheeler, M. E., 811, 813, 814, 817, 820 Wangchug Dorje,´ 501–502, 505, 507, 508, 511, 512, Wheeler, P., 605 513, 515, 517, 518, 519 White, G., 633 Ward, L. M., 535 White, N. S., 711 Warrenburg, S., 500, 533, 535, 783 White, R. J., 789 Warrington, E. K., 237, 252, 253, 272, 336, 717, 718, White, R. W., 448, 584 810–811 White, T. D., 582, 583, 585 Washburn, D. A., 160 Whitehead, A. N., 29 Wason, P., 583 Whitehouse, K., 301 Watanabe, T., 215 Whiten, A., 578, 590, 597, 605, 606–607, 608, 612 Watkins, K. E., 587 Whitlow, J. W., 777 Watson, J. B., 409, 419, 572, 574, 611, 620 Whitten, A., 870 Watson, J. M., 821 Whittington, M. A., 532 Watt, D. F., 796 Whittlesea, B. W., 297, 314, 315 Watt, I., 377, 396 Whorf, B. E., 640 Watts, D. J., 746 Wicker, A. W., 559 Way, E., 650 Wicker, B., 733, 757 Weatherill, R., 681 Widarman, K. F., 269, 817 Webber, M., 643, 648 Wider, K. V., 113, 526 Weder, B., 710 Widiasmoro, 584 Weekes, N., 778, 789 Widing, E. L., 817 Wegner, D. M., 54, 271, 315, 329, 330, 341, 461, 556, Wiebe, V., 587 558 Wielapp, P., 710 Weinberg-Eliezer, A., 311 Wierzbicka, A., 359 Weinberger, D. R., 484, 489, 490, 493 Wiesel, T. N., 712 Weinberger, J., 467, 673, 680, 683, 684, 685, 686, Wieser, H. G., 789 693 Wiest, G., 790 Weiskrantz, C. M., 237, 239 Wiggins, J. S., 447 Weiskrantz, L., 252, 253, 272, 336, 342, 418, 717, 718, Wiggs, C. L., 811, 814, 820 720, 759, 793, 794, 810–811, 832 Wightman, R. M., 780 Weiss, F., 458 Wijesinghe, R. S., 534 Weiss, J. N., 680, 732 Wilcox, S. E., 587 Weiss, M., 393 Wild, B., 306 Weitzenhoffer, A. M., 461 Wildschiødtz, G., 541, 711 Weldon, H. L., 253 Wilenius-Emet, M., 753 Weldon, K. L., 618 Wilke, M., 306 Weldon, M. S., 272, 273, 274 Wilkes, K., 9, 129, 132 Weldon, S. M., 267, 268 Wilkinson, I. D., 693 Wellman, H. M., 290, 409, 419, 611, 619, 620 Willard-Schroeder, D., 484, 490 Wells, G. L., 307 Williams, B., 396 Wenger, M. A., 533, 537 Williams, C. J., 559, 560 Wertheim, N., 790 Williams, E. M., 600 Wesensten, N. J., 441, 709, 710 Williams, P., 505 Wessels, P. M., 819 Williams, S. C., 465, 720 P1: JzG 0521857430aind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:36

938 author index

Williamson, P., 332 Yamada, 710 Williford, K. W., 61 Yang, C., 232 Willingham, D., 158 Yaniv, I., 295, 313 Wilson, F., 28 Yasuda, Y., 790 Wilson, G. M., 375 Yasue, K., 905 Wilson, M., 599, 615, 621 Yates, R., 584 Wilson, T. D., 271, 527, 560, 688 Yaxley, S., 843–844 Wimmer, H., 870 Ye,H.H.,414 Wimsatt, W. C., 780 Yellen, J. E., 584 Wineland, D., 897, 907 Yi, D., 721, 722 Wingeier, B. M., 532, 534 Yingling, C. D., 792 Wingrad, T., 129 Yokoyama, R., 790 Winkelman, M., 633, 634, 648, 653, 660 Yonelinas, A. P., 255, 256, 257, 260, 267, 269, 274, Winkielman, P., 291, 293 483, 487, 488, 493, 817 Winman, A., 289, 291, 306 Yorifuji, S., 790 Winocur, G., 821 Yoshida, H., 525 Winograd, P., 291 Yoshiura, T., 795 Winograd, T., 129 Young, A. W., 228, 229, 810–811, 872 Winter, L., 561 Young, C., 494 Winward, L., 817 Young, D., 648 Wiseman, M., 441 Young, M. P., 746, 747 Wishaw, I. Q., 787 Younger, J., 535 Witherspoon, D., 253 Yun, L. S., 818, 820 Wittgenstein, L., 85, 147 Yzerbyt, V. Y., 291, 454 Wixted, J. T., 255, 818 Woddis, D., 685, 686 Zacks, J. M., 363 Wojciulik, E., 228, 230, 718 Zahavi, D., 46, 61, 72, 76, 81, 82, 83, 84, 526, 762 WoldeGabriel, G., 585 Zaidel, E., 783, 789 Wolfe, J. M., 211, 753 Zajonc, R. B., 221 Wolkstein, M., 718 Zakay, D., 298 Woloshyn, V., 268 Zaks, M., 750 Wolpaw, J. R., 328, 329, 343 Zanella, F. E., 718 Wolpert, D. M., 339, 343, 344, 345, 560, Zangwill, O. L., 467 564 Zarcone, V. P., Wolters, G., 714 Zechmeister, E. B., 260, 296, 301, 304 Wood, B., 210 Zeki, S., 203, 543, 718, 779, 781, 793, 794 Woodruff, G., 573, 574, 577, 870 Zelazo, P. D., 391, 393, 409, 410, 411, 412, 416, 420, Woodruff, P. W., 465 422, 423, 425 Woods, R. P., 874 Zelazo, P. R., 407, 412 Woody, E. Z., 447, 455, 461, 464, Zeman, A., 531, 715 465 Zephrani, Z., 458 Woolf, Virginia, 381 Zhang, K., 414 Worthy, J. S., 295 Zhao, M., 811, 820 Wrangham, R. W., 390, 606, 612 Zheng, Y., 693 Wreschner, E. E., Zhu, J., 328, 329 Wright, D., 460 Zichel, S., 310 Wright, E. W., 330, 331, 783, 795 Ziesser, M., 339, 340 Wright, J. J., 736, 737 Zihl, J., 337, 747 Wright, R., 297, 310 Zijdenbos, A., 414 Wundt, Wilhelm, 26, 571 Zilles, K., 563 Wyer, R. S., 561 Zimbardo, P. G., 556 Wyzinski, P. W., 439 Zimmerman, J., 456 Zontanou, A., 718 Xie, X., 414 Zucker, R., 903 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

Subject Index

AA. See alphabetical numeric (AA) task in IDA, Abhidharma tradition, 93, 503 ideomotor theory of, 187, 339 on attention, 99–100 immediate vs. delayed, 338 on dependent origination, 93–94 intentional, 327, 329, 332 Dharmak¯ırti vs., 109–111 language and, 386 discernment for, 99 memory for, 811 emotion in, 99, 100–101 motor, 340 ﬁrst-person approach in, 94 movement vs., 328 intention in, 96, 99 neural activity of, 332 James, W., vs., 94–95 overlearning in, 343 mind in, 94, 97–101 perceived effects of, 339–341 on non-substantiality, 93–94 perception and, 336, 339, 347 Sankhya¯ vs., 94 prefrontal cortex role in, 564 types of awareness in, 97 quantum mechanical model of, 885 Western phenomenology vs., 94–95 sensorimotor representations and, 339 abnormalities, 343–346 sensory consequences of, 339 absorption, will and, 328, 329, 332, 445 access, 182 willed vs. ideomotor, 187 ACT* model, 153, 161 activation action(s) conscious vs. unconscious, 694–695 anarchic hand sign and, 343–344 factors in, 695 in blindsight patients, 342 active consciousness, 363 causal chain of, 329 active intellect choices about, 883 Aristotle on, 15 condition for, 341 ACT-R, 161 conscious vs. intentional, adaptation, 597, 600–601, 610–614, 615. See also control of, 343–346 adaptive behavior; psychological adaptation disownership of, 343–344 ancestral, 615 dual routes to, 847–852 behavioral response in, 620 emotional route to, 834 by-products vs., 602 goal and effects of, 339 cognition and, 189, 608–610 higher order thought and, 609 deﬁned, 602

939 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

940 subject index

adaptation (cont.) overt behavior in, 449 domain-specific mechanisms of, 604 psychophysiological indices of, 449 emotion and, 852, 855 reports of, 648 in evolutionary psychology, 602 subjective experience of, 449 exaptations and, 602–603 transpersonal anthropology and, 647–649 in Homo sapiens, 618–619, 870 altruism human vs. chimpanzee, 616–618 intentionality and, 580 intentionality and, 622–623 reciprocal, 580 introspection and, 614 ambiguity primate vs. human psychological, 615–616 linguistic phenomenon of, 365–366 social, 870 American Anthropological Association, 633 of speech, 586 amnesia, 2. See also anterograde amnesia; hypnotic unconscious processes of, 613 amnesia; posthypnotic amnesia; source amnesia The Adapted Mind (Barkow, Cosmides, Tooby), 609 implicit memory and, 810 adaptive behavior in learning, 453 ancient, 601 memory and, 452 self-consciousness impact on, 600–601 neuropsychological studies of, 810 adolescence amnesia critique DCx during, 427 inattentional blindness and, 211 adualism amnesic patients in children, 407 recall and recognition in, 252 adults remember-know paradigm in, 269 belief-desire reasoning in, 611 amygdala, 848 consciousness in, 418 action-perception coupling by, 873 executive function in, 418 brain mechanisms of, 839 explicit memory in, 418 damage to, 873 verbal communication in, 418 emotion and, 378, 527, 528 Advaita Vedanta,¯ 92 perception role of, 229 adverbial expressions, 370 post-perceptual processing by, 872 affect, 359, 831. See also emotion(s) research on, 872 affective consciousness, social judgements role of, 873 affective neuroscience, 501 thalamus and, 229 affectivity visual information processing by, 873 receptivity vs., 74 anaesthesia afflictive mentation, 97 brain activity during, 714 agency. See self-agency brain structure theory of, 711–714 aggregate judgement, 304 information processing under, 712 agrammatism, 782 memory processes under, 714 ahamk¯˙ ara, 92, 97 neuroimaging data on, 712 AI. See artificial intelligence thalamus role during, 711 AIM dimensional model analogy of brain states, computer, 650 of conscious states, 440 in Nyaya,¯ 101 5-D, 443 analytic philosophy air hunger. See shortness of breath Anglo-American, 35 alert state higher order monitoring theory of, 46 gamma synchrony in, 779 phenomenology and, 83 almost-rational model anarchic hand sign of judgement, decision making, action disownership in, 343–344 alpha waves SMA and, 343–344 in EEG, 182 utilization behavior and, 344 in occipital cortex, 182 an¯atman, alphabetical numerical (AA) task anatomically localizable mechanism in CLARION, 160 C dependence on, 782 altered states, 633–634. See also hypnosis; meditation Anaxagoras, 13 Buddhism and, 648 Anglo-American philosophy, 35 cross-cultural, 647 self-representational theory in, 50 definition of, 448–450 animal consciousness, 387, 436, 574 diagnosis of, 449–450 animals induction procedure in, 449 conscious states of, 48–49 methodological issues of, 648 consciousness in, 387, 436, 574 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

subject index 941

gamma synchrony in, 753 apperception, 104, 633 intentionality in, 573–575, 577 perception vs., 636–637 introspection in, 572 self-cognition and, 103 language in, 573 The Architecture of the Language Faculty (Jackendoff), mental modelling for, 135 357 mental representation in, 598 Aristotle, 15 mental time travel for, 579 on active intellect, 15 recursion in, 577–581 Descartes vs., 19 self-awareness in, 579 on matter and form, 15 social behavior in, 580–581 on mind, 15 synchrony in, 749 mind-body debate and, 19 tactical deception by, 578–579 on soul, 15 anoetic consciousness articulated emulator conceptual tests of, 272–273 in simulation theory of mind, 874–875 implicit memory tests of, 271 artificial intelligence (AI), 382 perceptual tests of, 271–272 consciousness studies and, 146 priming in, 271 critics of, 136, 146 anosognosia, 794 defined, 117 A-not-B task weak vs. strong, 142 infant performance of, 420 artificial learning (AGL) tasks antahkarana, 92 in CLARION, 160 anterograde amnesia, 842 Asa˙nga, 93 anthropology. See also cognitive anthropology; types of awareness for, 97 phenomenological anthropology; psychological ascending activation, 791–792 anthropology Mc vs., 791–792 American schools of, 632 Asian perspective, 89 behavioral environment in, 641 asked prime approach bias in, 649 stimuli and judgements tasks in, 216 evolution, biology, culture in, 632 aspect of experience, 642–647 cognition and, 102–103 methodology of, 634 assertoric states modern, 642–647 thetic vs. telic, 47 natural attitude in, 634 atman¯ , 91, 503 neurophenomenological approaches to, atomism, 19 658–660 attachment theory participant observation in, 634 in psychoanalysis, 681 phenomenalistic approach to, 643 styles of, 681 psychic unity of mankind in, 635 attention, 1, 358, 359. See also Focused Attention; psycho-cultural, 654–656 visual awareness senses and, 636–637, 656–658 Abhidharma on, 99–100 since mid-19th century, 632 alpha and theta activation with, 536 symbols in, 643–644 concentration vs., 100 universalist vs. relativist, 642 consciousness and, 199, 722–723 anthropology of consciousness to experience, 44 contemporary developments in, 647–649 information and, 199 cross-cultural research in, 661 multiple mechanisms of, 792 cultural schemas in, 649 neuronal synchrony impact on, 749–750 ethnological research in, 647 norepinephrine and, 439 ethnoscientific approaches to, 649 ordering of, 199 methodology in, 657–658 perceptual processing and, 722 object focus in, 660 somatic modes of, 659 structuralist vs. relativist, 654 thalamic contribution to, 792–793 anytime algorithm attention and awareness self-model and, 135 late selection models of, 208 apoha theory. See exclusion theory attentional amplification, 199 apparent consciousness, 41 attentional blink, 200, 721 appearance gamma synchrony impact on, 752 being vs., 40–41 Augustine, St. reality vs., 22 Descartes vs., 16 subjectivity and, 69 phenomenology of consciousness and, 16 appearance properties, 41 problem of other minds and, 16 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

942 subject index

autism Baar’s global broadcasting social behavior in, 871 Edelman-Tonini Dynamic Core Hypothesis vs., autobiographical memory, 183 Baars’ Global Workplace (GW) Theory, 162, automaticity 194–197 effortlessness in, 461 Baddeley, A., 187 incorrigible execution in, 461 WM for, 184, 188 inevitable evocation in, 461 Barkow, J., 609 in posthypnotic suggestion, 462–463 Baron-Cohen theory of mind speed in, 461 EDD and, 611 unavailability in, 461 ID in, 611 autonoetic consciousness, 263, 576 basal ganglia nuclei confidence in, 255–256 connectivity structure of, 746 defined, 254 Bastian, Adolf episodic memory in, 576 collection representation deduction of, factors impacting, 262 635 impairment of contents in, 485–486 elementary idea deduction of, knowing in, 255–256 fieldwork of, 635 measurement issues of, 254–255 folk idea analysis of, 635 memory and, 255–256, 810 psychic unity of, 636 in schizophrenia, 484 scientific psychology application of, 635–636 theoretical accounts of, 254–255 behavior. See also adaptive behavior; biobehavioral autonomic system processes; emergent behavior; social behavior; behavioral responses and, 852 utilization behavior availability consciousness, 36 in altered states, 449 awareness. See also experience; noetic awareness; under anaesthesia, reflective self-consciousness; self-awareness; anarchic hand sign and, 344 sensory awareness; somatosensory awareness; in animals, 580–581 visual awareness anthropology environment of, 641 in Abhidharma systems, 96, 97 attitudes impact on, 556 Asaçga on, 97 awareness of, 560–561 attentive vs. inattentive, 48 brain size impact on, 603 autonoetic vs. noetic, 482 complex, 12 of behavior, 560–561 conception of, 12 for children, 419, 424–425 concepts of, 12 of conscious states, 48 confidence judgements impact on, 309 direct availability for behavioral control of, 841 conscious choice in, 555 environment and, 869 conscious will and, 564–565 failure in neurophysiological cases of, 185 control of, 12, 556, 560–561, 841 of feeling, 573 disorders of, 708 grammar learning without, 209 emotion impact on, 836 integrated field theory of, 780 evolution and, 599 of intention, 564 human, 590, 622 mental processing outside, 693 imitation of, 407 objective vs. subjective thresholds of, 211–214 intentionality impact on, 619 of perception, 573 memory impact on, 252 perception without, 208–209 mental states and, 82 procedural knowledge without, 209 moral code of, 621 in schizophrenia, 484, 491 motivated vs. conscious, 675 of self, 573 neural correlates of, 378, 420 6-alternative test for, 233, 234 non-REM sleep disorders of, 708 skill learning without, 209 posthypnotic, 463 temporal, 424–425 projective aspects of, 407 of thought, 573 proximate vs. distal causes of, 601–602 unconscious, 418 psychological mechanisms of, 601 unconscious vs. conscious, 418 REM-sleep disorders of, 708 without perception, 208–209, 244 retrospective monitoring impact on, 309 ayahuasca, 3 in social psychology, 561 transformation of, 408 BA. See Brodmann Area will and, 564–565 Baars, Bernard, 124, 182, 184, 201 without consciousness, 378 global broadcasting concept of, behavioral approaches GW of, 162, 194–197 to implicit perception, 207–209, 225 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

subject index 943

behavioral environment Bradley, F., 24 in anthropological theory, 641 brain, 37 behavioral response during anaesthesia, 711–714 in adaptation, 620 behavior control systems of, 12 autonomic system and, 852 Broca’s area of, 782 behavioral studies during Buddhist meditation, 543 of implicit perception, 210 chaos in, 735–736 behavioral-inevitability communication pathways in, 523–524 in computational models of consciousness, 121 computational model of, 118, 145 behaviorism, 30, 572 connections in, 745–746, 747 being consciousness as secretion of, 27–28 appearing vs., 40–41 consciousness mediation by, 442 Being and Becoming Human (Count), corticocortial reentry in, 201 Being and Nothingness (Satre), 73 cultural givens in, 642 belief as dynamic neural system, 732 in adults, 611 as dynamic structure, 732 in children, 611 environment relational interactions with, conscious experience vs., 866 867 desire reasoning in, 611 of evolution, 603–606 false, 49, 611, 870, 873 evolution of, 603–606 targetless thought and, 49 during Focused Attention, 542 in theory of mind, 611 glial cells in, 780 Berkeley, George, 23 global theories of, 780 bias growth of, 582 in anthropology, 649 interconnectedness of, 746 in dissociation paradigm, 244 introspection and, 37 in implicit perception studies, 225 during judgement, 872 monophasic, 647–648, 649 during Kundalini Yoga, 542 in social behavior, 560 linguistic ability in, 26 binocular rivalry, 527, 719 mathematical considerations of, 748–749 in implicit perception, 230 MBSR of, 522 biobehavioral processes meditation impact on, 499 subjective experience vs., 757 mind supervening on, 866–869 biofeedback, 182 as model generator, 651 biogenetic structuralism, 652 neurodynamic approach vs. computer metaphor for, biology 736 of hypnosis, 463 neurophysiological states of, 43 species, 35 neuroplasticity of, 521–523, 530 studies of, 152 personal vs. conventional, 651 Bi-thalamic occlusion, 790 phase-synchronization patterns in, 737 The Blank Slate (Pinker, S.), 609 production of consciousness by, 37, 118 blindsight, 2, 327, 336–338, 342, 793. See also change recording techniques of, 747–748 blindness; inattentional blindness; sound blindness scientiﬁc understanding of, 378 implicit perception and, 237–239 sensory systems of, 12 in monkeys, 572 spatiotemporal structures in, 737 Block, N. J. structural vs. functional, 746 globalist argument of, 202 symbolic systems in, 134 Inverted Earth case of, 45 Template for Action of, 889, 907 Bloomﬁeld, Leonard, 356 thalamocortical complex of, 192, 201 Boas, Franz, 636, 637 unseen stimulus impact on, 234 sensory impression anthropology of, brain damage 636–637 causes of, 794 body consciousness impact by, 787 experience of, 71, 78, 80 memory and, Bohm Approach spontaneous, 441 to quantum theory, 894–896 studies on, 336–339 Bourdieu, Pierre brain design practice theory and, 644–645 reward, punishment, emotion in, 836–838 Bower, G. brain dynamics CLARION, 161 quantum theory of, 888–889 bracketing brain functioning of objective reality, 70 dynamicist view of, 743 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

944 subject index

brain growth Broad, C. D., 30 neuron production and, 603 Broca, Paul, 26 postnatal, 603 Broca’s area brain imaging, 441 lesions around, 782 averaging procedures in, 736 Brodmann Area (BA) first-person approaches and, 493 of parietal cortex, 811 of hypnosis, 445 BSRF. See brainstem reticular formation of ideas, 363 Buddha, 95 of meditation, 540 Buddha S´ akyamuni,¯ 503 of synoptic analgesia, 464–465 buddhi, 90, 92, 97 brain lesions. See lesions Buddhism, 90–91, 509. See also Tibetan Buddhism brain motor areas Abhidharma, 93 of intention, 330, 332 altered states of consciousness and, 648 brain oscillatory rhythms basic forms of, 503–505 EEG studies of, 533 common trends in, 503 brain physiology cultural context of, 503 amygdala, 839 Dharmak¯ırti, 101 during anaesthesia, 711–714 discursive strategies in, 505 of C, 781 distinct practices within, 519 in change blindness, 720 emotional states in, 505 figures of speech in, 789–790 epistemology of, 107–109, 110–111 long-distance connectivity among, 193 history of, 503–506 during resting state, 709 identity in, 504 in sleep, 708–711 meditation in, 499 social information and, 876 memory in, of WM, 563–564 mental and material in, 96 brain processes mind-body dualism in, 96 in time-consciousness, 78 nirvana¯ in, 502 unconscious, 357 post-meditative changes in, 508 brain recordings practices of, 93, 94 interdependent activations from, 747–749 reflexive nature of mental events in, 102–104 brain size, 571 samatha´ in, 504, 506 behavior impact by, 603 Sankhya¯ vs., 96 changes in, 603–605 self in, 93, 503, 526 cognitive benefits of, 602 theoretical accounts of, 506 delayed development and, 605 theory of perception in, 110–111 language and, 581 thought and language in, 107–109 learning and, 605 typology of, 101 Neanderthal vs. Homo sapiens, 585 Buddhist meditation, 499 of primates, 388 brain activity during, 543 recursive thought and, 581 clarity vs. intensity in, 506–507 social complexity and, 605–606, 870 contemporary practice of, 508–510 brain states cultural context of, 503 AIM dimensional model of, descriptive precision emphasis of, 503 conscious states and, 38 diverse forms of, 503 states of consciousness and, 465–466 neuroscientific research on, brain synchrony. See synchrony object perception in, 506–507 brain systems Open Presence in, 507 for consciousness vs. language, 841 as relaxation, 507 memory and, 276–277, 810 samatha´ vs. vipa´syan¯a in, 506 brainstem reticular formation (BSRF), 786 terminology problems of, 508–510 breathing. See also respiratory drive traditional descriptions of, 502 control of, 504 vipa´syan¯a, 504–505, 506 during meditation, 504 visualization in, 505 mind and, 504 Buss, D. M., 609 thought and, 370 in vipa´syan¯a, 508 C breathlessness. See also shortness of breath anatomically localizable mechanism dependence of, neurons in, 791 782 breath-meditation, 504 brain structures of, 781 Brentano, Franz, 26, 96, 126 complex self-concept in, 776–778 self-representation theory of, 50 concept of, 776 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

subject index 945

consciousness vs., 778 childhood development cortical neuronal clusters and, 781 conscious awareness in, 419, 424–425 emergent property of, 781 first- and third-person knowledge during, 415 endowing emotions with, 796 imitation in, 407 episodic memory impact by, 797 information-processing bias in, 600 hemispheric doubling of, 789 neural activity in, 420 interruption of access to, 794 Piaget on, 408 Mc and, 779, 781 sensory orders in, 658 me-ness in, 777 stages of consciousness in, 394 NAP of, 778–779 children neocortex and, 782 adualism in, 407 neural mechanism needed for, 775 belief-desire reasoning in, 611 not included in, 778 cognition in, 291, 425 of representations, 794 concept of Now in, 423 time dependent process of, 779 consciousness development in, 391 ventral vs. dorsal stream and, 781 current state consciousness for, 421 CA. See confidence-accuracy relationship DCCS by, 422 Cabanis, P., executive function in, 423–424 Cartesian philosophy, 18, 38. See also Descartes, Rene´ false belief in, 611 epistemology and, 23 first-person perspective in, 408 mind-body interaction and, 18 future research areas on, 426 Cartesian Theatre, 124, 378 higher order consciousness in, 413–426 Cartwright, J., 609 highest level of consciousness development in, Carvak¯ a,¯ 90 413–426 categorical inference (CI) task imagination in, 393 in CLARION, 160 knowledge of mental states by, 619 categories knowledge of other minds by, 619 phonological representation of, 367 language and, 361, 382, 392, 408, 619 causal indexicality, memory in, 310, 415 causality mental states of, 419 action and, 329 metacognition in, 295 conscious intention and, 330 meta-representation by, 419 of phenomenal properties, 39 mind theory of, 384 primates’ understanding of, 387 narrative in, 391 triggering vs. structuring in, prefrontal cortex in, 414 CCC. See Cognitive Complexity and Control theory refC2 in, 423–424 central nervous system (CNS) representational change task for, 423 lesions in, 786 self conceptualization by, 393 centre mediane (CM) self continuity in, 421 function of, 786 self representation in, 415 ILN vs., 785–786 temporal awareness for, 424–425 centrencephalon as unconscious, 406 cerebral commissurotomy and, 783 chimpanzee consciousness mechanism of, 783–785 adaptation in, 616–618 cerebral commissurotomy cognition in, 606 centrencephalon and, 783 consciousness in, 387, 389–390 cerebral cortex. See cortex culture formation among, 387 Chafe, Wallace, 355, 367, 370, 372 intentionality in, 387–388 on imagery, 356–357 learning by, 387 Chag-zôg mental modelling by, 388 Tibetan Buddhism practice of, 510 mind theory in, 577–579, 870 Chalmers, David, 39–41 mirror self-recognition in, 606–607 conceivability of zombies of, 40–41, 53 perspective of other in, 607–608 naturalistic dualism of, 39–41 referential gesturing by, 617 positive theory of consciousness of, 42 social intelligence of, 578, 608 change and succession, 77 Chinese Room argument, 141 change blindness, 867, 873 Chomsky, Noam, 356, 366 frontal brain regions in, 720 Christian church chaos systems, 735 materialists and, 19 control of, 735 Churchland, Paul, 38 strange attractors in, 735 CI. See categorical inference task chaotic dynamics, 152 citta, 92, 96 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

946 subject index

CLARION. See Connectionist Learning with Rule cognitive functions Induction ON-Line of mind, 98–99, 598 classical physics cognitive neuroscience. See also social cognitive choice in, 882 neuroscience consciousness within, 882 first-person in, 71 difficulties with, 884 of social visual signals, 872 mind and matter in, 881 cognitive processes origins of, 884 mental processes vs., 684 science impact by, 883 cognitive psychology clauses, 362–363 computationalism in, 118 Clifford, William, 25 ideomotor approach in, 339 CM. See centre mediane cognitive revolution CNS. See central nervous system implicit processes impact on, 674 co-consciousness psychology, 1 trance logic and, 459 cognitive science cognition, 633. See also metacognition; self-cognition; abstract representation in, 598 situated cognition; social cognition computer analogy in, 650 adaptation and, 189, 608–610 connectionist theory in, 650–651 aspect and, 102–103 dynamical, 733 in children, 291, 425 neuroscience vs., 3 in chimpanzee, 606 social cognition in, 81 computer programs vs., 650 cognitive systems conscious and unconscious, 342 access to, 182 cortical coordination dynamics in, 740–741 appearance of, 600 EEG of, 749 cognitive/emotional constraint satisfaction emotion and, 10, 19, 76, 836 cognitive vs. hedonic, 694 environmental information impact on, 610 coherence. See synchrony explicit teaching requirement of, 607 collective consciousness, 632 externalization of, 869 collective representations in hypnosis, 448 collective effervescence impact on, 638 implicit assumption of, 608–610 color blindness Indian thinkers on, 90 in hypnosis, 457–458 materialistic account of, 17 coma, 271 me-ness and, 777 VS and, 716 mental factors and, 98 lesions causing, 715 metacognition vs., 290 communication methods of, 178 in adults, 418 neural synchrony in, 533 brain pathways of, 523–524 pre-linguistic, 99 of emotion, 835 relational interactions in, 867 imagination in, 362 as self-luminous, 103 sensory information in, 362 serial processing models of, 673–674 verbal, 418 situated, 866–868 comparative psychology social, 869–875 theory of mind in, 610–611 unconscious processes and, 2 compassion cognitive abilities Tibetan Buddhist cultivation of, 519 biologically primary vs. secondary, 612 component process model cognitive anthropology, 649–653 of memory, 276 cognitive architectures compromise formation, in information models of mental processes, 182 computation cognitive assessment defined, 152 of emotional states, 10 computational model cognitive closure, 37 of brain, 118 in mysterianism, 38 of mind, 118 Cognitive Complexity and Control (CCC) theory, computational models of consciousness, 410 151 cognitive constraints ACT* model, 153 experimental manipulation of, algorithm + instance view, 153–154 to judgement, 692 behavioral-inevitability argument in, 121 cognitive evaluation computers in, 131–132 emotion impact on, 836 critics of, 146 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

subject index 947

of Dennett, 124–127 confidence judgements encoding in, 131 free- vs. forced-report testing impact on, 312 Hofstadter, Minsky, McCarthy, 121–124 confidence-accuracy relationship (CA) IDA, 162, 166 mnemonic cues impact on, 307 localist distributed representation view, 154 research on, 307 mechanistic explanations in, 152 connectionism, memory element in, 131 architecture of, 145 mental models and, 135 computationalism vs., Moore/Turing inevitability, 120–121 Connectionist Learning with Rule Induction ON-Line of Perlis and Sloman, 127 (CLARION), 155 PS + SN view, 153 AA task in, 160 representational difference view, 155 AGL, 160 self-modelling in, 122 Bower and, 161 of Smith, B. C., 129–130 CI task in, 160 SN + PS view, 153 Hunt and Lansman’s model vs., 162 symbols and tokens in, 132 implementation of, 157–159 syntax and semantics in, 131 McClelland and, 161 synthetic summary of, 130–139 memory in, 160 Turing test and, 140–141 metacognitive subsystem in, 160 computational systems PC in, 160 code breaking in, 133 representation in, 157, 160 intentionality of, 126, 132–133, 137 Schacter model vs., 168–169 notion of, 131–132 skill learning tasks with, 160 self-modelling as conscious in, 134–139 SRT tasks in, 160 symbols in, 134 TOH task in, 160 computational theory. See computational models of connectionist theory consciousness in cognitive science, 650–651 computationalism conscience. See superego Chinese Room argument in, 141 conscious in cognitive psychology, 118 mental representation as, 43 defined, 117 unconscious vs., 154, 197, 332, 418, 459–460, evidence for, 118 466 in linguistics, 118 conscious activation. See activation phenomenal consciousness and, 118–119, 130 conscious agents Turing machines, 129 physical world interaction of, 883 Turing’s Test of, 140–141 conscious beings computer mental attributes of, 12 analogy of, 650 conscious broadcast The Computer Revolution in Philosophy (Sloman), IDA, 165 computers conscious cognition. See cognition brains as, 145 conscious content derived intentionality of, 126 primitive, noncognitive components of, 782–783 emotion feelings and, 838 conscious control human-level intelligence and, 140 in LOC model, 410 in models of consciousness, 131–132 conscious controller natural phenomenon and, 131 in Schneider & Pimm-Smith’s Message-Aware neural nets as, 132 Control Mechanism, 189 in psychological modelling, 145 conscious experience symbols attributed to, 131 beliefs vs., 866 conceivability argument, 41 in child development, 419, 424–425 concentration content of, 43 attention vs., 100 cultural and, 639, 644 on object, 504 dynamic core hypothesis of, 742 conception. See also thought dynamics of, 731 of behavior, 12 extended, 868–869 imagery and, 357 from external environment, 869 conceptual tests eye tracking indicator of, 776 of anoetic consciousness, 272–273 first-person perspective on, 481–482 Concordia Group, 657 higher order representation of, 47 confidence higher order thoughts and, 419 performance impact of, 309 linguistic expression and, 355 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

948 subject index

conscious experience (cont.) threshold view, 155 neural correlates of, 230 working definitions of, 178 in neurodynamic models, 763 consciousness. See also active consciousness; affective objective events and, 866 consciousness; apparent consciousness; availability problem of, 863–866 consciousness; C; co-consciousness; core qualitative vs. subjective, 46 consciousness; creature consciousness; social relations impact on, 863 development of consciousness; extended subjectivity of, 864 consciousness; fringe consciousness; Helmholzian substantive vs. transitive, 755 consciousness; higher order consciousness; subvenience base of, 869 inner-time consciousness; Meadean consciousness; conscious intention minimal consciousness; narrative-of-consciousness; causation and, 330 ordered states of consciousness; pathological conscious processes consciousness; phenomenal consciousness; direct accessibility of, 156 phenomenology; phonological consciousness; conscious recollection. See memory pre-reflective self-consciousness; primary conscious representation(s). See also representation(s) consciousness; pure-thought consciousness; frontoparietal cortices in, 723 recursive consciousness; reflective consciousness 1; neuronal processes in, 723 reflective consciousness 2; representing neuropsychological approaches to, 716–719 consciousness; species consciousness; structural conscious retrieval model of consciousness; temporary consciousness; neuroimaging studies of, 815 time-consciousness; trait consciousness; conscious states, 436 Vygotskyan consciousness; Woolfian consciousness AIM dimensional model of, 440 abnormalities in, 343–346 of animals, 48–49 access view on, 155, 162 awareness of, 48 adaptation and, 597, 610–614, 615 brain states and, 38 in adults, 418 causally inert, 42 affect in, 359 first-person knowledge of, 49 agency and, 292–293 four-dimensional model of, 441 all-or-none, 203, 406–407 frontoparietal cortical regions impact on, 707 alternative states of, 633–634 immediacy of, 47 anatomically specifiable mechanism for, 789 intentional direction of, 50 in animals, 436 memory tasks and, 251 anthropological studies of, 632 neuronal structures of, 708 artificial intelligence researchers on, 119 physiological processes in, 435 attention vs., 199, 722–723 self-directed element in, 51 autonoetic vs. noetic, 271 self-intimating nature of, 15–16 behavior vs. introspective evidence for, 572 thalamus impact on, 712 behavioral indices of, 418 conscious system in biological species vs. individual organisms, 35 in Freud’s topographic mind model, 676 brain damage impact on, 787 conscious vision brain mediation of, 442 frontoparietal networks role in, 720 as brain secretion, 27–28 conscious vs. unconscious, 154, 197, 332, 418, 459–460, C vs., 778 466 causal properties of, 378, 380, 854 access view, 155, 162 cerebral cortex and, 775 activation in, 694–695 challenges of, 610–615 activation strength, functional connectivity and, in chimpanzees, 387, 389–390 229 within classical physics, 882 assessment of, 178 cognitive methods and, 178 attractor view, 155 cognitive neuroscience of, 822 capability contrasts between, 181 collective, 632 chunking view, 155 comparative psychology of, 606–608 cognition, 2, 342 computational models of, 151 computational explanations of, 152 continued loss of, 715 emotional processing, 855 cortex vs. thalamus in, 742 empirical methods used in, 178 culture and, 634, 640, 650, 655, 660, 661 Hunt and Lansman’s model of, 162 definition, components of, 435–437 in intention, 341 development of, 407 memory in, 673 development psychology of, 391–394 PS in, 153 developmental perspective of, 407 SN in, 153 discontinuity in, 675 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

subject index 949

dissociable levels of, 407 phenomenological side of, 18 as dynamical operator, 738–739 of phonology, 361 in eighteenth century English novels, 396 as physical phenomenon, 18, 30 as emergent phenomenon, 136 physiology of, 437 emotions and, 383, 389–390, 838–839 positive theory of, 42 ethnocentric conception of, 633 precursors to, 606–608 evolution vs. role of, 597 present vs. absent, 180 evolutionary impact of, 598–599 primate behavior and, 598 explanatory targets with, 10 priorities within, 365 explicitness in, 645 process fractionation of, 716 in fetus, 417 properties of, 359 first-order vs. meta-level of, 406–407 property vs. contents of, 776 first-person subjectivity of, 17 psychoanalytic clinical observation of, 673 folk theory of, 378, 385 psychological functions of, 182 gamma synchrony and, 750–754 as psychological processes trait, 716–723 as global workplace, 744–745 purpose of, 565 globalist models of, 184, 201, 203 quantum vs. classical approaches to, 889–890 hierarchical structure of, 409 recursive operations of, 571 higher vs. lower, 410 reductive approaches to, 53 history of inquiry about, 1 relational properties and, 876 Homeric vs. Classical, 396 in religion, 621–622 human uniqueness of, 589 representational theories of, 18 human vs. animal, 436, 571 research on, 146, 251–252 ideational, 362 Sapir-Wolf hypothesis and, 641 imagination and, 391 scientific explanation of, 53 imagistic, 362 scientific integration of mind with, 29 immediate vs. displaced, 355, 359, 369, 372 self-consciousness vs., 136, 406 in Indo-European languages, 633 self-regulation of, 682 information-processing models of, 406 self-transcending nature of, 72 intentional alteration of, 649 semi-synonyms for, 776 intentionality and, 26, 346, 387–388, 633 sensory experience and, 359 interdisciplinary investigation of, 2 as serial processing system, 693 introspective, 386 shamanic vs. pathological, 649 language and, 126, 573 shared experience of, 582 in later psychoanalytic models, 678–682 simultaneous representation in, 716 matter vs., 17 single brain hemisphere requirement of, 789 meanings of, 707 social influence and, 384, 448, 573, 661 metacognition and, 313 structure vs. nature of, 26 microconsciousness vs., 203 substrate of, 526–528 modern problem of, 29 syntactic binding requirement of, 853 modern rational bureaucratic, 649 technological skill development and, 611–612 monitoring vs. controlling, 447, 451 temporal structure of, 647 monophasic bias about, 647–648, 649 thalamocortical system and, 715, 775 multiple modalities in, 654–655 theory of, 839–847 narrative forms of, 375 of thought, 361–362, 366 naturalistic, 16, 661–662 thought outside of, 370–372 as necessary condition, 341 of thoughts vs. sounds, 364 network theories of, 190–194 transitive vs. intransitive modes of, 526 neural basis of, 26, 203, 426–427, 743–744 Tulving’s distinctions of, 254 neuroanthropological approaches to, 652–653 unitary property of, 853 neurobiology of, 499 usages of, 778 neurodynamical models of, 738–745 as variable, 178 neuroscientific perspective of, 544 of voices in debate, 384 in nineteenth century European novels, 396 in Western imaginative literature, 394–398 noetic vs. autonoetic, 576 without behavior, 378 objective aspect of, 103 Consciousness and Cognition, 83, 177 as order parameter, 738–739 Consciousness and the Computational Mind as organism state, 708–716 (Jackendoff), 357 oscillatory neural synchrony and, 531–534 Consciousness and the Novel (Lodge), 397 as patterns of system-wide activity, 190–194 Consciousness Explained (Dennett), 383, personal, interpersonal functions of, 386–387 397 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

950 subject index

consciousness studies in language, 640 interdisciplinary cooperation in, 177 cultural psychology content consciousness in anthropology and, 653–660 poise and, 44 pluralistic contextualism in, 653 context proponents of, 653 GW of, 196–197 culture Continental philosophy, 68 cognitive theory of, 651 contrastive analysis, 178 conscious experience impact by, 639 data from, 180 consciousness and, 634, 640, 650, 655, 660, 661 method of, 180 emotion impact by, 660 in perception and imagery, 180 internalization of, 650 control intrapersonal, 650 of action, 343–346 meaning cycle of, 652 of behavior, 12, 556, 560–561, 841 migration and growth of, 590–591 of breath, 504 mind and body in, 661 in meditation, 504 narrative in, 396 memory and, 310 non-mentalistic system of, 643 metacognition and, 307–312 scientific phenomenology of, 644 subjective experience and, 314–315 self impacted by, 385 control sensitivity sensation and, 636 in memory, 310 sensory experience impact by, 638 conversation signification theory in, 650 language as, 389 social behavior impact of, 560 topics of, 389 as text, 644 conversion disorder, 450–451 transmission in humans of, 590 Conway, John, 30 Cooney-Gazzaniga Damasio, A., 855 globalist argument of, 202 globalist argument of, 201 Copenhagen Quantum Theory. See also quantum neuroanatomically motivated model of, theory 169 observer/observed system in, 885–886 Damasio theory of emotion, 855 core consciousness, 406 James-Lange theory vs., 855 cortex. See also motor cortex; neocortex; occipital Darwin, Charles cortex; orbitofrontal cortex; parietal cortex; evolutionary theory of, 24 prefrontal cortex; right anterior prefrontal cortex Darwinism, 24. See also neural Darwinism 35–75 Hz firing in, 169 mind-body debate and, 25 alpha waves in, 182 Davidson, Donald consciousness and, 775 swampman of, 865–866 ILN and, 783, 794 DCCS. See Dimensional Change Card Sort intelligence and, 775 DCx. See dimensional complexity lesions in, 778 De Anima (Aristotle), 15 memory and, 775, 809 decision-making neuron clusters in, 746 emotional constraint satisfaction in, 687–693 during non- vs. REM-sleep, 708–711 implicit vs. explicit systems in, 849 as reverberant cell assembly, 737 judgement and, 289 as seat of consciousness, 775 neural activity and, 125 thalamus and, 201, 742 perceptual system vs., 136–137 cortical coordination dynamics declarative knowledge, 695 in cognition, 740–741 deconstruction cortical neurons phenomenology impact on, 68 C and, 781 Deese-Roediger-McDermott paradigm (DRM) timing and binding in, 743 in false remembering, 277 cortical regions. See also non-sensory cortical Dehaene, Stanislas, 184, 198, 200, 202 networks Dehaene’s Global Neuronal Network Theory, 184, 198, Cosmides, L., 609 200, 202 creature consciousness, 35 Democritus, 13–14 Crick, F., 169 Dennett, Daniel, 378, 383, 397 Criterion of Ultimate Concreteness, 130 computational models of consciousness of, 124–127 cross-cultural psychology on first-person point of view, 124 Piagetain research in, 653 globalist argument of, 201 cross-modality priming, 276 intentional stance idea of, 132 cryptotypes on language, 125, 137 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

subject index 951

multiple drafts argument of, 202 modern applications of, 210, 215–216 dependent origination N400 and, 232 Abhidharma on, 93–94 neuroimaging and, 227, 234 Descartes, Rene,´ 90 null explicit sensitivity of, 212 Aristotle vs., 19 regression approach in, 220 dualism argument of, 18, 23, 38 relative sensitivity alternative to, 221–222, on emotion, 19 223 method of doubt of, 10 dissociative disorders, 450 on mind-body problem, 17, 19 hypnosis, 451 on scientiﬁc revolution, 17 taxonomy of, 451 sensory perception for, 19 double brain, 789 St. Augustine vs., 16 dream bizarreness, 438 theory of mind of, 18 dreaming, 123 development of consciousness bizarreness, 438 differentiation and integration in, 407 brain state of, 438 early accounts of, 407 emotion during, 440 Representational Redescription model, 409 internal perception during, 440 developmental psychology, 289, 391–394 lucid, 441 metacognition research in, 290 REM-sleep vs., 709–710 theory of mind in, 610–611 Driesch, Hans, 24 Dharmak¯ırti, 101 DRM. See Deese-Roediger-McDermott paradigm Buddhism, 101 DSM. See Diagnostic and Statistical Manual for Mental concept formation for, 107–108 Disorders mind for, 102 dual recognition memory models, 488 on non-deceptiveness, 101–102 dualism, 35, 122. See also mind-body dualism; property Nyaya¯ vs., 105 dualism; subject-object duality; substance dualism Diagnostic and Statistical Manual for Mental Disorders case for vs. against, 39–43 (DSM) of Descartes, 18, 23, 38 hysteria in, 450 McGinn’s mysterianism and, 36–37, 38 DICE. See Dissociable Interactions and Conscious of mind and body, 19 Experience model modern, 39 dichotic listening method monism vs., 38–39 implicit perception and, 209, 211 naturalistic, 39–41 differentiated self, 415 Plato’s arguments for, 14 Dignaga,¯ 101, 103, 107 positive theory of consciousness and, 42 Dimensional Change Card Sort (DCCS) reincarnation and, 14 childrens’ performance on, 422 Sankhya¯ vs. classical, 91–92 dimensional complexity (DCx) substance vs. property, 19 during adolescence, 427 dual-process recognition models direct perception remember-know paradigm and, 256–258 nonsensory symbolic forms impact on, 641 Durkheim, Emile, 638 discernment collective representations for, 638 for Abhidharma, 99 Dynamic Core Hypothesis, 184, 192 discrepancy-reduction model in of conscious experience, 742 in JOLs, 308 dynamic large-scale integration displaced consciousness, 355, 369, 372 radical embodiment and, 739–740 past tense compatibility with, 369 dynamic neural systems displaced immediacy, 369 autonomy in, 732 Dissociable Interactions and Conscious Experience brain as, 732 (DICE) model, 184, 185–186 deﬁned, 733–734 dissociation mathematics of, 734 alternatives to, 220 noise component of, 734 bias and motivation in, 244 nonlinear, 734 critiques of, 214 observed behavior vs. external constraints in, exhaustiveness requirement of, 214, 235 732 explicit perception measurement in, 211 order parameters of, 734 in hypnosis, 458–459 properties of, 732 implicit perception and, 210–211, 222, 241 segregation vs. integration in, 732 between intention and action, 562–564 self-organization of, 732 Know judgements in, 267 spatial extension in, 736 in learning, 454 dynamic unconscious, 675 merits and assumption of, 210–211 dynamical cognitive science, 733 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

952 subject index

ease-of-learning judgements consciousness and, 383, 389–390, 838–839 in metacognition, 293 cultural patterning of, 655–656, 660 Ebbinghaus, H., 252 Damasio’s theory of, 855 Ebbinghaus illusion, 334 Descartes on, 19 Eccles-Bech during dream consciousness, 440 quantum theory of, 902–904 elicitation of autonomic responses by, 834 EDD. See eye-direction detector elicitation of endocrine responses by, 834 Edelman-Tonini Dynamic Core Hypothesis, emulation of, 875 193 episodic memory impact by, 836 Edelman, G., 191 examples of, 831 educational psychology, 290 of fear, 389–390 EEG. See electroencephalographic patterns of fear vs. pain, 833 EEG alpha blocking functions of, 834–836 meditation and, 537 hyper- vs. hypocognized, 656 ego about judgement, 687–693 superego vs., 851 learning and, 834 ego psychology MD in, 786 in psychoanalytic theory, 679–680 memory impact by, 260–262, 836 Einstein-Maxwell theory, 884 in mentality, 643 electroencephalographic (EEG) patterns, 747. See also motivation impact by, 835, 836 EEG alpha blocking; synchrony neurophysiology of, 543 alpha waves in, 182 primary vs. secondary enforcers of, 833 of brain’s oscillatory rhythms, 533 regulation of, 517, 522 of cognition, 749 reinforcing stimuli impact on, 832 coherent, 420 rewards and punishment in, 832 fetal, 417 secondary social, 618 of infants, 420 self-regulation of, 873 of meditation, 534–538 in social bonding, 835 of sleep stages, 709 as states, 831–834 Stage 1, 437 unconscious, 458, 680, 855 of waking consciousness, 710 vicissitude and, 383 elementary consciousness emotional coherence, 691 concept of, 776 emotional consciousness, 10 The Elementary Forms of Religious Life (Durkheim), emotional constraint satisfaction 638 in judgement, decision-making, 687–693 Elliotson, John, 27 emotional processing embodiment paradigm conscious vs. unconscious, 855 in phenomenological anthropology, 658 emotion-based goals. See also goals; intention emergence, 13, 24, 29–30, 53, 136. See also emergent interaction and, 386 behavior; epistemological emergence empathy, 10, 82 definition of, 758, 780 Husserl on, 82 of language, 590 phenomenological conception of, 82 of mind, 875 Empedocles, 13 property dualism of, 30 encoding emergent behavior defined, 131 aggregative vs., 780 ensephalization quotient (EQ) defined, 780 humans vs. primates, 603 emotion(s), 633. See also implicit emotion entromedian nucleus (CM), 785 Abhidharma on, 100–101 environment adaptive value of, 852, 855 brain relational interactions with, 867 amygdala and, 378, 527, 528 conscious awareness from, 869 appraisal in, 832 information in, 867 behavior direction impact by, 836 information processing of, 867 behavioral responses in, 834 mind impact by, 868 biocultural approaches to, 653 non-biological vs. social, 870 in Buddhism, 505 EPAM model, 142 C and, 796 epilepsy, 738. See also temporal lobe epilepsy cause of, 832 split-brain phenomenon in, 379 cognition and, 10, 19, 76, 836 epiphenomenalism, 42 communication of, 835 episodic buffer in computers, 838 in WM architecture, 188 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

subject index 953

episodic memory, 137, 252, 405, 410, 810 Pleistocene, 582–583 in autonoetic consciousness, 576 social learning theories and, 600 C impact on, 797 underlying assumptions of, 599–600 emotion impact on, 836 Evolutionary Psychology (Buss, D. M.), 609 in IDA, 165 exaptations illusory recollection and, 278 adaptation and, 602–603 language and, 137 evolutionary psychologists on, 602 long-term, 183 exclusion theory mental time travel and, 577 of language, 108 MTL and, 276 executive function neuroimaging studies, 812 in adults, 418 semantic vs., 576–577 in children, 423–424 in Shallice’s Supervisory System Model, 187 outside of consciousness, 292 spatial-temporal context in, exhaustiveness requirement episodic recollection. See episodic memory of dissociation paradigm, 235 epistemic arguments, 39 existential phenomenology, 71 epistemological asymmetry existentialism between mind and matter, 28 phenomenology impact on, 68 epistemological consciousness, 10 experience epistemological emergence, 30 access to, 759–760 epistemological mysterianism, 36 anthropology of, 642–647 epistemology attention to, 44 Cartesian philosophy and, 23 of body, 71, 78, 80 higher order monitoring theory and, 49 computers ability to, 140 modern, 17 cultural influence on, 633 epoche´ dynamic aspects of, 733 first-person methods vs., 71 enabling restraints to, 600 in phenomenological reduction, 70 first-person quality of, 75 as practical procedure, 71 fleeting, 755–757 EQ. See ensephalization quotient intentional quality vs. matter of, 75 Equanimous Exchange of Self and Other misrepresentation of, 45 in Tibetan Buddhism, 518 ongoing, 755–757 ERP. See event-related potential proprioception in, 415 ERTAS. See Extended Reticular Thalamocortical scientific accessibility of, 759–760 Activating System stable vs. fleeting, 756 Ethics (Spinoza), 20, 21 syntheses of identity in, 77 event-related potential (ERP) experimental psychology for implicit perception, 227 first-person in, 71 P300 and, 233 explanatory gap, 36 in priming experiments, 217, 232–234 explicit knowledge, 419 events explicit memory as causally inert, 42 in adults, 418 evolution. See also Darwinism; human evolution MTL role in, 811 behavior and, 599 neuroimaging studies of, 811 of brain, 603–606 explicit perception. See also perception by-products in, 602 in dissociation paradigm, 211 Darwin, 24 implicit perception vs., 236 emergence and, 25 intentional/automatic dichotomy and, 243 heredity and, 11, 24 measurement of, 211, 213 inclusive fitness in, 599 objective vs. subjective criterion of, 211–214 of language, 137, 390, 571 performance and, 223 noise in, 602 explicit processing Pleistocene, 599 implicit processing vs., 226 of recursive thought, 571, 581–582, 590 parietal cortex and, 822 Evolution and Human Behavior (Cartwright, J.), prefrontal cortex and, 822 609 explicit recognition evolutionary psychology Know judgements and, 269 adaptationist approach in, 602 perceptual priming and, 269 books on, 609 extended consciousness, 406 higher order cognition in, 609 Extended Reticular Thalamocortical Activating System modern, 583–586 (ERTAS), 781 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

954 subject index

externalization realism vs. idealism, 494 of mind, 643 on recognition memory, 482, 488–491 extrospective, 37 on schizophrenia, 483–484 eye gaze structural invariants approach to, 761–762 social visual signal of, 871 subjectivity and, 2, 17, 76, 481, 493 eye-direction detector (EDD) third-person vs., 757 in Baron-Cohen theory of mind, 611 topological approach to, 760–761 eyewitness memory, 296 Turing’s Test and, 140 Flexible Item Selection Task, 425 false belief, 870 fMRI. See functional Magnetic Resonance Imaging emergence of, 873 Focused Attention targetless thought and, 49 brain during, 542 false recognition object of perception in, 511 MTL and, 822 Open Presence vs., 517 false remembering, 277 results of, 513 familiarity in Tibetan Buddhism, 511–513 left frontal/parietal activity of, 815 visualization in, 511 recollection vs., 814 FOK. See feeling-of-knowing schizophrenia impact on, 491 folk psychology, 12 fear, 389–390 forensic psychology, 289 Feature Integration Theory, 753 foresight bias Fechner, Gustav, 26 in JOLs, 305 feeling(s). See also emotion(s) FOXP2 gene in Abhidharma tradition, 99 articulate speech and, 586 human awareness of, 573 human vs. animal, 587 neural basis of, 838–839 language impact by, 590 feeling-of-knowing (FOK) Franklin’s IDA, 184 accessibility model of, 299, 315 free- vs. forced-report testing answer retrieval speed in, 309 confidence judgements impact by, 312 deceptive questions impact on, 302 free will, 290, 331, 846–847. See also will direct-access view of, 294–295 illusion of, 346 familiarity and, 298–299, 300, 309 intention and, 327 heuristic-based accounts of, 298–301 self-model and, 138 knowing vs., 301–303 Freeman, W. J., 736 in memory research, 290, 293 Freeman’s Dynamical Systems approach, 194 test alternatives impact on, 306 frequency-synchronization. See synchrony FEF. See frontal eye field Freud, Sigmund, 2, 27 Fellows of the American Association for Artificial mental processes structural model of, 675–678, Intelligence, 119 679 fetus on representations, 676 consciousness in, 405–406, 417 unconscious processes theory of, 675 EEG patterns of, 417 fringe consciousness, 10 pain in, 405 frontal eye field (FEF) first-person method transcranial magnetic stimulation over, epoche´ vs., 71 722 for investigating consciousness, 71 frontal lobes first-person perspective memory and, 276–277 in Abhidharma, 94 frontoparietal cortical regions of body, 80 in conscious representation, 723 brain imaging and, 493 conscious state impact by, 707 in children, 408, 415 during sleep, 724 in cognitive neuroscience, 71 frontoparietal networks of conscious states, 49, 481–482 conscious vision role of, 720 Dennett on, 124 functional Magnetic Resonance Imaging (fMRI), 2, in experimental psychology, 71 747, 811 of meditative states, 544 in meditation research, 540 of mental states, 409 memory studies with, 813 in metamemory, 494 PET vs., 813 in neurodynamic models, 763 technical limitations of, 814 phenomenological reduction in, 70 fusiform gyrus, 229, 872 quality of experience of, 75 implicit perception and, 228 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

subject index 955

Galileo, Galilei theoretical constructs of, 195–197 Hobbes impact by, 20 unconscious specialized processor in, 195 primary, secondary properties for, 16 goals Gall, Franz, 25 action and, 339 Game of Life, 30 in behavior hierarchy, 834 gamma synchrony, 750–754 conscious choice in, 565 in alert state, 779 nonconscious, 563–564 in animals, 753 in social psychology, 562 attentional blink effect by, 752 state representation of, 383 consciousness and, 750–754 will and, 565 in perception, 752 WM during, 563–564 problems concerning, 753–754 Godel’s¨ Incompleteness Theorem, 907 for sensory awareness, 754 grain size in visual system, 753 memory accuracy impact by, 311, 312–313 Gassendi, Pierre, 19 grammar Gazzaniga, Michael, 3 language without, 586 Geertz, Clifford learning of, 209 interpretivist/hermeneutic approach of, 643–644 Greeks. See preclassical Greeks General Relativity, 891 Grew, Nehemiah, 24 generalization synchronization. See synchrony gunas, 91 genetics GW. See Global Workplace Theory information based, 24 German and French philosophy habitus phenomenology impact on, 68 deﬁned, 644, 645 ghost in the machine in perception, 638 of mind, 643 unconscious acquisition of, 645 Giving and Taking Hacker, 382 in Tibetan Buddhism, 518 Hallowell, Irving glial cells behavioral environment of, 641–642 in brain, 780 hallucination global access hypnosis and, 465 agenda-based control and, 118 realism viewpoint on, 494 blackboard model of, 118 Harvard Group Scale of Hypnotic Susceptibility, Form global theories, 200. See also Global Workplace A (HGSHS:A), 446, 462 theory health Baars, 201 mental training and, 523 Block, 202 Hebbian reverberant cell assembly, 737 of brain, 780 Hegel, G., 24 of consciousness, 184, 203 Heidegger, Martin, 68, 71 Cooney-Gazzaniga, 202 on moods, 73 Dehaene-Naccache, 202 Heisenberg Uncertainty Principle, 885, Damasio, 201 907 Dennett, 201 Heisenberg, Werner, 883, 885 dissenting views, 202 Helmholtz, Hermann, 26, 380 of information-processing theories, 194–200 Helmholzian consciousness, 380–381 John, 202 hemispherectomy, hemicerebrectomy Kanwisher, 201 functional, 789 Llinas, 201 total, 788 Marcel, 202 heredity of network theories, 194–200 evolution and, 11, 24 Rees, 202 HGSHS:A. See Harvard Group Scale of Hypnotic Tononi-Edelman, 201 Susceptibility, Form A Varela, 202 high retrieval content Zeki, S., 203 neural activity of, 822 Global Workplace Theory (GW), 36, 162, 184, higher order cognition 194–197, 198, 201, 744–745 in evolutionary psychology, 609 Baars vs. Dehaene, 200 higher order consciousness, 406 context in, 196–197 in children, 413–426 intentional agents and, 203 language and, 425 modularity and frame problems in, 197 prefrontal cortex in, 412–413 neural analogues in, 195–196 thought and, 425 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

956 subject index

higher order global states theory Husserl, Edmund, 75, 79, 96, 110 of Van Gulick, 51 on empathy, 82 higher order linguistic processing, 854 on inner-time consciousness, 78 higher order monitoring theory (HOME), 16, 35 on kinaesthetic systems, 79–80 animal and infant consciousness and, 48 on natural vs. phenomenological attitude, 68 case against, 48–50 Husserlian phenomenology, 27, 70 epistemic argument against, 50 Huxley, Thomas, 12 epistemology of consciousness and, 49 hypnosis. See also hypnotic age regression; hypnotic master argument for, 47–48 amnesia; hypnotic analgesia; hypnotic deafness; Rosenthal’s higher order thought theory and, 46–47 hypnotic hypermnesia; posthypnotic amnesia higher order representation absorption in, 446 targetless, 49 altered state of, 447, 462 higher order syntactic thoughts (HOSTs), 831, automaticity in, 460–461 843–844 biology of, 463 vs.HOTs,852 brain imaging of, 445 language in, 852 brain states and, 465–466 higher order thoughts (HOTs), 292 cognition in, 448 action and, 609 color blindness in, 457–458 conscious experience and, 419 controversy over, 447 HOSTs vs., 852 dissociation in, 458–459 about lower order thought, 841 dissociative phenomenon of, 445, 451 Hilbert space, 887 enhancing, 446 Hindu philosophy hallucinations and, 465 Sankhya¯ tradition in, 91 hysteria and, 450–451 Vedanta¯ school of, 91 ideomotor theory of, hippocampus imagination impacted by, 445 in memory, 811 implicit emotion and, 458 Hitch, G. J., 187 individual differences in, 446 Hobbes, Thomas involuntariness in, 447, 461–462 Galileo impact on, 20 limits of, 460 Hofstadter, Minsky, McCarthy memory impact by, 460 computational model of consciousness of, 121–124 mental imagery and, 447, 465 Hofstadter, Richard, 121, 124 in mind and body, 463–464 holonomic theory motivational context of, 448 receiver in, 191 neodissociation theory and, 447 HOME. See higher order monitoring theory neurophysiology of, 463–464 Homo sapiens neutral, 464 migration of, 584 for pain, 454–455, 465 social adaptability of, 870 physiological correlates for, 448 tool making by, 611 psychometric instruments and, 446 HOSTs. See higher order syntactic thoughts research on, 446 HOTs. See higher order thoughts social interaction of, 445 How the Mind Works (Pinker, S.), 609 sociocognitive theory of, 447 human evolution state-nonstate dichotomy in, 448 phylogeny for, 603 Stroop Effect in, 457–458, 460 reinterpretation process in, 617 trance logic in, 447–448, 459 human nature unconscious vs. conscious in, 459–460 relativist notions of, 637 uses of, 459 humans. See also Homo sapiens hypnotic age regression, 459–460 behavior of, 590, 622 hypnotic amnesia culture transmission by, 590 posthypnotic amnesia vs., 451 educability of, 612 hypnotic analgesia, 445 EQ in, 603 brain imaging of, 464–465 intentionality in, 573–575, 618 explicit vs. implicit perception in, 455 language in, 11, 573 implicit perception in, 454–455 learning in, 599 hypnotic blindness social cognition by, 608 implicit perception in, 456–457 taxonomic classiﬁcation of, 616–617 perceptual couplings in, 456–457 unique psychological adaptations in, 618–619 hypnotic deafness, 445 Hunt and Lansman’s model delusion in, 456 conscious/unconscious process in, 162 implicit perception in, 455–456 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

subject index 957

priming in, 456 impersonal intentionality, 127 psychological studies of, 455 implicit body self hypnotic hypermnesia, 459 in infants, 415 hysteria implicit emotion in DSM, 450 hypnosis and, 458 hypnosis and, 450–451 implicit knowledge, 414. See also implicit perception suggestibility in, 451 in neurophysiological cases, 186 implicit memory ID. See intentionality detector amnesia and, 810 IDA. See Intelligent Distribution Agent evidence for, 717 idealism, 23. See also transcendentalism in posthypnotic amnesia, 451–454 Kant and, 22 Implicit Memory and Metacognition (Reder), 314 material world existence for, 19 implicit memory tests vs. materialism, 19 of anoetic consciousness, 271 psychology and, 26 contamination in, 273 ideas, 18 experimental methods in, 273–275 brain imaging of, 363 PDP in, 274 constancy of, 368 priming in, 252–253, 272–273, 279 of events and states, 362–363 responding on, 271 literary, 396 implicit perception, 207. See also explicit perception; representational content of, 19 implicit processing ideational experiences arguments for, 213 imagery vs., 360 attentional focus and, 224 identity behavioral approaches to, 207–209, 243 in Buddhism, 504 behavioral studies of, 210 in Open Presence, 513 believers’ vs. skeptics’ interpretations of, 242 self-representation and, 51 binocular rivalry in, 230 theory, 31 blindsight and, 237–239 ideomotor theory, 340 in brain damaged subjects, 235–236 of action, 339 critiques of, 212 illusion, 633. See also visual illusions dichotic listening method and, 209, 211 illusory recollection, 277. See also visual illusions direct vs. indirect equivalency measures in, 221 episodic memory and, 278 dissociation and, 210–211, 214, 222, 241 ILN. See intralaminar nuclei emotional stimuli and, 229 ILN subserve Mc empirical literature on, 208 bithalamic paramedian strokes and, 790 ERPs for, 227 multiple functional roles of, 787–788 evidence for, 209, 215, 222, 227, 244 widespread connections of, 787–788 exclusion studies and, 224–225 imageless thought, 367 exhaustiveness in, 220 imagery, 372. See also visualization explicit perception vs., 208, 210, 236 concepts and, 357 of faces, 228–230 contrastive analysis in, 180 failed exclusivity in, 220 experience of, 359–360 fMRI for, 227, 234 ideational experiences vs., 360 fusiform gyrus and, 228 Jackendoff vs. Chafe on, 356–357 in hypnotic analgesia, 454–455 language and, 356 in hypnotic blindness, 456–457 nature of, 359–360 in hypnotic deafness, 455–456 nonverbal, 360 inattentional blindness and, 213 imagination, 368, 633 inferences of, 227 in children, 393 intentional/automatic dichotomy and, 243 in communication, 362 masked priming and, 210, 211, 215 consciousness and, 391 motor interference vs. semantic priming in, 219 hypnosis impact on, 445 neuroimaging of, 207–209, 227, 228, 234 in narrative, 391 neuroimaging processing, stimulus classes for, object-directed intentionality and, 73 227–228 perceptual experience of, 71 neuroimaging vs. behavioral approaches to, 227 recollection vs., 73–74 neuropsychology and, 207–209 imitation objective-subjective continuum of, 213 in child development, 407 parietal neglect in, 239–242 immediate consciousness, 355, 372 patient data evidence for, 235–236 displacement in, 370 perceptual learning approaches to, 215–216 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

958 subject index

implicit perception (cont.) in Sankhya,¯ 101 performance and, 214 information popular notions of, 208 attention and, 199 presentation duration and, 224 location of, 867 prime stimulus impact on, 210 information models of mental processes priming effects and, 213 cognitive architectures in, 182 process dissociation procedure for, 220 information processing, 406 qualitative performance differences and, 223 under anaesthesia, 712 relative differences methodologies in, 220 of environment, 867 relative sensitivity procedure for, 220, 221 globalist models of, 194–200 response compatibility in, 241 implicit, 414 signal detection theory in, 226 intelligence of, 136 skeptics vs. believers, 209 levels of, 357 subjective experience in, 211, 212 modular processes in, 184 symbol manipulation, representation in, 237 in Shallice’s Supervisory System Model, 186 two pathways argument in, 237–239 inhibition. See unbalanced inhibition word primes and, 219 inner speech implicit perception studies, 209 thought and, 357 behavioral approaches to, 214, 225 inner-time consciousness biases and motivational factors in, 225 Husserl on, 78 masked prime approach to, 210, 211 temporality and, 76–78 numerical stimuli in, 219–220 input-output gating regression technique in, 226 sleep and, 440 response compatibility in, 226 Insight Meditation. See vipa´syan¯a implicit processing integrated ﬁeld theory cognitive revolution impact by, 674 of awareness, 780 explicit processing vs., 226 intellect. See also active intellect pervasiveness of, 219 representational ability of, 92 of prime stimulus, 215 intelligence semantic content and, 210 cerebral cortex and, 775 of words and numbers, 230–232 computers, 140 inattentional blindness, 756 consciousness and, 136 amnesia critique and, 211 Machiavellian, 851, 870 implicit perception and, 213 in nonhuman primates, 851 inconclusiveness, 367 Intelligent Distribution Agent (IDA), 184 Independence Remember Know (IRK), 257 architecture of, 163 Indian tradition. See also Buddhism behavior net in, 166 on cognition, 90 coalition manager in, 165 on mind, 90 cognitive process model of, 162, 166 individual organisms episodic memory in, 165 consciousness in, 35 modules of, 163 induced gamma response, 750 psychology/neurobiology and, 162 inevitable evocation SDM in, 164 posthypnotic suggestion and, 463 slipnet of, 164 infants intention A-not-B task performance by, 420 brain motor areas of, 330, 332 EEG activity of, 420 intention and action, 328 implicit body self in, 415 dissociations between, 562–564 intermodal body schema of, 415 emotive, 76 me vs. I in, 420 executive functions in, 341–343 memory in, 415, 417–418 intentional object and, 72 mirror self-recognition paradigm in, 420 moments in, 75 NIRS of, 420 performance of, 333–339 PET of, 420 subjective experience of, 329, 340 refC1 in, 422 intentional lobotomy, 441–442 self consciousness in, 420 intentional object self-differentiation process in, 415 intentional act and, 72 time-orientation in, 418 pain as, 73 inference intentional stance, 127, 618 about others’ minds, 385 Dennett’s idea of, 132 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

subject index 959

intentional states internal perception evidence for, 851 during dream consciousness, 440 psychopathology impact by, 615 interpretation intentional/automatic dichotomy of sensory information, 362 explicit perception and, 243 intersubjectivity, 80–83. See also subjective experience implicit perception and, 243 objectivity and, 83 intentionality, 26, 96, 346, 387–388, 633. See also phenomenologists on, 81–82 intention and action; intentional stance; problem of other minds and, 80–81 perceptual intentionality intralaminar nuclei (ILN) Abhidharma on, 99 activation relay by, 792 action and, 327, 329, 332 changed definition of, 786 adaptation and, 622–623 CM function vs., 785–786 altruism and, 580 cortex and, 783, 794 in animals, 577 VS impact of, 794 awareness of, 564 Mc impact by, 785 behavior impact by, 619 selective processing by, 792 brain motor areas in, 330, 332 subdivisions of, 785 in chimpanzees, 387–388 thalamic obstructive strokes and, 790 of computational systems, 126, 132–133, 137 visual attention role of, 788 conscious and unconscious in, 341 intransitive self-consciousness, correlational structure of, 72 intrapersonal culture, 650 defined, 126 introspection, 26, 38, 186, 386, 841 derived vs. intrinsic, 126, 131, 133 adaptive value of, 614 detection of, 610 brain and, 37 empathy and, 82 intentionality and, 574–575 event sequence of, 331 memory and, 278, 303 free will and, 327 mental activity of, 9 goal-directed conscious, 328 metacognition and, 295 in humans vs. animals, 573–575 model mediation of, 122 impersonal, 127, 132–133 in nonhuman animals, 572 impression formation and, 561–562 about perception, 138 internal harmoniousness and, 133–134 pitfalls of, 762–763 introspection and, 574–575 Inverted Earth case, 45 in language, 590 involuntariness of mental representations, 126 in hypnosis, 447, 461–462 in narrative, 383–384 neuropsychology of, neural state of, 329–330, 332 involuntary memory, 252, 275–277 object-directedness of, 71, 72–73, 110–111 involuntary recollection, 252 observer-relative, 127 ipseity, 76, 526 orders of, 575–576 IRK. See Independence Remember Know phenomenological concept of, 71 of physical systems, 133 Jackendoff, Ray, 355, 357, 364, 366, 370, pre-linguistic, 126 372 in preverbal infants, 619 on imagery, 356–357 recursion impact on, 575 Jackson, Frank referential opacity in, 574 Knowledge Argument of, 39–40 robots and, 134 James, William, 27, 178, 182, 381, 436 Searle on, 141 Abhidharma vs., 94–95 self-direction and, 51 stream of consciousness of, 359, 733 self-monitoring of, 345 James-Lange theory of emotion signitive, pictorial, perceptual, 73 Damasio’s vs., 855 in social behavior, 561–562 Jibu-Yasue subjectivity and, 70, 340 quantum theory of, 902–904 use of, 613 jñ¯a, 90 utilization behavior sign in, 343 John, E. intentionality detector (ID) globalist argument of, 202 in Baron-Cohen theory of mind, 611 Johnson-Laird’s Operating System Model of interdisciplinary investigation, 2 Consciousness, 184–185 intermodal body schema JOLs. See judgements of learning of infants, 415 Journal of Consciousness Studies, 83, 177 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

960 subject index

judgement. See also metacognitive judgements type vs. expressed, 695 aggregate, 304 Knowledge Argument, 39–40 almost-rational model of, materialists on, 40 brain region activation in, 872 Koch, C., 169 constraints to, 692 Kumarila,¯ 108 decision-making and, 289 Kundalini Yoga emotional satisfaction in, 687–693 brain activity during, 542 in metacognition, 291, 293 Kurzweil, Raymond, 120 perceptual experience of, 71 about social information, 871 La Mettrie, Julein de, 22–23 judgements of learning (JOLs) language(s), 184. See also narrative language; calibration of, 304 programming languages discrepancy-reduction model in, 308 of action, 386 experience vs. theory-based, 302 brain size and, 581 experiences based, 298 in Buddhist epistemology, 107–109 foresight bias in, 305 case grammar and, 388 high frequency words and, 301 children and, 361, 382, 392, 408, 619 intrinsic vs. extrinsic, 295–296 components of, 360–361 in metacognition, 293 consciousness and, 126, 573 paired-associate learning in, 304–305 as conversation, 389 processing fluency impact on, 302 cryptotypes in, 640 retention interval impact on, 302 Dennett on, 125, 137 UWP with, 304 differences in, 370 emergence of, 590 kalpan¯a, episodic memory and, 137 Kant, Immanuel, 638 evolution of, 137, 390, 571 idealism and, 22 exclusion theory of, 108 on mind, 23 familiar vs. unfamiliar, 364–365 transcendentalism of, 23, 69, 494 first-order intentionality in, 590 Kanwisher, N. FOXP2 gene impact on, 590 globalist argument of, 201 generativity of, 573 kinaesthesis gestural theory of, 587 Husserl on, 79–80 higher order consciousness and, 425 perception and, 79–80 HOSTs in, 852 Know judgements human use of, 11 explicit recognition and, 269 in humans vs. animals, 573 fluency-familiarity interpretation of, 266–267 imagery and, 356 instructions impact on, 264 inadequacies of, 366 interpretation of, 270–271 Indo-European, 633 perceptual priming and, 267–269 intention verbs in, 388 knowing. See also Know judgements intentionality in, 590 cognitive studies of, 810 limitations of, 367 confidence and, 269 little-studied, 365 FOK vs., 301–303 mental functioning impact by, 653 guessing and, 269–270 myth and, 390–391 illusions of, 277–278 within narrative frameworks, 377, 393 knowing of, 388 neuronal synchrony impact on, 749–750 for man vs. animal, 388 outside consciousness, 356 metamemory judgements and, 269–270 perception impact by, 640 neural activity of, 816–818 percepts valuations in, 358 remembering vs., 810 preadaptation for, 388, 390 knowledge. See also implicit knowledge processing system of, 842 declarative vs. procedural, 695 programming, 129 explicit, 419 progression of, 362 explicit vs. implicit, 695 proto-, 586 knowing-self vs., 102 protolanguage vs., 586 knowledge of, 294 recursion in, 576 meaning and, 126 recursiveness of, 576 neural events for, 873 relations in, 371–372 quasi-declarative, 696 Sapir-Wolf hypothesis and, 641 representation of, 137 sensory experience stream of, 640 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

subject index 961

in Shallice’s Supervisory System Model, 186 Levy, Robert signed, 587 on hyper- vs. hypocognized emotions, social behavior impact by, 611–612 656 sounds of, 360, 364 Lewes, George, 29 stages in production of, 360 LFP. See local ﬁeld potentials subjectivity and, 573 LGN. See lateral geniculate nucleus thought and, 134, 137, 356, 358, 366, 426 L’hermitte’s syndrome, 564 of unconscious system, 677 L’Histoire Naturelle de l’Ame¯ (La Mettrie), 23 vocabulary and, 182 L’Homme Machine (La Mettrie), 22 without grammar, 586 Libet experiments written symbols of, 365 in quantum theory, 899–900 Language Instinct (Pinker), 609 linguistic expressions Lateral Dorsal (LD), 784 autonomous speech impact on, 590 lateral geniculate nucleus (LGN) conscious experience and, 355 of thalamus, 747 representational content of, 53, 99 lateralized readiness potential (LRP), 231, 331–332 linguistic processing in masked priming, 217 in mind, 641 LD. See Lateral Dorsal in neural networks, 854 learning linguistics action and, 343 brain regions and, 26 brain size and, 605 cognition and, 99 by chimpanzees, 387 computationalism in, 118 dissociation in, 454 intentionality and, 126 emotion and, 834 phonological organization in, 361 in humans, 599 in twentieth century, 356 implicit vs. explicit, 651 literary idea instrumental, 834 evolution of, 396 metacognition and, 290 Lived body (Leib), 79 monitoring, control processes in, 291 living systems norm of study in, 308 quantum theory of, 887 procedural, 651 Llinas, R. rote, 365 globalist argument of, 201 self-paced, 312 lobotomy. See intentional lobotomy source amnesia in, 453 LOC. See Levels of Consciousness stimulus-reinforcement type of, 832, 834 local associations study time allocation in, 308 in IDA, Leibniz, Gottfried, 17, 20–22, 104 local ﬁeld potentials (LFP) on mental states, 22 of neurons, 747–748 on monads, 21 Locke, John, 104 Spinoza vs., 21 locked-in syndrome, 442 Leibniz’s Law, 17 Lodge, David, 397 lesions, 441–442 Loftus misinformation paradigm, 277 VS and, 715 Logical Investigations (Husserl), 75 around Broca’s area, 782 logical thought Bi-thalamic, 790 in primitive mentality, 639 causing MCS, 715 lower order thought cerebral cortical, 778 higher order thought about, 841 in CNS, 786 LRP. See lateralized readiness potential coma and, 715 lucid dreaming, 441 neuropsychological studies of, 716 Levels of Consciousness (LOC) model, M* 409 M** about, 122 conscious control in, 410 Machiavellian Intelligence Hypothesis, 851, 870 mindfulness in, 427 magnetic resonance imaging (fMRI) minimal consciousness in, 414 for implicit perception, 227, 234 prefrontal cortex development in, 421 of perception, 230 rule complexity in, 410 magneto-encephalography (MEG), 2, 747 SelfC in, 421 Mahavibh¯as¯a, 95 Levi-Strauss,´ Claude manas, 92, 97 on myth, 642 Marcel, A. structuralism and, 642–643 globalist argument of, 202 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

962 subject index

masked priming. See also priming meaning compatibility effects in, 216, 221 knowledge and, 126 ERP us in, 217 past orientation of, 647 implicit perception and, 210, 211, 215 standardization and personalization of, 656 LRP in, 217 value vs., 647 N400 modulation by, 232–233 mechanist models of consciousness. See computational neural activation in, 217 models of consciousness regression use in, 218 medial dorsal nucleus (MD), 784 response window and, 216–220 in emotion, 786 SOA in, 218 explicit memory role of, 811 speeded response in, 218 structure of, 786 studies of, 216 medial-temporal lobe (MTL) visibility of prime in, 216 episodic information and, 276 with words, 216 false recognition and, 822 Mass Action in the Nervous System (Freeman), memory impact by, 811 736 meditation. See also Buddhist meditation; meditative master module of consciousness, 170 states; samatha´ ; Tibetan Buddhism; material world Transcendental Meditation; vipa´syan¯a as appearance, 22 academic research interest in, 530 existence for Idealists of, 19 active vs. passive, 541 perception and, 23 actual vs. post-meditative states of, 507 materialism. See also Carv¯ aka¯ advanced, 521, 530 appearance vs. reality in, 41 advanced hypotheses of, 521 Christian church and, 19 attention regulation in, 535–536 cognition in, 17 benefits of, 545 conceivability of zombies and, 40–41 brain imaging techniques and, 540 existence of mind in, 19, 27 brain impact of, 499 idealism vs., 19 breath control in, 504 origin of consciousness and, 12 in Buddhism, 499 vitalism vs., 24 control in, 504 matter EEG and, 534–537, 538 consciousness and, 17, 24 effects after, intrinsic mental properties of, 28 effects during, 520 La Mettrie on, 23 effort in, 516 third person objective of, 17 empirical research on, 499 matter and form excitement in, 512 Aristotle on, 15 fMRI of, 540 Mauss, Marcel, 638 focused attention/mindfulness-awareness, MBSR. See Mindfulness Based Stress Reduction 542–543 Mc forms of, 500 ascending activation vs., 791–792 guided, 541 C and, 779, 781 hypoarousal in, 500 cerebral mechanism of, 779 improvement over time in, 502 ILN requirement of, 785 intentional modality in, 520 localization evidence for, 782 learned practice of, 502–503 NAP of, 778 long-term impact of, 544 thalamic structures in, 779 mind-body impact of, 513 McCarthy, John, 123 neuroimaging correlates of, 501, 530–538, McClelland, J. 539–543 CLARION and, 161 neuroscience and, 521–530 McDermott, D., 134 as neuroscientific explanandum, 502–503 McGinn, Colin neuroscientific research on, 499, 501 mysterianism theory of, 36, 38 objectless practice of, 507, 538 McGurk effect operational definition of, 500 in speech, 588 passive, 535 MCQ. See Memory Characteristics Questionnaire physiological baselines and, 528–530 MCS. See minimally conscious state practices of, 520 MD. See medial dorsal nucleus predictable effect of, 502 MDI. See motion direction information pure compassion and loving-kindness, 542–543 Mead, Margaret, 633 relaxation practice, 540–541 Meadean consciousness, 384–385 research on, 499, 519–521, 534–538, 540, 544 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

subject index 963

setting for, 511 implicit vs. explicit, 452, 810 specific traditions of, 501 in infants, 415, 417–418 stability vs. clarity in, 520 introspection and, 278, 303 staged progression of, 512 involuntary conscious, 275–277 states of, 502 involuntary recollection, 252 subject-object duality impact of, 502 MTL impact on, 811 tantric “Wind,” 505–506 neural activity of, 809, 812 terminology of, 519 neural substrates of, 812–816 in Tibetan Buddhism, 518–519 neuroimaging evidence of, 809, 811 traditions of, 501 neuronal synchrony impact on, 749–750 Meditations (Descartes), 18 noetic vs. autonoetic consciousness and, 255–256, meditative states 810 brain physiology during, 541 for odors, 811 first-person vs. cultural descriptions of, 544 perceptual experience of, 71 neuroelectric correlates of, 531–538 performance of, 312 Tibetan Buddhism in, 510–511 PET studies of, 812–813 medulla posthypnotic amnesia and, 452 respiratory centers of, 791 prefrontal and parietal cortices relation to, 814 MEG. See magneto-encephalography prefrontal cortex and, 816 meme, 125 procedural, 810 memory, 633. See also autonoetic consciousness; quantity-accuracy tradeoff in, 310 dual-process recognition models; episodic recollection vs. familiarity in, 258 memory; explicit memory; involuntary memory; during REM sleep, 438 involuntary recollection; mental activity; mental remember-know paradigm in, 253, 254 events; metacognition; metamemory; recall and research on, 1, 251–252 recognition; recognition memory; recollection; retrieval content of, 809 remembering; remember-know; semantic memory retrieval vs. storage strength in, 301 accuracy of, 300, 309–310 in schizophrenia, 485 for actions, 811 semantic, 810 under anaesthesia, 714 sensory reactivation in, 812, 818–821 autobiographical, 183 serotonin and, 440 autonoetic consciousness and, 255–256, 810 short-term, 187, 753, 849 behavior impact by, 252 for sounds, 811 brain systems and, 276–277, 810 source frameworks for, 262–263, 311 in children, 310, 415 storage of, 183 in CLARION, 160 study time allocation impact on, 312 cognitive neuroscience of, 822 subjective confidence in, 297 component process model of, 276 subjective experience and, 811 conceptual influences on, 259–260 in Vipassana tradition, 509 conscious recollection vs. familiarity in, 488 visual short-term, 753 conscious states and, 251 for visual stimuli, 811 conscious vs. unconscious expression of, 673 volition and, 277 constructive view of, 811 voluntary recollection, 252 control sensitivity in, 310 working, 855 cortex and, 775, 809 Memory Characteristics Questionnaire (MCQ), distinctiveness impact on, 260 263 emotion impact on, 260–262, 836 memory performance empirical studies of, 252 grain size impact on, 311, 312–313 episodic, 810 metacognition regulation impact on, 312–313 episodic vs. semantic, 576–577 memory research, 289 explicit vs. implicit, 253 FOK in, 290, 293 eyewitness, 296 priming and, 252–253 false, 277, 310 TOT in, 290 fMRI studies of, 813 memory retrieval fMRI vs. PET studies of, 813 emotion impact on, 836 free report of, 310, 311 prefrontal cortex and, 811 frontal lobes and, 276–277 success vs. attempt, 812, 815 high vs. low retrieval content of, 809, 822 memory storage hippocampus in, 811 emotion impact on, 836 hypnosis impact on, 460 memory systems illusions of, 277–278 declarative, 852 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

964 subject index

memory tests first- and third-person in, 409 implicit vs. explicit, 271 Leibniz on, 22 me-ness of mind, 90 in C, 777 mind as succession of, 96 cognition and, 777 Nagel on, 74 mental activity. See also mental events; mental states neural events in, 329 of introspective consciousness, 9 non-introspective conscious, 841 in monkeys, 572–573 of others, 610 of reflective self-consciousness, 9 phenomenally conscious nature of, 75 thalamocortical interaction dependency of, 776 physical processes of, 90 of time travel, 579 representational content and, 27 waking vs. dreaming, 710 in self-representational theory, 50 mental architecture subject awareness of, 46 unconscious, 675 suffering and happiness and, 104 mental attributes unconscious, 53 of conscious beings, 12 unobservable entities of, 81 of matter, 28 Western vs. Indian philosophy of, 90 mental content mental time travel externalist view of, 866 human vs. animal, 579 mental disorders mental training schizophrenia, 483–484 physical health and, 523 mental events. See also mental processes mentalism, 356 in Buddhist epistemology, 102–104 mentality. See also primitive mentality declarative/procedural, 695–696 elementary vs. higher functions of, 653 implicit/explict, 695–696 emotion in, 643 reflexive nature of, 102–104 modern, 654 type vs. expression mode of, 696 motivation in, 643 mental factors primitive vs. modern, 639 in Abhidharma, 97–101 Merleau-Ponty, Maurice, 68, 71, 110–111 cognitive awareness and, 98 meta-awareness cognitive vs. affective, 100 in vipa´syan¯a, mental functioning metacognition, 313. See also olfactory metacognition language impact on, 653 children vs. adults, 295 non-conscious, 199 cognition vs., 290 mental imagery comprehension judgements in, 293 hypnosis and, 447, 465 control function of, 307–312 neural process to, 381 crossover model in, 315 mental modelling definition of, 290 animal’s capacity for, 135 developmental psychology research on, 290 by chimpanzees, 388 developmental vs. memory research on, mental models 290–291 computational models of consciousness, 135 ease-of-learning judgements in, 293 of others, 388–389 experimental memory research of, 290 mental processes experimenter vs. self-initiated strategies in, 292 cognitive processing vs., 684 introspection and, 295 conscious, 180 JOLs in, 293 Freud’s topographic model of, 675–678 judgement in, 291, 293 by function, 679 learning and, 290 information-processing models of, 182 monitoring impact on, 308 outside of awareness, 693 nonoccurrence decisions in, 297 unconscious, 180 at object-level vs. meta-level, 290 mental properties output monitoring in, 293 robots and, 124 philosophy of mind and, 290 mental representation probability judgements in, 291 as conscious, 43 recollection rejection in, 297 humans vs. animals, 598 remember-know judgements in, 293 naturalist account of, 53 research on, 289 mental representations RJR in, 293 intentionality of, 126 social cognition and, 297 mental states, 16 social psychology and, 291 behavior and, 82 source monitoring in, 293 of children, 419 subjective feelings and, 292 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

subject index 965

metacognitive judgements. See also confidence; computational model of, 118 confidence judgements as computer program, 1 accuracy of, 303–304 Descartes theory of, 18 bases of, 294 Dharmak¯ırti view of, 102 calibration in, 303 emergence of, 875 confidence and accuracy in, 298 environment impact on, 868 correction process in, 296–297 externalization of, 643 cue-utilization view of, 295 fluctuations of, 504 direct-access view of, 294–295 Freudian theory of, 27, 655 experience-based, 297–301 Freud’s topographic model of, 675–678 imagination inflation effect impact on, 300 ghost in the machine of, 643 inference vs. memory in, 301 Hobbs on, 20 information vs. experienced-based, 295 Indian traditions on, 90 mnemonic cues in, 298 inference about, 385 over- vs. underconfidence in, 306–307 internal social world of, 382 resolution in, 303 Kant on, 23 self efficacy in, 296 limitations of, 37 self-reflective inspection of, 298 linguistic systems in, 641 subjective/objective dissociations in, 301–303 materialist account of, 19, 27 theory-based, 295–297 mental states of, 90 trace-access view of, 303 in naturalistic terms, 16 unconscious basis of, 295 neuroplasticity of, 521–523, 530 validity of, 303–307 object directed nature of, 96 metacognitive knowledge, 295 Patanjali’s˜ Yoga view of, 92 metacognitive monitoring recursion in, 576–577 goal vs. data driven, 315–316 reflective, 395 metacognitive myopia, 306 in Sankhya¯ tradition, 91–92 metacognitive processes scientific integration with consciousness of, experience- vs. theory-based, 313–314 29 in learning, 290 situated in world, 868 metacognitive skills sociocultural complexity of, 635 age, individual differences in, 291 as sole reality, 22 metamemory, 290 as succession of mental states, 96 accuracy of, 300 supervening on brain of, 866–869 experimental paradigms in, 293–294 time travel by, 577 first-person perspective in, 494 mind and body, 16 research on, 291 alternative theories on, 19 metamemory judgements Aristotle vs. Descartes on, 19 knowing and, 269–270 Cartesian dualism and, 18, 23 meta-representation cultural orientations to, 661 by children, 419 Darwinism and, 25 method of doubt of Descartes, 17, 19 of Descartes, 10 dualism of, 19 microconsciousness, 203 hypnosis in, 463–464 consciousness vs., 203 interaction mechanisms of, 523–525 middle ages Leibniz on, 21 philosophy of, 16 meditation impact on, 513 Mˆıg-me´ Nying-je.´ See Non-Referential Compassion separation of, 17 Mill, John Stuart, 28, 29 Spinoza on, 21 mimesis, 390 Spinoza vs. Leibniz, 21 mind. See also problem of other minds; theory of mind mind and matter 20th century views of, 30 epistemological asymmetry between, 28 in Abhidharma, 94 The Mind of Primitive Man (Boas), 637 apart from consciousness, 15 mind-body dualism Aristotle on, 15 Buddhist, 96 Augustine, St., on, 16 mind-brain problem breath and, 504 brain state control in, 435 capacities of, 436 mindfulness, 99 Cartesian theory of, 38 in LOC model, 427 children’s theory of, 384 research on, 427 of chimpanzee, 577–579 Mindfulness Based Stress Reduction (MBSR), 531 cognitive functions of, 98–99, 598 brain activity of, 522 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

966 subject index

The Mind’s Past (Gazzaniga), 3 Mueller-Lyer illusion, 334 minimal consciousness, 391 multiple drafts argument in LOC model, 414 Dennett’s, 202 minimally conscious state (MCS) multiple personality disorders, 379 VS vs., mysterianism, 35 lesions causing, 715 case against, 38 Minskey, Marvin, 122, 123, 134 cognitive closure in, 38 mirror neurons, 341, 388, 874 dualism and, 36–37, 38 in social psychology, 563 master argument for, 37–38 mirror self-recognition ontological vs. epistemological, 36 in chimpanzees, 606–607 myth in infants, 420 language and, 390–391 misrepresentation Levi-Strauss´ on, 642 of experience, 45 mnemonic cues N400, 234 CA impact by, 307 dissociation paradigm and, 232 in metacognitive judgements, 298 masked primes modulation of, 232– modelling 233 of modelling, 139 for unseen stimuli, 235 subpersonal, 138 Nagel, Thomas, 12, 39–40 modern dualism on mental states, 74 properties in, 39 NAP. See neuronal activity patterns modern epistemology narcissistic disorders invention of, 17 self in, 682 modern structuralism, 642–647 narrative(s). See also narrative language monads agents in, 376 Leibniz on, 21 in children, 391 monism. See also neutral monism coherence in, 384 dualism vs., 38–39 as cultural objects, 376 materialist vs. idealist in, 38 cultural use of, 396 monitoring and control evolution of, 387–391 cause-and-effect relation between, 315–316 imagination in, 391 in learning, 291 intention in, 383–384 monkeys. See also chimpanzee interaction-type elements of, 376 blindsight in, 572 literary, 377 mental activity in, 572–573 mutuality and, 385–386 subjective experience in, 572 myth as, 390–391 moods, 73 stories vs., 376 Moody, Todd, 123 story-type elements of, 376 Moore’s Law, 121 vicissitudes in, 376 Moore/Turing inevitability, 120–121 narrative language Moravec, Hans, 120 development of, 377, 393 Morgan, Conwy Lloyd, 30 narrative self-consciousness motion direction information (MDI) for preclassical Greeks, 395 in STS, 793 narrative-of-consciousness, 375. See also narrative(s) motivation, 621 definition, 375 of behavior, 675 narratology, 376, 377 consciously intended, 675 narrator emotion impact on, 835, 836 self as, 384 in hypnosis, 448 natural psychologists, 386 in mentality, 643 natural selection reflective awareness and, 600 evolution and, 11 self-consciousness and, 601 naturalistic dualism, 39–41 in social psychology, 562 nature. See also human nature unconscious, 675 classical-physical models of, 884 motor actions observer vs. observed in, 886 sensory effects of, 340 psychological vs. physical in, 889 motor cortex scientific approach to, 13 TMS and, 340 NCCs. See neural correlates of consciousness movement NCE. See negative compatibility effect action vs., 328 near infrared spectroscopy (NIRS) MTL. See medial-temporal lobe of infants, 420 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

subject index 967

neocortex correlational strategy in, 757–758 C and, 782 ﬁrst-person data in, 763 neurons in, 746 neural activity in, 763 neodissociation theory, 451 qualiﬁcations to, 754–757 hypnosis and, 447 self-referring perspective in, 762 neonatal pain, 405 subjectivity in, 756–757 nervous system neuroimaging, 2 neural activity variability in, 736 of anaesthesia, 712 nonlinear dynamical system of, 734 of conscious retrieval, 815 synaptic pattern in, 734 of Democrats and Republicans, 692 network theories dissociation paradigm and, 227, 234 globalist models of, 194–200 of explicit memory, 811 system-wide activity in, 184 fMRI, 747 neural activity implicit perception and, 207–209, 215, 227, 228, of action, 332 234 in child development, 420 of meditation, 530–538, 539–543 decision making and, 125 of memory, 809, 811 distributed, spatiotemporal pattern of, 745 of objectless meditation, 538 dynamics of, 731 PET, 747 high retrieval content of, 822 of relaxation practice, 540–541 of knowing, 816–818, 873 of speech, 588 large-scale models of, 736 Transcendental Meditation, 534–538 in masked prime approach, 217 of visual extinction, 228 of memory, 809, 812, 816–818 neurology in neurodynamic models, 763 of consciousness, 203 perception and, 80 of speech, 587 spatiotemporal patterns of, 745 of subjective experience, 499 synchrony, 532–533 of vocalization, 586 variability of, 736 neuron clusters in waking consciousness, 449 in cerebral cortex, 746 neural circuitry neuron production of meditation, 501, 530–538, 543 brain growth and, 603 neural correlates of consciousness, 743–744 neuronal activity neural Darwinism, 651 of sleep, 708–711 neural events neuronal activity patterns (NAP) in mental states, 329 of C, 778–779 timing of, 329 Mc of, 778 neural models neuronal networks, 142 development of, 652 as computers, 132 neural processes, 2 Dehaene’s theory of, 184, 198, 200, 202 to mental image, 381 higher order linguistic processing in, 854 neural states information processing in, 839 of intentionality, 329–330 modularity of, 199 transitions between, 755 syntactic capability of, 853 neural synchrony, 532–533 syntax in, 853 in cognition, 533 neuronal processes frequency ranges of, 533 of conscious access, 724 neural systems. See also dynamic neural systems in conscious representation, 723 consciousness and, 26, 426–427 of reportability, 724 mediation based upon, 17 neuronal representation neuroanatomy in orbitofrontal cortex, 839 of remember-know paradigm, 269 neuronal synchrony social cognition and, 606 language impact by, 749–750 neuroanthropology syntactic binding through, 853 of consciousness, 652–653 neurons. See also cortical neurons; mirror neurons neurobiological theory, 36 behavior impact of, 378 neurobiology in breathlessness, 791 of consciousness, 499 discharge synchronization of, 743 IDA and, 162 functional columns of, 746 neuro-cognitive theories of consciousness, 178 LFP of, 747–748 neurodynamic models, 738–745 membrane potential of, 734 conscious experience in, 763 microtubular sub-structure of, 892 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

968 subject index

neurons (cont.) Non-Referential Compassion in neocortex, 746 emotional state of, 517 synchrony of, 749–750 Open Presence and, 517–518 transient functional coupling between, 737 practice sequentiality of, 518 neurophenomenology, 71 in Tibetan Buddhism, 517–519 neurophysiological states non-sensory cortical networks of brain, 43 in phenomenal perception, 724 failures of awareness in, 185 non-substantiality implicit knowledge in, 186 Abhidharma on, 93–94 introspection of, 38 nonverbal imagery physical states of environment vs., 43 thought impact by, 357 neurophysiology norepinephrine of conscious states, 708 attention and, 439 of emotion, 543 norm of study of hypnosis, 463–464 in learning, 308 of pain, 543 Now of simulation, 874 for children, 423 of sleep, 438–440 NREM sleep neuroplasticity stages of, 437 of mind and brain, 521–523, 530 nuclei neuropsychological architecture of thalamus, 785 consciousness in, 200 null explicit sensitivity neuropsychology, 2, 17, 289 of dissociation paradigm, 212 of amnesia, 810 Nyaya,¯ 101, 107 applications in, 900–902 analogy in, 101 of conscious representations, 716–719 Dharmak¯ırti vs., implicit perception and, 207–209 perception in, of involuntariness, lesion studies in, 716 object neuroscience. See also cognitive neuroscience concentration on, 504 affective, 501 givenness through phenomenology of, 70 cognitive science vs., 3 linguistic distinction about, 99 of consciousness, 544 mental content as, 516 dynamical approach in, 734–735 mental factor characterization of, 96 meditation and, 521–530 subjectivity as, 516 phenomenology and, 76 object perception pragmatic, 890 in Buddhist meditation, 506–507 social cognition and, 871 time-consciousness and, 77 neuroscientiﬁc research object relations theory on meditation, 499, 501, 534–538, 540 attachment theory in, 681 neutral hypnosis, 464 in psychoanalysis, 680–682 neural correlates of, 464 relational motives in, 681 neutral monism, 20 representational world in, 681 New Essays Concerning Human Understanding object-as-intended (Leibniz), 104 as object-that-is-intended, 72 New Look object-directedness defense/vigilance in, 683 imagination form of, 73 in perception, 683 of intentionality, 71, 72–73, 110–111 subception in, 684 pre-cognitive, operative intentionality and, Nice Work (Lodge), 397 74 NIRS. See near infrared spectroscopy recollection form, 73 nirvana, 93 object-experience Buddhist notions of, 502 self-experience in, 79 noema, 72 objective events noetic awareness, 576 conscious experience and, 866 at experimental level, 263 objective reality memory and, 810 bracketing of, 70 schizophrenia impact on, 491 objectivity semantic knowledge and, 263 intersubjectivity and, 83 non-deceptiveness objectless meditation Dharmak¯ırti on, 101–102 neuroelectric correlates of, 538 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

subject index 969

objects parietal neglect in Open Presence, 513 in implicit perception, 239–242 perception of, 10 participant observation physical emergence of, 130 in anthropology, 634 as relational correlates of experience, Pashler’s Analysis 69 in quantum theory, 898–899 observation Patanjali’s˜ Yoga reliability of, 94, 95 mind in, 92 observer vs. observed system, pathological consciousness, 649 occipital cortex PC. See process control tasks alpha waves in, 182 PDP. See process-dissociation procedure old-new recognition test, 815 Penrose-Hameroff olfactory metacognition, 293 quantum theory of, 890–894 On Alternating Sounds (Boas), 636 percept to preconscious buffer On the Origin of Objects (Smith), 130 IDA, ontological mysterianism, 36 perception, 38. See also direct perception; explicit Open Presence perception; implicit perception; internal basic practice of, 515–517 perception; visual perception in Buddhist meditation, 507 in, 101 discursive strategies in, 516 action and, 336, 339, 347 Focused Attention vs., 517 alterations in, 445 identity in, 513 amygdala role in, 229 Non-Referential Compassion and, 517–518 apperception vs., 636–637 objects in, 513 in Buddhist epistemology, 110–111 subjectivity in, 515 conscious element of, 10 theoretical background of, 513–515 consciousness vs., 10 in Tibetan Buddhism, 501, 513–515 context of, 78–79 openness to experience. See absorption contrastive analysis in, 180 operative intentionality, 72 co-subjectivity in, 83 activity vs. passivity in, 74 dissociations in, 242–245 optic ataxia, 336 dynamic sensorimotor approach to, 80 orbitofrontal cortex embodiment and, 78–80 brain mechanisms of, 839 fMRI and, 230 neuronal representation in, 839 gamma response in, 752 ordered states of consciousness, 47 habitus in, 638 origin of consciousness human awareness of, 573 mainstream view about, 11 implicit, 207 materialism and, 12 implicit vs. explicit, 208, 210, 236 other introspection about, 138 understanding perspective of, 607–608 kinaesthesis and, 79–80 output monitoring language impact on, 640 in metacognition, 293 of material world, 23 neural activity and, 80 P300 “New Look” in, 683 ERP patterns and, 233 as non-conceptual, 106 pain, 73. See also neonatal pain; suffering of objects, 10 hypnosis for, 454–455, 465 phenomenal consciousness and, 138 neonatal, 405 as presentational, 74 neurophysiology of, 543 prior entry phenomenon in, 331 paired-associate learning qualia and, 138 in JOLs, 304–305 social, 558–560 PANIC THEORY. See Tye’s PANIC Theory thought vs., 381 panpsychism, 13, 21 thresholds vs. signal detection of, 211 emergence vs., 12–13 of time, 633 neutral monism vs., 29 without awareness, 208–209, 244 parietal cortex, 811 perception of space, 633 BA area of, 811 perceptual couplings explicit retrieval and, 822 in hypnotic blindness, 456–457 familiarity and, 815 perceptual experience memory and, 814 of imagination, 71 number representation and, 231 of judgement, 71 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

970 subject index

perceptual experience (cont.) analytic philosophy and, 83 of memory, 71 constitution in, 69–70 of time-consciousness, 71 current, 68 perceptual intentionality deconstruction impact by, 68 differentiation of, 74 empathy in, 82 perceptual priming empirical vs. transcendental, 69 explicit recognition and, 269 existentialism impact by, 68 Know judgements and, 267–269 German and French philosophy impact of, 68 perceptual processing givenness of object through, 70 attention and, 722 hermeneutics impact by, 68 response compatibility in, 216 Husserlian, 27, 70 perceptual system intentionality in, 71 decision-making system vs., 136–137 intersubjectivity and, 81–82 perceptual tests neuroscience and, 76 of anoetic consciousness, 271–272 philosophical movement of, 67–68 Perennialism, 500 post-structuralism impact by, 68 performance rise of, 26 implicit and explicit perception and, 214 St. Augustine, 16 performance measures structuralism impact by, 68 direct vs. indirect tasks in, 221 Phenomenology and the Cognitive Sciences, 83 performed action phi phenomenon, 202 perceived effects of, 339–341 Philosophical Investigations (Wittgenstein), Perlis, Sloman philosophy. See also Anglo-American philosophy; computational models of consciousness of, Continental philosophy 127 identity theory and, 31 persistent vegetative (VS) state, 714–715. See also philosophy of mind vegetative state metacognition and, 290 lesions causing, 715 phonetic imagery personality, 446 thought impact by, 358 personality disorders Phonological Loop consciousness self-regulation in, 682 of WM model, 187 multiple, 379 phonology, 360 representations in, 681 of categories, 367 perturbation paradigm, 334 thought priority over, 364–366 PET. See Positron Emission Tomography phrenology, 25 phantom limb, 344–345 physical systems phantom recollection, 277 intentionality of, 133 phenomenal consciousness, 36 physical theory computationalism on, 118–119, 130 classical vs. quantum, 888 fleeting experiences of, 756 physical world perception and, 138 as mental construct, 22 properties of, 74 physics, 130 reductive vs. causal explanation of, 39 emergenist position impact of, 30 transitive, 76 relational properties and, 42–43 phenomenal perception PI. See Pulvinar non-sensory cortical networks in, 724 Piaget, Jean, 355 phenomenal properties, 39 A-not-B task of, 420 causal efficacy of, 43 on child development, 408 causal laws of, 39 cross-cultural psychology by, 653 phenomenological anthropology Pibram’s Holonomic Theory, 191 embodiment paradigm in, 658 Pimm-Smith, M., 188 phenomenological observation Pinker, S., 609 representationalism and, 44 Planck’s Constant, 884, 906 phenomenological reduction, 69 Plato, 14–15 epoché in, 70 dualism arguments of, 14 in first-person, 70 pleasure for Husserl, 27, 70 subjective state of, 845 purpose of, 70 Pleistocene steps of, 69 evolutionary period, 599 phenomenology, 22, 28. See also existential pluralistic contextualism phenomenology; phenomenal consciousness in cultural psychology, 653 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

subject index 971

poise prime stimulus functional role property of, 44 implicit perception impact of, 210 positive theory of consciousness of, 42 implicit processing of, 215 Positron Emission Tomography (PET), 747, 811 priming. See also perceptual priming in meditation research, 540 in anoetic consciousness, 271 memory studies with, 812–813 cross-modality, 276 of prefrontal cortical function, 420, 441 ERPs in, 232–234 SPECT vs., 540 in hypnotic deafness, 456 Posner, Michael, 199 in implicit memory experiments, 252–253, 272–273, posthypnotic amnesia, 445 279 hypnotic amnesia vs., 451 Primitive Classification (Durkheim, E.), 638 implicit memory in, 451–454 primitive mentality memory impact by, 452 law of participation in, 639 process dissociation paradigm and, 453 logical thought in, 639 psychological research on, 452 The Principles of Psychology (James), 28 spared priming during, 452 prior entry phenomenon state-dependent memory vs., 452 in perception, 331 posthypnotic behavior, 463 problem of other minds posthypnotic suggestion Augustine, St., and, 16 automaticity in, 462–463 intersubjectivity and, 80–81 inevitable evocation and, 463 problem solving and creativity, 290 post-structuralism. See also structuralism procedural knowledge, 695 phenomenology impact on, 68 without awareness, 209 practice theory, 642–647 procedural memory, 810 of Bourdieu, P., 644–645 Process and Reality (Whitehead), 29 pradh¯ana, 91 process control tasks (PC) pragmatic neuroscience, 890 in CLARION, 160 prakriti, 91 process-dissociation procedure (PDP) pram¯ana, 101–102 in implicit memory tests, 274 preclassical Greeks for implicit perception, 220 narrative self-consciousness for, 395 implicit vs. explicit performance in, 223 pre-cognition posthypnotic amnesia and, 453 object-directed intentionality and, 74 relative sensitivity approach vs., 223 preconscious/conscious system, 677 in schizophrenia, 491 in Freud’s topographic mind model, production system (PS) prefrontal cortex, 811 in conscious/unconscious distinction, 153 action role of, 564 programming languages in children, 414 reflection in, 129 consciousness and, 409 pronouns, 363 explicit retrieval and, 822 linguistic expression of, 368 in higher order consciousness, 412–413 property dualism, 39 in LOC model, 421 case for, 39 memory and, 814 of emergentism, 30 memory attempt vs. success and, 816 propositional attitudes memory retrieval control and, 811 propositional content vs., 75 PET of, 420 propositional veracity pre-reflective self-consciousness, 46 in self-consciousness, 619 presentation proprioception re-representation vs., 74 in self experience, 415 Presocratic philosophers, 13–14 protolanguage, 586 reductionist strain in, 14 PS. See production system Pribram, Karl, 184, 190 psychedelic drugs, 3 Pribram’s Holonomic Theory, 184 experiences with, 648 primal impression-retention-protention, 78 psychoanalytic theory primary consciousness, 406 contribution of, 693 primates ego psychology in, 679–680 brain size of, 388, 603 object relations in, 680–682 causality understanding by, 387 representations in, 682 EQ in, 603 subliminal psychodynamic activation in, intelligence in, 851 684–686 suicide among, 618–619 unconscious fantasy in, 679–680 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

972 subject index

psycho-cultural anthropology, 654–656 vector spaces of, 886 psychological adaptation Von Neumann’s, 881 ancient, 601 Von Neuman/Stapp Approach to, 896–898 self-consciousness impact by, 615 Quantum Zero Effect, 906–907 psychological anthropology Piagetain research in, 653 R. See reticular nucleus psychological functions, 182, 716–723 radical embodiment psychological states dynamic large-scale integration and, 739–740 physical conditions and, 26 Rapports du Physique et du Moral de l’Homme psychological studies (Cabanis), 27 of hypnotic deafness, 455 reality on posthypnotic amnesia, 452 appearance vs., 22 psychologists, 386 defined, 360 psychology. See also comparative psychology; formal vs. objective, 18 cross-cultural psychology; cultural psychology; givenness or appearance of, 69 developmental psychology; educational reasoning, 633 psychology; ego psychology; experimental computers capability for, 140 psychology; forensic psychology; neuropsychology; in thought, 370 social psychology recall and recognition adaptation in, 615–616 in amnesic patients, 252 cognitive revolution in, 1 predictions vs. actual, 302 computational theory of, 131 Recall-Judgement-Recognition (RJR) conscious vs. unconscious in, 673 in metacognition, 293 IDA and, 162 recC. See recursive consciousness idealism and, 26 receptivity as scientific discipline, 25 affectivity vs., 74 Psychology from the Empirical Standpoint (Brentano), recognition memory 26 first- vs. third-person perspective on, 488–491 psychopathology first-person perspective on, 482 biological intentional states impact on, 615 item/associative recognition comparison in, 489 neural foundation for, 674 process-dissociation procedure in, 490–491 scientific analysis of, 2 recall/recognition comparison in, 488–491 of shamanism, 648 remember-know procedure in, 482–483 psycho-physical functionalism, 15 recognition vocabulary, 182 psycho-physical law, 26 recollection. See also memory psychophysiological indices familiarity vs., 814 of altered states of consciousness, 449 imagination vs., 73–74 Pulvinar (PI), 785 object-directed intentionality and, 73 pure-thought consciousness, 10 recollection rejection purusa, 91 in metacognition, 297 recursive consciousness (recC) qualia, 118, 138, 633, 839 emergence of, 417 cross-personal comparison of, recursive thought Sloman on, brain size and, 581 quantum jumps, 905–906 in consciousness, 571 quantum law of motion, 885 evolution of, 571, 581–582, 590 quantum mathematics human vs. animal, 580 of causal structures, 883 intentionality impact by, 575 quantum mechanics, 152. See also Copenhagen in language, 576 Quantum Theory in mind, 576–577 action in, 885 in other species, 577–581 quantum theory self vs. other, 590 Bohm Approach to, 894–896 in social behavior, 576 of brain dynamics, 888–889 ToMM and, 579 Eccles-Bech approach to, 902–904 WM and, 575 Jibu-Yasue approach to, 902–904 Reder, L. M., 314 Libet experiments and, 899–900 reductive explanation, 53 of living systems, 887 representationalism and, 44 Pashler’s Analysis in, 898–899 Rees, G. of Penrose-Hameroff, 890–894 globalist argument of, 202 physical vs. phenomenal events in, 888 refC1. See reflective consciousness 1 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

subject index 973

reflection repetition suppression, 231 consciousness impact by, 409 representation(s). See also conscious representation; in programming languages, 129 conscious representation(s); higher order thoughts rule hierarchy formation through, 410 Cof,794 reflective consciousness by children, 415 skills and habits in, 643 collective vs. mental, 655 during sleep, 437 conscious vs. unconscious, 337, 682 reflective consciousness 1 (refC1) Freud, S., on, 676 in infants, 422 of other minds, 870 reflective consciousness 2 (refC2) in personality disorders, 681 in children, 423–424 psychoanalytic concept of, 682 reflective mind representations of, 783 emergence of, 395 of self and other, 682 reflective self-consciousness sensorimotor vs. sensor, 867 mental activity of, 9 simultaneous, 716 self-awareness vs., 76 representational change task regression technique for children, 423 critics of, 220 representational difference in dissociation, in computational models of consciousness, in implicit perception studies, 226 155 reincarnation Representational Redescription model dualism and, 14 of consciousness development, 409 relation representational theories constitutive, non-contingent, of consciousness, 18 relative sensitivity procedure representationalism, 35 dissociation paradigm alternative of, 221–222 case against, 45–46 exhaustiveness assumption in, 223 intellect role in, 92 goal of, 221 master argument for, 44 for implicit perception, 220 phenomenological observation and, 44 problems with, 222 reductive explanation and, 44 relaxation sensuous qualities and, 45–46 Buddhist meditation as, 507 transparency of experience and, 44 REM sleep Tye’s PANIC Theory of, 43–44 behavior disorders of, 708 representing consciousness, 370 dream vs., 709–710 re-representation EEG characteristics of, 709 presentation vs., 74 memory during, 438 research stages of, 438 consciousness vs. memory, 251–252 thought during, 438 resonant cell assemblies. See reverberant cell assemblies remember judgements respiratory drive instructions impact on, 264 measurement of, 791 remembering response compatibility cognitive studies of, 810 in implicit perception, 226, 241 knowing vs., 810 response window neural activity of, 816–818 masked priming and, 216–220 remember-know, 279, 481. See also Know judgements; resting state Remember judgements brain regions during, 709 in amnesic patients, 269 defined, 709 in amnesic subjects, 256 retention and protention criterion shift model in, 255–256 recollection and expectation vs., 78 dual process recognition models and, 256–258 reticular activating system Know judgements in, 267 neuromodulators of, 438–439 measurement issues, instructions role in, 258 reticular nucleus (R), 784 in memory, 253, 254 of thalamus, 792 in metacognition, 293 retrospective confidence judgements, 306 neuroanatomical regions in, 269 retrospective monitoring in recognition memory studies, 482–483 behavior impact by, 309 in schizophrenia, 486, 493 reverberant cell assemblies, 737. See also Hebbian sequential vs. simultaneous, 255 reverberant cell assembly signal detection vs. dual-process model, 257 binding in, 737 studies of, 821 cortex, 737 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

974 subject index

right anterior prefrontal cortex science of consciousness retrieval-related activation in, 813 theoretical frameworks for, Rig-pa Chôg-zhag. See Open Presence scientific analysis RJR. See Recall-Judgement-Recognition of psychopathology, 2 robots, 129–130 scientific law intentionality and, 134 classification of, 29 mental properties and, 124 scientific realism self-model of, 141 consciousness and, 30 Roelofs effect, 333 scientific revolution Rosenthal’s higher order thought theory Descartes and, 17 higher order monitoring theory and, scientific view 46–47 consciousness and, 29 rote learning, 365 subjectivity and, 10 Russell, Bertrand, 28, 355 SDM. See sparse distributed memory Searle, J. R. saccadic suppression, 335 Chinese Room argument of, 141, 143–144 SAD. See shared-attention mechanism on intentionality, 141 Sankhya¯ tradition, 91 second-person Abhidharma vs., 94 in subjectivity, 82 Buddhist tradition vs., 96 See semantic networks (SN), 153 inference in, 101 SELF. See Simple Ego-type Life Form mind in, 91 self perception in, 101 as agent, 380 sources of knowledge in, 101 autonomy of, 385 verbal testimony in, 101 in Buddhism, 93, 503, 526 samatha´ culture impact on, 385 in Buddhism, 504, 506 differentiation of, 415 meanings of, as episodic, 379 Sapir-Whorf hypothesis human awareness of, 573 hard vs. soft versions of, 640–641 in narcissistic disorders, 682 language, consciousness and, 641 as narrator, 384 SAS. See Supervisory Attentional System model neuronal substrates of, 777 Satre, Jean-Paul, 68, 73 as novelist, 380 sattva, rajas, tamas, 91 other vs., 415 Schacter, D., 168–169 of others, 386 DICE model, 184, 185–186 situated, 415 schizophrenia as social, 385 alien control in, 345 subconceptual aspect of, 777 autonoetic awareness in, 484 word usage of, 776–777 awareness in, 484, 491 self and other, 385, 415 false memories in, 485 psychological frame of, 598 familiarity impact by, 491 recursive thought impact on, 590 first-person perspective on, 483–484 representations of, 682 noetic awareness impact by, 491 self conceptualization process-dissociation procedure in, 491 by children, 393 remember-know procedure in, 486, 493 self consciousness source recognition memory task in, 487 in infants, 420 will in, 345 terminology of, 138 word frequency effect in, 484 self continuity Schneider & Pimm-Smith’s Message-Aware Control in children, 421 Mechanism, 184, 188–190 self-agency, 292–293, 380, 384, 415, 436 conscious controller in, 189 self-awareness Schneider, W., 188 cognitive adaptations to, 598 Schopenhauer, A., 24 human vs. animal, 598 Schroedinger equation, 885 McCarthy on, 123–124 science reflective self-awareness vs., 76 19th century, 29 SelfC human activity of, 883 in LOC model, 421 origins of, 13 self-cognition quantum mechanics impact on, 883 apperception and, 103 social prestige of, 29 Tibetan scholars on, 103–104 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

subject index 975

self-concept semantics, 361 in C, 776–778 in computationalism, 131 self-consciousness thought vs., 364, 367 adaptive behavior impact by, 600–601 sensation, 633 consciousness and, 127, 136 access to consciousness via, 37 counterintuitive detriments of, 613–614 anthropology of, 636–637, 656–658 as epiphenomenon, 600–602 assumptions about, 657 intransitive, 76 culture and, 636 mental states of others and, 610 non-visual, 656 mirror self-recognition evidence of, 606–607 physical basis of, 26 motivation and, 601, 613 research on, 657 propositional veracity in, 619 sensorimotor representations psychological adaptation impact on, 615 action and, 339 self-deception and, 613 sensory acuity subjectivity vs., 864 primitive vs. civilized, 637 self-continuity sensory convergence zones, 169 subjective experience of, 422 sensory experience self-experience action and, 339 in object-experience, 79 consciousness and, 359 selfhood culture impact on, 638 development of, 375 Descartes on, 19 self-identity emergence of, 406 continuity of, 847 gamma synchrony requirement for, 754 self-modelling, 134–135 sensory feedback anytime algorithm and, 135 predicted states and, 344 in computational models of consciousness, 122, sensory information 134–139 in communication, 362 exterior vs. interior, 135 interpretation of, 362 free will in, 138 sensory materials of robots, 141 thinking and, 10 self to model in, 128 sensory orders self-organization in child development, 658 spatiotemporal pattern emergence and, 736–738 sensory reactivation self-recognition in memory, 812, 818–821 in chimpanzees, 606–607 sensory systems self-referentiality, 76 of brain, 12 self-referring perspective sensuous qualities in neurodynamic models, 762 representationalism and, 45–46 self-representation sentient soul in Anglo-American philosophy, 50 Gassendi on, 19 of Brentano, 50 serial reaction time (SRT) tasks epistemic argument on, 52 in CLARION, 160 identity and, 51 serotonin intrinsic vs. derivative, 53 memory and, 440 linguistic expressions in, 53 Sevenfold Causal Instructions master argument for, 51–52 in Tibetan Buddhism, 518 mental states in, 50 Shallice’s Supervisory System Model, 184, 186 naturalist account of mental representation vs., 52 episodic memory in, 187 necessary vs. sufﬁcient condition of, 53 language system in, 186 part-whole relation in, 51 supervisory system in, 186 problems for, 52–53 shamanism targetless higher order representations in, 52 neurophenomenology of, 653 type-token distinction in, 50 psychopathology of, 648 varieties of, 50–51 Shannon, Benny, 3 semantic knowledge, 239, 240 shared-attention mechanism (SAD) noetic consciousness and, 263 in Baron-Cohen theory of mind, 611 semantic memory, 810 shortness of breath (SOB) episodic vs., 576–577 source of, 791 knowing and retrieval from, 264–266 subjective sensation of, 790 semantic networks (SN) short-term memory, 187, 753, 849 in conscious/unconscious distinction, 153 SHRDLU system, 129 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

976 subject index

SHSS:C. See Stanford Hypnotic Susceptibility Scale, recursion in, 576 Form C situational forces impact on, 556 sign language, 587 social perception and, 558–560 signal detection spontaneous trait inference effect in, 561 in remember-know paradigm, 257 stereotyping and prejudice in, 560 signal detection theory trait-concept in, 561 in implicit perception, 226 transgression of, 616 Simple Ego-type Life Form, 856 social brain hypothesis, 869–871 simulation, 874–875 social cognition, 869–875 neurophysiology of, 874 chimpanzee vs. human, 608 articulated emulator in, 874–875 in cognitive science, 81 theory-theory and, 81 metacognition and, 297 Single Cell Recording Studies neuroanatomy and, 606 in cats, 439 neuroscience studies of, 871 Single Photon Emission Computed Tomography social cognitive neuroscience, 871–874 (SPECT) social complexity PET vs., 540 brain size and, 870 situated cognition, 866–868 social influence situated self consciousness and, 448 emergence of, 415 social information 6-alternative test brain structure and, 876 for awareness, 233, 234 social judgements. See also social behavior awareness measurement by, 233 amygdala role in, 873 skill learning social learning theories without awareness, 209 evolutionary psychology and, 600 sleep. See also sleep-wake cycle social life brain structures of, 708–711 dialogical basis of, 396 consciousness during, 437 social perception EEG of, 709 attitude activation in, 559 experimental studies of, 443 behavior trait activation impact on, 562 frontoparietal activation during, 724 causal attribution in, 559 input-output gating and, 440 dispositional attribution in, 559 neuronal activity of, 708–711 social psychology, 289 neurophysiology of, 438–440 automatic goal pursuit in, 562 NREM stages of, 437 automatic vs. controlled processes in, 557 reflective consciousness during, 437 cognitive miser in, 558 visual imagery during, 437 conscious choice in, 556 sleep-wake cycle, 437–438 conscious vs. automatic processes within, 557 Sloman, Aaron, 128–129, 134 cortical pathways in, 563 SMA. See Supplementary Motor Area cross-cultural, 636 Smith, Brian Cantwell, 130 dual process models in, 555, 557–562 computational models of consciousness of, 129–130 efficient processing in, 558–560 Criterion of Ultimate Concreteness of, 130 effortful processing in, 557–562 Smith, David Woodruff goal pursuit models in, 557 modal account of self-direction, 51 ideomotor action in, 561 SN. See semantic networks imitative behavior in, 561 SOA. See stimulus onset asynchrony impression formation in, 558, 559–560 SOB. See shortness of breath information overload conditions in, 559, 560 social behavior. See also social perception; social metacognition and, 291 psychology mirror neurons in, 563 in animals, 580–581 monitoring processes in, 556 attitude impact on, 559 nonconscious action control in, 563 in autism, 871 research evolution in, 557 automatic attitude activation model of, 559 self-regulation research in, 557 bystander intervention studies of, 556 studies in, 556 emotion in, 835 unconscious motivation in, 562 external vs. internal forces in, 555 social relations intentionality in, 561–562 agent, theme and goal in, 581 language impact on, 611–612 conscious experience dependence on, 863 nonconscious components of, 557 social sciences obedience experiments in, 556 narrative frameworks in, 377 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

subject index 977

social visual signals, 871 stimulus cognitive neuroscience of, 872 measurement of, 215 eye gaze, 871 stimulus onset asynchrony (SOA) The Society of Mind (Minsky), 123 in masked priming approach, 218 somatoform disorders, 450 stimulus-reinforcement somatosensory awareness in learning, 832, 834 defined, 659 store-consciousness, 97, 111 neuroelectric activity of, 752 stream of consciousness, 28, 77, 381, soul. See also sentient soul 733 Aristotle on, 15 Abhidharma on, 94–95 sound of James, W., 359, 733 defined, 360 multiple simultaneous, 853 sound blindness, 636 temporality and, 77 source amnesia stroke in learning, 453 impact of, 794 source monitoring Stroop Effect, 225 in metacognition, 293 in hypnosis, 457–458, 460 source recognition memory task structural model of consciousness in schizophrenia, 487 of Freud, S., 679 sources of knowledge id, ego, superego in, 679 in Sankhya,¯ 101 morals, values, ideals in, 679 space, 21 topographic mind model vs., 679 kinaesthetic systems and, 79–80 structuralism perceptual experience of, 71, 633 of Levi-Strauss,´ 642–643 visual consciousness in, 203 phenomenology impact on, 68 sparse distributed memory (SDM) STS. See superior temporal sulcus in IDA, 164 study time allocation sparsening in learning, 308 by Mc, 782 subjective experience spatiotemporal pattern emergence of altered states, 449 self-organization and, 736–738 contamination of, 301 species’ biogram, 652 control function of, 314–315 species consciousness, 35 development of, 409–410 SPECT. See Single Photon Emission Computed first-person expertise and, 525–530 Tomography genesis of, 313–314 speech in implicit perception, 212 advantages of, 588 of intentional action, 329, 340 FOXP2 gene and, 586 memory and, 811 gestural theory of, 588 in monkeys, 572 linguistic expressions impact by, neural counterpart of, 499, 525–530 590 neuroscientific study of, 544 manual gesture vs., 588–589 of self-continuity in time, 422 McGurk effect in, 588 subjective feelings neural adaptation of, 586 metacognition and, 292 neuroimaging of, 588 subjective state neurological basis of, 587 of pleasure, 845 Spinoza, Baruch de, 20, 21, 22 subjectivity, 10. See also intersubjectivity Leibniz vs., 21 appearance and, 69 on mind and body, 21 co, 83 on substance, 20 of conscious experience, 864 split-brain phenomenon, 3 defined, 481, 864 in epilepsy, 379 first-person and, 2, 17, 76, 481, 493 spontaneous brain damage, 441 intentionality and, 70, 340 SRT. See serial reaction time tasks inter-, 80–82 Stanford Hypnotic Susceptibility Scale, Form C language and, 573 (SHSS:C), 446 location of, 870 states, 35–36 in neurodynamic models, 756–757 brain, 465–466 as object, 83, 516 of emotion, 831–834 ontological vs. epistemic, 864 intentional, 851 in Open Presence, 515 other directed, 10 role of body in, 78 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

978 subject index

subjectivity (cont.) syntactic binding scientific view and, 10 for consciousness, 853 second-person perspective in, 82 neuronal synchronization impact on, 853 self-consciousness vs., 864 syntax, 361 substrate of, 526 in computationalism, 131 subject-object duality emergence of, 585–586 meditation impact on, 502 informational encapsulated modules and, subliminal activation, 673 197 in politics, 686 in neural networks, 853 unconscious process of, 682–687 syntheses of identity, 77 subliminal politics, 686 System of Logic (Mill), 29 subliminal psychodynamic activation in psychoanalytic theory, 684–686 tantric “Wind” meditation, 505–506 research on, 684–685 targetless thought substance false belief and, 49 extended vs. thinking, 38–39 taxes for Spinoza, 20 in organism design, 836 substance dualism telic states, 47 alternatives to, 20 Template for Action, 907 property dualism vs., 19 of brain, 889 suffering and happiness temporal awareness mental states and, 104 for children, 421, 424–425 root cause of, 503 temporal lobe epilepsy, 442 suggestion effect. See involuntariness temporality suicide inner time-consciousness and, 76–78 in primates, 618–619 stream of consciousness and, 77 superego temporary consciousness, 367 ego vs., 851 tense superior temporal sulcus (STS) importance of, 371 MDI in, 793 thalamic nuclei, 785 Supervisory Attentional System model (SAS), 341, 342 thalamic obstructive strokes supervisory system ILN and, 790 in Shallice’s Supervisory System Model, 186 thalamocortical disconnection Supplementary Motor Area (SMA) vegetative state impact of, 794 anarchic hand sign and, 343–344 thalamocortical system swampman, 865–866 of brain, 192 symbolic systems consciousness and, 715, 775 direct perception impact by, 641 development of, 417 of human brain, 134 in Mc, 779 human vs. nonhuman primates, 641 mental activity and, 776 symbolic thought thalamus, 742 emergence of, 421 amygdala and, 229 symbols anaesthesia role of, 711 in anthropology, 643–644 conscious state impact by, 712 in computational systems, 132, 134 cortex and, 201 of computer, 131 LGN of, 747 direct perception and, 641 nuclei of, 785 implicit perception and, 237 reticular nucleus of, 792 in language, 365 short-latency pathway from, 229 meaning and, 633 simplified image of, 785–786 written, 365 stylized drawing of, 784 sympathy, 10 theory of literary narratives, 377 synchrony, 737 theory of mind, 873 defined, 749 childrens’ false belief in, 611 gamma, 750–754 of chimpanzees, 577–579, 870 in human vs. animals, 749 in developmental and comparative psychology, neural, 532–533, 749–750 610–611 oscillatory, 749 Theory of Mind Mechanism (TOMM) temporal relations of, 750 in Baron-Cohen theory of mind model, 611 synoptic analgesia levels of, 579 brain imaging of, 464–465 recursion and, 579 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

subject index 979

theory-theory time-consciousness, 21. See also inner-time simulation theory and, 81 consciousness; space; temporal awareness; thetic states, temporality Thing and Space (Husserl), 79 brain processes in, 78 Thinks (Lodge), 397 childrens’ experience of, 421, 424–425 third person objective measurement of, 95 of matter, 17 mental travel of, 579 third-person knowledge for mind, 577 childhood development of, 415 object perception and, 77 thought(s), 361–362, 366. See also logical thought perception of, 633 assertoric, 47 perceptual experience of, 71 bodily movement and, 328 retention vs. protention in, 77–78 breathing and, 370 retentional-protentional structure of, 70 in Buddhist epistemology, 107–109 visual consciousness in, 203 categorization of, 364 timing and binding constancy of, 367 in cortical neurons, 743 higher order, 48, 425 tip-of-the-tongue (TOT) experience, 294, 358 human awareness of, 573 direct-access view of, 294–295 imageless, 367 within memory research, 290 inner speech and, 357 phonology connection of, 366–367 Jackendoff on, 364 Titchener circle illusion. See Ebbinghaus illusion language and, 134, 137, 356, 358, 366, TM. See Transcendental Meditation 426 TMS. See transcranial magnetic stimulation meaning of, 370 TOMM. See Theory of Mind Mechanism non-inferential, 47 Tononi, G., 191 nonverbal imagery impact on, 357 Tononi-Edelman occurrent mental states of, 47 Dynamic Core Hypothesis of, 184 organization of, 362–363 globalist argument of, 201 orientations of, 370–371 Tooby, J., 609 outside of consciousness, 370–372 topographic mind model paradigmatic, 376 conscious system in, 676 perception vs., 381 conscious vs. preconscious system in, 676–677 phonetic imagery impact on, 358 defenses in, 678 quasi-perpetual state of, 46 of Freud, S., 675–678 reasoning in, 370 mental processes in, 676 recursive, 577 preconscious system in, 676–677 during REM sleep, 438 strengths vs. weaknesses of, 678 semantics vs., 364, 366–367, 369 structural model of consciousness vs., 679 sensory materials and, 10 unconscious system in, 677–678 targetless, 49 TOT. See tip-of-the-tongue temporary consciousness of, 367 Tower of Hanoi (TOH) task, 717 thinking about, 849, 854 in CLARION, 160 unconscious, 357–358, 370 trait consciousness 35–75 Hz ﬁring conscious representations vs. conscious access in, in cerebral cortex, 169 707 Tibetan Buddhism, 528. See also Buddhism 530 trance logic Chag-zˆog practice of, 510 co-consciousness and, 459 compassion in, 519 in hypnosis, 447–448, 459 Equanimous Exchange of Self and Other in, 518 Transcendental Meditation (TM), 530 Focused Attention in, 511–513 neuroelectric correlates of, 534–538 Giving and Taking in, practitioners of, 530 518 research on, 530 meditations in, 518–519 technique of, 530 meditative states in, 510–511 transcendental phenomenology, 69, 72 Non-Referential Compassion in, 517–519 transcendentalism, 22 Open Presence in, 501, 513–515 of Kant, 23, 69, 494 postures in, 511 transcranial magnetic stimulation (TMS), 758 self-cognition in, 103–104 motor cortex stimulation with, 340 setting in, 511 over FEF, 722 Sevenfold Causal Instructions in, 518 transparency of experience Tumo practice, 525 representationalism and, 44 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

980 subject index

transpersonal anthropology present orientation of, 647 altered states of consciousness and, 647–649 Van Gulick, R. Tsé-cig Ting-ngé-dzin. See Focused Attention higher order global states theory of, 51 Tulving, E. Varela, Francisco, 71, 78 consciousness distinctions of, 254 globalist argument of, 202 Tuner, Victor Vasubandhu, 93, 96 anthropology of experience and, 645–647 vegetative state (VS). See also coma Turing, Alan, 121 ILN impact on, 794 Turing Test, 129 MCS vs., 715 computational models of consciousness and, 140–141 persistent, 714–715 of computationalism, 140–141 thalamocortical disconnection impact on, 794 first-person and, 140 verbal communication intelligence vs. consciousness detection for, 141 in adults, 418 Tye’s PANIC Theory verbal testimony of representationalism, 43–44 in Sankhya,¯ 101 type-token distinction vicissitudes in self-representation theory, 50 consciousness impact of, 390 emotion and, 383 Uddyotakara, 108 vipa´syan¯a unbalanced inhibition Buddhist meditation technique of, 504–505, function loss from, 794–795 506 unconscious. See also dynamic unconscious focus on breath in, 508 affect regulation processes of, 673 memory in, 509 brain processes, 357 vision. See conscious vision children as, 406 visual agnosia, 327, 336 conscious vs., 154, 197, 332, 418, 459–460, 466 visual awareness. See also social visual signals; visual dynamic vs. descriptive, 675 illusions emotion, 458, 680, 855 foveal vs. peripheral, 48 habitus and, 645 gamma synchrony in, 753 by mental architecture vs. motivation, 675 ILN role in, 788 metacognitive judgement impact by, 295 unconscious, 237 phylogenetic history of, 613 The Visual Brain in Action (Milner and Goodale), in social psychology, 562 80 visual awareness, 237, 418 visual consciousness unconscious activation, 673 space and time, 203 unconscious inference, 673 visual extinction unconscious processes neuroimaging studies of, 228 adaptive role of, 613 visual illusions, 346 cognition and, 2 perceptual vs. motor measures of, 327 psychodynamic views of, 208 visual imagery of subliminal activation, 682–687 during sleep, 437 unconscious processes theory social, 871 of Freud, S., 675 visual information unconscious specialized processor amygdala processing of, 873 in GW theory, 195 visual input pathways unconscious system perception vs. action, 563 in Freud’s topographic mind model, 677–678 visual perception, 79–80 language of, 677 unconscious awareness in, 418 preconscious/conscious system vs., 677 visual short-term memory, 753 unconscious thought, 357–358, 370 visual stimuli underconfidence-with-practice (UWP) effect memory for, 811 in JOLs, 304 semantic processing of, 239, 240 utilization behavior visualization anarchic hand sign and, 344 in Buddhist meditation, 505 in intention, 343 in Focused Attention, 511 UWP. See underconfidence-with-practice effect Visuo-spatial Sketchpad of WM model, 187 valuations of percepts vitalism, 24, 25 in language, 358 vocabulary. See recognition vocabulary value vocalization meaning vs., 647 neural control of, 586 P1: JzG 0521857430sind CUFX049/Zelazo 0 521 85743 0 printer: cupusbw February 22, 2007 16:46

subject index 981

volition, 186. See also will WM. See working memory conviction of, 795–796 Woolﬁan consciousness, 381 memory, 277 working memory (WM), 187, 188 voluntary action, 332 Baddeley’s models of, 188 stages of, 327, 346 brain structures of, 563–564 Von Neumann conscious processing in, 188 physical systems for, 887–889 episodic buffer of, 188 quantum theory of, 881 nonconscious operation of, 563–564 Von Neuman/Stapp Phonological Loop of, 187 quantum theory of, 896–898 recursion and, 575 VS. See vegetative state role of, 855 Vygotskyan consciousness, 381–384, 653 Visuo-spatial Sketchpad of, 187 Wundt, William, 26 waking consciousness EEG of, 710 yoga, 530. See also Kundalini Yoga; Patanjali’s˜ Yoga neural correlates of, 449 advanced, 530 Walter Freeman’s Dynamical Systems Approach, 184 Yogac¯ ara,¯ 97 Whitehead, Alfred North, 29 will, 883. See also intention Zeki, S. and action, 328, 329, 332, 445 globalist argument of, 203 behavior and, 564–565 Zen. See Buddhism in goal pursuit, 565 Zen Buddhism, 530 in schizophrenia, 345 zero point problem, 178 self-model and, 138 zombies Winograd, T. conceivability of, 40–41, 864 SHRDLU system of, 129 McCarthy on, 123 Wisconsin Card Sorting Task, 310 self-representing, 53