<<

THE PHILOSOPHICAL DEVELOPMENT OF

A Study of His Published and Unpublished Writings

© Charlotte Vrijen 2007

Illustrations front cover: 1) Ryle’s annotations to Wittgenstein’s Tractatus 2) Notes (miscellaneous) from ‘the red box’, Linacre College Library Illustration back cover: Rodin’s Le Penseur

RIJKSUNIVERSITEIT GRONINGEN

The Philosophical Development of Gilbert Ryle

A Study of His Published and Unpublished Writings

Proefschrift

ter verkrijging van het doctoraat in de Wijsbegeerte aan de Rijksuniversiteit Groningen op gezag van de Rector Magnificus, dr. F. Zwarts, in het openbaar te verdedigen op donderdag 14 juni 2007 om 16.15 uur

door

Charlotte Vrijen

geboren op 11 maart 1978 te Rolde

Promotor: Prof. Dr. L.W. Nauta Copromotor: Prof. Dr. M.R.M. ter Hark

Beoordelingscommissie: Prof. Dr. D.H.K. Pätzold Prof. Dr. B.F. McGuinness Prof. Dr. J.M. Connelly

ISBN: 978-90-367-3049-5

Preface

I am indebted to many people for able to finish this dissertation. First of all I would like to thank my supervisor and promotor Lodi Nauta for his comments on an enormous variety of drafts and for the many stimulating discussions we had throughout the project. He did not limit himself to deeply theoretical discussions but also saved me from grammatical and stylish sloppiness. (He would, for example, have suggested to leave out the ‘enormous’ and ‘many’ above, as well as by far most of the ‘very’’s and ‘greatly’’s in the sentences to come.) After I had already started my new job outside the academic world, Lodi regularly – but always in a pleasant way – reminded me of this other job that still had to be finished. I owe debt to my copromotor Michel ter Hark for his valuable comments on my writings. Brian McGuinness, James Connelly and Detlev Pätzold kindly agreed to be on my reading committee, and gave their approval to this dissertation. I would like to thank them for both. I am also grateful to Brian McGuinness for showing me an early unpublished letter from Ryle to his former tutor H. J. Paton (later published as McGuinness and Vrijen 2006). James Connelly greatly helped me to acquire some of Collingwood’s philosophical work and stimulated me to get in touch with other experts on the . Our discussions about the correspondence between Ryle and Collingwood were very useful as well. One of the things that made working on this dissertation a varied and stimulating was the fact that it involved doing research at Linacre College Library and the Bodleian in Oxford, as well as talking to some of Gilbert Ryle’s former colleagues, friends and relatives. Except for my conversations with Brian McGuinness I also benefited from personal stories and anecdotes by Rom Harré, John Rogers, David Pears, , John North, Jerry Cohen and Tony Palmer. Ryle’s twin-sister’s stepdaughter Janet Beckley and his nephew Michael Ryle and his wife Bridget Ryle kindly invited me into their houses in respectively Dyffed (Wales) and Exmoor. Their anecdotes greatly helped me to get a better image of Ryle as a person. Louise Trevelyan and her fellow librarians at Linacre College Library enabled me to make use of the ‘Gilbert Ryle Collection’ to its full extent. My AIO-colleagues at the Department of in Groningen who made working there a pleasant experience, not only focused on thinking and writing but also on talking, discussions, drinks and fun, are Kim van Gennip, Martine Prange, Casper Zijlstra, Jan-Willem Romeyn, Arend Jagersma, Barteld Kooi, Jan Albert van Laar, Katherine Gardiner, Alice Stollmeyer, Daan Franken, Menno Rol, Marc van Duijn and Constanze Binder. Other members of the department whom I would like to thank for their stimulating discussions and for creating a pleasant environment for doing research are Job van Eck, Pieter Sjoerd Hasper, Karin de Boer, Eddo Evink, Lodi Nauta, Michel ter Hark, Detlev Pätzold, Jeanne Peijnenburg, Jeroen Bartels, Jos Lensink, Theo Kuipers, Erik Krabbe, Martin van Hees, René Boomkens, Hans Harbers, Allard Tamminga, Gyan Otto, Gerda Bosma, Trijnie Hekman, Jorine Janssen, Marga Hids, Kirsten van der Ploeg, Janny Moesker, Miran Huizenga, Benno Ticheler, Eva-Anne le Coultre and Hauke de Vries. The board members of the GAIOO/GRASP! and my former badminton friends at Amor helped me to my to other things than philosophy. My dear friends Barbara van der Pol, Reinier Michiels, Janien Kamps, Monique Heringa, Marco Visser and Liesbeth Schipper I thank for their support and friendship (and for not always asking when this dissertation was going to be finished). Finally I would like to thank my father and Bas, and my mother and Henk for always supporting me and giving me the opportunity to make my own mistakes. My brother Wouter and his girlfriend Gabie for their support, friendship and, occasionally,

talking postcards. And, last but definitely not least, Wouter van Alst for being a great friend and boyfriend at the same .

Contents

INTRODUCTION...... 5

CHAPTER 1 GILBERT RYLE – A BIOGRAPHICAL SKETCH ...... 9 Family Background and Childhood...... 9 Life as a Student...... 10 Christ Church College ...... 11 Ayer and the Circle...... 16 Second World War ...... 17 Magdalen College...... 18 Ryle as a Teacher ...... 23 Ryle as a Person ...... 25

CHAPTER 2 RYLE’S EARLY WRITINGS ...... 29 Position 1: Denotationalism and the Reformulation of Systematically Misleading Expressions...... 31 Influences and Philosophical Context...... 31 Ryle’s Realism and Russell’s Influence Illustrated by Two Papers...... 35 ‘Systematically Misleading Expressions’...... 40 Position 2: Rejecting Denotationalism...... 44 -Mistakes: the of Philosophical Problems...... 46 Philosophy as a Method and the Reductio ad Absurdum as the Ultimate Philosophical Argument ...... 50 Conclusion ...... 53

CHAPTER 3 THE OF MIND: MAIN AIMS, METHOD AND RECEPTION 55 Ryle’s Method: Starting With What We Already Know...... 56 Ryle’s Primary Target...... 59 Theorizing as the Primary and Silent Activity of the Mind...... 62 One : Not Two ...... 64 The Issues Under Discussion ...... 66 Case-study: Self-Knowledge...... 67 A More Accurate Account of the Category-mistake of Cartesian dualism ...... 70 The Reception of ...... 71 Ryle’s Contemporaries ...... 72 The Concept of Mind as a Non-Behaviourist Attack on Dualism...... 82 Ryle’s Behaviourism According to Modern Handbooks ...... 82 Suggestions of Behaviourism in The Concept of Mind...... 83 Ryle’s Rejection of a Behaviouristic Interpretation...... 85

Two Interpretations: Shelley M. Park and Rowland Stout...... 86 Concluding Remarks ...... 91

CHAPTER 4 GILBERT RYLE’S LATER WRITINGS...... 93 Ryle’s Meta-Philosophical Papers...... 94 Continuity in Ryle’s on Philosophical Method ...... 95 The Ordinary Use of ...... 96 Language versus Speech and Words versus Sentences...... 97 Philosophy as an Inter-Level Activity ...... 99 Dilemmas: The ...... 99 Philosophical Dilemmas...... 100 Discussions of Concrete Dilemmas...... 102 ‘It Was To Be’ ...... 103 An Adverbial Account of Thinking...... 105 The Broader of Thinking and the Problems with the Thinking of ‘Le Penseur’...... 107 The Thinking of ‘Le Penseur’...... 110 Criticisms...... 116 Concluding Remarks ...... 118

CHAPTER 5 RYLE AND COLLINGWOOD: THEIR CORRESPONDENCE AND ITS PHILOSOPHICAL CONTEXT ...... 121 Part 1: The Ryle-Collingwood Correspondence: Misinterpretation and Structural Disagreements...... 123 Historical Background...... 124 Definitions, the and Self-reflexivity...... 126 The Essence of the ...... 130 Part 2: Ryle and Collingwood: Similarities Revisited...... 138 Refuting Cartesian Dualism...... 138 and ‘Knowing How’ versus ‘Knowing That’...... 142 Logic of Question and Answer...... 143 Conclusion ...... 145

CHAPTER 6 RYLE AND WITTGENSTEIN...... 147 Historical context ...... 147 Ryle’s Interpretation of Wittgenstein...... 149 Wittgenstein’s ‘Overriding Worry’ ...... 150 The Tractatus ...... 151 From the Tractatus to the Philosophical Investigations ...... 154 Ryle’s Critical remarks...... 157 Ryle’s attempt at tracing the origins of Wittgenstein’s ...... 160 The English Translation of the Tractatus...... 167 Ryle in Relation to Contemporary Wittgenstein Scholarship...... 169 A Comparison Between Ryle and Wittgenstein ...... 170 Peter Hacker: the Main Between Ryle and Wittgenstein is Depth ...... 170 O. K. Bouwsma: Ryle used Arguments whereas Wittgenstein did not...... 171

An Evaluation of Their Differences and Similarities...... 173 Concluding Remarks ...... 175

CHAPTER 7 SUMMARY AND CONCLUDING REMARKS...... 177

APPENDIX 1: RYLE’S WRITINGS – THE COMPLETE LIST ...... 185 Christ Church...... 185 Magdalen College ...... 186 After Ryle’s Retirement...... 189 Posthumous Publications...... 190

APPENDIX 2: GILBERT RYLE COLLECTIONS IN OXFORD...... 193 Minor Ryle Collections Outside Linacre College...... 193 The New Bodleian, Oxford...... 193 The Philosophy Library, Oxford ...... 193 Nuffield College, Oxford...... 194 Queen’s College, Oxford...... 194 Collections outside Oxford ...... 194 Ryle’s books and papers at Linacre College ...... 195

APPENDIX 3: RYLE’S LECTURE NOTES ON THE STRUCTURE OF WITTGENSTEIN’S TRACTATUS...... 199

DUTCH SUMMARY ...... 201

REFERENCES...... 207

Introduction

This dissertation is the first comprehensive study of the writings, both published and unpublished, of Gilbert Ryle (1900-1976), one of the most well-known and influential of the twentieth century.1 His The Concept of Mind became a classic almost as soon as it was published in 1949, and remained essential reading for generations of British philosophers. Ryle was also highly influential as editor of Mind and played an important role in the development of a new curriculum in Oxford and the discipline of philosophy in England. Even though he exercised an immense influence on post-war , his work is not much studied today. If he is mentioned at all, it is for his alleged behaviourism and for some basic ideas and notions such as the category mistake and his distinction between ‘knowing how’ and ‘knowing that’. Thus, in modern handbooks his position is usually labelled as ‘behaviouristic’, since Ryle wanted to show that most of our psychological or mental are dispositional rather than of processes or things occurring in a mysterious entity called the mind. However, Ryle’s philosophical career spans more than 50 years, and he wrote much more than The Concept of Mind. My first aim in this book, therefore, is to trace his philosophical development from the early 1930s till the time of his death in 1976. A second aim follows from this first one: in discussing not only The Concept of Mind but also his work predating and postdating this work, it will become clear that the label ‘behaviourist’ is highly infelicitous. Ryle was not a behaviourist, not even a logical behaviourist (we shall come to these descriptions later). The qualification is important, since it is this label which has done great harm to Ryle’s reputation at a time when behaviourism fell into discredit. By correcting this dominant interpretation of Ryle as a behaviourist I hope to show that his philosophical position and methods still have a lot to say to us. His analysis of what makes sense and in philosophy and how philosophical problems arise out of linguistic confusion has not lost anything of its topicality. In order to trace the philosophical views of Ryle I shall not only study his published writings but shall also use – for the first time – unpublished sources. This may come as a surprise, for it is commonly believed that Ryle had destroyed early drafts, papers and anything that could be used as a kind of by later historians. It was a collection of papers written by Ryle and posthumously edited by René Meyer (1993) that first made it clear to me that at least some material had escaped Ryle’s destructive hand. Soon after I had started my PhD-work in 2001, I found out that there does exist a sort of ‘Ryle Collection’ at Linacre College, virtually unknown to the scholarly world, which consists of a substantial part of what was once his own philosophical library and several unpublished documents. I also found several minor collections of Ryle material – letters, notes, typescripts of conversations, correspondences and papers – at various other places in Oxford, e.g. the New Bodleian, the Philosophy Library and Nuffield College. The correspondences include an important exchange between Ryle and Collingwood. These collections will be described

1 To my knowledge, only three book-length studies of Ryle have appeared: Lucie Antoniol’s book Lire Ryle Aujourd’hui (1993) which does not provide much historical and philosophical background ; Ramoino- Melilli’s Filosofia analisi in Gilbert Ryle (1983); and William Lyons, Gilbert Ryle. An Introduction to his Philosophy (1980) – a useful but typically introductory book. The first two appeared in, respectively, French and Italian, have not been translated into English and have been – probably at least partly for this – generally neglected by British and American commentators.

The Philosophical Development of Gilbert Ryle in Appendix 2, and I shall draw at extensively on it. It is to be expected that more material will turn up. Recently, Brian McGuinness, for example, drew my attention to an early letter written by Ryle in 1926 to his former tutor, H. J. Paton. It reveals Ryle’s early interests and readings. Influenced by Windelband, Ryle tried to transcend the opposition between realism and , an attempt which prefigures his famous attack on Cartesian dualism which was also motivated by a wish to transcend this opposition between realism and idealism or, to put it roughly, the reduction of mind to matter and the reduction of matter to mind. Since the publication-cum-study of this letter has been a co-production of McGuinness and me (McGuinness and Vrijen 2006), I shall leave it out of my dissertation. While this book is, I believe, the first study on this scale of Ryle’s philosophical development, I have not attempted to discuss each and every piece he wrote. I have omitted from consideration his reviews of other philosophers. Ryle is believed to have held a low opinion of historical scholarship: The Concept of Mind, for instance, does not display scholarly erudition and there are hardly any footnotes. Nonetheless, his reviews reveal a deep knowledge of such diverse philosophers as (on whom he wrote a book as well), , Kant, Heidegger, Frege, Russell, Wittgenstein and others. It would, however, require a book of its own to discuss his interpretations of all these philosophers and to examine all possible cross-connections between his and their works. I have made two notable exceptions: Wittgenstein, who was a good friend and colleague of Ryle, and Collingwood, with whom Ryle had an interesting and only recently published exchange of opinions. This dissertation starts with a biographical chapter on Ryle, based on written sources (both published and unpublished) as well as on interviews which I held with several of Ryle’s former friends, students, colleagues and relatives. Chapter 2, 3 and 4 follow a chronological order. They tell the story of Ryle’s earliest philosophical papers, The Concept of Mind and his later writings which focus mainly on thinking and related notions. We will see that Ryle’s position changes significantly in the 1930s but from around 1938 remains largely the same, even though the focus and tone could of course change from time to time, dependent on the problems at stake. Chapter 2 offers an analysis of Ryle’s earliest philosophical writings. Some of these early papers must be read against the background of the discussion between Idealists and Realists, which then dominated the philosophical scene. The chapter describes the development from his early denotational theory of – largely influenced by Russell – to a non-denotational and – as will be explained in this chapter – a ‘non-homogeneous’ theory of meaning. This went hand in hand with a change of focus from a search for an ideal language to a position more similar to that of the later Wittgenstein. The third chapter discusses The Concept of Mind, focusing on Ryle’s method, his main target (i.e. Cartesian dualism), and his attempt to provide what he called the ‘logical ’ of mental concepts (Ryle 1949a, 10). I shall then discuss in some detail the category-mistake which Ryle attributes to Cartesian dualism. This is followed by an examination of some main responses to The Concept of Mind, particularly the reviews by Hampshire (1950), Austin (1950) and Ayer (1970), in order to illustrate how the work was received then. This will also help us to understand Ryle’s later development. Finally, I shall argue that the label ‘behaviourist’, which was put on Ryle both then and now, is misleading. Chapter 4 deals with Ryle’s philosophical work after The Concept of Mind when he feels the need to clarify some of his positions. The theoretical claims which he had presented in his earlier papers such as ‘’ (1938) and ‘Philosophical Arguments’ (1945) were apparently not a sufficient justification for the method employed in The 6

Concept of Mind, which itself hardly made its theoretical framework explicit. Hence, Ryle expanded on his ‘old’ theoretical account of how to do philosophy, which resulted in several theoretical papers in the late fifties and early sixties. Additionally, in 1953 he gave the Tarner Lectures in Cambridge, published as Dilemmas (1954a). The purpose of these lectures was largely similar to that of The Concept of Mind, namely to show what ‘good philosophy’ is by providing examples. The great majority of his later writings, however, is dedicated to the notion of thinking, which kept fascinating and troubling him until the time of his death. In these later papers Ryle sharpens, and sometimes changes, his earlier thoughts from The Concept of Mind. Chapters 5 and 6 offer two case-studies of Ryle’s relationship with two eminent philosophers of the twentieth century: Robin Collingwood and . A discussion of the relationship between Ryle and Wittgenstein (Chapter 6) will, hopefully, shed light on the writings of both. Although resemblances between them have been noticed (e.g. by Peter Hacker, Rom Harré, O. K. Bouwsma and Waismann), this chapter aims at further illuminating similarities and dissimilarities between the two philosophers, using for the first time material from the Ryle archives. The ‘Ryle Collection’ contains three copies of Wittgenstein’s Tractatus, each of them heavily annotated by Ryle, as well as a less heavily annotated copy of Wittgenstein’s Philosophical Investigations. Another ‘new’ source of information, also found in the ‘Ryle Collection’, is an unpublished discussion of Wittgenstein’s early ‘Notes on Logic’ (1913). Obviously, this is a huge theme, and I have limited myself to some general lines of and interpretations without pretending in any way to give a complete and detailed account of Wittgenstein’s views and different Wittgenstein interpretations. The for Collingwood needs more explanation, since his name is not commonly associated with Ryle. Indeed, they never paid any particular attention to each other’s philosophical writings, except on one occasion in 1935. However, a few striking similarities between them (e.g. their rejection of Cartesian dualism and their of intelligence as ‘knowing how’) have been brought to our attention by Collingwood scholars such as Alan Donagan, David Boucher and W. J. van der Dussen. When first reading the correspondence between Ryle and Collingwood in 2002, still unpublished then, I was struck by their mutual misunderstandings and misinterpretations, which – as I shall argue in Chapter 5 – are the result of their different views on the of philosophy. At the same time there was the need to qualify the similarities between Ryle and Collingwood as they were presented by Donagan, Boucher and Van der Dussen. An analysis of the correspondence and the differences and similarities between Ryle and Collingwood will be offered in Chapter 5. Ryle put an important stamp on British philosophy in the twentieth century, but the reverse is true as well: his own philosophical career was shaped by the historical and philosophical developments of his time. In this book these developments will be taken into account, though it would be out of the question, given the nature of my project, to let the background become the foreground. I shall be pleased if my study helps to make Ryle a respectable again, whose views are still worth studying, once we have recognized that the label of behaviourism is misleading. An exposition of Ryle’s philosophical development, based on published and unpublished sources, is the central task I have set myself in this dissertation.

7

The Philosophical Development of Gilbert Ryle

8

Gilbert Ryle – A Biographical Sketch

Chapter 1

Gilbert Ryle – A Biographical Sketch

Family Background and Childhood

The roots of Gilbert Ryle’s family lie in Lancashire (in a place called Ruyhal or Ryehill). His ancestors moved to Cheshire in the 16th Century and became prominent in the spinning and silk industry in Macclesfield (one ancestor married a daughter of Sir Richard Arkwright, who invented the Spinning Jenny). Gilbert’s great-grandfather became a big man in the silk business, owned a private bank and was knighted as Sir John Ryle. He was a Member of Parliament for a short time. His eldest son, John Charles Ryle, was destined to follow his father into the business and , but unfortunately the latter lost nearly all his money through the actions of an untrustworthy partner. As a result, John Charles was sent into the Church. Here he was a great success, and eventually became the first Bishop of Liverpool and a leader of the Evangelical movement in the Church. Bishop Ryle’s eldest son Reginald (Gilbert’s father) became a general practitioner. Influenced by Darwin and the Huxleys he abandoned religion and was a declared agnostic. He married Catherine Scott (herself one of 14 children) and they had ten children; Gilbert and his twin-sister Mary were the eighth and ninth. A legend runs through the Ryle family that Gilbert was born first, but did not look too good and the midwife said: ‘leave him and concentrate on saving the second baby’, but the doctor father was not so pessimistic and, applying simple medical , picked young Gilbert up and swung him round his head to get the blood flowing through the brain – as Gilbert Ryle’s nephew Michael Ryle expressed it – ‘an early application of the concept of mind!’2 Forced upon them by their father, young Gilbert and his siblings always had an ice-cold bath in the morning. Gilbert kept faithful to this habit for a long time. His twin-sister Mary’s adopted daughter Janet would later also be introduced into this ritual by him. Besides being a general practitioner, Gilbert Ryle’s father was an amateur astronomer and a philosopher. This greatly influenced Ryle, who literally read everything he came across. He read many of the chiefly philosophical and semi-philosophical books in his father’s large and diverse library. Another thing that influenced Ryle was his parents’ agnosticism, which stimulated him and his siblings to have critical thoughts about the orthodoxies of Christianity, orthodoxies which naturally prevailed at school. Perhaps partly influenced by this departure from official thought, Ryle had a tendency to deviate from the official lines throughout the rest of his life. When Ryle was a schoolboy at , the First World War naturally played a large role in his life, as in everyone else’s life. In his autobiographical essay Ryle

2 Personal communication with Michael Ryle, 2002.

9

The Philosophical Development of Gilbert Ryle further mentions his early recognized and never overcome disability to study and remember facts.

I remember another master saying, “Ryle, you are very good on theories, but you are very bad on facts.” My attempts to repair this latter weakness were short-lived and unsuccessful. (Ryle 1970, 1)

Life as a Student

In 1919 Ryle went to Oxford with a classical scholarship from Queen’s College. He started working for Classical Honours Moderations, but after a few terms he knew that this was not really his type of study.

I lacked the ear, the nostrils, the palate, and the toes that are needed for excellence in linguistic and literary studies. (Ryle 1970, 2)

His main problem with the taught in Oxford at that time was that it did not provide students with unsolved problems to think about. To Ryle, Classics felt too much like an exercise, an exercise that could be done in preparation for job. Logic was such a real job. Ryle called it ‘a grown-up subject, in which there were still unsolved problems’ (Ryle 1970, 2). Unfortunately, at the time logic was as good as dead in Oxford.

Logic, save for Aristotelian scholarship, was in the doldrums. Little was heard now even of the semi-psychological topics discussed in Bradley’s mis-titled Principles of Logic. Russell’s Principles of Mathematics had been published when I [Ryle] was three; twenty-five years later it and Principia Mathematica were still only the objects of Oxonian pleasantries. The names of Boole, De Morgan, Venn, Jevons, McColl, Frege, Peano, Johnson, and J. M. Keynes did not yet crop up in lectures or discussions. In the bibliography of the Kneales’ The Development of Logic no Oxford entries, save contributions to scholarship, belonged to the half century from Lewis Caroll (1896). (Ryle 1970, 4)

Despite his problems with the classics taught in Oxford, in 1921 Ryle gained first honours in classical honours moderations. After that, he worked for Greats in ancient and modern philosophy, and in Greek and Roman , and in his fifth year he worked for the new school of Modern Greats, the honour school of Philosophy, Politics and (PPE). This school was primarily designed for undergraduates without a Greek-education and was meant to give them a study comparable with Greats, and to qualify them for a diversity of careers. In 1923 Ryle gained first class honours in Greats and in 1924 in Modern Greats. He was invited to sit the Modern Greats finals in order to set a standard for first class performance. The founding of PPE greatly influenced Oxford philosophy. The number of undergraduates reading honours philosophy expanded rapidly, and more teachers in philosophy were appointed. Because of the larger number of teachers per college it became quite common that one of them would be exempt from teaching Plato, and so could specialize, for example, in Kant, something which had been inconceivable before the founding of PPE. And because the demand for teachers grew fast whereas the number of available teachers only grew steadily, teachers often had the to teach whatever subject they wanted. This led to a great variety of subjects taught in Oxford. From the very start of his studies in Oxford, Ryle had a great philosophical curiosity. He became a member of the undergraduate’s Jowett Society and, while not

10

Gilbert Ryle – A Biographical Sketch appreciating Plato’s , which was at the time the ‘bible’ in Oxford, he read a lot of philosophical writings neglected at Oxford. Ryle’s tutor was H. J. Paton, a Crocean who knew how to motivate the young Ryle in doing philosophy.

(…) but for me his [Paton’s] untiring “Now, Ryle, what exactly do you mean by …?” was an admirable spur. (Ryle 1970, 2)

In 1924, probably influenced by his tutor, Ryle started acquiring some reading knowledge of Italian and read some works by and even more by . At the end of his philosophical career he did not think that these Italian Idealists influenced his later thinking much, and wrote that the only thing he could remember having learned from Croce was that philosophical thinking, which in Croce’s work is the highest floor of the theoretical part of the Spirit, is Good. This led Ryle into believing that ’s philosophy, which was after all philosophical thinking, could not be Bad, and that he should read it despite the fact that at the time most Oxford philosophers despised Russell’s philosophy.

Then, since philosophical thinking is a Good Thing, Bertrand Russell’s thinking, in so far as it is philosophical, cannot be a Bad Thing, yet that is what Oxford philosophers ostracise it for being. So, despite them, I ought to look at it, lest I miss something that ought not to be missed. (Ryle, 1970, 3)

Ryle studied Russell and was in fact influenced by him in several ways, as will be discussed later in this dissertation. During his time as a student, Ryle was not only interested in reading books and studying, but he also participated in many other academic activities. He was, for example, captain of the Queen’s College Boat Club in 1923.

Christ Church College

In October 1924, Ryle became junior lecturer in philosophy at Christ Church. At the time, the senior philosophy tutor was H. W. Blunt, who soon retired and made place for M. B. Foster. In 1925 Ryle was elected a Student and tutor in philosophy at Christ Church. From 1937 to 1938 he was junior proctor of the university. In 1924 the situation in Oxford was pretty much the same as during Ryle’s time as a student, except for the changes PPE had initiated. Still nothing of great philosophical interest was going on. The atmosphere in philosophical Oxford was stuffy, philosophers generally stuck to their own ideas, and publications were not encouraged or even really appreciated. Part of the reason for this attitude was the long Oxford tradition of studying classics.

Whereas the majority of the Cambridge philosophers were trained mathematicians or scientists, the Oxford realists had read classics, a training which, it seemed to Ayer, had given them a stubborn disapproval of new ideas. (Rogers 2000, 67)

After Harold Joachim had replaced the only Realist professor, , in 1915 as Wykeham Professor of Logic, all three professors were British Idealists. Harold Joachim was Wykeham Professor of Logic from 1919-1935. The other Idealist professors were John Alexander Stewart, who was White’s Professor of Moral Philosophy 1897-1927, and the

11

The Philosophical Development of Gilbert Ryle second was John Alexander Smith, who was Waynflete Professor of Metaphysical Philosophy 1910-1935. Both were influenced by the Italian Idealists Gentile and Croce. Other important Idealists were J. L. Stocks and H. J. Paton. The early predecessors of the Oxford Idealists were philosophers such as Green, author of the influential Prolegomena to , Edward Caird, Francis Herbert Bradley, Bernard Bosanquet, William Wallace and Robert Lewis Nettleship. Important Realists were Cook Wilson, H. A. Prichard and H. W. B. Joseph. After a long and glorious reign of Idealism under the presence of T. H. Green (1836-1882) and F. H. Bradley (1846-1924), Oxford philosophy had been turned into a parochial matter. The basic issue discussed was between Idealists, who thought that our was determined by mind, and Realists, who held that reality was independent of mind. Unlike their colleague Realists in Cambridge, Oxford Realists rejected epistemological ‘intermediaries’ in , such as sense-data. The historical approach which Oxford philosophy had adopted from humanities did not help to make Oxford an exciting place for philosophers. In 1914, T. S. Eliot, an American graduate student at Merton College, wrote:

I do not think that anyone would come to Oxford to seek for anything very original or subtle in philosophy. (Prest 1993, 218)

The fact that logic was a totally off-side subject which no one took seriously, did not make things better. The Oxford philosophers, such as Collingwood and Joachim, were totally isolated from the much more exciting developments that took place in Cambridge, with philosophers like G. E. Moore and Bertrand Russell, and elsewhere. Henry Price was the first link between Oxford and Cambridge. He went to Cambridge and soon tried to convince the philosophical society in Oxford that it could and should learn from Cambridge. Ryle realised that Price was right:

Soon Oxford’s hermetically conserved atmosphere began to smell stuffy even to ourselves. (Ryle 1970, 5)

In his autobiographical essay Ryle mentions two developments in the second half of the 1920s that helped to change its insulated situation. The first was that Ryle and some other junior philosophy tutors began to attend the annual Joint Sessions of the Mind Association and the , which enabled them to exchange ideas with their colleagues. At one of these Joint Sessions Ryle met Ludwig Wittgenstein, and the two men became friends.3 During the 1930s they occasionally went on walking holidays together. In his autobiographical essay Ryle describes his occasional visits to the Moral Sciences Club at Cambridge as exclusively devoted to Wittgenstein. The visitors of the Club only talked about him and references to other philosophers were jeered away. Although Ryle and Wittgenstein were friends, Ryle considered this attitude unhealthy for the students and for Wittgenstein himself.

It made me resolve, not indeed to be a philosophical polyglot, but to avoid being a monoglot; and most of all to avoid being one Monoglot’s echo, even though he was a genius and a friend. (Ryle 1970, 11)

3 After World War II, the friendship was not as close as before after Ryle had written a very positive article on one of ’s essays, in which Popper had been very negative about Wittgenstein. Wittgenstein could not appreciate this. Furthermore, Wittgenstein accused Ryle of borrowing other men’s thoughts, a type of accusation which was not untypical of Wittgenstein. Cf. Chapter 6. 12

Gilbert Ryle – A Biographical Sketch

It was not difficult for Ryle to avoid being a monoglot and study more thinkers than only Wittgenstein. Both as a student and as a teacher he had to study and teach other philosophers, such as Plato, Aristotle, Descartes, Hume, Kant and Husserl, thoroughly. Ryle considered it fruitful to compare some of the thoughts of these philosophers to Wittgenstein’s, because he thought that these comparisons could often be elucidatory of both. In Chapter 6 the philosophical relation between Ryle and Wittgenstein will be discussed. The second development that made philosophical life in Oxford less closed-off, and more exciting for Ryle, was the establishment of the ‘Wee Teas’ at the end of the 1920s. ‘Wee Teas’ was an informal dinner club with as its members Ryle and five other junior philosophers. Its name was a parody on a famous Scots sect, and also on ‘the Thursday Teas’, which was a weekly tea gathering of teachers of philosophy dominated by the much older senior lecturers. The members of ‘Wee Teas’ met once a fortnight (during term) and during these meetings the host of the evening provided a discussion-opening paper after dinner. The original members were Ryle, Price, W. F. R. Hardie, C. S. Lewis, T. D. Weldon and J. D. Mabbott. At various times, other members were H. H. Cox, O. S. Franks, William C. Kneale, W. G. Maclagan and M. B. Foster. The members of ‘Wee Teas’ tried out new ideas on one another until the middle of the 1960s. During the meetings there was a friendly and stimulating atmosphere.

We never aimed at unanimity or achieved it; but we could try out anything on one another without anyone being shocked or rude or polite. Each of us had five friends and no allies. (Ryle 1970, 6)

Ryle and some of his contemporaries also tried to free themselves from the isolation of Oxford by attending lectures at other universities and teaching off-centre subjects. Kneale attended Husserl’s lectures in Freiburg and was interested in Brentano, and Ryle taught a course on ‘Logical Objectivism: Bolzano, Brentano, Husserl, and Meinong’. Together with Oliver Shewell Franks (1905-1992) he also visited Husserl in 1927 in order to hear more about his ‘system’.4 After this meeting Husserl started sending them notes on his lectures – as shown by a postcard from Husserl in the ‘Ryle Collection’. At the time, British philosophers who were interested in continental philosophy were extremely rare. As Anthony Quinton expressed it:

British and American philosophers showed hardly any interest in what was going on in continental Europe. (…) A solitary and surprising exception is to be found in the earliest interests of that most British of modern British philosophers, Gilbert Ryle. He made a close and sympathetic study of Brentano, Husserl and the phenomenological movement and his first substantial publication was a long review of Heidegger’s Being and Time (…). (Quinton 1998, 61-62)

When Ryle was in his mid-twenties he and many of his contemporaries started to worry about what philosophy really was.

I must have been near my middle twenties when good-humoured fraternal scepticisms about the of my subject showed me that it really was part of my business to be able to tell people, including myself, what philosophy is. (Ryle 1970, 6)

4 The story is that when Ryle and Franks visited Husserl Frau Husserl wanted them to admit that her husband was as great as Plato, but the two men were only willing to admit that he might be as great as Kant.

13

The Philosophical Development of Gilbert Ryle

Philosophy had only become a separate academic subject (separated from theology, economics, and the other sciences) a relatively short time ago. Philosophy itself, as well as the other sciences, was asking for a clear of what philosophy, as distinct from the other sciences, was. First, over-influenced by ‘Socrates’ fruitless hunts for definitions’, as Ryle himself declares in his autobiographical essay, he was only prepared to declare vaguely that what philosophers examine is the meaning of expressions. Only much later was Ryle prepared to state that the relevant questions to philosophy are ‘Why does this or that expression make nonsense?’ and ‘What sort of nonsense does it make?’ instead of ‘What does this or that expression mean?’. Ryle thought that philosophising essentially incorporates argumentation. Therefore, he became interested in the theory of reasoning, primarily in the theory of Meanings (and Nonsense):

I laboured upon the doublets: – , and Extension, Concept and , and Constituents, Objectives and Objects, Facts and Things, Formal Concepts and Real Concepts, Proper Names and Descriptions, and Subjects and Predicates. (Ryle 1970, 7)

When Ryle became a don, he started to teach himself German and started thoroughly to study ’s Logische Untersuchungen, and later also , , and . Husserl interested Ryle because of the fact that he took seriously the opposition between Sense and Nonsense, but he was not happy with Husserl’s way of dealing with this opposition. Ryle was primarily interested in Husserl’s intentionalist, anti-psychologist theory of Meaning. He also read and discussed ’s Sein und Zeit. In the 1930s Ryle was influenced by the , which was formed around 1924 and was led by Schlick. Other members were , , Edgar Zilsel, Béla von Juhos, Felix Kaufmann, Herbert Feigl, Victor Kraft, Philipp Frank, Karl Menger, Kurt Gödel, Hans Hahn, and from 1926 . According to the Circle, the task of philosophy was logical analysis. Traditional philosophy and were to be replaced by the investigation of the logical syntax of the language of science. By clarifying meaningful concepts and propositions, the Circle wanted to lay the foundations for science and mathematics. The of verification, the principle that the meaning of a is its method of verification, played a large role in the Circle’s work. Ryle and his colleagues were not really troubled by the Circle’s demolition of Metaphysics – after all they were not doing metaphysics. They were interested in the Principles of Verifiability and , although they did not start using them as criteria. Because the Vienna Circle equated Metaphysics with Nonsense and Sense with Science, the question arose where Ryle and the other ‘anti-nonsense’ philosophers belonged. What were they doing? Were they dealing with metaphysical sentences? Ryle and others soon began to think that the Circle’s dichotomy ‘Either Science or Nonsense’ did not leave enough room for nuances. Why could it not be the case that the writings of philosophers in the past which did not meet the Circle’s qualifications for Science, contained a few significant things after all? They started to look in the history of philosophy for possible interesting ideas and notions.

Naturally we began, in a patronising mood, by looking for and finding in the Stoics, say, or Locke, primitive adumbrations of our own most prized thoughts. But before long some of them seemed to move more like pioneers than like toddlers, and to talk to us across the ages more like colleagues than like pupils; and then we forgot our pails of whitewash. (Ryle 1970, 11)

14

Gilbert Ryle – A Biographical Sketch

In 1933 their common interest in the development of led , his cousin J. O. Wisdom, Margaret MacDonald, Karl Britton, Max Black, Austin Duncan- Jones, A. J. Ayer and the seniors Richard Braithwaite, and C. A. Mace to approach Blackwell’s to publish a new journal: Analysis. The first number appeared in November 1933. Austin Duncan-Jones became the journal’s editor and the editorial board consisted of Gilbert Ryle, Susan Stebbing and C. A. Mace. Regularly, competitions were organized in Analysis for the best solutions to philosophical problems. In ‘Fifty Years of Philosophy and Philosophers’ (Ryle 1976c, 382-383), Ryle mentioned these competitions to show that, to his great pleasure, contributions to Analysis were often ‘trial-runs’ and not ‘didactic Messages’5. In 1938-1939 Ryle helped to move the Brentano Society from Prague to Oxford. In 1931 the Society was established by the Czech president T. G. Masaryk for the purpose of investigating, securing and further developing the work of his teacher and friend, Franz Brentano. He used part of a donation given to him by the Czecho-Slovakian nation on the occasion of the 1930 jubilee, to provide the financial means for the Brentano Society. In a letter dated 30 December 1938, Georgij Katkov, the Society’s archivist, wrote that sudden political changes threatened the future existence of the Society in Prague. Professor O. Kraus, the Society’s chairman, had been dismissed from his university chair and prevented from any other activity by the cultural policy of the government, which was inspired by the Nazis. And, suddenly the late president’s donation had become a source of danger to the Society’s existence. It was considered to transfer their activities to another country. Oxford University was not unwilling to offer a home for the continuance of the work. Ryle and David Ross were appointed to negotiate with the Brentano Society, and on 31 January 1939 invited the Society to move to Oxford, to be incorporated in the University. But first several problems needed to be conquered; the Czech government was strongly against any transfer of funds to Oxford. Finally they did agree with some sort of co-operation with the Oxford University, and the Brentano Society could then move to Oxford. The originals of the Brentano Archives of Prague were saved and transported by one of the last civilian aeroplanes which left for England, just before the war started. For a while the papers were kept at the Bodleian Library in Oxford. From there they were transferred to Northwestern University, Illinois, where at that time John Brentano (son of Franz Brentano) was teaching . Finally they were deposited at the Houghton Library in Cambridge, .6 Ryle’s most important and influential essay published during the time he lectured at Christ Church was probably ‘Systematically Misleading Expressions’ (1932b), followed by ‘Categories’ (1938). On the former he wrote much later:

5 An example of such a competition was the one in which the question contestants had to try to solve was: ‘If a distraction makes me forget my headache, does it make my head stop aching, or does it only stop me feeling it aching?’ As Ryle reported in Analysis 14, 1954: ‘The selection of my short list was not easy. Some competitors helped me by spending a lot of their short in pronouncing generalities about the Nature of a Philosophical Puzzle, or about the Deficiencies of Ordinary Usage and the like. I excluded them without hesitation. They were asked not to talk about doing philosophy, but to do a bit of it.’ (Ryle 1954c, 51) The ones that were short-listed ‘saw that the concepts of attending, heeding, noticing and concentrating are cardinal to the puzzle. But though they operated efficiently with these concepts, they did not operate enough upon them. Much more, indeed everything, remains to be said, and a good deal requires to be unsaid about this family of concepts. I had hoped to steal an idea or two on this matter from the entries, but have not been lucky.’ (Ryle 1954c, 51-52) 6 In the New Bodleian Library in Oxford there is an unpublished letter on the transfer (Top Oxon. C. 870, fol. 18- 34). Wilhelm Baumgartner, co-founder of the Internationale Franz Brentano Gesellschaft was so kind as to give me some additional information.

15

The Philosophical Development of Gilbert Ryle

It is certain that when I wrote “Systematically Misleading Expressions” I was still under the direct influence of the notion of an “ideal language”—a doctrine according to which there were a certain number of logical forms which one could somehow dig up by scratching away at the earth which covered them. (Ryle 1961b, 305)

When he wrote ‘Systematically Misleading Expressions’ Ryle clearly believed in objective logical forms of facts. Therefore, his early position is often referred to as logical atomism7. As early as in his ‘Categories’ (1938), Ryle abandoned the idea of objective logical forms of facts, as well as that of an ideal language. However, in the line of thought of ‘Systematically Misleading Expressions’ he still focused on syntactical and grammatical differences between expressions and viewed philosophy as trying to apprehend category differences between expressions. According to Ryle, to know to what category an expression belongs is to answer the questions, ‘In what sorts of non-absurd sentences and in what positions in them can the expression ‘so and so’ enter?’ (Ryle 1938, 180) and ‘What sorts of sentences would be rendered absurd by the substitution for one of their - factors of the expression ‘so and so’?’ (Ibid.). We shall discuss these and other early papers in Chapter 2. The style and method Ryle first developed in this period, which is most clearly present in ‘Systematically Misleading Expressions’, became characteristic of his philosophy. Ryle uses a lot of everyday examples to make his point, in deliberate contrast to the abstract and technical style of earlier generations of philosophers (Idealists and Realists) in Oxford. This background should be kept in kind when people later criticize him for using metaphors, analogies and examples too often.

Ayer and the Vienna Circle

In 1929 Ayer first met Ryle, who would become his tutor in philosophy at Christ Church. Ryle soon started to appreciate Ayer’s qualities.

(…) I must have shown some critical acumen, since it was not long before Ryle gave me to understand that he thought I had the makings of a philosopher. (Ayer 1977, 80)

For Austin Ryle was an important example of someone who knew more of philosophy than what was taught in Oxford only:

It was through Gilbert Ryle that I first learned the work of Ludwig Wittgenstein. Though the English translation of Wittgenstein’s Tractatus Logico-Philosophicus had been published in 1922, and he himself had been working in Cambridge since 1929, his ideas had hardly penetrated to Oxford. I think it likely that many of the college tutors had not even heard of him. Here too Gilbert was one of the exceptions and he put me on to the Tractatus early in my final year. (Ayer 1977, 115)

In Ayer’s last term as an undergraduate (probably Spring term 1932) he read a paper on Wittgenstein’s Tractatus to the Jowett Society:

I believe, however, that I am right in thinking that this was the first occasion in Oxford on which there had been any public discussion of Wittgenstein’s work. (Ayer 1977, 119)

7 Ryle himself in fact never described his early position in terms of ‘’. 16

Gilbert Ryle – A Biographical Sketch

After Ayer’s examination for Greats he became a lecturer at Christ Church. Ryle took him to Cambridge to introduce him to Wittgenstein. There were enough lecturers at the time, and Ayer was given permission to have two lecture-free terms. He would have liked to spend these terms in Cambridge to learn from Wittgenstein. Ryle suggested, however, that Ayer went to Vienna in order to learn as much as possible about the Vienna Circle, since hardly anything was known about it in England. Ryle had met Schlick at an international congress in Oxford in 1930 and had been enormously impressed by the leader of the Vienna Circle. Ayer gave in and went to Vienna in December 1932, only after Ryle had assured him that his deficient knowledge of German would not be a real problem. Schlick spoke English well and Ayer would have time enough to learn some German. He was immediately introduced into the Circle and regularly reported to Ryle.8 Ayer attended Schlick’s lectures at the , which were on the . He also attended the Vienna Circle meetings, held once a fortnight on Thursday. Another visitor of the Circle at that time was the American philosopher W. V. O. Quine. Wittgenstein was also in Vienna at that time and Ayer was interested in learning from him too. However, Wittgenstein refused to attend the Circle’s meetings and only communicated with the Circle through Schlick and Waismann, and this was how Ayer was informed of Witgensteinian developments. Waismann and Wittgenstein even planned to write a book together in which Waismann would write down Wittgenstein’s ideas in a systematic way, but because of many disagreements Wittgenstein ended their contact after Schlick’s death in 1936. In 1937 Waismann left for Cambridge with the manuscript and kept reworking the material, here and there including his own ideas. He did not succeed in publishing it during his life. Many years later Ryle would play an important role in preventing this manuscript from being lost forever. Ryle visited Waismann’s appartment only a few days after his death in Oxford in 1959 when he saw Waismann’s landlady throw away large amounts of paper into the trash. He decided to have a look and found papers, note cards and other material written by Waismann. He hailed a passing postman’s van and obtained two large mailbags in which he stuffed Waismann’s writings and stored them in Magdalen Tower. Later, Rom Harré worked on the Waismann legacy and in 1965 he published Friedrich Waismann’s, The Principles of Linguistic Philosophy.

Second World War

From the very beginning of the Second World War men were recruited to the armed forces. But university teaching was on the list of reserved occupations. When university teachers were over twenty-five they were not conscripted. Later this age was raised to thirty. So, at

8 One of the letters Ayer wrote to Ryle during his stay in Vienna has been published in the Linacre Journal (1999). Something that becomes clear from this letter is the Circle’s adoration of Wittgenstein. ‘Of Weisman’s [sic] philosophy I have learned very little except that he thinks Ramsey solved the problem of Universals. He is the expert on Wittgenstein, about whom he has been writing a book for the last five or ten years. Wittgenstein is treated here as a second Pythagoras and Weismann [sic] is the high priest of the cult. If one praises Ramsey they say ‘of course he got it all from Wittgenstein’ or Moore ‘we hear he attends all of Wittgenstein’s lectures but doesn’t always understand them’ or Braithwaite ‘Wittgenstein does not speak well of his philosophy’. They say here that Wittgenstein has changed his views a great deal since the publication of the Tractatus. It has only been divulged to a few what these changes are, and they are as secretive about it all as the initiates of a mystery religion.’ (Ayer 1933, 31)

17

The Philosophical Development of Gilbert Ryle first nothing really changed in Oxford because all teachers were over twenty-five and most by far over thirty. The Brigade of Guards did not participate in the normal process of recruitment, but had the privilege to choose its own officers. Jock Lewis, a former student of Ryle’s who had become an officer in the Welsh Guards, introduced Ryle to the colonel of his regiment. Ryle was offered a commission, despite the fact that his ancestors were not Welsh. He joined the Welsh Guards together with Ayer and Eric Gray, who was at the time a tutor in Roman History at Christ Church. Ryle, who was the oldest, joined his regiment directly as an officer, but Ayer and Gray had to go through a two-month training first. Many of Ryle’s fellow officers were artists and writers, e.g. Richard Llewellyn Lloyd, who was a novelist, and the painters and Simon Elwes. The officers did not have to occupy themselves with taking care of their equipment or their uniforms. They had so-called ‘batmen’ (soldier-servants) to take care of these jobs. Later, Ryle used to say that he had had a ‘good’ war. He seemed to have liked his time as an officer.

Gilbert was naturally at home in the officer’s mess, where he was called the Professor and very much liked and respected. (Ayer, 1977, 236)

First, Ryle stayed in the training battalion. Later, he joined the historian Hugh Trevor- Roper in a counter-espionage unit. Stuart Hampshire also worked within this unit. Ayer once claimed that perhaps the refreshing break from teaching philosophy positively contributed to Ryle’s later philosophical productivity.

He [Ryle] found intellectual refreshment in these breaks in the routine of academic life and it was almost immediately after the war that he began to write his masterpiece, The Concept of Mind. (Ayer 1977, 236-237)

When Ryle finally went home in 1945, he had the rank of major.

Magdalen College

After the war, the centre of analytic philosophy was moving from Cambridge towards Oxford. Undoubtedly, this had something to do with the unexpectedly high number of graduate students who came to study in Oxford9 and with one of Ryle’s greatest contributions to Oxford philosophy, namely his role in the invention of the B.Phil. This new degree was introduced in 1946 and first examined in 1948. Ryle and some of his colleagues, especially Mabbott, thought that both the D.Phil. and the B.Litt. had many weaknesses. The B.Litt. encouraged no discussions with other philosophers besides one’s supervisor. There were no relevant lectures either. All the students working on their B.Litt. were writing a thesis on a specialized subject in total isolation. And the statute of the D.Phil. first required that the thesis written was ‘fit for publication’, but this appeared to be too demanding, e.g. Collingwood failed every candidate he examined (cf. Mabbott 1986, 145). Then the requirement was altered in such a way that a thesis should be ‘in a form fit for publication’, which was a very vague criterion, leading to an enormous variation in the examination. Furthermore, both the B.Litt. and the D.Phil. were too narrow in order to be a proper preparation for teaching philosophy.

9 In pre-war Oxford (post)graduate studies in philosophy had been rare. Usually, undergraduates who did very well in their bachelor examinations were directly appointed to lectureships or even fellowships. (Rée 1993, 6) 18

Gilbert Ryle – A Biographical Sketch

Besides, these degrees were seen as superfluous for Oxford graduates because the Final Honour Schools offered them a proper preparation for teaching philosophy. The many disadvantages of the B.Litt. and the D.Phil. to philosophy made Mabbott and Ryle decide to try to found a new degree, the B.Phil. Mabbott’s task was primarily to lay out the main aims of the degree, whereas Ryle took great efforts in getting the degree accepted by the conservative establishment. The B.Phil. became a two-year course. Students had to choose one or two items from a given list of ‘Chosen Authorities’, such as Plato, Aristotle, etc. They further had to pick one or two (wide) fields of philosophy they were particularly interested in, e.g. moral philosophy, the theory of knowledge, optional periods in the history of science, or logic. This was to widen the students’ reading and to give them the opportunity to discuss problems with other graduates. The finals consisted of three examination papers and a thesis. The B.Phil. became a great success. It attracted many overseas students and enormously stimulated the philosophical activity of the tutors at Oxford University. Philosophy finally came to evolve into, what Ryle later called, ‘a full-scale academic profession’:

It now has its separate academic departments, with its own offices, secretaries, seminar-rooms, blackboards and library-facilities; it is managed by specially qualified and duly appointed Professors, Readers, Lecturers and Instructors; it has its proprietary manuals, journals, encyclopaedias and dictionaries; it has its conferences, colloquia, symposia, teach-ins, broadcasts and even quizzes; it is a field for examinations, degrees, higher degrees, scholarships and prizes; it distributes fees, stipends, pensions, royalties and honours; and so on and so on. From the point of view of a Vice-Chancellor or Registrar it is just one academic subject among others (…). (Ryle 1976c, 387)

From 1945 until his retirement in 1968, Ryle was Waynflete Professor of Metaphysical Philosophy and fellow of Magdalen College. From 1948 to 1971 he also edited Mind. During his time at Magdalen Ryle made a lot of effort to encourage young and unknown philosophers. This was also expressed in his editorship of Mind. Warnock, when discussing Ryle’s editorship, noticed the following development.

Two things, I think, strike one here. In 1948 the contributors of articles were, with only one or two exceptions, inhabitants of the British Isles. In 1971, two-thirds of them were not. Further, the contributors of 1948 were, on the whole, ‘well-known’, the bearers of names already respectedly familiar; those of 1971 were on the whole – and I hope they will forgive me for saying this – not. (Warnock 1976, 54)

Warnock thinks that these developments were partly ‘just because of the way things have gone’, but also to a large extent a deliberate policy of Ryle. Ryle tried to convince his colleagues from early on that they could and should learn from philosophers outside Oxford, and he was not the kind of person who would prefer essays of the elderly and respected over those of young and unknown philosophers. This can be illustrated by letters Ryle wrote to Russell and Popper. On February 24th, 1965 Ryle wrote to Popper:

The April MIND went to bed before I got your letter, but I am afraid I couldn’t have done anything about it as I have a very long string of accepted discussion notes ahead of yours in the queue. I don’t mind going [sic] minor shuffles with the queue but, except in very special circumstances, I have scruples about doing major ones.10

10 Hoover Institution Archives, Collection title: K. Popper; Box number: 89; Folder ID: 4

19

The Philosophical Development of Gilbert Ryle

He tended to accept the essays of those who needed it most, namely the not yet well-known philosophers whose essays were not automatically accepted by journals. Some people criticised Ryle for this attitude, but it is a fact that the number of members of the Mind Association more than trebled in the period he edited Mind. Another characteristic of Mind during Ryle’s editorship was the lack of footnotes. Ryle thought that philosophers should have the ability to write essays without the extensive use of footnotes. Ryle continuously strove for readable texts without technical language. In a letter dated April 15th 1964 Ryle wrote to Popper:

May I editorially beg you to eliminate all footnotes other than references from your reply? I am always having to tell young contributors that footnotes containing additions to the argument are sheer bad writings as well as being a nuisance to editors and compositors, and it’s embarrassing if they can come back and say “Well, Professor Popper does”.11

In 1976, five years after Ryle’s editorship had come to an end, Warnock remarked:

And I have even observed that, in some very recent issues, some articles have actually appeared with footnotes. As Vosper used to remark to Bradbury Fisher, ‘it would not have done for the duke’. (Warnock 1976, 56)

Ryle did not want the editorship of Mind to remain the relatively ponderous administrative machinery it had been under Moore. Therefore, he reduced the staff of Mind to the Editor and a secretary. Under Ryle’s editorship, Mind did not have committees, review editors or referees, simply because he thought that they would unnecessarily increase Mind’s administrative process, and Ryle’s ways of direct communication worked just fine without them.

I cannot easily believe that Mind will ever be edited again with such economy of administrative machinery, so incredibly little fuss. I should be glad to be proved wrong. (Warnock 1976, 56)

A consequence of Ryle running Mind almost entirely on his own was that he had to read an enormous number of books and articles in a short period of time. He became very good at this and invented the expression ‘Ryling a book’, which meant reading it very fast, only in order to get an idea of what it was about, what its main themes were, and, most importantly, whether it was worth a review or should be published in Mind. Michael Ryle remembers seeing his uncle with piles of letters and articles which he had to deal with at home during holidays or over the weekend. During Ryle’s editorship there was an incident which attracted the attention of almost the entire Oxford and Cambridge academic community. Ryle refused to have the book Words and Things, written by in 1959, reviewed in Mind. Russell had written an introduction to the book and was infuriated when he heard that Ryle had written a letter to the publisher, Victor Gollancz Ltd., in which he stated his refusal to have the book reviewed. Ryle found the book abusive and claimed that it could therefore not be considered a contribution to academic philosophy. Russell wrote a letter to The Times in which he claimed that Gellner simply disagreed with the opinions which he discussed, but that the book was not abusive. He accused Ryle of refusing to review the book, simply because he did not like Gellner’s ideas.

11 Hoover Institution Archives, Collection title: K. Popper; Box number: 345; Folder ID: 20 20

Gilbert Ryle – A Biographical Sketch

If all books that do not endorse Professor Ryle’s opinions are to be boycotted in the pages of Mind, that hitherto respected periodical will sink to the level of the mutual-admiration organ of a coterie. All who care for the repute of British philosophy will regret this.12

Ryle responded in The Times only four days after Russell’s letter had appeared, claiming that:

In the book referred to by Earl Russell (…) about 100 imputations of disingenuousness are made against a number of identifiable teachers of philosophy; about half of these occur on pages 159- 92 and 237-65.13

The letters by Russell and Ryle in The Times provoked letters by others, such as Gellner himself; John Wisdom, a Cambridge Professor of philosophy; Brian McGuinness, a fellow of Queen’s College, Oxford; and Conrad Dehn, a correspondent. By far most of the correspondents agreed with Russell and thought that Ryle should have had Gellner’s book reviewed in Mind, but Ryle refused to do so. In 1945, Ryle’s former tutor, Paton, who was then editor of a new series called Hutchinson’s Philosophical Library, invited him to contribute to it. Ryle agreed without knowing what exactly he was going to write about. He did know that he wanted to give an answer to the question ‘What constitutes a philosophical problem; and what is the way to resolve it?’, a question which had preoccupied him in the 1920s and the 1930s, by giving an application of it. Ryle thought it was time to apply the answer to that question to a real, current problem, in order to give an example of the method at work. First Ryle thought he would choose the problem of the Freedom of the Will, but finally he picked the Concept of Mind as the most suitable philosophical crux.14 Thus, The Concept of Mind was written with a meta-philosophical purpose. The Tarner Lectures, published as Dilemmas, which Ryle wrote five years later were meant to provide more, and especially more diverse, examples of his method at work. In analytic philosophy, The Concept of Mind became one of the most influential philosophical works for decades, and definitely Ryle’s best-known and most influential work. The German analytic philosopher Eike Von Savigny even claimed in 1974:

Das einfluβreichste Buch in der Philosophie der normalen Sprache haben nicht Wittgenstein und nicht Austin geschrieben, sondern Ryle; das äuβert sich unter anderem darin, daβ man seit dem Erscheinen des Concept of Mind jeder Arbeit, die zur Philosophie des Geistes von einem Philosophen dieser Richtung – und nicht nur dieser Richtung – geschrieben worden ist, anmerkt, daβ es dieses Buch gibt. (Von Savigny 1974, 123)15

According to Ayer, in the first years after the war Ryle was the strongest philosophical influence in Oxford, even before the publication of The Concept of Mind in 1949 (Ayer 1977, 295-296). He mentioned that Ryle had written an unpublished paper called ‘The

12 Russell, as quoted in Mehta 1963 (p. 12). 13 Ryle, as quoted in Mehta 1963 (p. 13). 14 This title only occurred to Ryle after writing the book. 15 ‘The most influential book in ordinary language philosophy was not written by Wittgenstein or Austin, but by Ryle; this can be shown, among other things, by the fact that since the appearance of The Concept of Mind every work about the , written by a philosopher of this direction, and not only this direction, notices its existence.’ (free translation by C. V.)

21

The Philosophical Development of Gilbert Ryle

Mind is its own Place’16 in which he had already formulated the central issues that later appeared in The Concept of Mind. He had sent this paper to Ayer in 1946. After The Concept of Mind and till his death, Ryle primarily focused on the notion of Thinking. In the second volume of his Collected Papers, he provides the reason for this choice.

But I have latterly been concentrating heavily on this particular theme for the simple reason that it has turned out to be at once a still intractable and a progressively ramifying maze. Only a short confrontation with the theme suffices to make it clear that and why no account of Thinking of a Behaviourist coloration will do, and also why no account of a Cartesian coloration will do either. (Ryle 1971b, viii)

Because of his editorship of Mind, and partly also because of the success of The Concept of Mind, Ryle became a dominant and powerful figure in English philosophy. He had a great number of graduate students in Oxford and was almost always asked for his opinion when academic staff had to be appointed.

And for years he [Ryle] had the reputation of being the chief king-maker when it came to academic appointments in philosophy, not just at Oxford but in universities all over the country. (Magee 1971, 100)

According to John Rogers, at the time, an extremely large percentage of the staff at English universities had been pupils of Ryle. Although Ryle did not have real ‘Ryleans’, real followers, it is clear that whole generations were influenced by him. Ryle’s enormous influence and high reputation is testified by the many academic honours he received. He was a foreign honorary member of the American Academy of Arts and Sciences (1968), an honorary Student of Christ Church, and an honorary fellow of Queen’s and Magdalen Colleges. He was awarded honorary doctorates by the universities of Birmingham, Warwick, Sussex, Hull, Keele, Trent (Ontario), and College (Dublin). Ryle retired in 1968 and from then on lived in Islip, near Oxford, with his twin- sister Mary. They had to move to a bungalow because Mary had arthritis. By that time, Ryle had cataracts and could not drive, but Islip was well-situated on bus routes to Oxford. His best friend Mabbott also lived in Islip and they did their morning walks together, several times a week. After their walks they would drink a ‘Prof’s. Special’, a secret between Ryle and the landlord, in an inn called the Swan. Upon his retirement Ryle donated his books and papers to the Linacre College Library17. They are currently still kept there as part of the ‘Gilbert Ryle Collection’. Ryle did not stop doing philosophy after his retirement and remained philosophically active until his death. He rendered his last service to Oxford philosophy by fighting for the founding of the Philosophical Centre. In his tribute to Ryle, J. D. Mabbott reminded of this great contribution to philosophy in Oxford.

His services to Oxford philosophers were crowned by his fight for a building as their centre with a library, seminar rooms, a meeting room and so on. It is sad that he did not see its complete

16 Unfortunately this paper is missing. It is not in the ‘Ryle Collection’ at Linacre College and not in the Ayer estate either. 17 Ryle was one of the Founding Delegates of Linacre College, and honorary fellow. He was also a close friend of Rom Harré, who was at the time Linacre College’s librarian. Moreover, Linacre urgently needed books. All of this may have prompted Ryle to donating his books and papers to Linacre College. 22

Gilbert Ryle – A Biographical Sketch

installation in its fine new home at 10 Merton Street; but it is most appropriate that a room there has been designated ‘Ryle Room’. (Mabbott 1976)

Ryle had his first stroke in 1975. He suddenly lost his speech and the use of his left arm and leg. After ten minutes the symptoms disappeared. From that moment on, one of Ryle’s biggest fears was to be perfectly conscious without being able to talk, move, or express himself in any other way. A year later he suffered a second and third stroke and died on 6 October 1976 on one of many walking holidays with Mabbott, who described Ryle’s last days in his Oxford (Mabbott 1986). Ryle left his library, that is, what remained after his first donation upon his retirement in 1968, to, once more, Linacre College and he left a considerable amount of money to Hertford College, which was at the time relatively poor compared with the other colleges. ‘The Gilbert Ryle Collection’ is still part of Linacre College Library.18 Ryle was a philosopher with his own unique method of doing philosophy and – as is not often mentioned – a great interest in the works of other philosophers, from ancient philosophers to his own contemporaries. During his philosophical career Ryle taught and wrote about Plato, Aristotle, Russell, Frege, Husserl, Brentano, Wittgenstein and many other philosophers. His interest was not in the first place of a historical nature; it was rather placing the ideas of other philosophers in the perspective of theories and discussions of Ryle’s own time that fascinated him. As a result, this gave rise to accusations of anachronism, but it can be doubted whether Ryle would have been bothered by this, since he was primarily interested in other philosophers’ for today’s problems. This, however, did not mean that his interpretations could not be (sometimes surprisingly) accurate and clear. In his book From Frege to Wittgenstein (2002) Erich H. Reck for example mentions that Ryle ‘offered a rather eloquent summary of Wittgenstein’s criticisms of an atomistic theory of meaning (…)’ in Ryle 1962d and that ‘Aside from a few notable exceptions, hardly anyone writing on the Tractatus over the subsequent several decades seems to have noticed early Wittgenstein’s repudiation of an atomist theory of meaning or noticed that Ryle noticed it.’ (Reck 2002, 447) In the same book Reck notes that it is surprising that Ryle ‘should have (at least occasionally) had such a firm grip on “Frege’s difficult but crucial point”’. (Reck 2002, 433) Ryle’s substantial knowledge of other philosophers is not only clear from his ‘historical’ papers, books and lectures, but is also reflected in the ‘Ryle Collection’. Some of its books are heavily annotated, which makes them an interesting object of study. Ryle’s annotations largely consist of attempts to trace influences of other philosophers. His annotations show his knowledge of the history of philosophy and give away thorough scholarship of numerous philosophers, from Aristotle and Plato to Heidegger, Husserl, Russell and Wittgenstein. Since his annotations are not seldom explanations of the writings of a philosopher in terms of examples or discussions that only became relevant long after they were written, again the term ‘anachronism’ comes up.19

Ryle as a Teacher

Ryle was a remarkable teacher. The method used in his lectures closely resembled that of his writings. He used many everyday life examples and was patient with his audience.

18 Cf. Appendix 2. 19 Cf. Appendix 2.

23

The Philosophical Development of Gilbert Ryle

When he noticed that something was not clear to his students he would reformulate his argument and rephrase his examples, or try to find others. He developed his thoughts throughout the lecture. To others, it sometimes seemed as if he was repeating himself over and over. In a way this was true, but Ryle refined and sharpened his thoughts continuously and adapted them, and his examples, to the public. During seminars, however, he would not say much. Ryle was not nearly as good in discussions as he was at recognizing conceptual differences. When counterexamples were presented during a seminar he would often say that he did not quite know how to answer them and then come up with some half-way house which looked like the counterexample but had a fault somewhere. Ryle would then reject the half-way house, ignoring the original counterexample, often not to the satisfaction of the student who had presented it, as his former student Rom Harré recalled.20 He did not want to teach his students a theory, and certainly did not want followers, but tried to teach his audience an activity. He taught them how to do philosophy, i.e. what methods they had to use in order to reach sound and sharp arguments.

He did not want philosophers to be looking forward to erecting intellectual edifices that would stand for decades, or even – God forbid – for centuries. (Moravcsik 1977, 3)

Ryle usually himself selected the PhD students who wanted to do philosophy in Oxford, in which he largely relied on his intuition. Intelligence and a feeling for philosophy were more important than scholarship. In 1960 in a letter to Professor Julius Moravcsik (at the time one of his students) Ryle responded to a manuscript of Moravcsik’s in a way that was in the spirit of his teachings and his approach to students in general.

(…) In particular, the B.Phil. Thesis frame of mind – “must keep both examiners from thinking up things that I ought to have known or considered” inevitably infects the first things that one writes with the of airing an idea or discovery which, one hopes, may make a difference to the progress of the subject. (and not to the award of the degree!) Go through the thing, asking yourself over each paragraph “is this meant to silence Cherniss, or is it part of what I want Mr. Snooks to lap up and digest?” (Mr. Snooks being a youngish man who is inquisitive about Plato, say.) This may mean not merely heavy excisions but actual re-writing. If so, then a) go through the piece, with a card in your hand, jotting down perhaps 5, perhaps 8 moves, that have got to be in. b) Then hide the MS, and write a third Programme talk bringing in those moves. c) Then, and then only, look at your MS and “add”? the indispensable citations, references, crucial arguments and points! And: THERE YOU ARE! (Ryle, quoted in Moravcsik 1977, 4)

Ryle used high standards and expected his students to be extremely precise in formulating their thoughts and arguments. He tried to teach them to be critical of their own formulations and to make sure that they actually stated what they meant. His exam questions were always a challenge and hardly ever what his students expected. They were usually aimed at making students philosophise themselves, instead of making them repeat arguments of authorities in philosophy. In his paper, Moravcsik quotes as Ryle’s most famous B.Phil. exam question: ‘If existence is not a predicate, what is it?’ (Moravcsik 1977, 5) A famous exam question Michael Ryle remembers is: ‘Q.5 Is Question 5 a fair question?’. Ryle always encouraged his students to have many philosophical interests, and to avoid being a monoglot. He himself taught – as one of the few – philosophers who were not

20 Interview with Rom Harré in June 2002. 24

Gilbert Ryle – A Biographical Sketch normally taught in Oxford, such as Husserl, Brentano and Wittgenstein. For a long time he was the only Oxford lecturer, and with respect to Husserl even one of the few English lecturers, who offered such courses.

The most distinguished English philosopher to take it seriously was Gilbert Ryle, who thought highly of Husserl’s Logical Investigations, first published at the turn of the century, but became critical of his later work, so much so that he did not think it worth recommending to me when I became his pupil in 1929. Otherwise Husserl’s name was hardly known in England (…). (Ayer 1984, 26)

Even when Ryle was preoccupied with both the editorship of Mind and the development of graduate programs, he did not neglect the level of practical teaching, and remained one of the best and most extraordinary Oxford teachers in philosophy.

For him [Ryle] philosophy did not become an “administrative activity”. He led the whole operation of “Mind”, graduate program, etc. from his “command post” at Magdalen College; but those same rooms were also the scene for some of the most stimulating tutorials that one could hope for. (Moravcsik 1977, 3)

Ryle as a Person

Enough about Ryle’s contributions to philosophy, but what about Ryle as a person? People who met him for the first time would not characterize him as a typical philosopher. He was strong, slim, athletic and his appearance was almost military. He had penetrating eyes, a high forehead, and a typically British look. He always smoked his pipe and had the characteristic cough between every two sentences. His loud voice and commanding presence made people always take notice of him when he entered a room. He made up his mind very quickly about things and issues. Ryle’s directness, friendliness, unconventionality and strong sense of justice were his most distinguished qualities. His negative traits of character were that he could be rather brusque and uncompromising. Moreover, Ryle was sometimes believed to have no ‘inner life’, presumably a joke about his supposed denial of an inner life in The Concept of Mind (1949).21 Ryle remained unmarried, but in a way he had his own family with his twin-sister Mary and her adopted daughter Janet. He much enjoyed walking and often went on walking holidays, e.g. with Wittgenstein and with Mabbott. In earlier years he had made his tours in the Alps and elsewhere with his favourite brother, Peter. They were very close when Gilbert was a boy and Peter wrote to him from the trenches in the First World War. This continued when Peter returned from the trenches and Gilbert was entering into adult life. Peter was an electrical engineer, and although their professional lives were very different, theirs was a life-long and deep friendship, with many shared values and a shared sense of humour. Gilbert and his twin-sister Mary were also particularly kind to Peter’s widow, Rebecca, when she came to live near them after Peter died in 1961. Gardening and rowing were two of Ryle’s favourite pastimes. He particularly liked the digging and ordering of his garden but was not particularly interested in the plants themselves. In Oxford there have always been the ‘wetbobs’ and the ‘drybobs’ and Ryle

21 Bryan Magee once remarked: ‘Gilbert Ryle was a person of life-enhancing intellectual brilliance, but he had no inner life worth speaking of.’ (Magee 1997, 87) According to Magee this was a standing joke among Ryle’s friends.

25

The Philosophical Development of Gilbert Ryle was without doubt a ‘wetbob’ and much preferred rowing over cricket. He did enjoy playing golf – again with his brother Peter – not in a special sports outfit, but in his normal shoes and tweed jacket, and often with his pipe in his mouth. And for a while he also tried – and enjoyed – gliding. Ryle was a sociable person. He liked to go to conferences and enjoyed hugely talking about philosophy over a beer.

The Senior Common Room atmosphere – any Common Room – fitted him [Ryle] like a glove. (Mehta 1963, 67)

Being quite a consumer, when he was in the army Ryle was famous for going into the sergeants mess and challenging anyone to drink him under the table. The story is that he never lost. He even gave his name to a measure of beer, roughly three-quarters of a pint. Christ Church had special glasses made for these ‘Ryle’s of beer’. He used to say that sometimes half a pint was too small a measure and a pint too large. Once he stated that he would prefer to be remembered for having invented a measure of beer than for having written The Concept of Mind. He had a great love for Jane Austen and P. G. Wodehouse. But he also, like his twin-sister Mary, liked doing The Times Crossword every day (usually in less than half an hour). On holiday he was also an avid reader of detective stories; it is said that he could devour an ordinary paper-back in less than an hour – and identify the criminal. Sometimes Ryle pretended to dislike intellectual matters and reading. He certainly did not like people who tried to look like intellectuals and boasted about the many books they had read. Art and music were two subjects he was not at all interested in. There is an anecdote that he once saw Isaiah coming out of the Sheldonian Theatre after a performance of Bach, and shouted to him across the street, ‘Isaiah, have you been listening to some tunes again?’ (Mehta 1963, 67). He always wore a suit for his role of Professor in philosophy, and generally liked dressing up – he could boast about a new cap he had bought. Only at the weekends did he wear a sports-jacket. There is a story about Ryle inviting Wittgenstein to his club in London, in order to persuade him to give some lectures in Oxford. When he was about to catch the train to London, Ryle suddenly realised that, whereas at his club people were expected to wear suits and ties, Wittgenstein did not own a suit, probably not even a tie, and that he would feel really uncomfortable in his leather jacket. He did not want Wittgenstein to feel embarrassed and decided to put on his old sports-jacket and ordinary trousers. Not looking smart at all, he then waited in the club for Wittgenstein, who arrived wearing a perfectly nice suit which he had hired especially for the occasion. Naturally, Wittgenstein thought this was quite funny. Ryle was always a bit of a rebel and did not like the establishment much. He was offered a knighthood once, but turned it down. He did not want to have a title. For him philosopher would do. He was always kind to vulnerable people and was very good with children. He took care of people, such as his twin-sister Mary, never philosophically insulted those who were not intellectually well-established and sometimes even tried to keep others from doing so. For example, he once tried to convince Russell to rewrite his reply to a Mr. Emmet’s discussion of infinity and leave out some sharp remarks.

26

Gilbert Ryle – A Biographical Sketch

I am not very happy about your scolding of Mr. Emmet. He is a schoolmaster, and doubtless is as innocent as you say, but I don’t shut the door against amateur’s difficulties, even if they are amateurish, and I certainly don’t restrict MIND to things that I myself agree with (May 1st, 1957)

My qualms, as I hoped my first letter indicated, were due to the fact that it seemed to me that you spread yourself too much in the castigation of a young amateur. Very likely he merited being told shortly and sharply that he had not done his prep, or else was unequipped to benefit from it, but from a person like you one short sharp reproof of this sort would be enough. The length of your scolding of him will be bound to seem like an expression of pique. The lion may squash the mouse, but it’s embarrassing for his friends if he seems to be savaging it for the pleasure of it. It’s for you to choose; I’ll publish it if you want it published, but I think you would tarnish your name, not as a philosopher, but as a sage, if you appeared in print doing everything you could to try to make a young amateur feel small. So I ask you, non-editorially, to cut down the punitive remarks to one short sharp rebuke, and have the bulk of the reply devoted to explaining the point which experts doubtless would not need to have explained. (May 8th 1957)22

As I mentioned before, Ryle donated his books to a college that desperately needed them and left part of his estate to the poorest college. It is also said that he used his own money to arrange grants for at least two of his students. He called the grant ‘the Oxford and Cambridge philosophy fellowship’ and made sure that the students did not know that the money came from him. If I wanted to describe Ryle in only a limited amount of words I would say that he was a typical British philosopher, that he was kind to others, eager to help young and talented philosophers and had a strong sense of justice. On the other hand, that he could be brusque and uncompromising and was sometimes believed to have no ‘inner life’. He disliked footnotes, unnecessary technical language, bureaucracy, dogmas and uncritical followers/fans; he liked Jane Austen, P. G. Wodehouse, rowing, gardening, crosswords, detectives, walking and drinking beers in a social environment. He regarded philosophy as an activity rather than a theory and always tried to think beyond conventional solutions. In the following chapters many of these characteristics can be recognized.

22 Both letters are kept in the Russell Archive at McMaster University, Hamilton, Ontario. Russell’s reply was published in July 1958 (Russell 1958b) and although its tone is still very sharp, Ryle’s suggestions seem to have been taken seriously.

27

The Philosophical Development of Gilbert Ryle

28

Ryle’s Early Writings

Chapter 2

Ryle’s Early Writings

This chapter is devoted to Ryle’s early work, ‘early’ in the sense that it predates his most famous work, The Concept of Mind (1949a). The label ‘early’ however should not be taken to mean that these papers solely prepare the way for and culminate in the latter work. While prefiguring themes, arguments and methods which became prominent in The Concept of Mind, these papers should primarily be seen in the philosophical context of the time. Written in the 1930s and 40s they take up themes which were being debated at that time, showing traces of the influence of Russell, Husserl, Moore, Wittgenstein and others, as will be clear from what follows. Like other philosophers at that time Ryle was sparing in giving references to other works and other philosophers, so that I shall have to pay some attention to their historical-philosophical context in order to bring out the force of Ryle’s views. Without aiming at a full exhaustive discussion of all these papers, I shall limit myself to some of the themes which recur time and again. My aim in particular is to describe the methods he used and to trace the most important influences on his own philosophy. I shall discuss his ideas on meaning, , the existence of propositions, philosophical methodology, categories and category-mistakes. Ryle’s excursions into the history of philosophy – as was already mentioned in the introduction – constitute a genre on its own and will be ignored here, as well as ‘Knowing How and Knowing That’ (1946a), which shall be discussed in Chapter 3, since the main ideas of The Concept of Mind (1949a) have their roots in this paper. When Ryle began to write his first philosophical papers, Oxford philosophy was still under the spell of the realism of John Cook Wilson (1849-1915), H. A. Prichard (1871- 1947) and H. W. B. Joseph (1867-1943). It will become clear that for Ryle, after these earliest Realist influences – and other early influences for which Ryle’s tutor H. J. Paton was at least partly responsible, e.g. Lotze, Windelband, Rickert, Croce and Gentile – other philosophers such as the early Wittgenstein, Brentano, Meinong, Husserl and Russell soon started to get the upper hand. Next to Wittgenstein, whose influence shall be discussed in Chapter 6, Russell seems to have influenced Ryle most directly. In order to understand Ryle’s early and later views as well as their connections to Russell and Wittgenstein, we may distinguish two different positions, and two corresponding types of , which he seems to have held during his philosophical career. Ryle’s method of philosophical analysis, roughly, started as a means to fix the imperfections of language, followed by analysis as a more conceptual and more positive enterprise. In his early papers he was still under the spell of an ideal language and a denotational theory of meaning, that is a theory of meaning according to which all statements have meaning in the same way, in the way ‘Fido’ means Fido (the dog). According to this theory words and sentences are treated as names. But in the years that followed Ryle’s focus changed from a homogeneous theory of meaning towards a heterogeneous theory of use, which is most clearly at work in The Concept of Mind (1949a) and later writings, but is in its essence already present much earlier. Ryle’s aim was no longer to reformulate misleading expressions and achieve an ideal – or more ideal – language, but rather to expose flawed theories and doctrines by showing that category-

29

The Philosophical Development of Gilbert Ryle mistakes have been made. The philosopher becomes a cartographer whose more positive task is to map the logical geography of our concepts. An understanding of Ryle’s development from a homogeneous, denotational theory of meaning to a heterogeneous theory of use is essential if we want to grasp Ryle’s position and the type of philosophical analysis he aimed at. In an article from 1994 Shelley M. Park showed, for example, that the misinterpretation of The Concept of Mind as a behaviouristic tract, can be exposed by pointing at Ryle’s anti-denotationalism. I think she is right about the importance of Ryle’s rejection of this type of theory of meaning – it should also be noted that anti-denotationalism is a recurrent theme in Ryle’s later discussions. He considered it to be essential for progress in philosophy (e.g. in his review of Carnap’s from 1949 and his discussion of the historical development of the notion of meaning in ‘The Theory of Meaning’ from 1957). Ryle’s early adherence to a denotational theory of meaning and his later position of anti-denotationalism have received scant attention – probably partly because Ryle’s main work, The Concept of Mind, has often been interpreted as a new theory of mind, and not as the meta-philosophical project it in fact was. And also because Ryle himself hardly mentions denotationalism with respect to his own position, but only in discussing other philosophers. Russell, who greatly influenced the early Ryle, also defended a denotational theory of meaning. His aim was to reformulate those expressions that were exceptions to the denotationalist rule in such a way that they would ‘fit’ the rule. As John G. Slater argues, Russellian analysis ‘continues until the variables employed in the symbolic transformations of the original proposition denote actual entities’ (Slater 1994). We shall see that the early Ryle of ‘Systematically Misleading Expressions’ (1932b) held a similar view, reformulating the misleading sentences that seem to refer to an object x in reality into non- misleading ones that, in fact, refer to quite another object in reality, say y, and not into expressions that do not refer to any entity in reality. For example, expressions that seem to refer to a Mr Pickwick in reality are not formulated into ones that just do not refer to anything in this way, but into ones that refer to Dickens or The Pickwick Papers. Ryle soon abandoned his early denotationalist position (which from now on will be referred to as position (1)) and exchanged it for a non-homogeneous and non- denotationalist one (which from now on will be referred to as position (2)). This development was completed at the time when ‘Categories’ appeared in 1938. In it Ryle no longer aims at reformulating systematically misleading expressions but characterizes philosophy in terms of categories and philosophical method in terms of the discovery and analysis of category-mistakes. A uniform denotationalist relation between language and reality is no longer implied. The heterogeneity of the different ways in which our expressions have meaning requires a different approach. Philosophy is no longer considered to be a mere negative enterprise and is no longer merely assigned the ungrateful task of detecting ‘the sources in linguistic idioms of recurrent misconstructions and absurd theories’ (Ryle 1932b, 61). It is no longer oriented on pressing linguistic expressions into the straightjacket of denotationalism. It is rather a conceptual task to determine whether two expressions have meaning in the same way. One of the why it is important to see what exactly Ryle tried to achieve in his early writings, besides of its historical value, is that – as we will see in the following chapter – an accurate interpretation can help us to avoid certain misunderstandings and misinterpretations of what is commonly regarded as his main work, The Concept of Mind (1949a). And his later, relatively neglected, papers on thinking could also benefit from an understanding of the early origins of some of its problems – by which of course I by no

30

Ryle’s Early Writings means intend to imply that Ryle was a ‘static’ philosopher and never changed his mind or altered his approach.

Position 1: Denotationalism and the Reformulation of Systematically Misleading Expressions

Ryle’s point of view at the time of his earliest papers (roughly, until ‘Categories’ from 1938) can be defined as position (1) but with already of a development towards position (2). It is relevant here to consider the most important influences on his philosophical thought. In the first part of this section I shall discuss these influences in general, followed by a more detailed discussion of Russell’s influence on Ryle’s papers ‘Negation’ (1929) and ‘Are there Propositions?’ (1930). Finally, I shall have a look at ‘Systematically Misleading Expressions’ (1932b), which, of his earliest papers, most clearly shows his ideas on the nature of philosophy and philosophical analysis.

Influences and Philosophical Context Given the fact that the Realism-Idealism debate dominated Oxford philosophy at the time Ryle started his philosophical career, it is to be expected that philosophers such as Cook Wilson (considering the fact that Ryle is usually thought to be more of a Realist than an Idealist) at least partly influenced him. And in later years, Ryle indeed acknowledges the influence of Cook Wilson’s realism on his early philosophical views:

What was wanted was (a) Realism without additional entities to apprehend or (b) Realism without fabricated apprehendings. (…) So far my was that of a would-be antibiotic epistemologist. ‘Fidgetty Cook-Wilsonian’ would, so far, have been a fair title for me. (Ryle 1968b, 114)

Furthermore, some of Cook Wilson’s claims – such as that knowing makes no difference to what is known – bear close to Ryle’s ideas, as well as his typically Oxonian brand of logic, that is an informal logic as a philosophical investigation into thinking, rather than an attempt to construct a logical calculus.23 Moreover, Cook Wilson had great respect for ordinary usage and sharply distinguished between grammar and logic. He tried, for example, to show that ordinary usage tells against the Idealist theory of judgement, and distinguished between the logical and the grammatical subject and predicate. However, soon – as early as ‘Systematically Misleading Expressions’ (1932b) – Cook Wilson’s influence disappeared in the background, making way for Russell, the early Wittgenstein, Brentano, Meinong and Husserl. Perhaps Cook Wilson and the other Oxford Realists would have exercised a stronger influence on Ryle, if Ryle’s tutor, H. J. Paton, had been a Realist too, rather than an Idealist who introduced Croce and Gentile to Ryle; or if Cook Wilson had not – rather successfully, with the exception of Ryle – tried to keep Oxford far from any Russellian influence.24

23 This tradition can partly be explained by the classical education of philosophers in Oxford, focusing on Aristotle and having a strong linguistic interest; partly also by the fact that they lacked a scientific or mathematical background – contrary to their Cambridge colleagues. Cf. Passmore 1957 (p. 240); Hacker 1996 (p. 88). 24 In a letter to Bosanquet, Cook Wilson once wrote about Russell’s paradox: ‘I am afraid I am obliged to say that a man is conceited as well as silly to think such puerilities are worthy to be put in print: and it’s simply exasperating to think that he finds a publisher (where was the publisher’s reader?), and that in this way such

31

The Philosophical Development of Gilbert Ryle

An early letter from 1926 from Ryle to Paton – Ryle’s former tutor – shows that Ryle studied Croce and Gentile, and also Lotze, Windelband and Rickert.25 As he mentions in this letter, Ryle considered the Neo-Kantians’ idea of philosophy as the science of validity and value very attractive. His rejection of the existence of two levels of reality could well have been inspired by Windelband. His interest in Windelband and Rickert may also have led him to study Husserl. The ‘Ryle Collection’ at Linacre College reveals his thorough study of Husserl.26 In Ryle’s copy of Husserl’s Logische Untersuchungen he noted in the margin that Windelband and Rickert had anticipated an important remark of Husserl (but they would have made the unfortunate mistake of looking for a psychological foundation of logic). Ryle’s interest in Husserl – and also in Frege – may perhaps also be said to be, at least partly, inspired by an offprint of an article written in 1926 by Paul F. Linke. It is entitled ‘The Present State of Logic and in German’ and draws attention to the fundamental importance of the work of Frege in showing that logic could not be a psychological feature of man, but had to be the laws of itself. According to Linke, Frege’s work was relatively unknown and he considered Husserl to be the one who first developed this thought of Frege in the first volume of Logische Untersuchungen. This passage was side-lined in Ryle’s copy of the offprint. At the end of the 1920s and the beginning of the 1930s the questions in which Ryle becomes interested are primarily the logical ones raised by Frege, Meinong, Russell and the early Wittgenstein. Phenomenology remains influential, but mainly on the points it coincides with Cambridge ideas; therefore my focus will be on the influence of Russell and Wittgenstein.27 Ryle also follows in the footsteps of Russell and Wittgenstein in displaying an interest in the nature of philosophy, characterizing it as a logical versus a psychological enterprise28, and, like Wittgenstein but unlike Russell, points at the differences between philosophy and the other sciences. Something Ryle does not share with Frege, Russell and Wittgenstein was their interest in the problem of the foundations of mathematics.29 In this respect he is more a ‘pure philosopher’, like Moore. From his early writings it is clear that Ryle was not so much interested in epistemology, but rather in meaning and logic; especially the cases where logic seems to break down caught his attention. As he himself later puts it, ‘I found the pack-ice of logical theory cracking’ (Ryle 1970, 7). He thought that Wittgenstein’s Tractatus had started from the same cracks of logical theory. It was not formal logic that interested Ryle, but informal logic, the logic of our use of words, concepts, sentences and expressions. He did not – in the way of formal logicians – abstract from all differences in subject-matter, but, on the contrary, tried to lay bare the logic of different types of subject-matter. In terms of purely formal logic, there is no difference between ‘The king of France is bald’ and ‘The king of England is bald’, but, according to Ryle, there is an essential difference in informal logic. Ryle’s early papers are firmly rooted in Russell’s Principles of Mathematics (1903) and ‘’ (1905). He was vexed by Russell’s paradoxes – such as that the set of contemptible stuff can even find its way into examinations.’ (Cook Wilson 1926, 739) See also Marion 2000a (p. 304). 25 Cf. McGuinness and Vrijen 2006 26 See also Appendix 2. 27 For more on the phenomenological influences on Ryle see Thomasson 2002. 28 The Oxford Realists had argued for a view not unlike what was later called anti-psychologism – Prichard in ‘A Criticism of the Psychologists’ Treatment of Knowledge’ (1907, 27-53) and Joseph in his series of papers on ‘The Psychological Explanation of the Development of the Perception of External Objects’ (1910/1911). See Marion 2000b (p. 501). 29 He was interested in parts of Russell’s Principles of Mathematics, but not in his more formally mathematical Principia Mathematica. 32

Ryle’s Early Writings sets which are not members of themselves is neither a member of itself, nor not a member of itself (both alternatives being absurd). And he was also interested in problems of reference in cases such as ‘The present king of France is bald’. The phrase cannot refer to an entity called ‘the present king of France’, because no such entity exists. These are the kinds of problems Russell attempted to solve, trying to avoid the need to create a Fregean or Meinongean third realm of entities that neither exist in the mind, nor in reality, but merely subsist. Ryle, too, refutes the idea of a third realm and tries to solve the apparent paradoxes or puzzles by showing ‘what it really means to say so and so’ (Ryle 1932b, 61). What do negative expressions or expressions about imaginary objects such as Mr Pickwick really mean? We shall see that in Ryle’s earliest papers a denotational theory of meaning is present. But how did this denotationalism start to dominate the philosophical theories of Frege and Russell in the first place, putting a spell on philosophy for many generations? It can be understood best by seeing it as a response to logical idealism and psychologistic theories that were unable to rescue the of logic and truth. Logical realism – the view that logical entities exist independently of the mind – and anti-psychologism – the rejection of the view that the laws of logic are ultimately psychological laws – were used to safeguard this objectivity. Meanings had to be objective too, since otherwise two people would be unable to mean the same thing and therefore unable to agree or disagree. From this perspective, denotationalism and the in the existence of objective propositions and negative facts seem to be necessary in order to avoid and psychologism. In order to safeguard denotationalism and explain the existence of objective propositions, negative facts, numbers, etc., Frege and Meinong created a Third Realm of entities which are real and subsist but do not exist as objects in reality. Russell proceeded differently, claiming that ordinary language misleadingly suggests that our expressions, which ostensibly designate entities such as Mr Pickwick, numbers or golden mountains, are in fact referring terms. According to Russell these seemingly referring expressions are merely ‘incomplete ’ that have meaning within appropriate contexts but are meaningless on their own. He argues that while logically proper names do have ‘real’ referents30, descriptive phrases – such as ‘The present king of France is bald’ – are merely collections of quantifiers and propositional functions. As such they do not denote. This enables Russell to explain the truth of a negative existential without being committed to the belief that the subject term has reference. He is able to rescue the law of excluded middle and the law of without being committed to create Meinongean modes of being.31 Ryle would later show that denotationalism and the creation of an extra realm of entities cause antinomies, and moreover, that in the end they are not necessary – the objectivity of logic and truth and sameness of meaning can be guaranteed without them. Upon realizing that meaning and language are much richer and cannot – and need not – be captured in a denotationalist relation between language and reality, the need of a third realm disappears. Denotationalism is also an important target of The Concept of Mind (1949a) and remains a central theme in several of his later writings. In addition, Ryle’s discussions of

30 At least this was Russell’s position at the time of ‘On Denoting’. Later he argues that proper names such as ‘Moses’ and ‘Socrates’ (since there is no object in reality for them to refer to) had to be regarded as uniquely determining descriptions, a position Wittgenstein would later criticise in his Philosophical Investigations. 31 Claiming that ‘the author of Waverley’ by itself does not mean anything – it only has a meaning in an appropriate context – Russell is able to escape the problematic that ‘The author of Waverley’ means the same thing as ‘Scott’, without being forced to give up the law of identity (the idea that when a and b are identical everything that is true with respect to a is true with respect to b, and the other way around). And because ‘The present King of France is bald’ as such does not denote, the law of the excluded middle is no longer endangered.

33

The Philosophical Development of Gilbert Ryle other philosophers often focus on their attitude towards denotationalism. Early traces of Ryle’s later rejection of denotationalism can be found in his discussion of Husserl’s phenomenology:

And as Husserl seems, anyhow latterly, to reject Platonic or Meinongian subsistence theories, it becomes very hard to see in what sense he holds that ‘intentional objects’ really are genuine objects or subjects of attributes at all. He should hold (I believe) that what we miscall ‘the object or content of an act of ’ is really the specific character or nature of that act, so that the of an act is not a relation between it and something else, but merely a of it so specific as to be a differentia or in some cases an individualizing description of it. He does in fact, however, continue to speak as if every intentional act is related, though related by an internal relation, to a genuine subject of attributes. (Ryle 1932b, 175)32

And in his discussion of Wittgenstein Ryle describes that at the time of the Tractatus, although ‘one foot was already free’, ‘Wittgenstein still had one foot in the denotationist camp’ (Ryle 1957, 363). He already realized that logical constants do not stand for objects and do not have the meanings they have qua designating objects, and that sentences are not names and should not be treated as if they are. Only after the Tractatus did Wittgenstein remove his other foot from the denotationist camp – when he started to define ‘meaning’ in terms of use. The use of an expression became ‘the role it is employed to perform, not any thing or person or for which it might be supposed to stand’ (Ryle 1957, 364). Aside from Russell, Wittgenstein and the other philosophers mentioned above, we should also make mention of the Vienna Circle, which Ryle himself in 1970 recognized as an important early source of inspiration in the 1930s.33 The Circle’s ideas confirmed and perhaps strengthened Ryle’s suspicions of metaphysics. ’s ideas on philosophy in 1930, e.g. seem to be close to Ryle’s views.

Philosophy is not a system of statements; it is not a science. (…) The great contemporary turning point is characterized by the fact that we see in philosophy not a system of , but a system of acts; philosophy is that activity through which the meaning of statements is revealed or determined. By means of philosophy statements are explained, by means of science they are verified. The latter is concerned with the truth of statements, the former with what they actually mean; the philosophical activity of giving meaning is therefore the Alpha and Omega of all scientific knowledge. This was indeed correctly surmised when it was said that philosophy supplied both the foundation and the apex of the edifice of science. It was a mistake, however, to suppose that the foundation was made up of “philosophical” statements (the statements of theory of knowledge), and crowned by a dome of philosophical statements (called metaphysics). (Schlick 1930/31, 56-57)

As we shall see, however, the influence of the Vienna Circle on Ryle should be qualified a bit. First of all, there were crucial differences of opinion between the different members of the Circle. Schlick, for example, was close to Wittgenstein and clearly distinguished between science and philosophy, whereas Carnap and Neurath aspired to a Russellian scientific philosophy. In this respect, Ryle sided with the former group. Secondly, Ryle did indeed display an interest in verification and falsification, but he never actually started to use the Circle’s methods, and – although sympathetic towards the general idea of – in a paper from 1936 Ryle attacks, on several grounds, the form in which

32 Ryle does not stand alone in this interpretation of Husserl. In his paper from 1993 António Zilhão claims: ‘Husserl still accepts uncritically the empiricist , namely, the idea that the meaning of an expression lies in the relation of designation it has with the object or phenomenon of experience to which it refers.’ (António Zilhão 1993, 957) 33 See also Chapter 1. 34

Ryle’s Early Writings the verifiability-principle was stated. Moreover, later in 1951 he is far more critical, claiming that, although it is important to determine what does or might verify or falsify certain hypotheses, this does not justify the verification principle, i.e. that the meaning of a proposition is identical with its method of verification. There is more to the meaning of a proposition than the way to test it.

We test a recipe by seeing whether cakes made in accordance with it are good or bad, but this is not the whole point of the recipe. The normal reason for following a recipe is that we want to have cakes to eat. Not all cooking is experimenting. We could say that understanding a recipe is knowing how to make cakes of a certain sort; we could not say that understanding a recipe is knowing how to tell from the cakes made according to it whether it is a good or bad recipe – though this is, of course, the way to find out whether it is a good or bad recipe. Similarly we could say that understanding a law-statement is knowing what concrete inferences to draw from certain concrete premises, but not that it is knowing how to tell from the fates of such inferences whether the statement is true or false – though, again, this is, of course, the way to test the statement. (Ryle 1951b, 291-292)

Ryle’s basic idea is that although verification is the right way to test propositions, the meaning of a proposition is not identical with its method of verification. Just as understanding a recipe is knowing how to make cakes and understanding a recipe is not identical with knowing how to test whether it is a good or a bad recipe, understanding a proposition is not identical with knowing how to verify it. Identification of meaning with verification is unsatisfactory since this would just be a variant of the paradox of ‘The Liar’ – what a recipe tells us would be ‘the way to find out whether what it tells us is acceptable or unacceptable’ (Ryle 1951b, 292). Thus, whereas Ryle still considers verification to be of great importance, he no longer thinks (that is, if he ever fully believed it) that the meaning of a proposition is its method of verification. At the most, the method of verification of a proposition provides a clue to its meaning. Therefore, we should be wary to associate Ryle too strongly with logical . As we will see in Chapter 3, such a close association has been partly responsible for the later misinterpretation of Ryle as a behaviourist.

Ryle’s Realism and Russell’s Influence Illustrated by Two Papers As an illustration of Ryle’s early realism and his creative use of Russellian ideas, let us now look at two early papers: ‘Negation’ (1929) and ‘Are there Propositions?’ (1930).

‘Negation’ ‘Negation’ was presented at a symposium on this subject, organized by the Aristotelian Society in 1929.34 Ryle’s contribution is partly a response to Mabbott’s idealist claim that all negative judgments are subjective and inferior to knowledge. Ryle’s realism, and his dislike of Mabbott’s idealist terminology, is expressed in his claim: ‘We are not excluding B-ness from A – the nature of A does that – we are only coming to know or believe that A’s nature does so, i.e. that A is not B.’ (Ryle 1929, 2) Whereas Ryle and Mabbott would both agree that a negative proposition can never be the complete answer to the question ‘What is A?’, their ideas diverge on what the negative proposition can do in this case. Mabbott claims negative propositions to be merely expressions of doubt or ignorance, whereas Ryle argues that ‘what a negative sentence states may be a real fact and one which is both knowable and worth knowing’ (Ryle 1929, 2). His realism forces him to show that ‘the

34 The other speakers were J. D. Mabbott and H. H. Price.

35

The Philosophical Development of Gilbert Ryle negative element in negative sentences does denote, or may denote, some objective character of known facts and not merely some subjective limitations in our intellectual powers and equipment’. (Ryle 1929, 6) Ryle claims that the sentence ‘Mrs Smith’s hat is not green’ expresses a genuine proposition, since it is an answer to the question ‘What colour is Mrs Smith’s hat not?’. Although it is not a statement of what is the colour of the hat, it would be incorrect to say that nothing is asserted of the hat. What is asserted is ‘Mrs Smith’s hat is not green but some other colour’. Ryle argues that this ‘but some other…’ is always present, either explicitly or implicitly, in negative propositions. He proposes to call negative facts ‘abstract’ – such as ‘Mrs Smith’s hat is not green’ – and knowledge of such facts can be ‘instrumental to us in getting to know some further and less “abstract” fact’ (Ryle 1929, 2). It is abstract in the sense that it is a fact about a fact about a thing and not directly a fact about a thing. ‘Mrs Smith’s hat is not green’ positively characterises the determinate colour of the hat as being one of the colours other than green – ‘being coloured’ positively characterizes the hat. The ‘but some other…’-clause is part of what is meant by a negative sentence. Ryle further criticises Mabbott for not having clear criteria for distinguishing between judgements which are genuine and those which are not. The criterion Ryle himself employs is whether a particular proposition is an answer to a real question – a real question being one ‘which can be known necessarily to have an answer, though the answer may not be known.’ (Ryle 1929, 3) ‘Virtue is not square’ is according to Ryle not nonsensical because of its being negative, but because it is not an answer to a real question, since virtue does not have a shape. Ryle thinks that Mabbott would agree with him so far, but that the latter would claim that it is precisely because all negative sentences are not answers to real questions that all negative sentences are nonsensical. Mabbott would probably argue:

‘Mrs Smith’s hat is not green’ is nonsense unless it is true that the hat is either red or blue or green or yellow… But that is just the point. It is not true. A particular hat can not have a disjunctive colouring or hover between alternatives. If it is, e.g., blue, then it isn’t any other colour, and so there is no ‘either-or’ about it at all.’ (Ryle 1929, 8)

Ryle’s reply is that, contrary to what Mabbott suggests, there is no question of a particular hat with a disjunctive colouring. ‘Not green’ ascribes a character to a character (namely its being coloured) of Mrs Smith’s hat and only indirectly ascribes a character to Mrs Smith’s hat itself. Negation and disjunction are of a higher level than concrete statements such as ‘this hat is yellow’. The basic idea of different levels – different types – was already formulated by Russell in order to avoid logical paradoxes, but Ryle applies it here in his own original manner. According to Ryle, and disjunctions are ‘abstract’; they are not directly about facts of objects such as hats or trains, but about facts about a particular character (e.g. being coloured) of objects such as hats (or trains). Genuine negative and disjunctive propositions presuppose and express knowledge, independent of whether we do in fact know the ‘fact of’ – e.g. what the actual colour of the hat is. In 1914 Russell already proclaimed the existence of negative facts, which almost started a riot at Harvard (he also claimed that atomic and general facts existed).35 His idea

35 In 1918, during a course of eight lectures delivered in London, Russell again plead for the existence of negative facts, taking into account Mr. Demos’ attempt to explain why there are no negative facts – Demos was someone among Russell’s audience in Harvard (Mind, april 1917). Russell claimed: ‘I really only ask that you should not dogmatise. I do not say positively that there are [negative facts], but there may be.’ (1918, 67) Russell’s main objection against Demos’ view is that it makes incompatibility fundamental. Demos adheres to the Bradleyan view that when we assert ‘not-p’ we are really asserting that there is some proposition q which is true and is an opposite 36

Ryle’s Early Writings of negative facts must be seen against the background of his logical atomism and his idea of analysis as a means to discover the simple and ultimate constituents of the world. He took negative facts to be among these ultimates. He either had to admit negative facts or abandon the correspondence theory. The Russellian idea of negative facts as ultimates seems to imply the problematic existence of non-linguistic negative things in reality, such as ‘not-triangles’, ‘not-green’ or ‘not-square’, in order to be able to make negative propositions such as ‘This is not a triangle’ true. Mabbott and Ryle would both argue against this acceptance of negative facts. But whereas Mabbott denies the existence of negative facts altogether, Ryle wishes to defend ‘the position that a sentence involving a negative may be the expression of something that I know – in other words, that there are real negative facts.’ (Ryle 1929, 1) Ryle’s talk of the existence of negative facts is indebted to Russell, and so does his idea that these negative facts are second order facts, which reminds us of Russell’s Theory of Types. However Ryle’s application of this Russellian idea to negative statements is innovative and does not occur in Russell or – to my knowledge – in other contemporary philosophers.36

‘Are there Propositions?’ In his paper on the existence of propositions (1930) Ryle pays attention to the question whether there are propositions, not because he that there are philosophers who still adhere to the view that propositions are genuine entities (the so-called ‘proposition theory’), but because the and errors of the psychological theory (e.g. that our knowledge is merely subjective) which was refuted by the proposition theory, seem to crop up again. If there are no such things as objective propositions for us to believe in, what are the objects of our beliefs? And what do we know? Must the conclusion be that we have to return to the undesirable position that knowledge is merely subjective? Ryle’s aim is to offer a theory which avoids the ambiguities and errors of the psychological theory as well as the defective proposition theory. In A Hundred Years of Philosophy (1957) claims that the problem whether there are propositions is ‘one of the most controverted points in recent philosophy’ (p. 559, f. 7). The arguments of Moore and Russell against the proposition theory are similar to Ryle’s. According to the proposition theory, our false beliefs – as do our true ones – require the existence of objective propositions for us to believe in. Otherwise, how can we retain the objectivity of our knowledge, and of logic and truth? Both Moore and Russell argue against this theory that the very essence of a false belief is that what we believe is not; that there is no such object as the one we believe to be there. Otherwise, so Russell claims in The Problems of Philosophy (1912), if there were such a genuine object as the proposition ‘Desdemona loves Cassio’, Othello’s false belief that ‘Desdemona loves of p – which Russell interprets as being incompatible with p. Incompatibilities rather than negative facts become objective facts. Russell does not consider this solution simpler or less problematic than allowing negative facts. One of Russell’s objections is that incompatibilities hold between propositions, which are not ‘real’; not between facts. Therefore it needs a lot of ‘dressing up’ (Russell 1918, 70) before it can be taken as ‘an ultimate fact of the real world’ (ibid.). Russell thinks that it would be better to take negative facts as ultimates, rather than incompatibilities. 36 Many years later Ryle discussed the idea that negative facts exist as things in reality in Meinong (Ryle 1973, 13). While Ryle then still rejects the existence of ‘un-things’ (ibid.) in reality, he does appreciate Meinong’s many discriminations, that is his discussions of different types of ‘negative objects’, and believes that there is still a lot of work to do for philosophers. As he says: ‘We cannot cry off the task of providing our own antiseptic analyses of such varieties on the score that Meinong had housed them all together in a single box, and that was a wrong one. His variegated object-differences have to be equally methodically re-cast as equally variegated predicate- differences and, behind these, as equally variegated syntax-differences.’ (Ryle 1973, 13)

37

The Philosophical Development of Gilbert Ryle

Cassio’ would be true instead of false. The very reason why his belief is false is that there is no such object as ‘Desdemona loves Cassio’. And if false beliefs are not beliefs in objective propositions, it seems likely that true beliefs are not beliefs in propositions either. As Moore argues:

(…) the theory as to the analysis of belief which I wish to recommend (…) may be expressed by saying that there simply are no such things as propositions. That belief does not consist, as the former theory held, in a relation between the believer, on the one hand, and another thing which may be called the proposition believed. (Moore 1910-11, 265)

However, this does not mean that every statement we make about propositions is nonsense. Moore thinks that we should ‘continue to talk as if there were such things as propositions’ (Ibid., 265).

All that our theory compels us to say is that one part of this expression, namely the words ‘The proposition that 2 plus 2 equals 4’, though it seems to be the name of something, is not really a name for anything at all, whereas the whole expression, ‘The proposition that 2 plus 2 equals 4 is true’ is a name for a fact and a most important fact; and (…) that (…) we must not suppose that this fact can be analysed into a fact called ‘the proposition that twice two are four’ and a relation between this fact on the one hand and truth on the other. (Ibid., 266)

Moore claims that these kinds of sentences are not in general, as a rule, misleading, but only if we wrongly analyse the fact which they express, or regard every seemingly naming expression necessarily as a name of something.37 While reaching the same conclusions as Russell and Moore, Ryle uses a different strategy. Rather than merely showing that the existence of objective propositions is implausible because of the of the existence of objective negative propositions, he focuses on the idea that we do not need the existence of objective propositions. He discusses different kinds of arguments in favour of the existence of objective propositions, e.g. the argument from the intentionality of acts of thinking (‘‘accusatives’ of acts of thinking have come to be called ‘propositions’’ (Ryle 1930, 14)) and arguments for the independence of propositions from thinking (since otherwise there would merely be subjective and no objective knowledge). All these arguments lead to the same conclusion, namely that there are two main kinds of being, namely existing and subsisting being (the latter constitutes the so-called Third realm, consisting of universals, , numbers, and objective and falsehoods or objective propositions). Ryle expresses his fear that antinomies necessarily arise from treating propositions as substances. He does not want to admit a Third realm, showing that his own alternative theory renders objective propositions neither plausible nor necessary. In this way he tries to solve the problems which the proposition theory had tackled, e.g. that knowledge is necessarily subjective and in this respect does not differ from belief. Ryle’s theory is that when we know something what we know is a fact, ‘fact’ not denoting any new entity. There is no intermediating something (e.g. an objective proposition) between my knowing and the fact. And the fact that I know this fact, and the meaning of my statement what I know are one and the same thing. According to Ryle, statements of what I merely think (all varieties of apprehension other than knowing, e.g. believing, supposing) belong to a different category; they are not statements of known facts. This, however, does not make them meaningless. What I believe is ‘still something

37 In this respect, from ‘Categories’ (1938) onwards, Ryle seems to resemble Moore. He no longer aims at reformulating misleading expressions but focuses on a more conceptual type of analysis. 38

Ryle’s Early Writings

“accusative” to my thinking, and in its identity thinkable, too, to any other persons or to me on any other occasions’ (Ryle 1930, 27), or by ‘one or many persons in many sorts of attitude’ (Ibid., 28). Ryle regards Meinong to be right that there is a neutral common ‘thinking that’ (Annahmen) present in all of our thinking attitudes, namely ‘thinking of …as’. To think ‘x is y’ is to think x as if it is/were y.

If I am right, the air is now cleared to this extent. We have only two states of mind to deal with. At the top we have knowing, the ‘accusative’ of which is a fact. At the bottom we have ‘thinking of …as’ or ‘entertaining’, the ‘accusative’ of which is still sub judice. (Ibid., 29)

Ryle asks himself of what a statement (say ‘x is y’) is presentative, or what it symbolizes, in the case of ‘thinking of …as’. The statement cannot present or symbolize a fact known to me, because this would make it knowledge instead of merely ‘thinking of …as’, and presumably it is not the statement of a fact at all. When I think (and not know) the statement ‘Smith is taller than Jones’, the statement is not presentative of anything but functions as if it does.

The image does not present the fact that Smith is taller than Jones – there may be no such fact – but it is as if it did. It [the image or, in other cases, the statement] is therefore a constituent of the hypothetical fact that the image depicts their relative heights if and only if Smith is taller than Jones. (Ibid., 36)

Knowledge differs from mere thinking in that the first means thinking in a and the second in a quasi-symbol.

The symbol proper – i.e. the statement of a known fact – symbolizes a fact in which it itself is not a constituent. The quasi-symbol exhibits but does not symbolize a fact which is in part about itself, namely, that it has such grammatical properties that it may state a fact and does so if there is such and such a fact. (Ibid., 36)

Symbols and quasi-symbols have the same grammatical structure but not the same symbolizing or presentative function.38 According to Ryle, propositions are not what we think, but what we think in. There are no propositions, but there are facts, symbols and quasi-symbols by means of which the undesirable consequences of psychologism can be avoided. Without assuming the existence of propositions Ryle thus finds a way to rescue the essential difference between knowing and thinking, while still being able to explain the possibility that several persons, or one person at several different occasions or in different attitudes may ‘think the same thing’. Because of the fact that at the time he wrote ‘Are there Propositions?’ Ryle still adhered to a denotational theory of meaning, and did not want to accept the existence of propositions in reality, he was forced into the position that false sentences do not have a meaning. He writes: ‘In the strict sense, only those statements mean something which state a fact to some one who knows the fact.’ (Ryle 1930, 34) And false sentences do not state facts. Ryle does distinguish between the meaninglessness of false sentences and meaningless ‘in the way in which a haphazard collocation of words is meaningless’ (Ibid.). A false statement is a quasi-symbol, not genuinely symbolizing a reality. In his book Outlines of a Nominalist Theory of Propositions (1972) Paul Gochet refutes Ryle’s theory

38 As will be clear from Ryle’s description of Wittgenstein’s views on how a sentence has a meaning (see Chapter 6, p. 155), the view he attributes to Wittgenstein closely resembles his own view. He may well have been influenced by Wittgenstein on this point.

39

The Philosophical Development of Gilbert Ryle because he considers it to be a mistake to let truth-value determine meaning. (Gochet 1972, 125-6) Later, after he had rejected his rigid denotational theory of meaning, Ryle would also be able to set aside his earlier idea that false sentences do not have a meaning. Thus, both these early two papers, ‘Are there Propositions?’ and ‘Negation’, show clear traces of a Russellian way of thinking, such as a Theory of Types, , and the idea that there is no such thing as an intermediary proposition between knowledge and the known fact. But Ryle operates with these instruments in his own way, claiming that negative facts are of a higher order – that is, they are of a higher level of – than positive ones. And as Russell avoids absurdities by introducing the idea of incomplete symbols, Ryle’s quasi-symbols perform a similar task.

‘Systematically Misleading Expressions’ Ryle’s ‘Systematically Misleading Expressions’ (1932b) is without doubt one of his most well-known contributions to philosophical method. The central idea of the paper, that there are expressions that are systematically misleading, and that these expressions should be reformulated, was certainly not new. Ryle was influenced by the Russellian idea that a philosopher’s job is to construct a logically ideal language which lacks the misleading structure of ordinary language.39 Such an ideal language would exhibit the true nature of the world, containing only those expressions which unambiguously and correctly reflect this ontological structure. There was supposed to be just one word for every simple object; ‘there will always be a certain fundamental identity of structure between a fact and the symbol for it’ (Russell 1918, 52). In the early 1930s – following in Russell’s footsteps – Ryle thought that many expressions, though they are clearly understood by their everyday users, are nevertheless couched in grammatical or syntactical forms which are in a demonstrable way improper to the states of affairs which they record. When an expression is of such a syntactical form that it is improper to the fact recorded, it is systematically misleading in that it naturally suggests to some people – though not to ordinary people – that the state of affairs recorded is of quite a different sort from that which it in fact is. Such systematically misleading expressions can be reformulated and – for philosophy – must be reformulated into expressions of which the syntactical form is proper to the facts recorded (or the alleged facts recorded). Systematically misleading expressions are not false or senseless. For them to be false (or true) they would have to be proper to the states of affairs recorded. And for them to be senseless they would not record any state of affairs, for example ‘virtue is square’. According to Ryle, philosophers cannot help treating grammatical forms as clues to the logical structures for which they are looking. And these clues are often misleading. If a statement cannot be really about x when it is either true or false it is systematically misleading. E.g. ‘God does not exist’ cannot be really about God because if it were true there would be no God for the statement to be about. Ryle roughly distinguishes between three different types of systematically misleading expressions:

1) quasi-ontological statements (e.g. ‘God does not exist’, ‘Mr Pickwick is a fiction’); whereas the grammatical form of ‘x exists’ and ‘x does not exist’ suggests that a given subject of attributes x does or does not have the attribute of

39 The general idea that the grammatical and the of an expression sometimes differ was already put forward by Husserl, Bradley, Cook Wilson and others. 40

Ryle’s Early Writings

existing, in fact these statements assert or deny the attribute of being x-ish or being an x of something not named in the statement. ‘Mr Pickwick is a fiction’ seems to be about a Mr Pickwick who does not have the attribute of existing, but the statement is really about Dickens (or about his Pickwick Papers). The systematically misleading statement should be reformulated into one which does not even appear to refer to a subject ‘Mr Pickwick’, but unambiguously refers to Dickens or the Pickwick Papers. 2) statements seemingly about universals and quasi-platonic statements (e.g. ‘unpunctuality is reprehensible’ and ‘equality is, or is not, a real entity’); whereas ‘unpunctuality’ seems to denote the subject of which an attribute is being asserted, in fact ‘is unpunctual’ signifies the having of an attribute. Universals are not objects in the way in which the Mt Everest is one. We cannot speak of ‘equality’ and ‘justice’ as if they were objects, since general nouns, adjectives, etc., are not proper names. But because of the grammatical form of these quasi-platonic statements we tend to treat them as if the universals – such as ‘unpunctuality’ – denote objects, although in reality they signify the having of attributes. 3) descriptive expressions and quasi-descriptions (e.g. ‘the eldest son of Jones was married to-day’ and ‘Poincaré is not the King of France’); genuine unique descriptions – such as ‘the eldest son of Jones was married to-day’ are not intrinsically misleading but philosophers often do mistakenly assume that (1) descriptive phrases – such as ‘the eldest son of Jones’ – are proper names and therefore denote in virtue of being called ‘the so and so’, and (2) that a description means what it describes – e.g. that ‘the eldest son of Jones’ means Tommy since Tommy is the one that is described. But descriptive phrases are not proper names. They denote in virtue of possessing and being the sole possessor of the attribute signified by the descriptive phrase. And they mean what is meant by the predicative expression – being a son of Jones and being older than Jones’ other sons. Whereas (1) and (2) are errors with respect to descriptive expressions and are not caused by misleading grammatical clues, more interesting mistakes are made with respect to quasi-descriptions, that is ‘the’-phrases which behave grammatically as if they were unique descriptions referring to individuals, when in fact they are not denotational phrases at all. Examples are ‘Poincaré is not the King of France’, ‘the top of that tree’, ‘Jones hates the thought of going to hospital’, ‘the defeat of the Labour Party has surprised me’ and ‘the whale is not a fish but a mammal’.

All three types of systematically misleading expressions are misleading in roughly the same way. They suggest the existence of new sorts of objects. Expressions are misconstrued as denoting when in fact they do not denote, but only look grammatically like denoting expressions. Perhaps at first sight Ryle’s account seems to amount to a rejection of denotationalism. However, a closer look at his examples shows that it does not. He reformulates ‘Mr Pickwick is a fiction’, which seems to refer to a Mr Pickwick in reality, into a statement that refers to a Dickens, or The Pickwick Papers. So the expression does refer, but not to what it seems to refer to at first sight. It therefore should be reformulated into an expression that does not disguise this ‘real’ reference. Ryle also reformulates expressions of type (2) and (3) in this way. They refer to some object in reality, but not to the one they seemed to refer to at first sight. According to Ryle, systematically misleading expressions play a central role in the forming of mistaken philosophical doctrines:

41

The Philosophical Development of Gilbert Ryle

I suspect that all the mistaken doctrines of concepts, ideas, terms, judgements, objective propositions, contents, objectives and the like derive from the same fallacy, namely, that there must be something referred to by such expressions as ‘the meaning of the word (phrase or sentence) x’ on all fours with the policeman who is really referred to by the descriptive phrase ‘our village policeman is fond of football’. And the way out of this confusion is to see that some ‘the’-phrases are only similar in grammar and not similar in function to referentially used descriptive phrases, e.g., in the case in point, ‘the meaning of “x”’ is like ‘the King of France’ in ‘Poincaré is not the King of France’, a predicative expression used non-referentially. (Ryle 1932b, 56)

The reformulation of systematically misleading expressions is what philosophy for Ryle is. However, this does not make philosophers into philologists. The task of philosophy is not to substitute nouns for nouns or verbs for other verbs, but to make clear ‘what it really means to say so and so’ (Ryle 1932b, 61). ‘Its restatements are transmutations of syntax.’ (Ryle 1932b, 61) Russell’s influence is clearly present, both in the idea of reformulating misleading expressions into ones that are not, and in the choice of examples. In ‘On Denoting’ (1905) Russell introduced ‘the King of France’, which became a stock example in the literature. Ryle, too, uses it in ‘Systematically Misleading Expressions’. According to Russell, ‘the present King of France’ seems to be a denoting phrase – by virtue of its form – but it does not denote anything (since France is not a monarchy). He was in fact the first to recognize that denoting phrases containing ‘the’ do not automatically refer to entities. Some of Ryle’s other examples are of the same kind as Russell’s, e.g. Ryle’s ‘God does not exist’ and Russell’s ‘The golden mountain does not exist’.40 But their discussions show, besides similarities, also important differences. First of all, there is a difference in focus; whereas Russell uses the language of logic and talks about avoiding a breach of the law of contradiction and the law of identity, and about ‘incomplete symbols’, Ryle neither mentions laws of logic, nor other technical terms, using everyday language to make his point.41

In this paper I have deliberately refrained from describing expressions as ‘incomplete symbols’ or quasi-things as ‘logical constructions’. Partly I have abstained because I am fairly ignorant of the doctrines in which these are technical terms, though in so far as I do understand them, I think that I could re-state them in words which I like better without modifying the doctrines. But partly, also, I think that the terms themselves are rather ill-chosen and are apt to cause unnecessary perplexities. But I do think that I have been talking about what is talked about by those who use these terms, when they use them. (Ryle 1932b, 62)

Secondly, Ryle dissociates himself from the early Wittgensteinian and Russellian thought that there is a real and unconventional one-to-one relation between the composition of the expression and that of fact. According to Ryle, a fact is not a collection: ‘I do not see how a fact or state of affairs can be deemed like or even unlike in structure a sentence, gesture or diagram.’ (Ryle 1932b, 59) But the alternative is not easy for him to accept either. It cannot be just by convention that a given grammatical form fits facts of a given logical form, because customary usage obviously tolerates systematically misleading expressions. Ryle is not sure which alternative to choose, but he claims that the relation between the

40 Ryle does not use examples similar to ‘The author of Waverley is Scott’, whereby Russell attempts to show that this is more informative than ‘Scott is Scott’ and that ‘the author of Waverley’ by itself does not mean anything. It only means something in an appropriate context. 41 As we have seen, in ‘Are there Propositions?’ (1930) Ryle used technical language, e.g. ‘quasi-symbols’. From ‘Systematically Misleading Expressions’ (1932b) onwards, this type of language was almost entirely absent from his writings. 42

Ryle’s Early Writings grammatical form of expressions and the logical form of the facts they record is ‘more nearly conventional than natural’ (Ryle 1932b, 60). Another way in which Ryle distances himself from Russell is his refusal to adopt the latter’s idea of positively determining an ideal language – that is a language that exhibits the true nature of the world. Ryle’s method is a negative and more modest one which starts with absurdities and antinomies indicating that certain expressions are systematically misleading, and results into a reformulation of these expressions into those that are proper to the facts recorded, without inferring ontological claims. Ryle does aim at an ideal language, though not ideal in the sense of a formal logical language. He uses a negative method and deals with the misleading expressions one by one.

I conclude, then, that there is, after all, a sense in which we can properly enquire and even say ‘what it really means to say so and so’. For we can ask what is the real form of the fact recorded when this is concealed or disguised and not duly exhibited by the expression in question. And we can often succeed in stating this fact in a new form of words which does exhibit what the other failed to exhibit. And I am for the present inclined to believe that this is what philosophical analysis is and that this is the sole and whole function of philosophy. (…) But, as confession is good for the , I must admit that I do not very much relish the conclusions towards which these conclusions point. I would rather allot to philosophy a sublimer task than the detection of the sources in linguistic idioms of recurrent misconstructions and absurd theories. But that it is at least this I cannot feel any serious doubt. (Ryle 1932b, 61)

Ryle does not attempt to provide an exhaustive list of all possible types of systematically misleading expressions – ‘I fancy that the number is in principle unlimited, but that the number of prevalent and obsessing types is fairly small’ (Ryle 1932b, 61). We can only look at specific expressions or types of expressions and free them of their systematically misleading character. After ‘Systematically Misleading Expressions’ (1932b) the differences between the two philosophers enlarged. As I mentioned before, Russell thought that meaning was usually determined referentially – in the sense that ‘Fido’ (the name) means Fido (the dog) – but that there were exceptions which are misleading in the sense that grammatically they were similar to the ‘regular’ cases. This is why he called expressions such as ‘the king of France’ incomplete symbols which are meaningless by themselves and only get a meaning in an appropriate context. Meaning is symbolizing, it is referring to atomic facts in reality, and Russell tries to account for those cases which are exceptions to the rule. One may say that Russell’s question (and also Ryle’s in 1932) was: given a general account of how expressions mean (e.g. a denotational account of meaning), is there a fundamental identity of structure between a certain expression and the fact described? (Russell), or, is the syntactical form of a certain expression proper to the facts recorded? (Ryle) If not, how should the expression be reformulated? As Ryle himself claimed many years later at the Royaumont Colloquium of 1961, his later views on philosophical method were different from the ones he had developed in ‘Systematically Misleading Expressions’.

It is certain that when I wrote “Systematically Misleading Expressions” I was still under the direct influence of the notion of an “ideal language” – a doctrine according to which there were a certain number of logical forms which one could somehow dig up by scratching away at the earth which covered them. I no longer think, especially not today, that this is a good method. I do not regret having travelled that road, but I am happy to have left it behind me. (Ryle 1961b, 305)

43

The Philosophical Development of Gilbert Ryle

Soon after 1932 Ryle was moving towards his later position of rejecting the denotational theory of meaning altogether – instead of merely showing that there is a minority of cases in which the denotationalist theory leads to absurdities and thereby in fact confirming that denotationalism constitutes the rule. It is difficult to determine exactly when this change of position occurred, but it must have been rather early; perhaps it started as early as ‘Imaginary Objects’ (1933a). In this paper Ryle uses the example of Mr Pickwick again, but no longer reformulates expressions about Mr Pickwick into ones that refer to Dickens or The Pickwick Papers. The paper ‘About’ (1933b) also shows traces of position (2). In this paper Ryle distinguishes between three different sorts of ‘about’ of which two are relevant here: ‘about-referential’ (about (r)) and ‘about-linguistic’ (about (l)), the first being about an object in reality in the sense of referring to it, while the second is a mere linguistic ‘about’. ‘Charlotte is writing her thesis’ is both ‘about (r)’ me and ‘about (l)’ me, whereas a sentence cannot be ‘about (r)’ Mr Pickwick, but can only be ‘about (l)’ him (since there is no object in reality to which ‘Mr Pickwick’ refers). In this paper, as well as in all his other early papers – especially ‘Negation’ (1929), ‘Are there propositions?’ (1930) and ‘Imaginary Objects’ (1933a) – the relation between language and reality is fundamental. If all language were ‘about (r)’ objects, the existence of negative facts and even of negative nonlinguistic entities would be required, and imaginary objects would have to be objects in reality, that is if we want to talk about them in any meaningful way. Likewise, for our statements to be ‘about (r)’ propositions, propositions would have to be objects in reality. The consequence of treating all our expressions as if they are ‘about (r)’ is that we have to create more objects than seems to be desirable. And doing so leads to paradoxes or antinomies. Although his approach changed, this general problem remained a central theme throughout Ryle’s career. Ryle’s later question (around 1938) became: are expressions that seem to have the same meaning (for example because they have the same grammatical structure) of the same category? That is, do they mean what they mean in the same way? At the time he wrote ‘Categories’ Ryle held the view that there is no general account of how expressions mean and therefore no (uniform) connection with the level of reality. He now went so far as to reject the whole idea of a homogeneous theory of meaning.

Position 2: Rejecting Denotationalism

Russell’s influence gradually seems to make way for a more (late) Wittgensteinian influence. This becomes apparent in ‘Taking Sides in Philosophy’ (1937) and ‘Categories’ (1938) when Ryle abandons Russell’s idea of the need and the possibility to seek for an ideal language which is logically perfect – and therefore not misleading – in that it exhibits the true nature of the world. Ryle changed his mind about an ideal language and seems to think at the time of ‘Categories’ that our everyday language is fine the way it is, as long as we realise that grammar does not provide a reliable tool for establishing meaning and sameness or difference of category; the word ‘misleading’ disappears from his vocabulary. He no longer focuses on reformulating expressions like ‘The present King of France is bald’, but insists that philosophers should realise that ‘the present King of France’ and ‘the present King (or, nowadays better, Queen) of England’ belong to different categories. Reformulation is no longer required or even desirable. This position, to which he was to

44

Ryle’s Early Writings adhere throughout the rest of his career, constitutes an important difference between him, and Russell. The later Wittgenstein also moved away from an ideal formal language, but kept rejecting the application of ‘true’ and ‘false’ to philosophical propositions, restricting their use to scientific propositions. At a Royaumont Colloquium in 1961, when asked what was the fundamental difference between himself and the later Wittgenstein, part of Ryle’s reply was:

Wittgenstein did not like using words such as “true” or “false” because he wanted to avoid blurring the line of demarcation between philosophy and science, as had been done in the past. For reasons which seemed to him sufficient, he thought the word “true” belonged, or at least should belong, to the scientist. For myself, I do not see any good reason why the use of the word should be restricted in this way. I think that to say that a philosophical proposition is true, and to say that a scientific proposition is true, does not entail that the two propositions are of the same order. And I see no reason for not using the word “true” – or, more often, the word “false” – in both cases. (Ryle 1961b, 305)

Ryle clearly let go of his earlier wish for an ideal language – focusing on recognizing category differences rather than on reformulating. Ryle’s change of method after ‘Systematically Misleading Expressions’ was based on his rejection of denotationalism, or hypostatization, that is ‘treating as names or other sorts of mentions expressions which are not names or other sorts of mentions’ (Ryle 1949b, 234). Denotationalism was a popular position which Ryle considered to be responsible for many misunderstandings and paradoxes. He appreciated Frege, e.g. for pointing at the difference in meaning of ‘morning star’ and ‘evening star’, as well as Russell for his theory of descriptions, which solved problems concerning expressions such as ‘The present King of France is bald’. However, he criticised them – Frege more than Russell – for not having realized that these examples had much greater consequences than they had ever imagined: in fact they proved the denotationalist view to be mistaken. Russell and Frege conceded that there were some exceptions to the referentialist rule, but they did not question denotationalism as such. And Carnap, while claiming to reject denotationalism, in fact made the same kind of mistake. He, too, was guilty of ‘hypostatization’.

So though in fact only a minority of sentence-fragments, namely mentioningly used substantival expressions, can be said to have extensions, Carnap has to assimilate the jobs even of sentences to this special job of a species of sentence-fragments. And this is precisely parallel to the Frege- Meinong mistake of treating sentences as names. These theorists assimilated saying to calling; Carnap assimilates saying to mentioning. Yet both mentions and names (which are a species of mention) are ordinarily used only as fragments of42 sentences. They enable us to say certain sorts of things, but when we have uttered them by themselves we have not yet said anything. (…) Now by ‘hypostatization’ we mean treating as names or other sorts of mentions expressions which are not names or other sorts of mentions. And just this is the tenor of the whole of Carnap’s meaning-analysis. (Ryle 1949b, 233-234)

Perhaps all theories of meaning in some way or another make this mistake of hypostatization, or more in general, of treating all expressions alike. This may also have been the reason why Ryle later moved from a theory of meaning to a theory of use, thereby stressing the heterogeneity of our sayings. In what follows we shall look at the shift from a Russellian type of analysis to that of categories and category-mistakes, which became the core of Ryle’s later views, e.g. in

42 The printed text reads ‘fragments or sentences’, but Ryle presumably meant ‘fragments of sentences’.

45

The Philosophical Development of Gilbert Ryle

The Concept of Mind (1949a).43 An important paper here is ‘Categories’ from 1938, and we shall therefore discuss it in some detail. To complete our picture of the period before The Concept of Mind, we shall look at the development of these ideas in the 1940s.

Category-Mistakes: the Essence of Philosophical Problems In his early paper ‘Categories’ (1938) Ryle defines the essence of what he thinks philosophy is by means of categories or types.

The matter is of some importance, for not only is it the case that category-propositions (namely assertions that terms belong to certain categories or types) are always philosopher’s propositions, but, I believe, the converse is also true. So we are in the dark about the nature of philosophical problems and methods if we are in the dark about types or categories. (Ryle 1938, 170)

We need to pursue category-mistakes because it is the only way we can get rid of the antinomies we are confronted with, both in our everyday life (e.g. I see a bent stick and the stick is straight) and with respect to technical concepts (e.g. the problem of the internality of relations44). But what kinds of categories does Ryle have in mind? What are his categories categories of? Some philosophers, e.g. , interpreted them as ontological categories and in doing so associated Ryle with a rather Russellian approach. Williams claimed at a Royaumont Colloquium in 1961 that Ryle’s point of view was one according to which ‘there are ontological categories which can be discovered by looking at linguistic expressions’ (Williams 1961, 304). This however may be questioned. Rather than ontological categories, they are semantical, as I will suggest in what follows. Ryle himself gives us important clues as to what his categories are categories of:

We try, then, to say that absurdities result from the improper coupling not of expressions but of what the expressions signify, though the coupling and mis-coupling of them is effected by operating upon their expression. But there is not and cannot be any univocal title for all the significata of expressions, since if there was such a title, all these significata would be of one and the same type. And just this is what was at bottom wrong with the Lockean terminology of ‘ideas’ and the Meinongian terminology of ‘objects’, words which were employed to perform exactly this impossible task. (…) So I use ‘proposition-factor’ (intending it to have all possible type-ambiguities), to collect whatever is signified by any expression, simple or complex, which can be a complement to a gap- in some propositional factor or other (or which can be a variable is some propositional function or other). And, if asked such questions as Do proposition-factors exist? How many of them are there? Are they mental? What are they like?, my answer is, ‘All such questions are ridiculous, since “factor” is and is meant to be the meeting- place of all type-ambiguities.’ (Ryle 1938, 180-181)

Ryle’s terminology of significata and signification may suggest that he accepts some sort of denotationalism after all, but this is not the case. His point is that the significata of expressions are multifarious and cannot be reduced to one type. In order to show this multifarious character he chooses to use the very general term ‘proposition-factor’, by which it becomes immediately clear that Ryle’s categories are linguistical or semantical

43 Whether, and in what way, Ryle later changed his attitude towards categories and their importance for philosophy, which at least seems to be implied by Dilemmas (Ryle 1954a), will be dealt with in Chapter 4. 44 As Ryle tries to show in his paper ‘Internal Relations’ (Ryle 1935c), antinomies resulting from the technical concept of relation caused the problem of the internality of relations. He thinks that arguments from the internality of relations to or the coherence theory of truth are caused by a wrong of the technical concept of relation. 46

Ryle’s Early Writings ones. Ryle wants to emphasise that what is signified by various expressions does not belong to one single category, but to several. He uses the word ‘proposition-factor’ for whatever is signified by any expression. This view expresses Ryle’s idea of the heterogeneity of our statements – the endless variety of different categories of ‘meaning’ – and his rejection of a theory of meaning according to which the meaning of all statements is determined alike, viz. denotationalism. Precisely because expressions mean what they mean in many different ways it is not possible to limit category differences to ontological ones. Category differences between two expressions are differences in how these expressions mean what they mean. Not all the significata of expressions are of the same category – and therefore cannot have a univocal title – and insufficient acknowledgement of this is exactly where philosophy goes wrong. How can we know whether two expressions belong to different categories of meaning?

To ask the question To what type or category does so-and-so belong? Is to ask In what sorts of true or false propositions can so-and-so enter? Or, to put it semantically, it is to ask In what sorts of non-absurd sentences and in what positions in them can the expression ‘so and so’ enter? And, conversely, What sorts of sentences would be rendered absurd by the substitution for one of their sentence-factors of the expression ‘so and so’? (Ryle 1938, 180)

We can determine to which category a proposition belongs by looking at its relations or ‘liaisons’, i.e. what the proposition ‘implies, what is implied by it, what it is compatible with and what it is incompatible with’ (Ryle 1938, 183). And this is precisely what philosophical analysis is – the activity of argumentation and not merely that of paraphrasing.

Now the operation of formulating the liaisons of a proposition just is the activity of ratiocination or argumentation (…). And this is why philosophising is arguing, and it is just this element of ratiocination which, as a rule, is left out of the latter-day definitions of philosophy as ‘analysis’. For these generally suggest that analysing is some sort of paraphrasing. But some sorts of paraphrase throw no philosophical light, for they fail to exhibit just those features of propositions and their factors, obscurity about which involves us in antinomies, namely their liaisons which flow from or constitute their logical types and forms. Mere increase of prolixity is not enough. (Ryle 1938, 183-184)

Ryle claims that in a way the formal structure of a proposition and its liaisons are the same thing. It is essential that Ryle does not mean ‘formal structure’ and ‘logical type’ in the sense of formal logic. Whereas formal logic abstracts from all differences in subject-matter, to Ryle these differences are essential.45 From a purely formal point of view there is no difference between ‘all virtues are square’ and ‘all tables are square’.46 Therefore, Ryle’s kind of cannot be captured in purely formal logic. Logical form for Ryle is logic by virtue of meaning (or, perhaps, already this early, logic by virtue of use) and not a logic that is merely based on content-neutral logical constants. In this respect Ryle clearly stands in the long tradition of the Oxford use of informal logic47, as opposed to the Cambridge tradition of formal logic. More specifically, the way to test whether two proposition-factors are of different categories is to check whether:

45 Cf. Strawson 1970 (p. 185). 46 It is important to realize that both ‘allvirtues are square’ and its denial are equally absurd. 47 Bosanquet, Cook Wilson, etc. Cf. p. 31 of this dissertation.

47

The Philosophical Development of Gilbert Ryle

(…) there are sentence-frames such that when the expressions for those factors are imported as alternative complements to the same gap-signs, the resultant sentences are significant in the one case and absurd in the other. (Ryle 1938, 181)

For example, ‘I’ and ‘the writer of this paper’ belong to different categories, since the sentence-frame ‘… never wrote a paper’ is significant if we import ‘I’, but absurd in the other case – viz. ‘the writer of this paper never wrote a paper’. (Ryle 1938, 181) Now that we have seen what Ryle has to say about ‘categories’ and ‘category- mistakes’, it is time to ask the question whether his descriptions suffice to show what exactly are ‘category’ and ‘category-mistake’. Has he been successful in convincing us that they are useful concepts? In an excellent paper on Ryle’s categories, P. F. Strawson claims that it is plausible that Ryle’s early works need some kind of general theoretical account of the notions of category, category-difference and category-mistake, but that, unfortunately, Ryle himself does not provide it. Ryle’s explanations, descriptions and examples of what a category is, and what a category-mistake is, lack the precision that is required for the concept to be of real use. If Ryle’s notion of category-mistake turns out to be a mere garbage-bin for all kinds of mistakes which are not or cannot be further identified, it is hard to see how the notion can be of much use. Ryle’s frequent use of reductio ad absurdum arguments does not seem to help, but only adds to the problem. As Strawson notes, there are different kinds of absurdities and only some of them are the result of category-mistakes. In order to see whether Strawson is justified in his claim that Ryle needs a general theory we should first address the following questions: (1) By virtue of what do different Rylean categories differ? And (2) what kind of absurdities are category-mistakes? As Strawson argues, the descriptions above do not suffice to show that ‘category’ and ‘category-mistake’ are useful concepts, at least not without further explanation, since there are cases in which Ryle’s descriptions allow us to infer that two proposition-factors belong to different categories, whereas Ryle must have considered them to belong to the same category. For example, following Ryle’s instructions we are justified in claiming that ’27’ and ‘37’ belong to different categories. If we import them in the sentence-frame ‘She was over … and under 33’, in the first case the resulting sentence is significant, in the second absurd. But surely, Ryle did not mean to claim that ‘27’ and ‘37’ belong to different categories.48 As a matter of fact, Ryle’s description seems to force us to say that all proposition-factors belong to different categories, since for two proposition-factors we can always find a sentence which is significant if the one is imported to the gap-sign and absurd when the other is imported. Let us infer from the examples Ryle uses what kinds of category-differences and absurdities he had in mind. He does not aim at purely formal absurdities such as ‘A and not A’ or ‘there is a number over 37 which is under 33’. The absurdities he has in mind are ‘virtue is square’, ‘the writer of this paper never wrote a paper’ (Ryle 1938, 181) and ‘time began a million years ago’ (Ryle 1945, 204), concluding that ‘virtue’ does not belong to the same category as, for example, ‘table’; ‘the writer of this paper’ not to the same category as ‘I’ and ‘time’ not to the same category as ‘human life’. Only certain sorts of differences in logical form imply category-differences. The question then is: what sort of differences? According to Strawson, a category-mistake is a specific kind of mistake, an absurdity that takes the form of :

48 This is the example Strawson uses in Strawson 1970 (p. 187). 48

Ryle’s Early Writings

the inappropriateness of some range of predicates to some class of subjects (…). “Xs are not the sort of thing that can be ф”; “xs neither ф nor fail to ф”; “‘фing’ (or ‘being ф’) cannot be significantly predicated of xs.” (Strawson 1970, 193)

‘Square’, for instance, cannot be predicated of ‘virtue’, and ‘never having written a paper’ cannot be predicated of ‘the writer of this paper’. In my opinion, Strawson’s example of ‘27’ and ‘37’ belonging to different categories is perhaps a bit unfair to Ryle, because the complex sentence ‘She was over … and under 33’ contains a predicate which belongs to the same category as both ‘27’ and ‘37’. The question should obviously not be whether ‘27’ and ‘37’ can be predicated of a subject of which another predicate is already predicated (namely being under 33), but simply whether they can be predicated of the subject. And I cannot think of any simple sentence that is significant if ‘27’ is imported into the gap-sign and absurd if ‘37’ is imported. However, Strawson’s example does work well for discovering Ryle’s implicit criteria of category differences. So far, defining characteristics of sameness or difference of categories have been mentioned. It is by virtue of meaning (and not by virtue of content-independent formal logic) that some range of predicates is (in)appropriate to some class of subjects. An important question that remains to be answered is whether Ryle’s view on categories and category-mistakes can be seen as a general theory in the sense that they explain exactly which kinds of categories and category-mistakes Ryle is after and distinguish them from philosophically irrelevant ones, such as ‘someone blind neither spots nor overlooks the fly in the ointment’ (Strawson 1970, 205). According to Strawson, without such a theory it could be argued that:

The particular interest excited by some cases of these correlations and not by others reflects nothing but a depth or strikingness or, sometimes, of philosophical importance, as attaching to some cases and not to others; and this is why theories of categories tend to move in circles. For this is the result that inevitably ensues when an attempt is made to present as resting on a clear and general distinction a habit of classification which in fact rests on nothing but differences in degree of impressiveness. (Strawson 1970, 206)

It may be questioned, however, whether Ryle holds a theory of categories in this sense. He does distinguish between philosophically important and philosophically less important category-mistakes, e.g. he thinks that abstractness causes philosophically interesting category-mistakes, but he does not see the need to exclude those that are not, or less, interesting to philosophy. They are only less interesting because they do not tempt us into making mistakes. Concepts or that are not concrete but are of a higher level of abstraction, such as ‘the common man’ or ‘the average tax-payer’, constitute the more interesting cases. Why are they more interesting? Simply because such abstract concepts are more likely to cause antinomies or absurdities in philosophy – these absurdities being less obvious. But in principle ‘any uncharted concept is liable to generate antinomies, for ignorance of its chart is ignorance of some of the implications and compatibilities of the propositions containing it.’ (Ryle 1938, 182) ‘Someone blind neither spots nor overlooks the fly in the ointment’ is as philosophical as ‘In saying “I am now lying” I am neither lying nor telling the truth’, but not as interesting. According to Ryle the latter sort is interesting since ‘their absurdities are not obvious but manifest themselves in the generation of contradictions or vicious circles, whereas ‘Saturday is in bed’ is obviously absurd before any contradictions are seen to result from the hypothesis that it is true’ (Ryle 1938, 179). Clues to interesting cases are therefore abstractness and that the absurdity is not recognized as such beforehand but only after contradictions and absurdities have presented themselves.

49

The Philosophical Development of Gilbert Ryle

If Strawson means to show that Ryle does not have an independent, external rod to measure whether a category-mistake is philosophically relevant, he is right. However, such a claim would suggest that Ryle needs one. Strawson hopes to find some clues in grammar. If we know why an expression that mixes up different categories makes grammatical sense, perhaps we will be able to understand more about the nature of categories and category- mistakes. Why is it that ‘Saturday is in bed’ makes grammatical sense even though in virtue of its meaning, Saturday is not the kind of thing to be in (or out) bed? But does this help? How does into why an expression is grammatical help us to understand why it is absurd? Does ‘Saturday is in bed’ even make grammatical sense? (And does ‘yellow is in bed?) Does any singular term make sense when imported into ‘… is in bed’? It seems to me that we either interpret grammatical rules in a very weak sense, which means that ‘yellow is in bed’ and ‘Saturday is in bed’ are grammatically correct, or in a stronger sense, thereby smuggling in semantics. In the first case grammar is not likely to give clues to semantics; in the second any clue would be, at least partly, semantical, which would endanger the non-circularity of the resulting general theory. According to Ryle, there is no way in which the different categories can be ‘discovered’ by looking at the world or other non-semantical sources. There is no external rod to be discovered, and philosophy does not rely on finding one, as science usually does, not even a linguistical one. Ryle does not give a general theoretical account because he simply does not think that such an account is possible in the first place. Categories and category-mistakes are not the kinds of concepts that can be defined in the sense of giving a theoretical description.

Philosophy as a Method and the Reductio ad Absurdum as the Ultimate Philosophical Argument From ‘Categories’ (1938) it is clear that Ryle considers philosophy to be a method rather than a theory, argumentation being the essence of philosophical analysis. Another paper from about the same time, ‘Taking Sides in Philosophy’ (1937b), presents this view in explicit terms. Philosophers do not discover new matters of fact – they discover new philosophical arguments. ‘The philosopher throws new light, but he does not give new information’ (Ryle 1937b, 165) At this point Ryle is much nearer to the later Wittgenstein than to Russell.

Philosophers do not make known matters of fact which were unknown before. The sense in which they throw light is that they make clear what was unclear before, or make obvious things which were previously in a muddle. And the dawning of this desiderated obviousness occurs in the finding of a logically rigorous philosophical argument. Something that was obscure becomes obvious to me in the act of seeing the force of a particular philosophical argument. (…) Anyone who appreciates the argument ipso facto gets the clarification. Though, of course, it is often very hard to appreciate involved and abstract arguments, like that which constitutes the . (Ryle 1937b, 166-167)

This is why a philosopher’s discovery cannot be summarized by the conclusion of the argument – as scientists’ discoveries can be summarized by stating what new facts they have discovered – since the argument is what is discovered. And if a philosopher in fact succeeds in finding a new philosophical argument the problem completely disappears.

The obscurity which he has overcome is, apart from collapses of cultures, dead from that time on. His arduously achieved discovery becomes a public truism, and, if it is of any importance,

50

Ryle’s Early Writings

becomes crystallized in the diction and the thought of educated people, even though the great majority of them have never read a word of him. (Ryle 1937b, 167)

This is why it is often difficult to determine the influence of philosophers. On retrospect the discoverer of a new truism, will seem to have been talking platitudes.

The historian who wants to find out what Aristotle or Locke ‘discovered’ must see what public truisms existed after the philosopher’s work was done which were not even the topic of a clearly recognized question before he began it. (…) And just this is his great achievement, so to emancipate men from an obscurity that they can regard as a platitude what their predecessors could not even contemplate clearly enough to regard as a paradox. (Ryle 1937b, 167)

Ryle himself, as well as Wittgenstein, has been accused of talking ‘platitudes’. In his inaugural lecture ‘Philosophical Arguments’ (1945), which he delivered right after the Second World War, Ryle gave further thought to the nature of philosophy and philosophical method, as opposed to the methods of the special sciences which are aimed at discovering and explaining facts. In this lecture he formulates the method he was to use a few years later in The Concept of Mind (1949a) to analyze the ‘logical geography’ of different mental concepts. As Hacker writes:

‘Philosophical Arguments’ (1945) was, in effect, a fresh declaration of principles, replacing ‘Systematically Misleading Expressions’ and developing further ideas in ‘Categories’. The task of philosophy, he declared, is ‘the charting of the logical powers of ideas’. In a metaphor reminiscent of the Blue Book (BB 57; cf. AWL 43; LFM 44), he observed that ‘People often know their way about a locality while being unable to describe the distances or directions between different parts of it or between it and other localities… Our workaday knowledge of the geography of our ideas is a similar case.’ (Hacker 1996, 149)

According to Hacker, Ryle may have been influenced by Wittgenstein, whose Blue Book circulated in Oxford in the 1930s, but apart from this ‘no clear ancestors can be traced of this new conception of philosophy’ (Hacker 1996, 100) Ryle proceeds with the method he started in ‘Categories’ (1938), showing more in detail how categories and category-mistakes are the essence of philosophical thinking. If philosophy is different from science and does not aim at discovering facts, its methods must be different as well. What are the proper methods of philosophy? In ‘Philosophical Arguments’ (1945) Ryle tries to exhibit the logical structure of a type of arguments which are proper to philosophical thinking, praising his predecessor R. G. Collingwood for realising that the apparent antithesis between natural sciences and human studies is an illusion. They simply give their own answers to their own specific questions about the world, not rival answers to the same questions.49 Philosophy, too, gives its own answers to its own specific questions. Ryle claims that philosophical arguments are neither inductions, nor deductions from axioms50; they are of a different category in that they ask different questions and use different methods. Ryle considers the reductio ad absurdum to be proper to philosophical thinking, by which he refers to what he calls the strong reductio, as opposed to the weak reductio which Euclid sometimes used – which proves only ‘either that the required theorem is true if the axioms are true or that both are false, that is, that the contradictory of the required theorem is not compatible with the axioms’ (Ryle 1945, 197). Ryle’s strong reductio does not

49 In Chapter 5 I shall discuss the relation between Ryle’s and Collingwood’s conceptions of philosophy. 50 This was already argued by Ryle in his early paper ‘Phenomenology’ (1932a, 169-170).

51

The Philosophical Development of Gilbert Ryle determine truth or falsity but shows that a proposition is illegitimate because it has logically absurd consequences. It is important to note that propositions cannot themselves be absurd – this would imply that there can be no such propositions. Only expressions can be absurd. What the reductio does is to disclose ‘that a given expression cannot be expressing a proposition of such and such a content with such and such a logical skeleton, since a proposition with certain of these properties would conflict with one with certain of the others.’ (Ryle 1945, 203) The strong reductio is the type of argument Ryle already employed in ‘Categories’ (1938) and ‘Systematically Misleading Expressions’ (1932b). He denies that this type of argument has only a destructive effect, and describes the reductio as having something in common with the destruction-tests by which engineers discover the strength of materials.51

(…) philosophical arguments of the type described have something in common with the destruction-tests by which engineers discover the strength of materials. Certainly engineers stretch, twist, compress, and batter bits of metal until they collapse, but it is just by such tests that they determine the strains which the metal will withstand. In somewhat the same way, philosophical arguments bring out the logical powers of the ideas under investigation, by fixing the precise forms of logical mishandling under which they refuse to work. (Ryle 1945, 197-98)

The ‘logical powers’ of a proposition are its logical relations to other propositions. Since intelligent people can never grasp all the logical relations of a certain proposition, mistakes occur. In this sense they have an imperfect understanding of any proposition they use, although they often learn from practice all the logical powers they need in order to use certain propositions in the limited sense in which they are ordinarily used, without making mistakes which would indicate an imperfect understanding of such propositions. There is the risk of ending up in paradoxes by performing operations with propositions or ideas which deviate from their familiar use, all the more so because the grammatical form of propositions does not provide clues to their logical powers. Ideas or propositions with different logical powers often have the same grammatical structure. ‘Men naturally, therefore, tend to be blind to the fact that different ideas have different logical powers or at least they tend to treat the varieties of logical types as being few in number.’ (Ryle 1945, 200) Thus, we do have ‘workaday knowledge’ (Ibid., 201) of the geography of our ideas, but this is ‘knowledge without system and without checks’ (Ibid., 201) – it is not knowledge by rules and therefore provides a mere naïve, preliminary account of the logical powers of these ideas. The real logical powers are only discovered by paradoxes which are evidence for type-confusions (or category-mistakes) and in this way force us to turn back in our tracks and to start using the idea in a different way – ‘determining with method and with definitive checks the rules governing the correct manipulation of concepts’ (Ibid., 201. This is what Ryle describes as ‘the charting of the logical powers of ideas’ (Ibid., 201). From this moment, the metaphor of maps becomes central to Ryle’s idea of philosophical analysis, and it also plays an important role in The Concept of Mind. The metaphor helps to make clear what he means by the mapping of the logical powers of our ideas. People often know their way in their own neighbourhood, but they are often unable to describe distances or directions between different parts of their neighbourhood, or between the neighbourhood and other locations. In a similar way we often have workaday knowledge of when and how to use certain propositions or ideas, but we are unable to

51 This example reminds us of Wittgenstein’s engineering examples, e.g. in Philosophical Investigations 67 and the Tractatus 4.461. 52

Ryle’s Early Writings describe logical relations between these propositions or between these and other propositions or ideas. Furthermore, just as a map does not show individual objects but various kinds of features of the area – such as roads, bridges and altitudes – as well as relations with other areas, so the logical geography of certain ideas or propositions is about the cross-bearings of many ideas that belong to the same or adjacent areas. Again, Wittgenstein’s voice seems to be present in this position.

The problem, that is, is not to anatomize the solitary concept, say, of liberty but to extract its logical powers as these bear on those of law, obedience, responsibility, loyalty, government and the rest. Like a geographical survey a philosophical survey is necessarily synoptic. Philosophical problems cannot be posed or solved piecemeal. (Ryle 1945, 202)

However, there is also an essential difference between the mapping of propositions or ideas and . Whereas the cartographer has both a positive and a negative method at his disposal, the philosopher only has a negative one. The correctness of a geographical map can be determined negatively in that a cartographical contradiction would prove the map to be incorrect, and positively in that visual observations are positive evidence of the map’s correctness. The latter process, that of visual observation, has no counterpart in philosophy. The charting of the logical powers of our ideas should not be interpreted as a plea for an ideal language of any kind. As we have seen, Ryle had abandoned this idea already at the time of ‘Categories’ (1938). In ‘Philosophical Arguments’ (1945) he felt the need to make clear, once and for all, that he no longer believed in the use or possibility of an ideal language.

The suggestion that men should coin a different diction to correspond with every difference in the logical powers of ideas assumes, absurdly, that they could be aware of these differences before being taken aback by the paradoxes arising from their naïvely attributed similarities. It is like suggesting that drill should precede the formation of habits or that children should be taught the rules of grammar before learning to talk. (Ryle 1945, 207)

The reductio ad absurdum is the instrument we use to determine our geographical map of propositions or ideas negatively. As we have seen, Ryle no longer preaches an ideal language, no philosophical doctrine, and no uniform relation between language and reality. He propagates a philosophical method which is essentially different from scientific ones in that it has a conceptual task, as opposed to the task of discovering new information. Denotationalism is banned for causing absurdities, a conclusion Frege and Russell should have drawn as well, but did not.

Conclusion

Ryle remained faithful to his early notion of philosophical analysis as an activity of detecting the sources of absurdities, although his more specific methods changed and his views on philosophical analysis became more positive throughout the years. Philosophy was first described as a mere negative activity, but later he described the reductio as having something in common with the destruction-tests by which engineers discover the strength of materials (Ryle 1945, 198). By 1938 Ryle’s early decompositional or reductive philosophical analysis had already changed into a connective type of analysis, focusing on the diversity of different ways in which our concepts have meaning and the connections between these concepts.

53

The Philosophical Development of Gilbert Ryle

We have seen that Ryle moved from his early Russellian search for an ideal language to a position more similar to that of the later Wittgenstein. Ryle soon abandoned the Russellian method of analyzing and reformulating expressions, as well as denotationalism, as he began to realize that denotationalism was responsible for many misunderstandings, sham-questions and problematic theories in philosophy. In terms of the positions I outlined at the beginning of this chapter, we may say that his philosophical method and approach changed from ‘Systematically Misleading Expressions’ (position 1), to his rejection of denotationalist theories of meaning (position 2). Although in 1938 Ryle already rejected the idea of a homogeneous theory of meaning, it took him a few more years to refine his thoughts into a heterogeneous ‘theory’ of use. ‘Conscience and Moral Convictions’ (1940) is probably the first paper in which the ‘use’ of a word (in this paper the word ‘conscience’) plays an important role, moving the focus away from ‘meaning’, which for Ryle implied a homogeneous relationship between language and reality, or that all expressions mean what they mean in the same way. ‘Use’, on the other hand, underscores the many different ways in which expressions ‘mean’, that is, are used. Ryle’s use-theory was further developed in The Concept of Mind (1949a) – which will be discussed in the next chapter. Throughout the rest of his philosophical career, he rejected homogeneous theories of meaning, moving towards a heterogeneous use theory of meaning. And it is primarily by means of this general idea, his rejection of the ‘Fido’ Fido theory of meaning, and his thoughts on the verifiability-principle that I will attempt to show that his later work – primarily The Concept of Mind – has often been misinterpreted and misunderstood.

54

The Concept of Mind: Main Aims, Method and Reception

Chapter 3

The Concept of Mind: Main Aims, Method and Reception

In the previous chapter we have seen that, although the early Ryle adhered to a denotational theory of meaning – that is a theory of meaning according to which all statements have meaning in the same way, in the way ‘Fido’ means Fido – at the time of The Concept of Mind (1949a) he had since long rejected denotationalism and in fact any type of homogeneous theory of meaning. By now he stressed the heterogeneity of our use of language. His method of philosophical analysis had changed from a logical and decompositional or reductivist type of analysis to a connective type, largely analogous to the task of a cartographer. The method employed in The Concept of Mind aims at showing that certain sorts of operations with concepts of mind go against the rules which these concepts ought to follow according to the logical geography of these concepts. No longer using a reductive type of analysis, Ryle does not talk about the deeper level of the logical form of facts, nor does he aim at an ideal language: his focus is on the possible and impossible connections between different (types of) concepts. Wittgenstein was probably an important source for Ryle – other sources of this new conception of philosophy are difficult to identify, which perhaps partly explains the frequent misinterpretations of the book. An understanding of Ryle’s development from a homogeneous, denotational theory of meaning to a heterogeneous theory of use is essential to a proper interpretation of The Concept of Mind. The hardheaded misinterpretation of this book as a behaviouristic tract can for example be exposed by pointing at Ryle’s anti-denotationalism and his ideas on what philosophy is. It should be noted that anti-denotationalism is a recurrent theme in Ryle’s later discussions of other philosophers; real progress implies a rejection of a denotational theory of meaning altogether (e.g. in his review of Carnap’s Meaning and necessity from 1949b and in his discussion of the historical development of the notion of meaning in ‘The Theory of Meaning’ from 1957). This development in Ryle has been relatively neglected.52 This is partly because Ryle’s main work has often been interpreted as a new theory of mind rather than as the meta-philosophical project which it in fact was. It may also be due to the fact that Ryle himself hardly mentions denotationalism with respect to his own position, but only in discussing other philosophers. However, an exclusive focus on this aspect of The Concept of Mind would not do justice to the plurality of the problems Ryle raises. Rejecting denotationalism is not his starting-point, although a denotational theory of meaning frequently turns out to be a poor instrument for charting mental concepts and indeed leads to absurdities. Ryle does not, at least not explicitly, focus on denotationalism since he may have realized that his theme is much richer and more complex than that. Moreover, his primary target is Cartesian dualism, which is not exclusively a philosopher’s problem, but is relevant to science and everyday life as well. And since people in general do not have views on denotationalism,

52 With the exception of MacDonald 1951, Passmore 1957, Ramoino-Melilli 1983, Antoniol 1993, Park 1994 and Rey 1997.

55

The Philosophical Development of Gilbert Ryle they cannot be persuaded of the absurdity of Cartesian dualism by proving that a denotational theory of meaning is wrong. Furthermore, we see a change of method; Ryle no longer wants to talk about theoretical, technical concepts (and denotationalism would be such a technical concept) but wants to start with what we already know. That is, he starts with the mental concepts as we use them, thereby avoiding the meta-level at which a discussion about denotationalism would take place. This chapter starts with a discussion of The Concept of Mind, focusing on Ryle’s method, his main target, that is Cartesian dualism, and his positive contribution to determining the correct ‘logical geography’ (Ryle 1949a, 10) of mental concepts. It is important to give a detailed account of the main category-mistake which Ryle attributes to Cartesian dualism. Secondly, I discuss some responses to The Concept of Mind, particularly the reviews by Hampshire (1950), Austin (1950) and Ayer (1970), which give a valuable insight in the reception of The Concept of Mind in the 1950s and which also help us to understand Ryle’s later development. Finally, after showing that Ryle was – and still often is – interpreted as a behaviourist, I shall criticise this persistent interpretation.

Ryle’s Method: Starting With What We Already Know

It has often been noticed that, compared to other philosophical books written at the time, as well as to some of the earlier writings of Ryle himself, The Concept of Mind stands out for its lack of technical language and footnotes, and for its clear and appealing examples. Although nowadays his use of examples may seem somewhat excessive, his style should be seen as a response to a technical style of philosophy common in his early days. Philosophers such as Bosanquet, Collingwood (at least in his early days) and Prichard had a strong aversion against using common sense examples, considering them inferior to abstract thinking. The discussion between Realists and Idealists which dominated Oxford for a long time was a typical example of the high level of abstraction and technicality. Ryle, however, considered technical language in philosophical texts (as well as an extensive use of footnotes) to be a weakness instead of an ideal. As a result, The Concept of Mind is pleasant to read. At first sight his examples seem plausible – sometimes they even seem trivial – and are not difficult to understand. However, Ryle does not explicitly state what he wants to put in the place of Cartesian dualism, i.e. he does not present a positive position that should replace the dualist one. In general, this makes it difficult to see what he exactly wants to achieve by means of his examples. In his review from 1950 Stuart Hampshire strikingly characterizes Ryle’s method and style:

The thought and the style are indissolubly linked in a manner which constitutes both the strength and, as it seems to me, also the weakness of the book; its strength, in that the reader is carried from the beginning to end by a single sustained impetus; its weakness, in that its argument seems somehow to fade and to lose some of its force when, laying the book down, one probes it again in some other and less powerful idiom. (…) There are many passages in which the argument simply consists of a succession of epigrams, which do indeed effectively explode on impact, shattering conventional trains of thought, but which, like most epigrams leave behind among the debris in the reader’s mind a trail of timid doubts and qualifications. (Hampshire 1950, 17-18)

An unsatisfactory feeling is indeed what many people have experienced after reading Ryle’s book. As will be discussed in the next section about the reception of The Concept of Mind, such a feeling is not only caused by Ryle’s frequent use of ‘epigrams’, his somewhat

56

The Concept of Mind: Main Aims, Method and Reception excessive use of examples, and the lack of a positive position and theoretical framework, but also by his looseness of expression and his use of vague concepts and expressions. Ryle’s style and method are closely connected with the intention he had in writing The Concept of Mind. The book was written with a meta-philosophical purpose in mind and it is important to realize that it was not primarily a theoretical interest in mental concepts as such that made him write the book; it was above all his intention to produce an example of good philosophy. He wanted to show what philosophical arguments are rather than to write a theoretical story about the requirements they should meet. His reason for choosing the concept of mind as the subject of his book was that he considered it to be a ‘notorious and large-sized Gordian knot’ (Ryle 1970, 12). In the introduction of The Concept of Mind, Ryle makes it perfectly clear that the assumptions which he argues against in this book are assumptions which he himself once embraced. Another preliminary observation he makes is that he does not intend to provide us with new information about . He does not want to increase the knowledge we possess about minds, but he wants to ‘rectify the logical geography of the knowledge we already possess’ (Ryle 1949a, 9). We already know so many things about minds, and in our daily life we can deal with them perfectly well. We can understand other people; we can discern their motives and influence their minds; and we can correct the mistakes we sometimes make in dealing with our own minds and those of others. So this we already ‘have’ before we start reading Ryle’s book. What we do not yet have, according to Ryle, is knowledge how we can correlate our concepts of mind with one another and with other types of concepts. He compares this situation to the one in which people, who have been living in a certain area for years, know all the roads in the area but cannot construct or even read a map of the region in which their area lies. Likewise, Ryle’s goal is to chart the ‘logical geography’ of our mental concepts. Why is it important for philosophy to do this? Ryle’s answer is that throughout the ages various disciplines – ethics, epistemology, political theory, aesthetics, etc. – have provided us with flawed arrangements of mental concepts into categories, the doctrine of Cartesian dualism underlying almost all of them. Ryle therefore wants to explode the Cartesian myth and replace it with a new map of mental concepts. He wants to:

reveal the logic of the propositions in which they are wielded, that is to say, to show with what other propositions they are consistent and inconsistent, what propositions follow from them and from what propositions they follow. (…) It is important to see to what logical type or category a concept belongs, because this gives us “the set of ways in which it is logically legitimate to operate with it”. (Ryle 1949a, 10)

In other words, the method we should use is to expose category-mistakes, that is, to show that some of the logical operations we perform with concepts of mind are not allowed because of the logical types or categories these concepts belong to – these operations go against the logical rules they should follow according to the logical geography of concepts of mind. As we have seen in Chapter 2, Ryle no longer aims at the construction of a formally logical calculus of any kind. He uses ‘logic’ in the sense of informal logic, signifying the ‘logic’ – that is the rules – of our use of language. Ryle provides us with several examples of category-mistakes. The most famous one is that of a foreigner visiting Oxford for the first time who, after being shown the colleges, libraries, sports facilities, departments, etc. asks: ‘But where is the university?’ The foreigner makes a category-mistake here, assuming that ‘university’ belongs to the same category as ‘college’, ‘library’, etc., whereas university is just the way in which all the

57

The Philosophical Development of Gilbert Ryle institutions that were shown to the foreigner are organized. ‘University’ belongs to a different category and cannot be treated in the same way as, for example, ‘library’ or ‘college’. In Ryle’s early papers, as we have already seen in Chapter 2, Ryle did not provide a precise definition of category-mistake; he will not do so in The Concept of Mind either. The more interesting examples of category-mistakes are, according to Ryle, the cases in which people are competent to apply concepts in at least some situations, but get confused when it comes to abstract thinking. Then they suddenly treat the concepts as belonging to logical types to which they do not belong, e.g. someone who can talk intelligibly about ‘the average taxpayer’ in discussions concerning tax, perhaps does not have a clue as to what to answer to the question: ‘why can’t you come across him in the street?’ If one treats ‘the average taxpayer’ as a fellow-citizen, one will attribute all kinds of ghostly qualities to him. It will be someone who is at the same time everywhere and nowhere. Showing that some of the logical operations performed on mental concepts lead to absurd conclusions, Ryle’s main type of argument is the reductio ad absurdum53. In this way he tries to show both to what logical types or categories the different concepts of mind belong and that certain operations are not allowed. This method is not merely destructive, but also achieves positive results, ‘bringing out part of the correct logic of mental-conduct concepts’ (Ryle 1949a, 24). As we have already noticed, Ryle compares it with the destruction-tests by which engineers discover the strength of materials. (Ryle 1945, 197- 198)54 He does not limit himself to the reductio, sometimes using other arguments which are – as he himself characterizes them – ‘of a less rigorous sort’ (Ryle 1949a, 10). These arguments are used in a therapeutic sense. People simply need them to get used to the possibility of convincing arguments against the intellectual habits they have had for a long time:

Philosophy is the replacement of category-habits by category-disciplines, and if persuasions of conciliatory kinds ease the pains of relinquishing inveterate intellectual habits, they do not indeed reinforce the rigorous arguments, but they do weaken resistances to them. (Ryle 1949a, 10)

Showing that mental happenings are not by definition conscious, Ryle for instance suggests that no one ever replies to the question ‘but do you really remember x?’ by answering ‘oh yes, for I am conscious of doing so’. According to Ryle, ‘this [argument] is not intended to be more than a persuasive argument’ (Ryle 1949a, 154). Another way to weaken people’s resistance against his more forceful arguments is to give a diagnosis of how the Cartesian dualists lost their way. Why do they, for example, see the mind as a world separate from, and opposed to, the one governed by mechanical theory? In showing that a conceptual confusion led the theorists to adopt Cartesian dualism in the first place, Ryle hopes to make people sensitive to the possibility that the view may be mistaken altogether. It is important to take into account the difference between rigorous arguments and those that are merely convincing. One of the misinterpretations partly responsible for the persistent idea of Ryle as a behaviourist – as will be shown in the next section – was that

53 As Ryle explained in ‘Philosophical Arguments’ (1945), he uses the strong rather than the weak reductio, showing that a proposition is illegitimate because it has logically absurd consequences. Cf. Chapter 2, pp. 52. 54 Cf. Chapter 2, p. 52. 58

The Concept of Mind: Main Aims, Method and Reception his verificationist arguments were considered to be of the rigorous type instead of being merely persuasive. The theory of meaning underlying Ryle’s arguments is a heterogeneous theory of use. The meaning of an expression is its ordinary use – use in the logical rather than the grammatical sense; and ordinary not in the sense of what people actually say in practice but rather in the sense of the norms of correct usage that are followed if one speaks (or writes) with care, that is our stock or plain use. Ryle frequently tries to show that the implications of what an expression means according to the official theory – that is the doctrine of Cartesian dualism – do not match with the ordinary use of this expression. E.g. according to the official theory we can never determine whether someone is intelligent or inventive because this would require us to look at shadow actions inside the person’s head. However, the use of labels such as ‘intelligent’ or ‘inventive’ in everyday life does not correspond with this. We use these labels precisely because we can test whether someone is intelligent; we do it all the time. The mistakes Ryle tries to lay bare are conceptual confusions, e.g. people fail to recognize metaphors, press analogies too far, treat abstract concepts as if they are things, and look for objects behind every word. Ryle’s examples and his diagnoses show that our constant urge to uniformity, homogeneity and simplification are mainly responsible for the mistakes that are being made. These mistakes arise, for example, from treating statements about the mental on a par with those about the body; from treating task-verbs (e.g. ‘hunt’, ‘look’, ‘treat’, ‘kick’) in the same way as success-verbs (e.g. ‘find’, ‘see’, ‘cure’, ‘score’) – assuming that task-verbs, too, indicate that ‘some state of affair obtains over and above that which consists in the performance’ (Ryle 1949a, 143); and from thinking that all sentences in the indicative report facts. If we believe that the dispositional sentence ‘Peter knows French’ reports a fact, we will look for a special occurrence inside Peter’s head. However, the disposition functions as an ‘inference-ticket’ (Ryle 1949a, 119), allowing us to infer from ‘Peter read a telegram in French’ to ‘Peter understood the telegram’, and as such does not imply an occurrence inside Peter’s head. Ryle does not make theoretical claims concerning uniformity and homogeneity, but they are recurrent themes implicitly present in his examples. These mistakes, at least the harmful ones, merely occur when one starts theorizing. There is nothing wrong with our ordinary use of expressions, in the sense of our stock or plain use. However, when we start talking ‘about’ these expressions – rather than ‘with’ them – as is the case when we start theorizing, we are in the danger of mistakenly, and dangerously, treating them analogous to others. We run this risk even more so when we assume – as we usually do – that sameness in grammar implies sameness in logical functioning (e.g. in the case of treating task-verbs as if they are success-verbs).

Ryle’s Primary Target

As was already mentioned, Ryle’s arguments are aimed at what he sometimes calls ‘the official doctrine’ or ‘the official theory’, at other times ‘Cartesian dualism’ or ‘Descartes’ myth’, ‘the double-life theory’ or ‘the ’. He does not attack the historical Descartes, but anyone who embraces this myth, such as Kant, Husserl, Frege and Russell; and not just philosophers but also scientists. In fact practically anyone who starts theorizing about the mind seems to ascribe to Cartesian dualism in one way or the other.

59

The Philosophical Development of Gilbert Ryle

The myth primarily consists of the idea that every human being has both a mind and a body. Human bodies are in space and obey mechanical laws, whereas minds are not in space and are not governed by mechanical laws. During a human life, body and mind are always together, but after a person’s death his mind continues to exist whereas the body dies. The myth implies that a person leads two different lives: one which consists of all the mental events, usually called internal events; and one which consists of all the bodily events, that is, all the external events. Bodily events consist of matter or are functions of matter, whereas mental events consist of consciousness or are functions of consciousness. Just as the foreigner visiting the expected it to be something like another college or playground, but also considerably different, so the official theorist expects the mind to be something like another body, though considerably different. Mental processes are causes and effects, but causes and effects of a different sort than those of bodily processes. According to the ‘official theory’, in normal circumstances we have a direct and infallible access to the workings of our own mind by means of our consciousness. On top of this direct access, a person is from time to time also able to use introspection. In other words, he can reflectively ‘watch’ what goes on in his own mind. This active ‘watching’ of one’s inner is also insusceptible to confusion or mistakes. When we actively ‘watch’ the workings of our own minds, we cannot be mistaken in what we ‘see’. Ryle not only wants to show what is wrong with the official doctrine, but also to explain how the official doctrine came into being. Why did people believe it in the first place? According to him, the basis of the doctrine lies in Galileo’s discovery that his methods of scientific discovery could provide a mechanical theory which covered everything in space. From a scientific point of view, Descartes believed in the claims of mechanics, but from a moral and religious point of view he did not want to believe that humans do not differ in kind but only in perplexity from mechanistic brute animals or automata. His escape-route led to the official doctrine. Defenders of this doctrine try to describe minds in what is merely a vocabulary obverse to the vocabulary in which they try to describe bodies, e.g. ‘not in space’, ‘not accessible to other people’, ‘not visible’.

The differences between the physical and the mental were thus represented as differences inside the common framework of the categories of ‘thing’, ‘stuff’, ‘attribute’, ‘process’, ‘change’, ‘cause’ and ‘effect’. Minds are things, but different sorts of things from bodies; mental processes are causes and effects, but different sorts of causes and effects from bodily movements. And so on. (Ryle 1949a, 20-21)

The belief that there is a polar opposition between matter and mind implies that both terms are of the same logical type, that they belong to the same category. Thus, according to ‘the official theory’, minds belong to the same category as bodies. And because bodies are rigidly governed by mechanical laws, minds must be rigidly governed as well (by non- mechanical laws); and because the material world is a deterministic system, some theorists think that the mental world must be deterministic as well, and in this world there is, as a consequence, no room for responsibility, freedom of choice, etc. So, the ‘official doctrine’ is, next to being the cause of many problems concerning theories of mind, also – at least partly – responsible for the problem of the freedom of the will. The view that mind and body belong to the same category reflects a specific theory of meaning. ‘Since mental-conduct words are not to be construed as signifying the occurrence of mechanical processes, they must be construed as signifying the occurrence of non-mechanical processes.’ (Ryle 1949a, 20) Apparently the Cartesian myth presupposes that there is one homogeneous account of the signification of words. 60

The Concept of Mind: Main Aims, Method and Reception

Ryle claims that the polar opposition between internal and external events, between mind and matter, which is implied by Descartes’ myth, causes immense trouble, even if taken metaphorically. One of the main problems is how mind and matter could possibly influence one another. The causal connections between mind and matter can be neither material nor mental, so what do they consist of? Neither the physiologist nor the psychologist can provide us with an answer to that question. Another problem is that whereas bodies are mechanically connected and can therefore influence one another, this is not the case with minds. There do not exist any direct causal connections between different minds. The consequence is that people cannot know what goes on in other people’s minds. Of course they can make inferences from a person’s observed behaviour to the person’s state of mind, but these inferences are, at the least, problematic. We cannot even claim to know for sure that there are minds other than our own. Based on the official doctrine, there is just no way of confirming this claim. Thus, according to the official logical geography of mental concepts, these concepts, Ryle argues, cannot be used effectively in our descriptions of, and prescriptions for, other people’s minds. Because this totally conflicts with our everyday life in which we do successfully employ mental concepts, the official logical geography of mental concepts must be mistaken. Ryle’s aim is to show that when we describe people in terms of mental concepts such as intelligent, careful, inventive, etc., we are not referring to mystical or occult episodes as causes of which observable actions and utterances are the effects. We rather talk about these actions and utterances themselves. Ryle hopes ‘to prove that [Cartesian dualism] is entirely false, and false not in detail but in principle’ (Ryle 1949a, 17), that it is one big category-mistake. But what exactly is this mistake? Does Ryle think that the dualism he attacks should be replaced by monism? The fact that he refrains from explicitly stating what he considers the category- mistake to be in the case of Cartesian dualism, creates problems with respect to the interpretation of the entire book. Different interpreters have attributed to Ryle different category-mistakes, e.g. those who interpreted him as a behaviourist have considered the category-mistake to be the mistake of treating statements about the mind as if they belong to a different category than statements about the body. For now, I take the ‘big mistake’ Ryle refers to, to be, roughly, that of treating statements about the mind in the same way as bodily statements. Statements about the mind are mistakenly supposed to have meaning in the same way as those about the body, namely by their . I shall qualify this claim at the end of this section. Why did Ryle not explain what exactly this category-mistake was? Probably because this would mean telling a more theoretical story than the one Ryle wanted to do. His project is to start with what we already know, with our use of everyday language, and from there on to tackle Cartesian dualism. In doing so, Ryle attacks the following ideas which belong to the official doctrine in some detail, thinking that they are primarily responsible for the absurdities it brings about: firstly, the idea that theorizing is the primary activity of minds and that it is intrinsically a silent, internal and private operation; secondly, the view that intelligent doings necessarily consist of two processes – one of acting and one of theorizing. The well-known and influential distinction he makes between ‘knowing how’ and ‘knowing that’ plays a central role in his attack of these two ideas, running through the book like a continuous thread. Starting from our use of language, these dualist ideas and others which follow from them are tackled by means of the reductio ad absurdum as well as ‘less rigorous’ – ‘therapeutic’ – arguments.

61

The Philosophical Development of Gilbert Ryle

Theorizing as the Primary and Silent Activity of the Mind

In his rejection of the idea of theorizing as the primary activity of the mind, Ryle first introduces the distinction between ‘knowing how’ and ‘knowing that’. ‘Knowing that’ is what we normally call theorizing; it is knowledge of facts or operations and is also characterized as the consideration of propositions. Our knowledge that the proposition ‘2+2=4’ is true is an instance of ‘knowing that’. ‘Knowing how’, on the other hand, refers to knowing how to do something, indicating an ability rather than the fact that something is the case or that something is happening. A clown knowing how to make people laugh is a clear example of ‘knowing how’. The clown has the ability to make people laugh but this ability does not, at least not exclusively or necessarily, consist in his consideration of propositions. Ryle tries to show that ‘knowing how’ and ‘knowing that’ belong to different categories by demonstrating some differences in our use of them, for example that we cannot partly understand or partly know a fact, whereas we can partly know how to play chess and partly understand how to play chess. Another difference is that ‘knowledge how’ to play chess is acquired gradually, whereas ‘knowledge that’ is acquired at a certain moment. It is perhaps tempting to describe Ryle’s difference between ‘knowing that’ and ‘knowing how’ as that between ‘explicit knowledge’ and ‘implicit knowledge’. At first sight this seems to be plausible, since most of the time we use ‘explicit knowledge’ for being able to recite certain rules, for example grammatical ones, and we use ‘implicit knowledge’ for the cases in which we cannot recite the exact rules but in which our practice betrays a certain type of knowledge after all. For example, ‘implicit knowledge’ is used for the kind of knowledge we express in correctly using our native language while we cannot (or may not be able to) recite its grammatical rules. However, we do have to be careful in using ‘knowing that’ and ‘explicit knowledge’, and ‘knowing how’ and ‘implicit knowledge’ interchangeably, since ‘explicit’ and ‘implicit’ may trick one into thinking that they are the same types of knowledge, but one is conscious and the other unconscious. In the case of our use of our native language one could be lead into believing that both our implicit and our explicit knowledge consist of the same set of grammatical and syntactical rules (propositions) inside our head, the only difference being that in the case of explicit knowledge we are conscious of their presence and in the case of implicit knowledge we are not. This would not at all be what Ryle wanted to propose. If one wants to equate ‘implicit knowledge’ with ‘knowing how’ at all one has to let go of the idea that it is of the same type as ‘explicit knowledge’, the only difference being that it exists on an unconscious level. Thus, the difference between ‘knowing how’ and ‘knowing that’ can be described as that between ‘implicit knowledge’ and ‘explicit knowledge’, but this will probably not contribute to a better understanding since the second distinction would then have to meet the specific requirements Ryle contributed to the first. According to Ryle, the adherents of the official doctrine only take into account ‘knowing that’ and totally neglect ‘knowing how’, which in fact plays a much larger role in our everyday life. We are much more concerned with other people’s abilities than with their knowledge of facts. Consider a boy who learns how to play chess. He can either learn how to play chess by looking at people who play the game correctly, or he can learn the rules by heart and then learn how to apply them. And finally, it is not what the boy does in his head, but what he does on the board that shows whether or not he knows how to play chess, just as cleverness at fighting is exhibited in the giving and parrying of blows, not in the acceptance or rejection of propositions about blows. Adherents of the official doctrine can

62

The Concept of Mind: Main Aims, Method and Reception only claim that theorizing is the primary activity of minds, because they simply neglect the fact that minds are also, and even more, concerned with ‘knowing how’, which neither is theorizing nor depends on it. What is it like to know how to do something? The first condition Ryle mentions is to perform well: for us to describe a person as knowing how to do something his performances must meet certain standards. The second condition is that the person must be responsible for his performance; he must be able to detect and correct mistakes, learn from others and improve upon successes, which is often described as: ‘the agent is thinking what he is doing while he is doing it, and thinking what he is doing in such a manner that he would not do the action so well if he were not thinking what he is doing.’ (Ryle 1949a, 29) This does not mean that the person who is performing an intelligent action, e.g. reading, at the same time has to consider certain propositions and additionally has to put them into practice. It simply means that the agent must pay attention to what he is doing (heedingly do what he is doing). There is only one action, namely reading, and this action has a special character. By providing examples which do not by themselves give conclusive arguments against the official doctrine, Ryle tries to weaken people’s resistance to his attacks to come. To give his conclusive argument right away would probably have aroused too much aversion. Ryle tries to let the reader get used to the idea that perhaps not every intelligent action consists of first considering certain relevant propositions and then acting upon them, even though this may be so in the case of chess.

Yet the general assertion that all intelligent performance requires to be prefaced by the consideration of appropriate propositions rings unplausibly, even when it is apologetically conceded that the required consideration is often very swift and may go quite unmarked by the agent. (Ryle 1949a, 30)

There are situations in which people know how to execute intellectual operations without knowing that the relevant propositions are true. A good clown knows how to act during his performances, but may not be able to teach others how to do it, or even mention all the relevant propositions to himself. And Aristotle knew perfectly well how to avoid and detect fallacies even before he tried to formulate the maxims for reasoning intelligently. These examples are not meant as conclusive evidence – and should not be judged as such – but are merely supposed to weaken the reader’s resistance to Ryle’s attacks on the doctrine. Finally, only after these less forceful arguments does Ryle present his, what he calls, ‘crucial objection’ (Ryle 1949a, 31) to the intellectualist legend (roughly, the idea that intelligence refers primarily to the class of theorizing-operations):

The consideration of propositions is itself an operation the execution of which can be more or less intelligent, less or more stupid. But if, for any operation to be intelligently executed, a prior theoretical operation had first to be performed and performed intelligently, it would be a logical impossibility for anyone ever to break into the circle. (Ryle 1949a, 31)

Ryle claims that the idea that considering relevant propositions is a necessary condition for performing an intelligent action is absurd. In performing an intelligent action A we would have to consider the propositions relevant to A, but considering these propositions is itself, if done correctly, an intelligent action (let us call it B) and should therefore be preceded by considering those propositions relevant to B, etc. Knowing how to apply A cannot be reduced to or derived from the acceptance of the relevant maxims, because for a person to intelligently consider these maxims, he has to intelligently consider other maxims relevant

63

The Philosophical Development of Gilbert Ryle to considering the relevant maxims for knowing how to apply A. The regress is infinite and therefore the intellectualist legend is absurd, because it implies that for an operation to be intelligent it should be preceded by a prior intellectual operation. Consequently, ‘knowing how’ cannot be reduced to ‘knowing that’. Ryle considers the view that theorizing is essentially a silent, internal and private operation to be mistaken as well, although he does not provide a rigorous reductio here. First of all, silence is not relevant; the criteria by means of which our actions are judged as intelligent or not intelligent are the same in both cases, e.g. intelligent argumentations should be cogent, clear, relevant and well organized, independent of whether one is arguing aloud or silently. Moreover, Ryle argues that before we can think silently, and before we can talk to ourselves in silence, we have to be able to talk out loud. Silent thinking cannot be done without effort. It makes no sense to claim that theorizing is intrinsically a private or internal operation, since we have to theorize out loud before we can do it silently.

One Action: Not Two

So, theorizing – ‘knowing that’ – is not the primary activity of the mind, nor are intelligent performances essentially silent, internal and private operations. A second, related, dogma is that intelligent operations must consist of two fundamentally different processes; one of doing or acting and one of theorizing. The argument goes like this: actions concern muscular activities, and therefore they cannot be mental operations. So, to be able to attribute intelligence to an act, we have to refer to processes different from muscular activities; we have to refer to a process of theorizing. Against this, Ryle argues:

What distinguishes sensible from silly operations is not their parentage but their procedure, and this holds no less for intellectual than for practical performances (…) When I do something intelligently, i.e. thinking what I am doing, I am doing one thing and not two. My performance has a special procedure or manner, not special antecedents. (Ryle 1949a, 32)

Ryle uses an example of a clown to show that although it is true that there may be no overt difference between the act of a clown and someone who is very clumsy, this does not mean that the clown is performing some extra secret act. The tripping of a clown, which he does on purpose, is both a mental and a bodily process, but it is not two processes. It is not the case that the fact that the clown’s operations are intelligent implies that in his head there must occur a counter performance to the overt performance that is taking place. In showing what then constitutes the difference between the clown and a clumsy man, Ryle introduces the concepts ‘skill’ and ‘disposition’. In contrast to the clumsy person, the clown acts intelligently because he has a certain skill. And since a skill or disposition is not an event, it cannot be seen (or unseen) by others – not because it is an extra, secret event, but because it is not an event at all. Ryle later gives the following description of ‘disposition’: ‘Moreover, merely to classify a word as signifying a disposition is not yet to say more about it than to say that it is not used for an episode.’ (Ryle 1949a, 112) To say that someone has a certain disposition to do something is not to say that he is at a particular moment in the process of doing (or undergoing) something, but rather that he is able or prepared to do certain things when certain situations occur. Hypothetical propositions describe dispositions whereas categorical propositions describe episodes. However, many propositions are neither purely categorical, nor purely hypothetical. These propositions are what Ryle calls ‘mongrel categorical’, or ‘semi-

64

The Concept of Mind: Main Aims, Method and Reception hypothetical statements’ (Ryle 1949a, 135). ‘The bird is migrating’ is an example of such a mongrel-categorical proposition, being more categorical (or episodic) than ‘the bird is a migrant’ and more hypothetical (or dispositional) than ‘the bird is flying to Africa’. The simplest dispositions can be described by a simple single-track disposition, such as: ‘this glass is brittle’ (meaning: ‘if this glass ever is or ever had been struck or stained, it would fly, or have flown, into fragments.’), or ‘I am an habitual smoker’ (meaning: ‘I am permanently prone to smoke when I am not eating, sleeping, lecturing or attending funerals and I have not recently been smoking’), and all other habits and reflexes. These propositions are called single-track dispositions because their actualisations are nearly uniform. There also exist higher-grade dispositions, that is, dispositions the exercises of which are indefinitely heterogeneous, and these are what Ryle is mainly concerned with in The Concept of Mind. Dispositions like ‘knowing’ and ‘believing’ are examples of higher-grade dispositions. Many epistemologists who expect all dispositions to be single- track ones, suppose that there are specific intellectual processes by which these dispositions are actualized. They expect to find an action which realizes ‘Peter is intelligent’, similar to the one that realizes ‘Peter is a smoker’. Because ‘Peter is a smoker’ is true by virtue of ‘Peter smokes now’ being true regularly, they are convinced that there must be a specific episode by virtue of which ‘Peter is intelligent’ is true. Since they cannot find such an episode, they assume that it must be a secret unobservable one inside Peter’s mind. Consequently, doing something intelligently is supposed to mean that two processes are going on: an observable and public one of ‘doing’ and a private one inside one’s head. Ryle claims that adherents of the official doctrine have wrongly argued from the type-distinction between disposition and episode to, respectively, secret and hidden mental causes and their overt physical effects, whereas the difference between dispositions and episodes should have left them with dispositions that are neither witnessable nor unwitnessable and episodes that are witnessable. Once more, it seems that the theorists have been allured into treating mental concepts as if they are bodily ones. The idea that logically complex propositions – such as ‘mongrel-categorical ones’ (Ryle 1949a, 135), that is, those that are neither purely categorical, nor entirely hypothetical – are descriptions of complex processes also contributed to the mistaken view that intelligent operations consist of two separate processes (one internal and one external). But the greater complexity of the description of a bird as migrating compared to the description of a bird flying in the direction of Africa, does not consist in describing a larger number of processes. ‘Only one thing need be going on, namely that the bird be at a particular moment flying south. “It is migrating” does not tell more stories, but tells only a more pregnant story than the one told by “It is flying south”. It can be wrong in more ways and it is instructive in more ways.’ (Ryle 1949a, 136) Similarly, if an agent is described as doing something (say reading) which he knows how to do, he is described as paying attention to what he is doing (heedingly do what he is doing). There is only one action, namely reading, and this action has a special character. But then, if it is not the case that there are two processes, how can we reply when asked why a person is reading a book: ‘because he is interested in what he is reading’ (Ryle 1949a, 137)? The answer is that we should not consider the interest to be the real cause of the reading. ‘The interest explains the reading in the same general way, as the migrating explains the flying south.’ (Ryle 1949a, 137) It is a matter of explanation, not causation. The migrating is not an additional process but ‘it describes an event in terms which are law- impregnated’ (Ryle 1949a, 137), warranting the inference ‘it is a migrant’. So we now know that we have to describe intelligence in terms of higher-grade dispositions, and we also realise that these dispositions do not have uniform actualizations

65

The Philosophical Development of Gilbert Ryle and that complex descriptions do not signify complex processes. But how do we distinguish between intelligent and unintelligent actions? How can we know whether someone has a certain skill, e.g. whether he knows how to shoot? How do we know that he does not accidentally meet the standards, e.g. shoots in the bulls eye by pure luck? According to Ryle, we have to look beyond the performance itself and determine the skills and abilities of the person whose actions we are judging. This does not mean that we have to make inferences to occult causes. It is only going ‘beyond’ in the sense of considering the powers and propensities of which people’s performances are exercises.

But in looking beyond the performance itself, we are not trying to pry into some hidden counterpart performance enacted on the supposed secret stage of the agent’s inner life. We are considering his abilities and propensities of which their performance was an actualization. Our enquiry is not into causes but into capacities, skills, habits, liabilities and bents. (Ryle 1949a, 45)

But is trying to grasp someone’s skills and abilities not simply an attempt to spot secret occurrences inside the person’s head? According to Ryle, a consideration of a modest number of heterogeneous performances usually suffices to establish whether someone who just shot in the bull’s eye really knows how to shoot or was just lucky. For example, if the same person after his one shot tries to shoot in the bullet’s eye again and succeeds, we will normally be prepared to say that this person knows how to shoot. Ryle claims that: ‘The mind is not the topic of sets of untestable categorical propositions, but the topic of testable hypothetical and semi-hypothetical propositions.’ (Ryle 1949a, 46) In understanding each other’s arguments or actions we are not inferring to the workings of other people’s minds, but we are, according to Ryle, ‘following’ them. Understanding other people, knowing about their capacities, interests and methods, is nothing more than appreciating how their overt doings are conducted. We are merely following and not causally explaining other people’s minds.

The Issues Under Discussion

In The Concept of Mind Ryle applies the method described above to several complex problems which are connected to the concept of mind, treating ‘the will’, ‘’, ‘dispositions and occurrences’, ‘self-knowledge’, ‘sensation and observation’, ‘imagination’, ‘the intellect’ and ‘psychology’. The assumption of extra things, objects and processes is what Ryle argues against in all chapters. He rejects the idea of a separate faculty of the will and that processes occur which correspond to acts of the will. And our explanation of other people’s behaviour in terms of does not entail references to ghostly inner states or processes. Most emotions are not states or processes but propensities, which are dispositions rather than episodes. Consequently, they provide us with reasons for the actions of other people (and our own); not with causal explanations. The most important claim Ryle makes with respect to sensation and observation is that sensations are of the wrong category to be observed. Sensation and observation belong to different categories and neglecting this difference was partly responsible for the sense- datum theory – according to which a squinter who reports that he sees two candles is said to be seeing two ‘candle-looks’ – as well as . Both postulate extra entities next to ordinary objects; sense-datum theorists postulate images and phenomenalists postulate sensible objects, these images and sensible objects being what is really observed. Ryle’s

66

The Concept of Mind: Main Aims, Method and Reception account of people who upon encountering a round plate describe it as elliptical, is that the round plate looks ‘as if’ it is elliptical. This does not require introducing new objects, such as elliptical images or sensible objects. Ryle’s rejection of the theory of special status images, that is, the view that if we visualize or imagine something we see a picture with a special status, runs along the same line. Again, visualizing is not observing, just as sensations are not observings. We make a category-mistake if we describe our imagining a dragon as seeing dragon-fantasies. Creating these kinds of extra objects is what Ryle is arguing against. If a child imagines that his doll is smiling he does not see an image with a special status of the smile. As sham- murders in a play are not murders, in the same way imagined images and sounds are not really images and sounds. As we have already seen, Ryle rejects the view that ‘knowing that’ is primary rather than ‘knowing how’, which is clearly reflected in his chapter on the intellect (chapter 9). With respect to the intellect the category-mistake often made by epistemologists is that they take concepts such as ‘deduction’ and ‘judgement’ to refer to actions of pondering (that is, ‘path-making’) rather than to refer to the results of our ponderings (that is, ‘path- using’). As a result, they ended up with extra actions: acts of judgement, acts of deduction and acts of abstraction. The same Occamising zeal can be found in Ryle’s chapter on self-knowledge, which will be discussed in some detail below. This subject will prove to be particularly relevant for the remainder of this chapter, since a considerable part of the comments of contemporary reviewers were directed at Ryle’s treatment of self-knowledge. His later work on the nature of thinking – which will be discussed in the next chapter – is also related to it.

Case-study: Self-Knowledge

An important theory that belongs to the Cartesian dualist doctrine is that there are specific ways of discovering what is going on in the mental world, as counterparts to the ones that reveal what happens in the physical world. This ‘official doctrine’ puts forward consciousness and introspection as the ways to achieve self-knowledge, that is knowledge of our own mental occurrences. In order to show how Ryle makes different types of arguments work together, I discuss some of his arguments concerning his rejection of self- knowledge as immediate, private and infallible knowledge. He uses the reductio as well as arguments which are less rigorous, and, as part of the latter type, diagnoses of how people came to believe these doctrines. He both wants to show that the official theories of consciousness and introspection are ‘logical muddles’ (Ryle 1949a, 149), and how we can have self-knowledge without these doctrines. The official dualist theory of consciousness is that a person’s mind cannot help being constantly aware of the contents of the private room in the person’s head, which has little affinity with the more common uses of consciousness (e.g. being awake or paying attention). The official story of introspection is the idea that a mind can, by means of some sort of non-sensuous perception, deliberately examine and acquire knowledge of at least some of its own states and operations. Introspection differs from consciousness in that everyone is constantly conscious, while awake, whereas introspection is an act of attention; at the times we introspect we are interested in what is going on in our mind. Both consciousness and introspection are supposed to be mistaken-, as opposed to

67

The Philosophical Development of Gilbert Ryle knowledge of the inner states and operations of other people. Furthermore, adherents of the two-world myth do not claim that the fact that mental processes are conscious means that we do or could report on them post mortem but that we are aware of what is engaging our notice exactly at the moment it is engaging our notice. It is considered to be a non- dispositional type of knowledge. Ryle rejects this idea of self-knowledge and considers it to be nothing more than a myth. This does not, however, entail that we cannot know what there is to know about ourselves but only that this is not fundamentally different from what there is to know about other people. There are some differences in degree between what I can know about myself and what I can know about other people, because there are differences in the requirement of the relevant data in both cases. But it is definitely not the case that it is always easier for me to find out a certain thing about myself. And self-knowledge is certainly not infallible. An important advantage of dropping the privileged-access view is that we lose the danger of epistemological isolationism, the idea that we can only ‘really’ know things about ourselves and not about other people and that we are in this sense completely isolated from all other people. Challenging the claim that mental states and operations are by definition conscious, Ryle makes use of several arguments. The first one depends upon his rejection of the dogma of . If there is no such thing as a mental world, then, obviously, no occurrences can take place in that world. The other arguments do not depend on a rejection of the dualist dogma, asserting for instance that in everyday life no one ever tries to vindicate any of his assertions of fact by saying that he found it out ‘from consciousness’ (Ryle 1949a, 154). No one ever replies to the question ‘but do you really remember x?’ by answering ‘oh yes, for I am conscious of doing so’. Ryle explicitly states that ‘this is not intended to be more than a persuasive argument’ (Ryle 1949a, 154). Another argument is aimed at the alleged infallibility of consciousness, purporting to show that ‘there is no contradiction in asserting that someone might fail to recognize his frame of mind for what it is’ (Ryle 1949a, 155). One may think of people deceiving themselves about their motives, or those who do not know that they are dreaming when they are in fact dreaming. If consciousness were infallible, as adherents of the dogma of the ghost in the machine claim it to be, it would be logically impossible for such failures and mistakes to take place. Finally, Ryle presents his most rigorous argument, claiming that ‘it makes sense to ask whether, according to the doctrine, I am not also conscious of being conscious of inferring’ (Ryle 1949a, 156). This would lead to infinite regression, to an infinite number of ‘onion-skins of consciousness’ (Ryle 1949a, 156). The only way to avoid such a regress is to claim that there are elements in mental processes which we cannot be conscious of, but this would be to admit that ‘conscious’ cannot be part of the definition of ‘mental’. What about introspection? Ryle’s main objection against it closely resembles his infinite regression argument against consciousness. Our acts of introspection are themselves also mental processes and should therefore also be introspectable. An infinite regression of acts of attention seems unavoidable. Or else we would have to admit that there is a limit to the number of synchronous acts of attention, which means that there must be at least some mental processes which are unintrospectable, namely when the limit to the number of synchronous acts of attention is reached. This means that a person’s knowledge of his own mental states does not always rest on introspection, and then why would it ever? In the cases where it does not, an escape-route would be to appeal to consciousness for our knowledge that we introspect. However, the theory of consciousness has already been shown to be far from unproblematic. 68

The Concept of Mind: Main Aims, Method and Reception

Moreover, as was first put forward by Hume, there are some states of mind which cannot be ‘coolly’ introspected, since our being in those states involves precisely that we are not cool, for example being in a state of panic or fury. According to Ryle, states of mind like these can be examined, but only in retrospect. He even goes further, asking: if retrospection, that is, to catch oneself doing so and so, can give us the data we need for our knowledge of some states of mind, why cannot retrospection give us the data we need for our knowledge of all states of mind? Retrospection is a genuine process which neither has problems concerning a limit to the number of synchronous acts of attention, nor does it have the problem that some uncool states of mind cannot be the objects of cool introspection. These arguments purportedly show that the claim that people can in principle have direct and failure-proof access to their own mental events is problematic. What people do have is the ability to retrospect, but this is not limited to ‘mental events’, and therefore cannot serve as evidence of the authenticity of mental events. The reports which a person makes about himself are not free from bias or carelessness, but are subject to the same kind of defects as his reports on other people. According to Ryle, part of what people have in mind when they talk about introspection is really retrospection, the objects of which can be both silent, private acts and overt acts. I can retrospect seeing things, as well as imagining things, and I can report both the calculations that I have been doing in my head and the ones I have been doing out loud in the classroom. Ryle compares a person’s retrospection to a person’s diary: ‘a diary is not a chronicle of ghostly episodes, but it is a valuable source of information about the diarist’s character, wits and career.’ (Ryle 1949a, 160) Keeping in mind the idea of the diarist, we should not speak of someone’s mind knowing this or that, but of someone knowing this or that, just as we cannot speak of someone’s eyes seeing this or that, but only of a person seeing this or that. My mind is not another organ, but it signifies my ability and proneness to do certain things. Therefore, questions about the relations between a person and his mind, like those about the relations between a person’s body and his mind, are improper questions: ‘We ought to follow the example set by novelists, biographers and diarists, who speak only of persons doing and undergoing things.’ (Ryle 1949a, 161) But if there is no such thing as direct and infallible self-knowledge, then what is the difference between self-knowledge and knowledge of others? For Ryle, there is no difference in principle, only a difference in degree: ‘The ascertainment of a person’s mental capacities and propensities is an inductive process, an induction to law-like propositions from observed actions and reactions.’ (Ryle 1949a, 164) In real life we are quite familiar with the techniques of assessing persons and accounting for their actions, whereas according to the standard theory no such techniques could exist. And we never in fact appeal to privileged access:

(…) in none of these senses in which we ordinarily consider whether a person does or does not know something about himself, is the postulate of a Privileged Access necessary or helpful for the explanation of how he has achieved, or might have achieved, this knowledge. There are respects in which it is easier for me to get such knowledge about myself than to get it about someone else; there are other respects in which it is harder. But these differences of facility do not derive from, or lead to, a difference in kind between a person’s knowledge about himself and his knowledge about other people. (Ryle 1949a, 173)

We know quite well how to find the answer to the question whether person x has understood the argument. We must observe this person’s conduct, remarks and tone of voice. Normally, ordinary day-to-day observation suffices. Although there is no single

69

The Philosophical Development of Gilbert Ryle performance which could determine that person x had understood the argument, this does not, according to Ryle, imply that there is no finite set of sub-sets that does.

A More Accurate Account of the Category-mistake of Cartesian dualism

We may now develop our initial account of the category-mistake as treating statements about the mind as if they were about the body, that is, as treating mental statements analogous to bodily ones. To describe Ryle’s category-mistake in this way would be to characterize his position as conceptual dualism. And we have now seen that his category- mistake in fact refers to a variety of mistakes, which reveal that not even all bodily statements function alike, and neither do all mental statements; hence it would not do to qualify Ryle’s position as conceptual dualism. In my discussion so far I have mentioned several mistakes, pointed out by Ryle, which mainly consist in treating one type of statements as if they are of another type, for example because grammatical similarity between two statements is taken as a clue of sameness in type of meaning. The most important ones are:

1) treating dispositional statements as if they are categorical (e.g. our assumption that ‘Peter is intelligent’ has a meaning in the same way as ‘Fred is feeding the dog’ is responsible for our urge to look for an episode or process taking place inside Peter’s head, analogous to the one of ‘feeding the dog’.); 2) treating higher-grade dispositions as if they are single-track (e.g. if we treat ‘Peter is intelligent’ analogous to ‘Peter is a smoker’ we mistakenly look for a uniform actualization of ‘is intelligent’.); 3) treating mongrel-categorical or semi-dispositional statements as if they are categorical (e.g. treating ‘Peter is driving carefully’ analogous to a categorical one leads us to analyse the statement into two doings: ‘driving’ and an additional doing corresponding to ‘carefully’.)

These category-mistakes are often caused by our general creed for uniformity, homogeneity and simplification, and, more specifically, our employment of a homogeneous, denotational theory of meaning. We make the mistake of treating all concepts as names, whereas mental concepts are in fact qualifying; mental predicates are adverbial rather than substantive. We are tricked into treating ‘intelligent’ as referring to a thing or process and in doing so we create a realm of things and processes that do not exist in the bodily world, but merely in a ghostly, mental world. Whereas our complex descriptions of intelligent actions are considered to reflect a complexity in processes, they should rather be regarded as higher- level descriptions of these same processes. The mental predicate ‘intelligent’ does not describe a process but rather the way in which all kinds of processes are managed, neither does ‘carefully’ in ‘Peter is driving carefully’ refer to a process other than driving; there is no process corresponding to ‘carefully’. This interpretation of the main category-mistake of Cartesian dualism as treating all concepts as names is supported by several passages in The Concept of Mind:

But to speak as if the discovery of a law were the finding of a third, unobservable existence is simply to fall back into the old habit of construing open hypothetical statements as singular

70

The Concept of Mind: Main Aims, Method and Reception

categorical statements. It is like saying that a rule of grammar is a sort of extra but unspoken noun or verb, or that a rule of chess is a sort of extra but invisible chessman. It is to fall back into the old habit of assuming that all sorts of sentences do the same sort of job, the job, namely, of ascribing a predicate to a mentioned object. (Ryle 1949a, 117-18)

The recommended restoration of the trade-names of traditional epistemology to their proper place in the anatomy of built theories would have a salutary influence upon our theories about minds. One of the strongest forces making for belief in the doctrine that a mind is a private stage is the ingrained habit of assuming that there must exist the ‘cognitive acts’ and ‘cognitive processes’ which these trade-names have been perverted to signify. (Ryle 1949a, 299)

My earlier description of the category-mistake as treating mental statements and bodily ones in the same way can now be made more precise. If we made a scheme consisting of types of statements, and we were to put the categorical ones at the bottom of the page and work our way up to the purely hypothetical ones, it would be possible to say that the category-mistakes Ryle mentions are all specimens of treating statements as if they are lower in the scheme than they really are, that is treating them as if they are more ‘categorical’ or more ‘bodily’ than they in fact are. By ‘treating statements as if they are (like) …’, I mean treating them as if they mean (what they mean) in the same way. In the end, a rigid, homogeneous theory of meaning – the ‘Fido’-Fido theory of meaning – is underlying these mistakes and is therefore largely responsible for the official doctrine of Cartesian dualism. So, Ryle does not appear to be a dualist, not even a conceptual dualist; rather, he is what we may call a conceptual pluralist. It is not the case that the meaning of all statements about the body are determined in one way and the meaning of all statements about the mind in another way; there are not just two categories of meaning. It is rather that categorical statements get their meaning in a different way than hypothetical ones, or semi- categoricals, etc. There are many different types – or categories – of meaning. As Ryle puts it later, in 1957, in his paper ‘The Theory of Meaning’: ‘there is an indefinitely large variety of kinds of roles performed by the expressions we use in saying things. (…) there is an endless variety of categories of sense or meaning.’ (Ryle 1957a, 365) Confusing these categories with each other results in conceptual mistakes and wrongheaded theories, such as Cartesian dualism.

The Reception of The Concept of Mind

The Concept of Mind has been of great importance to analytic philosophy. The book has become a classic and has been studied by virtually every philosophy student in England for many years. It is therefore interesting to see how the book was received by Ryle’s contemporaries, not in the least because some of these comments led Ryle to develop and change his views, especially on thinking, in his later writings. It is also valuable to see how more recent philosophers have interpreted Ryle, and how his book is received today. To this end I shall describe the main characterizations of Ryle in handbooks in philosophy and cognitive science. The interpretation of Ryle as a behaviourist will be shown to dominate contemporary interpretations, causing a general and swift disregard of his views. After the rise of cognitive science and the influential rejection of behaviourism by Chomsky (1959) and Fodor (1975), any association with behaviourism became damaging.

71

The Philosophical Development of Gilbert Ryle

Ryle’s Contemporaries In what follows I shall briefly enumerate some of the main points of praise and criticism made by reviewers and commentators in the years following on the publication of The Concept of Mind. I have used:

• P.S. Maclellan, ‘Professor Ryle and the Concept of Mind’ (1951); • Albert Hofstadter, ‘Professor Ryle’s Category-Mistake’ (1951); • Margaret MacDonald, ‘Professor Ryle on The Concept of Mind’ (1951); • Hugh R. King, ‘Professor Ryle and the Concept of Mind’ (1951); • J.L. Austin, ‘Intelligent behaviour. A critical review of The Concept of Mind’ (1950); • A.J. Ayer, ‘An honest Ghost?’ (1970); • Stuart Hampshire, ‘Critical review of The Concept of Mind’ (1950); • Bertrand Russell, ‘What is Mind?’ (1958a); • , ‘Professor Ryle’s Logical Behaviourism’ (1951); • John Wisdom, ‘The Concept of Mind’ (1950); • Dickinson Miller, ‘Descartes’ Myth and Professor Ryle’s Fallacy’ (1951).

Besides showing appreciation for Ryle’s project, his commentators also expressed different types of criticism. Ryle was, for example, accused of being unclear about his target, of using vague concepts and distinctions, of remaining at the superficial level of description rather than providing proper explanations, of unsuccessfully trying to reduce the mental to the physical, and of making the counterintuitive claim that there was no fundamental difference between self-knowledge and knowledge of other people. Roughly, we may distinguish two types of criticism: on the one hand the critical remarks of philosophers such as Austin, Wisdom and Hampshire who are in general in agreement with Ryle’s ideas on the nature of philosophy and its relation to the other sciences; on the other hand Hofstadter, Russell and others who criticize Ryle for not taking into account scientific explanations (for example, of human behaviour) when doing philosophy. In what follows these comments will be discussed and put in perspective.

Positive Remarks: Common Sense Philosophy, and the Avoidance of Technical Language All of the commentators think that The Concept of Mind makes a significant contribution to analytic philosophy. For example, Hampshire considered The Concept of Mind to be the ultimate realization of analytic philosophy.

[The Concept of Mind] actually uses the now rapidly changing methods of linguistic analysis to cut the root of a large metaphysical problem. (…) if analytical philosophy really is what it claims to be, it must ultimately issue in plain and pointed prose, with most of the workshop apparatus of technical distinctions left in the background; this, in the tradition of Locke and Berkeley, Mill and Russell, Professor Ryle supremely achieves (…). (Hampshire 1950, 18-19)

And commenting on Ryle’s style, Hampshire writes:

The avoidance of technical jargon, and the disdain of footnote and historical allusion, are evidently parts of a design to restore philosophy to common sense in the manner of the eighteenth century (…), though the derision is generally of a more robust, and sometimes even knock-about, character than was natural to Hume. (Hampshire 1950, 17)

72

The Concept of Mind: Main Aims, Method and Reception

Wisdom admires ‘the power, the and the grace of Ryle’s work. It is an achievement and a part of the progress of philosophy’ (Wisdom 1950, 189); and Austin considers The Concept of Mind to be ‘stimulating, enjoyable and original (…), standing head and shoulders above its contemporaries’ (Austin 1950, 45). Hofstadter praises Ryle for ‘the most brilliant attack on the mentalism in mind-body dualism that has appeared in a long time’ (Hofstadter 1951, 257), claiming that ‘it has already raised quite a dust (…) in the intellectual world (…), and it will undoubtedly be read and discussed as a classic of its kind’ (Ibid.). The book is further complimented for ‘its brilliance and penetration, its wit and clarity, and the merciless incisiveness of much of its criticism of dualism’ (Ibid.). The other reviewers used similar forms of praise.

An Unclear Target One of the main criticisms of Ryle’s commentators is that it is never clear precisely whom Ryle has in mind when he attacks the dogma of the ghost in the machine. He himself claims that he is not attacking the historical Descartes, but what he calls ‘the official theory’, ‘which hails chiefly from Descartes’ (Ryle 1949a, 13) and which ‘continues to distort the continental geography of the subject’ (Ibid., 10). But he is not always clear which specific aspects of this theory he is attacking, and on what grounds. Thus, Hampshire argues that Ryle does not even attack a philosophical theory, but rather the following feature of ordinary language:

(…) That most of its forms of description have been and are being evolved by the constant transfer of terms from application in one kind of context to application in another, and in particular by the transfer of what were originally physical descriptions (…) into psychological descriptions. (Hampshire 1950, 21)

This would mean that Ryle cannot have refuted the dogma merely on the basis of the authority of ordinary language – since ordinary language itself is the source of the dogma: dualism is firmly embedded in it. So something else is needed as well. Hampshire considers this ‘something else’ to be an a priori theory of language, conflicting with our dualistic ordinary language. Because Ryle is unclear about the exact nature of his target, Hampshire finds it difficult to judge the adequacy of his arguments: ‘it is never clear precisely whom he is attacking when he attacks the Ghost and therefore what weapons are appropriate’ (Hampshire 1950, 23). Miller, too, claims that it is not clear what Ryle does and does not believe in The Concept of Mind. ‘He [Ryle] cannot help alluding occasionally to the facts of private consciousness, though he denies them. He never draws for us the urgently required line between what he accepts and what he does not.’ (Miller 1951, 273)

Conceptual Confusions and Vague Distinctions Unfortunately, the target is not the only thing that is unclear. A point of criticism often voiced by reviewers (for instance, MacDonald, King, Austin, Russell and Hampshire) is that Ryle himself – in his crusade against conceptual mistakes – was not able to avoid using confused or vague expressions and concepts himself. This, of course, makes it difficult to unravel his arguments and assess their power and convincingness. Hampshire asserts that confusions arise because Ryle borrows certain distinctions from logic in the strict sense of the word, e.g. the one between categorical and hypothetical, and uses them to characterize distinctions which are only logical in a loose sense, e.g. the one between occurrence and disposition. According to Hampshire, the latter distinction is

73

The Philosophical Development of Gilbert Ryle not strictly logical because it depends on Ryle’s identification of the meaning of a statement with its method of verification: ‘Because overt behaviour often constitutes for most people the best, and, in some conditions of utterance, the sole available evidence for statements about mental activities and states of mind, such statements come to be identified with hypothetical statements about behaviour’ (Hampshire 1950, 32). Hampshire’s objections against such a verificationist theory of meaning will be discussed later in the section on behaviourist interpretations of Ryle. Furthermore, Hampshire attributes to Ryle a naive correspondence theory of language, which led Ryle ‘to write as if there were a real answer, independent of the conventions of a particular language, to such questions as ‘Does the verb “mind” or “try” designate a single, distinct activity or a complex of activities?’ – as though the world consisted of just so many distinguishable Activities (or Facts or States or Things) waiting to be counted and named.’ (Hampshire 1950, 26) Ryle uses this non-logical terminology of ‘different kinds of things designated’ (Hampshire 1950, 25) to characterize, among others, the distinction between occurrence and disposition. As a consequence, his use of occurrence and disposition is confused, which is, according to Hampshire, problematic since what he considers the central thesis of the book – i.e. that statements ordinarily construed as categorical statements about ghostly events should be re-interpreted as hypothetical statements about overt happenings – heavily depends on the distinction between occurrences and dispositions.

It suddenly becomes clear that there must be many verbs (…) which stand for an occurrence as opposed to a disposition in some of the many senses given to this central distinction, namely, that they name incidents which can be significantly dated or clocked, or that certain adverbs are applicable to them, but which in another and predominant sense of the distinction are dispositional verbs, namely, that any statement containing them can be accepted as true only if some testable hypothetical statements about the future are accepted as true. (Hampshire 1950, 26-27)

Consequently, almost any statement can be said to be both categorical and hypothetical, which according to Hampshire makes the distinction between categorical and hypothetical unworkable. He even goes so far as to claim that this destroys Ryle’s whole account of the relation between mind and body. Hofstadter and MacDonald also direct their criticism against the distinction between categorical and hypothetical. More specifically, Hofstadter asserts that Ryle’s line between purely categorical and mongrel-categorical statements is too vague; as a matter of fact he does not believe that, on Ryle’s own definitions, one can give an example of a pure categorical statement that is not at the same time a mongrel-categorical statement.55 According to MacDonald, Ryle seems to be uncertain about what precisely to think of mongrel-categoricals such as ‘being careful’, ‘paying heed’ and ‘minding’. For example, for Ryle to say that someone is ‘paying heed to what he is doing’ is to say categorically that he is doing what he is doing in a certain ‘frame of mind’, and hypothetically what he will

55 It may be doubted whether the criticisms by Hampshire and Hofstadter cut ice, since, although they are certainly right that the line between categorical and hypothetical statements is vague, there may be nothing wrong with such a fuzzy line. Ryle sometimes seems to, but does not need to, claim that both types of statements exist in their pure form. It would also do to regard all statements as more or less hypothetical (and less or more categorical), his aim being to show that we mistakenly regard many statements as more categorical than they in fact are. 74

The Concept of Mind: Main Aims, Method and Reception do (with respect to the object of his paying heed). MacDonald interprets this ‘frame of mind’ as forcing Ryle to concede the existence of something like a private, inner life.56 King points at another terminological confusion found in Ryle, namely the ‘often indiscriminate use of the ambiguous word “intelligent”, which may mean “a directed action” but can also mean “one species of directed action”’ (King 1951, 283), that is, a way of performing such an action, namely intelligently – that is, skilfully – rather than stupidly. Two episodically indistinguishable shots of a skilled marksman and a novice are both directed, or conscious, actions, but one is performed more intelligently, in the sense of skilfully, than the other. Similarly, a bad car driver may still be heeding what he is doing. King takes Ryle’s confusion of the two different uses of ‘intelligent’ to have far-going consequences, accusing him of shifting the reader’s attention from the real problem of distinguishing between intelligent and non-intelligent action – in the sense of directed action and nondirected action – to the problem of distinguishing between different intelligent, that is, different directed, actions. In The Concept of Mind Ryle provides an explanation of ‘intelligent’ by means of dispositions, but this is only relevant to the second problem – not to the first, according to King more pressing, problem. When we are in doubt whether a specific action is directed or not an appeal to dispositions is of no, or at least hardly any, use. Dispositions are not able to tell us whether a given action is done unconsciously or consciously, that is, whether the action is directed or not. Russell, too, thinks that Ryle neglected to explain the difference between mental and nonmental concepts: ‘(…) Professor Ryle never explains, or seems to think it necessary to explain, what is the difference between “brittle” and “intelligent” that makes the latter mental and the former not.’57 (Russell 1958a, 636) Austin mentions Ryle’s (occasional) looseness of expression, e.g. with respect to the word ‘disposition’, as a defect of The Concept of Mind, claiming that this looseness of expression, which also occurs several times in the formulation of critical points in Ryle’s work, contradicts with Ryle’s ‘admirable sensitivity to the nuances of words’ (Austin 1950, 51-52) in most parts of the book. Moreover, it makes it very difficult for the reader to grasp what exactly is Ryle’s purpose. Austin, however, at the same time, thinks this looseness is a direct consequence of the praiseworthy untechnical style of Ryle. He is less positive about Ryle’s suggestion that he uses a single technique or type of argument in his rejection of Cartesian dualism as one big category-mistake. Ryle does not show that he attacks a single clear type of mistake at all, unless it is so very general as

56 Perhaps Macdonald takes ‘frame of mind’ too literally here. Ryle most likely does not mean that someone who is in a certain ‘frame of mind’ is having a private experience, but merely that the person acts – or would act – upon the objects of his paying heed in a certain way. ‘(…) while we are certainly saying something dispositional in applying such a heed concept to a person, we are certainly also saying something episodic. We are saying that he did what he did in a specific frame of mind, and while the specification of the frame of mind requires mention of ways in which he was able, ready or likely to act and react, his acting in that frame of mind was itself a clockable occurrence.’ (Ryle 1949a, 134) 57 But perhaps Russell expects too much. He later claims: ‘I suppose Professor Ryle might agree that the main purpose of this book is to give a new definition of the adjective “mental”.’ (Russell 1958a, 641) But Ryle does not seem to aim at such a definition, he just rejects the existing ones. This does not mean that he would not distinguish between ‘brittle’ and ‘intelligent’. In contrast to ‘brittleness’, ‘intelligence’ is a mental concept. In discussing mental capacities and tendencies Ryle argues: ‘Our concern is with a restricted class of dispositional terms, namely those appropriate only to the characterization of human . Indeed, the class we are concerned with is narrower than that, since we are concerned only with those which are appropriate to the characterization of such stretches of human behaviour as exhibit qualities of intellect and character. We are not, for example, concerned with any mere reflexes which may happen to be peculiar to men, or with any pieces of physiological equipment which happen to be peculiar to human anatomy.’ (Ryle 1949a, 121) It is clear that this description of mental concepts excludes ‘brittleness’.

75

The Philosophical Development of Gilbert Ryle

‘misunderstanding the ways in which words are intended to be, and normally are, used, and so wrongly inferring that they can be made to behave like other, very different, words, and that what they “stand for” (dangerous phrase) is “like” what those other words stand for’ (Austin 1950, 48). This is neither peculiar to the myth, nor its only basis. Ryle in fact uses a number of ‘most effective’ (Austin 1950, 49) methods, showing, for instance, that words which seem to stand for episodes in fact stand for dispositions; he also uses methods more peculiar to the topic concerned, such as showing how a wrong treatment of index-words leads to absurd theories of self-knowledge. For instance, treating ‘I’ and ‘you’ as if they are proper names may easily result in the idea that these are names of a special kind, naming extra, ghostly beings hidden inside the individuals which are indicated by ‘regular’ proper names, such as ‘Charlotte’. The difficulties in interpreting Ryle’s ‘one big category-mistake’ are not only revealed by the critical remarks of his reviewers, but also by the fact that this category- mistake has been interpreted in many different ways. MacDonald, for instance, argued that ‘The fundamental mistake of the Cartesian theory is to construe “Mind” (and, indeed, “Matter” too) as a name and invent a simple substance for it to name.’ (MacDonald 1951, 82) According to Hampshire and Austin the mistake Ryle aimed at was the two-worlds- view, that of the ghost and that of the body, whereas we should have been talking of only one world – namely that of the body.

Description versus Explanation: Philosophy versus Science A criticism often expressed by the more science-minded interpreters is that Ryle does not give explanations, but remains at the superficial level of description, and therefore, as Miller puts it, he:

never comes in contact with the main body of the enemy. (…) we still find that he has not touched the facts that decide the question at issue, the facts that force us to form the conception of consciousness. He does not prove his point, he can not, for he does not talk about this point. He talks about something else that there is no necessity to question, a set of facts most familiar. His fallacy is that of presenting an argument irrelevant to his conclusion. His book is one long ignoratio elenchi. (Miller 1951, 271)

Maclellan, Russell and Hofstadter voice similar sentiments, arguing that The Concept of Mind does not satisfy our hunger for explanation. King believes that Ryle does try to give explanations, but fails because of his appeal to dispositions. Whereas Ryle thinks that there is no physiological difference between the description of a bird as ‘flying south’ and of another as migrating because to an observer they are indistinguishable, King asserts, however, that scientific explanation has now shown that physiologically the two episodes are not identical in character: something more is going on inside a migratory bird than is going on inside one accidentally flying south.58 We may appeal to scientific explanation in order to distinguish between the two

58 King may be right that physiologically the bird flying south and the bird migrating are not identical, but this is not something that Ryle rejects (or affirms). In claiming that they do not differ physiologically (Ryle 1949a, 50), Ryle seems to have meant that when we look at a bird migrating and one flying south we do not see a difference in their physical behaviour. They make the same movements with their wings, etc. It is unlikely that he meant differences on the level of neurons, especially since he identified meaning with use, and our use of ‘the bird is migrating’ and ‘the bird is flying south’ does not depend on differences between the neurological processes that take place in the birds. Ryle merely claims that on the level of descriptions the two birds are identical because to an (ordinary) observer they are indistinguishable, and philosophy works on a different level than neurology and (scientific) physiology do. (A quotation which supports this interpretation can be found on p. 90.) We can doubt 76

The Concept of Mind: Main Aims, Method and Reception birds; no appeal to dispositions is required. Moreover, King asserts that if we assume that the two episodes are identical, our appeal to dispositions would be nothing less than an appeal to the ghost, which Ryle wanted to exorcize.59 Miller holds that whereas Ryle claims that he will disprove the myth of consciousness, he never actually does. He gets stuck at the level of describing our ordinary practice, instead of disproving or explaining anything.

(…) this author’s philosophy is “the most natural thing in the world.” When a man approaches in the street (…) we do not see his vision of us, but also we do not think of it. We can imagine (…) that he is only a body there behaving in the natural way; smiling, nodding, speaking, answering, but wholly without consciousness. But this is precisely thus that we do conceive him on all ordinary occasions and for all ordinary purposes. Other people’s fields of appearance – we do not pause to give ourselves the trouble of imagining them at all. We believe in them at all times potentially, we are ready to think of them, we are on the very verge of doing so, but we rarely go over the verge. (…) Professor Ryle’s book simply erects this ordinary oblivion into a philosophy, gives it a sweeping and imposing formulation. (Miller 1951, 277)

Russell and Hofstadter also criticise Ryle for remaining at the level of common sense talk. Ryle should have given a more scientific explanation to satisfy his goal. By limiting himself to the question ‘How do we use such expressions as “he saw a robin”?’, while dismissing the question ‘How do we see robins?’, he is, as Russell asserts, ‘dismissing important scientific knowledge in favour of verbal trivialities’ (Russell 1958a, 640). Hofstadter’s complaint is that Ryle offers ‘a philosophy of mind in the sense of a logical analysis of talk about mind, and indeed not so much of scientific as of common- sense talk.’ (Hofstadter 1951, 257-258) And if dualism is false, or at least improbable, ‘this will have to be shown to be so on grounds of the sort used to invalidate a scientific theory, not on grounds of the sort the author [Ryle] alleges. If dualism is a mistake, it is not merely a logical one.’ (Hofstadter 1951, 259) He claims that the best way to attack dualism or mentalism is to develop a better explanatory theory. Ryle’s strong preference for description, according to Hofstadter, is caused by his excessive and the fact that he turns logical analysis – which Hofstadter considers nothing more than a useful but subordinate activity in philosophy – into philosophy itself. Rather than appealing to the ideal of scientific explanation, Maclellan refers to our ‘deep-seated psychological need’ (1951, 139) for a causal explanation of the mind. We should not try to dispose of the mind as cause and substitute it for the mind as description of the acts and performances of human beings. What is at the root of the idea of cause is more than a matter of logic: it is a fundamental psychological necessity:

The notion of cause admittedly sometimes turns out to be a phantom and problems often have to be re-stated, but the expectation of cause is something demanded by the intellect. It would rather have an erroneous explanation (as long as it was unaware that it was erroneous) than no explanation at all. (Maclellan 1951, 140) whether Ryle’s use of ‘physiological’ is not misleading here. And even if we take him to mean that there was no physiological (in the sense of neurological) difference between a bird migrating and one flying south, we would indeed have to say that he was wrong here, but this would be but a minor mistake which by no means causes damage to his main ideas. It is not particularly relevant to Ryle’s claims whether or not the two birds differ on the level of neurons. Even nowadays, with the progress of neurology and physiology, when we, in our everyday life, wonder whether a particular bird is migrating or (just) flying south we do not – and do not need to – catch the bird and try to look inside its brain by means of PET scans, MRI, or other methods to map the activity in the bird’s brain. 59 This suggests that King thought that mentality either had to be explained by means of bodily happenings or by means of ghostly happenings.

77

The Philosophical Development of Gilbert Ryle

In other words, Ryle does not meet our needs when he denies that any causal concepts should be involved in expressing the relationship of mind to body. Ryle clearly does not provide us with the kind of explanation many scientists would wish for. He does not occupy himself with physiology, but rather with our everyday life, and gives us ‘explanations’ which are entirely different from those produced by physiologists or other scientists. Scientific explanations are not the kind of explanations philosophy can offer, which is evidently something not all of Ryle’s contemporaries agreed with

Ryle’s Account of Self-knowledge Several of Ryle’s reviewers critically assessed his ideas on self-knowledge. Especially Ryle’s disqualification of introspection and its replacement by retrospection caused a great deal of controversy. The claim that there is no fundamental difference between the things I know about myself and the things I know about others seems counterintuitive and does not do justice to our ordinary language practice. For example, how do we account for the fact that it does not make sense to ask ‘How do you know?’ if someone says ‘I am thinking of riding my bike’? Hampshire asserts that whereas Ryle effectively undermines the thesis of Privileged Access, he fails to account for the difference between my own and other people’s knowledge of what I am thinking. Ryle rightly shows that the possibility of you finding sufficient evidence against one of my reports about my own state of mind is never in principle, though often in practice, excluded. If I claim to have been in pain yesterday morning, you can object: ‘You could not have been in pain, you were anaesthetised’ (Hampshire 1950, 38). But he neglected to pay attention to the fact that I always know both whether I am thinking and what I am thinking of without ever needing inferences, contrary to other people’s knowledge of what I am thinking. If I say “I am thinking of climbing Helvellyn”, it is senseless for you to ask “how do you know?”. In asserting: ‘The sorts of things that I can find out about myself are the same as the sorts of things that I can find out about other people, and the methods are much the same’ (Ryle 1949a, 149), and: ‘In fact they [people] are relatively tractable and relatively easy to understand’ (Ryle 1949a, 110), Ryle ‘proves too much’ (Hampshire 1950, 33), ignoring the conflict between what people report about their own states of mind and what we observe. In doing so, he neglects the important function of the word ‘really’ in combination with mental concepts, e.g. in the utterance ‘you do not really enjoy gardening’ (Hampshire 1950, 39) – for example because when it rains you never work in the garden – and ‘I think this is my real attitude, motive or feeling, but I am not certain’ (Ibid.). The fact that Ryle does not discuss the conflict between people’s reports about their own states of mind and what we observe is, according to Hampshire, a serious lack in his treatment of the mind. He argues that neither Privileged Access (only I can know about my own states of mind) nor Ryle’s Open Access can be true. We can always think of situations in which further evidence would make the subject withdraw his claim, for example in the case of ‘You could not have been in pain since you were anaesthetised’; but, on the other hand, it is also not the case that my justification of ‘I am in pain’, expressed in the present tense, and your justification of me being in pain are of the same type. Similar to Hampshire’s criticism is that of John Wisdom, who writes that Ryle does not help us to understand ‘the facts which lead people to say that a person has a way of knowing how he feels which no one else has, has a right to say what he does about how he feels which no one else ever had or ever will have.’ (Wisdom 1950, 195) In his article

78

The Concept of Mind: Main Aims, Method and Reception

‘The Concept of Mind’ (1950), Wisdom attempts to continue Ryle’s work on this point by trying to determine why it is the case, and what it means to say, that ‘A never has that reason for a statement about how things seem to B which B has.’ (Wisdom 1950, 200)60 Ayer, too, thinks that there are still reasons to believe that a knowledge-claim which is based on personally having the experience has authority over and above knowledge-claims based on observation, even though Ryle may have been right in thinking that there is only a difference in degree between knowledge of one’s own mental processes and knowledge of another’s mental processes. Sceptical arguments concerning ‘external knowledge’ will still have to be met. According to Ayer, Ryle has totally ignored this problem. It is more than a trivial point of logic that another person cannot feel my twinges, just as he cannot smile my smile. King’s criticism is that self-knowledge by retrospection neither touches upon introspection nor upon the problem of the privacy of consciousness, and that retrospection is not even possible without introspection:

If we were not conscious of what we were doing when we did it, retrospection upon how as a matter of fact we did act would be impossible. Retrospection is dependent on , and memory is dependent largely upon the fact that we make conscious observations and are aware of our performances. (King 1951, 292)

Moreover, when it is ambiguous whether a certain action was an intelligent one or not, only the performer can tell us for sure whether it was. We must not forget that our overt actions are not all there is to our mind; they are nothing but a small fragment of the workings of our minds. King rejects – what he calls the ‘modern’ – assumption that we can clarify the nature of observation and sensation without something like the ghost in the machine. The problems for philosophy have just begun by eliminating this ghost, which ‘has long been a convenient bearer of all sorts of phenomena which our natural , or physiological and physical theories, could not hope to explain’ (King 1951, 294). Self-knowledge is not the only subject Ryle is considered to have mistreated by many of his interpreters. His ideas on imagination, voluntary action, motive and belief were also criticised, particularly his positive claims.61 But rather than dealing with these issues I shall focus on behaviouristic interpretations of The Concept of Mind, which were – at least partly – responsible for the later disregard of Ryle’s work.

Behaviourism Many of Ryle’s contemporaries characterized his philosophy as behaviouristic. Some of these interpretations suggest a strong, reductionist, type of behaviourism, others only a weak, non-reductionist, one; some take into account Ryle’s explicit denial of aiming at a reduction of mind to matter, others do not. In this section I present several behaviouristic interpretations as well as those which go against this dominant interpretation. Hampshire considers Ryle’s book to be written in transition from a two-world conception of logic and philosophical method to a one-world view. He criticizes Ryle for not realising that the logical arguments he directs against the form and generality of two- world philosophical statements just as well apply to one-world statements – it is a matter of

60 This does not mean that Wisdom does not agree with Ryle about whether we can sometimes, or usually, have knowledge about other people’s minds; but he considers this a question of fact and not of philosophy. 61 Cf. Hacker 1996 (pp. 170-171).

79

The Philosophical Development of Gilbert Ryle different sides of the same coin. That Hampshire does have a point may be seen from Ryle’s letter to Thomas W. Bestor, which was written many years later in 1976:

The target I aimed at was the big and famous crux, the Cartesian Split. Only incidentally and half-heartedly did I shoot at what I now see to be the Reductionist backside of the same target. I suspect that this half-heartedness was partly due to strategic one-eyed-ness, but partly to the patent pettiness of Watson, Neurath, Hebb, Quine, and Skinner [all of whose] was sketchy and mean compared with the Cartesian’s splendid ampleness. (Ryle 1976a, 241)

Thus, Ryle, too, considers Cartesian dualism and reductionism to be different sides of the same coin, and admits in retrospect that in The Concept of Mind he had not paid enough attention to the reductionist side. But Ryle did not in fact try to reduce the two-world view to a one-world view. Hampshire may have been led in thinking so by his attribution of a naive correspondence theory of language to Ryle (see above.). But as I have shown in the previous chapter, Ryle had abandoned this type of theory already in the late 30s. Because Ryle constantly tries to show that statements involving mental concepts are testable propositions, Hampshire argues that Ryle’s project of laying bare the logic of our mental concepts sets out to determine the standard tests appropriate to these concepts. Because overt behaviour often gives people the best, or even the sole, available evidence for statements about the mental activities and conditions, such statements then become identified with hypothetical statements about behaviour. Hampshire’s ascription of this verificationist position to Ryle provides him with another reason to think that he held a behaviouristic one-world view. For Hampshire this identification of meaning and verification is problematic because it does not allow for exceptions. And the possibility of exceptions plays an important role in our everyday life.

(…) we usually test a child’s concentration by testing his performance, but we do not identify the performance and the concentration, precisely because although the first may often (not always or necessarily) be properly accepted as sufficient evidence for the other, it is not a necessary condition in the sense of being in all circumstances the only kind of relevant evidence available, either to the child himself or to the teacher. (Hampshire 1950, 32)

Hampshire’s idea that the meaning of a statement is not to be identified with its verification is explicitly supported by Ryle later in 1951.62 And it may even be argued that Ryle held this position already in The Concept of Mind. But it has to be admitted that Ryle was at least unclear about this, and some passages suggest such an identification. But Hampshire’s critique went further: Ryle would not properly have carried out his behaviouristic project. Ryle’s position is ambiguous since he has a behaviouristic aim while at the same time pointing out that re-interpretations of mental statements into hypothetical ones about overt happenings are not always possible, e.g. in the case of emotions, thrills and imaginings. Ryle sometimes even suggests that it would be a logical mistake to do so. Austin, too, maintains that Ryle’s aim is to present a doctrine of ‘one world’, namely a bodily world. He criticizes Ryle for not eliminating ‘the ghost’ – e.g. he did not succeed in disproving that sensation is not a ‘private’ or ‘occult’ experience. According to Austin, Ryle’s conviction that he must show that there exists only one world, combined with his belief that he employs one single method – ‘a single Excalibur, clothed in the name of Logic’ (Austin 1950, 49) – caused a delusion in his thinking.

62 See Chapter 2, p. 35. 80

The Concept of Mind: Main Aims, Method and Reception

Miller and Russell also consider Ryle’s aim in The Concept of Mind to be behaviouristic. The latter focuses on Ryle’s denial of introspection as a source of knowledge, which connects him to behaviourism, and claims that Ryle only disagrees with behaviourism in so far as it results in mechanistic explanations. Miller heavily criticizes Ryle’s behaviourism, which he considers to result in the denial that we exist.

You cannot for a moment identify yourself with a body without consciousness. You are your conscious self. What Professor Ryle is doing (without knowing it) is denying that we exist. (…) is the tenet that I do not exist but other things do. This is the culmination of human modesty. It surpasses even “Excuse me for existing”. Professor Ryle is substantially a behaviorist. (Miller 1951, 272)

Wisdom, too, characterizes Ryle as a behaviourist, although he does notice that Ryle does not deny that there are mental processes, e.g. doing long division and making a joke, and that he does not seek to reduce mental states and processes to physical states and processes. He attributes to Ryle a weaker type of behaviourism, namely that there are not mental processes over and above physical processes (Wisdom 1950, 19).

He [Ryle] does, however, often use words which suggest that he does wish to say that in the sense in which the average man is reducible to individual men the mind is reducible to the body, that in the sense in which the passing of a division is reducible to the passing of men and guns, mental processes are reducible to bodily processes, that consciousness is to its manifestations as electricity to its manifestations.’ (Wisdom 1950, 191)

That many of his reviewers characterized Ryle as a behaviourist does not mean that they disagreed with Ryle on this point. Some liked his views despite his behaviouristic tendencies, and others, such as Morris Weitz, appreciated the form of behaviourism which they thought Ryle advocates: ‘his basic impresses me as essentially correct and profoundly first-rate’ (Weitz 1951, 301) There were also opposing views: Ayer and MacDonald denied that Ryle held a behaviouristic position. The latter claimed that Ryle neither intended nor needed to accept such a metaphysical conclusion.

What he does, I think, admit as the ultimate reference of discourse is a world, “the ordinary world”, but if to do this is to be a metaphysical monist then so is every plain man, which is absurd. The world of ordinary life is not one, and might have been two or more, but simply the world. But we talk about this world in many idioms. (MacDonald 1951, 85)

Ayer, too, denies that Ryle is a behaviourist. Ryle does not attempt to completely reduce people’s inner lives to behaviour (or behavioural dispositions), since he often refers to inner occurrences, such as feelings of despair and pride, internal monologues, tunes running in one’s head, and mental arithmetic. Ayer argues that Ryle does admit a ghost, but it is a friendly one in that it does not require the stage of a private theatre – there is no observing of privates states or processes by the person to whom they are private, nor is it logically impossible for them to be observed by other people. ‘(…) He nowhere suggests that his silent soliloquies are to be equated with physiological states or processes: nor does he make any attempt to translate them into behavioural dispositions.’ (Ayer 1970, 65) Ayer thinks that a thesis more consistently held throughout the book is that ‘whether or not the programme of logical behaviourism can be carried out in its entirety, it does give a correct account of a great deal of what is ordinarily classified as talk about the mind’. (Ayer 1970, 67) Silent thought can occur but is not a necessary condition for intelligent action – its role

81

The Philosophical Development of Gilbert Ryle is minimized. Ayer believes that Ryle takes logical behaviourism ‘as far as it can legitimately go’ (Ayer 1970, 74), but admits that the stronger (in fact, too strong) thesis of behaviourism is suggested in some parts of The Concept of Mind. The discussion about Ryle’s alleged behaviourism was often blurred because the reviewers often failed to distinguish between different types of behaviourism and neglected to make clear what type they used. Or when they did, they did not make the effort to discuss other types. In the next section I hope to provide a more complete account of Ryle’s position, taking into account different types of behaviourism.

The Concept of Mind as a Non-Behaviourist Attack on Dualism

As we have seen, many of Ryle’s contemporaries interpreted him as a behaviourist in one way or another. This reception can partly be explained by the scientific Zeitgeist of the 1940s and 50s. The ideal of objectivity rendered behaviourism an attractive position, since it rejected unobservable, internal causes, as well as the intensionality of cognitive constructs. In modern handbooks Ryle is usually also mentioned as a behaviourist. At the time he wrote The Concept of Mind Ryle expected that he would be interpreted as a behaviourist, claiming that ‘the general trend of this book will undoubtedly, and harmlessly, be stigmatized as “behaviourist”’ (Ryle 1949a, 308). He could not have been more wrong about the harm this label would cause. In the 1960s the cognitive revolution killed behaviourism, which from then onwards carried the stigma of being an outdated and uninteresting view.63 After showing several examples of characterizations of Ryle as a behaviourist in modern handbooks, I shall discuss which type of behaviourism, if any, can be attributed to him.

Ryle’s Behaviourism According to Modern Handbooks In modern handbooks in the philosophy of mind, as well as introductions or companions, Ryle, if discussed at all, is almost without exception characterized as a behaviourist (although the type of behaviourism differs), often resulting in him being put aside as outdated. In E. J. Lowe’s An Introduction to the Philosophy of Mind (2000) Ryle is mentioned only briefly in three footnotes. And in Philosophy of Mind (1996) mentions The Concept of Mind as ‘the influential classic work representing behaviorism as a philosophical doctrine’ (Kim 1996, 45) – the only reference to Ryle. In of Mind (1997) Georges Rey characterizes Ryle’s view as analytical behaviourism:

Ryle and others did propose a more positive approach, which emerged as the doctrine of Analytical Behaviorism (“Abism”). The general idea is that claims about the mind should be understood as equivalent to various sorts of dispositional or conditional claims about how an agent would behave if she were in such and such circumstances, on the model of dispositional elsewhere: “Salt is soluble” presumably means something like, “If salt is put into water in certain

63 William O’Donohue and Richard Kitchener argue that behaviourism is by no means dead nowadays. However, this primarily has to do with the fact that they call Ryle and Wittgenstein behaviourists too, and thus use behaviourism in a broad sense. (O’Donohue and Kitchener 1999, 7) 82

The Concept of Mind: Main Aims, Method and Reception

normal conditions, then it dissolves”; “Glass is fragile” something like “If struck in normal circumstances, it breaks”. (Rey 1997, 151)

In Lowe’s An introduction to the philosophy of mind (2000) it is mentioned in a footnote that in The Concept of Mind ‘a sophisticated version of logical behaviourism is developed by Ryle’ (Lowe 2000, 42). In his introductory Minds and Bodies (2000) Robert Wilkinson claims that ‘philosophical (or analytical) behaviourism is most thoroughly set out in The Concept of Mind (1949)’ (Wilkinson 2000, 50). From an earlier date is Armstrong’s contribution to William G. Lycan’s Mind and . A Reader (1990), in which he asserts that ‘in The Concept of Mind (1949), Gilbert Ryle, although he denied that he was a Behaviorist, seemed to be upholding an account of man and his mind that was extremely close to Behaviorism’ (Armstrong 1990, 38). In the Blackwell Guide to Philosophy of Mind (2003) six – rather general – references to Ryle can be found. The most interesting remark is of the hand of William G. Lycan, who claims that in 1956 Place was the first to formulate a middle way between dualism and behaviourism: ‘So matters stood in stalemate between dualists, behaviorists, and doubters, until the late 1950s, when U. T. Place (1956) and J. J. C. Smart (1959) proposed a middle way, a conciliatory compromise solution.’ (Lycan 2003, 50) According to Lycan, before 1956 one was either a dualist or a behaviourist – there was no way of rejecting both.64 Tim Crane’s The Mechanical Mind (1995) does not contain any reference to Ryle, and The Oxford Companion to the Mind (2004) no more than two – one that he argued against those who ‘dubbed the mind “the Ghost in the Machine”’ (2004, 285); the other is a general account of Ryle’s most well-known ideas with a total length of one page. And what about handbooks concerning the ? Of course Ryle wrote about the mind and about mental concepts, but was he not principally a philosopher of language? After all, The Concept of Mind had a meta-philosophical purpose, and was about concepts of mind, not about the mind itself. Ryle’s aim was to expose a defective theory by showing the true logical geography of mental concepts, not by offering an alternative theory of mind. However, in handbooks about philosophy of language Ryle does not frequently occur either, e.g. in A Companion to the Philosophy of Language (1997), his name is not mentioned once (in comparison: Wittgenstein is referred to over a hundred times; Russell counts eighteen references). Now, it is important to see why there is still such a strong urge to interpret Ryle’s work behaviouristically, and – accordingly – to classify him as a philosopher of mind rather than a philosopher of language. There are obviously passages in The Concept of Mind which invite similar interpretations.

Suggestions of Behaviourism in The Concept of Mind References to behaviour and suggestions of behaviourism first occur in ‘Conscience and Moral Convictions’ (Ryle 1940). In this article, Ryle argues that public tests of whether a person really knows or is really convinced are behavioural, suggesting that the word ‘conscience’ is used ‘for those moral convictions which issue not in verdicts but in behaving or trying to behave’ (1940, 188). ‘Conscience is not something other than, prior to or posterior to moral convictions; it is having those convictions in an operative degree, i.e. being disposed to behave accordingly.’ (Ibid., 189) And in ‘Knowing how and knowing

64 I hope to show in the remainder of this section that in The Concept of Mind Ryle already found a way to reject both dualism and behaviourism – not by providing an alternative theory of mind, but by rejecting what was at the basis of both ‘isms’, i.e. a denotational theory of meaning.

83

The Philosophical Development of Gilbert Ryle that’ (Ryle 1946a) Ryle argues that knowledge does not refer to a thing or process – as physical verbs such as ‘to walk’ do – but knowledge is an ability, a disposition to behave such and such. The main source of Ryle’s reputation as a behaviourist is of course The Concept of Mind (1949a). First of all, from his claim that he wanted to get rid of the ghost in the machine it was easy to imply that he just wanted to get rid of the ghost, while retaining the machine, whereas what he really wanted was to get rid of the whole mind-body dichotomy. His description of the ghost in the machine in terms of a category-mistake did not help to avoid misinterpretation: ‘It represents the acts of mental life as if they belonged to one logical type or category (or range of types or categories), when they actually belong to another’ (Ryle 1949a, 17). It is tempting and easy – but wrong – to interpret this as if Ryle means to say that the facts of mental life, rather than belonging to the category of ghostly, immaterial being, actually belong to the category of the material world. Moreover, as we have seen, Ryle does not point at one single category-mistake but rather presents a whole cluster of category-mistakes, some of which are underlying others, though never so general that they are underlying our whole geography of mental concepts. Only the category- mistake of denotationalism seems to cover the whole picture. Other category-mistakes, which operate in a more limited area, only show more specific mistakes and do not, for example, rule out behaviourism; sometimes they even seem to suppose it. Ryle’s frequent appeals to behaviour in his analysis of category-mistakes strengthened the suggestion of behaviourism. He considered the methods of verification of an expression to be important indications of its meaning, and this verification generally involved observable behaviour. For example, the cleverness of a clown is not a hidden performance taking place inside the clown’s head. What his spectators admire is not a secret, hidden performance, but the visible performance. Another example is Ryle’s claim that knowledge statements are dispositional statements: ‘to say that this sleeper knows French, is to say that if, for example, he is ever addressed in French, or shown any French newspaper, he responds pertinently in French, acts appropriately or translates it correctly into his own tongue’ (Ryle 1949a, 119). It may be tempting to infer from this that there is nothing but behaviour, or at least that our mental concepts can be reduced to ones about (bodily) behaviour. It is not difficult to see how verification and behaviourism are connected. When I want to verify that someone has a dream, is in pain or is sad, the only thing I can do is to collect testimony and observe external symptoms. What could this testimony and these external symptoms be but this person’s overt behaviour? Ryle, however, did not identify verification and meaning, but only saw verification as providing indications for the meaning of expressions. This would mean that Ryle’s verificationist arguments are nothing but therapeutic, less rigorous, arguments. The following type of logical behaviourism lies in ambush when an identification of meaning and verification is attributed to Ryle, and his verificationist arguments are interpreted as logically rigorous instead of therapeutic:

Logical Behaviorism: every mentalistic term M refers to (means) a set of behaviors B and/or behavioral dispositions BD, which are the verification basis (evidence) for the application of M. (Kitcher 1999, 400)

Hacker (who did not himself consider Ryle a behaviourist) claims that the main reason why Ryle is so frequently interpreted as a behaviourist is that his cause – ‘Exorcize the ghost in the machine’ (Hacker 1996, 171) – was misleading in that it suggested the view ‘that all mental predicates signify behaviour, behavioural dispositions, tendencies or inclinations’

84

The Concept of Mind: Main Aims, Method and Reception

(Hacker 1996, 172). Moreover, he calls the category-mistake ‘double-edged’ (Hacker 1996, 172) and therefore very much susceptible to misinterpretation. Ryle’s idea that the Cartesian myth was one big category mistake ‘could all too easy be read as arguing that the mistake was to construe psychological predications as categorical ascriptions to one substance, namely the body – which was not what he meant at all’ (Hacker 1996, 172). What further helped to establish the view that Ryle was a behaviourist was that ‘he arguably overworked the dispositionality of many psychological expressions (e.g. belief, motive)’ (Hacker 1996, 171) and failed ‘to give due attention to the detailed analysis of the concept of disposition itself’ (Ibid.). The misconception of Ryle’s theory of meaning as denotational or of his theory of language as a naive correspondence theory of language has also contributed to behaviouristic interpretations. In Chapter 2, I showed that Ryle had in fact already abandoned the idea of a denotational theory of meaning before 1938 and that from then onwards he worked on a heterogeneous theory of use. However, as Hampshire pointed out in his review of The Concept of Mind, Ryle’s terminology is (at least sometimes) still that of a naive correspondence theory of language (or of denotationalism). Rather than to jump to the conclusion that Ryle must therefore still have held this view, I would argue that in this respect his use of language was sometimes misleading, and that it was – at least partly – this confused use of certain concepts which led Hampshire and others to their behaviouristic interpretations.

Ryle’s Rejection of a Behaviouristic Interpretation Already in The Concept of Mind itself we find clear indications of Ryle’s non- behaviouristical stance:

If my argument is successful there will follow some interesting consequences. First, the hallowed contrast between Mind and Matter will be dissipated, but dissipated not by the equally hallowed absorptions of Mind by Matter or of Matter by Mind, but in a quite different way. For the seeming contrast of the two will be shown to be as illegitimate as would be the contrast of ‘she came home in a flood of tears’ and ‘she came home in a sedan-chair’. The belief that there is a polar opposition between Mind and Matter is the belief that they are terms of the same logical type. (Ryle 1949a, 23)

This passage is incompatible with the claim that Ryle’s project was to reduce statements about the mental to statements about behaviour, or to identify the two: if both kinds of statement belong to different categories such a move would clearly be absurd. Ryle’s aim was to differentiate statements about the mind from those about matter, rather than to reduce the one to the other. Such a reduction was precisely what he was arguing against.65 As Hacker argues:

Ryle did not deny that nouns such as “pain”, “twinge”, or “tickle” signify sensations, or affirm that they are reducible to dispositions to behave. He did not, save in incautious moments, deny that words for various kinds of “agitation” (e.g. the shock of surprise, throb of compassion, thrill of anticipation) signify episodes which have phenomenological features, or claim that they connote merely dispositions and susceptibilities. He did not deny that people have afterimages, visualize things and picture things to themselves or have tunes running through their heads, but only that there are mental pictures or mental tunes which only their subject can see or hear. And he asserted that some of our thinking is conducted in internal monologue, which we can keep to ourselves. (Hacker 1996, 171)

65 See also Park 1994 for this argument.

85

The Philosophical Development of Gilbert Ryle

In fact, from the introduction to The Concept of Mind it is clear that Ryle talked about the meaning of mental concepts; he did not intend to increase our factual knowledge about minds, but to ‘rectify the logical geography of the knowledge which we already possess’ (Ryle 1949a, 9). In other words: his aim was conceptual rather than factual, which in itself shows the untenability of the view that Ryle is an ontological behaviourist, since he was not interested in making ontological claims. However, in order to discard a logical behaviouristic interpretation as well, we should keep in mind the meta-philosophical purpose of the book, in combination with his more explicit ideas of philosophy and philosophical method in his papers written from the 1930s onwards.

Two Interpretations: Shelley M. Park and Rowland Stout I shall now discuss what I consider to be the most charitable interpretations of Ryle’s alleged behaviourism, since these take seriously both the suggestions of behaviourism in The Concept of Mind and the clues to the opposite. I start with Shelley Park’s interpretation, which denies that Ryle is a behaviourist. Rowland Stout, on the other hand, denies the presence of certain types of behaviourism in The Concept of Mind, but does not reject a behaviouristic interpretation altogether.66 Park denies that Ryle was an ontological behaviourist, that is, the position of ontologically reducing minds to physical behaviour or of an ontological elimination of minds, since this did not seem to fit with his meta-philosophical aim and the fact that he considered philosophy to be a conceptual rather than a factual inquiry. Ryle could not have been a logical behaviourist either – logical behaviourism being defined as, roughly, the position that ‘all talk about mental events is translatable into talk about actual or potential overt behaviour’ (Park 1994, 276), or that of ‘attempting to “identify” the meaning of mental-conduct statements with their verification’ (Ibid.). She mentions other versions of logical behaviourism as well, but limits her discussion to the version just mentioned. According to Park, Ryle’s position in The Concept of Mind would be better characterized as ‘linguistic Antirealism’, that is:

The meaning of sentences about the mind is not a function of the denotation of their constituent mental-conduct terms, and sentences about the mind are not to be judged as true or false (if true

66 Ullin T. Place, who wrote a chapter on Ryle’s behaviourism in the Handbook of Behaviorism (Place 1999) should be briefly mentioned as well. He considers Rylean behaviourism to be the weak view that the primary function of our ordinary concepts is public talk and talk about the behaviour of other people, rather than talk about our own private . I do not think that Park or Stout would reject that this weak form of non-reductive behaviourism can be attributed to Ryle, although Park would argue against Place’s suggestion that Ryle at the time still held a denotational theory of meaning. Place namely claims Ryle’s behaviourism to be a deduction from the following two principles: (1) ‘The primary function of language is to enable human beings to communicate with one another across the space between them in such a way as to provide the listener with instructions and information relating to objects, events, and states of affairs in the physical environment common to both parties to which the listener would otherwise have no access.’ (Place 1999, 363-4) (2) ‘The lexical words (names) that comprise a language user’s basic vocabulary acquire their meaning by a process somewhat misleadingly referred to as “ostensive definition”, whereby the child learns the meaning of a word by being shown instances to which it applies.’ (Ibid.) It seems that Place suggests that this position entails some kind of denotational theory of meaning – after all, an ostensive definition is a way of defining a term by pointing to its referent. Or at least, he does not clearly explain that Ryle did not ascribe to a denotational theory of meaning. Of course Place is right that Ryle believes that language is in a way learned by means of instances to which it applies, but ‘ostensive definition’ seems to be too simple and homogeneous for characterizing this position. Language is learned by showing how it is used, but there is no such uniform relation between lexical words and objects in reality as Place’s account suggests. Ryle’s point is precisely that the meaning of sentences about the mind is their – heterogeneous – use, rather than their reference to objects, processes, etc. in reality. 86

The Concept of Mind: Main Aims, Method and Reception

means, as above stipulated, true by virtue of denotation) because their constituent terms are not meant to denote: and thus The question concerning whether mental-conduct terms denote material or immaterial entities is a pseudoquestion. (Park 1994, 282)

Therefore, Ryle refutes dualism and behaviourism alike, since both depend on Linguistic Realism, which is, according to Park’s definition, ‘the view that the meaning of sentences about the mind is a function of the denotation of their constituent mental-conduct terms, and such sentences are to be judged as true or false depending upon whether or not those terms successfully denote’ (Park 1994, 282). Logical behaviourism is ‘seeking to translate statements which are ostensibly about (referentially) minds into statements which are really about (referentially) observable behaviors’ (Park 1994, 278-79), but this was not Ryle’s aim. At the time of The Concept of Mind Ryle did not hold a denotational theory of meaning. The meaning of mental concepts and statements is not their denotation: ‘we should not be misled into thinking that these terms denote, name, refer to, or stand for states, episodes, happenings or events, acts, incidents, occurrences, processes, performances, operations or things’ (Park 1994, 278). It is not the case that the meaning of ‘person x is intelligent’ is determined by the denotation of ‘the intelligence of person x’ in the sense in which ‘Fido’ denotes Fido – it does not denote an object in reality (either material or immaterial). This is simply not how sentences about the mind mean according to the Ryle of Park’s interpretation.67 This interpretation holds that a denotational theory of meaning has influenced discussions about mind and body and is responsible for the most important claims of Cartesian dualism, e.g. that theorizing is the primary, and intrinsically private, silent and internal, operation of the mind; and that intelligent performances always consist of doing two things: one of considering the appropriate propositions and another of putting these into practice. I believe Park to be right, since looking for the thing or process in reality behind the mental concept seems to have made us accept these dualist claims. We treat mental concepts just like the name ‘Fido’ and believe that they mean in the same way, or the same sort of thing as ‘Fido’, which results in our desperately trying to find Fido’s in order for our mental concepts to have meanings. Because we are unable to find the Fido’s we are looking for, we assume that there are two worlds: the material world of the Fido’s, which mean ‘Fido’ by denotation, and the mental world in which mental concepts also mean by denotation – but denotation of mental objects rather than of material ones. If we accept that statements about the mind do not get their meaning from denoting in the way ‘Fido’ denotes Fido, we are not tempted to accept Cartesian dualism in the first place. The mistake made by the interpreters who attribute a behaviouristic position to Ryle is that they (implicitly) consider him to be a ‘linguistic Realist’ (see above) rather than a ‘linguistic Antirealist’. But according to Park, Ryle’s main objection to Cartesian dualism is precisely that it entails linguistic realism, which ‘renders mind-language unlearnable’ (Park 1994, 285) in that ‘we could never be assured that our comments about the mental conducts of others “have any vestige of truth”’ (Ibid.). But we do know how to make these comments, and it is precisely because we know how to make them that theories of meaning about mental concepts and the reality they describe have been invented in the first place: ‘Hence, learning the meaning of mental-conduct sentences must be possible and, contrary to the claims of linguistic Realism, it must consist simply in learning the appropriate and inappropriate conditions of their use’ (Ibid.), not in learning their denotation. In this way,

67 In fact, it is not how some sentences about material things mean either: the brittleness of glass does not denote in this way either.

87

The Philosophical Development of Gilbert Ryle

Park is able to explain Ryle’s identification of meaning with use, as well as his frequent appeals to ordinary language.

Ryle wants to show us that we are operating, in the philosophy of mind, with a wrongheaded theory of meaning. And one of the ways to convince us of this is to show that the consequences of that theory of meaning are things that “no one ever does or would say” (CM 161). We are talking nonsense about mental-conduct concepts, if our philosophical theory of those concepts undermines our ability to talk sense with them. (Park 1994, 289)

Park’s paper gives a very good analysis of The Concept of Mind, since it can account for its meta-philosophical aim as well as for the fact that Ryle believes philosophy to be a conceptual rather than a factual enquiry. Park considers the Cartesian myth to make the mistake of ‘treating a philosophical problem – in this case the mind-body problem – as a matter for factual investigation’ (Park 1994, 186) and regards The Concept of Mind as ‘a case study in the great difference it makes to apply (consciously or unconsciously) one, rather than another, theory of meaning to philosophical statements’ (Ibid.). Thus, she convincingly argues that Ryle wanted to show that the ‘Fido’-Fido theory of meaning – the theory that statements are true or false ‘depending upon whether the entities and occurrences they refer to are as they describe them’ (Park 1994, 279) – is an incorrect model of interpretation for statements about the mind. The meaning of statements about the mind is their use, not their reference to things, occurrences, entities, etc. One minor point of criticism may be that Park’s interpretation is a bit too Wittgensteinian. Although she is strictly correct in claiming that ‘sentences about the mind are not to be judged as true or false (if true means, as above stipulated, true by virtue of denotation68) because their constituent terms are not meant to denote’ (Park 1994, 282), she suggests that Ryle considers ‘true’ and ‘false’ to be inapplicable in philosophy. Whereas this view can indeed be linked to Wittgenstein, Ryle always maintained that ‘true’ and ‘false’ are applicable to philosophical statements as well. And he would have found no reason to exclude them from the realm of mental statements. We just have to keep in mind that ‘true’ and ‘false’ in philosophy are not used in the way in which they are used in science. Park’s coherent and plausible interpretation may also be too charitable, for Ryle’s views were not so clear and coherent, as he himself acknowledged many years later in 1976 in a letter to Thomas W. Bestor, when he admitted that he had neglected to reject the reductionist backside of Cartesian dualism in a systematic way. Not only this ‘one-eyed- ness’ (Ryle 1976a, 241), but also Ryle’s tendency to overwork certain distinctions was responsible for misinterpretations. As Ryle himself confessed to Bestor: ‘having started my anti-Cartesianism and anti-Lockeanism over the special concepts of Knowing and Believing, I was over-captivated by the luminousness of the contrast between Having (or Owning or Being equipped to …) and, for example, Doing Undergoing, Trying, Succeeding, Failing, etc.’ (Ryle 1976a, 235). The fact that Ryle was mistakenly considered to hold a denotational theory of meaning was also partly responsible for the persistent interpretation of Ryle as a behaviourist. As noticed above, other things played a role as well: his numerous appeals to overt behaviour; his focus on Cartesian dualism; his rather loose use of ‘category-mistake’ and other concepts; and, above all, the unclarity of his real target.

68 Italics are mine. 88

The Concept of Mind: Main Aims, Method and Reception

In a recent paper another philosopher, Rowland Stout, argues that there are sufficient reasons to call Ryle a behaviourist – at least as long as we realize that he was not a reductionist behaviourist. He attributes to Ryle the following claims:

1) ‘That our entitlement to describe an individual’s state of mind depends on being entitled to make certain inferences about their behaviour’ (Stout 2003, 38). 2) ‘The interpretation of behaviour rather than introspection was the primary method of access to the mind.’ (Ibid.) 3) ‘There is nothing more to one’s assertion that someone is in a certain state of mind than that one is entitled to make certain inferences about their behaviour.’69 (Ibid.)

According to Stout, the problem underlying both and dualism, which Ryle tackled in The Concept of Mind, did not start with the assumption of a special realm of the mental, but rather one step further back: with the assumption that there is a special physical realm, that is a non-mental realm which is fully describable without using mental language. This assumption of a purely physical realm only leaves open three options for mental language, namely: dualism, reductive materialism – that mental language can be reduced to non-mental language – and the position ‘that mental language describes nothing at all’70 (Stout 2003, 39). But Ryle would have opted for a fourth position, namely ‘that such talk [mental language – CV] is talk of the ordinary world and is an eliminable part of our talk of the material world.’ (Stout 2003, 39) This fourth position was only possible because Ryle rejected the existence of a physical world which is fully describable by non-mental language. There is only one world and this is neither a ‘mental world’ nor a ‘physical world’, but it is our ordinary world ‘of people’s doings, sayings, imaginings, etc.’ (Stout 2003, 39). Both mental and non-mental language describes this world – that is, our ordinary world.71 Support for such an interpretation comes from the following quotation (which Stout, surprisingly, does not give):

But when a person talks sense aloud, ties knots, feints or sculpts, the actions which we witness are themselves the things which he is intelligently doing, though the concepts in terms of which the physicist or physiologist would describe his actions do not exhaust those which would be used by his pupils or his teachers in appraising their logic, style or technique. He is bodily active and he is mentally active, but he is not synchronously active in two different ‘places’, or with two different ‘engines’. There is the one activity, but it is one susceptible of and requiring more than one kind of explanatory description. Somewhat as there is no aerodynamical or physiological difference between the description of one bird as ‘flying south’ and of another as ‘migrating’, though there is a big biological difference between these descriptions, so there need be no physical or physiological differences between one man as gabbling and another as talking sense, though the rhetorical and logical differences are enormous. (Ryle 1949a, 50)

And another quotation supports this interpretation as well:

69 At first sight this may seem a strong claim that does suggest some type of reductive behaviourism. However, we will see that Stout’s use of behaviour is broader than just bodily behaviour. 70 Stout attributes this third position to Park. However, this seems to be cutting too many corners. Park merely claims that mental language does not describe in the way ‘Fido’ describes Fido (the dog), rather than saying anything about weaker (more commonsense) versions of ‘describing’. 71 This interpretation is similar to the one given by MacDonald 1951 (p. 85). It is also the position Thomas W. Bestor attributes to Ryle (Bestor 1979).

89

The Philosophical Development of Gilbert Ryle

His [a person’s] life is not a double series of events taking place in two different kinds of stuff; it is one concatenation of events, the differences between some and other classes of which largely consist in the applicability or inapplicability to them of logically different types of law- propositions and law-like propositions. Assertions about a person’s mind are therefore assertions of special sorts about that person. So questions about the relations between a person and his mind like those about the relations between a person’s body and his mind, are improper questions. (…) It follows that it is a logical solecism to speak, as theorists often do, of someone’s mind knowing this, or choosing that. The person himself knows this and chooses that, though the fact that he does so can, if desired, be classified as a mental fact about that person. (…) ‘my mind’ does not stand for another organ. It signifies my ability and proneness to do certain things and not some kind of personal apparatus without which I could or would not do them. (…) Where logical candour is required from us, we ought to follow the example set by novelists, biographers, and diarists, who speak only of persons doing and undergoing things. (Ryle 1949a, 160-61)

According to Stout, Ryle’s mental language ‘serves a descriptive function that we have no reason to suppose can be served without such language’ (Stout 2003, 39), which he illustrates by the following quotation from The Concept of Mind:

To talk of a person’s mind is not to talk of a repository which is permitted to house objects that something called the “physical world” is not permitted to house; it is to talk of the person’s abilities, liabilities and inclinations to do and undergo certain sorts of things, and of the doing and undergoing of these things in the ordinary world. (Ryle 1949a, 190)

Stout makes two distinctions: 1) between limiting oneself to discussing overt phenomena and discussing phenomena that can be measured by scientific instruments; 2) between a reductive conception of behaviour and a non-reductive one.

Distinction (1) is between limiting oneself to, for example, a discussion of stimuli and responses, and limiting oneself to overt phenomena, such as doings, sayings and gestures, which can, however, not ‘be read off a scientific measuring device’ (Stout 2003, 40). It is the difference between saying that ‘the way someone is disposed to behave can be specified in terms which make no use of mental or intentional language’ (Stout 2003, 40) – i.e. that statements about the mind can, without loss of meaning, be translated into ones about the body – and saying that this is not possible. Distinction (2) is between specifying behaviour in a way which is neutral ‘between the description of a machine’s behaviour and the description of a person’s behaviour’ (Stout 2003, 40) and specifying it in a way which does differ between machines and persons. In this way he can call Ryle a behaviourist while still allowing him to make allusion to things beyond people’s purely bodily behaviour. Ryle is regarded as a non-reductive behaviourist who limits himself to overt phenomena, but does not go so far as to limit himself to ones that can be measured by scientific instruments. Stout characterizes Ryle’s behaviourism as follows:

The non-reductive (Rylean) behaviourist operates with an irreducibly mentalistic or intentional conception both of bits of behaviour and of ways of behaving. One way of spelling out such a non-reductivist behaviourist view would be the following: What it is to say that someone is in a certain state of mind is to say that they are disposed to behave in a way characteristic to that state of mind. It may not be possible to describe the characteristic way of behaving without referring to that state of mind considered abstractly. So this is not the sort of behaviourism that would explain to someone what certain mental terms meant if they did not already know. It is instead the sort of behaviourism that reveals the nature of the systematic relationship between being entitled to describe states of mind and being entitled to make behavioural inferences. (Stout 2003, 41)

90

The Concept of Mind: Main Aims, Method and Reception

At first sight the views of Park and Stout may seem irreconcilable, since Stout rejects Park’s position, but there are important similarities between their positions. They both argue against the idea that Ryle was a reductionist behaviourist, as well as against the thought that Ryle would have aimed at characterizing states of mind in a purely scientific way. However, Stout mainly focuses on the exact type of behaviourism which may be attributed to Ryle, whereas Park concentrates on the importance of meaning and language for Ryle, and on his rejection of a denotational theory of meaning. That Stout nevertheless rejects Park’s position is due, in my opinion, to the fact that he did not clearly distinguish between different ways in which ‘description’ (and perhaps also ‘denotation’) is used. Stout attributes to Park the position that mental language describes nothing at all, holding her to claim that mental language serves another purpose, which may have been caused by Park’s – somewhat misleading – description of Ryle’s position as linguistic antirealism. However, Park merely claims that Ryle’s mental terms do not denote in the sense in which ‘Fido’ denotes Fido, and surely she attributes to Ryle the position that mental terms do not describe in the way ‘Fido’ describes Fido. But in the more general sense of ‘denotate’, ‘signify’ or ‘describe’, she would not have denied them of mental concepts. Another point of contention is that whereas Stout suggests that Ryle was a (non- reductive) behaviourist, Park thinks he was not a behaviourist at all. This difference may be due to Park’s interpretation of behaviourism in terms solely of one version (or, at best, one cluster of versions of the same type) of logical behaviourism. What she rejects is the view that Ryle seeks to translate statements about the mind into ones that are about observable behaviour, which contain no reference to the state of mind considered abstractly – which differs from Stout’s description of Rylean behaviourism: ‘The non-reductive (Rylean) behaviourist operates with an irreducibly mentalistic or intentional conception both of bits of behaviour and of ways of behaving’ (Stout 2003, 41). Park does not take into account the kind of behaviourism Stout has in mind, but there is no reason why she would not have accepted it. Her arguments against ‘regular’ – in the sense of reductive – logical behaviourism do not work against the non-reductive type of behaviourism Stout attributes to Ryle. As said before, we should not forget that in 1949 Ryle did not have such a complete and coherent position and was not totally clear about it himself. He did not aim at rejecting behaviourism, focusing his attention rather on the dualist theory – which he considered to be far more dangerous. Ryle himself later admitted that at the time of The Concept of Mind he had not yet realised the full scope of behaviourism, nor had he been fully aware that dualism and behaviourism were two sides of the same coin. And although it is certainly true that Ryle could have criticised behaviourism theoretically by means of his rejection of a denotational (homogeneous) theory of meaning, he did not in fact do so, at least not in a systematic way.

Concluding Remarks

In this chapter I have given an interpretation of The Concept of Mind by focusing on Ryle’s method and main aims, and by giving an overview of the most important interpretations of the book. The reviews of The Concept of Mind exposed confusions and a vague, sometimes misleading, use of some of the basic concepts and main distinctions. Several commentators wrote, for example, that it is never clear precisely whom Ryle has in mind when he attacks the dogma of the ghost in the machine. Another point of criticism often voiced by reviewers

91

The Philosophical Development of Gilbert Ryle

(for instance, MacDonald, King, Austin, Russell and Hampshire) is that Ryle himself – in his crusade against conceptual mistakes – was not able to avoid using confused or vague expressions and concepts himself. This, of course, made it difficult to unravel his arguments and assess their power and convincingness. Furthermore, some reviewers demanded a method of explanation instead of mere descriptions and argued that Ryle never in fact touched upon the arguments of his opponents. Above all, Ryle was criticised for insufficiently defining his target and for an unclear use of his basic tool – the category- mistake. A proper understanding of the category-mistake which Ryle attributes to Cartesian dualism seems to hold the key to a proper interpretation. As I argued, in The Concept of Mind category-mistakes occur on many different levels, and this partly explains the difficulties in interpreting the book. The most general category-mistakes are: treating mental concepts as names, and thinking that mental language describes a mental world and that there is something like a physical world which is fully describable by non-mental language. These two category-mistakes expose both dualism and behaviourism, although Ryle was not fully aware of this at the time. The mistakes as they are described are of different types. The first concerns the question how specific expressions mean what they mean. The answer is that there is an endless diversity of categories of meaning. We have to look at how expressions are used; the heterogeneity of our use of language has replaced Ryle’s earlier homogeneous, denotational theory of meaning. Mental concepts should no longer be treated as names. The second category-mistake concerns the question what our language – as a whole – is ultimately about (where ‘being about’ has to be interpreted in the broad sense rather than in the narrow sense of ‘Fido’ being about Fido). Our language is ultimately about the world; not about a mental or a bodily world but about our ordinary world. Physical and mental language are just different ways to describe the same world. This is a conceptual rather than an ontological position. Ryle replaces the view of having one type of relation between language and reality – reality consisting of many different worlds – by that of an endless diversity of relations between language and reality. According to the latter view reality is defined as just one world, that is the world. These two category-mistakes were not, as such, identified by the reviewers of The Concept of Mind, except by MacDonald (1951). The main reason why they did not recognize the first category-mistake was that Ryle was still, at times, employing a denotational terminology, as was pointed out by Hampshire (1951). Ryle’s rejection of the dichotomy of a mental and a physical world, instead of merely rejecting a mental world, and his transcending of this dichotomy by replacing it with the ordinary world, about which we can talk in terms of mental and in terms of bodily language, were also neglected. This was mainly due to Ryle’s rather misleading claim that he wanted to get rid of the ghost in the machine – thereby suggesting that the machine could be retained – and his frequent appeals to behaviour and references to the verification of mental statements by means of overt behaviour. This lack of recognition of the ‘real’ category-mistake(s) which Ryle tried to expose was largely responsible for the behaviouristic interpretations then and now, and these interpretations still prevent Ryle from being taken seriously in contemporary philosophy of mind and language. In the following chapter we will see that in his later papers Ryle becomes more explicit – and accurate – about his rejection of behaviourism. Ryle’s non-reductive method of connective philosophical analysis, stripped of its alleged behaviourism, may still be of value for modern philosophy as will be suggested in the concluding chapter of this thesis. It is, for example, relevant for discussions about the foundations of . 92

Gilbert Ryle’s Later Writings

Chapter 4

Gilbert Ryle’s Later Writings

After The Concept of Mind (1949a) Ryle felt the need to clarify some issues. After all, the book was written to provide an example of what he considered to be good philosophy and hardly contained any theoretical claims concerning the nature of philosophy – Ryle’s explanation of the category-mistake probably came the closest to any such thing. The theoretical claims which Ryle had presented in his earlier papers, such as ‘Categories’ (1938) and ‘Philosophical Arguments’ (1945), could not fully carry the weight of the method of The Concept of Mind. Hence, in the late fifties and early sixties Ryle wrote a series of papers meant to add to and clarify his earlier work: ‘Ordinary Language’ (1953b), ‘Proofs in Philosophy’ (1954b), ‘The Theory of Meaning’ (1957a) and ‘Use, Usage and Meaning’ (1961a). Furthermore, in 1953 Ryle gave the Tarner Lectures in Cambridge, which were published as Dilemmas (1954a). His purpose with these lectures was similar to that of The Concept of Mind, namely to show what good philosophy is by providing examples. Ryle later said that the Tarner Lectures ‘were fairly explicitly dedicated to the consolidation and diversification of what had been the meta-theme of the Concept of Mind’ (Ryle 1970, 12). A difference between the two books is that in the later one Ryle qualifies his notion of the category-mistake; Dilemmas is also more explicit on the theoretical framework backed up by the numerous examples, which makes it easier to grasp. But the majority of Ryle’s writings after The Concept of Mind is dedicated to the notion of thinking, which kept fascinating and troubling him until his death. In his later writings on thinking Ryle sharpens, and sometimes changes, his earlier thoughts in The Concept of Mind with a focus on persuading people, using analogies rather than the reductio ad absurdum. Ryle’s most well-known writings on thinking are his papers written between 1951 and 1968, which were printed in the second volume of his Collected Essays (1971b)72. However, there are many other papers, most of which were published posthumously:

• On Thinking (Kolenda 1979): a collection of posthumously published papers and lectures, containing papers from the last period of Ryle’s life, up until 1976 – the year of his death.73 • Aspects of Mind (1993), a collection of posthumously published papers edited by René Meyer, Professor at the University of Pretoria and in 1964 a student under

72 ‘Thinking and Language’ (1951a), ‘Thinking’ (1953a), ‘A Puzzling Element in the Notion of Thinking’ (1958), ‘A ’ (1962a), ‘Thinking Thoughts and Having Concepts’ (1962c), ‘Thinking and Reflecting’ (1966-67), ‘The Thinking of Thoughts – What is ‘le Penseur’ doing?’ (1968a) 73 ‘Adverbial Verbs and Verbs of Thinking’ (1979a), ‘Thought and Soliloquy’ (1979b), ‘Thought and Imagination’ (1979c), ‘Thinking and Self-Teaching’ (1972a), ‘Thinking and Saying’ (1972b), ‘Mowgli in Babel’ (1974), ‘Negative “Actions”’ (1973), ‘Improvisation’ (1976b)

93

The Philosophical Development of Gilbert Ryle

Ryle in Oxford. It contains papers to which Ryle gave Meyer access in 196474, papers discovered by Meyer in the ‘Ryle Collection’ in Linacre College Library and in the Philosophical Library in Oxford75, and Meyer’s own notes of Ryle’s lectures and seminars from 1964. • Finally, the separately published paper ‘Courses of Action or the Uncatchableness of Mental Acts’ is relevant to Ryle’s views on thinking. This paper was prepared for delivery at the annual conference of the Experimental Analysis of Behaviour Group at Bangor in April 1974 (but was in the end read by Professor T. R. Miller because Ryle was not able to attend), and was only published in 2000 in Philosophy.

This chapter is divided into three parts: I start my discussion of Ryle’s later writings by giving an account of the theoretical papers postdating The Concept of Mind, followed by a discussion of Dilemmas. Finally, the focus will be on Ryle’s account of thinking, on which relatively little has been written. It is particularly this account that bears close resemblances to Wittgenstein’s views and makes one realize that a behaviouristic interpretation of Ryle is not just a minor mistake, but rather a serious and damaging misinterpretation.

Ryle’s Meta-Philosophical Papers

The Concept of Mind (1949a) had a meta-philosophical purpose, but did not contain an explicit theoretical account on how to do philosophy. The book was supposed to show what philosophy was instead of providing a theory of it. As we have seen in the reviews of The Concept of Mind, this relative absence of theoretical claims left the readers of the book with many unanswered questions and difficulties. In the 1950s and early 1960s Ryle wrote four theoretical papers which, on the one hand, can be regarded as responses to contemporary developments in the field of philosophy – e.g. the tendency to think of the enterprise of philosophy as revealing the logic of our ordinary use of language as a mere verbal enterprise, and the rising criticisms of the general program of the philosophy of language (e.g. on the paradigm case argument). On the other hand they were further explications of his Concept of Mind. An examination of these later papers will, I think, not only shed light on Ryle’s later views but also – in retrospect – on The Concept of Mind; for he now presents a more explicit theoretical framework. Moreover, the complaint – which has often been expressed – that Ryle merely describes whereas he should have given explanations can also be countered by Ryle’s later account of the essential difference between philosophy and science (and mathematics).

74 ‘Is Induction a Sort of Inference?’, ‘Induction’, ‘Deductive and Inductive Thinking’, ‘Our Thinking and Our Thoughts’ and ‘Reason’, all of which were written in the late fifties and early sixties. Since my focus will be on thinking, I shall only discuss the last two papers in this chapter. 75 ‘Ontological and Logical Talk in Wittgenstein’s Tractatus’, according to Meyer written in the late thirties; ‘The Meno’, written in preparation for delivery to the Cambridge B Club in October 1976 (however, Ryle died early in that month); and a Paper Read to the Oxford Philosophical Society 500th Meeting in 1968. The first paper will be discussed in Chapter 6, the others are only briefly mentioned. 94

Gilbert Ryle’s Later Writings

Continuity in Ryle’s Ideas on Philosophical Method Many of the ideas which Ryle outlined in his later theoretical papers had already been – in some form – present in the earlier ones. As was already mentioned in the preceding chapters, at the time of The Concept of Mind (1949a) Ryle adhered to what may be called a use theory of meaning. The homogeneity of denotationalism had been replaced by the heterogeneity of use, and the need of the existence of a Third Realm housing meanings was rejected. It was no longer assumed that all words, phrases and sentences have the same homogeneous role of naming extra-linguistic objects or processes. On the contrary, as Ryle later alleged, ‘there is an endless variety of categories of meaning’ (Ryle 1957, 365). Ryle also continued to express his view that philosophy was about informal logic; he did not want to translate our everyday expressions into formal notations. He made this point in various works, e.g.:

Of those to whom this, the formaliser’s dream, appears a mere dream (I am one of them), some maintain that the logic of everyday statements and even the logic of the statements of scientists, lawyers, historians and bridge-players cannot in principle be adequately represented by formulae of formal logic. The so-called logical constants do indeed have, partly by deliberate prescription, their scheduled logical powers; but the non-formal expressions both of everyday discourse and of technical discourse have their own unscheduled logical powers, and these are not reducible without remainder to those of the carefully wired marionettes of formal logic. (Ryle 1953b, 316)

Ryle not only showed the difference between philosophy and formal logic, but in his inaugural lecture (Ryle 1945) he also pointed at the difference between philosophical arguments and the types of arguments that are used in mathematics and science. Philosophical arguments are neither inductions nor deductions. An argument which is especially appropriate to philosophy is the reductio ad absurdum. In ‘Proofs in Philosophy’ (1954b) Ryle followed up on this issue and claimed that, as opposed to mathematics or science, there are no proofs in philosophy – by which he does not mean that philosophy is not argumentative or rhetorically persuasive. Proofs are built upon true premises, but philosophers’ arguments do not work like that. Whereas philosophers usually reject the idea of building anything on this or that particular premises, they should reject the whole idea of premises-theorem proofs in philosophy. ‘Only we have been shy of saying anything of the sort, because we have inadvertently assumed that any argument with any degree of logical powerfulness must have the shape of a premises-theorem proof.’ (Ryle 1954b, 321) Another important difference between philosophy and science is stressed in ‘Philosophical Arguments’ (1945), The Concept of Mind (1949a) and ‘Abstractions’ (1962b) by drawing a comparison between doing philosophy and drawing maps. We should not put philosophical problems on a par with, for example, those of the chemist which can be solved piecemeal. Philosophical quandaries are of a different nature in that they are intertwined in all sorts of ways. There is no such thing as first solving philosophical problem a and then starting on philosophical problem b, etc.

Philosophers’ problems do not in general, if ever, arise out of troubles about single concepts, like that, say, of pleasure or that of number. They arise, rather, as the traffic-policeman’s problems arise, when crowds of conceptual vehicles, of different sorts and moving in different directions meet at some conceptual cross-roads. All or a lot of them have to be got under control conjointly. This is why, in its early stages, a philosophical dispute strikes scientists and mathematicians as so messy an affair. It is messy, for it is a traffic-block – a traffic-block which cannot be tidied up by the individual drivers driving their individual cars efficiently. (Ryle 1954b, 325)

95

The Philosophical Development of Gilbert Ryle

Different non-synonymous words and phrases have different, what Ryle calls, ‘implication threads’ (Ryle 1962b, 443), that is, they are ‘different in having different implications, in requiring different tests for truth or falsehood, in being compatible and incompatible with different affiliated statements, in being evidence for or against different corollaries, and so on’ (Ryle 1962b, 442-43). Since the things we say are often mixtures of different words or different phrases, and ‘sometimes the implication threads generated by one of them pull or seem to pull across or away from the implication threads generated by another’ (Ryle 1962b, 443), it is no longer sufficient to be familiar with the separate contributions of single words and phrases. We need to be able to say something about the relation between them, for instance about why their apparent conflict is not a real one. Ryle argues that:

when two or twenty familiar implication threads seem to pull across and against one another, it is no longer enough to be able unperplexedly to follow along each one by itself. We need to be able to state their directions, their limits and their interlockings; to think systematically about what normally we merely think competently and even dexterously with. (Ryle 1962b, 444)

Philosophical questions for the crossing implication threads are questions of a new sort, not just new questions of an old sort. The ideas discussed so far have been present from the early works onwards, but in Ryle’s later theoretical papers there are some relatively new ideas as well. They are new not in the sense that they reveal a radical change of mind, but rather in that they were not made explicit before. To these we will turn now.

The Ordinary Use of Language In The Concept of Mind Ryle frequently appeals to our ordinary use of expressions, for example, when he refers to the ordinary use of concepts such as ‘intelligent’, ‘being a good car driver’, ‘knowing how to do x’. But what does philosophy have to do with our ordinary use of concepts in the first place? Do philosophers have to limit themselves to what ordinary people say, thereby neglecting technical concepts? And what is this ‘use’ – does it require a sociological investigation into people’s customs and practices? These are some of the misunderstandings which Ryle attempts to rectify in his papers on ‘ordinary language’, ‘use’, ‘linguistic usage’ and related concepts. It is important that ‘the use of ordinary language’ should not be identified with ‘the ordinary use of the expression’. ‘Ordinary’ means ‘stock’ or ‘standard’ rather than ‘everyday life’ or ‘non-specialist’, as it qualifies ‘use’ rather than ‘language’. The ordinary use of a fish-knife is to cut up fish with; a non-stock use would be to cut tomatoes with it. Technical as well as everyday life concepts have an ordinary or stock use, and often they also have many non-stock uses. ‘Use’ in the phrase ‘the ordinary use of the expression x’ is all about what is done with the expression: ‘(…) the enquiry is an enquiry not into the other features or properties of the word, coin or pair of boots, but only into what is done with it, or with anything else with which we do the same thing.’ (Ryle 1953b, 305) To use the analogy of the coin, the enquiry is into the use of the coin – that is what I can and cannot do with it – and not into, for example, its date, shape or colour. This reveals why it is misleading to categorize philosophical questions as either linguistic or non-linguistic ones – the use of ‘cause’ is the same, according to Ryle, as that of ‘Ursache’; they are not about the English (or the German) language. When Wittgenstein said (according to Ryle somewhat misleadingly) that philosophical problems were linguistic problems he meant

96

Gilbert Ryle’s Later Writings problems about the logic of how expressions work, that is the description of the rules of use, not to be confused with philological or rhetorical problems.76 Thus, ‘use’ should be interpreted in a logical rather than in a philological sense and ‘the ordinary use of an expression’ is not the same thing as ‘the use of ordinary expressions’. But of course there is some kind of primacy of the stock use of our everyday expressions over that of scientific or specialist expressions in that explanations of technical terms depend on our knowledge of the use of ordinary expressions. Moreover, Ryle does sometimes recommend that philosophers use everyday language. First of all, they should only use technical terms if useful, which does not just apply to philosophers but to anyone. A reason more connected to the nature of philosophy is that a philosopher’s job is generally to determine the cross-bearings between the concepts of different theories or disciplines, and in doing so, ‘he cannot naively employ the dictions of either theory. He has to stand back from both theories, and discuss the concepts of both in terms which are proprietary to neither. He may coin neutral dictions of his own, but for ease of understanding he may prefer the dictions of Everyman.’ (Ryle 1953b, 314) Furthermore, since there is no peculiar field of knowledge or skill that belongs to philosophy, it is often worth to appeal to everyday language, whereas this is much more difficult in the case of science or law, which do have their own special objects. Although it is certainly clarifying in important respects, our talk about the use of expressions sometimes mislead us as well; a few possible misunderstandings have to be dealt with. First of all, ‘use’ should not be interpreted in the sense of ‘utility’ or ‘usefulness’. To learn the utility of a sparking-plug is not yet knowing how to operate with it; and the way, method or manner of whistling can be known without attributing any usefulness to the trick. A second confusion is between ‘use’ and ‘usage’, that is custom or practice. Knowing how to operate with a canoe-paddle, a traveller’s cheque or a word is not identical with knowing sociological generalities about other people. Moreover, in order to describe usages we have to have a description of uses. In order to avoid the confusions between ‘use’ and ‘usefulness’ as well as between ‘use’ and ‘usage’, Ryle recommends to use ‘employ’ (or ‘employment’) instead of ‘use’.

Language versus Speech and Words versus Sentences Although Ryle has rejected denotationalism, he warns us against turning the scale in favour of philosophy as some kind of philology: ‘The difficulty is to steer between the Scylla of a Platonistic and the Charybdis of a lexicographical account of the business of philosophy and logic.’ (Ryle 1957, 371) We have already seen that philosophical problems are not to be treated as a kind of philological ones. If we put the stress on ‘use’ in ‘the ordinary use of language’ we may be able to avoid this misunderstanding. However, this mistake can also arise if we confuse speech with language, or rather ‘using “Language” and “linguistic” ambivalently both for dictions and for dicta, i.e., both for the words, etc. that we say things

76 This point of view was heavily criticized, for example by Zeno Vendler. He argued that no such clear distinction could be made between conceptual and linguistic inquiries: ‘How could Ryle know, without an exhaustive study of both , that the use of Ursache is the same as that of cause?’ (Vendler 1967, 252) However, as makes clear in his discussion of the claim that philophical inquiries are language-neutral, it is irrelevant to Ryle’s claim whether or not the use of ‘Ursache’ is exactly the same as the use of ‘cause’. They have to correspond in the relevant aspects, but not in every respect. ‘What matters is that the same questions can be discussed by a German philosopher with reference to the use of “Ursache” as are discussed in English with reference to the use of “cause”.’ (Hanfling 2000, 66-67)

97

The Philosophical Development of Gilbert Ryle in and for what we say in them’ (Ryle 1961a, 413). Therefore the distinction is in need of clarification.

A language, such as the French language, is a stock, fund or deposit of words, constructions, intonations, cliché phrases, and so on. ‘Speech’, on the other hand, or ‘discourse’ can be conscripted to denote the activity or rather the clan of activities of saying things, saying them in French, it may be, or English or some other language. (Ryle 1961a, 407)

In other words, language is what we say things in – e.g. in French or English – whereas speech is used to indicate what we say in them. Speech involves language but cannot be reduced to it: ‘The reproof “You cannot say that and speak good French” is generically different from the reproof “You cannot say that without absurdity”. (…) Speech-faults are not to be equated with Language-faults. (Ryle 1961a, 410-11) Philosophy is concerned with exposing absurdities and therefore with speech-faults rather than language-faults – showing why we cannot say what we are tempted to say rather than showing why we cannot say things in the language we say them in, or are tempted to say them in. A philosopher’s task is to look at the dicta that are made or can be made with certain problematic words, and to consider possible speech-faults. ‘E.g., what sorts of dicta could not be significantly made with them [these words – CV], and why; what of argument pivoting on these live dicta would be fallacious, and why; what kinds of verification-procedures would be impertinent, and why; to what kinds of questions such live dicta would be irrelevant, and why; and so on.’ (Ryle 1961a, 413) Closely connected with the difference between speech and language is that between words and sentences – a difference which should also not be neglected. Only words have a use, or better an employment; we cannot ask whether a person knows how to employ a certain sentence. Ryle compares words and sentences respectively with the salt, sugar, flour, beans and bacon, and the pie for which they are used. ‘Sentences are things we say. Words and phrases are what we say things with.’ (Ryle 1953b, 311) Just as it does not make sense to ask how the cook uses the pie, we cannot ask how sentences are used; only how words are used. It follows that only words (and phrases) can be misused. Furthermore:

(…) sentences and clauses make sense or make no sense, where words neither do nor do not make sense, but only have meanings; and that pretence-sentences can be absurd or nonsensical, where pretence-words are neither absurd nor nonsensical, but only meaningless. I can say stupid things, but words can be neither stupid nor not stupid. (Ryle 1953b, 313)

Ryle considers words to be the atoms of a language; sentences are the units of speech. The first we have to learn in order to master a language; the second we construct when we say things. ‘I am its author, not its employer. Sentences are not things of which I have a stock or fund.’ (Ryle 1961, 408) Therefore we have to clearly distinguish between words and sentences, a distinction not always observed by philosophers. We should avoid talking about the use of sentences in the way we talk about the use of words or expressions. In learning a language we learn how to employ these words and in doing so we construct sentences. And the truth which is conveyed by a given sentence is not just an assemblage of the several meanings of the several words in the sentence. In a different order these very same words might have produced a falsehood (or a different truth or nonsense). In short, we should be careful not to confuse speech with language, and not to treat words and sentences in the same way. We do not want to end up with a philosophy that is mere philology, any more than that we want to turn philosophy into an enterprise of discovering platonic entities.

98

Gilbert Ryle’s Later Writings

Philosophy as an Inter-Level Activity As was mentioned above, Ryle described philosophy as opposed to science in that philosophical problems cannot be solved piecemeal and in that philosophy does not give proofs. But if philosophy does not concern giving proofs, what are we doing when philosophising? According to Ryle to do philosophy is to make explicit the previously implicit logic of the employment of expressions. He compares this activity with someone who tries to ‘formulate verbal recipes or instructions for correct operations with numbers’ (Ryle 1954b, 322), which is different from merely being able to operate with numbers as children do. The second activity always precedes the first. We already know how to operate with certain expressions – but philosophical arguments are second-order; they are operations upon operations: ‘we are trying to codify in words of one level the rules observed in the employment of words of another level’ (Ryle 1954b, 323). As an example of proper philosophy Ryle mentions Frege, who showed, for example, that those were wrong who suggested that just as adjectives such as ‘green’ and ‘square’, adjectives such as ‘one’ and ‘two’ also stood for qualities of things. He showed that ‘numerical expressions will not go through all the same inference-hoops as - expressions’ (Ryle 1954b, 324) – ‘“numbering 1” is not, as “honest” is, the sort of predicate which can characterise me. Numerical predicates – such as ‘numbering 1’ – can characterise only such subjects as “the Oxford Professors in this room”’. (Ryle 1954b, 324) By pointing at a specific fault in the suggested codification (the fault of characterizing adjectives such as ‘one’ and ‘two’ as standing for qualities of things) Frege fixed a positive element of the required codification. As opposed to proofs, philosophical arguments concern multiple levels: ‘(…) they are operations not with premisses and conclusions, but operations upon operations with premisses and conclusions. (…) Proving is a one-level business; philosophical arguing is, anyhow sometimes, an inter-level business.’ (Ryle 1954b, 324) This is exactly what constitutes the difference between philosophy and the sciences. Whereas the sciences explain in the sense that they are trying to discover new facts and directly operate on them, philosophy does not concern the discovery of facts but operates on operations. In this sense philosophy is concerned with the level of descriptions, rather than with the level of explanations. ‘Abstractions’ (1962b) is one of the papers in which Ryle is concerned with this difference between concrete or factual questions and abstract or conceptual questions. He mentions the example of St Augustine who was perfectly capable of answering concrete questions about time, e.g. how to use a calendar, but unable to answer abstract questions such as ‘Why could time never have started?’ and ‘What is Time – is it a thing, a process or a relation?’. Abstract or conceptual questions are not about things or processes but about concepts. Seen in this light, the criticisms frequently voiced that Ryle merely provides his readers with descriptions whereas he should have given explanations seems to be misdirected – since from Ryle’s point of view explanations do not belong to the realm of philosophy.

Dilemmas: The Tarner Lectures

In order to illustrate his views on the nature of philosophy, Ryle introduced real philosophical problems and the methods for solving them. As we have seen, he did this in The Concept of Mind with respect to various kinds of problems about the mind, or more accurately: about our descriptions of mind. He then decided to widen his scope, using a

99

The Philosophical Development of Gilbert Ryle variety of other examples in the Tarner Lectures77 held in Cambridge in 1953, and published as Dilemmas (1954a). The book contains eight chapters: one on the theory followed by seven practical applications of it. After discussing the theoretical part, I will briefly sketch the case studies, concluding with a more detailed discussion of one in particular – ‘It Was To Be’, which deals with the dilemma between the idea that we are responsible for at least some of our actions and the equally stringent – and seemingly conflicting – thought that everything that is was to be.

Philosophical Dilemmas The general title, Dilemmas, suggests that the book is about conflicts between different theories. At the beginning of the book Ryle specifies the particular type of dilemmas in which he is interested. His aim is to discuss a specific type of conflict between theories, namely ‘quarrels between theories, or, more generally, between lines of thought, which are not rival solutions of the same problem, but rather solutions or would-be solutions of different problems, and which, none the less, seem to be irreconcilable with one another’ (Ryle 1954a, 1). Therefore, the conflict or dilemma does not exist in the sense that there is one problem or question and two competing solutions or answers, but rather in the sense that we are confronted with two seemingly contradictory statements – accepting the one seems to logically entail rejecting the other – both of which we strongly insist on. In reality, however, these are only apparent conflicts and not real ones, since the statements or lines of thought in question are not answers to the same question but to different ones. Take for example the different account of perception given by the neurophysiologist and by the ordinary man. Whereas the neurophysiologist arrives at a theory of the mechanisms of perception according to which we never perceive what is really there, the ordinary man finds out what is there (e.g. that the clock has stopped) by perceiving, so they seem to be in conflict. However, according to Ryle, there is no real competition between the different accounts of perception of the neurophysiologist (what are the mechanisms of perception?), the philosopher (what is perception?) and ordinary people (how did you find out that the clock had stopped?); they are simply different questions. How are we to deal with these seeming conflicts? Once again, Ryle stresses that we need to reveal the informal logic of our ordinary and our technical concepts and to show the different relations between these concepts in different – partly overlapping – areas of thinking. What we do not need is more ‘internal’ evidence, e.g. more scientific research, in order to back up any of the seemingly conflicting lines of thought. It is tempting, though, to look for more evidence of that kind, since this is precisely what would be needed if we were dealing with real conflicts, that is, between rival solutions to the same problem. But in the cases which Ryle discusses – seeming conflicts between solutions to different problems – we need philosophical enquiries that do not concern the internal aspects of these lines of thought, but rather the informal logic of the relations between them. It is not an issue of one line of thought triumphing over another; ‘what is at stake is not which shall win and which shall lose a race, but what are their rights and obligations vis à vis one another and vis à vis also all other possible plaintiff and defendant positions’ (Ryle 1954a, 5). It is not just, or even primarily, the technical concepts themselves that trick us into making logical mistakes, but rather the underlying non-technical concepts. In fact, since highly technical concepts have relatively well-defined functions, they are in general easier

77 These public lectures had been established in 1916 by Mr. Tarner and were supposed to deal with the philosophy of the sciences. Three of Ryle’s most well-known predecessors were Broad (1923), Russell (1926) and Moore (1929). 100

Gilbert Ryle’s Later Writings to map than non-technical concepts. The latter belong to no specific discipline and their functions are far more intricate, involving a high complexity of many overlapping and intertwined areas of thinking. As a result, highly technical concepts, such as those of Bridge, are philosophically uninteresting because they usually do not generate philosophical puzzles. More interesting are less technical concepts that have a place in many different fields – these are the ones that generate dilemmas. One may think that dilemmas can be avoided altogether, since, after all, they are seeming conflicts rather than real ones. This, however, is not the case: inevitably we make the attempt to fix particular concepts, such as pleasure, by describing them in terms of concepts which belong to theories in other fields, e.g. political or psychological theories. However, this does not mean that we are to blame for borrowing them in the first place, since according to Ryle ‘we learn the powers of a borrowed tool side by side with learning its limitations, and we find out the properties of the material as well when we find out how and why the borrowed tool is ineffectual upon it, as when we find out how and why it is effectual’ (Ryle 1954a, 66). Moreover, there is no other option than working on concepts that are still uncharted. In the end, only trial and error shows us what we can and cannot do with our concepts. This description of philosophical method explains Ryle’s frequent use of the reductio ad absurdum. We only know that we mishandle concepts when absurdities arise, which indicates that we treat them as belonging to one category when they in fact belong to quite another one. Real conflicts between lines of thought do not reveal category-mistakes whereas dilemmas do. Clearly, Ryle still thinks of doing philosophy in terms of categories and category- mistakes, a method he developed in ‘Categories’ (1938) and The Concept of Mind (1949a). However, he now feels the need to qualify his use of category and category-mistake, probably (at least partly) as a response to the criticisms that his use of these concepts in The Concept of Mind (1949a) had provoked. He does not use ‘category’ in an exact way, as some sort of ‘skeleton-key’ that ‘will turn all locks for us’ (Ryle 1954a, 9) but rather as a mere ‘coal-hammer’ which ‘will make a satisfactory knocking noise on doors we want opened to us’ (Ryle 1954a, 9). ‘Category’ should be used in an inexact – and amateurish – way, ‘[giving] the answers to none of our questions but it can be made to arouse people to the questions in a properly brusque way’ (Ryle 1954a, 9). The philosopher’s job by no means ends here. It still remains to be shown in detail that and how the discipline or sub- discipline of the concept which was supposed to function in a way similar to that of another discipline or sub-discipline in fact differs from it more, or less, than was initially supposed. And this is what Ryle puts into practice in the remaining chapters of the book. Ryle’s qualification of ‘category-mistake’ should not be interpreted as a departure from his earlier account. He felt the need to qualify the strictures of its applicability because he realised that he had not been sufficiently clear about this in The Concept of Mind (1949a), in which it is sometimes suggested that Rylean ‘categories’ and ‘category-mistake’ perform the role of a ‘skeleton-key’. But his overall position did not change in any substantial way. It is commonly accepted that the main aim of Dilemmas is meta-philosophical. His message that philosophy is heterogeneous, that there is no such thing as the philosophical method, has been noticed by various scholars. As A. M. Quinton wrote: ‘There is no determinate set of precisely articulated logical patterns by reference to which all philosophical paradoxes can be resolved. Philosophy is inescapably ad hoc.’ (Quinton 1954, 89) In Revue de Métaphysique et de Morale a reviewer describes Ryle’s meta- philosophical aim as follows: ‘On ne peut mettre une fin à ces disputes d’une façon stéréotypée. Le mérite du livre c’est de nous initier aux méthodes convenables, chaque fois

101

The Philosophical Development of Gilbert Ryle différentes.’ (1958, 499) More recently, Jennifer Szalai has credited Ryle with an existentialist position, in the sense that ‘the world does not contain a “theory of everything” awaiting our discovery’ – there is no one particular theory that can provide the answer for everything: ‘The possibilities for understanding our world are infinite, and the freedom to choose means the freedom to misunderstand and, ultimately, the freedom to fail.’ (Szalai 2002, 57) Whatever the merits of Ryle’s solutions of individual dilemmas – and as we shall see, they have been criticized – his meta-philosophical message has proven to be of great value.

Discussions of Concrete Dilemmas Ryle’s dilemma’s are not all of the same type. According to William Lyons (1980a), we can distinguish between three different types of dilemmas: the first type is caused by different answers to different questions; the second by a seeming competition between science and an everyday account of the world; the third type is caused by turning an unproblematic view into a problematic one by making a category-mistake. On closer inspection, however, it seems that all dilemmas are, in a broad sense, examples of being answers to different questions. The first two dilemmas – ‘It was to Be’ and ‘Achilles and the Tortoise’ – are different from the others in that they are more like riddles than like problems that really worry us; they can be solved in a black-and-white kind of way, e.g. either Achilles will or will not catch the Tortoise; either some things are our fault or not. The other dilemmas do not admit such a yes-or-no answer. Furthermore, the first three dilemmas – the third one concerns pleasure – are rather marginal or somewhat peripheral tangles, whereas the last four are more fundamental; they are ‘out in the middle of the room’ (Ryle 1954a, 68). Whereas Dilemmas was widely appreciated for its meta-philosophical value (cf. Quinton 1954 and Szalai 2002 mentioned above), its reviewers were more critical about Ryle’s specific treatment of individual dilemmas. The main disadvantage of his discussions, according to his reviewers, was that he did not even start to solve the philosophical problems underlying the dilemmas, and merely aimed at recognising the dilemmas and providing proper descriptions of them. As J. O. Urmson puts it, ‘the treatment of the more important problems is too brief and, deriving from the lecture-hall, too loosely-knit to constitute the main interest’ (Urmson 1955, 554). And A. M. Quinton claims: ‘Professor Ryle’s technique is rather more suggestive than convincing. His rejection of the disputes of the schools and of the paraphernalia of erudition (…) seems to exclude more valuable things as well – in particular rigorous methods of argument and precisely defined expression of problems.’ (Quinton 1954, 90) In order to assess these criticisms, we have to keep in mind that the main aim of Dilemmas was of a meta-philosophical nature. As L. J. Russell correctly remarks: ‘It is not vital to his [Ryle’s] task that he should convince us that he has resolved all the difficulties. What he is doing is to show how philosophical enquiry is appropriate to situations where the clearing up of perplexities does not depend on our getting further facts but on more careful use of our concepts.’ (Russell 1955, 348) Barnes also admits that ‘Ryle in raising this topic is not claiming to have solved it and wishes merely to be provocative’ (Barnes 1955, 358) Perhaps the conclusion should be that whereas Ryle’s specific discussion of the individual dilemmas was often too brief and superficial to be completely satisfactory, his treatment of the dilemmas did fit his meta- philosophical purpose. Ryle’s distinctive style of presenting numerous everyday life examples and metaphors undoubtedly gave flair to his writings. Although generally appreciated, this style

102

Gilbert Ryle’s Later Writings was sometimes considered to be a disadvantage as well. As Winston H. F. Barnes remarked: ‘Perhaps it is a little overdone. On one or two occasions the hitching of so much theoretical harness made me a little hot under the collar; and the metaphors served rather to distract me with unwanted, though amusing, imagery than to strike out light.’ (Barnes 1955, 364) And according to Bambrough: ‘the manner is sometimes lost amid the mannerisms’ (Bambrough 1994a, 380). The weaknesses of Ryle’s style are expressed more accurately by Erik Götlind: ‘There are risks of letting the analogies be answers of the original questions and for hiding weak points in the reasoning for the reader by knocking him down with a super-drastic expression, just when his reflection begins to protest.’ (Götlind 1956, 70) Sometimes Ryle’s style rather than the content of his arguments is what convinces people; but at the same time this style makes it at times difficult to see the value of his arguments.

‘It Was To Be’ We shall now look at one particular dilemma, ‘It was to be’, in more detail. Ryle seeks to find a way to reconcile our idea that we are responsible for our actions with the thought that ‘whatever is, was to be’. The problem of the freedom of the will already interested him before 1953. In fact, this problem was an alternative project to that which resulted in The Concept of Mind, as Ryle later recalls: ‘For a time I opted for the problem of the Freedom of the Will as the most suitable Gordian Knot; but in the end I opted for the Concept of Mind.’ (Ryle 1970, 12) He chose to discuss this particular dilemma as the first in a series because it is not a burning issue and can therefore be studied like a puzzle without strong passions playing a role and muddying the debate. A second advantage is that it is not a technical controversy; thirdly, only a few concepts are involved which makes the question relatively easy to deal with. What does it mean to say ‘whatever is, was to be’? It says that ‘for whatever takes place it was antecedently true that it was going to take place’ (Ryle 1954a, 15) – henceforth referred to as ‘the principle of antecedent truth’ – which clearly cannot be doubted. The dubious step is the one that leads us from here to the claim that ‘everything, including everything that we do, has been definitively booked from any earlier date you like to choose’ (Ryle 1954a, 15), and subsequently to the Fatalist assertion that ‘nothing that does occur could have been helped and nothing that has not actually been done could possibly have been done’ (Ryle 1954a, 15). This assertion is incompatible with our equally strongly held opinion that we are responsible for our own actions. Exposing the flaws of the argument from the principle of antecedent truth to , Ryle tries to dissolve the apparent conflict between the freedom of the will and the principle of antecedent truth. In order to do so he starts by analysing what is exactly meant by ‘it was true a thousand years ago that a thousand years later these things would be being said in this place’ (Ryle 1954a, 17). It does not mean that there is an identifiable person who knew that these things would be said in this place, that is it is not required that there was someone who at one time had foreknowledge of these things. Neither do we have to suppose that someone did in fact make the prediction. What the above phrase means is something like: ‘if anybody had made a prediction to this effect, though doubtless nobody did, he would have been right.’ (Ryle 1954a, 17) Essential is Ryle’s assertion that what we talk about is not an actual prediction having come true, but a conceivable prediction that would have been made true by a certain event if the prediction had been made. ‘Came true’ and ‘was fulfilled’ apply to predictions actually made, but get us into trouble if we apply them to ‘might-have-been predictions’ – just as ‘might-have-been bullets’ cannot hit targets. Furthermore, since ‘true’ and ‘false’ have the connotation of ‘sincere’ and ‘insincere’ which

103

The Philosophical Development of Gilbert Ryle is improper to the realm of guesses, Ryle suggests to use ‘correct’ and ‘incorrect’ instead, reformulating the principle of antecedent truth into: ‘for any event that takes place, an antecedent guess, if anyone had made one, that it was going to take place, would have been correct, and an antecedent guess to the contrary, if anyone had made it, would have been incorrect’ (Ryle 1954a, 19). Now, what the fatalist in fact does is to assume that since the antecedent truth requires the event of which it is true, this event must be in a way caused by its antecedent truth. Necessity is automatically considered to be causal necessity. The mistake can be exposed by comparing the fatalist’s claim to the claim that ‘for everything that happens, it is true for ever afterwards that it happened’ (Ryle 1954a, 20), which does not trick us into fatalist assumptions. In this case it is clear that the event renders the posterior truths about it true, and not the other way around. In the case of posterior truths we are not tempted to say that they are the causes of the events of which they were true, because causes precede their effects and therefore posterior truths can never cause the events which make them true– it is a matter of the relative dates of event and truth. Ryle thinks that the fatalist’s mistake arises out of an interplay between the concepts ‘event’, ‘before and after’, ‘truth’, ‘necessity’, ‘cause’, ‘prevention’, ‘fault’ and ‘responsibility’. The fatalist goes wrong in trying to apply the inescapability of the conclusions of valid arguments, which is an or logical inescapability, to events, which belong to a different category and are only practically (in)escapable. Our talk of singular statements in the future tense strengthens these fatalist urges. We cannot say of any designated event that it was averted (for instance, this collision was averted) and this leads us to conclude that it is logically impossible to avert any events. Ryle argues that singular propositions do not belong to the realm of statements in the future tense, and that only statements in the past and present tense can be singular. The mere fact that we can ask questions such as ‘Could this collision have been prevented?’ or ‘Could the Battle of Waterloo have been unfought?’ is because they happened – otherwise we would not have been able to talk about ‘this collision’ and ‘the Battle of Waterloo’. It is absurd to ask the question whether the Battle of Waterloo was fought or unfought, since: ‘That it was fought goes with our having an it to talk about at all.’ (Ryle 1954a, 26) Singular propositions do not belong to the realm of statements in the future tense and the fact that we cannot say of any specific event that it was averted is a logical fact about future tense statements which does not have anything to do with an impossibility to keep events from happening. Ryle draws two general points from this dilemma: firstly, that there is no such thing as the method for solving philosophical problems. We have to use a method of trial and error to discover which moves are to be made in this or that specific situation. Secondly, this dilemma illustrates once more that philosophy does not consist in untying logical knots one at the time:

the quandary, though relatively simple, does depend upon a smallish number of concepts, namely, in the first instance, upon those of event, before and after, truth, necessity, cause, prevention, fault and responsibility. Now there is not just one of these concepts which is the logical trouble-maker. The trouble arises out of the interplay between all of them. (Ryle 1954a, 31)

In this dilemma, Ryle reveals the conceptual mistakes that are made when one argues from the innocent principle ‘that whatever is, was to be’ to a fatalist conclusion. He illustrates the complexity of the interplay of different concepts and the need for philosophy to take up the task of exposing ‘speech-faults’ (see above) and showing how seeming conflicts – of a 104

Gilbert Ryle’s Later Writings seemingly non-conceptual nature – vanish if conceptual mistakes are laid bare. No new facts are discovered and no real proofs are given. His treatment of this dilemma is a school- example of what philosophy purports to be, as described in his theoretical papers. Despite Ryle’s witty and original, inventive manner of approaching this dilemma, it may be doubted whether he in fact has solved it. A general point of criticism is that Ryle only disentangled philosophical puzzles without solving the underlying problems (Kaplan 1955, Ambrose 1955). In this case, he may be said not even to have started to solve the philosophical problem underlying the argument from ‘that whatever is, was to be’ to a fatalist conclusion. Further, some more specific points of criticism were directed to his treatment of ‘It Was to Be’ as well. For example, Barnes argues that Ryle never pays attention to the relation between foreknowledge (if possible) and , which seems to constitute an important part of the underlying problem of ‘It Was to Be’, neither does he sufficiently clarify the difference between ‘true’ and ‘correct’. (cf. Barnes 1955) He also seems to use ‘come true’ and ‘was true’ interchangeably without recognising the difference between these notions. In her review, Alice Ambrose, too, suggests that Ryle just solves the rather superficial problem which was caused by a confusion of categories, without solving the underlying problem. In order to avoid the category confusion the problem can perhaps be reformulated as follows: ‘everything there is was from eternity causally necessitated’ is incompatible to ‘we often should do certain things and not others’, while we strongly consider both statements to be true. (Ambrose 1955, 157) Although these criticisms perhaps need not have bothered Ryle too much, for his main aim was a meta-philosophical one, it should be noted that virtually all of his specific solutions to the dilemmas have been disputed, mainly for their superficiality.

An Adverbial Account of Thinking

From the 1950s onwards Ryle wrote papers and lectures on a variety of subjects – e.g. Dilemmas, papers dealing with induction, moral philosophy, the history of philosophy, etc. – but what kept intriguing him most strongly were conceptual puzzles concerning mental concepts. The majority of his later writings is devoted to this theme. As we have already seen, in The Concept of Mind (1949a) Ryle regards mental concepts to be adverbial rather than substantive; they are qualifying and should not be treated as names. Complex descriptions of intelligent actions should be regarded as higher- level descriptions of processes, not as descriptions of complex processes. The mental predicate ‘intelligent’ does not describe an extra process, but rather the way in which all kinds of processes are managed, just as ‘carefully’ in ‘Peter is driving carefully’ does not refer to a process other than driving. Ryle’s later writings on thinking deal primarily with what it precisely means to say that our descriptions of thinking are higher-level descriptions of processes. Ryle usually distinguished between a broad and a narrow concept of thinking. The first includes every type of thinking, ranging from our practical thinking in everyday life to highly intellectual thinking, e.g. in scientific discovery and philosophy. The narrow concept – characterized as what Le Penseur does when he is thinking – seems to include higher intellectual activities such as philosophical and scientific theorizing, and the composition of music, but not practical activities such as intelligent car driving. The thinking of Le Penseur is a mental concept, and hence, according to Ryle, does not describe a process, but rather a higher-level description of processes – i.e. the way in which certain processes are managed. It may

105

The Philosophical Development of Gilbert Ryle perhaps be distinguished from the broader notion in that it does not seem to have a specific, identifiable action to supervene upon. Ryle mainly concentrates on the narrow notion of thinking, since it is the most problematic one. A question that rises immediately is: if the thinking Le Penseur is doing is (merely) the way in which processes are managed, then what kind of processes are we talking about here? In the case of ‘intelligent’ and ‘carefully’ it is clear that there are many different processes which these higher-level concepts can be about, e.g. ‘driving a car’, ‘playing chess’. Whereas someone who drives a car intelligently can be said to perform the actions of driving a car in a special way, this cannot be said of Socrates, Mozart and the scientist. Ryle uses ‘the thinking of Le Penseur’ precisely to indicate this category of thinking of which it is most difficult to grasp that thinking is adverbial and is supervening upon actions instead of itself being qualified as an action. If the ‘thinking’ of Socrates, Mozart and the scientist does not describe a process but a manner of ‘X-ing’, what can this X-ing possibly be? Is it ‘pondering’, ‘reflecting’ or ‘meditating’? What ‘happens’ when we do that? To present an answer to this question – or better: to this group or family of related questions – is the main objective in Ryle’s later papers on thinking. These papers have generally been neglected by later scholars. The only paper which thoroughly discusses the development of Ryle’s later thoughts on thinking is Sibley’s ‘Ryle and Thinking’ (1970) in which Sibley sketches Ryle’s development from a multiple-activity account of thinking to a polymorphous one, and finally to an adverbial account of thinking. According to Sibley, all three positions must be seen in the light of Ryle’s negative project to reject accounts ‘which find thinking to consist in any single ingredient or unitary activity’ (Sibley 1970, 76). The difference between the first two positions is that the polymorphous account is stronger than the multiple-activity account in the sense that in the polymorphous account these activities can also be performed when one is not thinking. What both positions have in common is that ‘thinking comprises a collection of activities with no common strand of importance’ (Sibley 1970, 76). The difference between these positions and the third one is that in his last papers Ryle argues that thinking is not a separate activity (or even group of activities) at all. ‘Verbs of thinking are really “adverbial”; they do not themselves denote activities, but ways or manners in which other, often overt, activities are performed’ (Sibley 1970, 76). According to Sibley the third position was introduced not until ‘Thinking and Reflecting’ (Ryle 1966-67). In addition to Sibley’s paper, I should also mention a short paper by William Lyons, entitled ‘Ryle’s Three Accounts of Thinking’ (Lyons 1980a). Similar to Sibley, Lyons also recognizes three stages in Ryle’s development of the notion of thinking: (1) a mongrel- categorical account of thinking; (2) a polymorphous account; and (3) an adverbial one. Both Sibley and Lyons argue that at the time of The Concept of Mind Ryle did not yet hold an adverbial account of thinking. As has been shown in Chapter 3, however, Ryle already held an adverbial account of thinking in The Concept of Mind (1949a). What he did in the later papers was only sharpening and strengthening his position on thinking, without radically changing his mind. Although he puts it rather negatively, Gareth B. Matthews seems correct, I think, when he argues that the book makes clear that ‘alas, he [Ryle] developed very little new to say on the subject’ (Matthews 1981, 443). In 1949 when Ryle described thinking as ‘mongrel- categorical’ his description included mention of an occurrence or doing and of a disposition, i.e. a frame of mind whereby one is disposed to purposefully, methodologically, attentively, etc. do what it is that one is doing in that frame of mind. Except for the problematic concept of ‘frame of mind’ which Ryle stopped using shortly after 1949, Ryle’s position does not seem to have radically changed between The Concept of Mind and 106

Gilbert Ryle’s Later Writings his last writings on thinking. Problems concerning Ryle’s use of ‘frame of mind’ were put forward by his reviewers. According to MacDonald (1951), it suggested something internal to the mind, which was exactly what Ryle wished to avoid. Ryle may have toned down (or perhaps it is better to say ‘reformulated’) his earlier dispositional account because of these unsolvable problems with ‘frame of mind’ or, as Lyons suggests, because the propositional account (partly due to the notion of ‘frame of mind’) would not have enabled him to succeed in his later project of accounting for original or creative thinking. Although admittedly more prominent in his later writings on thinking, the adverbial element seems to me to have been present already in 1949. In On Thinking Ryle introduces several concepts and connects them to the notion of thinking, thereby making an attempt to pinpoint more exactly what this ‘thinking’ amounts to. The most important concepts and distinctions which Ryle introduces for this purpose are ‘abstract verb’, ‘polymorphous’, ‘higher-level description’, ‘parasitic action’ versus ‘host action’, ‘thin description’ versus ‘thick description’, ‘pathfinding’ (‘trying’, ‘delving’) versus ‘path-following’ (‘borrowing’), ‘imagination’, ‘improvisation’, ‘self- teaching’, and ‘lower-order act’ (or ‘infra-act’) versus ‘higher-order purpose’ (or ‘supra- policy’). We shall see that the notions of ‘abstract’, ‘higher-level’, ‘parasitic’, ‘thick description’, higher-order purpose’ and ‘supra-policy’ are all used with the same purpose and are more or less identical. Basically, they are all used to point at one and the same distinction; the only difference lies in the manner in which this distinction is presented to us. Using the ‘higher-order’ concepts (as we may call them) Ryle tries to show – more explicitly and also more accurately than in The Concept of Mind – that a description of the thinking of Le Penseur is not a description of complex processes, but rather a complex (or higher-order) description of processes. This enables him to explicitly reject both dualism and behaviourism, carefully steering between these views which both – mistakenly in Ryle’s view – take for granted that all descriptions are treated as if they were names and directly refer to processes or things, rather than that some descriptions do not function as names at all, but are higher-order descriptions. In what follows I shall discuss Ryle’s position and use of concepts in more detail. It is interesting to see how he tries to avoid the main pitfalls of The Concept of Mind, e.g. the suggestion of behaviourism. I shall pay attention to the strong as well as the weak points of his treatment of the notion of thinking, mainly by reference to how it was received by his contemporaries.

The Broader Notion of Thinking and the Problems with the Thinking of ‘Le Penseur’ To start with, it is important to look at Ryle’s ideas on the general notion of thinking, not just because it is interesting in its own right, but also because it helps to show why Ryle had difficulties with the notion of thinking of Le Penseur: prima facie it is hard to explain how this type of thinking fits the criteria Ryle formulated for thinking in general. A more detailed explanation of the nature of thinking in general provides clues for solving problems with respect to the more restricted notion of the thinking of Le Penseur. Moreover, in ‘Thinking and Reflecting’ (Ryle 1966-1967) Ryle intends to show that the thinking of the tennis-player or the person who carefully drives his car is the basic notion, the thinking of le Penseur being secondary. Therefore, this narrow notion of thinking needs to be explained via the more general, more basic, notion of thinking. A difficult question is what thinking (in general) is. An answer that Ryle quickly dismisses in ‘Thinking thoughts and having concepts’ (1962c) is ‘having-concepts-in-

107

The Philosophical Development of Gilbert Ryle mind’. Even if there were incidents of ‘having-concepts-in-mind’ – but Ryle claims that there are not – simply put together they would still not amount to thinking. His concept of thinking amounts to acquiring concepts; he focuses on the notion of thinking as travelling, rather than focusing on thinking in the sense of the result. In his later writings Ryle keeps paying attention to two of the basic ideas of The Concept of Mind: firstly, that the existence of a particular homogeneous X-ing is not necessary in order for a person to be thinking; secondly, that the concept of thinking is adverbial rather than substantial. Our inability to grasp the second often makes us reject the first. In ‘Adverbial Verbs and Verbs of Thinking’ (1979a) Ryle claims, as he already did in The Concept of Mind with respect to the broader notion of thinking, that there is not one homogeneous X-ing that necessarily exists for a person in order to reflect, meditate or ponder. And in ‘Thinking Thoughts and Having Concepts’ (1962c) he argues that to point at what the different kinds of thinking have in common ‘is not to point to a separately recollectable experience or a separately performable act’ (Ryle 1962c, 448-449). In ‘Thinking and reflecting’ (1966-67) he provides one of the most important reasons why we often believe there to be such an X-ing after all. We often assume that all active, tensed verbs are verbs of performing an action. The fact that we often treat ‘think’ in this manner explains our search for an homogeneous X-ing. However, ‘think’ should not be treated as a verb of doing any more than ‘perish’, ‘inherit’, ‘sleep’, ‘know’ and ‘forget’. Thinking is not an autonomous action and the verb ‘to think’ needs a proper verb to qualify, which is why it should be treated as an adverbial rather than an active verb, analogue to the verb ‘to hurry’.

(…) to obey or disobey the command to hurry, I must do some autonomous X, like eating or humming, etc., for there to be a hurried or an unhurried X-ing that tallies with or flouts the understood command (…). Trying, scamping, succeeding and failing are, in generically similar ways, not autonomous activities. There must be an X-ing, if there is to be successful, unsuccessful, difficult or easy, industrious or scamped X-ing. (Ryle 1966-1967, 467)

Just as driving carefully is not performing two actions, as driving while singing a song is, neither is X-ing hurriedly. ‘To think’ is also to be classified as an adverbial verb, not being an autonomous action, ‘nor a concurrent procession of autonomous anythings’ (Ryle 1966- 1967, 469). No one homogeneous X-ing needs to be going on for a person to be thinking. In ‘Thinking and Language’ (1951a) this view on thinking is used to explain why our reportings of our thinkings are graphic – for instance ‘go round in circles’, ‘grasp’ – as we do not seem to be able to find ungraphical, literal idioms. Literal, concrete idioms cannot be used precisely because there is no one homogeneous X-ing underlying all thinkings.78 At this stage it is important to pay attention to Ryle’s use of the notions of ‘autonomous’ and ‘homogeneous’. Sometimes it seems as if Ryle is using them interchangeably, when he claims that thinking is not an autonomous doing and is not, or does not require, an homogeneous X-ing. He uses his notion of thinking as non-autonomous and non-homogeneous with the purpose of showing that thinking is a heterogeneous higher- order concept. Although closely related, there is a difference between Ryle’s use of these two concepts. When he claims that thinking is not an autonomous doing he means to show that, as in the case of ‘hurrying’, we are in need of another action, of a lower level, which is autonomous. In order to be able to hurry at all, I must do some autonomous X-ing. Similarly, in order to think at all, I must do some autonomous X-ing. At the end of the chain (from higher-level to lower-level) there must always be at least one autonomous X-ing. Ryle’s thought that thinking is non-homogeneous refers to the idea that there is not one

78 Cf. Ryle 1958 (pp. 399-400). 108

Gilbert Ryle’s Later Writings specific autonomous doing which is necessarily at the end of the chain in the case of thinking. Many different types of autonomous actions could do, for example one can be pronouncing syllables while thinking or moving chess pieces while one is thinking, which is why he calls thinking ‘heterogeneous’. So, because thinking is not itself an autonomous action and there are many possible autonomous actions at the bottom, the concept of thinking is heterogeneous rather than homogeneous (‘smoking’ would be an example of a homogeneous verb because there is only one autonomous action at the bottom of the chain, namely smoking). Introducing the terminology of ‘parasites’ and ‘hosts’, and ‘thickly’ and ‘thinly’ described actions in ‘Thought and Soliloquy’ (1979b) and ‘Thinking and Reflecting’ (1966- 67), Ryle makes an attempt to sharpen his account of the higher-order character of thinking. If we take the example of someone practising jumping or parodying jumping, the thin description entails his ‘bare’ doings, whereas the thick description also includes his . The thin description of bare jump-doings enters into each of the thick descriptions (e.g. practising jumping and parodying jumping) without exhausting them. Moreover, the thick descriptions stand to the thin ones as parasites stand to hosts. We have to know what jumping is before we can grasp the idea of someone practising or parodying jumping. Yet, without the thick description which includes the intentions, the thin descriptions alone do not amount to thinking. We may add further descriptions, further layers; take for example a boy who is rather clumsy in winking: the thick description of what a boy is doing who is parodying the one who is winking in a clumsy way, adds another layer. One more layer is added when the boy is described as practising parodying the other boy.79

The thinnest description of what the rehearsing parodist is doing is, roughly, the same as for the involuntary eyelid-twitch; but its thick description is a many-layered sandwich, of which only the bottom slice is catered for by that thinnest description. Taking the word ‘only’ in one way, it is true enough that the rehearsing parodist is, at this moment, only contracting his right eyelids. Taken in another way, this is quite false; for the account of what he is trying to effect by his eyelid-contraction, i.e. the specification of its success-conditions, requires every one of the successively subordinate ‘try’ clauses (…). (Ryle 1968a, 482-83)

‘Practising’ and ‘preparing’ form a category of thinking with problems of its own, since practising an action, e.g. a chess-move, is not yet performing the action itself. The question rises what someone who is practising or preparing an action is doing. Again, it is in the separation of parasite-actions from host-actions that Ryle expects to find the defining characteristics of practising something, e.g. boxing, or preparing something, e.g. an oration.

It is the very fact that the solitary shadow-boxer is not now boxing against an opponent and yet is purposefully and self-coachingly doing something that forces us to find a name and a home for the something that he is now engaged in, and for what the composing or the rehearsing orator, musician or conjuror is now engaged in, and for what the undecided chess-player is now engaged in while tentatively making dummy-moves with his queen. All of them are meditating – meditating actions which they are not yet performing, and which, when they are being

79 In the paper ‘Courses of Action or the Uncatchableness of Mental Acts’ (Ryle 2000), Ryle tries to explain the nature of our ponderings by introducing the concepts ‘infra-act’ and ‘higher-order purpose’, which function in the same way as ‘thin description’ and ‘thick description’. Infra-acts can be regarded as actions, whereas higher order purposes cannot. ‘Exploring is conducting ten thousand variegated infra-acts with one governing and complex supra-policy or Higher Order purpose.’ (Ryle 2000, 336) These infra-acts could all on their own have been prosecuted by persons with other higher order purposes, or by persons without any higher order purpose.

109

The Philosophical Development of Gilbert Ryle

performed, will not themselves be bits of meditating, that is they will not themselves be parasite- actions but host-actions. (Ryle 1979b, 38)

But what is this ‘something’ we are hoping to find a name for? What actions are the shadow-boxer and the rehearsing orator performing which can function as the hosts on which their ‘shadow’-actions are parasitic? Ryle’s explanation is that the host-actions are the ones they are not yet performing (fighting a boxing match or giving an oration) and the parasite-actions are the ones they perform in order to perform their host-actions well, e.g. making dummy-moves. The success of their parasite-actions, which are of a higher order than their host actions, can solely be measured by the success of the host-actions. Thus, according to Ryle, mental verbs are adverbial and stand, as it were, as parasites to hosts in the way in which ‘to practice jumping’ stands to ‘jumping’. But what consequences does this have for the thinking of Le Penseur? To which thin description does the thick description of the thinking of Le Penseur stand?

The Thinking of ‘Le Penseur’ We have distinguished two different notions of thinking: a more general notion and the thinking of Le Penseur. But in what sense precisely do they differ? Ryle argues against the idea that these types of thinking differ merely in complexity. In the case of the tennis-player who is trying to win the game, the tennis-player is concerned with his play, but not with his thinking how to play. He may blame himself for playing carelessly, but not for ‘thinking carelessly how to play’ (Ryle 1962a, 428). In contrast, the more specific notion of thinking as reflecting is thinking in which the thinker is concerned to think properly. And this thinker’s findings are ‘disengaged from any particular urgencies’ (Ryle 1962a, 429) What he discovers is neutral between the different practical purposes his discoveries may (later) appear to have – in a way he is exploring for its own sake. And while the tennis-player is more successful upon being more responsive to what happened outside him, Le Penseur’s concentration reveals itself in not paying attention to the outside world.

Both are absorbed, but the tennis-player’s absorption is in his and his opponent’s momentary playing, while le Penseur’s absorption is in something detached from the rock-squatting that he is momentarily doing, and the rain-drops that are momentarily wetting him. His quick and appropriate responses to what occurs around him on the tennis-court show that the player is concentrating. His non-responses to what occurs around him show, or help to show, that le Penseur is concentrating. (Ryle 1966-67, 466)

Whereas some of the actions of the tennis-player cannot be done without concentrating, it seems that le Penseur is not doing anything except tackling his problem. And he is detached from the other things he may be doing, such as breathing or scratching his head. Ryle labels this disengaged or detached thinking ‘reflecting’. Le Penseur is – unlike the tennis-player – completely, or at least nearly completely detached from external circumstances.80 What is important according to Ryle is not to characterize someone who is merely composing a speech or translating a passage as Le Penseur. This Thinker with capital ‘T’ should not just be thinking in the sense of trying to decide only what to say and how to say

80 Ryle does not wish to go so far as to consider any dependence on external circumstances as a disqualification of thinking as the thinking of le Penseur. Someone who is drawing dots and lines on paper depends on external circumstances for his possession of pen and paper as well. However, external circumstances do not force this person into drawing specific dots and lines rather than others. 110

Gilbert Ryle’s Later Writings it but he should utmost and for all be trying to make up his mind what to think. His primary task is not rhetorical, although it is not totally devoid of . Ryle presents ‘trying to decide what to think’ as a difference between Le Penseur (that is, an intellectual explorer, for instance, Euclid, Sherlock Holmes, Gibbon and Kant) and the composing orator. The first is working on a higher or deeper level and if successful ‘will have found something that, thenceforward, they and their hearers or readers will know, which they did not know before, but needed to know.’ (Ryle 1979b, 46) Connected with this are Ryle’s views on what the thinking of Le Penseur has to do with imagination and improvisation. Ryle argues against the idea that a person’s mind is divisible into faculties, the intellect being one such department and the imagination another. He thinks that imagining is neither an activity to be contrasted with thinking, nor a species of thinking. It is rather the innovative, risk-taking aspect of thinking.81 An important feature of the thinking of Le Penseur vis-à-vis the thinking of a more general kind is its explorative character. Ryle explains this in his remarks on teaching, learning and improvising. A pupil, for instance, qualifies as thinking if he makes his own applications and misapplications of the lessons of his teacher in new tasks. ‘Thinking is trying to better one’s instructions; it is trying out promising tracks which will exist, if they ever do exist, only after one has stumbled exploringly over ground where they are not.’ (Ryle 1972a, 78) A thinking pupil is ‘experimentally playing himself with might-be cues, clues, reminders, snubs, exercises, spurs, and the like, of types which are sometimes or often employed experimentally by teachers who are teaching what they do know.’ (Ryle 1972a, 67) So, the difference between teaching and thinking is that the teacher often knows what he tries to teach his pupil, whereas Le Penseur is thinking just because he does not know what he wants to know. Sometimes, however, from earlier experiences he does have an idea about which directions are more promising, what makes his attempt not just a matter of guesswork. He poses his questions experimentally, in order to find out whether they are the right questions to pose, or, in other words, whether they will be heuristically rewarding. The thinking of Le Penseur involves finding out by trial and error which questions are the best ones to ask. What Le Penseur is trying to do in saying things to himself is trying, by success/failure tests, to find out whether or not the things that he is saying would or would not be a ‘guiding track’ (Ryle 1968a, 494). His thinking is a trying- out: ‘when something is obvious no thinking, no task is required. But thinking is required when things are not obvious.’ (Ryle 1964, 148) Improvisation is essential to Le Penseur’s thinking. Ryle wants to reject our everyday ‘step-after-step’ picture of thinking which does not leave room for improvisation in thinking. Sometimes it is suggested that thinking is a neat step-by-step procedure like the presentation of a finished argument. However, we have to keep in mind that ‘what passes through the barrister’s mind when thinking out his argument is nothing like the argument he presents to the court’ (Ryle 1964, 163); ‘it is a baseless dream that our ponderings can be described in the idioms used in describing the results of our thinkings’ (Ryle 1964, 160).

81 Describing both an imaginative biographer and an unimaginative one, he argues that not all thinking deserves the label ‘imaginative’. The unimaginative biographer and the imaginative one properly employ certain concepts alike, for instance ‘heavy rain’. However, the unimaginative biographer starts and ends with these concepts whereas the imaginative one proceeds to a higher level and actually starts to work with them; he, for example, starts to think, after finding records of heavy rain, about the depth of the mud and the belatedness of the mules. (Ryle 1979c, 59).

111

The Philosophical Development of Gilbert Ryle

Thin and Thick Descriptions, Hosts and Parasites: Applied to the Thinking of Le Penseur We have seen that thinking in a more general sense can be explained in terms of thin and thick descriptions and in terms of hosts and parasites. What about the more restricted notion of the thinking of Le Penseur? What can be considered as a thin description of what Le Penseur is doing? And is it possible to describe the activities of Le Penseur in terms of hosts and parasites? In ‘Thought and Soliloquy’ Ryle introduces ‘saying so-and-so’ as a possible host on which the thinking of Le Penseur parasites. ‘Saying so-and-so’ can, just as jumping, function as a host to parasites of different levels of sophistication. The description of a child trying to echo ‘the mouse ran up the clock’ parasites on ‘the child uttered “the mouse ran up the clock”’. At a higher level of sophistication the child for example wants to be original in telling the story of the mouse and is disappointed upon finding out that his hearer heard it before. The same utterances function as the host: ‘This crude ladder of sophistication-rungs is a crude ladder of parasitic sayings that are all, in different ways, parasitic on the host saying of the phonemes “the mouse ran up the clock”.’ (Ryle 1979b, 49) ‘Saying so-and-so’ is, according to Ryle, just one optional host on which the thinking of Le Penseur parasites, but there are more. Not all thinking requires saying things to oneself; no homogeneous X-ing, such as for example saying things to oneself, exists which is the host on which ‘thinking’ parasites. Tackling philosophical problems usually does require saying things to oneself, but the thinking of Le Penseur comprises more types of thinking than merely this one. The thinking of Mozart, for example, is not parasitic on Mozart saying things to oneself. But what is at least clear is that the question what Le Penseur is doing is a question about the sophistication-level of the doing, for instance of the saying, which he is engaged in. Although, according to Ryle, no one specific autonomous X-ing exists for the thinking of Le Penseur, purely as a matter of logic, we do seem to need a non-adverbial – or at least partly non-adverbial – verb for the adverbs to attach to. And although adverbial verbs have the characteristic that they may pyramid, in the sense that the one adverbial verb can be parasitic on the other, the bottom one needs to be attached to a (partly) non-adverbial verb. For example:

If I am eating my breakfast, you may tell me to hurry over my breakfast. If I obey you, I do so not by breakfasting, since I am doing that already, but by accelerating the rate of my breakfasting. I am then obediently hurrying over my breakfast. If I resent your command but dare not disobey it, then I am with reluctance obeying your command to hurry over my breakfast. I am reluctantly obediently hurriedly breakfasting. I am not reluctantly breakfasting, nor necessarily reluctantly hurrying over my breakfast; I am with reluctant obedience hurrying over my breakfast, though I might have cheerfully hurried over it if you had ordered me not to do so. And so on, in principle, indefinitely. (Ryle 1966-67, 473-74)

In this example the bottom adverbial verb is attached to the non-adverbial verb ‘breakfasting’. There is not one specific verb to which ‘hurrying’ is always attached; like ‘thinking’ the adverb ‘hurrying’ can be attached to many different non-adverbial verbs, as long as it is always connected to one. It is important to realise that the verb ‘to say’, contrary to what is commonly thought, cannot function as the bottom-level-verb, that is the thinnest action-report, of the thinking of Le Penseur. ‘To say’ is an active verb not used for the bottom-level; at a lower level does something like ‘to voice syllables’ come. (see Ryle 1968a, 487) Le Penseur’s saying things to himself (if he is) is therefore not the thinnest action-report of his doings, which would be something like ‘murmuring syllables under his breath’ (Ryle 1968a, 487).

112

Gilbert Ryle’s Later Writings

Thus, the conclusion is that the thinking of Le Penseur, like all adverbial verbs, does need a non-adverbial verb. However, which verb it is that takes this place depends on the situation; there is no such thing as one homogeneous non-adverbial verb which is always attached to the thinking of Le Penseur and without which the person in question is not thinking in the sense Le Penseur is thinking. In some, but not other, circumstances the non-adverbial verb is something like ‘murmuring syllables under his breath’. We have to keep in mind though that a proper characterization of Le Penseur mentions what he was trying to achieve while murmuring those syllables, rather than to describe this murmuring itself. The (possible) sayings of Le Penseur are propositions with a heuristic point, rather than remarks with a conversational point – they are steps for him to get somewhere but they do not function as necessary steps for possible hearers. They are part of what Ryle calls ‘pathfinding’ (Ryle 1972b, 92), rather than of the process of path-following.82 ‘Thinking, then, can be saying-things-tentatively-to-oneself with the specific heuristic intention of trying, by saying them, to open one’s own eyes, to consolidate one’s own grasp, or to get oneself out of a rut (…)’ (Ryle 1972b, 92). Le Penseur is not just saying things to himself, but he is saying things to himself ‘with a special governing purpose, with a specially directed vigilance, resolution, interest, readiness for failure, and so on’ (Ryle 1972b, 92). Thus, thinking is not just saying things to oneself, nor is it saying things to oneself and doing something else as well. It is saying things to oneself with an experimental intention, ‘experimentally’ not adding an extra action but rather the ‘specific-intention-to-find-out- what-happens-when-the-tap-is-turned’ (Ryle 1972b, 92).83 This is also the reason why in ‘Thinking and Language’ – a paper which he presented in 1951 at the symposium ‘Thinking and Language’ of the Aristotelian Society – Ryle rejects the idea that thinking is to be equated with using ordinary language; it is rather using ordinary language ‘with a special governing purpose’.

What a thinker is doing with (or about or to) the expressions that occur makes all the difference; and there are scores of widely different things that he may be doing with (or about or to) them. Until we have heard his economic story, his technical story, his tactical story, or his artistic story, we do not know what he was trying to do or what he accomplished or failed to accomplish. And if we have not got its plot, we have not got any part of the story but only a welter of, so far, pointless details. A pure chronicle of the occurrence of expressions (or of images, hummed notes, glimpses or fingerings of plasticine) would not yet be even the beginning of a history of the thinker’s pondering.84 (Ryle 1951a, 265)

We do not want a detailed description of our simplest utterances, which would amount to sequences of letters or sounds; what we need is ‘idioms which, while neutral between the disparate details of the activities, are appropriate to the similarities and differences between

82 This distinction was already introduced in The Concept of Mind (1949a) where it was described as the difference between path-making and path-using (Ryle 1949a, 273). 83 The fact that thinking in general requires saying things to oneself does not mean that what is being done is thinking in these words and phrases. In fact, this claim would be absurd. How could someone who is preparing a speech possibly have thought in the phrases and sentences that finally constitute his speech while still composing his speech? Similarly, it does not make sense to say that the translator who is trying to render into German a passage from Gibbon’s Decline and Fall of the Roman Empire is thinking in Gibbonian English or would-be Gibbonian German words, phrases and sentences. (Ryle 1979b, 41-42) Cf. Ryle 1951a. 84 Ryle distinguishes between three different accounts of thinking, as an alternative to the three levels of thinking put forward by A. C. Lloyd, who would later become Professor of Philosophy at the University of Liverpool and was a scholar in late ancient philosophy, at the symposium ‘Thinking and Language’ (1951a). The first level is the chronicles account, that is, an account of the things that went on (without anything like a plot); secondly, there is the historian’s account, which is an account of the plot, often by using graphic idioms, neglecting most of what happened; thirdly, the scorer’s account, which is the outcome or results of the thinkings.

113

The Philosophical Development of Gilbert Ryle the policies, techniques, and clumsiness of the agents’ (Ryle 1951a, 267) Accordingly, in 1958 Ryle describes the reporting of one’s thoughts as not being a matter of merely chronicling internal, simple, items; but rather as reporting the complex items such as a man’s thought of these roses as his wife’s favourite flowers, or of the Oxford rowing crew as insufficient to beat Cambridge. But what is this ‘as’?

To describe someone as thinking of something as so and so is to say of him, at least inter alia, that it would be natural or easy for him to follow up this thought in some particular direction. (…) A person who thinks of something is, ipso facto, primed to think and do some particular further things; and this particular possible future that his thinking paves the way for needs to be mentioned in the description of the particular content of that thinking. (…) Roughly, a thought comprises what it is incipiently, namely what it is the natural vanguard of. Its burthen embodies its natural or easy sequel. (Ryle 1958, 403)

One of the ways to achieve this preferred way of talking about thinking is using non-literal, metaphorical expressions. Thus, a description of our thoughts cannot be done merely in bottom-level sequences but needs higher order concepts too. ‘What qualifies an undertaking as one of pondering or, not very differently, as one of discussing, is not any catalogue of simple qualities and simple relations, whether rude or refined, but some nexus of statable because statement-shaped conditions.’ (Ryle 1972b, 92-93)

Steering Between Dualism and Behaviourism The misleading interpretation of The Concept of Mind as a behaviouristic work led Ryle to be more explicit about his rejection of both Cartesian dualism and behaviourism, especially since by now he fully realized that both make the same type of mistake. Dualists duplicate whereas behaviourists reduce, but – as Ryle now recognized – they do so on similar grounds. They both assume that mental verbs (for instance, calculating, deliberating, etc.) behave similarly to, for instance, verbs like eating and typing. Both feel the need to create extra events; the dualists create events which exist in a second, separate ghostly world, whereas the behaviourists fully reduce the mental verbs to descriptions of events in the material world. Ryle considers both to be mistaken and, as must be clear by now from the discussions of his notion of thinking above, on several occasions he tries to make this point: ‘It is a gross error in the one direction to say that Le Penseur is, for instance just saying things. It is a gross error in the other direction to say that since he is not just saying things, he is therefore doing something else as well’. (Ryle 1979b, 49 and Ryle 1972b) This is a useful qualification of Ryle’s statement in The Concept of Mind that our descriptions of intelligent behaviour, such as driving carefully, describe one action; not two, which still leaves room for behaviourism and sometimes perhaps even suggests it. If we realize that mental verbs are of a different, more abstract level than eating and typing, there is no need for committing ourselves to either reduction or duplication. And this lesson is precisely what Ryle wants to inculcate in his later writings on thinking. In ‘Thinking and Saying’ (Ryle 1972b) Ryle rejects one central point both reductionists and duplicationists agree upon, that is, that the actions of a mimic can only differ from the actions of his victim by the former’s performance of extra actions which the victim did not perform. The fact that no such actions are witnessed leads the reductionist to ignore the different intentions and skills of the mimic, whereas it leads the duplicationist to believe that these differences must consist in extra, unobservable actions. One of the things Ryle tries to show is that actions often are not, in the strict sense, observable. For instance, a referee cannot strictly see that the player has scored a goal, since scoring a goal is not just kicking the ball between two posts. Scoring is kicking a ball between two posts when

114

Gilbert Ryle’s Later Writings several complex conditions or rules are satisfied. Neither is scoring kicking the ball between two posts and performing an additional action. It is rather performing one action which is of a higher order than merely kicking a ball between two posts. What is higher order about it is that the ball is kicked between the two posts satisfying complex conditions and rules. Similarly, a penny is neither merely a metal disc, nor is it two articles – a metal article and a non-metallic article. It is a disc that is qualified for some quite specific sorts of transactions. We need higher order concepts to describe these qualifications: ‘the formulation of these qualifications would require not just some simple auxiliary nouns, simple adjectives, or simple verbs but a whole batch of syntactically variegated subordinate clauses.’ (Ryle 1972b, 88) In the same way as scoring is neither just kicking the ball between two posts nor performing the kicking-action plus an additional action, and a penny is neither just a metal disc nor a metal disc and some additional thing, so a word is neither just a noise, nor is it a noise and something else as well.

The word is not a noise and something else as well; and it is not just a noise. It is a complexly qualified noise, a noise endowed with a quite specific saying-power, endowed sometimes by institutional regulations, generally by accumulating public custom, slightly rigorized by pedagogic disciplines; and so on. (Ryle 1972b, 88)

Thus, Ryle is neither a dualist nor a behaviourist, but what we may term a ‘conceptual pluralist’. Concepts are not to be treated in one single way, or even in two different ways, but in many different ways all depending on context. Concepts do not link to or represent our states of affairs in one or two uniform ways or on a limited amount of levels, but they are of multiple representational kinds. The assumption of one or two different types of concepts should be abandoned altogether.

Pitfalls in the Experimental Investigation of Thinking For Ryle, this philosophical analysis of thinking implies a rejection of a scientific- psychological investigation of it. In his paper ‘Thinking’ he argues explicitly that there is no place for a scientific approach to or a scientific method in exploring the nature of thinking. This brought him into conflict with Professor George Humphrey (1889-1966) who at the time was director of the newly founded Oxford Institute of Experimental Psychology, and had been working with Wilhelm Wundt in Leipzig. Let us briefly look at this debate. Ryle argues that ‘the experimental investigation of thinking had been, on the whole, unproductive, because the researchers have had confused or erroneous notions of what they were looking for.’ (Ryle 1953a, 300). These researchers looked in vain for things such as ideas or images which philosophers had assumed in their theories. Philosophical confusions had misled experimental researchers. According to Ryle, ‘there are no ingredient activities common and peculiar to (…) thinking’ (Ryle 1953a, 296), and this can only be showed by conceptual rather than experimental method. Humphrey responded to Ryle’s points in the same issue of Acta Psychologica. While agreeing with Ryle’s point of view that ‘many different activities are classed as thinking’ (Humphrey 1953, 199), as well as with the idea that the experimental investigation of thinking has been rather unproductive so far, Humphrey questions Ryle’s conclusion that therefore there is no place for an experimental method in the research on thinking. He argues that we need to do experiments since in this way we can prove the

115

The Philosophical Development of Gilbert Ryle falsity of apparently obvious things, as has been frequently done in the past. While agreeing with Ryle that ‘thinking (…) uses various techniques, and different operations are involved in it’ (Humphrey 1953, 199), Humphrey wants to claim that ‘our purpose of scientific investigation is exactly to discover whether they all involve common principles or not; that is to say, to find theoretical rules by which the practice of thinking works’ (Humphrey 1953, 199). Behind this difference of opinion lies a difference in their views on science and philosophy. Ryle would reply that ‘to find theoretical rules by which the practice of thinking works’ is impossible before philosophy first defines the concepts which are relevant, for instance the concept of ‘thinking’. And this project of defining thinking constitutes, according to Ryle, the main problem. Science will never be able to determine by means of scientific research whether ‘thinking’ denotes in the way ‘Fido’ denotes Fido. The (stock) use of concepts is the domain of philosophy and it partly sets the conditions for scientific research, determining for example the specific category the instances of which are considered to qualify as ‘thinking’.

Criticisms Since Ryle’s later papers are neglected in the literature, I have chosen first to describe rather than critically discuss his views. This neglect may be due to the fact that he continued on the same tracks after The Concept of Mind. Although Ryle refined his thoughts, made several important qualifications and later changed his focus to the thinking of Le Penseur, his method remains largely the same. He continued to regard philosophical problems as conceptual tangles which could only be dissolved by conceptual tools, and started to repeat himself, inventing ever new examples to illustrate his views on philosophy. But cognitive scientists and philosophers were interested in discussions on a different level; they were interested in the contents of theories and ideas on thinking rather than in meta- philosophical discussions. The lack of attention for Ryle’s later writings may also have resulted from the critical attitude towards post-war analytic philosophy in general. As Peter Hacker has suggested, there were various objections to analytic philosophy, for example its neglect of metaphysics, its defence of common sense and its commitment to the paradigm case argument. Other objections against post-war connectivist analysis are:

first, that (…) the subject-matter of philosophy is ordinary language rather than the nature of things; secondly, that it held that the problems of philosophy arise exclusively from ordinary language or from the ordinary use of words; thirdly, that it invites investigations of language for their own sake, investigations that belong more properly to linguistic theory than philosophy; fourthly, that it encourages philosophical (…), the view that (…) almost every philosophy is really right, inasmuch as it brings out some kind of insight. (Hacker 1996, 232)

Critics of a Rylean type of philosophy objected to the view that ordinary language cannot be improved upon and to the view that this ought to be the language in which science and mathematics are to be conducted. They also protested against – what they considered to be – a dogmatic defence of common sense against advanced science. Ryle was frequently criticized by his reviewers on these accounts. Disagreement with his meta-philosophical views seems to underlie at least some of their comments. Sibley, for example, criticizes Ryle for not looking for any candidates of essential ingredients of thinking in his later papers on thinking. (Sibley 1970, 82-3) But Ryle’s aim, we may say, was not of such a nature – to look for essential features of thinking – but rather

116

Gilbert Ryle’s Later Writings of a linguistic-conceptual kind: what we mean by thinking. His purpose is not ontological but concerns the meaning, or use, of our words. And the way in which we use ‘thinking’ – according to Ryle – does not require, or even suggest, even one essential ingredient of thinking. Relevant for philosophical purposes is not whether or not thinking is something but that ‘thinking’ has a meaning, that is, use. This is why Ryle persuades rather than proves, and why he uses his everyday life examples rather than scientific theories, and why he talks about concepts rather than brain processes. Sibley and others also expressed criticisms similar to those that were given by the reviewers of Dilemmas: ‘Unfortunately, the vagueness and variety of his locutions make it impossible to say what exactly his view is, while providing some easily demolished straw men.’ (Sibley 1970, 98) And in 1979, in his review of On Thinking Bernard Williams claims that:

many of the arguments here fail, because they rest on no coherent conception of the relation between mental phenomena and the language that describes them. (…) The mannerisms seem to have provided a substitute for the theoretical backing which he [Ryle] was so reluctant to give his arguments. (…) His (…) philosophy came increasingly to depend on an idiosyncratic . (Williams 1979, 6)

He holds that Ryle’s later papers are rather ‘mannered, empty and unconvincing’ (Ibid., 6), also compared to his earlier papers. Sibley draws attention to what is often considered to be a problem for Ryle:

the occurrence, and manner of occurrence, of my relevant and sequential x-ings seem to be the outcome of my trying, controlling and guiding; and the continuance or dismissal of my x-ings is the result of my assessings of what I call up. It therefore sounds decidedly odd to try to equate or identify these tryings and controllings, which are responsible for and explanatory of the manner of x-ing in certain circumstances, with the manner of x-ing in those circumstances. (Sibley 1970, 89)

Ryle’s reply would probably be that this is to mix up thinking in the sense of the result, i.e. a thought, and thinking in the sense of a trying to achieve something, of ‘path-finding’. He might also have wanted to argue that this problem only exists when one employs a causal- mechanistic view on thinking. Since thinking is a higher order concept, the simple terminology of cause and effect will not do. Gareth Matthews, in his review of On Thinking, blames Ryle for not taking into account recent views on thinking in philosophy and cognitive science:

perhaps the most dated feature of these discussions is their total innocence of the possibilities of computational functionalism as a sophisticated philosophy of mind (…). And his only consideration of , an exciting topic in very recent , comes in his dismissive treatment of (alleged) introspectibles. (Matthews 1981, 444)

He was right that Ryle neglected these views when writing his later papers on thinking. However, what Matthews did not know, and probably could not have known, when he wrote his review in 1981 is that Ryle had written a letter to a few months before his death on 22 February 1976, in which he did discuss Fodor, Dennett and mental representationalism (Ryle 1976d). This letter was part of the correspondence between Dennett and Ryle between 1968 and Ryle’s death in 1976, which mainly concerned personal and non-philosophical issues. This last letter was an exception. Ryle comments on Dennett’s review of Fodor and

117

The Philosophical Development of Gilbert Ryle

Dennett’s papers ‘Law of Effect’ and ‘Toward a Cognitive Theory of Consciousness’. About the third he says that he hopes – but doubts – that ‘“inner life” for you [Dennett] just equates with what is done by the adverbs “consciously”, “wittingly”, “guiltily” etc., & has no special tie to what now and then goes on “in my head”.’ (Ryle 1976d) And Ryle’s remarks on ‘Law of Effect’ express his appreciation for the fact that Dennett avoids both reductionism and duplicationism. His comments on Dennett’s review of Fodor express Ryle’s incomprehension of internal representations, or at least of their relevance to thinking (in the sense of trying to get somewhere). Ryle critically wonders what these representations would be representations of, for example when one tries to solve a puzzle, as well as what ‘internal’ here means. Is it literally internal or merely metaphorically? If the latter, it does not seem to add anything; if the former it seems to be ‘a bogus notion’ which is simply postulated. (Ryle 1976d) Not all scholars consider Ryle to be outdated and unfit to stand up to computationalism, functionalism (or a combination). In a recent paper on Ryle’s Le Penseur, Jeff Coulter, for example, defends the view that:

(…) in laying out many of the dimensions of the geography of these concepts, his [Ryle’s] work can today still be appealed to profitably in resisting the temptations of computationalist and functionalist of mind. Far from elucidating the nature of the phenomena of thinking and of thought, contemporary analogies with computing machinery only obscure from us the rich multi-facetedness of these aspects of our lives which Ryle so powerfully illuminated. (Coulter 2003, 78)

According to Coulter – as opposed to someone like Matthews – Ryle does have a role to play in philosophy of mind and cognitive science of the twenty-first century.

Concluding Remarks

Because the theoretical claims which Ryle had presented in his earlier papers could not fully carry the weight of the method of The Concept of Mind (1949a), in the late 1950s and early 60s Ryle wrote additions to his ‘old’ theoretical account of how to do philosophy. Ryle also wrote Dilemmas (1954a) in which he attempted to show how we are prone to be misled by ‘the delusions’ of language. He continued to reject the idea of a denotational theory of meaning, preferring a heterogeneity of use, and still focused on informal logic and the idea that philosophical problems are of a different nature than scientific ones and cannot be solved piecemeal. Newly introduced elements in Ryle’s theoretical papers were his instructions for interpreting ‘the ordinary use of the expression x’, namely that ‘use’ should be interpreted in a logical rather than in a philological sense and that ‘the ordinary use of an expression’ is not the same thing as ‘the use of ordinary expressions’. Furthermore, Ryle found the need to distinguish between language and speech and between words and sentences. Language is what we say things in – e.g. in French or English – whereas speech is used to indicate what we say in them. Speech involves language but cannot be reduced to it. And words are the atoms of a language whereas sentences are the units of speech. He warned us to be careful not to confuse speech with language, and not to treat words and sentences in the same way, since we would not want to end up with a philosophy that is mere philology, any more than that we want to turn philosophy into an enterprise of discovering platonic entities. Also new – in terminology though not in spirit – was Ryle’s

118

Gilbert Ryle’s Later Writings description of philosophy as an interlevel-activity. This is exactly what constitutes the difference between philosophy and the sciences. We have seen that Ryle’s later writings on thinking deal primarily with what it precisely means to say that our descriptions of thinking are higher-level descriptions of processes. If the thinking Le Penseur is doing is (merely) the way in which processes are managed, then what kind of processes are we talking about? His answer was that although no one specific autonomous X-ing exists for the thinking of Le Penseur, purely as a matter of logic, we do seem to need a non-adverbial – or at least partly non-adverbial – verb for the adverbs to attach to. I showed that Ryle mainly sharpened and strengthened his earlier position from 1949, without radically changing it. By using what I have called ‘higher- order’ concepts Ryle tried to show that a description of the thinking of Le Penseur is not a description of complex processes, but rather a complex (or higher-order) description of processes, which enabled him to reject both dualism and behaviourism, which both (mistakenly) take for granted that all descriptions are treated as if they were names and directly refer to processes or things. He showed us that a description of our thoughts could not be done merely in bottom-level sequences but needed higher-order concepts too; these higher-order concepts, however, do not imply the existence of an extra .

119

The Philosophical Development of Gilbert Ryle

120

Ryle and Collingwood: Their Correspondence and its Philosophical Context

Chapter 5

Ryle and Collingwood: Their Correspondence and its Philosophical Context

In this chapter the only recently published and hardly studied correspondence between Gilbert Ryle and R. G. Collingwood (1889-1943) will be discussed and placed in its historical and philosophical context.85 Ryle and Collingwood are not commonly associated with each other. From Ryle’s autobiographical essay (1970) we know that he saw Collingwood as an exact personification of pre-war Oxford: stuffy, hermetically closed off and philosophically dull. And, as did most of his analytically minded colleagues in philosophy, Ryle regarded Collingwood’s philosophy to be a relic of past metaphysical theory.86 Collingwood, in turn, never paid any particular attention to Ryle’s philosophical writings, except on one occasion in 1935 when Ryle and Collingwood discussed their philosophical views with one another.87 At the time, Ryle was a lecturer at Christ Church College in Oxford and was already regarded as a promising philosopher; Collingwood was University lecturer in Philosophy and Roman History and became Waynflete Professor of Metaphysical Philosophy in 1935, probably just after the correspondence. The tone in their brief correspondence was often biting and chilly, and the two philosophers seemed to disagree about almost everything. On the other hand, a few striking similarities between Ryle and Collingwood have been brought to our attention by Collingwood scholars such as Alan Donagan, David Boucher, W. J. van der Dussen and Giuseppina D’Oro. These philosophers pointed at similarities between Ryle and Collingwood with respect to their rejection of Cartesian dualism and their idea of intelligence as ‘knowing how’. As Donagan writes:

All theories of the relation between body and mind betray a philosophical misconception. Body and mind are not two related substances: they are man as investigated in two different ways, physiologically and historically. There is no conflict between physiology and history. (…) Here Collingwood strikingly anticipated Gilbert Ryle’s view as expressed in The Concept of Mind (New York, 1950). (Donagan 1967, 143)

Boucher notices the following resemblance between the two philosophers:

85 An earlier version of this chapter was published in The British Journal for the History of Philosophy (Vrijen 2006). 86 Collingwood’s philosophical writings are now more appreciated than they were during his lifetime. This is probably due to a radical shift in the interpretation of his philosophical work. Generally, modern interpreters regard Collingwood’s ideas as, broadly speaking, analytical instead of embracing the traditional view of Collingwood as a Hegelian Idealist. 87 I studied the manuscript of the Ryle-Collingwood Correspondence, which is kept at the Bodleian Library in Oxford. At the time I wrote this chapter the correspondence was still unpublished, but it has recently been published as part of a new edition of Collingwood’s An Essay on Philosophical Method (Oxford, 2005) by two Collingwood scholars, James Connelly and Giuseppina D’Oro. I shall give reference to the original manuscript as well as to this new edition.

121

The Philosophical Development of Gilbert Ryle

Ryle like Collingwood was avowedly anti-Cartesian. (…) In The New Leviathan, like Ryle, Collingwood traces the problem back to Descartes. Collingwood argues that the mind does not inhabit a body as a person inhabits a house. (Boucher 1995, 9) 88

Ryle was not the only philosopher to whom Collingwood was compared. In an attempt to represent Collingwood’s philosophical ideas as more akin to modern analytic approaches rather than metaphysical in a Hegelian sense, Collingwood’s name has also been bracketed with that of philosophers such as Wittgenstein and Kuhn. As Van der Dussen writes:

The relevance of Collingwood’s views is increased, moreover, by the fact that it is realized more and more that many of them are remarkable anticipations of influential modern theories. Donagan and Hayden White, for instance, have pointed out similarities between the views of Collingwood and Wittgenstein, Mink and Toulmin have noted those between the theory of “absolute ” and Kuhn’s theory of “paradigms”, while Mink has also referred to Collingwood’s affinities with and . I would add to this list the similarity of Collingwood’s view of science as “problem-solving” (using the “logic of question and answer”) to Popper’s theory of science, as well as the use by both of an idea of “situational analysis” in historical explanation. These modern interpretations of Collingwood differ sharply from the traditional ones, which consider him primarily an idealist and a follower of Hegel or Croce. (Van der Dussen 1981, 2-3)

In this chapter I shall argue that whereas remarks of this sort on the similarities between Ryle and Collingwood are useful in developing a more analytic interpretation of Collingwood, they nonetheless run the risk of seriously misrepresenting the philosophical relation between the two men. The similarities which Collingwood scholars have detected sometimes even suggest that the two philosophers were striving for the same philosophical cause. But the Collingwood commentators, whose main aim was to acquire a better understanding of Collingwood’s position, have systematically neglected Ryle’s point of view. I shall maintain that there are indeed similarities between Ryle and Collingwood, but that qualifications need to be made, primarily with respect to Ryle’s views. The aim of this chapter is twofold: to show that, on the one hand, it is not correct to maintain the view that Collingwood’s and Ryle’s positions were polar opposites, and, on the other, equally mistaken to argue that the two philosophers were striving for the same cause. Instead I shall develop a reading which recognizes convergence as well as divergence. In the first part of this chapter I shall discuss the correspondence between Ryle and Collingwood, which took place in 1935 and consists of three comprehensive letters with a total of over forty pages. My main aim is to analyse the disagreements between Ryle and Collingwood, showing that some of them were merely a question of definition, while others were based on more fundamental issues, such as their differences in method and their views on philosophy in general. My discussion will show that their philosophical positions are not so far apart as they might seem on a first, one-dimensional, reading of the Correspondence

88 Cf. Donagan 1962, D’Oro 2003 and Van der Dussen 1981. With respect to Collingwood’s rejection of Cartesian dualism and his claim that body and mind are only indirectly related, Donagan argues that ‘Ryle, who in The Concept of Mind approached the mind-body problem in much the same way, described Collingwood’s pioneer work very justly in his Inaugural Lecture.’ (Donagan 1962, 292) D’Oro argues that ‘the reason why Collingwood’s contribution to the philosophy of mind has been neglected is due to the fact that his philosophy of mind is widely, even if mistakenly, regarded as the target of Ryle’s attacks on the dogma of the ghost in the machine’. She tries to undermine the assumption that Collingwood is a twentieth century adherent of Cartesian dualism. Collingwood, like Ryle, rejected Cartesian Dualism. (D’Oro 2003) Van der Dussen remarks with respect to Collingwood’s review of Spearman’s The Nature of ‘Intelligence’ and the Principles of Cognition (London 1923) that ‘It is further a striking anticipation of Ryle’s description of intelligence as “knowing how”, as developed in The Concept of Mind.’ (Van der Dussen 1981, 367) 122

Ryle and Collingwood: Their Correspondence and its Philosophical Context and that the context of the Correspondence must be taken into account in order to reach a more accurate and comprehensive interpretation. Until now only a few Collingwood scholars have touched upon it in discussing Collingwood89, but they have focused on placing the Correspondence in the context of Collingwood’s complete philosophical oeuvre. Therefore, they (unfortunately but understandably) neglected Ryle’s point of view. Giuseppina D’Oro gives a more comprehensive treatment of the Correspondence90, but she too does not say much about Ryle. Since the Correspondence has only recently been published, it is not surprising that it has not yet attracted attention from Ryle scholars. In the second part of this chapter, I shall discuss some similarities between Ryle and Collingwood, as mentioned by Donagan, Boucher and Van der Dussen. They describe Collingwood as an anticipator of Ryle with respect to the rejection of Cartesian dualism and the idea of intelligence as ‘knowing how’. D’Oro thinks that both philosophers reject Cartesian dualism, but claims that their positions are very different. She regards Collingwood’s approach as a promising non-reductive one, whereas Ryle is depicted as a logical behaviourist and semantic reductivist. I shall suggest that the remarks of these Collingwood scholars should be qualified, thus offering, I hope, a more accurate and complete interpretation of Ryle’s hitherto neglected point of view.

Part 1: The Ryle-Collingwood Correspondence: Misinterpretation and Structural Disagreements

In their correspondence and in the papers that preceded it Ryle and Collingwood disagreed on almost everything, and gave vivid expression to their disagreement, and sometimes despair:

I shall find it hard to condense within reasonable limits my objections to this argument. (Ryle 1935a, 111)

I am sorry that I have not made it clear to you; the fact is that (perhaps unreasonably) I expected the reader of ch. VI to remember ch. V (…). (Collingwood 1935a, letter 1, 5 = eds. Connelly and D’Oro 2005, 259)

In the course of their correspondence, Ryle and Collingwood failed to convince each other on any point whatsoever. Moreover, they did not even understand each other’s position. And yet Ryle and Collingwood asked each other questions, begged for clarification, and revealed their definitions. How can it be that all of this failed to contribute to a mutual understanding of each other’s position? My answer to this question is that their clarifications and quarrels did not touch the essence of their fundamentally different views of philosophy. As we shall see, Ryle and Collingwood discussed the various different kinds of propositions: universal, particular, categorical and hypothetical propositions. They attacked each other’s definitions and

89 For example, Rubinoff 1970 (pp. 200-208). It is remarkable that no thorough analysis of Collingwood’s An Essay on Philosophical Method (1933) was available until the appearance of Rex Martin’s paper, ‘Collingwood’s Essay on Philosophical Method’ in 1974. Not even prominent interpreters of Collingwood’s writings such as Alan Donagan, Louis O. Mink and Lionel Rubinoff had done more than to give a brief survey of An Essay on Philosophical Method. See Rik Peters 1998 (p. 391). 90 Giuseppina D’Oro, ‘On Collingwood’s Rehabilitation of the Ontological Argument’, Idealistic Studies 30 (2000), and her chapter ‘Collingwood’s ‘rehabilitation’ of the ontological proof’, in D’Oro 2002.

123

The Philosophical Development of Gilbert Ryle quarrelled about the Ontological Argument, but at bottom they were – to use Wittgenstein’s language – playing different games. The only way to mutual understanding would have been an attempt to grasp the game the other one was playing. On a few occasions Ryle and Collingwood tried to do this, but even then they only looked at some of the rules of the other’s game, which they then rejected; they failed to look at the underlying structure. Collingwood begged Ryle several times to stop judging him by the Rylean rules, which he did not want to follow in the first place. Ryle, on the other hand, was unable to do so because of Collingwood’s (in my view) idiosyncratic and sometimes obscure use of language. Thus, Ryle had no other option than to judge Collingwood by his own terminology. First, I shall provide some background to the Correspondence. Second, the definitions upon which Collingwood and Ryle based their arguments will be discussed, focussing on types of propositions, the Ontological Argument and the idea that philosophy is self-reflexive. Finally, I will try to evaluate Ryle’s and Collingwood’s arguments both in their own right and with respect to one another.

Historical Background Being colleagues in Oxford from 1924 until Collingwood’s resignation in 1941, Ryle and Collingwood found themselves in each other’s company on various occasions, such as Sub- Faculty Committee meetings and at the so-called ‘Thursday Teas’91. But they hardly had any contact, either personally or professionally.

I surmise that he [Collingwood] had quite early been lacerated by the Joseph-Prichard treatment, but lacked the resilience to retaliate; and that he then, very unwisely, deemed all philosophical colleagues to be unworthy. When, in 1935, I launched in Mind an unconciliatory criticism of the version of the Ontological Argument in his Essay on Philosophical Method, I had never heard him say a word on this or any contiguous matter; and though my article resulted in some correspondence between us, I am pretty sure that we never met to reduce or to liquidate our differences. If I knew his Christian name, it certainly never tripped off my tongue, even behind his back. (Ryle 1970, 13)

The correspondence between Ryle and Collingwood was occasioned by the publication of Collingwood’s An Essay on Philosophical Method in 1933. One of its chapters was entitled ‘Philosophy as Categorical Thinking’. This chapter, together with Ryle’s fierce attack on it in the journal Mind (1935a), led to the correspondence between Ryle and Collingwood in the same year. Ryle challenged Collingwood to respond to his attack publicly. Collingwood, however, replied in the form of a letter, starting with an apology and justification for his refusal to publish it.92

(…) this letter, incomplete and desultory as it is, already runs twice the length of the paper itself, and is a document which no editor of a journal would dream of publishing. (Collingwood 1935a, letter 1, 1 = eds. Connelly and D’Oro 2005, 254)

91 The ‘Thursday Teas’ was a weekly tea gathering of teachers of philosophy dominated by the senior lecturers. At the end of the 1920s Ryle and a few of the other younger philosophers started their own discussion group, which they called the ‘Wee Teas’. 92 Someone who did take up the challenge was Errol Harris. His reply, ‘Mr. Ryle and the Ontological Argument’, was printed in Mind in 1936. Ryle responded with his ‘Back to the Ontological Argument’ (Ryle 1937a). Harris’ second reply to Ryle was refused by Mind. 124

Ryle and Collingwood: Their Correspondence and its Philosophical Context

This letter, dated 9 May 1935, was the first, and by far the longest, of the three letters of which the Ryle-Collingwood Correspondence consists. Ryle’s response is dated 21 May 1935. The last letter was written by Collingwood on 6 June 1935. Collingwood’s chapter ‘Philosophy as Categorical Thinking’ should be considered in the light of the main aim of An Essay on Philosophical Method (1933), which was to determine what philosophy is by focussing on the similarities and differences between philosophy and science. In this chapter Collingwood tried to show that, by contrast with scientific propositions, philosophical propositions (which in his view are universal) are ‘not merely hypothetical’ but, as he calls it, ‘in essence categorical’. As will be argued below, Collingwood’s claim that philosophical propositions are in essence categorical must be understood against the background of his general approach towards philosophy. He wanted to formulate a metaphysics which is concerned with the presuppositions of our different scientific disciplines and different ways of doing research, and which does not remain as abstract and intangible as, for example, Hegel’s metaphysics. The chapter is divided into three parts; in the first Collingwood maintains that philosophy as categorical thinking has been present in philosophies of all times, from the systems of Plato and Aristotle to those of Kant and Hegel, and beyond. In the second part, he focuses on Anselm’s Ontological Argument, using this ‘famous argument which has stood in the forefront of metaphysical discussion for nearly nine hundred years’ (Collingwood 1933, 124) as another and (because of the important role this argument played in the history of philosophy) more fundamental illustration of the presence of the idea of philosophy as categorical thinking in the history of philosophy.93 The history of philosophy is crucial to Collingwood, who saw his own work as building upon the philosophical achievements of the past. In the third part of the chapter, Collingwood argues that philosophical propositions are not merely hypothetical but in essence categorical because all philosophical propositions are both normative and descriptive. Thus, the three distinctions that play a key role in Collingwood’s chapter are the distinction between universal and particular, between hypothetical and categorical and between descriptive and normative propositions. I will mainly discuss Collingwood’s central claim that philosophical propositions are both universal, and in essence categorical, and address the question as to whether he was justified in using the Ontological Argument, because the Ryle-Collingwood Correspondence is primarily about these issues. The third distinction between descriptive and normative propositions will only briefly be touched upon. Ryle attacked Collingwood’s arguments in his Mind article, and his most important claims are: first, that there are no universal categorical statements; and, second, that Collingwood ignored the criticisms of the Ontological Argument made by philosophers like Hume, Kant and Russell. Many years later, in 1971, when he wrote the introduction to the first volume of his Collected Papers, Ryle recalled the fierceness of his attack on Collingwood.

The vehemence of my onslaught on Collingwood came partly from a local patriotism. We Oxford philosophers were, I thought, by the mid-thirties, sufficiently abreast of the century’s advances in logic and meta-philosophical theory to have the right to dissociate ourselves from

93 That Collingwood had a longstanding interest in the Ontological Argument is evident from his unpublished ‘Lectures on the Ontological Proof of the Existence of God’ from 1919, in which he tried to show that the Ontological Proof is not out of date, but is, on the contrary, of great importance. Collingwood provides a detailed historical discussion of the proof, from Plato to Descartes, Kant and Hegel, concluding with a discussion of the contemporary Realists’ treatment of the argument.

125

The Philosophical Development of Gilbert Ryle

the Ontological Argument that Collingwood had recently exhumed. We were not ex officio dedicated to this lost cause. (Ryle 1971a, vol. 1, viii)

Definitions, the Ontological Argument and Self-reflexivity The differences in opinion between Ryle and Collingwood in their correspondence can partly though not completely be explained by their competing definitions of the terms universal, particular, categorical and hypothetical proposition. As we shall see below, their interpretations of the Ontological Argument and their differing views on the self-reflexivity of philosophy also played a crucial role.

The Universal-Particular and the Categorical-Hypothetical Distinction For Collingwood, universal propositions are general propositions which are not empirical generalisations. Thus, although its form suggests otherwise, ‘All people in this room have blond hair’ is not a universal proposition. It is a general proposition, but also an empirical generalisation and therefore it is particular rather than universal. According to Collingwood, propositions such as ‘Mind exists’ are universal propositions, since they are general propositions and not empirical generalisations. He claims that all philosophical propositions are universal ones. Ryle does not agree with Collingwood. He believes that particular and existence- propositions are the same thing.94 Therefore ‘Mind exists’, which is clearly an existence- proposition, is a particular proposition. Ryle further claims:

By an universal proposition I mean one which is not (and is not a compound of) of the form ‘This is…’ or ‘This is not…’, or ‘Something is…’ or ‘Something that is…is not…’ or ‘There is a…’. Nor yet is an universal proposition a compound of such singular or particular 95 propositions. Instead, an universal proposition is one of the form ‘Whatever is …is …’96 or ‘Any x is y’ or ‘If something is x, it is y’ or ‘Nothing is both x and y’ or ‘Being x involves being y’. (Ryle 1935b, letter 2, 35 = eds. Connelly and D’Oro 2005, 306)

Therefore, propositions such as ‘Mind exists’, which for Collingwood are universal propositions, are for Ryle particular and not universal. With respect to the second distinction, the one between categorical and hypothetical propositions, Collingwood claims that whereas hypothetical propositions are indifferent to the existence of their subject-matter, categorical propositions are not. An example of a hypothetical proposition is: ‘All triangles have three sides’. The truth of this proposition does not depend on the existence of actual triangles. For Collingwood, ‘Mind exists’ is a categorical proposition because its truth does depend on the existence of mind, which is its subject-matter. He claims that all philosophical propositions are not merely hypothetical97, but in essence categorical, that is, they presuppose the existence of their subject-matter; their essence involves existence.

94 ‘Existence-proposition’ is the term used by Ryle and Collingwood in the Correspondence. 95 Singular propositions are of the form ‘That man yonder is a red-haired cardinal’. They entail (but are not quite equivalent to) particular ones, such as ‘a red-haired cardinal exists’ but not vice versa. 96 The original phrase was ‘Whatever is, …is, ….’, but I removed the two commas for the sake of clarity. 97 It seems rather strange that Collingwood first makes a clear distinction between categorical and hypothetical propositions and subsequently partly annuls this distinction by using the word ‘merely’ in his claim that philosophical propositions are not merely hypothetical, but in essence categorical. More on this in the section ‘the nature of philosophy’. For now, I shall ignore the word ‘merely’, which is what Ryle generally did in the Correspondence. 126

Ryle and Collingwood: Their Correspondence and its Philosophical Context

Ryle agrees that ‘Mind exists’ is categorical, but his distinction between categorical and hypothetical propositions is based on different criteria than Collingwood’s distinction. Ryle claims that by definition categorical propositions are particular, singular or a combination of both. Because Ryle also considers propositions such as ‘Mind exists’ to be particular ones, he does not argue against Collingwood’s claim that ‘Mind exists’ is categorical. As mentioned before, Ryle does dispute both Collingwood’s particular claim that ‘Mind exists’ is universal, and the general claim that universal categorical propositions are possible. ‘‘No universal propositions are categorical’ is, therefore, as I use these terms, a pure verbal tautology.’ (Ryle 1935b, letter 2, 35 = eds. Connelly and D’Oro 2005, 306).

The Role of the Ontological Argument How did Collingwood employ the Ontological Argument to support his conclusion that philosophical propositions are in essence categorical, that is, that they are not indifferent to the existence of their subject-matter? He does not use the Ontological Argument as originally formulated by Anselm, in which the existence of God is deduced from a consideration of the concept ‘that than which nothing greater can be thought’98. Instead, Collingwood uses a revised non-theological version of the Argument.

Divesting this argument [the Ontological Argument] of all specially religious or theological colouring, one might state it by saying that thought, when it follows its own bent most completely and sets itself the task of thinking out the idea of an object that shall completely satisfy the demands of reason, may appear to be constructing a mere ens rationis, but in fact is never devoid of objective or ontological reference. (Collingwood 1933, 124-125)

He strips Anselm’s Ontological Argument of its theological implications and formulates what he considers to be the essential truth of the argument, namely that in metaphysics (and in philosophy in general) essence implies existence. Philosophy is not indifferent to the existence of its subject-matter and is therefore in essence categorical.

Reflection on the history of the Ontological Proof thus offers us a view of philosophy as a form of thought in which essence and existence, however clearly distinguished, are conceived as inseparable. On this view, unlike mathematics or empirical science, philosophy stands committed to maintaining that its subject-matter is no mere hypothesis, but something actually existing. (Collingwood 1933, 127)

According to Collingwood, this essential core of truth in the Ontological Argument plays a large role in philosophy, and was never successfully refuted. In line with Collingwood’s revised version of the Ontological Argument, the metaphysical proposition ‘Mind exists’ is categorical in the sense that it is not indifferent to the existence of its subject-matter. As mentioned above, Ryle criticises Collingwood for ignoring important developments in the history of philosophy. Kant, Russell and others had already shown that an a priori argument, such as the Ontological Argument, cannot be used to establish the existence of particular matters of fact. For Ryle, Collingwood’s updated version of the Ontological Argument, devoid of theological implications, is still an a priori argument which wrongly purports to establish the truth of an existence-proposition. Ryle simply cannot accept this and does not understand why Collingwood does not recognise the force of the criticisms made by Russell and the others. Perhaps it is, as Ryle suggests ironically:

98 Anselm, Proslogion, Chapter 2.

127

The Philosophical Development of Gilbert Ryle

‘((…) because it [Russellian logic] happens to use Greek letters for some of its symbols instead of the canonised S, M and P.)’ (Ryle 1935a, 106)

The Self-reflexivity of Philosophical Judgements In the last part of his chapter Collingwood claims that all philosophical judgements are both descriptive and normative. Logic, for instance, cannot be merely descriptive, because then it would be a psychology of thinking; and it cannot be merely normative because then it would not be self-reflexive, that is, it would not be about itself. For Collingwood believes that, by contrast with the exact sciences, logic is in a way about itself.

(…) whereas in geometry, for example, the subject-matter is triangles, &c., and the body of the science consists of propositions about triangles, &c., in logic the subject-matter is propositions, and the body of the science consists of propositions about propositions. In geometry the body of the science is heterogeneous with its subject-matter; in logic they are homogeneous, and more than homogeneous, they are identical; for the propositions of which logic consists must conform to the rules which logic lays down, so that logic is actually about itself; not about itself exclusively, but at least incidentally about itself. (Collingwood 1933, 129)

In the Correspondence Collingwood explains this as follows:

Logic not only discusses, it also contains, reasoning; consequently, whenever a logician argues a point in the theory of inference, he is producing an instance of the thing under discussion; and, since he cannot discuss without arguing, he cannot discuss any point in the theory of inference without doing so. Consequently, in so far as it necessarily contains reasoning, the theory of reasoning cannot be indifferent to the existence of its own subject-matter (…) For example, if a logician could believe that no valid reasoning anywhere existed, he would merely be disbelieving his own logical theory. (Collingwood 1935a, letter 1, 11-12 = eds. Connelly and D’Oro 2005, 269)

As Ryle rightly remarks, this boils down to saying that logic is necessarily self-reflexive. In doing logic we are producing an actual instance of logic and we cannot practise logic without doing so. Not just logical judgements, but all philosophical judgements are not only normative but also descriptive according to Collingwood.99 They are self-reflexive and this self-reflexivity of philosophical judgements, their being instances of themselves, provides an argument for Collingwood’s claim that all philosophical propositions are categorical. Philosophical propositions cannot be indifferent to the existence of their subject-matter because they are instances of themselves; that is, they cannot be indifferent to their own existence. Against Collingwood, Ryle holds that there is no such thing as ‘propositions in general’, but only this kind of proposition and that kind of proposition. And it is possible to talk about a certain kind of proposition without using that kind of proposition. Ryle does not give any examples, but one may think of the following one. One of Ryle’s claims is: ‘Singular propositions entail particular propositions but not vice versa.’ (Ryle 1935b, letter 2, 35 = eds. Connelly and D’Oro 2005, 306) This claim can serve as an example of talking about singular propositions without using that kind of propositions (Ryle used a universal proposition to talk about a singular proposition). Ryle claims that there is no self-reflexivity in philosophy that differs from the accidental self-reflexivity in disciplines such as grammar

99 Another example Collingwood provides is moral philosophy, which he also considers to be self-reflexive. He claims that moral philosophy ‘is both normative and descriptive; it describes, not actions as opposed to ideas about actions, but the moral consciousness [how people think they ought to behave]; and this it is forced to describe as already being in some sense what it ought to be.’ (Collingwood 1933, 132) 128

Ryle and Collingwood: Their Correspondence and its Philosophical Context or poetry; in other words, it is not necessary that logical propositions instantiate the principles which they themselves propound. Philosophical propositions are not necessarily self-reflexive because (at least sometimes) one kind of proposition can be used to talk about another kind of proposition, just as grammar is not necessarily self-reflexive: one may write about Latin grammar in English. Collingwood’s reply is that in philosophy, by contrast with, for example, grammar, the different kinds of propositions are interrelated. An individual proposition can only be assented to in a logical context, and this context is a necessary one.

The context is not (may I say?) a merely psychological context, consisting of anything else that we may happen to be thinking at the time; it is a logical context, consisting of other things which if we didn’t think we couldn’t think what ex hypothesi we are thinking. (Collingwood 1935a, let 1, 17 = eds. Connelly and D’Oro 2005, 277)

The context of an individual proposition is not a psychological one in the sense that it contains anything that we accidentally associate with that specific proposition (for example our association of ‘he gave her a rose’ with ‘red’ and ‘love’), but it is a logical context, consisting of all the kinds of propositions (that is, element-types) which together determine what is assented to when we assent to an individual proposition. Collingwood regards his own theory as building upon the idea which originated from ‘old-fashioned formal logic’ (Collingwood 1935a, letter 2, 18 = eds. Connelly and D’Oro 2005, 278) that there were four element-types, viz. quality, , relation and modality; and many varieties on these element-types. Kant modified this doctrine to a theory of twelve element-types. Collingwood himself is unable to determine how many element-types exist, admitting that he cannot give an exhaustive enumeration of them. His idea of logical context remains rather abstract. He claims that the logical context of an individual proposition has a certain logical structure that always contains all element-types. A few examples of element-types Collingwood mentions are: affirmative and negative elements, categorical and hypothetical elements, and propositional and inferential elements. He claims that these elements are always present in real thinking (this does not mean that they are all expressed in words). Collingwood would claim that in Ryle’s universal proposition ‘Singular propositions entail particular propositions but not vice versa’, a singular element is present as part of the necessary logical context of the proposition assented to, as well as the more obvious universal element. Therefore, in talking about this kind of proposition (in the example: singular propositions), we are always using this kind of proposition (in the example: singular propositions), at least as part of the logical context of the proposition, so we cannot use only that kind of proposition (in the example: universal propositions). In contrast to Ryle, Collingwood maintains that the self- reflexivity of philosophy is necessary and essentially differs from the accidental self- reflexivity of, for example, grammar. Ryle thinks that even if Collingwood is right that logical propositions are necessarily self-reflexive, he is still wrong about the necessary existence of the subject- matter of logical propositions.

(…) I’ll grant (again for the sake of the argument) that the enquiry called logic can’t be done properly save in arguments which exemplify the forms of implication to be studied – and still I say that this is no case of Essence involving Existence100 – it is only a case of a big hypothetical

100 Collingwood claimed that in the case of philosophical propositions ‘Essence involves Existence’, in the sense that their essence involves that they are not indifferent to the existence of their subject-matter.

129

The Philosophical Development of Gilbert Ryle

proposition implying a smaller hypothetical proposition. We are saying ‘the Essence of Implication involves that any logical enquiry into it would have such and such a character’ which is exactly parallel to ‘the Essence of triangularity involves that the internal angles of a triangle are equal to two right angles’. (Ryle 1935b, let 2, 37 = eds. Connelly and D’Oro 2005, 311-12)

For Collingwood, the difference between the case of mathematics and the case of logic is that mathematics does not presuppose that triangles exist ‘in reality’, whereas logic does presuppose the existence of logical arguments ‘in reality’. In fact, logic is constantly construing these logical arguments itself. In the following section I will show that Ryle’s claim that ‘it is only a case of a big hypothetical proposition implying a smaller hypothetical proposition’ can be reconciled with Collingwood’s position. I shall use it as an illustration of how the differences in their use of language and their different basic assumptions led to misunderstandings and the rejection of each other’s position.

The Essence of the Matter So far, the differences in opinion and misunderstandings between Ryle and Collingwood concern their definitions of categorical, hypothetical, universal and particular propositions, their different views on the Ontological Argument and the self-reflexivity of philosophy. These are primarily the issues which were discussed in the Correspondence. I shall now argue that these and other disagreements and misunderstandings arise from their differing uses of the concept of ‘existence’ and their views on the nature of philosophy in general. First, I turn to ‘existence’. What does Collingwood mean by this concept? And what does Ryle think Collingwood means? It is clear from Ryle’s arguments that he considers ‘Mind exists’ to be a strong ontological claim101 and assumes that Collingwood does so as well. And on various occasions, Collingwood’s use of language certainly suggests a strong ontological import to his claims. For example, he answers Ryle’s question as to what he means by ‘essence involves existence’ as follows: ‘The slogan, “essence involves existence” of course means that the essence of something x involves the existence of x.’ (Collingwood 1935a, letter 1, 9 = eds. Connelly and D’Oro 2005, 266) However, recent interpretations of Collingwood’s work suggest that ‘existence’, as he used it, should not be interpreted in such a strong ontological sense.102 The Correspondence seems to confirm these more recent interpretations. For example, Collingwood’s interpretation of Anselm’s Ontological Proof seems to be a rather unusual and, at most, a weak ontological one. Whereas Anselm’s argument is commonly thought of as reaching the strong ontological conclusion ‘God exists’, Collingwood describes the argument as follows: ‘(…) that in conceiving a perfect being we are conceiving a subject possessed of all positive predicates, including that of existence, so that to think of this is already to think103 of it as existing (…).’ (Collingwood 1933, 125). He claims that Anselm’s view amounts to the following: ‘(…) in the special case of metaphysical thinking the distinction between

101 By a strong ontological claim I mean an absolute claim about reality; about the nature of things. In my opinion, Collingwood is merely making weak ontological claims, which are relative to certain structures of knowledge. See below, section ‘The Nature of Philosophy’. 102 See D’Oro 2000, D’Oro 2002, Martin 1995 and Tariq Modood 1995. D’Oro is the only scholar who explicitly claims that Collingwood is not making strong but merely weak ontological claims in his An Essay on Philosophical Method (1933). The arguments of Martin and Modood, although they do not explicitly state it as such, support a position such as D’Oro’s: certainly their claims are incompatible with a strong ontological interpretation of Collingwood’s claims. 103 Italics are mine. 130

Ryle and Collingwood: Their Correspondence and its Philosophical Context conceiving something and thinking104 it to exist is a distinction without a difference.’ (Collingwood 1933, 125). Thus, Collingwood refers to our ‘thinking that x exists’, rather than to the existence of x itself.105 If he had wanted to make a strong ontological claim, it would have made more sense for him either to preserve this ontological aspect of Anselm’s argument, or simply not to use the argument. But how should Collingwood’s arguments in the Correspondence be understood if they cannot be interpreted in a strong ontological sense? And what about Ryle’s response? In order to answer this question we must look at their views of philosophy in general.

The Nature of Philosophy To Collingwood, philosophy is a second order activity which has as its subject-matter the domains of enquiry of first order disciplines such as history and natural science. In other words: philosophy is about history and natural science. Without going into Collingwood’s complicated ideas on history, we may say that, according to Collingwood, history is the study of mind and therefore ‘Mind exists’ is a philosophical proposition; it has mind, which is the domain of enquiry of history, as its subject-matter.106 Philosophy has as its task to reveal the nature of the presuppositions of the first order disciplines.107 Collingwood’s idea of philosophy sheds light on how we should interpret his claim that all philosophical propositions are categorical, in the sense that they are not indifferent to the existence of their subject-matter. Philosophy has as its subject-matter, or, is about, the domains of enquiry of first order disciplines. This means that the subject-matter of philosophy is present as a of the first order disciplines. The subject-matter of the philosophical proposition ‘Mind exists’, which is mind, exists, not in a strong ontological sense, but as a presupposition of history. Collingwood’s philosophical propositions are therefore not absolute claims about reality itself, but they are relative to certain structures of knowledge, e.g. historical knowledge.108 It appears that Collingwood’s claim that all philosophical propositions are categorical is not a strong ontological claim, but a much weaker one, which seems perhaps even trivial. (I cannot imagine that anyone who was willing to accept Collingwood’s ideas on philosophy and history could disagree with him about his claim that ‘Mind exists’, as a presupposition of the domain of enquiry called history.) It is important to note that Collingwood does not hold that philosophical propositions are categorical and therefore not hypothetical. What he does claim is that philosophical propositions are not merely hypothetical, but in essence categorical. This

104 Italics are mine. 105 This interpretation is supported by Collingwood’s discussion of the proposition ‘God exists’, in An Essay on Metaphysics (1940). Collingwood argues: ‘(…) his [Anselm’s] exchange of correspondence with Gaunilo shows beyond a doubt that on reflection he regarded the fool who ‘hath said in his heart, There is no God’ as a fool not because he was blind to the actual existence of un nommé Dieu, but because he did not know that the presupposition ‘God exists’ was a presupposition he himself made.’ (Collingwood 1940, 189) See also Collingwood’s unpublished ‘Lectures on the Ontological Proof of the Existence of God’ from 1919. 106 ‘Matter exists’ is also a philosophical proposition, since, on Collingwood’s view, matter is the domain of enquiry of the empirical sciences. 107 This description of the task of philosophy is not yet present in Collingwood’s An Essay on Philosophical Method (1933), and first occurred in Collingwood’s An Essay on Metaphysics (1940), as a description of metaphysics. Although Collingwood had, by the time of the correspondence, begun to expound views close to his later account of metaphysics as the uncovering of presuppositions, it is unlikely that Ryle would have been aware of this. 108 Henceforth, I will refer to these kinds of ‘ontological’ claims, the ones Collingwood is making, as ‘weak ontological’ claims.

131

The Philosophical Development of Gilbert Ryle seems to mean that philosophical propositions are in a way both categorical and hypothetical. But, how is this possible? The answer to this question must be found in Collingwood’s idea of an overlap of classes in philosophy. For Collingwood, in philosophy – in contrast to exact and empirical science – the specific classes of a genus do not exclude each other, but overlap. For example, in mathematics the classes ‘straight’ and ‘curved’, into which the genus line is divided, do not overlap.

A line is either straight or curved; these are the two species into which the genus line is divided, and they are exclusive and exhaustive: no line can be both, and there is no third species. (Collingwood 1933, 29)

These rules of classification and division do not apply to philosophical statements. For example, when we classify our actions into those done from duty and those done from inclination, instances of the one are in a way instances of the other as well. It all depends on the particular features of actions we wish to emphasise. In the same way, categorical and hypothetical constitute separate classes in non-philosophical judgements, but in philosophical judgements they overlap. Collingwood explains this overlap as follows:

(…) if the body of philosophical knowledge consists of categorical judgements it must at least be surrounded, as it were, by a scaffolding of hypotheticals; I mean that in order to decide that a certain theory is true, our affirmation of this theory must be supported by considering what the consequences would have been, had any of the alternative theories been true. In this sense the working-out of conclusions from purely hypothetical premisses is a very necessary part of philosophical thinking, though a subsidiary part. (Collingwood 1933, 133)

In the light of Collingwood’s account of philosophy and the overlap of classes, Ryle’s objection to Collingwood’s claim that philosophy is categorical because it is self-reflexive seems misdirected.109 Collingwood does not claim that the subject-matter of logical propositions exists absolutely (in the sense of the existence of particular matters of fact), but only that it exists as a presupposition of logic. Perhaps this could, as Ryle argues, indeed be reformulated into what Ryle calls a hypothetical proposition, but this is irrelevant to Collingwood, since he does not deny that (some) logical propositions are hypothetical. What he denies is that they are merely hypothetical. They are in essence categorical, in the sense that their subject-matter exists as a presupposition of the domain of enquiry called logic. And this claim is not refuted by Ryle, at least not in the argument quoted. Ryle could of course still reject Collingwood’s argument by withdrawing his assent (which, after all, he gave only for the sake of the argument) to Collingwood’s claim that logic can only be done properly in arguments which exemplify the forms of implication to be studied. The nature of Ryle’s arguments against Collingwood can be explained partly by Ryle’s claim that all existence-propositions are systematically misleading. As is well- known, at the time according to Ryle one of philosophy’s major tasks is to discover what is really meant by so-called systematically misleading expressions and to reformulate them into expressions which are not systematically misleading. In his paper ‘Systematically Misleading Expressions’ from 1932 Ryle discusses existence-propositions and other systematically misleading expressions. A systematically misleading expression is one which has a syntactical form improper to the fact recorded. Although such expressions may function perfectly well in everyday life, when we start philosophising they create problems

109 Ryle 1935b, let 2, 37 = eds. Connelly and D’Oro (2005), 311-12, already quoted above, see section ‘The self- reflexivity of philosophical judgements’. 132

Ryle and Collingwood: Their Correspondence and its Philosophical Context and muddles. An example of a systematically misleading expression is: ‘Satan does not exist’. Grammatically ‘Satan’ denotes the thing of which ‘exists’ is a predicate, but when we look at the meaning of the expression it does not. For if I assert ‘Satan does not exist’, supposing that my assertion is true, I cannot really be talking about Satan (at least not in the sense in which ‘I am sleepy’ is about me), because Satan does not exist. Ryle suggests that the phrase ‘Satan does not exist’ should be reformulated as ‘nothing is both devilish and alone in being devilish’, or ‘‘Satan’ is not the proper name of anything’. Thus, Ryle claims that all existence-propositions are systematically misleading and should therefore not be used in philosophy. In the Correspondence Ryle did not discuss Collingwood’s idea that there is an overlap of classes in philosophy, but in his 1935 Mind paper he clearly expressed his suspicions: ‘I fear that the principle of the overlap of Classes will be brought in to give us carte blanche to have it both ways when it suits our convenience!’ (Ryle 1935a, 105). For Ryle, ‘particular’, ‘universal’, ‘categorical’ and ‘hypothetical’ are formal properties of propositions. These cannot overlap; no matter in which discipline they are used to classify propositions. He probably was not against the basic idea that underlies Collingwood’s overlap of classes; the idea that particular actions can be classified in different classes (which would be exhaustive and mutually exclusive outside philosophy) depending on which features of these actions we wish to emphasise. But Ryle would describe it differently. He would not describe it in terms of an overlap of classes, since he would probably consider this to be an instance of misleading language. Why would one claim a proposition to be categorical and hypothetical, if what one really wants to say is that the proposition in question is categorical if we emphasise aspect X of the proposition and hypothetical if we emphasise aspect Y of the proposition? The claim that a proposition is both categorical and hypothetical implies that it is both categorical and hypothetical, with the proposition being used in the same sense in both cases, which is nonsense. If the proposition is not used in the same sense in both cases, the formulation ‘proposition Q is both categorical and hypothetical’ is misleading and should be reformulated. Lionel Rubinoff attempts to show that Ryle’s and Collingwood’s different attitudes towards an overlap of classes in philosophy, and their misunderstandings of each other’s position, are rooted in their different philosophical and logical aims. He explains the difference between Ryle and Collingwood as follows: Collingwood aimed at a logic which could ‘adequately regulate the various ways of enquiring into the nature of the dialectical processes of mind’ (Rubinoff 1970, 204), whereas Ryle mistakenly treated Collingwood’s enterprise as a search for a logic that can best describe the world of externally related facts. (Rubinoff 1970, 204). Collingwood looks at the various ways of enquiring and notices that whereas each way of enquiring has its own point of view (its own questions), philosophy is about all these points of view. This is why he needs an overlap of classes in philosophy. For example, something can belong to class X according to one way of enquiring and to class Y according to another way of enquiring, but because philosophy is about these various ways of enquiring, in philosophy the thing in question belongs to both classes; to class X according to one way of enquiring and to class Y according another way of enquiring. More important than saying that Ryle misjudged Collingwood’s project is to show the cause of his misinterpretation. As we have seen, this is at least in part due to Collingwood’s particular use of concepts such as ‘ontological’ and ‘existence’, which on a natural reading imply that he was making strong ontological claims – which he was not. Not surprisingly, Ryle was misled by Collingwood’s language, and mistakenly treated Collingwood’s enterprise as a search for a logic as an instrument to describe external matters of fact.

133

The Philosophical Development of Gilbert Ryle

Furthermore, Ryle’s rejection of an overlap of classes in philosophy is not sufficiently explained by Rubinoff’s distinction between the different kinds of logic Ryle and Collingwood were aiming at. It is not the case that if Ryle had only understood that Collingwood was not aiming to develop a logic which described the world of externally related facts, but one which could account for our various ways of enquiring, he would have accepted Collingwood’s overlap of classes. The different roles of logic in Ryle’s philosophy and Collingwood’s metaphysics are relevant as well. Collingwood’s historical metaphysics of course needs logic (in the way in which all our arguments need logic) but he regards logic as (at least in part) peculiar to particular domains of enquiry, whereas for Ryle logic – for it to be logic at all – holds good for all domains of enquiry. Ryle views logic as a science with its own rules, not as a science of which the rules change depending on which area or discipline it is used in. Ryle’s commitment to a logic which is applicable in all disciplines and his continuous attempt to unmask systematically misleading expressions do not leave him any room for a Collingwoodian overlap of classes.

Philosophical Method Ryle’s and Collingwood’s approaches to method were in many ways quite different. Collingwood tended to adopt what might be broadly viewed as an Hegelian or dialectical approach. This implies a commitment to the history of the subject and explains his wish to retain the philosophical of the past (suitably modified) in the present. Thus, Collingwood reformulates the old metaphysics into a modern one and therefore, in a way, his improved metaphysics encapsulates the old one. The same goes for his use of the Ontological Argument; he radically reformulates the famous argument, but still considers his version of the Ontological Argument to be built on the achievements of Anselm and others. Ryle, on the other hand, generally does not pay any particular interest to the history of philosophy, at least when tackling philosophical problems in a systematic way. Although Ryle certainly was interested in the history of philosophy, he always tried to separate these two interests. This difference in method can be used to explain how and why Collingwood used the Ontological Argument, and why this was incomprehensible from Ryle’s point of view. Whereas Ryle judges arguments by their validity and clarity, Collingwood attaches great importance to the influence of particular arguments throughout the history of philosophy. He considers Anselm’s Ontological Argument to have played an important role through the ages and therefore prefers not to simply ignore or reject it, but rather to adjust it in order to rid it of its theological implications. Ryle, on the other hand, dismisses the argument right away on Russellian grounds and does not take into account its ‘historical value’. In their styles the two philosophers were also extremely different at the time of the Correspondence, though Collingwood’s later style became more accessible. He started using more examples – even common-sense examples – and his use of language became less technical and abstract, especially in The New Leviathan (1942). Consider the following example in which Collingwood refers to the grammatical differences between ‘I see a blue colour’ and ‘I kick a bad dog’ in order to explain why according to Descartes the question ‘Are sensa mind-dependent?’ would have been nonsensical.

(…) Descartes denied the blue colour to be the object of a transitive verb to see, as a dog may be the object of a transitive verb to kick. It means that for Descartes the grammar of the sentence ‘I see a blue colour’ is not like the grammar of ‘I kick a bad dog’ but like the grammar of ‘I feel a transient melancholy’ or ‘I go for a fast walk’. The colour, the melancholy, the walk, are not objects of an action, they are modes of an action; their names have an adverbial function in the sentences in which they occur. If the Cartesian answer is right, the question which Berkeley 134

Ryle and Collingwood: Their Correspondence and its Philosophical Context

answered in one way (‘sensa are mind-dependent’) and Moore, like so many others in the present century, in the opposite way (‘sensa are not mind-dependent’) is a nonsense question: a question to which no possible answer is right because it arises logically from an assumption that is not made. (Collingwood 1942, 30-31)

The role of grammar in this passage bears some resemblance to the famous methods of Wittgenstein and Ryle. Ryle often used the clues of grammar to determine whether or not two expressions belonged to the same category. Collingwood only started using this method at the end of his career, e.g. in The New Leviathan (1942), a work which has been relatively neglected by Collingwood scholars. Ryle’s and Collingwood’s different ways of practising philosophy contributed to their disagreement and misunderstandings. They had different starting-points and were unable to appreciate each other’s philosophical methods. Their dispute about the Ontological Argument in the Correspondence can almost entirely be blamed upon these methodological differences and resulting terminological confusions.

Pre-Humean versus Collingwoodian Metaphysics I have mentioned a few possible reasons why Ryle misinterpreted Collingwood’s enterprise as a strong ontological one, such as Collingwood’s terminology. Something else that may have played a role is the fact that Collingwood’s project was aimed at re-establishing metaphysics, whereas Ryle firmly rejected the possibility of metaphysics. Considering the fact that at the time of the Correspondence Ryle was under the influence of the Vienna Circle and Wittgenstein, it is not surprising that he was not sympathetic towards Collingwood’s project. According to the Vienna Circle, the task of philosophy was logical analysis. Traditional philosophy and metaphysics were to be replaced by the investigation of the logical syntax of the language of science. By clarifying meaningful concepts and propositions, the Circle wanted to lay the foundations for science and mathematics. The principle of verification, the principle that the meaning of a proposition is its method of verification, played a large role in the Circle’s work. Although – as he mentioned in his ‘Autobiographical Essay’ (Ryle 1970) – Ryle had soon started to think that the Circle’s dichotomy ‘Either Science or Nonsense’ did not leave enough room for nuances, the Circle’s influence was still clearly present at the time of the Correspondence. Collingwood’s attempt to breathe new life into metaphysics could have been one of the reasons leading Ryle to think that he was making strong ontological claims. Is that not what all metaphysics does after all? Does not all metaphysics make claims about reality? On closer inspection, Collingwood does not defend the kind of metaphysics Ryle rejects; their conception of metaphysics is entirely different. Because Collingwood’s ideas on metaphysics found only full expression in An Essay on Metaphysics from 1940110, it is not surprising that Ryle misunderstood Collingwood’s project of re-establishing metaphysics.111 Collingwood was not trying to rehabilitate the pre-Humean metaphysics which Ryle (and the Vienna Circle) critically attacked, but his aim was to set the conditions for a

110 In 1934 Collingwood did deliver two lectures on metaphysics, ‘The Permanent Problems of Metaphysics’ and ‘The Special Problems of Modern Metaphysics’, opening a course of 16 lectures on the subject by various speakers. And in 1935 he read the paper ‘Method and Metaphysics’ before the Jowett Society (henceforth to be referred to as Collingwood 1935b). However, neither the paper nor the lectures express Collingwood’s ideas on metaphysics as clearly as his later An Essay on Metaphysics (1940). 111 For the view that Collingwood’s thought did not undergo a radical change between his An Essay on Philosophical Method (1933) and his An Essay on Metaphysics (1940), see Martin 1995 (p. 236).

135

The Philosophical Development of Gilbert Ryle modern metaphysics, which was to be freed from the perplexities and obscurities of the old metaphysics. Collingwood’s metaphysics was to be a ‘historical science’112, its main goal being: ‘to find out what absolute presuppositions have actually been made by various persons at various times in doing various pieces of scientific thinking.’ (Collingwood 1940, 60) He claimed that: ‘The problems of metaphysics are historical problems (…). We must have no more nonsense about its being meritorious to inhabit a fog. A metaphysician is a man who has to get at facts.’ (Collingwood 1940, 62). Collingwood rejects the traditional idea that metaphysics is a science without presuppositions, that it is a ‘deductive’ science – in the sense in which mathematics is a deductive science – and that a metaphysician should aim at system-building.

(…) the idea got about that metaphysics must be a science with no presuppositions whatever, a science spun out of nothing by the thinker’s brain. This is the greatest nonsense. If metaphysics is a science at all it is an attempt to think systematically, that is, by answering questions intelligently disposed in order. The answer to any question presupposes whatever the question presupposes. And because all science begins with a question (for a question is logically prior to its own answer) all science begins with a presupposition. (…) The attempt at a metaphysics devoid of presuppositions can only result in a metaphysics that is no science, a tangle of confused thoughts whose confusion is taken for a merit. Not only has metaphysics quite definite presuppositions, but every one knows what some of them are, for as metaphysics is an historical science it shares the presuppositions of all history; and every one, nowadays, has some acquaintance with the principles of historical thought. (Collingwood 1940, 63-64)

Collingwood’s revised metaphysics does have presuppositions, but not in the way in which mathematics has presuppositions. He claims that the historical character of his metaphysics prevents it from being a deductive, quasi-mathematical science:

The ambition of ‘deductive’ metaphysics is to present a constellation of absolute presuppositions as a strainless structure like a body of propositions in mathematics. That is all right in mathematics because mathematical propositions are not historical propositions. But it is all wrong in metaphysics. A reformed metaphysics will conceive any given constellation of absolute propositions as having in its structure not the simplicity and calm that characterize the subject- matter of mathematics but the intricacy and restlessness that characterize the subject-matter, say, of legal or constitutional history. (Collingwood 1940, 76-77)

Metaphysics is systematic in the sense in which all historical thought is systematic. Its task should be to provide a clear and orderly way of stating and solving problems, but not in the sense of dealing with all of the problems, that is of building a system. The problems do not form a ‘closed repertory’ (Collingwood 1940, 65) and therefore the idea that metaphysics should deal with all the problems is to be rejected. According to Collingwood, aiming at system-building is the ‘purest illusion’ (ibid.). Ryle’s critique of metaphysics had therefore practically the same object as Collingwood’s critique: metaphysics as an absolute deductive science, aimed at system- building and completeness. Ryle criticised the idea of philosophy as a deductive science in his paper ‘Taking Sides in Philosophy’ from 1937:

112 I use the notion ‘historical science’ because this is the notion Collingwood himself employed. However, I do not intend to suggest that after his Essay on Philosophical Method (1933) Collingwood collapsed philosophy into history. I think that Collingwood does not historicise philosophy but considers his metaphysics to be historical in the sense that it is an enquiry into the logical grounds of knowledge, its conclusions being judged against what we find in experience and what their explanatory value is in helping us to understand our experience. However, this is not the place to enter the debate on whether or not Collingwood historicises philosophy. For the interpretation that Collingwood does not, see Connelly 1990 and 2003, Modood 1989, Oldfield 1995 and Martin 1995. 136

Ryle and Collingwood: Their Correspondence and its Philosophical Context

Philosophical argument (…) is not demonstration ordine geometrico. (…) For we have no agreed or evident axioms to start with. In the sense of the word ‘presupposition’ in which philosophy is concerned with presuppositions, the goal of its labours is to reveal them. They are not the premisses of its arguments. (Ryle 1937b, 162)113

In my opinion, Ryle would have interpreted Collingwood’s chapter ‘Philosophy as Categorical Thinking’ differently had he fully recognised that Collingwood was not attempting to re-install a pre-Humean kind of metaphysics, but rather to reformulate metaphysics as a modern ‘historical science’.

Concluding Remarks The fundamental differences between Ryle’s views and those of Collingwood were neither about the Ontological Argument, nor about the way in which this argument can be used to show that philosophical propositions are categorical. Whereas an important part of Ryle’s attack was aimed at the presumed strong ontological aspect of Collingwood’s version of the Ontological Argument, there was in fact no such strong ontological aspect there to be attacked. Collingwood did not aim at establishing the existence of particular matters of fact by means of an a priori argument, but tried to show that there is a set of concepts which is explanatorily, rather than ontologically necessary. That is, there is a set of concepts necessary for explaining how a given area of knowledge or experience is possible. As has become clear, many of the differences between Ryle and Collingwood can be traced back to their fundamentally different views about the nature of philosophy and logic, the nature of metaphysics and the definition of a universal proposition.114 Essential is the difference between Collingwood’s overlap of classes in philosophy and Ryle’s idea of one general logic being applicable in all disciplines. In addition, Collingwood was committed to a belief in the value of the history of philosophy as a part of philosophy itself, whereas Ryle always tried to separate the two; further, Ryle’s concern with systematically misleading expressions does not have a counterpart in Collingwood. It is therefore not surprising that the Correspondence did not bring much mutual understanding and enlightenment. Ryle and Collingwood had radically different positions, methods and terminology. In particular Collingwood’s use of terms was idiosyncratic and was not ideally suited to the position he was in fact trying to maintain, leading Ryle to interpret it in a strong ontological sense.

113 Whereas Ryle and Collingwood both argue that philosophy is not a deductive science, they attribute different roles to empirical evidence in philosophy. Ryle rejects the idea of philosophy as an inductive science on the following grounds: ‘Both the premisses and the conclusions of inductions can be doubted or denied without absurdity. Observed facts and plausible hypotheses have no more illustrative force in philosophy than is possessed by fictions or guesses. Nor have either facts or fancies any evidential force in the resolution of philosophical problems. The evidential force of matters of fact is only to increase or decrease the probability of general or particular hypotheses and it is absurd to describe philosophical propositions as relatively probable or improbable.’ (Ryle 1945, 196) Collingwood’s historical metaphysics suggests that empirical evidence does play a role, in the sense that a philosopher’s conclusions are judged against what we find in experience and whether they explain or help us to understand our experience. 114 Their discussion on what universal propositions are was one of the rare occasions in which the discussion between Ryle and Collingwood was in fact about fundamental differences in opinion.

137

The Philosophical Development of Gilbert Ryle

Part 2: Ryle and Collingwood: Similarities Revisited

So far we have mainly looked at the differences between Ryle and Collingwood. What about the similarities which Collingwood scholars have detected? In the second part of this chapter I shall argue that there are indeed similarities between Ryle and Collingwood, but that Collingwood scholars in their attempt to cast a different, ‘modern analytic’, light on Collingwood, have neglected Ryle’s position and thus misconstrued the similarities between the two philosophers. I shall discuss the similarities concerning their rejection of Cartesian dualism and their idea of intelligence as ‘knowing how’ which were noticed by Donagan, Van der Dussen and Boucher, as well as D’Oro’s views. In addition, I will show that there are important resemblances with respect to a ‘logic of question and answer’. This part does not directly concern the Correspondence, since the similarities I discuss come mainly from the later writings of Ryle and Collingwood, such as Ryle’s The Concept of Mind (1949a), Collingwood’s An Essay on Metaphysics (1940) and his The New Leviathan (1942), but my discussion will indirectly contribute, I hope, to a better understanding of the Correspondence. The most fundamental differences I analysed in the first part of this chapter, such as Collingwood’s and Ryle’s different approaches to logic, will play an important role again in the second part.

Refuting Cartesian Dualism Several Collingwood scholars have pointed to the rejection of Cartesian dualism which Ryle and Collingwood seem to have in common.115 This in itself is interesting enough because, as D’Oro writes, ‘many of Ryle’s contemporaries were quick to identify Collingwood as a twentieth century adherent of the dogma [of the ghost in the machine]’ (D’Oro 2003). She rightly argues that it is a mistake to regard Collingwood’s position as a dualist one and therefore also a mistake to see Collingwood as the target of Ryle’s attack on dualism. However, she subscribes to an influential interpretation according to which Ryle is a logical behaviourist and a semantic reductivist. But this interpretation does not do justice to Ryle’s philosophy which is better characterised as a semantic non-reductive one. With this interpretation in hand, a rather different picture emerges of Ryle and Collingwood, suggesting far more similarities between their positions than D’Oro’s interpretation does. Let me start with a comparison between the refutation of Cartesian dualism by Collingwood and Ryle, after which I shall discuss D’Oro’s interpretation. Finally, I will pay attention to Donagan’s claim that Ryle’s and Collingwood’s rejection of Cartesian dualism had the same form, that is, that they tried to refute it in similar ways. Ryle and Collingwood described the dualism they rejected as follows.

A person (…) lives through two collateral , one consisting of what happens in and to his body, the other consisting of what happens in and to his mind. The first is public, the second private. The events in the first history are events in the physical world, those in the second are events in the mental world. (Ryle 1949a, 13)

Most people, probably, have thought of man’s mind as inhabiting his body somewhat as he inhabits a house. (Collingwood 1942, 8)

As we have seen in Chapter 3, Ryle claims that Cartesian dualism – which he also mockingly referred to as ‘the ghost in the machine’ – is a philosopher’s myth. Dualism

115 See above, footnote 88. 138

Ryle and Collingwood: Their Correspondence and its Philosophical Context makes the mistake of representing the facts of mental life as belonging to one logical category, whereas in fact they belong to another. In other words: mind-body dualists make a category-mistake. They wrongly assume that statements about mind and those about body belong to the same category, or: they mistakenly treat mental concepts as names. According to Collingwood, dualism is a very childish belief, since ‘nothing can inhabit a house made of matter except something else made of matter’ (Collingwood 1942, 8). It is a belief that can only survive by ignorance of its subject-matter; it can have no scientific interest, and therefore Collingwood calls it an ‘old wives’ tale’, a description which resembles Ryle’s ‘Cartesian myth’. He claims that there is only an indirect relation between body and mind and not a direct one, and that ‘The problem of the relation between body and mind’ is a bogus problem which cannot be stated without making a false assumption.’ (Collingwood 1942, 10). This false assumption is that man is partly body and partly mind, and that mind acts on body and body on mind, which Collingwood considers to be another ‘old wives’ tale’. He denies that body and mind are two different substances: they are one and the same thing, namely man himself, as known in two different ways. Donagan claims that, put in Rylean terms, the idea that mind and body are two different substances is the category-mistake from which Collingwood’s ‘old wives’ tale’ derives.116

Not a part of man, but the whole of man, is body in so far as he approaches the problem of self- knowledge by the methods of natural science. (…) Not a part of man, but the whole of man, is mind in so far as he approaches the problem of self-knowledge by expanding and clarifying the data of reflection. (…) The ‘indirect relation between body and mind’ is the relation between the sciences of body, or natural sciences, and the sciences of mind; that is the relation inquiry into which ought to be substituted for the make-believe inquiry into the make-believe problem of ‘the relation between body and mind. (Collingwood 1942, 11)

Both approaches, natural sciences and sciences of mind, are valid according to Collingwood. Each has its own problems and must solve them by using its own methods. One has to avoid what Collingwood calls the ‘fallacy of swapping horses’. If you are trying to solve a problem in the science of mind, you cannot halfway decide to ‘swap your horse’ for the one called natural science. In his inaugural lecture Ryle credits Collingwood for his understanding of the relation between natural science and human studies, that is, sciences of mind.117 He saw that Collingwood denied the existence of a feud between the objectives of the natural sciences and those of the sciences of mind; and further, that he showed that these enquiries were not giving ‘rival answers to the same questions about the same world’ (Ryle 1945, 195), nor ‘separate answers to the same questions about rival worlds’ (Ibid.). According to Collingwood natural sciences and sciences of mind were ‘giving their own answers to different questions about the same world.’ (Ibid.)118 So far, it seems that there are striking similarities between Ryle’s and Collingwood’s rejection of dualism. However, D’Oro in her paper ‘Collingwood and Ryle on the Concept of Mind’ (2003), showing that Collingwood was not a modern adherent of

116 Donagan 1962 (pp. 290-291) 117 Cf. Donagan 1962 (p. 292) 118 In Dilemmas, Ryle tries to unravel dilemmas which originated from the committing of fallacies similar to Collingwood’s ‘fallacy of swapping horses’. ‘There often arise quarrels between theories, or, more generally, between lines of thought, which are not rival solutions to the same problem, but rather solutions or would-be solutions of different problems, and which, none the less, seem to be irreconcilable with one another. A thinker who adopts one of them seems to be logically committed to rejecting the other, despite the fact that the inquiries from which the theories issued had, from the beginning, widely divergent goals.’ (Ryle 1954a, 1)

139

The Philosophical Development of Gilbert Ryle the dogma attacked by Ryle, drives a wedge between the two philosophers by characterising Ryle’s philosophical position as logical behaviourism and semantic reductivism, and Collingwood’s position as a non-reductive one. Whereas D’Oro still depicts the Ryle of 1949 as being heavily influenced by , Stout rightly argues that by that time ‘he [Ryle] wanted to distance himself from the positivistic thought that we should restrict our description of the world to a description of those phenomena that can be determinately measured by scientific instruments’ (Stout 2003, 40). D’Oro suggests that while both philosophers refuted Cartesian dualism, their final positions did not have a lot in common. Against this I argue that there is greater similarity between their positions than she maintains. This conclusion is based upon my claim from Chapter 3 that Ryle should not be interpreted as a reductive logical behaviourist, that is as someone (to use D’Oro’s definition) committed to maintaining the position that ‘statements describing mental phenomena can be translated, without loss of meaning, into statements describing behavioural phenomena’ (D’Oro 2003, 19). According to D’Oro’s interpretation Collingwood is ‘anti-Rylean’ because he does not allow the semantic reduction of statements about mind to statements about behaviour. Following Park’s nonbehaviouristic interpretation, I would say that in this sense Ryle was ‘anti-Rylean’ himself. The wedge D’Oro creates between the two philosophers does not in fact exist. She claims that Collingwood’s motivation for his rejection of the semantic reduction ‘is to be found in his rejection of logical positivism and its adoption of (empirical) verifiability as a condition of meaning’ (D’Oro 2003, 26). But by 1949 Ryle had abandoned these views as well. And the idea that the logical structure of action explanations is different from the logical structure of event explanations, which D’Oro attributes to Collingwood, could also be ascribed to Ryle. Thus, there are more similarities between Ryle’s and Collingwood’s refutation of Cartesian dualism than D’Oro’s interpretation of Ryle leaves room for. Of course there are important differences as well. Donagan claims that Collingwood’s idea that body and mind are not two different substances and are only indirectly related through natural sciences and the sciences of mind, was used by Ryle in The Concept of Mind in much the same way for refuting Cartesian dualism. The same is true of Collingwood’s idea that the enquiry into the problem of the relation between body and mind should be replaced by an enquiry into the relation between natural sciences and the sciences of mind: ‘Ryle (…) in The Concept of Mind approached the mind-body problem in much the same way (…)’ (Donagan 1962, 292). This claim, however, needs to be qualified, as becomes clear when we look at what Ryle was in fact doing in The Concept of Mind (1949a). The idea that mind and body are not two different substances, and that explanatory descriptions in terms of mind and body are two different explanatory descriptions of one activity is indeed present in The Concept of Mind.

He is bodily active and he is mentally active, but he is not being synchronously active in two different ‘places’, or with two different ‘engines’. There is the one activity, but it is one susceptible of and requiring more than one kind of explanatory description. (Ryle 1949a, 50)

However, Ryle nowhere refers to the idea of the relation between sciences as an indirect relation between body and mind, and, secondly, the passage quoted above is the only place in The Concept of Mind where Ryle actually mentions explanatory descriptions. Throughout the book he tries to show that when we describe someone’s action in terms of mental conduct terms we are not claiming that this person is actually performing two

140

Ryle and Collingwood: Their Correspondence and its Philosophical Context actions, one bodily and one mental action. Ryle uses everyday examples and does not refer to scientific explanations at all. One of his examples is that of a clown tripping on purpose. The clown who trips on purpose differs from a clumsy man, who cannot help tripping, in that the clown has a certain skill which the clumsy man lacks. It is not the case that the clown who trips on purpose performs one bodily action, namely tripping, and another mental action, namely doing it on purpose. There is one action in which an ability is put into practice. The clumsy man also performs one action, but without an ability being put into practice. Ryle would probably have agreed with Collingwood’s arguments against Cartesian dualism which were based on the nature of scientific explanations, but he did not in fact use these kinds of arguments in The Concept of Mind. (If he had done so, the book would probably have been misinterpreted less often than has been the case. Ryle has been most frequently misinterpreted as a behaviourist.119) Ryle’s approach was different; he used common sense examples and tried to show that Cartesian dualism led to absurdities. For example, acceptance of the dualist’s claim that we can only have knowledge, viz. introspective knowledge, of our own mental states leads to absurdity, since then we would not be able to know whether another person is intelligent, sad, angry etc. This is absurd, because in everyday life we know perfectly well how to attribute mental concepts to our friends, family and colleagues. Whereas Collingwood starts his argument against dualism by looking at the relations between different kinds of scientific methods and explanations, Ryle starts at the level of what we already know in our everyday life.

This book offers what may with reservations be described as a theory of the mind. But it does not give new information about minds. We possess already a wealth of information about minds, information which is neither derived from, nor upset by, the arguments of philosophers. The philosophical arguments which constitute this book are intended not to increase what we know about minds, but to rectify the logical geography of the knowledge which we already possess. (Ryle 1949a, 9)

The similarities between Collingwood’s refutation of Cartesian dualism and Ryle’s rejection of it in The Concept of Mind, are not, then, as straightforward as Donagan seems to suggest. An important similarity between the two philosophers – not observed by Collingwood scholars – is their attempt to refute dualism without making ontological claims. Their approach to the study of mind is a telling example. They both speak about mind in terms of what it does, without being willing to make strong ontological claims.

A study of mind on the historical method involves two renunciations. First (…) It does not ask what mind is; it asks only what mind does. (…) Secondly, it renounces all attempt to discover what mind always and everywhere does, and asks only what mind has done on certain definite occasions. (Collingwood 1942, 61)

Collingwood does not claim that what mind is and what it does are one and the same thing:

You can have your cake and eat it too by holding that mind is ‘pure act’, so that the question what mind is resolves itself without residue into the question what mind does; but whether this is defensible I shall not ask. (Collingwood 1942, 61) 120

119 Cf. Park 1994 (p. 265) 120 In The New Leviathan Collingwood deliberately does not commit himself to the strong ontological claim to which he seems to have committed himself earlier, e.g. in his early Religion and Philosophy (1916, 34). In this work he claimed that the mind is what it does, that is, that mind is pure act. Some scholars, e.g. Connelly (1995)

141

The Philosophical Development of Gilbert Ryle

Ryle does more or less the same thing. For example, in The Concept of Mind (1949a) he wants to provide a logical geography of the different concepts of mind, which amounts to a thorough and detailed description of what the mind does. Like Collingwood, Ryle refrains from making ontological claims about the mind.

Intelligence and ‘Knowing How’ versus ‘Knowing That’ Another example of an alleged similarity between Ryle and Collingwood is the distinction between ‘knowing how’ and ‘knowing that’. For Ryle, this distinction is of great importance in attacking dualism. ‘Knowing that’ refers to theoretical knowledge and ‘knowing how’ to an ability, a competence. Ryle claims that ‘knowing how’ is the primary one and does not necessarily require ‘knowing that’, that is, theorising.

Efficient practice precedes the theory of it; methodologies presuppose the application of the methods, of the critical investigations of which they are the products. It was because Aristotle found himself and others reasoning now intelligently and now stupidly and it was because Izaak Walton found himself and others angling sometimes effectively and sometimes ineffectively that both were able to give to their pupils the maxims and prescriptions of their arts. It is therefore possible for people intelligently to perform some sorts of operations when they are not yet able to consider any propositions enjoining how they should be performed. Some intelligent performances are not controlled by any interior acknowledgements of the principles applied in them. (Ryle 1949a, 31)

Ryle criticises what he sees as the two basic assumptions of Cartesian dualism, viz. the idea that ‘knowing that’ is the primary activity of the mind, and the assumption that theorising is intrinsically a private and internal operation. He claims that ‘knowing how’ is the primary activity of minds, instead of ‘knowing that’. His strongest argument against the dualists is the following reductio ad absurdum (Ryle 1949a, 31):

Suppose 1) that every intelligent performance is preceded by the consideration of propositions. (This is the dualist’s claim.) 2) The consideration of propositions is itself an operation the execution of which can be more or less intelligent. 3) This leads to an infinite regression of consideration of propositions. 4) This is absurd. Therefore, (1) is rejected.

Ryle’s idea of intelligence as ‘knowing how’ has been linked by Van der Dussen to Collingwood’s concept of intelligence as used in his review (Collingwood 1923) of Spearman’s The Nature of ‘Intelligence’ and the Principles of Cognition (1923). Van der Dussen remarks: ‘It is further a striking anticipation of Ryle’s description of intelligence as “knowing how”, as developed in The Concept of Mind.’ (Van der Dussen 1981, 367). Here Van der Dussen refers to Collingwood’s rejection of the usefulness of intelligence-tests.

Intelligence-tests are meant to test intelligence, and intelligence, as Professor Spearman’s opening chapter shows, is not scientifically definable. The word denotes not a scientific concept but a vaguely-defined and fluctuating mass of attributes which we wish to find in persons who are to be entrusted with certain vaguely-defined responsibilities. To pretend, in such inquiries, to scientific accuracy is like trying to plot the edge of a fog with a theodilite. We can see, normally, and H. S. Harris (1995), have argued that Collingwood still subscribed to the ontological claim that mind is what it does later in his career. Be that as it may, it is in any case clear that in 1942 Collingwood does not use this ontological claim to refute dualism. 142

Ryle and Collingwood: Their Correspondence and its Philosophical Context

when we are in a fog and when we are not; so we can, after ordinary experience of a person, tell whether he is or is not a person of ‘intelligence’, and suitable for positions of responsibility. (Collingwood 1923, 425-426)121

This passage, however, does not unambiguously support Van der Dussen’s claim that Collingwood considered intelligence to be ‘knowing how’ in a Rylean sense. The quotation is about the fact that intelligence does not have well-defined criteria, and therefore intelligence-tests are useless. It is not about whether theorising plays a fundamental role in intelligent performances. Moreover, when we consider Collingwood’s claim, made many years later in The New Leviathan (1942), that thought is both theoretical and practical and that it is primarily practical, it seems plausible that different distinctions are at stake. At first sight, Collingwood’s theoretical and practical thinking seem to resemble Ryle’s ‘knowing how’ and ‘knowing that’. However, a closer study of the examples Collingwood provides of practical thinking suggests differently.

Man’s mind is made of thought. (…) thought is both theoretical and practical. Theoretical thought is, for example, thinking about the cold, or thinking about the difference between cold and hot, or thinking that yesterday was even colder than to-day. Practical thought is, for example, thinking whether to light a fire or thinking that you will go to bed, or thinking: ‘Why should I have the window open?’ (…) thought is primarily practical (…). (Collingwood 1942, 5)

Ryle’s ‘knowing how’ refers to an ability to do something, whereas ‘knowing that’ refers to theoretical knowledge. Collingwood, however, seems to argue that theoretical thinking and practical thinking are both forms of theorising in the Rylean sense of considering propositions – one is theorising in science and the other is theorising in everyday life. One of Ryle’s most important goals in The Concept of Mind was to try to show that theorising (i.e. considering propositions) is not the primary activity of the mind, whereas Collingwood merely seems to claim that scientific theorising (or ‘theoretical’ theorising) is not the primary activity of the mind.122

Logic of Question and Answer There is another similarity – hitherto undiscussed by other scholars – between Ryle and Collingwood, which Ryle himself refers to in his Inaugural Lecture. His tone is strikingly positive, compared to that of the early Correspondence.

Professor Collingwood saw more clearly, I think, than did his most eminent predecessors in the philosophy of history that the appearance of a feud or antithesis between Nature and Spirit, that is to say, between the objectives of the natural sciences and those of the human studies, is an illusion. These branches of inquiry are not giving rival answers to the same questions about the same world; nor are they giving separate answers to the same questions about rival worlds; they are giving their own answers to different questions about the same world. Just as physics is neither the foe nor the handmaid of geometry, so history, jurisprudence and literary studies are neither hostile nor ancillary to the laboratory sciences. Their categories, that is, their questions, methods and canons are different. In my predecessor’s word, they work with different presuppositions. To establish this point it is necessary to chart these differences. This task

121 See Van der Dussen 1981 (p. 367). 122 This is not to deny that practice does indeed play an important role in Collingwood’s writings. Collingwood does state that practice precedes theory, e.g. ‘But in using such language I do not mean to imply that a philosopher can first of all work out certain rules of method and then go on to apply them, as a ready-made instrument, to fresh problems. It is only by working at problems, of whatever kind, that one can learn the way to handle them (…).’ (Collingwood 1935b, 1) But ‘practice’ here has a different meaning than ‘practical thinking’.

143

The Philosophical Development of Gilbert Ryle

Professor Collingwood died too soon to complete but not too soon to begin. He had already made that great philosophic advance of reducing a puzzle to a problem. (Ryle 1945, 195)

The importance of questioning was already recognised by Collingwood at an early stage in his career. In rejecting what he called the doctrine of propositional logic (that truth or falsehood belongs to propositions as such), he developed his ‘logic of question and answer’. In his autobiography he claimed that his work in archaeology had shown him the importance of the ‘questioning activity’ (Collingwood 1939, 30) in knowledge. He considered question and answer as strictly correlative. According to Collingwood’s logic of question and answer, only complexes consisting of questions and answers could be true or false, not propositions as such.

The current logic maintained that two propositions might, simply as propositions, contradict one another, and that by examining them simply as propositions you could find out whether they did so or not. This I denied. If you cannot tell what a proposition means unless you know what question it is meant to answer, you will mistake its meaning if you make a mistake about that question. (…) No two propositions, I saw, can contradict one another unless they are answers to the same question. It is therefore impossible to say of a man, ‘I do not know what the question is which he is trying to answer, but I can see that he is contradicting himself’. (Collingwood 1939, 33)

Consider the following example:

Suppose, instead of talking about the world, the metaphysician were talking about the contents of a small mahogany box with a sliding top; and suppose he said, ‘The contents of this box are both one thing and many things’. A stupid critic may think that he is offering two incompatible answers to a single question, ‘Are the contents of this box one x or many x’s?’ But the critic has reconstructed the question wrong. There were two questions: (a) Are the contents of this box one set of chessmen or many sets? (b) Are the contents of this box one chessman or many chessmen? There is no contradiction between saying that something (…) is one, and saying that it is many. Contradiction would set in only if that something were said to be both one x and many x’s. (Collingwood 1939, 41)

Collingwood’s idea that philosophical propositions could be properly understood only in the light of the questions to which they were answers, does not seem so remote from Ryle’s idea in The Concept of Mind (1949a) that philosophers should be looking at the logical geography of concepts, that is, they should determine the set of ways in which it is logically legitimate to operate with them. One could argue that in order to determine the logical geography of a philosophical proposition we have to look at what questions they are proper answers to. This method could help to lay bare a considerable part of a concept’s logical geography.

He [Descartes] had mistaken the logic of his problem. Instead of asking by what criteria intelligent behaviour is actually distinguished from non-intelligent behaviour, he asked ‘Given that the principle of mechanical causation does not tell us the difference, what other causal principle will tell it us?’ (Ryle 1949a, 23)

Ryle tried to lay bare the logical geography of Cartesian mind-body dualism by showing to what question it was a proper answer. And in his early paper ‘Negation’, he claimed:

By a genuine proposition I mean one which is an answer (the true answer or a false one) to a real question. And by a real question I mean one which can be known necessarily to have an answer, though the answer may not be known. (Ryle 1929, 3)

144

Ryle and Collingwood: Their Correspondence and its Philosophical Context

Ryle generally does not explicitly use the logic of question and answer in his writings, in the sense that he tries to determine the logical geography of a proposition by searching explicitly for the questions to which this proposition is a proper answer. However, his method can often be reformulated into one which does make use of a logic of question and answer, for example when he discusses the logic of dispositional statements in The Concept of Mind.

When a cow is said to be a ruminant, or a man is said to be a cigarette-smoker, it is not being said that the cow is ruminating now or that the man is smoking a cigarette now. To be a ruminant is to tend to ruminate from time to time, and to be a cigarette-smoker is to be in the habit of smoking cigarettes. (Ryle 1949a, 113)

In other words: ‘this man is a cigarette-smoker’ is not an answer to a question of the type ‘is this man smoking now?’, but to a question of the type ‘is this man in the habit of smoking cigarettes?’ However, Ryle did not agree with Collingwood’s claim that propositions cannot be true or false on their own, that is, that only complexes of questions and answers can be true or false. Their different notions of logic and Ryle’s battle against systematically misleading expressions can, once more, serve as an explanation. Let us take the example of the small mahogany box again. Collingwood describes the situation in which the metaphysician says, ‘The contents of this box are one thing and many things’ as follows. Nothing is wrong with the metaphysician’s expression. The critic who thinks he is offering two incompatible answers to a single question reconstructs the question in the wrong way. The critic did not realise that there were two questions. The answer ‘one thing’ concerned the question ‘are the contents of this box one set of chessmen or many sets?’, and ‘many things’ was an answer to the question ‘are the contents of this box one chessman or many chessmen?’. Ryle, on the other hand, would have described the situation differently. He presumably would not have blamed the critic but would have accused the metaphysician of a misleading use of language. Logic should not adjust to misleading and ambiguous language, but the other way around. Nevertheless, the basic idea they share, viz. that in order to understand philosophical propositions we should consider to which questions they are answers.

Conclusion

Ryle’s and Collingwood’s different approaches to logic and to the history of philosophy have been shown to be crucial to the argument developed in this chapter. These differences, together with Collingwood’s sometimes idiosyncratic terminology, which was not ideally suited to the position he was in fact trying to maintain, led in the Correspondence to various misinterpretations – such as Ryle’s interpretation of Collingwood’s project as a strong ontological one – and disagreements. We cannot properly interpret their dispute about the Ontological Argument if we do not have access to its underlying structures. The same is true of their disagreement about ‘existence-propositions’. I hope that, through sketching the broader philosophical context and drawing out some of its implications for a proper understanding of the points at issue, my account has contributed to the formulation of a more accurate and comprehensive interpretation of the Correspondence. On the surface, the Correspondence deals with many disagreements without ever discussing the underlying, more fundamental, issues between Ryle and

145

The Philosophical Development of Gilbert Ryle

Collingwood. In the first part of this chapter I tried to uncover these fundamental issues. The second part was meant, on the one hand, to provide part of the philosophical context of the Correspondence, and, on the other hand, to unmask some of the misleading comparisons that Collingwood scholars have made between Ryle and Collingwood in their attempts to cast a different and more analytic light on Collingwood. I have shown that, whilst there are crucial differences between the two philosophers, nonetheless after disentangling and removing misunderstandings it can be seen that their views are not so far apart as they might first appear. At the same time, I have tried to temper the claims made by those who, like Donagan, Van der Dussen and Boucher, tend to argue for a closer resemblance between Collingwood and Ryle than can be sustained by the texts.

146

Ryle and Wittgenstein

Chapter 6

Ryle and Wittgenstein

No substantial research has been done on the intellectual and philosophical relationship between Ryle and Wittgenstein. This may be explained by the fact that Ryle did not publish much on Wittgenstein and Wittgenstein did not write anything on Ryle. However, the resemblances between the two which have been noted by philosophers such as Peter Hacker, Rom Harré, O. K. Bouwsma, Waismann, Von Savigny and John Shosky make it a topic worth investigating. Because of the scarcity of published comments by Ryle and Wittgenstein on each other’s philosophical ideas, I will largely rely on unpublished sources. This chapter will focus on their similarities and differences, which will illuminate, I hope, aspects of their philosophies. I shall start by giving a historical sketch of their friendship and professional relationship, showing that Wittgenstein indeed played a large role in Ryle’s philosophical life, and – although less obvious – that until a certain time Wittgenstein greatly appreciated Ryle’s ideas as well. Secondly, I will discuss Ryle’s interpretation of Wittgenstein. Important sources here are three papers which Ryle wrote about Wittgenstein and the ‘Ryle Collection’, which is kept at Linacre College Oxford (see appendix 2). The remainder of this chapter aims at further illuminating similarities and dissimilarities between the two philosophers, using for the first time material from the Ryle archives. Obviously, this is a huge theme, and I have limited myself to some general lines of thought and interpretations without pretending in any way to give a complete and detailed account of Wittgenstein’s views and different Wittgenstein interpretations.

Historical context

Ryle became interested in Wittgenstein’s philosophical ideas around 1925, when he was a junior lecturer at Christ Church. At that time Oxford was still, as Ryle would say, philosophically dull and hermetically closed off. His colleagues were in general not interested in what people outside Oxford had to say. The influence of great philosophers such as Russell and Wittgenstein on Oxford philosophy was therefore negligible. Ryle, however, put his mind to their writings and developed a great interest in Wittgenstein’s ideas. Alfred J. Ayer recalled in his autobiography that Ryle had introduced him to the philosophical thoughts of Wittgenstein (Ayer 1977). In Ayer’s last term as an undergraduate (probably Spring term 1932) he read a paper on Wittgenstein’s Tractatus to the Jowett Society. In his Part of my Life (1977) Ayer wrote: ‘I believe, however, that I am right in thinking that this was the first occasion in Oxford on which there had been any public discussion of Wittgenstein’s work.’ (Ayer 1977, 119) As we have already seen in Chapter 1, at one of the Joint Sessions of the Aristotelian Society, Ryle met Wittgenstein and they became friends. During the 1930s they occasionally went on walking holidays together. In a conversation with Bryan

147

The Philosophical Development of Gilbert Ryle

Magee123, Ryle’s answer to Magee’s question whether he would consider his own and Wittgenstein’s work as belonging to the same movement was: ‘(…) certainly I wouldn’t regard myself as independent of Wittgenstein. I was trying to, from say 1925 onwards, to learn whatever I could from him.’ (interview with Magee, 2) And Wittgenstein must have appreciated Ryle’s work as well. In 1942 he answered the question how many people he thought understood his philosophy by saying: ‘Two – and one of them is Gilbert Ryle.’124 (Monk 1990, 436). In his autobiographical essay (Ryle 1970) Ryle describes his occasional visits to the Moral Sciences Club at Cambridge and says that these meetings were totally devoted to Wittgenstein. Ryle thought this total rejection of the thoughts of all other people extremely unhealthy for the students and for Wittgenstein himself. Ryle made further reservations about Wittgenstein. As he writes in 1957:

He was like Socrates in rigidly separating the philosopher from the sophist; unlike Socrates in shunning the market-place; like Socrates in striving to convert his pupils; unlike Socrates in feeling the need to conserve his genius by insulation. He was hermetic, ascetic, guru and Führer. (…) He loathed being connected with academic philosophers, and he avoided academic chores. After 1929 he attended no conferences; he did no reviewing for journals; only once did he attend a philosophical meeting in Oxford; he was inaccessible to visiting philosophers; he read few, if any, of the philosophical books and articles that came out during his last 25 years. (Ryle 1957b, 259)

One of this visiting philosophers was Ayer, who was introduced to Wittgenstein by Ryle in 1932. As Ayer recalls:

They were personal friends though they had their differences, one of them arising from Gilbert’s refusal to admit that it was inconceivable that there should ever be a good British film. (…) I do not think that they often discussed philosophy, in which their style was very different although their thoughts were later to run on rather similar lines. (Ayer 1977, 120)

It sounds odd that Ryle and Wittgenstein would not have discussed philosophical issues much, but this is the picture suggested in Ayer’s autobiography and Monk’s biography of Wittgenstein. At a symposium on Ryle in November 2002 at Linacre College, Oxford, Stuart Hampshire, who was one of Ryle’s colleagues in philosophy in Oxford, claimed that Ryle and Wittgenstein certainly must have had philosophical discussions. Next to his personal contact with Wittgenstein, in 1932-33 Ryle’s other source of Wittgenstein’s thinking was Ayer who, on Ryle’s suggestion, spent his two lecture-free terms in Vienna to learn about the Vienna Circle, as hardly anything was known about the Circle in England. Ayer attended Schlick’s lectures on the philosophy of science at the University of Vienna. He also attended the Vienna Circle meetings, held once a fortnight on Thursday, which Wittgenstein refused to attend. However, Wittgenstein regularly met Schlick and Waismann to discuss philosophical matters, and communicated with the Circle through them. After World War II, the friendship between Ryle and Wittgenstein became less warm, after Ryle had written a positive article on one of Karl Popper’s essays, in which

123 I found a typescript of this conversation at the Philosophical Centre in Oxford. Unfortunately, I have been unable to trace its exact source. Asked by me, Magee could not remember when and where the conversation took place and suggested it may have been taped by a student. 124 Naomi Wilkinson, a cousin of Gilbert Ryle, asked Wittgenstein this question. Miss Wilkinson organized gramophone recitals at the hospital. Wittgenstein regularly attended these and he and Miss Wilkinson became friends. 148

Ryle and Wittgenstein

Popper had been rather negative about Wittgenstein (Popper 1945). Wittgenstein could not appreciate this. Furthermore, Wittgenstein later accused Ryle of ‘borrowing other men’s thoughts’, a type of accusation which was not uncommon of Wittgenstein.125

He [Wittgenstein] spoke of Gilbert Ryle. Ryle had been good when he was young. Now he just borrowed other men’s thoughts. I [Bouwsma] suggested that this was due to the burden of administrative duties. But W. said it was much worse. (Bouwsma 1986, 50)126

Ryle later described Wittgenstein’s tendency to suddenly break off all connections with someone as follows: ‘He remorselessly excommunicated persons of whom he disapproved.’ (Ryle 1957b, 259) Ryle did not boast about knowing Wittgenstein personally, either to his colleagues or his students. Harré – a former student and colleague of Ryle – writes: ‘At that time no one, so far as I can tell, knew much if anything of the fairly close personal relations that had once existed between Ryle and Wittgenstein.’ (Harré 1999, 39) In fact, he was rather secretive about it, perhaps because of his dislike of disciple-like inclinations Wittgenstein provoked. Harré recalls feeling ‘a sense of having been short changed’ (Harré 1999, 45) upon finding out from Monk’s biography how well Ryle in fact knew Wittgenstein. Ryle and Wittgenstein met during the year in which Wittgenstein started to lecture at Cambridge about ideas which differed from the views expressed in his Tractatus. In 1933-34 he dictated lecture notes, and during 1934-35 he dictated another longer manuscript, this time privately to Frank Skinner and Alice Ambrose. These dictations became to be called respectively the Blue Book and the Brown Book, because of the colours of their wrappers. It is likely that Ryle was one of the first Oxford philosophers to have access to the typescripts of The .127 Apart from The Blue and Brown Books, The Tractatus and Philosophical Investigations128, it is difficult to determine to which of Wittgenstein’s other writings Ryle had access and at what stage of his career. Below I will argue that Ryle also read Wittgenstein’s very early and at the time relatively unknown ‘Notes on Logic’ (September 1913), his ‘Lectures on the foundations of Mathematics’ (1939), his Notebooks, 1914-1916, and his letters to Russell.

Ryle’s Interpretation of Wittgenstein

During his philosophical career Ryle only wrote three papers on Wittgenstein: two rather general ones, ‘Ludwig Wittgenstein’ (Ryle 1951c) and ‘Review of Ludwig Wittgenstein: Remarks on the Foundations of Mathematics’ (Ryle 1957b), and a more specialized one which was published posthumously, ‘Ontological and Logical Talk in Wittgenstein’s Tractatus’ (Ryle 1993). When Ryle wrote his first published paper on Wittgenstein, which was an obituary in Analysis, Wittgenstein’s Philosophical Investigations had not yet been published, but Ryle already referred to Wittgenstein’s work postdating the Tractatus. Most

125 Cf. Hacker 1996 (p. 313), footnote 98 and Bouwsma 1986 (p. 50). 126 This conversation between Bouwsma and Wittgenstein took place in October 1949. 127 Cf. 1990 (pp. 336-37); 1999 (p. 15); Flew 1998 (p. 194); Ben Rogers 2000 (p. 159); Hacker 1996 (p. 86). 128 I have found no indications that Ryle also had access to early, unpublished versions of the Investigations, although in theory it is possible that he had early access to the first 188 remarks of the later Investigations (mainly written during the autumn of 1936 and 1937), 37 remarks of which can be traced back to earlier manuscripts from 1930-1934, and 24 to the attempted revision of the brown book (in German) in 1936 (further: 23 to a notebook from 1936, 13 to manuscript writings from 1937-38, and one to 1945.) (Luckhardt 1979, 142).

149

The Philosophical Development of Gilbert Ryle likely, these remarks were mainly based on the Blue and Brown Books and Lectures on the Foundations of Mathematics, at that time still unpublished though circulating. The title of Ryle’s second paper, ‘Review of Ludwig Wittgenstein: Remarks on the Foundations of Mathematics’, suggests that it is a review of Remarks on the Foundations of Mathematics. But since a review of this work would have been too specialized for the general reader of Scientific American, the editors had asked Ryle to provide a more general interpretation of Wittgenstein. By that time Philosophical Investigations had been published and when speaking about Wittgenstein’s philosophical thought after the Tractatus Ryle mainly had this work in mind. Other sources I shall be using in my discussion of Ryle’s interpretation of Wittgenstein are Ryle’s heavily annotated copies of the Tractatus129 and his, less heavily, annotated copy of Philosophical Investigations130, a transcription of Rom Harré’s notes of Ryle’s lectures on Wittgenstein during his course ‘History of Theories of Meaning’ (1954- 55), and several unpublished documents which are kept in the so-called ‘red box’ at Linacre College: a typed comment on Wittgenstein’s early ‘Notes on Logic’ (1913), (lecture) notes on the structure of Wittgenstein’s Tractatus131, and a bibliography which was probably intended for students studying the Tractatus.132 The annotations reveal Ryle’s extensive study of Wittgenstein’s thought and make it all the more regrettable that he did not publish more about Wittgenstein. From what he did publish and from the unpublished material in the ‘Ryle Collection’ I shall try to construe Ryle’s ‘Wittgenstein’.

Wittgenstein’s ‘Overriding Worry’ In the introduction to the first volume of his Collected Papers Ryle writes about his intentions in studying other philosophers:

Not all, but most of these Critical Essays issue from a common exegetic policy (…). Through nearly all (…) runs a common strategy, or, it may be, a common idée fixe. From the time of the Tractatus the question had been a live and insistent one: – what sort of an enquiry is philosophy as distinct from Natural Science, Mental Science, Mathematics, Theology and Formal Logic? What, if any, is its proprietary subject-matter? What, if any, is its peculiar method? (…) To elucidate the thoughts of a philosopher we need to find the answer not only to the question ‘What were his intellectual worries?’ but, before that question and after that question, the answer to the question ‘What was his overriding Worry?’ (viii-ix)

According to Ryle, Wittgenstein’s overriding worry – the central question running through all of his writings – was: ‘What can philosophers and logicians do, and how should they do it?’ (Ryle 1951c, 251) Ryle claims that except for some incidental discussions, the question what philosophy is, or how it differs from science began seriously to worry philosophers only after the publication of the Tractatus. Wittgenstein’s important role in the rise of discussions about the nature and methods of philosophy cannot be underestimated.

129 The ‘Ryle Collection’ contains three different editions of Wittgenstein’s Tractatus (London & Kegan Paul LTD 1922, 1951 and 1969). The second oldest is most heavily annotated; the most recent one hardly so. 130 Philosophical Investigations, Wittgenstein, trans. Anscombe, Basil Blackwell Oxford 1953. Another version of Philosophical Investigations also owned by Ryle has disappeared from Linacre Library. 131 See Appendix 3. 132 If the notes on the Tractatus are written around the same time as the bibliography, which was presumably meant as a reading-list for students studying the Tractatus, they are to be dated between 1955 and 1957; for in his bibliography Ryle refers to Wittgenstein’s ‘Notes on Logic’, which was published in the Journal of Philosophy in 1957, as ‘unpublished’, and he already mentions Moore’s articles in Mind on ‘Wittgenstein’s teaching in Cambridge, 1930-33’, which were published in 1954-1955. 150

Ryle and Wittgenstein

The Tractatus Because Wittgenstein’s initial interest was in the philosophy of logic as inspired by Russell’s and Frege’s logicistic programme regarding the foundations of mathematics, he tended to put all meaningful expressions in the mould of the logic of mathematical expressions, thus Ryle. His logical-mathematical terminology – by which Ryle does not only mean that Wittgenstein occasionally uses mathematical formulas, but also his apodictic and aphoristic style – makes the Tractatus a very difficult work, and for understanding it one must first master its terminology. Ryle considers the two main aims of the Tractatus to be: 1) showing what philosophy is and what it is not, and 2) showing what Formal Logic is. Because of Ryle’s lack of interest in the formal logical part of the Tractatus, and perhaps also his unfamiliarity with its technical details, he mainly focuses on its first aim, claiming that in Wittgenstein’s later writings the question of the status of logic disappeared to the background. It was rather ‘philosophy (…) that is pestering him [Wittgenstein] for justice’ (Ryle 1957b, 265). Ryle tries to make the Tractatus more accessible by proposing chapter headings and section titles, something Wittgenstein most likely would not have approved of. The ‘red box’ at Linacre College contains a sheet of (lecture) notes on the structure of the Tractatus (see Appendix 3), suggesting treating the whole numbers as chapters and the single decimal numbers as sections. In his paper from 1957 Ryle argues that ‘the Tractatus consists of a chain of sentences or short paragraphs, prefaced by numerical and decimal index-numbers signalling both the train of the argument and the relative weights in it of the successive items.’ (Ryle 1957b, 265) However, the sheet found in the ‘red box’ makes it clear that Ryle did not always consider this to be the case: ‘Certainly sometimes a very important thing is said in a paragraph or sentence with a very fiddling number of decimals in its number, e.g. in 4.0312.’ (Ryle, Appendix 3) Apart from explaining Wittgenstein’s main aims and method, in his papers Ryle also makes an attempt to place Wittgenstein’s ideas in their historical context. In his paper from 1951, for instance, he traces Wittgenstein’s use of the dichotomy sense-nonsense back to Russell. When Russell investigated the logical principles which were the foundations of mathematics, he ran into trouble. He could not avoid the construction of expressions with the unpleasant habit of being true if and only if they were false, and vice versa. In other words: he ended up with paradoxes. Russell’s solution was to recognize a distinction more fundamental than the one between true and false, namely the one between significant and nonsensical expressions. Both true and false expressions are significant, but there is also a class of expressions which are neither true nor false: the class of nonsensical expressions. According to Ryle, Wittgenstein generalised Russell’s distinction in the Tractatus. For Wittgenstein:

All logic and all philosophy are enquiries into what makes it significant or nonsensical to say certain things. The sciences aim at saying what is true about the world; philosophy aims at disclosing only the logic of what can be truly or even falsely said about the world. This is why philosophy is not a sinister science or a parent science; that its business is not to add to the number of scientific statements, but to disclose their logic. (Ryle 1951c, 252)

Already in the Tractatus Wittgenstein started to free himself from what Ryle called ‘the denotationist camp’ (Ryle 1957a, 363), that is, the idea that meaning is naming. He realised, as did Frege, that logical and philosophical questions ‘are not questions about the properties or relations of the denotata, if any, of the expressions which enter into the sentences whose logic is under examination.’ (Ryle 1957a, 363-4) Logical constants, such

151

The Philosophical Development of Gilbert Ryle as ‘all’, ‘not’, ‘is’, ‘and’, do not stand for objects. They have the meanings they have by virtue of the contributions they make to the structures ‘of the unitary senses of the sentences in which they function’ (Ryle 1962d, 187), not by virtue of designating extra objects or terms.133 This is why philosophers and logicians cannot construct significant sentences – that is, sentences which are either true or false – about the meaning of logical constants. Wittgenstein maintained not just that sentences about logical constants could not be constructed significantly, but thought that, more in general, everything philosophy wanted to say could not be said (at least not significantly).

The conditions of significant (true-or-false) assertion could not be the topics of significant assertions. That sentences of different sorts observe or break the rules of significance can be shown but not stated or explained. (Ryle 1950, 247)

Ryle suggests that Wittgenstein’s claim of the ineffability of philosophy was, perhaps, derived from the idea central to Russell’s Theory of Types that a proposition cannot be about itself, or a sentence cannot convey a truth or falsehood about what it itself says (‘this sentence is false’). Another important move away from denotationist theory, additional to the idea that logical constants do not function as names, was the tractarian idea that sentences are not names but expressions of thoughts.

He [Wittgenstein] saw, too, that all the words and phrases that can enter into sentences are governed by the rules of what he called, slightly metaphorically, ‘logical syntax’ or ‘logical grammar’. These rules are what are broken by such concatenations of words and phrases as result in nonsense. Logic is or includes the study of these rules. (Ibid.)

Ryle later claims to have learned from Wittgenstein that philosophical problems are problems of a special sort and not ‘problems of an ordinary sort about special entities’ (Ryle 1971b, vii). In his papers and lectures Ryle does not pay attention to Wittgenstein’s thoughts on the relations between logic and mathematics and between logic and mechanics, although he does acknowledge their importance for showing the positive nature of logic. Instead he points to philosophy as opposed to logic and science, discussing the problem that whereas we can talk sense ‘we cannot talk sense about the sense that we talk’, and this is exactly what we are required to do in philosophy.

The truths of the natural sciences are factual truths, while those of logic are purely formal. Their truth is neutral between the world as it is and as it might have been. (…) Thus logic is unconcerned with the actual truth or falsity of the factual statements which can be draped on its skeletons. Nonetheless logic is essentially concerned with the truth-or-falsity of these statements, since it has to work out how the truth or falsity of one would follow, if another were true or were false. (Ryle 1957b, 260-261)

Ryle tries to explain what this truth-or-falsehood of a statement is by referring to our understanding of the proposition ‘It is raining’. Even when we do not know whether the

133 Of course this is by no means a new theory. In medieval times one called words such as ‘all’, ‘none’, etc. ‘syncategorematic terms’, i.e. terms which do not have a meaning by themselves, but derive their meaning from their connection to ‘categorematic terms’ (‘all’ does not have a meaning when taken on its own, but does have meaning when added to ‘man’). However, later movements had started to treat them in the same way again as categorematic terms. This mainly in order to be able to maintain logical realism and anti- psychologism. 152

Ryle and Wittgenstein statement is true or not, we do understand its meaning. In grasping the meaning of ‘It is raining’ ‘you are getting not what the state of the weather is, but what-it-is-being-presented- as-being’ (Ryle 1957b, 261), or ‘what it would be like if it were a fact-stating sentence’ (Harré 1999, 47), as Ryle explained during one of his lectures on the Tractatus in 1954-55. The next question in need for an answer is how an expression means something for Wittgenstein. It means something in much the same way as a map represents, truly or falsely, the relative positions and distances of different towns.

A sentence has a meaning if its syntax could be the structural analogue of an actual state of affairs, even though, when false, it actually has no such factual counterpart. Ceasar did not kill Brutus, but ‘Ceasar killed Brutus’ makes sense, since there is, so to speak, room in reality, though unfilled room, for this uncommitted murder. (Ryle 1957b, 262)

A truth-or-falsehood, then, is an organized complex of symbols representing, by analogy of structure, a counterpart actual-or-possible state of affairs. It is, for example, a sentence, ‘in its projective relation to the world’. To find out whether it is actually true or actually false we have to match it against its should-be counterpart state of affairs in the world.134 This idea, together with the fact that this matching could only be done by observations and experiments, led, according to Ryle, to the verificationist refutation of metaphysical, theological and moral statements as nonsensical by logical positivists and by Wittgenstein. They claim that these statements ‘have no anchorage in facts and so say nothing; they are nothing but disguised gibberish’. (Ryle 1957b, 262) But then, what about truths of logic? It is true that these are factually empty, they are tautological, but it is not the case that they are of no use to us. They show us, ‘by evaporation of content’, how our ordinary statements are organized. A more detailed discussion of the Tractatus is presented in Ryle’s posthumously published ‘Ontological and logical talk in Wittgenstein’s Tractatus’135. René Meyer dates the paper to the late thirties, though he does not give an argument, and it might equally well have been written many years later. As we know from his annotated copies of the Tractatus, Ryle seems to have been intrigued by this philosophical work and kept reading it over and over again. As the title already suggests, in this paper Ryle tries to explain the relation between the ontological and the logical claims of the Tractatus, speaking of, respectively, the ‘ontological story’ and the ‘propositional story’. Ryle takes it that the remarks with index- number 1 tell the ‘ontological story’, the ones numbered 2 the ‘picturing story’ and the ones numbered 3 the ‘propositional story’. In his notes on the Tractatus in the ‘Red Box’ Ryle categorizes the remarks numbered 4-7. See Appendix 3. Ryle discusses three different ways of explaining the relation between the ontological and the propositional story: (1) what he calls the ‘Euclidean’ interpretation: the propositional story is a complex conclusion following from the ontological story as a complex premise; (2) the so-called ‘reversed Euclidean’ interpretation: the ontological story can be inferred from the propositional story. An example of a ‘reversed-Euclidean theory’ is Russell’s idea that simples are found inferentially as the limit of analysis; (3) the ‘Aesop’ interpretation: Wittgenstein considered the ontological and the propositional story to be the same story, the first being an allegory told in the ‘material mode’ and the second the same

134 Ryle’s description of Wittgenstein’s ideas on how an expression has a meaning closely resembles the description he gave of his own views in Ryle 1930. (See also Chapter 2, pp. 37-40.) He may well have been influenced by Wittgenstein on this point. 135 The title is René Meyer’s. See Meyer 1993 (p. 6).

153

The Philosophical Development of Gilbert Ryle story told in the ‘formal mode’. Or: ‘that the ontological story [stands] to the propositional story as an Aesop’s fable stands to the corresponding ethical doctrine.’ (Ryle 1993, 102) Ryle thinks that, if pressed, Wittgenstein would have opted for the ‘Aesop’ interpretation. Ryle’s most important argument against interpretation (1) and (2) is that whereas Wittgenstein uses formal concepts, i.e. category-words (‘object’, ‘substance’, etc.), as if they were proper concepts, i.e. ordinary words, in his ontological story, from 4.126 on he makes clear that using formal concepts as proper concepts generates nonsense. Thus, ‘he cannot then have thought that the ontological story was a legitimate premise or a legitimate conclusion in an inference to or from the propositional story.’ (Ryle 1993, 103) Ryle suggests that Wittgenstein left it in for another purpose: as ‘a sort of prefatory parable’ (Ryle 1993, 104) and not either as a premise or as a conclusion – as ‘something that would not do, as a lead in to saying something that would do or nearly do’ (Ryle 1993, 104). This he considers to be the most charitable interpretation.

I think that for nearly every seemingly factual statement Wittgenstein makes about objects, simples, complexes, etc., in the ontological story, we could find a corresponding meta-statement about propositions in the propositional story. I suggest that this correspondence is not that of premisses to conclusions or conclusions to premisses, but of nursery-statements to grown-up statements. (Ryle 1993, 105)

Ryle does not discuss the role of the picturing story in the Tractatus, but claims that his interpretation needs to be extended to include this role, showing the place of the picturing story between the ontological and the propositional one:

Before we can talk about caricatures and maps, we have to be able to talk about faces and terrains. So Wittgenstein had to produce some seemingly descriptive talk about things and facts before he could say anything about caricatures and maps being true or false. (Ryle 1993, 105)

From the Tractatus to the Philosophical Investigations Ryle considers the break between the Tractatus and the Philosophical Investigations less radical than most traditional interpreters do, although he does support the idea that Wittgenstein later broke with some of the claims of the Tractatus. He claims for example that at the time of the Tractatus, although ‘one foot was already free’, ‘Wittgenstein still had one foot in the denotationist camp’ (Ryle 1957a, 363). He already realized that logical constants do not stand for objects and do not have the meanings they have qua designating objects, and that sentences are not names and should not be treated as if they are. According to Ryle, only after the Tractatus Wittgenstein removed his other foot, ‘consciously’ and ‘deliberately’, from the denotationist camp when he started to ask for the use instead of the meaning. The use of an expression, like a piece of chess, is the role which it plays, ‘not any thing or person or event for which it might be supposed to stand’ (Ryle 1957a, 364). Another difference was that the later Wittgenstein no longer forces all meaningful expressions into the mould of the logic of mathematics and even develops a general aversion against moulds.

Philosophical elucidation is still inspection of expressions, but it is no longer inspection through the slots of a logician’s stencil or through the prisms of a scholastic classification-system. His diction has reverted from that of a Russell discussing esoteric matters with mathematicians to that of a Socrates discussing everyday ideas with unindoctrinated young men. Like Moore, he explores the logic of all the things that all of us say. (Ryle 1951c, 255)

154

Ryle and Wittgenstein

According to Ryle, after the Tractatus Wittgenstein quickly realized that not everything we say is a truth or a falsehood, since not everything we say is an assertion. Other sayings are for example: questioning, promising, reassuring, joking, warning and praising.

In the Tractatus we were told, in effect, that only those sentences made positive sense which could be premisses or conclusions of a bit of natural science. In the Philosophical Investigations the door is opened to anything that anyone might say. We are home, again, in the country of real discourse. (Ryle 1957b, 266)

Whereas the Tractatus was primarily about written sentences, the Investigations is about what sentences normally are: ‘things said, not written’ (Ryle 1957b, 266). In other words: ‘in contrast to the passive character of the Tractatus which depicts a listener’s or hearer’s world, in the Philosophical Investigations people use expressions when they want to say things.’ (Ryle in Harré 1999, 51) In the Tractatus only stages and patterns of compositionality counted as distinction-criteria, while in his later work much more distinctions are possible. ‘There is no limit to the number of different sorts of sentences.’ (Ryle in Harré 1999, 51) Correspondingly, the rather static notion of sense has developed into a more elastic one. Both in the Tractatus and in the Philosophical Investigations Wittgenstein avoids general statements about the nature of philosophy. In the Tractatus this would have been saying what cannot be said. In the Philosophical Investigations it is because generalisations bring along unclarity. The nature of philosophy is such that it can only be taught by producing examples. We can show what makes a philosophical puzzle a puzzle by showing how it is solved, not by telling it; by teaching an ability, not by dictating a doctrine. Ryle’s descriptions of Wittgenstein’s philosophical method closely resemble his own philosophical ideas. As Ryle says about the Philosophical Investigations:

A crude generalization of Wittgenstein’s new account of sense or meaning is that the meaning of an expression is the rules for the employment of that expression; that is, the rules licensing or banning its co-employment with other expressions, those governing its effective employment in normal and abnormal communication-situations, and so on. The dynamic notion of rules to be mastered has replaced the notion of an imposed structural congruence. (Ryle 1957b, 267)

Ryle further mentions several differences in method and style between the two works:

The Tractatus consists of a chain of sentences or short paragraphs, prefaced by numerical and decimal index-numbers signalling both the train of the argument and the relative weights in it of the successive items. Each sentence seems to be the product of an almost Chinese process of pruning and recasting. Many of them mystify, but the reader cannot get them out of his head. In many stretches the Tractatus presupposes familiarity with . The Philosophical Investigations is more like a conversation. It is a dialogue between the author and his own refractory self, and it presupposes no technical sophistication. It is split into relatively long paragraph-sections, the continuities between which are often hard to see. Indeed, they are not always there. Unfortunately the book contains no aids to the reader in the shape of table of contents, index or cross references. (Ryle 1957b, 265)

The last sentence of this quotation brings us to an important difference in method, to which I will come back below. Wittgenstein generally considered it to be harmful to try to present philosophical thoughts in the conventional form of deductive order. In order to avoid this he once even proposed to order the sentences of a philosophical work alphabetically.136 He

136 See Max Black 1964 (p. 2).

155

The Philosophical Development of Gilbert Ryle would not have appreciated Ryle’s suggestion to add a table of contents and an index.137 Ryle made his own, private index by using the last page both of the Tractatus and The Philosophical Investigations to list important subjects and names with their corresponding pages. In his annotations of the Philosophical Investigations Ryle regularly showed his discontent with the place of certain remarks, i.e. with the way in which the remarks were put together. He suggests for example that 54 is out of place and should follow 56, and 55 should be placed next to 41, so that 56 should come right after 53 which is about the various possibilities a particular simple language-game has and the fact that a rule of a language-game may have very different roles in the game. Ryle suggests placing 56 right after 53 because 56 starts with the continuation of an example from 53. The paragraphs in between are more appropriate to follow 56. Ryle further wonders whether the example ‘Moses did not exist’ in 79 should be placed near 58. Remark 86, at least the last part of it, has to be removed to 143. Other suggestions are that the beginning of 88 and 71 should be placed together, just as 114 and 134; 116 and 133; 181 and 183; and 182 and 138. Ryle thinks that the method of showing that a certain analogy is carried through too far is one of the most important methods of the Philosophical Investigations. Wittgenstein ‘would show how striking similarities may go with important but ordinarily unremarked differences, and how we are tempted to lean too heavily on their similarities and hence to be tripped up by their latent differences.’ (Ryle 1951c, 255) Despite the differences between the Tractatus and the Philosophical Investigations, Wittgenstein’s ‘overriding Worry’ remained the same: ‘What can philosophers and logicians do, and how should they do it?’ (Ryle 1951c, 251), although his focus changed from logic to philosophy.

Moreover, a great deal of the Tractatus survives, both in the later Wittgenstein and in us too. It comes natural to us now – as it did not 30 years ago – to differentiate logic from science much as Wittgenstein did; it comes natural to us not to class philosophers as scientists or a fortiori as super-scientists; it comes natural to us to think of both logic and philosophy as concerned not with any ordinary or extra-ordinary kinds of things, but with the meanings of the expressions of our thoughts and knowledge; and it is beginning to come natural to us, when we reflect about sense vs. nonsense, to take as the units of sense what is conveyed by full sentences, and not what is meant by isolated words, that is, with what is said, and not with what is, for example, named. (Ryle 1957b, 265)

Ryle obviously held Wittgenstein responsible for great progress in philosophy. He writes:

Wittgenstein has made our generation of philosophers self-conscious about philosophy itself. (..) We no longer try to use for our problems the methods of arguing which are the right ones for demonstrating theorems or establishing hypotheses. In particular we have learnt to pay deliberate attention to what can and cannot be said. What had, since the early days of this century, been the practice of G.E. Moore has received a rationale from Wittgenstein; and I expect that when the curtain is lifted we shall also find that Wittgenstein’s concrete methods have increased the power, scope and delicacy of the methods by which Moore has for so long explored in detail the internal logic of what we say. (Ryle 1951c, 256-257)138

137 Ryle himself did not add an index to his Collected Papers (contrary to his earlier The Concept of Mind, which did contain one). In the introduction to the second volume he tries to justify the lack of an index as follows: ‘I have refrained from charging myself or any colleague with the labour of compiling an index. Such an index would expedite the studies only, I like indolently to think, of those who will be writing Doctoral Dissertations; and for them the chore of rummaging for themselves will be more rewarding than would be their inheritance of the proceeds of other people’s rummagings.’ (Ryle 1971b, viii) 138 ‘When the curtain has lifted’ possibly refers to the large bulk of unpublished writings by Wittgenstein. 156

Ryle and Wittgenstein

Ryle’s writings on Wittgenstein’s Philosophical Investigations remained limited to the following: ‘Review of Ludwig Wittgenstein: Remarks on the Foundations of Mathematics’ (Ryle 1957b), which is of a very general nature; some loose critical remarks in On Thinking (Ryle 1979d) and Collected Papers (Ryle 1971a and 1971b); and a critical discussion of Bouwsma’s interpretation of Wittgenstein’s Philosophical Investigations (Ryle 1972d). Ryle’s critique of Bouwsma’s Wittgenstein interpretation, which will be discussed in the second part of this chapter, will help to bring out his own interpretation of Wittgenstein’s Philosophical Investigations.

Ryle’s Critical remarks In ‘Systematically Misleading Expressions’, Ryle felt the need to make it clear that this paper should not be regarded as supporting the Wittgensteinian idea that what makes an expression formally proper to a fact is some real and non-conventional one-to-one picturing relation between the composition of the expression and that of the fact.

For I do not see how, save in a small class of specially chosen cases, a fact or state of affairs can be deemed like or even unlike in structure a sentence, gesture or diagram. For a fact is not a collection – even an arranged collection – of bits in the way in which a sentence is an arranged collection of noises or a map an arranged collection of scratches. A fact is not a thing and so is not even an arranged thing. (Ryle 1932b, 59)

But Ryle did not like the alternative either, viz. that an expression’s being formally proper to a fact is a purely conventional matter. He did believe that it was more a matter of convention than of nature.139 The startling conclusion that all philosophical talk is nonsensical, which Ryle identified as the main problem for Wittgenstein in the Tractatus140 was due, according to Ryle, to Wittgenstein being overinfluenced by his analogies between saying things and drawing maps. Since in order to tell the difference between significant and nonsensical propositions one would have to cross the boundary between them; it cannot be expressed in significant propositions. Ryle maintains that whereas it is indeed the case that a nurse can only depict the course of a patient’s temperature and cannot on a second paper depict the rules for representing his temperature, there is no reason to argue that in the same way the philosopher cannot say (that is significantly) ‘what it is that makes things said significant or nonsensical’ (Ryle 1951c, 254).141

Just as the nurse can tell, though not depict, how the temperature chart represents or misrepresents the patient’s temperature, so the philosopher can tell why, say, a scientist’s statement makes or does not make sense. What alone would be absurd would be a sentence which purported to convey a comment upon its own significance or meaninglessness. (Ryle 1951c, 254)

139 Cf. Chapter 2, pp. 45. 140 See the famous words in Tractatus 6.54: ‘My propositions are elucidatory in the following way: he who understands me finally recognises them as senseless, when he has climbed out through them, on them, over them. (He must so to speak throw away the ladder, after he has climbed up on it.)’ 141 In his paper from 1957 Ryle does not really take a stand, but makes the more neutral comment: ‘Critics quickly pointed out that Wittgenstein managed to say many important and understandable things. So perhaps the language of maps has limitations from which the language of words is exempt; and perhaps the notion of sense is wider than the notion of truth-or-falsehood to empirical fact.’ (Ryle 1957b, 264)

157

The Philosophical Development of Gilbert Ryle

Closely related to Ryle’s criticism of Wittgenstein’s idea that all philosophical talk is nonsensical are his annotations to 4.121 ‘That which mirrors itself in language, language cannot represent’ and 4.1212 ‘What can be shown cannot be said’. Ryle wonders in both cases: ‘why not?’ He claims that after the Tractatus Wittgenstein realized that his comparison between saying things and mapping things could not be pressed too far. Wittgenstein later retained the use of the analogy of mapping but because he felt no longer bound to put all elements of saying things in the rigid mould of mapping things, he was no longer forced to hold on to the idea that all philosophical talk is nonsensical. As a tradition of scholarship has demonstrated, it is far from easy to interpret Wittgenstein’s early talk of sense and significance.142 Since the exact details are not relevant to my discussion I will not attempt to fully explain them. What is of importance is that Ryle struggled with interpreting this difficult subject as well. As Rom Harré already showed, Ryle notes in his 1922 copy of the Tractatus, 3.21: ‘W. confused about sense = significance sense = direction’. In the same copy of the Tractatus Ryle further asks himself in the margin at 3.144 ‘(Names resemble points; propositions resemble arrows: they have sense)’ whether this sense is to be interpreted as direction. In his 1951 copy he points again at this confusion by showing that whereas Wittgenstein in 4.461 says: ‘Tautology and contradiction are without sense. (Like the point from which two arrows go out in opposite directions.)’, in 4.4611 he seems to claim the opposite: ‘Tautology and contradiction are, however, not nonsensical’. Ryle takes Wittgenstein to mean that they are without direction (sense), but not without significance. Ryle must have been thinking that Wittgenstein tried (but perhaps failed) to explain how propositions had sense (were true or false) in virtue of their having sense qua direction. In general a meaningful proposition has to have some structure or organization, as opposed to a name or a set of names. A typical form of organization would be determinacy as to sense qua direction – bRa rather than aRb is one example. This could lead to truth or falsity according to its correspondence (or lack of it) with the sense of the corresponding fact. Originally the sense of an arrow (in 3.144) is determined by the poles a and b, as can be found in ‘Notes on Logic’ (Wittgenstein 1913). If a proposition says that things go in the direction from a to b, and if they actually turn out to go in the b to a direction then the proposition is false. It is having polarity that marks out the proposition. Polarity and bi- polarity (being definitely either true or false) go together.143 This is what Wittgenstein may have tried to convey and Ryle possibly felt that the appeal to direction was unsatisfactory.144 Ryle presents another, more specific, objection in his ‘Ontological and logical talk in Wittgenstein’s Tractatus’, claiming that there are certain things we can say which cannot be analysed in terms of the truth-functions of their component propositions.

A question that we shall have to debate is whether all compound propositions are simple resultants or truth-functions of their components propositions, i.e. whether we could in principle say everything sayable, using for our conjunctions only ‘and’, ‘or’ and ‘not’. I think this doctrine

142 As Harré testifies, this is ‘exactly one of the points at which the lack of physics misled at least some Wittgenstein scholars. The point was exactly the adoptation by Wittgenstein of the second sense from physics. The sense of a vector is its direction, thus AB and BA are the same in magnitude but differ in sense (that is direction).’ (Harré 1999, 42) 143 Wittgenstein’s sense is not like Frege’s but he does think with Frege that there is a difference between a ‘beurtheilbarer Inhalt’ and an object: Julius Caesar cannot be asserted. Only Wittgenstein puts this by saying that ‘Caesar’ has Bedeutung and ‘Brutus killed Caesar’ has sense. 144 Professor Brian McGuinness was so kind as to explain to me the origins and implications of Wittgenstein’s remarks of sense as direction, some of which are also mentioned in Harré 1999. 158

Ryle and Wittgenstein

is false. (…) If I am right, then there are lacunas in Wittgenstein’s logical alphabet. He has not allowed himself the equipment with which to say certain sorts of sayable things. (Ryle 1993, 107)

Examples Ryle gives of propositions which cannot be analysed into merely p’s, q’s, and’s, or’s and not’s are: ‘If p then q’145 and ‘q, because p’. Neither does he belief that all general propositions – which are compound propositions as well – can be analysed into conjunctions or disjunctions of non-general propositions. Ryle’s objection is aimed at the idea Wittgenstein expresses in 3.3441: ‘We can, for example, express what is common to all notations for the truth-functions as follows: It is common to them that they all, for example, can be replaced by the notations of “~p” (“not p”) and “pvq” (“p or q”). (Herewith is indicated the way in which a special possible notation can give us general information.)’ With respect to Wittgenstein’s Philosophical Investigations Ryle was sceptical of Wittgenstein’s ‘hinted diagnosis’, although greatly appreciating his method of locating philosophical problems.

I do not think that anybody could read the Philosophical Investigations without feeling that its author had his finger on the pulse of the activity of philosophizing. We can doubt whether his hinted diagnosis will do; not that he has located, by touch, that peculiar and important intellectual commotion – philosophical puzzlement. (Ryle 1957b, 267)

I shall return to this point in the second part of this paper. During his lectures Ryle tried to expose or at least qualify some of the analogies Wittgenstein used in his later writing, such as the famous ones between using language and playing a game, and between learning to use language and learning to use a box of tools, the different tools sometimes looking or feeling the same though their jobs are very different. Ryle considers well used expressions to be more than tools. He claims that technical expressions such as ‘H2O’ are more like tools than ‘water’, ‘pipe’ or ‘shoe’, since the latter have become part of our personal life. And with respect to the analogy between using language and playing a game Ryle writes the following:

Wittgenstein’s own favoured example of game-playing (though he unfortunately stressed the noun rather than the verb) seems to me a clear case of an adverbial verb being misdiagnosed as itself a highly hospitable verb of doing. The report that during a certain period John was playing is not, as Wittgenstein seems to have thought, just vague, unspecific, elastic or disjunctive; it is unfinished. For it to be true there must have been a positive, concrete, or per se something that he was doing with a name and a description of its own, like wheeling a wheelbarrow or hitting a ball. (Ryle 1966-1967, 20)146

Ryle considers ‘playing’ to be an adverbial verb which needs a verb of doing, whereas he believes that Wittgenstein saw it as a vague verb of doing. Ryle also thought the analogy to be misleading in the sense that many readers seem to think that language is some kind of game. But the whole idea behind the tag ‘language-games’ is the notion of rules. The stress should have been more on rules and less on games in order to avoid misunderstandings.

145 This example may sound odd, since by using truth-tables ‘if p then q’ can be analysed into merely p’s, q’s, and’s, or’s and not’s. This, however, is only true of material implication. Ryle must have aimed at a non-material notion of ‘if p then q’ which cannot be analysed into the components mentioned above. 146 Kolenda does not explicitly date Ryle’s posthumously published paper ‘Adverbial Verbs and Verbs of Thinking’, but suggests that it was written during the last years of his life, after ‘Thinking and Reflecting’ (Kolenda in Ryle 1966-67).

159

The Philosophical Development of Gilbert Ryle

In On Thinking (1979) Ryle fully develops his idea of adverbial ascriptions and here he also criticizes Wittgenstein’s notion of ‘’ as applied to adverbial descriptions. The thinking of le penseur is an example of an adverbial notion. According to Ryle, there can be no such thing as thinking (at least the thinking of le penseur) per se; something positive or concrete must be done. And the difficulties we have with this notion of thinking is precisely that there is no homogeneous doing (X-ing) such that for a person to be thinking he must be X-ing. There are many considerably different things someone who is thinking may be doing – in fact, they differ too much to justify the use of the notion of family resemblance: ‘How multifarious and how patchy or thinly spread can family- likenesses between these postulated actions become, before the action-family itself evaporates into thin air?’ (Ryle 1979a, 21) Since adverbial ascriptions are not pointing to any separable doings, it would be a mistake to look for ‘family resemblance’ features. In his lectures on the later Wittgenstein in 1954-55, transcribed by Rom Harré, Ryle criticises Wittgenstein’s speaking of ‘using words and learning rules for using them; and using sentences and learning rules for them. And people speak of using sentences just like using words.’ (Harré 1999, 52) According to Ryle, using sentences is different from using words. First of all, we learn words and we make up sentences. There are no dictionaries of sentences. Second – and connected with the first objection – ‘when we say the meaning of a word is its rules of use or employment we cannot say the same for sentences.’ (Harré 1999, 52) Ryle here seems to refer to compositionality: if the meaning of a word lies in its rules of use, and if, additionally, sentences are made up out of words, the meaning of a sentence cannot be something like its rules of use (that is, if ‘rules of use’ is used in the same sense). (We could also argue – as Ryle does not – that the meaning of a sentence is its rule of use, but then the meaning of a word can no longer be – at least not in the same sense – its rules of use.) Third, there is a difference between words and sentences in that a word can be and is ‘used or misused in some sentence’ (Harré 199, 52), whereas this cannot be said about sentences. (But a sentence can be said to be used or misused in a greater context of sentences.) Finally, Ryle thinks that Wittgenstein cannot account for sentences which are being misused as clichés. He claims, in contrast to Wittgenstein, that the meaning of a word is principally different from the ‘meaning’ (sense) of a sentence.

A sentence differs from a word because it is the fact that someone has tried to employ words in such and such an order to say such and such a thing. A sentence is a different sort of abstraction from what a word is. It follows that ‘meaning of a sentence’ does not147 have the same sense as ‘meaning of a word.’ A sentence has a sense, and a word has meaning. (Harré 1999, 52)

Some of the objections concerning the Tractatus were, at least according to Ryle, later recognised by Wittgenstein himself. The others show Ryle’s disagreement with some of the ideas of Wittgenstein’s Philosophical Investigations. But in general he greatly appreciated Wittgenstein’s philosophical thought and acclaimed the great progress Wittgenstein achieved for philosophy.

Ryle’s attempt at tracing the origins of Wittgenstein’s thoughts Apart from Ryle’s many corrections to Ogden’s translation of the Tractatus, which will be discussed below, Ryle’s annotations to the Tractatus and Philosophical Investigations consist mainly of attempts to trace Wittgenstein’s thoughts back to other philosophers. In his copies of the Tractatus there are many references to Russell and Frege, but there are

147 ‘Not’ is not in Harré’s transcription, but in personal communication he confirms that this is a mistake. 160

Ryle and Wittgenstein also references to Plato’s Sophist and , Bergson, Kant, Ernst Mach, Heinrich Hertz (1857-1894), (1849-1923) (who was a student of Mach), Chomsky, Husserl and Meinong (especially to J. N. Findlay’s study of Meinong). Thus, Ryle wondered whether Wittgenstein had derived his interest in negation from Meinong – from 2.02331 until 2.11 almost every remark is accompanied by a reference to Meinong. And he thought that 5.641 (‘Thus there really is a sense in which philosophy can talk about the self in a non-psychological way’) and 4.1121 (‘Psychology is no nearer related to philosophy, than it is to any other natural science’) resembled Husserl’s philosophical position. Ryle attributes Wittgenstein’s ‘primitive propositions’ to Frege. In his paper on Ryle and the Tractatus, Harré mentions Ryle’s agreement with Harré’s suggestion of a Hertzian origin of the Tractatus. This claim is supported by Ryle’s inclusion of Hertz’s Principles of Mechanics in a Tractatus bibliography he compiled for his students, adding in the side-line: ‘(esp. introduction)’, and the comment he wrote on the last page of his copy of the Tractatus (1922 edition, 5th impression 1951): ‘Hertz (esp for Picture Theory?)’. And Ryle’s annotations include references to Hertz at 4.04, where Hertz is referred to by Wittgenstein himself, and 6.34. This link is thoroughly discussed by Harré in the first part of his paper ‘Wittgenstein: Science and Religion’ (2001). In opposition to the common view that the Tractatus is an extremely refined version of logical atomism, Harré argues that: ‘more of the Tractatus than has yet been realized can best be understood as a generalization of a number of important theses and doctrines developed in the writings of Helmholtz, Hertz and Boltzmann, a propos of the nature of physics as a way of creating a symbolic representation of the world.’ (Harré 2001, 211)148 Not only do Ryle’s annotations, once again, show his wide knowledge of the history of philosophy, but also his extensive study of Wittgenstein. His annotations contain references to the Tractatus, the early ‘Notes on Logic’ (1913), Wittgenstein’s Notebooks, lecture notes (Ryle probably meant The Blue and Brown Books) and to his letters to Russell. For example, Ryle suggests that 2.0201 (‘Every statement about complexes can be resolved into a statement about their constituents and into the propositions that describe the complexes completely’) is present in the Notebooks as well and is later criticised in The Philosophical Investigations. And he also traces 4.0312 (‘My fundamental idea is that the “logical constants” are not representatives; that there can be no representatives of the logic of facts’) back to Wittgenstein’s Notebooks (Dec 1914, p. 37). He wonders why 6.113 (‘It is the characteristic mark of logical propositions that one can perceive in the symbol alone that they are true; and this fact contains in itself the whole philosophy of logic (…)’)

148 I illustrate the resemblances between Hertz and Wittgenstein which were suggested by Ryle by quoting three passages from Hertz’s introduction to his The Principles of Mechanics (1899). (1) We form for ourselves images or symbols of external objects; and the form which we give them is such that the necessary consequents of the images in thought are always the images of the necessary consequents in nature of the things pictures. In order that this requirement may be satisfied, there must be a certain conformity between nature and our thought. (Hertz 1899, 1) (2) We shall denote as incorrect any permissible images, if their essential relations contradict the relations of external things (…). (Hertz 1899, 2) The following quotation will show a remarkable resemblance between the Hertzian method of dealing with meta- scientific questions and Wittgenstein’s (and Ryle’s as well) views on the nature of philosophical method: (3) ‘But we have accumulated around the terms “force” and “electricity” more relations than can be completely reconciled amongst themselves. We have an obscure feeling of this and want to have things cleared up. Our confused wish finds expression in the confused question as to the nature of force and electricity. But the answer which we want is not really an answer to this question. It is not by finding out more and fresh relations and connections that it can be answered; but by removing the contradictions existing between those already known, and thus perhaps by reducing their number. When these painful contradictions are removed, the question as to the nature of force will not have been answered; but our minds, no longer fexed, will cease to ask illegitimate questions.’ (Hertz 1889, 8) The first two quotes are also given in Harré 2001, to which I also refer for additional quotations and arguments.

161

The Philosophical Development of Gilbert Ryle appeared in Wittgenstein’s Letters to Russell but not in his ‘Notes on Logic’. Ryle further traces Wittgenstein’s use of ‘primitive propositions’ back to Letters 19 and 20 from 1913. He refers several times to letter 20, which he obviously considers to be an important testimony of Wittgenstein’s early development. The letter contains Wittgenstein’s answers to some questions by Russell on Wittgenstein’s ‘Notes on Logic’. Ryle refers to it in his annotations to 1.1 and to 5 and 6 (5.472 and 5.476, 5.556, 6.113, 6.1232 and 6.1271 – which Ryle takes to concern the totality of facts, the function and application of logical operation-signs, and the propositions of logic). Ryle’s references to Russell’s Theory of Knowledge and Wittgenstein’s own ‘Notes on Logic’, which are to be found in Ryle’s annotations of the Tractatus, are numerous and he clearly considered both writings of great importance to Wittgenstein’s early development. Therefore, Ryle’s ideas of their influence on Wittgenstein’s Tractatus will be discussed below in two separate sections. In Ryle’s annotations of the Philosophical Investigations, there is naturally less stress on tracing Wittgenstein’s thoughts back to Russell and Frege and more on pointing at similarities to Husserl and Moore. He refers also to Plato, Augustine, Kant, Schopenhauer, Ramsey and James. Ryle, for example, recognizes Moore in part 2 of the Investigations, paragraph 217e: ‘Meaning is as little an experience as intending’, and in 218e: ‘Meaning is not a process which accompanies a word. For no process could have the consequences of meaning’. Husserl – though always accompanied by question marks – is also often referred to. Ryle, for example, wonders whether the following part of paragraph 437 can also be found in Husserl:

A wish seems already to know what will or would satisfy it; a proposition, a thought, what makes it true – even when that thing is not there at all! Whence this determining of what is not yet there? This despotic demand? (“The hardness of the logical must.”)

And a remark which Ryle partly traces back to Husserl can be found on page 217 (part 2):

But what distinguishes them [meaning and intending] from experience? – They have no experience-content. For the contents (images for instance) which accompany and illustrate them are not the meaning or intending.

Furthermore, in his Collected Papers Ryle several times refers to a similarity between Husserl and Wittgenstein with respect to their use of ‘logical grammar’ and ‘logical syntax’, e.g. (Ryle 1957a, 363-64 and 1961a, 413) Many of Ryle’s annotations in the Investigations contain references to the Tractatus. Ryle noted that the target of some of Wittgenstein’s remarks in the Investigations had been his own Tractatus, e.g. of 23:

But how many kinds of sentences are there? Say assertion, question, and command? – There are countless kinds: countless different kinds of use of what we call “symbols”, “words”, “sentences”. And this multiplicity is not something fixed, given once for all; but new types of language, new language-games, as we may say, come into existence, and others become obsolete and get forgotten. (Wittgenstein 1953, 23

He also mentioned other remarks which were not aimed at rejecting the Tractatus but which could be traced back to it, or to Wittgenstein’s lecture notes (probably The Blue and Brown Books). For example, Ryle traces 236 – ‘Calculating prodigies who get the right answer but cannot say how. Are we to say that they do not calculate? (A family of cases.)’ – back to Wittgenstein’s lecture notes. 162

Ryle and Wittgenstein

Russell’s Theory of Knowledge Ryle’s copy of the 1961 translation of The Tractatus contains many references to Russell’s Theory of Knowledge, which was at the time Ryle wrote these annotations still a (partly) unpublished manuscript, written by Russell in the spring and summer of 1913. The first six chapters had been published in The Monist in the period January 1914-April 1915 but then, after being heavily criticised by Wittgenstein, Russell decided not to publish the rest of it. The (complete) manuscript was discovered in 1968 upon transferring Russell’s papers to McMaster University, Hamilton, where they are still kept, and was published in 1983 in Volume 7 of The Collected Papers of Bertrand Russell. David Pears recalls requesting a photocopy in 1972 and providing Ryle with a copy.149 In Ryle’s first comment in his copy of the 1969 edition of the Tractatus, he refers to the availability of part of Russell’s Theory of Knowledge in Xerox, probably the copy he had received from Pears. Pears claims that in his earlier seminars Ryle said that he suspected a Russellian origin of Wittgenstein’s remark in the Tractatus 4.441: ‘There are no logical objects’, but was never able to determine it exactly. According to Pears, the typescript of Russell’s Theory of Knowledge vindicated Ryle.150 Ryle’s annotations to the Tractatus betray a thorough familiarity with Russell’s early typescript. They suggest that Wittgenstein was influenced by it both positively and negatively. Ryle claims, for example, that in his Theory of Knowledge Russell already talked about logical forms of complexes and logical propositions – meaning propositions of logic. But he was also frequently the target of Wittgenstein’s attacks. Some of the positive influences concern similarities in examples or expressions which Wittgenstein and Russell both used – e.g. propositions of the form ‘aRb’ frequently occur in Wittgenstein and in Russell’s manuscript, as well as the proposition ‘b is a successor of a’ which, for example, appears in 4.1273. Other observations concern the possibility of tracing back some of Wittgenstein’s basic statements to Theory of Knowledge. Ryle asks himself for example whether 2.222 (‘The agreement or disagreement of its sense with reality constitutes its truth or falsity’) and parts of 4.023 (‘A proposition must restrict reality to two alternatives: yes or no’) and 4.024 (‘To understand a proposition means to know what is the case if it is true (One can understand it, therefore, without knowing whether it is true)’) can already be found in Russell’s Theory of Knowledge. The same question is asked about 4.121 (‘Propositions cannot represent logical form: it is mirrored in them’). And ‘elementary propositions’, for example in 4.21, is similar to Russell’s ‘atomic propositions’. In addition, Ryle wonders whether in his Theory of Knowledge Russell was talking about ‘the general propositional form’ and the general form of a truth-function, which can be found in the Tractatus 4.5, 5.47 and 6. Last but not least, Ryle noticed that ‘sense=direction’ was already present in Russell’s early work151, which suggests that Wittgenstein might have borrowed the idea from Russell. Something in which Wittgenstein was not influenced by Theory of Knowledge was his notion of formal concepts (e.g. in 4.126). As I mentioned above, Ryle considered Russell, and more specifically Russell’s Theory of Knowledge to be the target of Wittgenstein’s remark in 4.441 and 5.4: ‘There are no logical objects’. But this was not the only criticism which Ryle found to be aimed at Russell. The question that according to

149 Personal communication with Pears in 2003 and 2004. 150 Ibid. 151 Ryle suggests in his 1922 copy that the idea that sense is direction originated from Bertrand Russell’s Nature of Truth and chapter 27 of Principles of Mathematics; in his 1951 copy he suggests that the idea originated from chapter 27 of Principles of Mathematics and chapter 12 of The Problems of Philosophy (1912). In his 1969 copy he traces it back to Russell’s Theory of Knowledge.

163

The Philosophical Development of Gilbert Ryle

Wittgenstein ‘cannot be asked’ in 4.1274, namely: ‘Are there unanalysable subject- predicate propositions?’ was, according to Ryle, asked in Russell’s Theory of Knowledge. And in Ryle’s view, Wittgenstein’s reference in 5.4731 to the notion of self-evidence about which Russell talked so much could be to the Theory of Knowledge, in which he devoted a complete chapter to self-evidence. Finally, Ryle considers Wittgenstein’s criticism to Russell in 5.5422 – viz. that his theory did not satisfy the criterion of explaining the form of the proposition ‘A makes the judgement p’ in such a way that ‘it shows that it is impossible to judge a nonsense’ – to be aimed at Russell’s Theory of Knowledge. This seems indeed plausible when taking into account Wittgenstein’s letter to Russell from June 1913. At the time of Wittgenstein’s letter Russell was still working on his Theory of Knowledge. Wittgenstein wrote:

I can now express my objection to your theory of judgement exactly: I believe it is obvious that, from the prop[osition] “A judges that (say) a is in the Rel[ation] R to b”, if correctly analysed, the prop[osition] “aRb. V .~ aRb” must follow directly without the use of any other premiss. This condition is not fulfilled by your theory. (Wittgenstein 1974, R.12)152

Notes on Logic Both in his 1951 copy of the Tractatus and in a short unpublished typescript – which is kept in the ‘Red Box’ in Linacre College Library and wrongly catalogued as ‘Notes on the Investigations’ – Ryle pays attention to Wittgenstein’s ‘Notes on Logic’ from 1913. In Ryle’s unpublished comments on ‘Notes on Logic’, which have not been discussed so far, he mainly draws parallels between this work and the Tractatus, which was written a few years later. ‘Notes on Logic’ exists in several different versions. The so-called ‘Russell Version’ (catalogued as 201a-1 of the Wittgenstein Nachlass and first published in 1979) consists of a summary dictated by Wittgenstein in English and four ‘Manuscripts’ which Russell had translated into English himself. The ‘Costello Version’ (catalogued as 201a-2 and first published in 1957) – copied by Harry T. Costello from Russell when the latter came to Harvard as a visiting lecturer in the spring of 1914 – is a subsequent rearrangement of the text by Russell.153 These versions consist of several items which were written at different times, which makes it plausible that there have been more than these two versions. And it has been established that a typescript of the ‘Russell Version’ which does not match item 201a-1 of the Nachlass (nor item 201a-2), is kept at the Wittgenstein Archives at the University of Bergen.154 Ryle’s annotations, which mainly concern Wittgenstein’s remarks concerning the nature of philosophy, theory of knowledge and logic, seem not to be based on either version.155 His reference to specific pages in the ‘Notes on Logic’ does not correspond to the pages in either the ‘Russell Version’ nor the ‘Costello Version’. For example, 4.111 says: ‘The word “philosophy” must mean something which stands above or below, but not beside the natural sciences.’ Ryle sees close resemblances with a remark on page 15 of ‘Notes on Logic’, but the only resemblance I can find is with ‘The word “philosophy”

152 Again, the exact details are not relevant to my discussion and I will not attempt to fully explain them. 153 See G. H. von Wright, Wittgenstein (Oxford 1982), 54, and Brian F. McGuinness, ‘Bertrand Russell and Ludwig Wittgenstein’s “Notes on Logic”’, Revue Internationale de Philosophie 26, 1972. 154 In his working paper ‘Editing Wittgenstein’s “Notes on Logic”’, Michael A.R. Biggs refers to this typescript as TSx, which is the title I will use as well. He kindly sent me the typescript to check it with Ryle’s references. 155 Ryle’s comments contain a reference to a ‘diagram on the back’, similar to the one in the Tractatus, 6.12. Since the ‘Costello Version’ does not contain this diagram, but the ‘Russell Version’ does, Ryle most likely used a further compilation of the Russell-version. 164

Ryle and Wittgenstein ought always to designate something over or under, but not beside, the natural sciences’, which occurs on page B21 of the ‘Russell Version’ and on page B23 of the ‘Costello Version’. Similarly, Ryle traces 4.1121 (‘The theory of knowledge is the philosophy of psychology’) back to page 15. However, the remark corresponds to ‘Epistemology is the philosophy of psychology’, which occurs on page B21 of the ‘Russell Version’ and page B22 of the ‘Costello Version’. ‘Philosophy limits the disputable sphere of natural science’ is also traced back to page 15, but cannot be found there. The content of 5.451 (‘If logic has primitive ideas these must be independent of one another. If a primitive idea is introduced it must be introduced in all contexts in which it occurs at all’) is found in the ‘Russell Version’ on page B6 and in the ‘Costello Version’ on page B20: ‘The ways by which we introduce our indefinables must permit us to construct all propositions that have sense from these indefinables alone.’ (201a-1 B6 and 201a-2 B20) However, Ryle refers to it as ‘on page 14’. Typescript TSx does not match Ryle’s annotations either. This strongly suggests that Ryle had access to yet another version. It is not unlikely that slightly different copies, perhaps even different arrangements, of these two versions were circulating in Oxford and Cambridge at the time. It is difficult to determine precisely when Ryle first had access to ‘Notes on Logic’, especially since it is not clear which version he used. David Pears recalls discussing ‘Notes on Logic’ – or more accurately, the chapter on Logical Objects – with Ryle in the late 1960s and says that he thinks it unlikely that Ryle had already seen ‘Notes on Logic’ in 1948 when he first gave his seminar on Wittgenstein in Oxford. Ryle’s annotations in his copy of Wittgenstein’s Tractatus which concern ‘Notes on Logic’ have probably been inserted after 1974. He occasionally talks about the Tractatus and ‘Notes on logic’ in relation to Wittgenstein’s letters to Russell, which were not published until 1974. Although the 1960 edition of Wittgenstein’s Notebooks 1914-1916 did contain an appendix with extracts from Wittgenstein’s letters to Russell, Ryle’s reference to specific letters, such as R 12, indicates that he was using the later edition from 1974.156 Ryle’s references to ‘Notes on Logic’ in his 1951 copy of the Tractatus differ both in type and in importance. Sometimes Ryle wanted to point at a difference in terminology, such as that in ‘Notes on Logic’ Wittgenstein used ‘type’ instead of ‘form’ – which was used in the Tractatus (e.g. in 3.315 and 4.1241) – and ‘denote’ instead of ‘signify’ (used in the Tractatus). In ‘Notes on Logic’ Wittgenstein used ‘indicate’ or ‘mean’ where he would later in the Tractatus use ‘designate’, e.g. in 4.063 and 4.111. Ryle also tried to find out which concepts were already present in ‘Notes on Logic’ and which ones were not. The concept ‘prototype’ – present in 3.315 – does not occur in ‘Notes on Logic’, neither do ‘picturing’, ‘showing’, ‘logical space’, ‘ethics’, ‘the unsayable’, ‘operations’, ‘world’, ‘Sachverhalt’, ‘mathematics’, ‘logical synthese’, ‘atomic propositions’, ‘colours’, ‘tautology’ (which however did occur in Wittgenstein’s letters to Russell from 1913) and ‘intentionality’. Especially the first category shows Ryle’s preoccupation with language and precise formulations.

156 However, all three copies of the Tractatus contain the statement by Linacre College Library that Ryle donated them upon his retirement in 1968. This would mean that Ryle could not have written his comments in the margins of these books after 1968. This problem can easily be solved. The librarian of Linacre College did not consider it to be unlikely that although the note in the front cover of the books suggest that they are donated in 1968, they were in fact part of the second donation, which took place in 1976. Another explanation could be that he did donate the books in 1968 but later borrowed them from the library – there is evidence that Ryle later borrowed several of what once used to be his books from the libraries to which he had donated them.

165

The Philosophical Development of Gilbert Ryle

It has to be noted that Ryle’s annotations in the Tractatus which refer to Wittgenstein’s ‘Notes on Logic’, are almost exclusively limited to index numbers 3, 4 and 5. Ryle notices for example that 4.061-4.0621 almost literally contain the remarks on page 2 of ‘Notes on Logic’157. The Tractatus remarks with index number 1, 2, 6 and 7 were not referred to by Ryle in his attempt to trace part of the Tractatus back to ‘Notes on Logic’. There is one exception: Ryle namely wonders why 6.113, that ‘it is the characteristic mark of logical propositions that one can perceive in the symbol alone that they are true; and this fact contains in itself the whole philosophy of logic’, only occurs in Wittgenstein’s letters to Russell and cannot be found in ‘Notes on Logic’. Wittgenstein’s criticism of Russell in 5.5422 ‘The correct explanation of the form of the proposition “A judges p” must show that it is impossible to judge a nonsense. (Russell’s theory does not satisfy this condition.)’ – which according to Ryle is probably aimed at Russell’s Theory of Knowledge – also occurs in ‘Notes on Logic’:

In my theory propositional formula p has the same meaning as propositional formula not-p but opposite sense. The meaning is the fact. The proper theory of judgment must make it impossible to judge nonsense. (Wittgenstein 1913, 201a-1; A4)

As Ryle shows, more of Wittgenstein’s criticisms of Russell can also be found – although sometimes not yet in the same form – in ‘Notes on logic’, such as the one in 4.441 that there are no logical objects.

The assumption of the existence of logical objects makes it appear remarkable that in the sciences propositions of the form “p < V> q”, “p < ⊃> q”, etc. are only then not provisional when “ < V> ” and “< ⊃> ” stand within the scope of a generality sign [apparent variable]. (Wittgenstein 1913, 201a-2; B16)

In his days, Ryle was the only one, or at least one of a very few, who considered ‘Notes on Logic’ to be important for Wittgenstein’s early development – it was generally neglected. Today Anthony Palmer, who was a friend and student of Ryle, maintains that ‘Notes on Logic’ can be used to show that the Russell interpretation of the Tractatus – which is, according to Palmer, still widely accepted and can partly be held responsible for the idea of two Wittgenstein’s – is unacceptable. Palmer uses ‘Notes on Logic’ to support the interpretation of and the ‘New Wittgensteinians’ that there is only one Wittgenstein.158 Wittgenstein’s early paper is supposed to show that a proposition and the state of affairs it represents must not be construed as a relation. ‘Propositions by virtue of sense cannot have predicates or relations.’ (Wittgenstein 1913, 201a-2; B6) Palmer claims that for Wittgenstein ‘the only things that stand in a relation to each other are the signs ‘a’ and ‘b’ and possibly a and b themselves’ (Palmer 2002, 8), which is supported by Wittgenstein’s discussion of symbolism: ‘more correctly: that a certain thing is the case in the symbol says that a certain thing is the case in the world’ (Wittgenstein 201a-2, A7). This description does not even suggest a relation. This would render the Russell interpretation of the Tractatus implausible because this interpretation clearly depends on the idea that Wittgenstein tried to establish a relation between propositions and the world.

157 Which seem to be page B2 and B3 both in the ‘Russell Version’ and in the ‘Costello Version’. 158 Cf. Winch 1969 and Crary and Read 2000. 166

Ryle and Wittgenstein

The English Translation of the Tractatus Both in ‘Ludwig Wittgenstein’ and in ‘review of Ludwig Wittgenstein: Remarks on the Foundations of Mathematics’, Ryle refers to the unreliability of the English translation of the first edition of the Tractatus (1922) by Ogden and Ramsey. In addition, Ryle’s annotated copy of the Ogden-Ramsey translation consists of many alternative translations of words and sentences. And he considered them important enough to start each class with dictating his corrections to the translation. (Harré 1999, 41) Ryle was not so critical of Anscombe’s translation of the Philosophical Investigations, suggesting only some minor alterations, such as ‘designate’ instead of ‘signify’ (and sometimes ‘mean’ as well) throughout the book, e.g. in paragraph 15. And several times, he suggests to replace ‘when’ by ‘if’, e.g. in 105 and 577. Ryle would have preferred ‘intended’ instead of ‘wanted’ in paragraph 76, ‘’ instead of ‘’ in 47, and he thinks that ‘without right’ in paragraph 289 can be better expressed by ‘wrongly’. He further suggests to replace ‘with’ by ‘to’ in 248 and in 265 to change ‘into a word Y’ into ‘by a word Y’. And in the last part of 659 and the first part 660 ‘going’ should be replaced by ‘meant’. Since Ryle had far more criticisms on the translation of the Tractatus, in the remainder of this section I shall focus on this earlier work. Although there are records of Wittgenstein’s own corrections of the English translation, they were unknown for a long time to philosophers working on the Tractatus. In 1960 Russell wrote the following to Ogden:

I learn that questions have arisen as to the authenticity and authority of the English version of Wittgenstein’s Tractatus. I know that this version was sanctioned point by point by Wittgenstein. There are places where it is not an exact translation of the German. When I pointed this out to him, he admitted it, but said that the translation as it stood expressed what he wished to say better than a more exact translation. It is, of course, open to anybody to make a new translation in a more modern idiom, but it would be misleading to suggest that such a translation gave a more accurate rendering of Wittgenstein’s thought at the time than that which was published. I say this from recollection of careful and minute discussion with Wittgenstein as to what he wished the English version to say. (Russell 1973, 10)

In the absence of contemporary evidence, it is not known when this ‘careful and minute discussion’ between Wittgenstein and Russell took place.159 However, Russell’s claim that Wittgenstein sanctioned the English translation and deliberately chose not to make a literal translation is supported by Wittgenstein’s letters to Ogden in 1922 together with his comments on early versions of the translation. On 23 March Wittgenstein writes:

The translation as you said, was in many points by far too literal. I have very often altered it such that now it doesn’t seem to be a translation of the German at all. I’ve left out some words which occur in the German text or put in others which don’t occur in the original etc. etc. But I always did it in order to translate the sense (not the words). (Wittgenstein 1922b, 19)

Ryle, like most other philosophers, was probably unaware of the fact that Wittgenstein himself had worked on the Ogden-translation. He would probably not have made at least some of his comments if he had known that the English was not meant as a ‘literal’ translation of the German original. Ryle sometimes criticised precisely those translations that were suggested by Wittgenstein himself in 1922. For example, in 3.141 Wittgenstein himself had suggested

159 Cf. Von Wright 1973 (pp. 10-11).

167

The Philosophical Development of Gilbert Ryle

‘mixture’ instead of ‘medley’ and ‘articulate’ instead of ‘articulated’, translations which were later criticised by Ryle.

I should propose to put “mixture” instead of “medley”. (…) In the end of that prop[osition] couldn’t one say “is articulate” instead of “articulated”? I didn’t mean yet to say that the prop[osition] is articulatED but I used the word “artikuliert” in the sense in which one might say that a man speaks articulate that is that he pronounces the words distinctly. Or do you in that case also say “articulated”? If so leave it as it stands, if not you put “articulate”. (Wittgenstein 1922b, 24)

Whereas Ryle suggests to use ‘a priori’ in 4.411, Wittgenstein writes in 1922: ‘“von vornherein” doesn’t mean “a priori”. It should be something like: “It is probable from the very beginning that the bringing in of …” or “It seems probable even on first sight…”’ (Wittgenstein 1922b, 29) And in 5.1361 Ryle replaces ‘Superstition is the belief in the causal nexus’ by ‘Belief in the causal nexus is superstition’.160 Here Ryle, however, is making a mistake as evidenced by the following quote:

“Belief in the causal nexus is superstition” isn’t right. It ought to be: “Superstition is the belief in the causal nexus”. I didn’t mean to say that the belief in the causal nexus was one amongst superstitions but rather that superstition is nothing else than the belief in the causal nexus. In the German this is expressed by the definite article before “Aberglaube”. (Wittgenstein 1922b, 31)

Interestingly, sometimes both Ryle and Wittgenstein suggest the same (sort of) corrections of the English translation. Wittgenstein had already criticised the obvious overabundance of ‘the’s’, though not as extensively as Ryle later did. E.g. in 4.464 the original translation contained ‘the tautology’ and ‘the contradiction’. Wittgenstein changed it into: ‘The truth of tautology is certain, of propositions possible, of contradiction impossible’ (Wittgenstein 1922b, 30). The same applies to 5.143 and 6. Even after Wittgenstein got rid of several ‘the’s’, Ryle still found many to be thrown out. In Harré (1999) some of Ryle’s comments concerning the translation were already mentioned, e.g. the many misplaced ‘only’s’ – e.g. in 6.3611 Ryle suggests ‘possible only’ instead of ‘only possible’. This is also true for Ryle’s annotations concerning issues of particular relevance for the interpretation of the Tractatus. In 2.03 (1922 and 1951) Ryle suggests replacement of ‘members’ as a translation of the German ‘Glieder’ by ‘links’. And he proposes to distinguish between ‘abbilden’ and ‘darstellen’ which are often both translated with ‘represent’ (by the early Ogden-Ramsey translation as well as the later Pears-McGuinness one). Ryle suggests the more concrete ‘depict’ for ‘abbilden’ and the more abstract ‘display’, ‘exhibit’ or ‘convey’ for ‘darstellen’, in order to make the English text less ambiguous. Some additional interpretative corrections which I found are the following: in 2.0271 Ryle suggests ‘constant’ instead of ‘fixed’ and ‘mutable’ instead of ‘variable’, and in 4.115 he prefers ‘unsayable’ over ‘unspeakable’. Most of Ryle’s corrections indeed seem to be improvements. However, he sometimes contradicts Wittgenstein’s own earlier suggestions. In any case, Ryle’s annotations on the translation of the Tractatus show an interest in – and eye for – what Wittgenstein had to say at a remarkably detailed level.

160 See also Harré 1999 (p. 42). 168

Ryle and Wittgenstein

Ryle in Relation to Contemporary Wittgenstein Scholarship It is interesting to note that Ryle’s interpretation of ‘the early Wittgenstein’ seems to be closer to the interpretation of the ‘New Wittgensteinians’ headed by Conant and Diamond, than to the standard interpretations of his time, which generally depicted Wittgenstein’s Tractatus as metaphysical realism and his later work as a rejection of this position. The ‘New Wittgensteinians’ argue that there is only one Wittgenstein rather than two, attempting to provide an overall coherent interpretation which fits his early as well as his later writings. They claim that his ontological claims are not really ontological claims and, following this line of thought, that the picture theory is not a theory but rather a method, which is – as I have shown above – what Ryle argues too. Ryle’s interpretation of the ontological story of the Tractatus as some sort of prefatory parable seems to be close to the view of the ‘New Wittgensteinians’:

the book presents us with metaphysical sentences which lead us to participate in an imaginative activity of articulating the structure of the illusion of an external standpoint on language – an imaginative activity through which we can come to recognize that illusion as an illusion. (Crary and Read 2000, 13)

Like Diamond and Conant, Ryle maintains that Wittgenstein did not aim at providing us with ontological knowledge, but used his ontological claims for another purpose. Ryle thought that making us realize that the ontological story will not do was not the only thing Wittgenstein tried to achieve; he replaced the ontological story with a propositional one which would do. However, Ryle would not have agreed with the interpretation of the ‘New Wittgensteinians’ that Wittgenstein merely provides a therapeutic method and no arguments. This interpretation is offered not only by the ‘New Wittgenstein’ movement, but also by more traditional interpreters such as Bouwsma, as we will see in the second part of this chapter, Baker and Waismann. They make it seem as if there is a strict dichotomy between therapeutic method and metaphysical claims, with nothing in between. Whereas the philosophers who believe there are two Wittgensteins think that the first was building a metaphysical theory and was making claims, and the second was only developing a therapeutic method without arguing or refuting claims (which I call the traditional interpretation), the ‘New Wittgensteinians’ argue that there never was a Wittgenstein who was a metaphysical theory builder. According to this interpretation, Wittgenstein has always been a therapist. The ‘New Wittgensteinians’ understand Wittgenstein as ‘aspiring, not to advance metaphysical theories, but rather to help us work ourselves out of confusions we become entangled in when philosophizing’ (Crary 2000, 1). They regard his method to be a purely therapeutic one. Ryle, however, agrees with Hacker and others that Wittgenstein did argue, refute and rectify claims, which does not mean, as we will see later, that Ryle considers Wittgenstein to be making metaphysical claims in the Tractatus or in his later writings. The use of reductio ad absurdum arguments is a way of rectifying and refuting but this does not mean that one makes ontological claims. Wittgenstein’s refutation of Russell and Frege also supports such an interpretation.

169

The Philosophical Development of Gilbert Ryle

A Comparison Between Ryle and Wittgenstein

Though similarities between Ryle and Wittgenstein have been noticed, there is more to be said on this. Apart from agreeing in their rejection of Cartesian dualism, their development of a non-denotational concept of meaning, and other similarities which will be mentioned below, they also frequently use the same examples. Ryle’s well-known example of the chess-player in The Concept of Mind can already be found in Wittgenstein’s Blue and Brown Books.161 And his use of engineering examples – such as in ‘Philosophical Arguments’ (Ryle 1945, 198) where Ryle describes the reductio ad absurdum as having something in common with the destruction-tests by which engineers discover the strength of materials – can also be traced back to Wittgenstein.162 Furthermore, Ryle’s description in ‘Philosophical Arguments’ (1945) of the task of philosophy as ‘the charting of the logical powers of ideas’ and the metaphor he uses for our workaday knowledge of the geography of our ideas, namely that ‘People often know their way about a locality while being unable to describe the distances or directions between different parts of it or between it and other localities’, are reminiscent of the Blue Book (1958, 57).163 In what follows I shall discuss the interpretation of Hacker and Bouwsma who have examined the relationship between the two philosophers in some detail. I will also point at the similarities between Ryle’s later writings on thinking and on the nature of philosophy and Wittgenstein’s views on these subject. It is especially in these papers – more than in The Concept of Mind or Ryle’s earlier writings – that Ryle comes close to Wittgenstein.

Peter Hacker: the Main difference Between Ryle and Wittgenstein is Depth Hacker claims that Ryle’s interest in the distinction between meaningful and nonsensical statements originated from studying Wittgenstein’s Tractatus and was probably further developed by his contact with Wittgenstein. Their conceptions of philosophy bear recognizable resemblances. Ryle’s conception of philosophy as the mapping of the ‘logical geography’ and ‘cross-bearings’ of concepts is similar to Wittgenstein’s search for a clear ‘representation of our use of words in a given domain of discourse.’ (Hacker 1996, 168) Hacker further thought that both philosophers were similar in their detailed methods of unmasking philosophical confusion by pointing to conceptual differences in previously assumed uniformities. Hacker mentions several resemblances between the subjects treated by both philosophers.

There are numerous parallels between Ryle’s treatment of problems and Wittgenstein’s – for example, the discussion of intelligent performances and the repudiation of the dual-process conception of thoughtful, intelligent activity; the analysis of understanding, partial understanding and misunderstanding; the attack on the myth of volitions; and the repudiation of the traditional picture of self-knowledge and introspection. (Hacker 1996, 169)

Ryle not merely imitated Wittgenstein but ‘applied these ideas in his own inimitable and brilliant way’ (Hacker 1996, 170).

161 The example of the chess-player was also used in the Blue Book (Wittgenstein 1958, pp. 7 and 13) and in the Brown Book (Wittgenstein 1958, pp. 77, 84, 147, 183-4 and 185). 162 Cf. Harré 1999 (p. 42). 163 Cf. Hacker 1996 (p. 149). 170

Ryle and Wittgenstein

There are also differences. The most important one, according to Hacker, is a difference in depth and subtlety, ‘but no deep difference’ (Hacker 1996, 168). He thinks it is due to Ryle’s ‘superficiality’ that he repudiated the therapeutic aspect of Wittgenstein’s philosophy and that he did not share Wittgenstein’s aversion against characterizing certain things philosophers say as true or false. Ryle thought that one could speak of philosophers’ statements being true or false though not in the same sense in which scientific statements are true or false, just as one can talk about theories in philosophy, as long as one does not hold that they are similar to scientific ones. This shows, Hacker thinks, that Ryle remained at the surface and never touched the underlying questions, such as ‘if we deny that a philosophical proposition could be false, what is our explanation of why this is so?’. Another difference is what Hacker calls Ryle’s ‘excessive reliance’ (Hacker, 168) on the idea of a category-mistake in The Concept of Mind, which is not to be found in Wittgenstein, although the latter does occasionally speak of categories and confusions of categories.164 Wittgenstein does not, as Ryle does, speak of a general notion of a category- mistake, but looks from case to case how philosophical confusion arises from applying the grammar of one concept to that of another. Ryle’s sometimes relatively quick, impatient and superficial dismissal of the Cartesian myth has no counterpart in Wittgenstein. Wittgenstein wanted to do justice to the temptations to stick with a certain picture and could not bring himself to homogenize matters and simply get rid of them. Hacker interprets Wittgenstein’s only recorded remark on The Concept of Mind – ‘all the magic has vanished’165 – as a critique on Ryle’s failure to do justice to the temptations of the Cartesian myth. As a result of Wittgenstein’s aversion against homogenizing issues, his writings occasionally contain, as Hacker puts it, ‘bewilderingly truncated arguments’ (Hacker, 169) as well as opaque counter-moves, which are not found in Ryle. Hacker further shows that there is no complete overlap of the subjects which Ryle and Wittgenstein discussed. Wittgenstein’s writings, but not Ryle’s, contain the private language argument as well as a ‘detailed discussion of the asymmetry between the first- and third-person, present-tense psychological utterances in terms of the absence of criteria in the first-person case and the need for criteria in the third-person case’ (Hacker, 170). Ryle’s work did not invoke the notion of a criterion, as Wittgenstein’s did, neither did he pay attention to certain issues of philosophical psychology that were addressed by Wittgenstein, e.g. aspect perception. Moreover, Ryle’s work lends itself more easily to a behaviouristic interpretation than Wittgenstein’s writings. What Hacker calls Ryle’s ‘tendency to drift incautiously towards a behaviouristic position’ cannot be found in Wittgenstein (some commentators notwithstanding).166

O. K. Bouwsma: Ryle used Arguments whereas Wittgenstein did not At a symposium on Ryle in the summer of 1972 O. K. Bouwsma presented his paper ‘A Difference between Ryle and Wittgenstein’. Roughly, he claims that whereas Ryle’s

164 For example in MS 135 Wittgenstein writes about the problems in philosophical psychology and says: ‘It is a mixture of categories’. (This remark was written as late as 1947.) 165 This remark can be found in the unpublished version of Bouwsma’s notes on his conversations with Wittgenstein. 166 While Hacker does not believe that Ryle should be regarded as a logical behaviourist, he thinks that Ryle himself was partly to blame for this persistent misinterpretation of his work. His ‘strategic error’ (Hacker 1996, 172) was his own characterization of what he was fighting against in The Concept of Mind. His fight against the ‘ghost in the machine’ can easily lead people to think that he was defending the view that all mental predicates signify behaviour. Cf. Hacker 1996 (p. 170).

171

The Philosophical Development of Gilbert Ryle

Concept of Mind provides us with ‘knowledge that’, Wittgenstein’s Philosophical Investigations is concerned with showing us a method, teaching us ‘knowledge how’, a skill. Bouwsma thinks that clues for this interpretation can be found in the fact that, contrary to Ryle, Wittgenstein did not provide proofs, did not use arguments or refutations and rectifies nothing. The difference in style plays a role as well, The Concept of Mind being ‘a well-ordered book proceeding from one subject to another in a rational order and without breaks’, and The Philosophical Investigations a work without a clear beginning or end; Bouwsma calls it ‘an endless book’ (Bouwsma 1972, 82). Bouwsma claims that what Ryle calls a mistake, Wittgenstein describes as the expression of a linguistic confusion. Philosophical problems are caused in particular by two things, namely: (1) a certain feature of language, namely ‘the presence in our language of surface analogies between different areas of that language’ (Bouwsma 1972, 83); and (2) the disposition to be misled on the part of the thinker. Since we cannot change our concepts at will, we would have to change the disposition in order not to fall prey to the mistakes or linguistic confusions. Bouwsma mentions two possible approaches (the only ones according to him), one represented by Ryle, the other by Wittgenstein. One is ‘to study the language’ (Bouwsma 1972, 84) and the other is ‘to alert the thinker’ (ibid.). Whereas Bouwsma sees Ryle as correcting mistakes by studying the language in a neutral, scientific way, he claims that Wittgenstein wanted to alert the individual thinker:

(…) Wittgenstein was not thinking of what he was doing as correcting mistakes. It was not mistakes, but an urge, a bewitchment, a fascination, a deep disquietude, a captivity, a disorientation, illusions, confusions – these, the troubles of the mixed-up intelligence, that Wittgenstein sought to relieve. (…) Wittgenstein’s interest was not in any particular problem but in the bothered individual, particularly in the hot and bothered. (Bouwsma 1972, 84)

Bouwsma even speaks of Wittgenstein as ‘getting under one’s skin’: It is as though Wittgenstein purposely hid the meaning of what he had to say and no one is to have it for the asking. He will have to dig for it. Nearly every paragraph is offered as an exercise, a challenge. (…) He [Wittgenstein] isn’t satisfied with telling the reader something. He nags. He intends to get under your skin, to get into your hair, to make you uncomfortable, to drive you to self-examination and improvement. (Bouwsma 1972, 85)

Bouwsma regards Ryle as taking up a problem which can be dealt with independently of anyone, a problem similar to a scientific problem. ‘If you want to know to what category a term belongs, go to Ryle. Ask him: To what category does the term “uneasy” belong? He will tell you: To the category “agitation”.’ (Bouwsma 1972, 86) Wittgenstein on the other hand, since his aim was to teach a skill, would never have given a list similar to the one Ryle gives in characterising words such as ‘uneasy’, ‘anxious’, ‘startled’, ‘shocked’ and ‘irritated’ as signifying agitations. Wittgenstein’s problem is always an individual in trouble. ‘There are no standard explanations which bring relief.’ (Bouwsma 1972, 87) Wittgenstein’s goal is not so much to hit upon the analogy that really explains the compulsion of saying things, but it is ‘to hit upon the analogy that one may assent to as what has compelled one to say this’ (Bouwsma 1972, 87). In 1972 Ryle responded to Bouwsma’s paper and his criticisms were primarily aimed at Bouwsma’s interpretation of Wittgenstein and not so much at his characterization of Ryle (Ryle 1972d). Ryle’s first objection is that he unjustly assimilates the Wittgenstein of the Philosophical Investigations to the psychoanalyst. Bouwsma claims that both are ‘treating the individual’ (Bouwsma 1972, 86) and that the object of both is to ‘hit upon the analogy that one may assent to as what has compelled one to say [or ‘do’ in the case of

172

Ryle and Wittgenstein psychoanalysis] this’ (Ibid. 86-87). Ryle thinks that, apart from 254-255 and (perhaps) 133, no indications of Wittgenstein practising a psychoanalyst model are present in his Philosophical Investigations. He would rather want to stress skills instead of the specific solutions; Ryle considers the psychoanalyst to focus primarily on the latter. Second, Bouwsma overstresses the coaching element in Wittgenstein’s ideas. Ryle admits that there is indeed a coaching element present, but argues that Wittgenstein does not provide us with a particular appropriate way of getting rid of our confusions, or even with a ‘particular pocket of confusion’ (ibid.). Bouwsma neglects the ‘exploring element’, i.e. that ‘behind the mentor there was the philosopher’ (Ryle 1972d, 109) – who was especially present in the Notebooks, where Wittgenstein explored ‘his own flybottles from inside’ (ibid.).

We are told with pathos that Wittgenstein ‘sought to bring relief, control, calm, quiet, peace, release, a certain power.’ Well! – what of the Wittgenstein who got us interested, fascinated, excited, angry, shocked? He electrified us. Whom did he ever tranquilize? (Ryle 1972d, 109)

Ryle’s last objection is aimed at Bouwsma’s claim that the Philosophical Investigations provides ‘no theory of mind, contains no arguments or refutations and rectifies nothing’ (Ryle 1972d, 109). Ryle strongly disagrees with this and gives examples of what he considers to be arguments, refutations, rectifications and claims about mind in the Philosophical Investigations.

No arguments? Yet, ‘I remember having meant (gemeint) him. Am I remembering a process or state? – When did it begin, what was its course, etc.?’ (661) No arguments? But lots of Wittgenstein’s wearisome interrogatives are, at least this (…) one, the rhetorically barbed conclusions of reductio ad absurdum arguments. (Ryle 1972d, 110)

Bouwsma’s interpretation does not do justice to the fact that the Investigations was a fierceful attack rather than merely a story not directed at anyone.

The clang of Wittgenstein’s metal against the metals of Frege, Russell, Ramsey, Brouwer, Moore, and the author of the Tractatus is here muted to a soothing bedside murmur. (Ibid., 110)

Ryle’s criticisms of Bouwsma’s interpretation are particularly relevant because they reveal with respect to some main issues what Ryle’s place would have been in the current debate concerning the interpretation of Wittgenstein. This will be clarified below, where Ryle’s Wittgenstein interpretation will be briefly discussed in relation to that of the so-called ‘New Wittgensteinians’.

An Evaluation of Their Differences and Similarities Underlying the different assessments of Bouwsma and Hacker regarding the differences between Ryle and Wittgenstein may be the fact that whereas Ryle is often tempted to making positive claims, Wittgenstein hardly ever seems to do so – and even when he does, his positive claims are presented as purely descriptive rather than normative, that is, based on a philosophical theory. And as positive claims are in general easier to attack than negative ones, it seems that Wittgenstein is less susceptible to attacks. On the other hand, Wittgenstein does not draw any direct conclusions and this sometimes makes it difficult to see his point. And, as we have already seen, Ryle wondered whether Wittgenstein’s ‘hinted diagnosis’ (Ryle 1957b, 267) would do.

173

The Philosophical Development of Gilbert Ryle

Hacker may be right in claiming that Ryle’s writings are often more superficial than those of Wittgenstein and do not in general touch upon the so-called limits of language. This can be illustrated by the fact that Wittgenstein’s discussion of ‘certainty’, which has to do with Wittgenstein’s interest in the limits of our language games, has no counterpart in Ryle. In Wittgenstein evidently touches upon the limits of language and the limits of our language-games. His basic claim is that within each language-game there are always propositions that are certain. Otherwise language-games could not exist and communication would be impossible. As regards psychological language games, Wittgenstein relates their limits to the notion of ‘instinctive’ or ‘prelinguistic’ behaviour, a notion absent from Ryle’s work. Ryle does not discuss this matter. Furthermore, Wittgenstein is also deeper with respect to the nature of philosophy. The difference between grammatical and empirical propositions and the confusion this distinction causes in philosophy is, for example, not discussed by Ryle. Not only On Certainty but also Wittgenstein’s treatment of private language and psychology deals with subjects which are neglected by Ryle. And Wittgenstein pays more attention to the constitutive role of language for the meaning of concepts. Finally, Wittgenstein more persistently and thoroughly pays attention to the analogies which mislead us. I find Bouwsma’s claim that Ryle provides us with ‘Knowledge that’ whereas Wittgenstein merely wants to teach us a method, less convincing. We have already seen that Ryle himself does not agree with Bouwsma’s interpretation of Wittgenstein as a philosopher who does not use arguments and does not want to rectify, etc. Moreover, in his introduction to The Concept of Mind Ryle mentions two different kinds of arguments he employs (rigorous logical arguments and less rigorous persuasive ones). The use of persuasive arguments seems to be a counterexample to Bouwsma’s claim that Wittgenstein focuses on the individual whereas Ryle does not. Bouwsma refers to, at most, only one of Ryle’s methods. Ryle’s persuasive arguments remind us of Wittgenstein’s idea that:

Das, was den Gegenstand schwer verständlich macht ist – wenn er bedeutend, wichtig, ist – nicht, dass irgendeine besondere Instruktion über abstruse Dinge zu seinem Verständnis erförderlich wäre, sondern der Gegensatz zwischen dem Verstehen des Gegenstandes und dem, was die meisten Menschen sehen wollen. Dadurch kann gerade das Nahe liegendste am allerschwersten verständlich werden. Nicht eine Schwierigkeit des Verstandes, sondern des Willens ist zu überwinden. (Wittgenstein, The Big Typescript, MS 213, 406)167

While I agree with Hacker’s suggestion about the relation between Ryle’s The Concept of Mind and Wittgenstein’s writings, I consider it unfortunate that he does not take into account Ryle’s later writings on thinking in his comparison. Bouwsma does not do so either. As was mentioned in the preceding chapter on Ryle’s later writings, his treatment of the concept of thinking and other related mental concepts basically amounts to showing in many different ways that concepts are of many different orders and that thinking is a higher order concept. Thinking itself is not a one-layered but a many-layered concept of a higher order. As we have seen, in these later writings Ryle focuses on the concept of thinking in relation to other concepts, concepts of other types, and he also discusses the plurality of different uses of the concept of thinking. More than in his earlier writings he

167 When an object is significant and important what makes it hard to understand is not the lack of some special instruction in abstruse matters necessary for its understanding, but the conflict between the right understanding of the object and what most people want to see. This can make the most obvious things the very hardest to understand. What has to be overcome is not a difficulty of the understanding but of the will. (Kenny 1994, 263) 174

Ryle and Wittgenstein pays attention to the complexity of the concept of thinking, which reminds us of Wittgenstein. In his treatment of thinking Ryle tries to show by using several expressions (e.g. ‘abstract’, ‘adverbial’, ‘host action’, ‘thick description’ and ‘higher order purpose’) that concepts of thinking are of a different – higher – order than others. Wittgenstein is trying to show something similar, but not by using or introducing new concepts in order to describe theoretically what is involved in thinking, but rather by giving examples, stimulating the reader to think for himself. For example, in Remarks on the Philosophy of Psychology (vol. 1) (Wittgenstein 1980) Wittgenstein talks about the conceptual relation between thinking and learning how to think and about perfectionizing methods, subjects which are also to be found in Ryle. While Wittgenstein focuses on the plurality of concepts and of different uses of these concepts, Ryle seems to be more interested in showing us different categories, within which concepts can be treated more or less the same. Their views on language, ordinary language and the use of language also show remarkable similarities. In particular Ryle’s paper ‘Ordinary Language’ (Ryle 1953b) shows similarities to Wittgenstein’s view. In this paper Ryle, as was discussed in the preceding chapter, clarifies what is meant by ‘use’ and ‘ordinary’, distinguishing between ‘the ordinary use of language’ and ‘the use of ordinary language’. Its purpose is to disarm Ryle’s (and also Wittgenstein’s) critics, who would for example argue that looking at what ordinary people say would turn philosophy into a mere sociological or philological enterprise. Concerning their methods, we could say that Ryle focuses primarily on the relation between different concepts, and on the fact that not all concepts mean in the same way, or rather: not all concepts are of the same level; some concepts are of a higher order than others. Wittgenstein, on the other hand, concentrates more on showing the diversity in our use of particular concepts.168 However, neither of them limits himself to discussing exclusively ‘their own’ type. Moreover, the later Ryle – but the earlier not so much – paid considerable attention to the different uses of particular concepts, e.g. the concept of thinking.

Concluding Remarks

In the first part of this chapter we have seen that Ryle studied Wittgenstein’s writings extensively, and that he gave him credits for at least part of his own philosophical development. From Ryle’s papers and annotations we have been able to distil his picture of Wittgenstein, which has enabled us to say something about what would have been his position in the current Wittgenstein debate. Ryle would have chosen the side of those who consider Wittgenstein to be one philosopher rather than two – an early and a later one – let alone three. Ryle did not deny that Wittgenstein’s position underwent changes, but he did not consider them so radical as to defend the position that the Wittgenstein of the Tractatus had an ontological aim. Particularly interesting was Ryle’s early study of ‘Notes on Logic’ (Wittgenstein 1913). From the ‘Ryle Collection’, and especially from Ryle’s annotated copies of Wittgenstein’s Tractatus and Philosophical Investigations, new facts about Ryle’s study of Wittgenstein have been revealed.

168 Using the terminology of Gellner – who wrote a polemical critique of what he called linguistic philosophy – the early Ryle’s primary goal was to show ‘the external polymorphousness of concepts’ and that of Wittgenstein was rather to point at ‘the internal complexity of concepts’ (Gellner 1959, 68-69).

175

The Philosophical Development of Gilbert Ryle

In comparing Ryle and Wittgenstein my aim has been a modest one: to point to some important differences and similarities in philosophical method and choice of subjects between the two philosophers. A comparison between them evidently depends largely on how one interprets their writings. Those who think Ryle to be a behaviourist are likely to miss some of the striking resemblances between the two philosophers. We have seen that Ryle did not like Wittgenstein’s talk of language-games. He also denied the value of the notion of family resemblances, and did not touch upon the limits of our use of language in general. Furthermore, Wittgenstein can be considered ‘deeper’ than Ryle with respect to the nature of philosophy. For example, the difference between grammatical and empirical propositions and the confusion this distinction causes in philosophy has no counterpart in Ryle. Wittgenstein also pays more attention to the constitutive role of language for the meaning of concepts. Despite these differences, I believe that their philosophical views and methods reveal close and important similarities. Ryle as well as Wittgenstein has a great interest in the notion of thinking and learning how to think, or how to think creatively. Both use arguments in order to refute and rectify, although the form Wittgenstein uses may tempt us into thinking that he does not. Both provide us with a method and express their views on Cartesian dualism, behaviourism, denotationalism, etc. – sometimes in an obvious way, sometimes in a more hidden way. Underlying all of this are their views on meaning and use of language, which are in essence similar.

176

Concluding Remarks

Chapter 7

Summary and Concluding Remarks

In this dissertation I have discussed the philosophical development of Gilbert Ryle against the background of some of his contemporaries, in particular Russell, Collingwood and Wittgenstein. To this end I have used the ‘Ryle Collection’, kept at Linacre College Library, a hitherto virtually unexplored source. One of the main aims of this dissertation has been to criticize some persistent interpretations of Ryle as a behaviourist. This not only adds to a better understanding of the historical Ryle, but also of the positions ascribed to him in current debates in the philosophy of mind. I started with a biographical sketch of Ryle (Chapter 1), for which I used not only published sources, which are fairly limited, but also stories and anecdotes which relatives, former students, colleagues and friends of Ryle provided. These interviews enabled me, I hope, to present a somewhat fuller picture of him, both as a person and as a professional philosopher, than is usually presented in print. This biographical sketch clearly shows the importance of Ryle for Oxford philosophy and philosophy in general. By his manner of teaching and writing and by his discussions of philosophers who were not commonly studied in Oxford at the time he inspired generations of students, and his The Concept of Mind (1949) became a classic to be studied by virtually every philosophy student in Britain and abroad. Attention was also given to his other important contributions to Oxford philosophy: his role in the founding of the B.Phil, which attracted many overseas students and greatly stimulated the activity of philosophy tutors in Oxford; his donation to Linacre College Library of a large part of his personal library (see Appendix 2); his role in the founding of Analysis; his editorship of Mind; his role in moving the Brentano Society from Prague to Oxford in 1938-9; and his single-handed rescue of the Waismann legacy. These and other activities show that Ryle was a powerful figure in British philosophy, who was almost always asked for his opinion when academic staff had to be appointed. Ryle’s development from a homogeneous denotational theory of meaning towards a heterogeneous theory of use constitutes the subject of Chapters 2 and 3. We have seen that Ryle moved from his early Russellian search for an ideal language to a position more similar to that of the later Wittgenstein. Ryle soon abandoned the Russellian method of analysing and reformulating expressions and he began to realize that denotationalism was responsible for many misunderstandings, sham-questions and problematic theories in philosophy. Thus Ryle’s focus was no longer on ‘meaning’, which for him implied the view that there is a homogeneous relationship between language and reality, and that all expressions mean what they mean in the same way. He now focused on ‘use’, underscoring the many different ways in which expressions are used. Consequently, his method of philosophical analysis changed from a logical and decompositional (or reductive) type of analysis to what may be called a connective type, somewhat analogous to the task of a cartographer (one of Ryle’s own favourite analogies). As analysed in Chapter 3, the method employed in The Concept of Mind (1949a) was meant to show that certain kinds of operations with concepts of mind go against the rules which these concepts ought to follow

177

The Philosophical Development of Gilbert Ryle according to the ‘logical geography’ of these concepts. Ryle no longer uses a reductive type of analysis, and does not talk about a deeper level of the logical form of facts, nor does he aim at an ideal language. He is now interested in the possible and impossible connections between different (types of) concepts. As we have seen, Ryle is often characterized as a (reductive) behaviourist, which is largely responsible for a general and swift disregard of his views. Since the rise of cognitive science and the influential rejection of behaviourism by Chomsky (1959) and Fodor (1975), any association with (reductive) behaviourism has become damaging. The reason why Ryle was so frequently and mistakenly considered to be a behaviourist was that his rejection of the idea of a denotational theory of meaning was not sufficiently recognized. But Ryle himself was to some extent to blame for this misinterpretation, for it was not always clear which target he attacked. Moreover, the fact that The Concept of Mind deals with category- mistakes on many different levels makes it difficult to interpret the book. The most general type of category-mistake arises because mental concepts are treated as names, and mental language as a description of a mental world. Once we recognize these category-mistakes we are able to expose both dualism and behaviourism, although Ryle, as he himself later declared, was not fully aware of the consequences for a behaviourist position when writing The Concept of Mind. In this chapter I have given several reasons why Ryle should not be characterized as a reductive behaviourist. The most important ones are: (1) a reductive behaviourist view does not fit Ryle’s meta-philosophical purpose; (2) reductive behaviourism presupposes something like a homogeneous denotational theory of meaning, a view rejected by him already in the late 1930s; and (3) the most charitable interpretation of the main category- mistakes Ryle attacks in The Concept of Mind tells against it. I hope to have argued convincingly for a non-reductivist interpretation of Ryle, since in this way his method can still be of value for modern philosophy, e.g. in discussions about the foundations of neuroscience. In Chapter 4 Ryle’s later writings were analysed: his theoretical papers, Dilemmas and his later writings on thinking. The theoretical claims which Ryle had presented in his earlier papers could not fully carry the weight of the method of The Concept of Mind. Hence, in the late 1950s and early 60s Ryle wrote additions to the ‘old’ theoretical account of how to do philosophy. He continued on his path of rejecting the idea of a denotational theory of meaning in favour of a heterogeneity of use, and still focused on informal logic and the idea that philosophical problems are of a different nature than scientific ones and cannot be solved piecemeal. New elements were Ryle’s instructions for interpreting ‘the ordinary use of the expression x’, namely that ‘use’ should be interpreted in a logical rather than in a philological sense and that ‘the ordinary use of an expression’ is not the same thing as ‘the use of ordinary expressions’. Furthermore, Ryle analysed the differences between language and speech and between words and sentences. Speech involves language but cannot be reduced to it. If we confuse speech with language, and sentences with words, we end up, according to Ryle, with a philosophy that is mere philology. Also new – in terminology though not in spirit – was his description of philosophy as an interlevel- activity. Ryle argued that philosophical arguments, as opposed to proofs, concern multiple levels; they are, in his vocabulary, ‘operations upon operations’. This is exactly what constitutes the difference between philosophy and the sciences. Whereas the sciences explain in the sense that they are trying to discover new facts and directly operate on them, philosophy does not aim at discovering facts but is concerned with description rather than with explanation.

178

Concluding Remarks

Like The Concept of Mind, Ryle’s Dilemmas (1954a) was also meant as an example of how to do philosophy. Although Ryle’s specific solutions to the dilemmas which he analysed may be (and have been) criticized as superficial and incomplete, his meta-philosophical message has proven to be of great value: connective philosophical analysis diagnoses the mistakes we make when we let language set traps for us, and how we can recognize and discard them. The remainder of Chapter 4 considers in some detail Ryle’s later work on thinking. Ryle’s basic question was: what does it mean to say that our descriptions of thinking are higher-level descriptions of processes? He considers the figure of the Thinker (Le Penseur) and asks: if the thinking which Le Penseur is doing is (merely) the way in which processes are managed, then what kind of processes are we talking about here? Ryle uses ‘the thinking of Le Penseur’ precisely to indicate the sort of thinking of which it is most difficult to see that it is adverbial (that is, it qualifies something else) and is supervening upon actions instead of itself being qualified as an action. If the ‘thinking’ of Socrates, Mozart and the scientist does not describe a process but a manner of ‘X-ing’, what can this X-ing possibly be? Is it ‘pondering’, ‘reflecting’ or ‘meditating’? What ‘happens’ when we do that? To present an answer to this cluster of questions was Ryle’s main objective in these later papers. His answer was that although no one single, specific autonomous X-ing exists for the thinking of Le Penseur, purely as a matter of logic we do seem to need a non- adverbial – or at least partly non-adverbial – verb for the adverbs to attach to. In these papers Ryle sharpened and strengthened his position without radically changing his mind; the adverbial element, for instance, was already present in 1949, though less prominent. We have seen that Ryle, by using what I have called ‘higher-order’ concepts, tried to show that a description of the thinking of Le Penseur is not a description of complex processes. It was rather a complex (or higher-order) description of processes. This enabled him to reject both dualism and behaviourism, which both (mistakenly) take all descriptions as functioning like sentences that refer to processes or things. It was important for Ryle to become more explicit about his rejection of both Cartesian dualism and behaviourism in his later writings, since his own The Concept of Mind was interpreted as a behaviourist work. We have analysed his position in some detail and suggested that Ryle is neither a dualist nor a behaviourist – not even a conceptual dualist or a conceptual behaviourist. His position may be described as ‘conceptual pluralism’, according to which concepts should not be treated in one single way, or in two different ways, but in many different ways depending on the context in which a concept is used. Though Ryle’s later work on thinking has been criticized, it is argued that his solution is original and the distinctions at which he points give us a clearer view on the ordinary use of our concept(s) of thinking. In this respect he certainly has a role to play in philosophy of mind and cognitive science of the twenty-first century. In Chapter 5 I discussed the philosophical relation between Ryle and Collingwood and their brief correspondence from 1935. The Correspondence shows us their disagreements on many points but the underlying, more fundamental, issues were not brought to the surface. In this chapter I have tried to uncover these fundamental issues by showing that Ryle and Collingwood had a different approach to and understanding of logic and the history of philosophy. Their disagreements can be traced back to these more fundamentally different views about the nature of philosophy and logic, the nature of metaphysics and the definition of a universal proposition. Essential is the difference between Collingwood’s overlap of classes in philosophy and Ryle’s idea of one general logic being applicable in all disciplines. Further, Collingwood’s somewhat idiosyncratic terminology led Ryle to think that Collingwood’s project was of a strong ontological

179

The Philosophical Development of Gilbert Ryle nature. Armed with this fuller understanding of the correspondence and its philosophical background, I have criticized some interpretations of the philosophical relationship between Collingwood and Ryle. Ryle’s study of Wittgenstein has been discussed in Chapter 6. In the first part of this chapter it was shown that Ryle had studied Wittgenstein’s writings extensively, and that he explicitly recognized the influence of Wittgenstein on his own philosophical development. From Ryle’s papers and annotations we have been able to learn about his views on Wittgenstein. Ryle’s position comes close to those who now, in contemporary debates on the philosophy of Wittgenstein, argue that there is only one Wittgenstein, not two (let alone three). He did not deny that Wittgenstein’s position underwent changes, but he positioned the Tractatus closer to Wittgenstein’s later work than traditional interpreters have done. Of particular interest is Ryle’s early study of Wittgenstein’s ‘Notes on Logic’ from 1913. From the ‘Ryle Collection’, and especially from Ryle’s annotated copies of Wittgenstein’s Tractatus and Philosophical Investigations and his comments on ‘Notes on Logic’ I have been able to reveal new facts about Ryle’s study of Wittgenstein. In comparing Ryle and Wittgenstein in the second part of Chapter 6, I have studied some important differences and similarities between the two philosophers. Ryle did not like Wittgenstein’s talk of language-games. He also denied the value of the notion of family resemblances, and did not talk about the limits of our use of language. Some scholars have argued that Wittgenstein is ‘deeper’ than Ryle with respect to the nature of philosophy. For example, the difference between grammatical and empirical propositions and the confusion this distinction caused in philosophy has no counterpart in Ryle. Wittgenstein also pays more attention to the constitutive role of language for the meaning of concepts. But despite these differences, I have also argued that their philosophical views and methods reveal close and important similarities. Both philosophers had a great interest in the notion of thinking and learning how to think, and how to think creatively. They attacked Cartesian dualism, behaviourism, denotationalism, etc. – sometimes explicitly, sometimes less affront. They used arguments in order to refute and rectify, although in the case of Wittgenstein this is less obvious. Underlying all of this are their largely similar views on meaning and our use of language. It has not been easy to give a final verdict on Ryle’s philosophical work that has been discussed in this dissertation. Scholars have frequently praised Ryle’s style, especially that of The Concept of Mind, while at the same time it was often noticed that it was too florid and mannerist so that the true purport of his arguments was unclear. Nonetheless, behind the style we can discern a clear message about the nature of philosophy. For Ryle philosophy is a conceptual study. However, he did not, in one sense of the word, solve the philosophical problems. He showed the implications of the common assumption that the mind is a secret grotto inside our head, and that thinking is an equally secret process inside this mysterious entity. Ryle rejected this view both in The Concept of Mind and in his later papers, but did not provide a new theory to replace the old ones. One may say that he offered doubts about the old views but no securities about new ones. This was part and parcel of his philosophy, which was essentially about philosophy and how it should be done. It was so much part of his approach that Ryle began to repeat himself, inventing ever new examples and notions to illustrate it. This makes reading his later papers a somewhat tedious affair, and it is not surprising that these papers have attracted far less attention than The Concept of Mind. Notwithstanding Ryle’s repetitions and his often in the end not entirely satisfactory solutions to specific philosophical problems, his meta-philosophical message remains of great value and influenced many philosophers.

180

Concluding Remarks

The focus of this dissertation has not allowed me to pay much attention to whom Ryle influenced. In this conclusion let me therefore say a few things about it and about his importance for today. Although today explicit reference to his writings are scarce, it is not hazardous to say that many philosophers have been influenced by them, in particular by his views on the nature and aims of philosophy and its proper methods. We may rightly say that Ryle is one of the founding fathers of a connective type of philosophical analysis. Daniel Dennett is one good example of someone who is influenced by Ryle. He was a friend of Ryle and corresponded with him for many years. From him Dennett adopts the idea that at the ‘folk psychological level’ – what Ryle would call the ordinary use of psychological concepts – psychological concepts function roughly in a dispositional way. According to what Dennett calls the intentional stance, talk about propositional attitudes (such as belief and ) is an instrument for predicting other people’s behaviour. An example of his acknowledgement of Ryle’s influence is the following quotation from an interview in 2002:

Ryle saw (brilliantly, intuitively, but not systematically) that in order to escape the mysteries, we had to turn the mind inside out, in a certain way, and forsake the quest for the golden nuggets of content or phenomenality (or something like that) hidden in secret places amongst the machinery. How do the personal level events of folk psychology map onto the events of sub- personal level ? Ryle didn’t try to address this question directly, because he knew he was master of only the personal level, but that still gave him plenty to say about what we shouldn’t look for. So my introduction to the stances, for instance, can be seen as part of my effort to illuminate the directness Ryle identified but could give no positive account of. (Dennett 2002, 10)

Dennett attacks, as does Ryle, a Cartesian dualist view and more in general rejects the idea of propositional attitudes as causally active inner states of people. However, they part company when it comes to what goes on ‘inside’. Unlike Dennett, Ryle almost entirely refrains from making claims about this subject, whereas Dennett attempts to exorcise the ghost in the machine in particular by claiming that mental (i.e. computational) representation ultimately cannot require an (intelligent) homunculus (to which Ryle would agree), but rather a series of stupid homunculi (to which Ryle would not agree).169 Without neglecting their differences, it is fair to say that Ryle greatly influenced Dennett and through him also Dennett’s readers and students. Furthermore, Ryle considerably influenced the debates about dispositions which are currently dominated by, D. M. Armstrong, C. B. Martin and U. T. Place, all of whom studied in Oxford during Ryle’s professorship.170 The main problem with dispositions seems to be to explain ‘how it can be true that something has a dispositional property when the disposition ‘points beyond’ itself, and never manifests itself – without committing oneself to the idea that dispositions are not actual’ (Crane 1996, 4). Place and Martin stay the closest to Ryle. Place gives a purely hypothetical explanation of dispositions, whereas Armstrong and Martin think that something categorical has to be brought in as well. According to Armstrong dispositional properties supervene on purely categorical ones – he generally identifies dispositions with categorical properties; both Martin and Place argue that dispositions cannot be reduced to categorical properties. Place adheres to the idea that there are two kinds of properties: categorical and dispositional (hypothetical) ones, whereas

169 See also Chapter 4, p. 120 for Ryle and Dennett about ‘internal life’ and ‘internal representations’. 170 Armstrong was in Oxford from 1952-1954 to gain his B.Phil.; Martin spent the year 1953-1954 in Oxford, just before he was appointed Lecturer at the University of Adelaide; Place graduated in philosophy (and psychology) at Oxford in 1949.

181

The Philosophical Development of Gilbert Ryle

Martin thinks that a property is never either entirely categorical or entirely hypothetical but always both. Dispositional explanations (and more specifically Ryle’s) have also played a role in discussions concerning the possibility of extending deductive models of explanation in the natural sciences to the human sciences. In the 1950s P. L. Gardiner, W. Dray, C. G. Hempel and P. Winch all paid attention to the dispositional model as an alternative to the covering-law model.171 Finally, besides Ryle’s distinction between categorical and dispositional, many of his other distinctions still play a role today, for example the one between ‘knowing how’ and ‘knowing that’. Many philosophers, with due recognition to Ryle’s ground-breaking discussion, adhere to such a distinction, though recent discussions show a reductionist tendency: ‘knowing how’ presupposes or can even be reduced to ‘knowing that’. Stanley and Williamson, for example, argue in ‘Knowing How’ (2001) that there is no fundamental distinction between the two, ‘knowing how’ being simply a species of ‘knowing that’. Other similar examples of Ryle’s influence could be given. When we turn to Ryle’s work on thinking and on conceptual method we find that his views are often ignored in present-day discussions. That he was believed to be a behaviourist did his reputation no good, but it is worth noticing, as I discovered at conferences and in conversations, that this interpretation is especially held by American and continental philosophers, whereas British philosophers – and especially those from Oxford – seem to know that this is a misleading label. However, they do not argue the case, since they think it is so self-evident, which, alas, it is not. This is regrettable because if a non- behaviourist (or at least non-reductive) interpretation of Ryle would be more widely defended it would show that his work is still valuable in current philosophical debates, for instance on the nature of cognitive science and psychology. For example in a recent book by Bennett and Hacker, The Philosophical Foundations of Neuroscience (2003), the authors use a method which is remarkably similar to that of Ryle and Wittgenstein. (Hacker is of course a widely recognized Wittgenstein specialist.) Their aim is to expose conceptual confusions in neuroscience and their starting- point is a non-reductive view on philosophy. One of the main claims is that empirical research is fundamentally unable to solve philosophical problems. Another is that conceptual confusions about how the brain relates to the mind the intelligibility of research carried out by , in terms of the questions they choose to address, the description and interpretation of results and the conclusions they draw. Hacker and Bennett adopt methods which are close to the ones practiced by Ryle and Wittgenstein and apply them to the practice of neuroscience. For example, they expose the conceptual error ‘that perception is a matter of apprehending an image in the mind’ and show the mistakes in neuroscientific research as a result of that error. And there are many other examples where their discussions of conceptual and linguistic mistakes show close affinity with Ryle and Wittgenstein. This then is another example of the importance I see Ryle still has in contemporary philosophy. And I am not the only one. As Jeff Coulter for example writes: the ‘lasting value of Ryle’s endeavours remains clear: his work stands as an enduring refutation of all forms of neurophysiological reductionist reasoning about what it is to think or to have a thought.’ (Coulter 2003, 77) Without doubt, others will remain sceptical about the significance of Ryle’s later work – like Matthews (1981), who claimed that Ryle has become someone of the past, whose work is outdated. Ryle’s specific solutions may indeed be outdated in the sense that they do not present conclusive solutions, and that in the case of thinking he did not provide a

171 Cf. Grazia Ramoino-Melilli 2003 on this and on other aspects of the value of Ryle’s account of dispositions. 182

Concluding Remarks completely satisfactory account of what the thinking of Le Penseur is. But Ryle has provided us with tools for approaching philosophical problems, for disentangling conceptual confusions, and for paying careful attention to the richness of our language in describing the many different layers of thinking. However, in order to recognize the merits and importance of these achievements, we first need to free Ryle from the mistaken label of behaviourism, to which I hope this dissertation has contributed.

183

The Philosophical Development of Gilbert Ryle

184

Appendix 1: Ryle’s Writings – the Complete List

Appendix 1: Ryle’s Writings – the Complete List

Some fairly complete lists of Ryle’s published writings can be found in Oscar P. Wood and George Pitcher’s Ryle. A Collection of Critical Essays (London 1971) and in Lucie Antoniol’s Lire Ryle Aujourd’hui (Brussels 1993). At the time of publication, the first list was probably complete, or at least nearly complete, since Ryle himself assisted in the making of the bibliography. It stops in 1968. Antoniol’s list includes post-1968 writings and posthumously published essays, but needs to be updated. What follows is a chronological list of Ryle’s published writings 1927-2007.172 The following fields of philosophy are represented in his works: philosophical methodology, , philosophy of mind, and the history of philosophy.

Christ Church 1927 - Critical Notice of Roman Ingarden, Essentiale Fragen, Mind, Vol. XXXVI

1929 - Critical Notice of Martin Heidegger, Sein und Zeit, Mind, Vol. XXXVIII173** - ‘Negation’, Proceedings of the Aristotelian Society, Supp. Vol. IX (symposium with J. D. Mabbott and H. H. Price)**

1930 - ‘Are there propositions?’, Proceedings of the Aristotelian Society, Vol. XXX, 1929-30** - Editor’s Foreword to the Proceedings of the Seventh International Congress of Philosophy, held at Oxford, September 1-6, 1930,

1931 - ‘Mr. Ryle on Propositions, Rejoinder’, Mind, Vol. XL (a reply to R. Robinson in the same volume)

1932 - ‘Systematically Misleading Expressions’, Proceedings of the Aristotelian Society, Vol. XXXII, 1931-32. Reprinted in Antony Flew (ed.), Logic and Language, First Series (Blackwell 1951)** - ‘ P h e n o m e n o l o g y ’ , Proceedings of the Aristotelian Society, Supp., Vol. XI (symposium with H. A. Hodges and H. B. Acton)** - Review of A. Wolf, Textbook of Logic, Philosophy, Vol. VII - Review of Ralph Eaton, General Logic, Philosophy, Vol. VII

1933 - ‘Imaginary Objects’, Proceedings of the Aristotelian Society, Supp. Vol. XII - ‘About’, Analysis, Vol. 1, 1933-34** - ‘Locke on the Human Understanding’, Tercentenary Addresses, Christ Church Oxford, Oxford University Press**

172 * indicates that the book or paper mentioned did not appear on the lists of Wood and Pitcher and Antoniol. ** indicates that the paper mentioned was reprinted in Gilbert Ryle, Collected Papers, Vol. 1 or 2. 173 In Gilbert Ryle, Collected Papers, Vol. 2, this paper on Heidegger is said to be published in 1928, but this is a mistake.

185

The Philosophical Development of Gilbert Ryle

- Review of J. N. Findlay, Meinong’s Theory of Objects, Oxford Magazine, Vol. LII, 1933-34 - Review of B. M. Laing, , Philosophy, Vol. VIII

1935 - ‘Internal Relations’, Proceedings of the Aristotelian Society, Supp. Vol. XIV (Symposium with A. J. Ayer)** - ‘Mr. Collingwood and the Ontological Argument’, Mind, Vol. XLIV174**

1936 - ‘Unverifiability by Me’, Analysis, Vol. IV, 1936-37** - Review of Bent Schultzer, Transcendence and the Logical Difficulties of Transcendence, Philosophy, Vol. XI

1937 - ‘Induction and Hypothesis’, Proceedings of the Aristotelian Society, Supp. Vol. XVI (Symposium with Margaret MacDonald)** - ‘Taking Sides in Philosophy’, Philosophy, Vol. XII** - ‘Back to the Ontological Argument’, Mind, Vol. XLVI (Reply to E. E. Harris, ‘Mr. Ryle and the Ontological Argument’, Mind, Vol. XLV)**

1938 - ‘Categories’, Proceedings of the Aristotelian Society, Vol. XXXVIII, 1937-38. Reprinted in Antony Flew (ed.), Logic and Language, Second Series (Blackwell 1959)** - Welcoming speech to the 4th International Congress for Unified Science, Erkenntnis, Vol. VII

1939 - ‘Plato’s ’, Mind, Vol. XLVIII. Reprinted in R. E. Allen (ed.), Studies in Plato’s Metaphysics (Routledge & Kegan Paul 1965)** - Review of Karl Britton, Communication, Philosophy, Vol. XIV - Review of F. M. Cornford, Plato and Parmenides, Mind, Vol. XLVIII**

1940 - ‘Conscience and Moral Convictions’, Analysis, Vol. VII. Reprinted in Margaret MacDonald (ed.), Philosophy and Analysis (Blackwell 1954)** - Review of W. M. Urban, Language and Reality; Philosophy, Vol. XV - Review of Brand Blanshard, The Nature of Thought; Philosophy, Vol. XV

Magdalen College 1945 - Philosophical Arguments (Oxford University Press) (Inaugural Lecture as Waynflete Professor of Metaphysical Philosophy). Reprinted in A. J. Ayer (ed.), Logical Positivism (The Free Press, Glencoe, Illinois)** - ‘Knowing How and Knowing That’, Proceedings of the Aristotelian Society, Vol. XLVI, 1945-46**

1946 - ‘Why are the Calculuses of Logic and Arithmetic Applicable to Reality?’, Proceedings of the Aristotelian Society, Supp. Vol. XX (Symposium with C. Lewy and K. Popper)** - Review of Marvin Farber, The Foundations of Phenomenology; Philosophy, Vol. XXI**

174 See Chapter 5 on the Ryle-Collingwood Correspondence, which was provoked by this paper. 186

Appendix 1: Ryle’s Writings – the Complete List

1947 - Review of Karl Popper, The Open Society and Its Enemies; Mind, Vol. LVI

1949 - The Concept of Mind (Hutchinson’s University Library) - Discussion of Rudolf Carnap, Meaning and Necessity; Philosophy, Vol. XXIV**

1950 - ‘’If’, ‘So’ and ‘Because’’, Max Black (ed.), Philosophical Analysis (Cornell University Press)** - ‘The Physical Basis of Mind’, Peter Laslett (ed.), The Physical Basis of Mind (Blackwell) (A symposium with Lord Samuel and A. J. Ayer). Reprinted in Antony Flew (ed.), Body, Mind and Death (Macmillan 1964) - ‘Logic and Professor Anderson’, The Australasian Journal of Philosophy, Vol. XXVIII** - Review of M. H. Carré, Phases of Thought in England; Philosophy, Vol. XXV

1951 - ‘Heterologicality’, Analysis, Vol. XI, 1950-51. Reprinted in Margaret MacDonald (ed.), Philosophy and Analysis (Blackwell 1954)** - ‘Thinking and Language’, Proceedings of the Aristotelian Society, Supp. Vol. XXV (Symposium with Iris Murdoch and A. C. Lloyd)** - ‘ F e e l i n g s ’ , The Philosophical Quarterly, Vol. 1. Reprinted in William Elton (ed.), Aesthetics and Language (Blackwell 1954)** - ‘The Verification Principle’, Revue Internationale de Philosophie, Vol. V** - ‘Ludwig Wittgenstein’, Analysis, Vol. XII, 1951-52. Reprinted in Irving Copi and Robert Beard (eds.), Essays on Wittgenstein’s Tractatus (Routledge & Kegan Paul 1966)**

1952 - Review of J. Wisdom, Other Minds; The Listener, Vol. LXIV - ‘Graduate Work in Philosophy at Oxford’, Universities Quarterly, Vol. VI

1953 - ‘Thinking’, Acta Psychologica, Vol. IX175** - ‘Ordinary Language’, Philosophical Review, Vol. LXII. Reprinted in Charles Caton (ed.), Philosophy and Ordinary Language (University of Illinois Press 1963)**

1954 - Dilemmas, The Tarner Lectures (Cambridge University Press) - ‘Proofs in Philosophy’, Revue Internationale de Philosophie, Vol. VIII (and contributions to discussion)** - ‘Pleasure’, Proceedings of the Aristotelian Society, Supp. Vol. XXVIII (Symposium with W. B. Gallie). Reprinted in Donald Gustafson (ed.), Essays in Philosophical Psychology (Anchor Books 1964)** - Review of L’enseignement de la philosophie (Publications Unesco, Paris 1953); Universities Quarterly, Vol. VIII

1956 - ‘Sensation’, H. D. Lewis (ed.), Contemporary British Philosophy, Third Series (Allen & Unwin). Reprinted in R. J. Swartz (ed.), Perceiving, Sensing and Knowing (Anchor Books 1965)**

175 In the same volume Prof. G. Humphrey responded to Ryle’s paper.

187

The Philosophical Development of Gilbert Ryle

- ‘Hume’, M. Merleau-Ponty (ed.), Les Philosophes Célèbres, Editions de l’art (Lucien Mazenod, Paris)** - Introduction to The Revolution in Philosophy (Macmillan) (the book consists of previously broadcasted talks by A. J. Ayer and others)

1957 - ‘The Theory of Meaning’, C. A. Mace (ed.), British Philosophy in Mid Century (Allen & Unwin). Reprinted in Charles Caton (ed.), Philosophy and Ordinary Language (University of Illinois Press 1963)** - ‘Predicting and Inferring’, Observation and Interpretation in the Philosophy of Physics, Proceedings of the Colston Research Society Symposium, Vol. IX (University of Bristol)** - Final discussion with Mary Warnock and A. M. Quinton, in D. F. Pears (ed.), The Nature of Metaphysics (Macmillan) (A series of programmes originally broadcasted in the Third Programme of the B.B.C. in 1955) - Review of Ludwig Wittgenstein, Remarks on the Foundations of Mathematics; Scientific American, Vol. CXVII**

1958 - ‘On Forgetting the Difference Between Right and Wrong’, in A. I. Melden (ed.), Essays in Moral Philosophy (University of Washington Press)** - ‘A Puzzling Element in the Notion of Thinking’, Proceedings of the British Academy, Vol. XLIV (Oxford University Press). Reprinted in Dudley Bailey (ed.), Essays in Rhetoric (Oxford University Press 1965), and in P. F. Strawson (ed.), Studies in the Philosophy of Thought and Action (Oxford University Press 1968)** - Reply to J. Garelli, ‘La notion de possibilité dans l’analyse logique de l’esprit de Gilbert Ryle’; Revue de Metaphysique et de Morale, Vol. LXIII - ‘T. H. Weldon’, The Oxford Magazine, Vol. LXXVI, 1957-58 (Obituary)

1960 - ‘Letters and Syllables in Plato’, Philosophical Review, Vol. LXIX** - ‘Epistemology’, in J. O. Urmson (ed.), The Concise Encyclopaedia of and Philosophers (Hutchinson) - Comment on Mr. Achinstein’s Paper, Analysis, Vol. XXI, 1960-61 - Review of Richard Wollheim, F. H. Bradley; The Spectator, Vol. CCIV

1961 - ‘Use, Usage and Meaning’, Proceedings of the Aristotelian Society, Supp. Vol. XXXV (Symposium with J. N. Findlay). Reprinted in G. H. R. Parkinson (ed.), The Theory of Meaning (Oxford University Press 1968)**

1962 - ‘A Rational Animal’, Auguste Comte Memorial Lecture (University of London, Athlone Press)** - ‘Abstractions’, Dialogue, Vol. 1** - ‘Thinking Thoughts and Having Concepts’, Logique et Analyse, No. 20** - ‘La Phénoménologie Contre The Concept of Mind’, La Philosophie Analytique; Cahiers de Royaumont Philosophie, No. IV (Les Editions de Minuit, Paris)**

1963 - Review of G. E. Moore, Commonplace Book 1919-53; New Statesman, Vol. LXV**

1965 - ‘The Timaeus Locrus’, , Vol. X**

188

Appendix 1: Ryle’s Writings – the Complete List

- ‘ in the Academy’, in R. Bambrough (ed.), New Essays on Plato and Aristotle (Routledge & Kegan Paul)**

1966 - Plato’s Progress (Cambridge University Press) - ‘Jane Austen and the Moralists’, The Oxford Review, No. 1. Reprinted in B. C. Southam (ed.), Critical Essays on Jane Austen (Routledge & Kegan Paul, 1968)** - Review of Stuart Hampshire, Freedom of the Individual; New Statesman, Vol. LXXI

1967 - ‘Plato’, in Paul Edwards (ed.), The Encyclopaedia of Philosophy, Vol. VI (Macmillan and Free Press) - ‘John Locke’, Crítica Revista Hispano Americana de Filosofía (Mexico), Vol. 1** - ‘Teaching and Training’, in R. S. Peters (ed.), The Concept of Education (Routledge & Kegan Paul)**

After Ryle’s Retirement 1968 - ‘Dialectic in the Academy’, Aristotelian Dialectic (Oxford University Press) (Proceedings of the third Symposium Aristotelicum)** - ‘Thinking and Reflecting’, in G. N. A. Vesey (ed.), The Human Agent, Royal Institute of Philosophy Lectures, Vol. 1, 1966-67** - ‘The Thinking of Thoughts’, University Lectures, No. 18 (University of Saskatchewan)** - ‘G. E. Moore’s “The Nature of Judgement”’, in Ambrose, A., and Lazerowitz, M. (eds.), G. E. Moore: Essays in Retrospect (London: Allen & Unwin 1968)

1970 - ‘Autobiographical’, in Oscar P. Wood and George Pitcher, Ryle. A Collection of Critical Essays (Macmillan)* - Review of ‘Symposium on J.L. Austin’, The Listener** - ‘Obituary Appreciation of Bertrand Russell’, The Listener - ‘Bertrand Russell (1872-1970)’, Proceedings of the Aristotelian Society LXXI, reprinted in Revue Internationale de Philodophie XXVI (1972) and Roberts, G. W. (ed.), Bertrand Russell Memorial Volume (London: Allen & Unwin 1979) - ‘Some problems about Thinking’, Keifer, H., and Munitz, M. (ed.), Mind, Science and History (New York: SUNY Press 1970)

1971 - ‘Interview with Brian Magee’, The Listener, reprinted in Magee, Brian (ed.), Modern British Philosophy (London: Secker & Warburg 1971) - Collected Papers, volume 1 (Hutchinson of London), with an introduction by Ryle - Collected Papers, volume 2 (Hutchinson of London), with an introduction by Ryle - ‘Phenomenology and Linguistic Analysis’, Neue Hefte für Philosophie, no. 1

1972 - ‘Thinking and Self-Teaching’, in Konstantin Kolenda, A symposium on Gilbert Ryle, Rice University Studies, Vol. 58 (William Marsh Rice University) - ‘Thinking and Saying’, in Konstantin Kolenda, A symposium on Gilbert Ryle, Rice University Studies, Vol. 58 (William Marsh Rice University)

189

The Philosophical Development of Gilbert Ryle

- ‘Postscript’, in Konstantin Kolenda, A symposium on Gilbert Ryle, Rice University Studies, Vol. 58 (William Marsh Rice University) (Ryle’s response to the contributions of Professor Kolenda and Professor Bouwsma to the symposium) - ‘Intentionality-Theory and the Nature of Thinking’, Haller, R. (ed.), Jenseits von Sein und Nichtsein (: Akademische Druck und Verlagsanstalt 1972), reprinted in Revue Internationale de Philosophie, Vol. 27 (1973). First presented at a Symposium on Meinong held in Graz. - ‘Negative Actions’, Hermathena, No. CXV

1974 - ‘Mowgli in Babel’, Philosophy, Vol. 49, No. 187

1976 - ‘Improvisation’, Mind, Vol. LXXXV, No. 337 - Preface, Contemporary Aspects of Philosophy (London: Oriel Press 1976)

Posthumous Publications 1976 - ‘Fifty Years of Philosophy and Philosophers’, Philosophy, Vol. 51 (Lecture given at King’s College London, on 14 May 1976, to celebrate the fiftieth anniversary of the foundation of the Royal Institute of Philosophy)

1977 - ‘Can Virtue be Taught’, Raimono-Mellili, Grazia (ed.), Studi filosofici e pedagogici, no. 1 (1977)

1979 - Konstantin Kolenda (ed.), Gilbert Ryle On Thinking (Basil Blackwell, Oxford) (with a preface by G. J. Warnock); This is a collection of Ryle’s most recent (previously published) essays on Thinking. - ‘Last Thoughts on The Concept of Mind’, Personal communication to Bestor, Thomas W. (June 28, 1976), The Personalist

1990 - ‘Logical Atomism in Plato’s Theaetetus’, Phronesis 35 (1990)*

1993 - René Meyer, Aspects of Mind. Gilbert Ryle (Blackwell); This collection contains until then unpublished papers by Gilbert Ryle, notes René Meyer made when he attended Ryle’s lectures and a seminar in 1964, and two tributes to Ryle.

1999 - The Gilbert Ryle Issue, The linacre Journal, Number 3; This issue contains, besides reflections on Ryle’s philosophy and the ‘Ryle Collection’ present at Linacre College Library by various authors, four papers written by Ryle. All four essays were published before. Three of them were published in René Meyer’s, Aspects of Mind, but the editors of The Gilbert Ryle Issue did not in all instances agree with Meyer’s renderings of Ryle’s handwriting.*

1999 - Ryle, Gilbert, 1976b, Letter to Daniel Dennett, 22 February 1976, published by Dennett in 1999 on EJAP: http://ejap.louisiana.edu/EJAP/2002/RyleLett.pdf*

2000 - ‘Courses of Action or the Uncatchableness of Mental Acts’, Philosophy, Vol. 75 (originally read at the annual conference of the Experimental Analysis of Behaviour Group at Bangor in April 1974)*

190

Appendix 1: Ryle’s Writings – the Complete List

2005 - ‘The Ryle-Collingwood Correspondence’, in Connelly, James and D’Oro, Giuseppina, An Essay on Philosophical Method, revised edition (Oxford: Clarendon Press 2005)*

2006 - McGuinness, Brian and Vrijen, Charlotte (eds.), ‘First Thoughts: An Unpublished Letter from Gilbert Ryle to H. J. Paton’, British Journal for the History of Philosophy 14(4) 2006*

191

The Philosophical Development of Gilbert Ryle

192

Appendix 2: Gilbert Ryle Collections in Oxford

Appendix 2: Gilbert Ryle Collections in Oxford

In this appendix I shall describe collections of Ryle material which I found or rediscovered in various places in Oxford. The ‘Gilbert Ryle Collection’ at Linacre College is by far the largest one. Some of the others are but minor collections, sometimes consisting of no more than a few sheets of paper, but as they are virtually unknown to the scholarly world, I list these too. This list, however, is not complete as more material is likely to turn up.

Minor Ryle Collections Outside Linacre College

The New Bodleian, Oxford One of the most valuable sources is the Ryle-Collingwood Correspondence, which is kept in the New Bodleian as part of the Collingwood legacy. The handwritten letters as well as typescripts of them – not in all respects accurate – are available upon request at the Modern Papers Room. The correspondence was still unpublished when I started studying it in 2002, but it has now been published (Connelly and D’Oro 2005). On this correspondence see Chapter 5 of my thesis. Besides the Ryle-Collingwood Correspondence, the New Bodleian keeps other documents of a biographical rather than philosophical nature.

1) MS Eng. C. 2714 is a paper by a Mr Morgan about the history of the Ryle family, focusing on their activities as bankers. The paper was presented at the 2nd European Congres of the International Banknote Society in London, on 7 May 1972, and the National Congres British Association of Numismatic Societies. Mr Morgan’s story about J.C. Ryle seems to have been unknown to Gilbert in 1970. Gilbert and a few of his brothers and nephews corresponded with Mr Morgan about the Ryle-family. Gilbert wrote a letter to Mr Morgan on 13 November 1970. 2) Top Oxon. C. 870, fol. 18-34, gives insight into the transfer of the Brentano Institute from Prague to Oxford right before the Second World War. See Chapter 1 for Ryle’s role in transferring the Brentano Institute from Prague to Oxford.

The Philosophy Library, Oxford Other philosophically interesting documents can be found in the Philosophy Library. Apart from papers that have been published176, which I shall not mention, the following documents can be found here:

1) a typescript of a conversation between J. O. Urmson, Brian Magee and Ryle. I have not been able to find out when this conversation took place. It is most likely a

176 E.g. René Meyer published Ryle’s ‘The Meno’ and a paper read to the Oxford Philosophical Society 500th meeting in 1968 (Meyer 1993), both of which were kept in the Philosophy Library in Oxford.

193

The Philosophical Development of Gilbert Ryle

typescript of a radio program or video recording, which could have been made by a philosophy student. Magee could not remember the conversation; 2) a paper by J. King Gordon, containing large quotations of letters from Ryle to him; 3) a paper on Ryle as a teacher, by Professor Julius Moravcsik.

The first document is particularly interesting because Ryle makes some explicit remarks about his debts to Wittgenstein. The other two papers are of a more biographical nature.

Nuffield College, Oxford As part of the Lord Cherwell Correspondence, Nuffield College keeps documents relating to Ryle’s Intelligence work in the Second World War. The index of correspondents of Lord Cherwell contains three references to Gilbert Ryle (D219, F415 and G442)177. From the beginning of the Nazi persecutions in 1933, Lord Cherwell – Frederick Alexander Lindemann (1886-1957) – took much trouble to help distinguished German Jewish scientists to find places in British Universities and to obtain grants. Holding the post of Paymaster-General, a ministerial post in Britain, from 1942-1945, he was Churchill's scientific adviser during the Second World War. From the three letters written by Ryle it seems that he did not correspond with Lord Cherwell about finding places for German Jewish scientists at British universities, but rather about aircraft recognition and engagement of low flying aircrafts, as taught at the A.A. (Anti-aircraft) (L.M.G. – Light machine gun) School, Northolt. Clearly as a response to a letter from Lord Cherwell, one of the letters Ryle starts as follows:

I’ve not lost interest in my former activity, & send you the enclosed to bring to your mind some of the problems, technical & especially administrative, which face those who want to make the lower air unsafe for enemy planes. (D219/2)

F415/5 is a pamphlet on the methods taught at the A.A. School. The principal subjects taught were engagement of low flying aircraft – mobile and static warfare – and elementary aircraft recognition. It is unclear whether Ryle wrote the pamphlet which he sent to Lord Cherwell himself.

Queen’s College, Oxford A letter from Ryle to his former tutor H. J. Paton, dated April 15 1926, is kept as part of the Paton archive at Queen’s College Library.178 The letter provides valuable insight in Ryle’s earliest philosophical thoughts.179

Collections outside Oxford I have found only a few items outside Oxford, which is not surprising considering the fact that Ryle spent his entire academic life in Oxford. The Popper Collection of Hoover Institution Archives, Stanford, and McMaster’s Russell Archive contain correspondences with Ryle, mostly concerning Mind. Furthermore, I know of at least one unpublished paper, which was sold to an anonymous buyer via Lameduck Books in 2001 by Professor Samuel

177 As well as one to his brother John A. Ryle (D220) and another to his brother Sir (D221). 178 I am grateful to Professor Brian McGuinness for bringing this to my attention. 179 For the text and a detailed discussion of this letter see McGuinness and Vrijen 2006. 194

Appendix 2: Gilbert Ryle Collections in Oxford

C. Wheeler III, Professor of Philosophy at the University of Connecticut.180 The paper is a five-page unpublished holograph essay, entitled ‘Consciousness’, prepared by Ryle as a contribution to the Collier’s Encyclopedia. Since Ryle was still working out his views on the matter at the time, the essay became much too long and technical for its intended purpose, and was in the end not included. It consists of almost 2000 words, with numerous corrections and emendations, written on five legal-size lined sheets. Unfortunately, it was not possible to obtain a copy.

Ryle’s books and papers at Linacre College

As was mentioned in Chapter 1, as a Fellow of Magdalen Ryle was a member of the committee which oversaw the establishment of Linacre as a college. He donated much of his library to this College when he retired in 1968, followed by another donation upon his death in 1976. These books and papers, together with the so-called ‘red box’ which contains letters, postcards, two notebooks and separate sheets of notes, form the ‘Gilbert Ryle Collection’. At first Ryle’s books did not receive separate treatment and could be borrowed like any other book. Today they are kept as a special collection in restricted and secured cabinets, only available upon request. One should be careful not to attribute much significance to ‘gaps’ in Ryle’s library. As Giles Barber, Linacre College’s first librarian, has noticed, it is not clear what books Ryle owned other than the ones he donated to Linacre, since the circumstances of his donations are largely undocumented.181 It is obvious that some books are missing, for example published Bphil-theses that he supervised and in which he was thanked. Furthermore, there is no sign of copies of the Blue and Brown Books, of ‘Notes on Logic’ or Russell’s Theory of Knowledge, which he must have possessed. And Ryle must have owned more novels than the collection contains. He was a great admirer of P. G. Wodehouse but the library does not contain any of his books. The books in this collection are very much the working library of a philosopher of his day with wide-ranging interests. There are about 1100 volumes, some of which valuable, the earliest in date being an Aristotle of 1588. They include philosophical studies such as G. E. Moore’s Principia Ethica, Ludwig Wittgenstein’s Tractatus logico- philosophicus (1922) and Philosophical Investigations (1953), Aristotle, Plato, Bacon, Bergson, Descartes, Berkeley, Locke, Hume, Bolzano, Brentano, Lotze, Collingwood, Bosanquet, Husserl, Frege, Meinong, Russell and many others. Eighteenth-century English literature is also represented. The books and papers tell their own story of Ryle’s personal and intellectual development throughout his life, containing, for example, books owned by his parents, books which he got for Christmas in the year he first went to college, books he bought when he was a don at Christ Church, and several presentation examples. He owned relatively many second-hand books. His copy of his great-great-grandfather John Charles Ryle’s, The Christian Leaders of England in the eighteenth century had previously belonged to an E. J. Furlong from Trinity College Dublin. There are also non-philosophical

180 Before he sold the manuscript he tried to find a Ryle archive in order to send a Xerox to it by posting a message on the internet (http://palimpsest.stanford.edu/byform/mailing-lists/exlibris/1995/01/msg00012.html). The fact that Wheeler did not know of the ‘Gilbert Ryle Collection’ at Linacre College and finally sold the manuscript to an anonymous buyer shows the need to report about the existence of this Ryle Collection. 181 Thus, Giles Barber in his paper ‘A Philosopher and his books’ in: The Linacre Journal, 1999, nr. 3, 17-26, which is the only paper about the ‘Ryle Collection’.

195

The Philosophical Development of Gilbert Ryle books, e.g. W. Robinson’s The English Flower Garden (London 1893), which reflects Ryle’s love of gardening, several issues of Astounding Science Fiction and Florence Nightingale’s Notes on Nursing: what it is and what it is not (1860). Some books are heavily annotated, e.g. Ryle’s three copies of Wittgenstein’s Tractatus (1922, 1951 and 1969), his copy of Husserl’s Philosophische Untersuchungen, and Russell’s Principles of Mathematics; others were not even cut open. In his autobiographical essay Ryle mentioned the influence which his father’s interest in philosophy had on him. Ryle was an avid reader and read many of the mainly philosophical and semi-philosophical books owned by his father. The ‘Ryle Collection’ contains several books that originally belonged to his father, e.g. late nineteenth-century editions of Aristotle, bearing his signature. Other examples of books that had clearly belonged to his father are: The republic of Plato (London 1879); Lettres de Pascal (Paris 1862); Karl Pearson’s The Grammar of Science (1911), which contains a handwritten remark from the author on the first page: ‘R. J. Ryle From the author with affectionate regards’; a translation of Mach’s Popular Scientific Lectures (Chicago 1895); John Burnet’s Early Greek Philosophy (London/Edinburgh 1892); a translation of Kant’s Prolegomena and Metaphysical Foundations of natural science (London 1891); some Spinoza, such as Sir Frederick Pollock, Spinoza his life and philosophy (London/New York 1899); W. Wallace’s, Life of (London 1890); of course Arnauld’s The Port-Royal logic; and T. H. Green’s Prolegomena to Ethics (Oxford 1890). R. J. Ryle did not only read philosophical works but also preceded his son by contributing to the Proceedings of the Aristotelian Society. He was in fact one of the first members of the Aristotelian Society. Some of the early volumes of the Proceedings contain papers or discussions by him. In volume 1 of the Proceedings (1888) R. J. Ryle is mentioned as a member and in the second volume from 1894 he has four contributions, one to a symposium concerning the question ‘Does Law in Nature exclude the possibility of miracle?’, one to a symposium concerning the question ‘Is religion pre-supposed by morality, or morality by religion?’, a paper ‘The nature of force and matter’, and another paper called ‘Epictetus’. The 1933 volume, which contains his contribution to the symposium on Imaginary objects, is the first one bearing Gilbert Ryle’s name. Examples of books that originally belonged to other members of the Ryle-family are: Burnet’s edition of the Greek text of Plato, 5 volumes (Oxford 1899), which had once belonged to E. Ryle (Ryle’s sister Effie), who also owned Liddell & Scott’s Greek-English Lexicon (Oxford 1883) which she had received from her mother’s Union members of S. Hilda’s, Newcastle; Philosophical Lectures and Remains of Richard Lewis Nettleship (London 1897) which was bought by a C.R. in 1898. And Notes on Nursing by Florence Nightingale was at one time owned by an Isabelle Ryle who bought or received it on 4 May 1880. Ancient philosophy is well represented and so are the philosophical classics, e.g. works of Hobbes, Descartes, Leibniz, Locke and Berkeley. Ryle’s library contains the works of Aristotle, translated into English under the editorship of W. D. Ross, all annotated slightly; 14 volumes of the translated works of Aristotle into English in the Loeb series, all slightly annotated as well. Further, Ryle owned Bekker’s edition of the Greek text of Aristotle’s Opera and Metaphysica from 1837; Aristotle’s Theory of Poetry and Fine Art, with a critical text and translation of ‘the ’ by S. H. Butcher (London 1911). Aristotle’s, Ethica Nicomachea, ed. I. Bywater (Oxford 1894 and 1897) contains many annotations. In his paper about the Linacre Collection Giles Barber writes that ‘it is surprising and unfortunate that the collection contains virtually nothing by Descartes, whether 196

Appendix 2: Gilbert Ryle Collections in Oxford annotated by Ryle or not. Equally it may seem surprising to find no works by that very Oxford figure, John Locke.’ (Barber 1999, 21) However, the following books by Locke can be found on Ryle’s bookshelves:

1) Locke, An Essay Concerning Human Understanding, ed. Peter. H. Nidditch, Clarendon Press 1975. This book contains few marginal annotations, e.g. ‘faculties’ in the margin, mostly in book III. 2) Locke, The Works of John Locke in four volumes, seventh edition (1768). This could well have been one of his father’s books. 3) Locke on Civil Government (London 1884) 4) Locke’s conduct of Human Understanding, intr., notes, etc. by Thomas Fowler, 2 edition (Oxford 1882) 5) An Early Draft of Locke’s Essay together with excerpts from his journals, ed. R.I. Aaron and Jocelyn Gibb (eds.) (Oxford 1936), containing few marginal annotations, e.g. word ‘induction’ in the margin, and references to other philosophers such as Aristotle.

Barber’s claim that Ryle owned virtually nothing by Descartes is also incorrect. The collection clearly shows his interest in the works of Descartes. His copy of the French translation of Descartes’ Regulae ad directionem ingenii (Paris 1933) contains a receipt, showing that Ryle bought this book for 5 GBP on Feb. 13th 1933 when he was a lecturer at Christ Church. He bought it at Blackwell’s, together with Weldauer’s, Kritik der Transzendentalphaenomenologie Husserls. He had these books on approval, but he at least did not return the Descartes, though he did not read it (the pages are uncut). Ryle also owned, Descartes, Discourse on Method. Meditations. (Edinburgh and London 1890), which had belonged to his father. Further, Ryle owned a copy of Descartes’ Lettres sur La Morale, which is a correspondence with princes Elisabeth (Paris 1935). The first part was probably read by Ryle; the second is uncut. Ryle also read Elizabeth S. Haldane and G.R.T. Ross, (transl.), The Philosophical works of Descartes, volume 1 and 2 (Cambridge 1911). He rather heavily annotated volume 1. There are hardly any annotations in ‘The ’. Some of his comments consist of several sentences, others only of a word, such as ‘induction’, or a name, such as ‘Aristotle’, or ‘Heraclitus’. Most are references to other philosophers. Volume 2 is missing from Linacre Library. There is also a copy of Boyce Gibson’s The Philosophy of Descartes (London 1932). Numerous works by and about Brentano, Husserl, Meinong, Russell and Frege illustrate Ryle’s early interest in philosophers other than the commonly respected ones in Oxford at the time. Some of them are heavily annotated, such as Husserl’s Logische Untersuchungen (1922), which was bought by Ryle in 1926. Husserl’s Philosophie der Arithmetic (Leipzig 1891) and Ideeen zu einer reinen Phaenomenologie, Erster Band (1922), which Ryle bought when he was a don at Christ Church, are also heavily annotated. In these books Ryle made his own index of names and subjects. With a few exceptions it seems that Ryle only made marginal notes. His annotations to Husserls writings are often short abstracts. Sometimes Ryle presents an example, such as, with respect to Husserl’s claim: “Wahr ist für jede Spezies urteilender Wesen, was nach ihrer Konstitution, nach ihren Denkgesetzen als wahr zu gelten haben” (LU1, 117). Ryle gives the following example: ‘What is true for men, false for cows’. Sometimes Ryle fills in names to make things more concrete and to connect Husserl’s words to the history of philosophical thought. On page 122 of Logische Untersuchungen he mentions Kant’s categories and

197

The Philosophical Development of Gilbert Ryle

Bolzano. Ryle’s copies of Wittgenstein’s Tractatus were also heavily studied and annotated. They are discussed in Chapter 6. The most interesting documents in the so-called ‘red box’, which is part of the ‘Ryle Collection’, are his notes on Wittgenstein. ‘Ontological and Logical Talk in Wittgenstein’s Tractatus’ is one of these documents – it was published by René Meyer in 1993. Others, still unpublished, are Ryle’s notes on the structure of the Tractatus (included in this dissertation as Appendix 3); notes on 5.5.41; a bibliography of relevant readings to the Tractatus; and Ryle’s comments on Wittgenstein’s ‘Notes on Logic’ from 1913. Also worth mentioning are several letters and postcards to Ryle, e.g. a letter from Mabbott and a postcard written by Husserl, and two notebooks containing Ryle’s lecture notes and first drafts of papers.

198

Appendix 3: Ryle’s Lecture Notes on the Structure of Wittgenstein’s Tractatus

Appendix 3: Ryle’s Lecture Notes on the Structure of Wittgenstein’s Tractatus

This unpublished sheet of paper can be found in The Red Box, part of the ‘Ryle Collection’, which is kept at Linacre College Library, Oxford 182

“Inserted in the margin, where the whole numbers first appear, I shall suggest some chapter headings in a moment. Next, I think we can treat the single decimal numbers, such as 2.1 or 5.4 as the beginnings of sections. So anything in, say, the 4.3s can be described as in Chapter 4, Section 3. I shall also suggest section titles for insertion in the margin. I don’t think we need bother about the lower decimal places. Certainly sometimes a very important thing is said in a paragraph or sentence with a very fiddling number of decimals in its number, e.g. in 4.0312.

Here are my Chapter Headings (to be taken with several pinches of salt) Chapter 1 (i.e. from the first sentence on p.26/27) ‘Facts and the World’. Chapter 2 (i.e. from the eighth sentence on the same page) ‘The Representation of Facts’. Chapter 3 (p.42/43) ‘The Propositional Representation of Facts’. Chapter 4 (p. 60/61) ‘Propositional Forms.’ Chapter 5 (p. 102/3) ‘The Logic of Propositional Forms’. Chapter 6 (p.152/3) ‘The Propositions of Logic Itself’. Chapter 7 (p.188/9) ‘General Moral’.

Now for the Section-headings. But first for a quite general point. Usually, though not quite always, the new move of a new chapter is not explicitly launched until the beginning of Section 1 of that chapter. Thus in Chapter 3, it’s not until the beginning of Section 1, i.e.at 3.1, that we get an explicit mention of sentences or propositions though this is what the whole Chapter is about. Similarly it is not until Section 1 in Chapter 6 that we get an explicit mention of the propositions of logic, though again this is what nearly the whole chapter is about (my interpretation). The sentences or paragraphs numbered, say, ‘3.0 … so and so’, or 6.0 ... so and so’ are usually something of a resumé of or extract from the previous chapter – the springboard for the new move, but not the new move itself. Chapter 5 seems to be an exception, in starting off directly with the new move.

Section Titles (to be taken with pinches of salt) In Chapter 1. Section 1 (i.e. 1.1 to 1.13) ‘The Totality of Facts’. Section 2 (i.e. 1.2 and 1.21) ‘Independence of Facts from one another’. In Chapter 2. Section 0 (i.e. 2 to 2.063) ‘Facts and Objects’.

182 If the notes on the Tractatus were written at about the same time as the bibliography also found in the ‘red box’, which was presumably meant as a reading-list for students studying the Tractatus, they are to be dated between 1955 and 1957; for in his bibliography Ryle refers to Wittgenstein’s ‘Notes on Logic’, which was published in the Journal of Philosophy in 1957, as ‘unpublished’, and he already mentions Moore’s articles in Mind on ‘Wittgenstein’s teaching in Cambridge, 1930-33’, which were published in 1954-1955.

199

The Philosophical Development of Gilbert Ryle

Section 1 (i.e. 2.1 to 2.19) ‘Pictures of Facts.’ Section 2 ‘Isomorphy of Pictures and Facts’. In Chapter 3. Section 0 ‘The Thinkable and the Unthinkable’. Section 1 ‘Sentences and Words’. Section 2 ‘Names and Objects’. Section 3 ‘Sentences and their Logical Syntax’. Section 4 ‘Sentences and “Logical Space”’. Section 5 ‘Sentences and Thoughts’. In Chapter 4. Section 0 Section 1 What Sentences Tell versus what they Show. Section 2 Atomic Propositions. Section 3 Molecular Propositions. Section 4 The Truth-conditions of Propositions. Section 5 The Most General Propositional Form. In Chapter 5. Section 0 Propositions as Truth-functions and Truth-arguments. Section 1 Truth-conditions and Inference. Section 2 Logical operations. Section 3 Truth-operations. Section 4 What Logical Operation-signs do and do not express. Section 5 Construction of all molecular propositions, generality, identity and special Forms. Section 6 The Limits of the Sayable. In Chapter 6. Section 0 The General Form of Truth Function. Section 1 The Propositions of Logic. Section 2 The Propositions of Mathematics (versus those of Logic). Section 3 The General Propositions of Natural Science (versus those of Logic). Section 4 The Propositions of Ethics (etc.) Section 5 The Propositions of Metaphysics.”

200

Dutch Summary

Dutch Summary

De Britse filosoof Gilbert Ryle (1900-1976) is één van de meest vooraanstaande filosofen van de 20e eeuw. Hij is vooral bekend geworden met zijn Concept of Mind (1949a) dat doorgaans als een behaviouristisch manifesto is gelezen. Ryle zou de wereld van de geest willen reduceren tot zichtbare gedragingen van de mens. Dit proefschrift heeft tot doel deze interpretatie te corrigeren, en doet dit door Ryle’s gehele filosofische ontwikkeling te bestuderen vanaf zijn vroegste papers, geschreven onder de invloed van Bertrand Russell, tot en met zijn latere werk over het menselijke denken, geïnspireerd onder andere door de ideeën van Ludwig Wittgenstein. Daarbij wordt voor het eerst gebruik gemaakt van ongepubliceerd materiaal uit het Ryle archief in Oxford, dat tot op heden praktisch onbekend is gebleven. Uit al dit gepubliceerd en ongepubliceerd materiaal komt een heel andere Ryle naar voren: een filosoof die de geest niet wilde reduceren tot de materie of tot uiterlijke gedragingen. Ryle is vooral een filosoof die de concepten waarmee wij over ons mentale leven praten in kaart wilde brengen. Zijn verdiensten blijken dan aanzienlijk te zijn, en gaan verder dan zijn beroemde onderscheid tussen ‘knowing how’/’knowing that’, zijn notie van de categorie-fout en zijn befaamde verdrijving van de ‘ghost’ uit de ‘machine’. Wanneer de interpretatie van Ryle als behaviourist eenmaal is ontzenuwd, is de weg vrij voor eerherstel van deze genuanceerde filosoof, wiens ideeën over betekenis, taal en denken een vruchtbare rol zouden kunnen spelen in hedendaagse debatten op het gebied van taalfilosofie, de filosofie van de geest en cognitieve psychologie. Hoofdstuk 1 biedt een biografische schets van Ryle, waarvoor niet alleen gebruik gemaakt is van gepubliceerde bronnen en het Ryle archief, maar ook van verhalen en anekdotes die Ryle’s familieleden, voormalige studenten, collega’s en vrienden ter beschikking hebben gesteld. Dit heeft het mogelijk gemaakt om een completer en sprekender beeld van hem te geven, als filosoof en als persoon. Uit dit hoofdstuk wordt duidelijk dat Ryle een grote invloed heeft gehad op de filosofische traditie in Oxford, maar ook op de filosofie in het algemeen. Door zijn manier van doceren en schrijven en zijn aandacht voor filosofen die destijds vrijwel niet in Oxford werden bestudeerd heeft hij generaties filosofen kunnen inspireren. Zo werd the Concept of Mind (1949a) een klassieker die bestudeerd is door vrijwel iedere student in Groot-Brittannië en ook daarbuiten. Ook had Ryle een grote rol bij het opzetten van de B.Phil in Oxford en doneerde hij een aanzienlijk deel van zijn persoonlijke bibliotheek aan Linacre College Library. Daarnaast zijn noemenswaardig zijn rol bij het opstarten van Analysis, zijn hoofdredacteurschap van Mind, en de hoofdrol die hij speelde bij de verhuizing van de Brentano Society van Praag naar Oxford in 1938-9 en het behoud van de Waismann erfenis voor toekomstige generaties. Deze en andere activiteiten hebben hem tot een belangrijke figuur binnen de Britse filosofie gemaakt, wiens mening vrijwel altijd werd gevraagd bij de invulling van academische posities. Hoofdstukken 2 en 3 laten Ryle’s ontwikkeling zien van een homogene, denotationele betekenistheorie – een betekenistheorie volgens welke alle concepten en uitdrukkingen op dezelfde manier betekenis hebben – naar een heterogene theorie van ‘gebruik’. Vanuit een positie dicht bij Russell’s zoektocht naar een ideale taal ontwikkelt Ryle zich naar een positie die meer in overeenstemming is met die van de latere Wittgenstein. Hij verlaat al snel de Russelliaanse methode van het analyseren en

201

The Philosophical Development of Gilbert Ryle herformuleren van uitdrukkingen en begint zich te realiseren dat denotationalisme verantwoordelijk is voor veel misverstanden, non-vragen en problematische theorieën binnen de filosofie. Ryle’s focus ligt dan niet langer op ‘betekenis’, wat voor hem impliceert dat er een homogene relatie tussen taal en werkelijkheid is en dat alle uitdrukkingen op dezelfde manier betekenis hebben. Zijn nadruk ligt nu op ‘gebruik’, en met name op de diversiteit van de manieren waarop uitdrukkingen gebruikt worden. Ryle’s methode van filosofische analyse verandert hierdoor van een logische en decompositionele manier van analyseren naar een zogenaamd connectivistisch analyse type. The Concept of Mind, Ryle’s meest bekende werk, heeft tot doel in kaart te brengen wat Ryle de ‘logische geografie’ van concepten van de geest noemt, dat wil zeggen de mogelijke en onmogelijke verbindingen tussen verschillende (types) concepten. Bepaalde operaties met concepten van de geest gaan in tegen de regels die gevolgd zouden moeten worden volgens de ‘logische geografie’ van deze concepten. Dit noemt Ryle categorie-fouten, namelijk het behandelen van uitdrukkingen alsof ze tot dezelfde categorie behoren (dat wil zeggen op dezelfde manier betekenis hebben) terwijl ze in werkelijkheid tot verschillende categorieën behoren. Ryle maakt niet langer gebruik van een reducerend analyse type en heeft het hier niet langer over een dieper niveau van logische vormen of feiten, noch probeert hij te komen tot een ideale taal. In The Concept of Mind wijst hij ons op categorie-fouten op verschillende niveaus, wat één van de redenen is waarom dit boek moeilijk te interpreteren valt. De meest algemene categorie-fouten ontstaan omdat mentale concepten behandeld worden als eigennamen en mentale taal als een beschrijving van een mentale wereld (denotationalisme). Met andere woorden, mentale concepten worden behandeld alsof ze op dezelfde manier betekenis hebben zoals ‘Fido’ (de naam) Fido (de hond) betekent. Er is echter een oneindige verscheidenheid aan betekeniscategorieën die niet in de mal die de logica van eigennamen ons oplegt kan worden geperst. Als we ons dit niet realiseren maken we een categorie-fout. Het is in dit kader belangrijk dat we kijken naar hoe uitdrukkingen gebruikt worden. De heterogeniteit van ons gebruik van taal neemt de plaats over van Ryle’s eerdere homogene, denotationele betekenistheorie. Mentale concepten dienen niet meer te worden behandeld alsof ze op dezelfde manier ‘betekenen’ als eigennamen. Als we mentale concepten behandelen als een beschrijving van een mentale wereld gaat het ook fout. Onze taal gaat uiteindelijk over onze normale wereld, en niet over een mentale en/of een fysieke wereld. Volgens Ryle zijn fysische en mentale taaluitingen verschillende manieren om dezelfde wereld te beschrijven. Dit is een conceptuele en geen ontologische positie. Uit hoofdstuk 3 wordt duidelijk dat Ryle vaak gezien is als een reductief behaviourist en dat dit misverstand grotendeels verantwoordelijk is voor het feit dat zijn denkbeelden in de regel snel van tafel worden geveegd. Sinds de opkomst van de cognitieve wetenschap en de invloedrijke verwerping van het behaviourisme door Chomsky (1959) en Fodor (1975) is elke associatie met het behaviourisme immers schadelijk geworden. De reden waarom Ryle zo vaak wordt gekenmerkt als behaviourist is dat zijn verwerping van een denotationele betekenistheorie niet voldoende wordt erkend. Ryle is tot op bepaalde hoogte zelf verantwoordelijk voor deze verkeerde interpretatie, want hij maakt niet altijd voldoende duidelijk waar hij zijn pijlen precies op richt. Bovendien is hij, ondanks de heldere stijl die hij over het algemeen hanteert, in The Concept of Mind niet in staat gebleken zelf het gebruik van vage concepten en uitdrukkingen te vermijden. De belangrijkste redenen waarom Ryle niet als reductief behaviourist getypeerd kan worden zijn de volgende: (1) een dergelijke interpretatie is niet in overeenstemming met zijn meta-filosofische doel; (2) reductief behaviourisme veronderstelt iets als een 202

Dutch Summary homogene, denotationele betekenistheorie en op het moment van schrijven van The Concept of Mind heeft Ryle deze positie reeds lang achter zich gelaten; en (3) de meest welwillende interpretatie van de voornaamste categorie-fouten die Ryle aanvalt in dit boek spreken tegen een behaviouristische interpretatie. Bevrijd van het juk van behaviourisme kan Ryle nog steeds een waardevolle bijdrage leveren aan filosofische discussies van vandaag de dag, zoals bijvoorbeeld discussies over de fundamenten en onderzoeksmethoden van de neurowetenschappen. In hoofdstuk 4 worden Ryle’s latere werken geanalyseerd: zijn theoretische artikelen, Dilemmas (1954a) en zijn latere werk over denken. De theoretische claims uit Ryle’s eerdere artikelen blijken niet in staat te zijn geweest het volledige gewicht van The Concept of Mind te kunnen dragen. Daarom schrijft Ryle in de late jaren 50 en vroege jaren 60 aanvullingen op zijn ‘oude’ theoretische denkbeelden over wat filosofie is en hoe men dient te filosoferen. Hij vervolgt hierbij zijn eerdere weg van het verwerpen van een denotationele betekenistheorie ten faveure van de heterogeniteit van ‘gebruik’. Zijn focus ligt nog immer op informele logica en het idee dat filosofische problemen van een andere aard zijn dan wetenschappelijke vraagstukken. Nieuwe elementen zijn Ryle’s instructies over de interpretatie van ‘het normale gebruik van uitdrukking x’, namelijk dat ‘gebruik’ op een logische in plaats van op een filologische manier gezien moet worden en dat ‘het normale gebruik van een uitdrukking’ niet hetzelfde is als ‘het gebruik van normale uitdrukkingen’. Daarnaast analyseert Ryle de verschillen tussen taal en spraak en tussen woorden en zinnen. Spraak heeft taal nodig maar kan niet tot taal gereduceerd worden. Als we spraak met taal verwarren komen we uit op een filosofie die eerder filologie is. Ook nieuw – als terminologie maar niet als gedachte – is zijn beschrijving van filosofie als activiteit die meerdere niveaus bestrijkt. In tegenstelling tot wetenschappelijke bewijzen zijn filosofische argumenten ‘operaties op operaties’, oftewel ze gaan over operaties. Dit is nu precies het verschil tussen filosofie en wetenschap: waar de wetenschap verklaart in de zin dat ze nieuwe feiten boven water probeert te krijgen en direct op deze feiten opereert, zo houdt de filosofie zich bezig met het niveau daarboven. Net als The Concept of Mind heeft ook Dilemmas tot doel een voorbeeld te geven van een goede manier om filosofie te bedrijven. Hoewel de specifieke oplossingen die Ryle aandraagt voor de dilemma’s vaak bekritiseerd zijn om hun oppervlakkigheid en onvolledigheid, is zijn meta-filosofische boodschap van grote waarde: zijn connectieve filosofische analyse legt de fouten bloot die we maken als de taal ons verleidt en in de val lokt en laat bovendien zien hoe we deze fouten kunnen herkennen en vermijden. In het laatste deel van hoofdstuk 4 staat Ryle’s latere werk over denken centraal, waarin hij bepleit dat onze concepten van denken beschrijvingen zijn (van processen) op een hoger niveau. Deze concepten zouden bijwoordelijk (‘adverbial’) zijn in de zin dat ze niet zelf direct processen beschrijven maar op een hoger niveau deze processen kwalificeren, net zoals bijwoorden (zoals ‘oplettend’) werkwoorden (zoals ‘autorijden’) kwalificeren. Ryle stelt zich de figuur van de Denker (Le Penseur) van Rodin voor en vraagt zich af: als het denken van Le Penseur slechts de manier is waarop processen worden bestuurd, over welk soort processen hebben we het dan eigenlijk? Ryle gebruikt ‘het denken van Le Penseur’ precies om die categorie van denken aan te duiden van welke het het moeilijkste te begrijpen is dat hij bijwoordelijk is (dat wil zeggen dat hij iets anders kwalificeert) en supervenieert op acties in plaats van zelf gezien te worden als een actie. Als het ‘denken’ van Socrates, Mozart en de wetenschapper geen proces beschijft maar een manier van ‘X-en’, wat kan dit ‘X-en’ dan zijn? Is het ‘reflecteren’ of ‘mediteren’? Wat ‘gebeurt’ er als we dat doen? Ryle’s voornaamste doel in zijn latere artikelen over denken is het beantwoorden van dit cluster van vragen. Zijn uiteindelijke antwoord is dat hoewel er

203

The Philosophical Development of Gilbert Ryle geen specifiek autonoom ‘X-en’ bestaat voor het denken van Le Penseur, we puur logisch gezien toch een niet-bijwoordelijk werkwoord nodig hebben om de bijwoorden aan te kunnen koppelen. In deze latere artikelen versterkt Ryle zijn eerdere positie zonder echt fundamenteel van gedachte te veranderen. Het bijwoordelijke element was bijvoorbeeld, hoewel minder prominent, al aanwezig in 1949. Hij probeert te laten zien dat een beschrijving van het denken van Le Penseur geen beschrijving is van complexe processen, maar een complexe beschrijving van processen. Dit stelt hem in staat zowel dualisme als behaviourisme te verwerpen, welke beide aannemen dat alle beschrijvingen functioneren als zinnen die refereren naar processen of dingen (denotationalisme). Ryle is dus geen dualist of behaviourist, ook geen conceptueel dualist of conceptueel behaviourist. Zijn positie kan het beste gekarakteriseerd worden als ‘conceptueel pluralisme’, de positie dat concepten niet op één enkele manier behandeld moeten worden, of op twee manieren, maar op oneindig veel verschillende manieren, afhankelijk van de context waarin een concept wordt gebruikt. Hoewel Ryle’s latere artikelen over denken kritische geluiden hebben opgeroepen zijn zijn oplossingen origineel en stellen de onderscheidingen waarop hij ons wijst ons in staat een duidelijker beeld te krijgen van ons normale gebruik van de concepten van denken. Het onderwerp van hoofdstuk 5 is de filosofische relatie tussen Ryle en Collingwood en hun korte correspondentie uit 1935. In deze correspondentie verschillen Ryle en Collingwood veelvuldig van mening maar de onderliggende en meer fundamentele punten komen niet boven water. Ik beweer in dit hoofdstuk dat hun verschillen van mening teruggevoerd kunnen worden op hun verschillende visies op de aard van filosofie en logica, de aard van metafysica en de definitie van een universele propositie. Collingwood’s overlap van klassen binnen de filosofie en Ryle’s idee dat er een algemene logica is die toepasbaar is op alle gebieden zijn hier van groot belang. Verder is ook Collingwood’s gebruik van idiosyncratische terminologie verantwoordelijk voor Ryle’s onterechte aanname dat hij sterk ontologische claims maakte. Gewapend met dit betere begrip van de correspondentie en haar filosofische context bediscussieer ik tenslotte een aantal interpretaties met betrekking tot de filosofische relatie tussen Collingwood en Ryle. Ryle’s studie en interpretatie van Wittgenstein worden behandeld in hoofdstuk 6. In het eerste deel zien we dat Ryle Wittgenstein’s werk uitgebreid bestudeerd heeft en de invloed van Wittgenstein op zijn eigen werk expliciet heeft erkend. Ryle’s interpretatie van Wittgenstein ligt dichtbij die van degenen die in het huidige Wittgenstein debat beweren dat er slechts één Wittgenstein is en geen twee (laat staan drie). Wat vooral opvalt is dat Ryle de Tractatus dichterbij Wittgenstein’s latere werk plaatst dan traditionele interpretatoren. Interessant is in dit kader ook Ryle’s vroege studie van Wittgenstein’s ‘Notes on Logic’ uit 1913. Door bestudering van de ‘Gilbert Ryle Collectie’ in Oxford, en dan voornamelijk Ryle’s geannoteerde exemplaren van Wittgenstein’s Tractatus, Philosophical Investigations en zijn commentaar op ‘Notes on Logic’, kan een aantal nieuwe feiten over Ryle’s studie en interpretatie van Wittgenstein worden onthuld. Het tweede deel van hoofdstuk 6 bestaat uit een vergelijking tussen Ryle en Wittgenstein. Noemenswaardige verschillen zijn dat Ryle een aversie heeft tegen Wittgenstein’s gebruik van ‘taalspelen’. Hij ontkent ook de waarde van diens ‘familiegelijkenissen’ en spreekt in tegenstelling tot Wittgenstein niet over de grenzen van ons taalgebruik. Sommige filosofen hebben beweerd dat Wittgenstein ‘dieper’ gaat dan Ryle met betrekking tot de aard van filosofie. Het onderscheid dat Wittgenstein maakt tussen grammaticale en empirische proposities is bijvoorbeeld niet terug te vinden bij Ryle. En Wittgenstein besteedt ook meer aandacht aan de constitutieve rol van taal voor de 204

Dutch Summary betekenis van concepten. Behalve deze verschillen zijn er ook belangrijke overeenkomsten tussen beide filosofen. Beiden hebben een grote interesse in de notie van denken, leren denken en creatief denken. Ze verwerpen Cartesiaans dualisme, behaviourisme, denotationalisme, etc. – soms expliciet en op andere momenten op een minder directe manier. Zowel Ryle als Wittgenstein gebruikt argumenten, verwerpt en rectificeert, hoewel dit bij Wittgenstein minder duidelijk is door de vorm en stijl van zijn werk. Ze hebben grotendeels met elkaar overeenkomende visies op betekenis en ons gebruik van taal. Hoewel Ryle’s behandeling van concrete filosofische vraagstukken niet altijd als afdoende werd gezien door tijdgenoten en hij soms, vooral aan het eind van zijn filosofische loopbaan, te veel verstrikt raakte in ‘mannerisms’, heeft hij een belangrijke invloed uitgeoefend op andere filosofen, bijvoorbeeld op A. J. Ayer, Daniel Dennett, D. M. Armstrong, C. B. Martin en U. T. Place. Bovendien wordt een aantal van de onderscheidingen die hij geïntroduceerd heeft nog steeds bediscussieerd, zoals het onderscheid tussen ‘knowing how’ en ‘knowing that’ en tussen ‘dispositioneel’ en ‘categorisch’. Zijn visies op de aard en methode van de filosofie en de instrumenten die hij ons heeft aangedragen voor het oplossen van filosofische problemen hebben grote waarde; niet alleen vanuit historisch oogpunt maar ook omdat ze relevant zijn in huidige filosofische discussies, bijvoorbeeld met betrekking tot cognitieve wetenschap of taalfilosofie. Dan moet Ryle echter wel eerst bevrijd worden van het foutieve label van behaviourisme; ik hoop dat dit proefschrift hiertoe heeft bijgedragen.

205

The Philosophical Development of Gilbert Ryle

206

References

References

Personal e-mail correspondence with Wilhelm Baumgartner, co-founder of the Internationale Franz Brentano Gesellschaft, 11 and 18 August 2004

An e-mail correspondence and conversations with Michael Ryle, Gilbert Ryle’s nephew, and his wife Bridget Ryle, between 15 January 2002 and 2006

Personal conversation and e-mailcorrespondence with John Rogers, student and friend of Ryle from 1962-76, between June 2002 and 2006

Personal conversations and e-mailcorrespondence with Rom Harré, David Pears and Brian McGuinness, Ryle’s former colleagues, between 2002 and 2005

Personal conversation with Jerry Cohen and Tony Palmer, Ryle’s former students

Notes Critiques Dilemmas, Revue de Métaphysique et de Morale, vol. 63, 1958, p. 499

ACKRILL, J. L., 1965, ‘Aristotle’s Distinction between Energeia and Kinēsis’, in Ackrill, J.L., Essays on Plato and Aristotle (Oxford: Oxford University Press 2001)

ACKRILL, J. L., 1970, ‘In Defence of Platonic Division’, in Wood and Pitcher 1971

AMBROSE, Alice, 1955, ‘Review Dilemmas by Gilbert Ryle’, in The Journal of Philosophy, vol. 6, 1955, pp. 155-159

ANTONIOL, Lucie, 1993, Lire Ryle Aujourd’hui (Brussels : DeBoeck-Wesmael 1993)

ANSELM, Monologion and Proslogion with the Replies of Gaunilo and Anselm, transl. Williams, Thomas (Indianapolis/Cambridge: Hackett Publishing Company 1996)

ARMSTRONG, D. M., 1990, ‘The Causal Theory of the Mind’, in Lycan, William G. (ed.), Mind and Cognition (Oxford and Cambridge Mass.: Blackwell 1990)

AUSTIN, J. L., 1950, ‘Intelligent behaviour. A critical review of The Concept of Mind’, in The Times Literary Supplement, April 7, 1950, reprinted in Wood and Pitcher 1971

AYER, A. J., 1933, ‘Letter to Gilbert Ryle’, in The Linacre Journal, The Gilbert Ryle Issue (Oxford November 1999)

AYER, A. J., 1970, ‘An honest Ghost?’, in Wood and Pitcher 1971

AYER, A. J., 1977, Part of my life (London: Collins 1977)

AYER, A. J., 1984, More of my life (London: Collins 1984)

207

The Philosophical Development of Gilbert Ryle

BAMBROUGH, Renford, 1994a, ‘Review Dilemmas by Gilbert Ryle’, in Philosophy, vol. 69, 1994, pp. 378-380

BAMBROUGH, Renford, 1994b, ‘Gilbert Ryle: Collected Papers‘, in Philosophy, vol. 69, 1994, pp. 376-378

BARNES, Winston H. F., 1955, ‘Tangles Unravelled’, in The Philosophical Quarterly, vol. 5, 1955, pp. 355-364

BENNETT, M. R. and HACKER, P. M. S., 2003, Philosophical Foundations of Neuroscience (Blackwell publishing, Oxford, Malden USA and Victoria Australia 2003)

BESTOR, Thomas W., 1979, ‘Gilbert Ryle and the Adverbial Theory of Mind’, in The Personalist 60 (1979), pp. 233-242

BLACK, Max, 1964, A Companion to Wittgenstein’s ‘Tractatus (Cambridge: University Press 1964)

BOUCHER, David, 1995, ‘The Life, Times and Legacy of R.G. Collingwood’, in Boucher, Connelly and Modood 1995

BOUCHER, David, CONNELLY, James and MODOOD, Tariq, 1995, Philosophy, History and Civilization. Interdisciplinary Perspectives on R.G. Collingwood (Cardiff: University of Wales Press 1995)

BOUWSMA, O. K., 1972, ‘A Difference Between Ryle and Wittgenstein’, in Kolenda 1972

BOUWSMA, O. K., 1986, Wittgenstein: conversations, 1949-1951 (Indianapolis: Hackett Publishers 1986)

CHOMSKY, Naom, 1959, ‘A Review of B. F. Skinner’s Verbal Behavior’, in Language, 35, 1959, pp. 26-58

COLLINGWOOD, R. G., 1916, Religion and Philosophy (London: Macmillan 1916), reprinted edition (Bristol: Thoemmes 1994)

COLLINGWOOD, R. G., 1919, ‘Lectures on the Ontological Proof of the Existence of God’, written in December 1919, delivered Hilary Term 1920, unpublished, kept at the Bodleian Library in Oxford

COLLINGWOOD, R. G., 1923, review of ‘Spearman’s The Nature of ‘Intelligence’ and the Principles of Cognition (London 1923)’, in The Oxford Magazine 42 (1923-1924), pp. 117- 118

COLLINGWOOD, R. G., 1933, An Essay on Philosophical Method (Oxford: Clarendon Press 1977)

208

References

COLLINGWOOD, R. G., 1934, ‘The Permanent Problems of Metaphysics’ and ‘The Special Problems of Modern Metaphysics’, two lectures held on 15 and 17 January 1934, opening a course of 16 lectures on Metaphysics by various speakers, kept at the Bodleian Library in Oxford (MS Collingwood Dep. 18.2); partly published as ‘The Nature of Metaphysical Study’ in Collingwood (1940)

COLLINGWOOD, R. G., 1935a, The Ryle-Collingwood Correspondence, letter 1 and 3, kept at the Bodleian Library in Oxford (MS Eng Lett d194); recently published in: Connelly and D’Oro 2005

COLLINGWOOD, R. G., 1935b, ‘Method and Metaphysics’, unpublished paper presented at the Jowett Society on 19 June 1935, kept at the Bodleian Library in Oxford (MS Collingwood Dep. 19(3))

COLLINGWOOD, R. G., 1939, An Autobiography (London/Oxford/New York: Oxford University Press 1970)

COLLINGWOOD, R. G., 1940, An Essay on Metaphysics, revised edition (Oxford: Clarendon Press 2002)

COLLINGWOOD, R. G., 1942, The New Leviathan, revised edition (Oxford: Clarendon Press 1992)

CONANT, James, 2000, ‘Elucidation and Nonsense in Frege and Early Wittgenstein’, in Crary and Read 2000

CONNELLY, James, 1990, ‘Metaphysics and Method: A Necessary Unity in the Philosophy of R. G. Collingwood’, in Storia, antropologica e Scienze del linguaggio, vol. 5/1-2 (Roma: Bulzoni editore 1990)

CONNELLY, James, 1995, ‘Art Thou the Man: Croce, Gentile or de Ruggiero?’, in Boucher, Connelly and Modood 1995

CONNELLY, James, 2003, Metaphysics, Method and Politics: The Political Philosophy of R.G. Collingwood (Exeter: Imprint Academic 2003)

CONNELLY, James and D’ORO, Giuseppina (eds.), 2005, in An Essay on Philosophical Method, revised edition (Oxford: Clarendon Press 2005)

COOK WILSON, John, 1926, Statement and Inference, 2 vols., A. S. L. Farquharson (ed.), reprinted (Oxford: Clarendon Press 1969)

COULTER, Jeff, 2003, ‘Ryle’s “Le Penseur”’, in Revue Internationale de Philosophie, vol. 57, 1/2003, pp. 67-78

CRARY, Alice and READ, Rupert, 2000, (London: Routledge 2000)

DEMOS, Raphael, 1917, ‘A Discussion of a Certain Type of Negative Proposition’, in Mind, vol. XXVI, 1917

209

The Philosophical Development of Gilbert Ryle

DIAMOND, Cora, 2000, ‘Ethics, Imagination and Method of Wittgenstein’s Tractatus’, in Crary and Read 2000

DENNETT, Daniel, 2002, ‘An Interview with Daniel Dennett, August 2002’, in The Dualist, Stanford’s Undergraduate Journal of Philosophy, vol. 9

DONAGAN, A., 1962, The Later Philosophy of R.G. Collingwood, reprinted edition (Chicago/London: The University of Chicago Press 1985)

DONAGAN, A., 1967, ‘R. G. Collingwood’, in Paul Edwards (ed.), The Encyclopedia of Philosophy (New York: Macmillan 1967), pp. 140-144

O’DONOHUE, William, and KITCHENER, Richard, 1999, ‘Introduction: The Behaviorisms’, in O’Donohue and Kitchener 1999

O’DONOHUE, William, and KITCHENER, Richard, 1999, Handbook of Behaviorism (San Diego: Academic Press 1999)

DUSSEN, van der, W. J., 1981, History as a Science: The Philosophy of R.G. Collingwood (The Hague/Boston/London: Nijhoff 1981)

FLEW, Antony, 1998, Philosophical Essays, Shosky, John (ed.) (Oxford: Rowman and Littlefield 1998)

FLEW, Antony, 1999, ‘Gilbert Ryle: A Personal Recollection’, in Harré and Shosky 1999

FODOR, Jerry A., 1968, Psychological Explanation: An introduction to the philosophy of psychology (New York: Random House 1968)

FODOR, Jerry A., 1975, The Language of Thought (New York: Crowell 1975), reprinted edition (Hassocks: The Harvester Press 1976)

GELLNER, E., 1959, Words and Things (London: Gollancz 1959)

GOCHET, Paul, 1972, Outline of a Nominalist Theory of Propositions (Dordrecht: D. Reidel Publishing Company 1980), transl. of Esquisse d’une théorie nominaliste de la proposition (Paris: Armand Colin 1972)

GÖTLIND, Erik, 1956, ‘Review Gilbert Ryle: Dilemmas’, in Theoria, vol. 22, 1956, pp. 69- 71

HACKER, P. M. S., 1996, Wittgenstein’s Place in Twentieth-Century Analytic Philosophy (Oxford/Malden: Blackwell Publishers 1996)

HALE, Bob and WRIGHT, Crispin (eds.), 1997, A Companion to the Philosophy of Language, reprinted edition (Oxford and Malden Mass.: Blackwell Publishers 2000)

210

References

HAMPSHIRE, Stuart, 1950, ‘Critical review of The Concept of Mind’, Mind, Vol. LIX, 1950, reprinted in Wood and Pitcher 1971

HANFLING, Oswald, 2000, Philosophy and Ordinary Language. The Bent and Genius of Our Tongue (London: Routledge 2000)

HARRÉ, R., and SHOSKY, J. (eds.), 1999, The Linacre Journal, The Gilbert Ryle Issue (Oxford November 1999)

HARRÉ, R., 2001, ‘Wittgenstein: Science and Religion’, in Philosophy 76 (Cambridge University Press 2001), pp. 211-237

HARRIS, Errol, 1936, ‘Mr. Ryle and the Ontological Argument’, in Mind Vol. XLV, 1936, pp. 474-480

HARRIS, H. S., 1995, ‘Croce and Gentile in Collingwood’s New Leviathan’, in Boucher, Connelly and Modood 1995

HERTZ, Heinrich, 1899, The Principles of Mechanics, reprinted edition, English translation by Jones, D. E., and Walley, J. T. (New York: Dover Publications 1956)

HOFSTADTER, Albert, 1951, ‘Professor Ryle’s Category-Mistake’, in The Journal of Philosophy 48, 1951

HUMPHREY, George, 1953, ‘Thinking’, in Acta Psychologica, vol. IX, 1953

JOSEPH, H. W. B., 1910/11, ‘The Psychological Explanation of the Development of the Perception of External Objects’, in Mind 19 (1910): pp. 306-21, 457-69 and Mind 20, 1911, pp. 1-14

KAPLAN, Abraham, 1955, ‘Review Dilemmas by Gilbert Ryle’, in (1955), pp. 644-646

KENNY, Anthony (ed.), 1994, The Wittgenstein Reader (Oxford: Blackwell 1994)

KIM, Jeagwon, 1996, Philosophy of Mind (Oxford and Boulder Colorado: Westview Press 1996)

KING, Hugh R., 1951, ‘Professor Ryle and the Concept of Mind’, in The Journal of Philosophy 48, 1951

KIRK, John R., 1955, ‘Of Words and Deeds. Dilemmas by Gilbert Ryle’, in The Humanist, vol. 15, 1955, p. 96

KITCHENER, Richard F., 1999, ‘Logical Behaviorism’, in Kitchener and O’Donohue 1999

KOLENDA, Konstantin, 1972, Studies in Philosophy. A Symposium on Gilbert Ryle (Rice University Studies 1972), Vol. 58, No. 3

211

The Philosophical Development of Gilbert Ryle

KOLENDA, Konstantin (ed.), 1979, On Thinking (Oxford: Basil Blackwell 1979)

LINKE, Paul F., 1926, ‘The Present State of Logic and Epistemology in German’, in the Monist, April 1926

LOWE, E. J., 2000, An Introduction to the Philosophy of Mind (Cambridge: Cambridge University Press 2000)

LUCKHARDT, C. G., 1979, Wittgenstein: Sources and Perspectives (Hassocks, Sussex: The Harvester Press 1979)

LYCAN, William G., 2003, in Devitt, Michael, Hanley, Richard (eds.) The Blackwell Guide to Philosophy of Mind (Malden: Blackwell 2003)

LYONS, William, 1980a, ‘Ryle’s Three Accounts of Thinking’, in Lyons 1980b

LYONS,William, 1980b, Gilbert Ryle. An Introduction to his Philosophy (Brighton Sussex and New Jersey: The Harvester Press Limited and Humanities Press Inc. 1980)

MABBOTT, J. D., 1929, ‘Negation’, symposium with Ryle and Price, Proceedings of the Aristotelian Society, suppl. vol. IX, 1929

MABBOTT, J. D., 1976, ‘Gilbert Ryle: A Tribute’, in: Meyer, R.S., Aspects of Mind. Gilbert Ryle (Oxford: Blackwell 1993)

MABBOTT, J. D., 1986, Oxford Memories (Oxford: Thornton’s of Oxford 1986)

MACDONALD, Margaret, 1951, ‘Professor Ryle on The Concept of Mind’, in Philosophical Review 1951

MACLELLAN, P. S., 1951, ‘Professor Ryle and the Concept of Mind’, in Hibbert-Journal 1951

MAGEE, B., 1971, Modern British Philosophy (London: Secker & Warburg 1971)

MAGEE, B., 1997, Confessions of a Philosopher (London: Weidenfeld & Nicolson 1997)

MARION, Mathieu, 2000a, ‘Oxford Realism: Knowledge and Perception I’, The British Journal for the History of Philosophy 8 (2) 2000, pp. 299-338

MARION, Mathieu, 2000b, ‘Oxford Realism: Knowledge and Perception II’, The British Journal for the History of Philosophy 8 (3) 2000, pp. 485-519

MARTIN, Rex, 1974, ‘Collingwood’s Essay on Philosophical Method’, Idealistic Studies 4.3, September 1974, pp. 224-250

MARTIN, Rex, 1995, ‘Collingwood’s Claim that Metaphysics is a Historical Discipline’, in Boucher, Connelly and Modood 1995

212

References

MATTHEWS, Gareth B., 1981, Review On Thinking by Gilbert Ryle, The Philosophical Review 1981, pp. 443-444

MCGUINNESS, Brian, 1972, ‘Bertrand Russell and Ludwig Wittgenstein’s “Notes on Logic”’, in Revue Internationale de Philosophie 26, 1972

MCGUINNESS, Brian and VRIJEN, Charlotte, 2006, ‘Early Thoughts: An Early Letter from Gilbert Ryle to H. J. Paton’, in The British Journal for the History of Philosophy, 2006

MEHTA, V. P., 1963, Fly and the Fly-Bottle (London: Weidenfeld & Nicolson 1963)

MEYER, René (ed.), 1993, Aspects of Mind (Oxford UK and Cambridge USA: Blackwell 1993)

MILLER, Dickinson, 1951, ‘Descartes’ Myth and Professor Ryle’s Fallacy’, in Journal of Philosophy 48, 1951

MINK, Louis O., 1969, Mind, History and Dialectic. The Philosophy of R.G. Collingwood (Bloomington: Indiana University Press 1969)

MODOOD, Tariq, 1989, ‘The Later Collingwood’s Alleged Historicism and Relativism’, in Journal of the History of Philosophy 27, 1989, pp. 101-125

MODOOD, Tariq, 1995, ‘Collingwood and the Idea of Philosophy’, in Boucher, Connelly and Modood 1995

MORAVCSIK, J., 1977, ‘Gilbert Ryle’, unpublished and kept in the Philosophy Library in Oxford

MONK, R., 1990, Ludwig Wittgenstein. The Duty of Genius, reprinted edition (London: Vintage 1991)

MOORE, George Edward, 1910-11, Some Main Problems of Philosophy, series of Lectures delivered in London in the winter of 1910-11 (London: George Allen & Unwin 1953)

OLDFIELD, Adrian, 1995, ‘Metaphysics and History in Collingwood’s Thought’, in Boucher, Connelly and Modood 1995

D’ORO, Giuseppina, 2000, ‘On Collingwood’s rehabilitation of the ontological argument’, in Idealistic Studies: An International Philosophical Journal 30(3) (The Hague: Nijhoff 2000), pp. 173-188

D’ORO, Giuseppina, 2002, Collingwood and the Metaphysics of Experience (London/New York: Routledge 2002)

D’ORO, Giuseppina, 2003, ‘Collingwood and Ryle on the Concept of Mind’, in Philosophical Explorations VI(I), 2003, pp. 18-30

213

The Philosophical Development of Gilbert Ryle

PALMER, Anthony, 2002, ‘Propositions, Properties and Relations. Wittgenstein’s ‘Notes on Logic’ and the Tractatus’, delivered to the Wittgenstein Workshop at the University of Chicago in April 2002

PARK, Shelley M., 1994, ‘Reinterpreting Ryle: A Nonbehavioristic Analysis’, in Journal of the History of Philosophy 32, 1994, pp. 265-290

PASSMORE, John, 1957, A Hundred Years of Philosophy, reprinted edition (Harmondsworth: Penguin Books 1970)

PETERS, Rik, 1998, The Living Past. Philosophy, History and Action in the Thought of Croce Gentile, de Ruggiero, and Collingwood (Nijmegen: [sn] 1998)

PLACE, Ullin T., 1999, ‘Ryle’s Behaviorism’, in O’Donohue, William, and Kitchener, Richard, Handbook of Behaviorism (San Diego: Academic Press 1999)

POPPER, Karl, 1945, The Open Society and its Enemies (London: Routledge 1945)

PREST, J. (ed.), 1993, The illustrated History of Oxford University (Oxford: Oxford University Press 1993)

PRICHARD, H. A., 1907, ‘A Criticism of the Psychologists’ Treatment of Knowledge’, in Mind 16, 1907, pp. 27-53

QUINTON, A. M., 1954, ‘Dilemmas by Gilbert Ryle’, The Hibbert Journal, vol. 53, 1954, pp. 89-91

QUINTON, A. M., 1998, From Wodehouse to Wittgenstein (New York: St. Martin’s Press 1998)

RAMOINO-MELILLI, Grazia, 1983, Filosofia e analisi in Gilbert Ryle (Pisa: ETS 1983)

RAMOINO-MELILLI, Grazia, 2003, ‘Ryle Revisited : The Dispositional Model Fifty Years After’, The British Journal for the History of Philosophy, 11(1) 2003, pp. 89-119

RECK, Erich H., 2002, From Frege to Wittgenstein (New York etc.: Oxford University Press 2002)

RÉE, J., 1993, ‘English Philosophy in the Fifties’, in: Radical Philosophy 65, autumn 1993

ROGERS, Ben, 2000, A.J. Ayer. A Life (London: Vintage 2000)

RORTY, Richard (ed.), 1967, The (Chicago: University of Chicago Press 1967)

RUBINOFF, Lionel, 1970, Collingwood and the Reform of Metaphysics (Toronto/Buffalo: University of Toronto Press 1970)

214

References

RUSSELL, Bertrand, 1903, The Principles of Mathematics (London: Bradford & Dickens 1948)

RUSSELL, Bertrand, 1905, ‘On Denoting’, Foundations of Logic 1903-05, in The Collected Papers of Bertrand Russell, vol. 4 (London: Routledge 1994)

RUSSELL, Bertrand, 1910-13, Principia Mathematica (Cambridge: Cambridge University Press, 1910-13)

RUSSELL, Bertrand, 1912, The Problems of Philosophy, reprinted edition (London, etc.: Oxford University Press 1967)

RUSSELL, Bertrand, 1913, Theory of Knowledge, first published in The Monist in 1914- 1915, reprinted in The Collected Papers of Bertrand Russell, vol. 7 (London: Routledge 1983)

RUSSELL, Bertrand, 1918, ‘The Philosophy of Logical Atomism’, reprinted in Pears, David (ed.), Russell’s Logical Atomism (London: Collins Sons & Co 1972)

RUSSELL, Bertrand, 1958a, ‘What is Mind?’, in Journal of Philosophy 55, 1958, reprinted in The Collected Papers of Bertrand Russell, volume II

RUSSELL, Bertrand, 1958b, ‘Mathematical Infinity’, Mind, Vol. 67 July 1958

RUSSELL, Bertrand, 1973, letter to Ogden, 21 November 1960, in Von Wright, G. H. (ed.), Letters to C. K. Ogden: with Comments on the English Translation of the Tractatus Logico- Philosophicus (Oxford: Blackwell and London: Routledge and Kegan Paul Ltd 1973)

RUSSELL, L. J., 1955, ‘Review Dilemmas by Gilbert Ryle’, The British Journal for the Philosophy of Science, vol. 5, 1955, pp. 346-348

RYLE, Gilbert, Interviews Urmson and Magee held with Ryle, date unknown, kept at the Philosophical Centre in Oxford

RYLE, Gilbert, Correspondence between Ryle and Popper, The Popper Collection of Hoover Institution Archives, Stanford

RYLE, Gilbert, Correspondence between Ryle and Russell, McMaster’s Russell Archive, Hamilton, Ontario

RYLE, Gilbert, 1928, ‘Heidegger’s ‘Sein und Zeit’’, Mind, vol. XXXVIII, 1928, reprinted in Ryle 1971a

RYLE, Gilbert, 1929, ‘Negation’, syposium with Mabbott and Price, Proceedings of the Aristotelian Society, suppl. vol. IX, 1929, reprinted in Ryle 1971b

RYLE, Gilbert, 1930, ‘Are There Propositions?’, Proceedings of the Aristotelian Society, vol. XXX, 1930, reprinted in Ryle 1971b

215

The Philosophical Development of Gilbert Ryle

RYLE, Gilbert, 1932a, ‘Phenomenology’, Proceedings of the Aristotelian Society, vol. XI, 1932, reprinted in Ryle 1971a

RYLE, Gilbert, 1932b, ‘Systematically Misleading Expressions’, Proceedings of the Aristotelian Society vol. XXXII, 1932, reprinted in Ryle 1971b

RYLE, Gilbert, 1933a, ‘Imaginary Objects’, Proceedings of the Aristotelian Society, suppl. vol. XII, 1933, reprinted in Ryle 1971b

RYLE, Gilbert, 1933b, ‘About’, Analysis, vol. 1, 1933, reprinted in Ryle 1971b

RYLE, Gilbert, 1935a, ‘Mr. Collingwood and the Ontological Argument’, in Mind XLIV, 1935, reprinted in Ryle 1971b

RYLE, Gilbert, 1935b, The Ryle-Collingwood Correspondence, letter 2, kept at the Bodleian Library in Oxford (MS Eng Lett d194); recently published in: Connelly and D’Oro 2005

RYLE, Gilbert, 1935c, ‘Internal Relations’, Proceedings of the Aristotelian Society, suppl. vol. XIV, 1935, reprinted in Ryle 1971b

RYLE, Gilbert, 1936, ‘Unverifiability-By-Me’, Analysis, vol. IV, 1936, reprinted in Ryle 1971b

RYLE, Gilbert, 1937a, ‘Back to the Ontological Argument’, in Mind XLVI, 1937, reprinted in Ryle 1971b

RYLE, Gilbert, 1937b, ‘Taking Sides in Philosophy’, Philosophy, vol. XII, 1937, reprinted in Ryle 1971b

RYLE, Gilbert, 1938, ‘Categories’, Proceedings of the Aristotelian Society, vol. XXXVIII, 1938, reprinted in Ryle 1971b

RYLE, Gilbert, 1940, ‘Conscience and Moral Convictions’, Analysis, vol. VII, 1940, reprinted in Ryle 1971b

RYLE, Gilbert, 1945, ‘Philosophical Arguments’, Originally delivered as the Inaugural Lecture as Waynflete Professor of Metaphysical Philosophy in 1945, reprinted in Ryle 1971b

RYLE, Gilbert, 1946a, ‘Knowing How and Knowing That’, Proceedings of the Aristotelian Society, vol. XLVI, 1946, reprinted in Ryle 1971b

RYLE, Gilbert, 1946b, ‘Why Are The Calculuses of Logic and Arithmetic Applicable to Reality?’, Proceedings of the Aristotelian Society, suppl. vol. XX, 1946, reprinted in Ryle 1971b

RYLE, Gilbert, 1949a, The Concept of Mind (London: Penguin Books 1990)

216

References

RYLE, Gilbert, 1949b, ‘Discussion of Rudolf Carnap: ‘Meaning and Necessity’’, Philosophy, vol. XXIV, 1949, reprinted in Ryle 1971a

RYLE, Gilbert, 1950, ‘Logic and Professor Anderson’, The Australasian Journal of Philosophy, Vol. XXVIII, 1950, pp. 137-53, reprinted in Ryle 1971a

RYLE, Gilbert, 1951a, ‘Thinking and Language’, symposium with Iris Murdoch and A. C. Lloyd, Proceedings of the Aristotelian Society, suppl. vol. XXV, 1951, reprinted in Ryle 1971b

RYLE, Gilbert, 1951b, ‘The Verification Principle’, Revue Internationale de Philosophie, vol. V, 1951, translated and reprinted in Ryle 1971b

RYLE, Gilbert, 1951c, ‘Ludwig Wittgenstein’, Analysis 1951, reprinted in Ryle 1971b.

RYLE, Gilbert, 1953a, ‘Thinking’, Acta Psychologica, vol. IX, 1953, reprinted in Ryle 1971b

RYLE, Gilbert, 1953b, ‘Ordinary Language’, The Philosophical Review, vol. LXII, 1953, reprinted in Ryle 1971b

RYLE, Gilbert, 1954a, Dilemmas (Cambridge: Cambridge University Press 1954)

RYLE, Gilbert, 1954b, ‘Proofs in Philosophy’, Revue Internationale de Philosophie, vol. VIII, 1954, reprinted in Ryle 1971b

RYLE, Gilbert, 1954c, report on ‘If a distraction makes me forget my headache, does it make my head stop aching, or does it only stop me feel it aching?’, Analysis 14, 1954

RYLE, Gilbert, 1957a, ‘The Theory of Meaning’, in Mace, C. A., British Philosophy in Mid- Century (1957), reprinted in Ryle 1971b

RYLE, Gilbert, 1957b, ‘Review of Ludwig Wittgenstein: Remarks on the Foundations of Mathematics’, Scientific American, vol. CXVII (1957), reprinted in Ryle 1971a

RYLE, Gilbert, 1958, ‘A Puzzling Element in the Notion of Thinking’, Proceedings of the British Academy, vol. XLIV, 1958, reprinted in Ryle 1971b

RYLE, Gilbert, 1961a, ‘Use, Usage and Meaning’, Proceedings of the Aristotelian Society, suppl. vol. XXXV, 1961, reprinted in Ryle 1971b

RYLE, Gilbert, 1961b, general discussion of Urmson’s paper ‘The History of Analysis’ at the Royaumont Colloquium of 1961, La Philosophie Analytique (Paris: Editions de Minuit, 1962), translated and reprinted in Rorty 1967

RYLE, Gilbert, 1962a, ‘A Rational Animal’, Originally delivered as the Auguste Comte Memorial Lecture in 1962, reprinted in Ryle 1971b

217

The Philosophical Development of Gilbert Ryle

RYLE, Gilbert, 1962b, ‘Abstractions’, Dialogue, vol. I, 1962, reprinted in Ryle 1971b

RYLE, Gilbert, 1962c, ‘Thinking Thoughts and Having Concepts’, in Logique et Analyse, no. 20, 1962, reprinted in Ryle 1971b

RYLE, Gilbert, 1962d, ‘Phenomenology versus “The Concept of Mind’ (french), La Philosophie Analytique, no. IV (Cahiers de Royamount Philosophie 1962), reprinted in Ryle 1971a

RYLE, Gilbert, 1964, Lecture notes by René Meyer, published in Meyer 1993

RYLE, Gilbert, 1966-67, ‘Thinking and Reflecting’, The Human Agent, Royal Institute of Philosophy Lectures, vol. I, 1966-67, reprinted in Ryle 1971b

RYLE, Gilbert, 1968a, ‘The Thinking of Thoughts. What is “Le Penseur” doing?’, University Lectures University of Sakatchewan, no. 18, 1968, reprinted in Ryle 1971b

RYLE, Gilbert, 1968b, ‘The Genesis of ‘Oxford’ Philosophy’, read to the Oxford Philosophical Society, 500th meeting, in 1968, printed in The Linacre Journal, number 3, november 1999

RYLE, Gilbert, 1970, ‘An Autobiographical Essay’, in Wood and Pitcher 1971

RYLE, Gilbert, 1971a, Collected Papers, Volume I (London: Hutchinson 1971)

RYLE, Gilbert, 1971b, Collected Essays, vol. II (London: Hutchinson 1971)

RYLE, Gilbert, 1972a, ‘Thinking and Self-Teaching’, Rice University Studies, vol. 58, no. 3, summer 1972, reprinted in Kolenda 1979

RYLE, Gilbert, 1972b, ‘Thinking and Saying’, Rice University Studies, vol. 58, no. 3, summer 1972, reprinted in Kolenda 1979

RYLE, Gilbert, 1972c, ‘Intentionality-Theory and the Nature of Thinking’, in Haller, R. (ed.), Jenseits van Sein und Nichtsein (Graz: Akademische Druck und Verlagsanstalt 1972), reprinted in Revue Internationale de Philosophie XXVII, 1973. First presented at a symposium held in Graz on Meinong in 1970.

RYLE, Gilbert, 1972d, ‘Reply to Bouwsma’, in Kolenda 1972

RYLE, Gilbert, 1973, ‘Negative Actions’, Hermathena , no. CXV, summer 1973, reprinted in Kolenda 1979

RYLE, Gilbert, 1974, ‘Mowgli in Babel’, Philosophy, vol. 49, no. 187, January 1974, reprinted in Kolenda 1979

RYLE, Gilbert, 1976a, personal communication, in Bestor 1979

218

References

RYLE, Gilbert, 1976b, ‘Improvisation’, Mind , vol. LXXXV, no. 337, January 1976, reprinted in Kolenda 1979

RYLE, Gilbert, 1976c, ‘Fifty years of philosophy and philosophers’, in Philosophy 51 (1976)

RYLE, Gilbert, 1976d, Letter to Daniel Dennett, 22 February 1976, published by Dennett in 1999 on EJAP: http://ejap.louisiana.edu/EJAP/2002/RyleLett.pdf

RYLE, Gilbert, 1979a, ‘Adverbial Verbs and Verbs of Thinking’, in Kolenda 1979

RYLE, Gilbert, 1979b, ‘Thought and Soliloquy’, in Kolenda 1979

RYLE, Gilbert, 1979c, ‘Thought and Imagination’, in Kolenda 1979

RYLE, Gilbert, 1979d, in Kolenda, Konstantin (ed.), On Thinking (Oxford: Basil Blackwell 1979)

RYLE, Gilbert, 1993, ‘Ontological and Logical Talk in Wittgenstein’s Tractatus’, in Harré and Shosky 1999

RYLE, Gilbert, 2000, ‘Courses of Action or the Uncatchableness of Mental Acts’, prepared for delivery at the annual conference of the Experimental Analysis of Behaviour Group at Bangor in April 1974 (but the paper was in the end read by Professor T. R. Miller because Ryle was not able to attend), first published in Philosophy (2000)

SAVIGNY, Von, E., 1974, Die Philosophie der normalen Sprache, second revised edition (Frankfurt am Main 1974)

SCHLICK, Moritz, 1930/31, ‘The Turning Point in Philosophy’, Erkenntnis, vol. 1, 1930/31, reprinted in Ayer, A. J., Logical Positivism 1959, reprint (Westport: Greenwood Press 1978)

SIBLEY, F. N., 1970, ‘Ryle and Thinking’, in Wood and Pitcher 1971

SLATER, John G., 1994, Bertrand Russell (Bristol: Thoemmes Press 1994)

STANLEY, Jason and WILLIAMSON, Timothy, 2001, ‘Knowing How’, in Journal of Philosophy 98, 2001

STOUT, Rowland, 2003, ‘Ryle’s Behaviourism’, in Revue Internationale de Philosophie 57, 2003, pp. 37-49

STRAWSON, P. F., 1970, ‘Categories’, in Wood and Pitcher 1971

STRAWSON, P. F., 1986 ‘Ryle, Gilbert’, in Dictionary of National Biography, Supplement 1971-1980, Lord Blake and Nicholls, C.S. (eds.) (Oxford 1986)

219

The Philosophical Development of Gilbert Ryle

SZALAI, Jennifer, 2002, ‘On Gilbert Ryle’s Dilemmas’, in New Statesman, 14 January 2002

THOMASSON, Amy, 2002, ‘Phenomenology and the Development of Analytic Philosophy’, in Southern Journal of Philosophy, Vol. XL, supplement, pp. 115-142

URMSON, J. O., 1955, Review Dilemmas by Gilbert Ryle, Mind, vol. 64, 1955, p. 554-556

URMSON, J. O., 1988, ‘Ryle, Gilbert’, in Parkinson, G.H.R. (ed.), An Encyclopaedia of Philosophy (London 1988)

VENDLER, Zeno, 1967, in Philosophy (Ithacan NY: Cornell University Press 1967)

VRIJEN, Charlotte, see also McGuinness

VRIJEN, Charlotte, 2006, ‘Ryle and Collingwood: Their Correspondence and its Philosophical Context’, in British Journal for the History of Philosophy 14 (1) 2006, 93- 131

WAISMANN, Friedrich, 1965, The Principles of Linguistic Philosophy, ed. Rom Harré (London: Macmillan, 1965)

WARNOCK, G. J., 1976, ‘Gilbert Ryle’s Editorship’, in Mind, Volume 85, Issue 337 (Jan. 1976)

WARNOCK, G. J., 1998, ‘Ordinary Language Philosophy, School of’, Craig, Edward (ed.), Routledge Encyclopedia of Philosophy (London and New York: Routledge 1998)

WEITZ, Morris, ‘Professor Ryle’s Logical Behaviourism’, Journal of Philosophy, 48, 1951

WILLIAMS, Bernard, 1961, paper presented at the Royaumont Colloquium of 1961, La Philosophie Analytique (Paris: Editions de Minuit, 1962), reprinted in Rorty 1967

WILLIAMS, Bernard, 1979, ‘Ryle Remembered’, The London Review of Books, vol. 1, 1979, no. 3, 22 November 1979, p. 6

WINCH, Peter G., 1969, ‘Introduction: The Unity of Wittgenstein’s Philosophy’, in Studies in the philosophy of Wittgenstein (London: Routledge 1969)

WISDOM, John, ‘The Concept of Mind’, Proceedings of the Aristotelian Society, 1950

WITTGENSTEIN, Ludwig, 1913, ‘Notes on Logic’, Wittgenstein Nachlass item 201a-1 and 201a-2, and manuscript TSx, kept at the Wittgenstein Archives at the University of Bergen

WITTGENSTEIN, Ludwig, 1922a, Tractatus Logico-Philosophicus (London: Routledge & Kegan Paul 1922, 1951 and 1969)

WITTGENSTEIN, Ludwig, 1922b, Letters to Ogden, published in Von Wright, G. H. (ed.), Letters to C. K. Ogden (Oxford: Blackwell 1973) 220

References

WITTGENSTEIN, Ludwig, 1939, Diamond, Cora (ed.), Lectures on the Foundations of Mathematics (Hassocks: Harvester Press 1976)

WITTGENSTEIN, Ludwig, 1953, Philosophical Investigations (Oxford: Basil Blackwell 1953), reprinted (Oxford: Blackwell 1997)

WITTGENSTEIN, Ludwig, 1958, The Blue and Brown Books (New York: Harper & Row Publishers 1958), reprinted (New York: Harper & Row Publishers 1965)

WITTGENSTEIN, Ludwig, 1961, Von Wright, G. H. and Anscombe, G. E. M. (eds.), Notebooks, 1914-1916 (Oxford: Blackwell 1961)

WITTGENSTEIN, Ludwig, 1980, Von Wright, G. H. and Anscombe, G. E. M. (eds.), Remarks on the Philosophy of Psychology (MS 135), Volume 1 (Oxford: Basil Blackwell 1980)

WOOD, O. P. and PITCHER, G., 1971, G., Ryle. A Collection of Critical Essays (London: MacMillan 1971)

WRIGHT, Von, G. H. (ed.), 1973, Letters to C. K. Ogden: with Comments on the English Translation of the Tractatus Logico-Philosophicus (Oxford: Blackwell and London: Routledge and Kegan Paul Ltd 1973)

WRIGHT, Von, G. H., 1982, Wittgenstein (Oxford: Basil Blackwell 1982)

ZILHÃO, António, 1993, ‘Ludwig Wittgenstein and Edmund Husserl’, Georg Meggle and Ulla Wessels (eds.), in Analyomen 1, Proceedings of the 1st Conference “Perspectives in Analytical Philosophy”, pp. 956-964

221