Language in Society 2016.Pdf

Society in Language, Language in Society Ruqaiya Hasan (17 February 2015) Photo by Ernest Akerejola Society in Language, Language in Society Essays in Honour of Ruqaiya Hasan

Edited by

Wendy L. Bowcher Sun Yat-sen University, China Jennifer Yameng Liang University of Science and Technology Beijing, China Selection and editorial content © Wendy L. Bowcher and Jennifer Yameng Liang 2016 Individual chapters © Respective authors 2016 Softcover reprint of the hardcover 1st edition 2016 978-1-137-40285-1

All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission. No portion of this publication may be reproduced, copied or transmitted save with written permission or in accordance with the provisions of the Copyright, Designs and Patents Act 1988, or under the terms of any licence permitting limited copying issued by the Copyright Licensing Agency, Saffron House, 6–10 Kirby Street, London EC1N 8TS. Any person who does any unauthorized act in relation to this publication may be liable to criminal prosecution and civil claims for damages. The authors have asserted their rights to be identified as the authors of this work in accordance with the Copyright, Designs and Patents Act 1988. First published 2016 by PALGRAVE MACMILLAN Palgrave Macmillan in the UK is an imprint of Macmillan Publishers Limited, registered in England, company number 785998, of Houndmills, Basingstoke, Hampshire RG21 6XS. Palgrave Macmillan in the US is a division of St Martin’s Press LLC, 175 Fifth Avenue, New York, NY 10010. Palgrave Macmillan is the global academic imprint of the above companies and has companies and representatives throughout the world. Palgrave® and Macmillan® are registered trademarks in the United States, the United Kingdom, Europe and other countries. ISBN 978-1-349-57286-1 ISBN 978-1-137-40286-8 (eBook) DOI 10.1057/9781137402868 This book is printed on paper suitable for recycling and made from fully managed and sustained forest sources. Logging, pulping and manufacturing processes are expected to conform to the environmental regulations of the country of origin. A catalogue record for this book is available from the British Library. Library of Congress Cataloging-in-Publication Data Society in language, language in society : essays in honour of Ruqaiya Hasan / Wendy L. Bowcher, Sun Yat-sen University, China ; Jennifer Yameng Liang, University of Science and Technology Beijing, China) pages cm Summary: “This is the first collection dedicated to research directly influenced by the innovative and groundbreaking ideas of the eminent linguist Ruqaiya Hasan. The collection offers an insight into the breadth and depth of Hasan’s distinctive linguistic approaches and theoretical concerns. The chapters cover areas such as verbal art, context of situation, semantic networks, cohesive harmony, text structure and literacy education. The volume contains an interview with Ruqaiya Hasan, and a section in which the contributors describe their connection with Ruqaiya Hasan and her work. This book is of particular value to scholars and students working in sociolinguistics, literary criticism, stylistics, functional linguistic theories, literacy pedagogy, social semiotics, multimodality and applied linguistics” — Provided by publisher. 1. Semiotics—Social aspects. 2. Communication—Social aspects. 3. Language and languages—Variation. 4. Language and culture. 5. Language and education. 6. Sociolinguistcs. I. Bowcher, Wendy L., editor. II. Liang, Jennifer Yameng, 1987– editor. III. Hasan, Ruqaiya, honoree. P325.5.S63S63 2015 401.4—dc23 2015012349 Contents

List of Figures vii List of Tables ix Preface x Acknowledgements xii Notes on the Contributors xiii

Part I Hasan’s Linguistics

1 The Ontogenesis of Rationality: Nigel Revisited 3 M. A. K. Halliday

2 ‘Construe My Meaning’: Performance, Poetry and Semiotic Distance 24 David G. Butt

Part II Verbal Art

3 Jakobson’s Place in Hasan’s Social Semiotic Stylistics: ‘Pervasive Parallelism’ as Symbolic Articulation of Theme 59 Donna R. Miller

Part III Semantic Networks

4 Can Semantic Networks Capture Intra- and Inter-Registerial Variation? Palliative Care Discourse Interrogates Hasan’s Message Semantics 83 Alison Rotha Moore

5 Hasan’s Semantic Networks Revisited: a Cantonese Systemic Functional Approach 115 Andy Fung

Part IV Context of Situation

6 Language and Society, Context and Text: the Contributions of Ruqaiya Hasan 143 Annabelle Lukin

v vi Contents

7 Multiscalar Modelling of Context: Some Questions Raised by the Category of Mode 166 Tom Bartlett 8 On Describing Contexts of Situation 184 Margaret Berry 9 Interfacing Field with Tenor: Hasan’s Notion of Personal Distance 206 Marvin Lam 10 Studying Language in Society and Society through Language: Context and Multimodal Communication 227 Rebekah Wegener

Part V Structure and Texture: Two Kinds of Unity 11 GSP and Multimodal Texts 251 Wendy L. Bowcher and Jennifer Yameng Liang 12 Construing Instructional Contexts 275 Carmel Cloran 13 ‘Threads of Continuity’ and Interaction: Coherence, Texture and Cohesive Harmony 300 Kristin M. Khoo

Part VI Literacy and Education 14 Reﬂection Literacy in the First Years of Schooling: Questions of Theory and Practice 333 Geoff Williams 15 Reﬂection Literacy and the Teaching of History 357 Mariana Achugar and Mary Schleppegrell

Part VII In Her Own Words 16 In Her Own Words: an Interview with Ruqaiya Hasan 381 David G. Butt and Jennifer Yameng Liang

Part VIII In Honour of Ruqaiya Hasan 17 In Honour of Ruqaiya Hasan: Notes from the Contributors 415

A Bibliography of Work by Ruqaiya Hasan 433 Author Index 439 Subject Index 444 List of Figures

2.1 Verbal art and language (Hasan, 1985, based on Hasan, 1979) 27 2.2 A schematic view of the logogenetic accumulation of tokens in Act 2 Scene 2 of Troilus and Cressida 33 3.1 The overlapping semiotic systems (based on Hasan, 1985, p. 99) 64 4.1 Hasan’s ‘Demand Information’ network (based on Hasan, 2013, p. 289) 95 5.1 Primary options of message semantics in Cantonese 126 5.2 Systemic options of RELATION ENACTMENT in Cantonese 127 5.3 A tentative semantic network of asking questions in Cantonese 134 6.1 Culture, meaning and situation (Hasan, 1985b) 148 6.2 A ‘rudimentary’ network for field (Hasan, 1999a) 150 6.3 Realization and Instantiation in Halliday’s linguistic model (Hasan, 2009c, p. 12, based on Halliday, 1999) 153 6.4 Class, codes and communication: an interpretation (Hasan, 2005c) 155 6.5 Hasan’s typology of sign types (Hasan, 2014, Figure 6.1, p.112) 158 8.1 Language event: Mode systems 191 8.2 Language event: Tenor systems 193 8.3 Language event: Field systems MK1 201 8.4 Language event: Field systems MK2 202 9.1 Sophie Neveu and Robert Langdon’s biography of interactions in Dan Brown’s The Da Vinci Code (Lam, 2010) 220 9.2 Accumulative frequencies of [ancillary] and [practical] vs [constitutive] (top) and [relation based] vs [reflection based] (bottom) (Lam, 2010) 223 10.1 Stratification: reproduced from Halliday and Matthiessen (2004, p. 25) 237 10.2 Stratification and the contextual plane representing the union of the social and biological 238 11.1 The GSP of the nursery tale (Hasan, 1996, p. 54) 253 11.2 GSP of multimodal print advertisements (Cheong, 2004, p. 164) 263 11.3 A tourist site entry ticket to Hua Shan (with ‘Enhancer’ circled) 265 11.4 A tourist site entry ticket with the Attributor circled 267

vii viii List of Figures

11.5 A tentative GSP of the register of tourist site entry tickets (TSET) 270 12.1 Relation of RU classes to a material base 276 12.2 GSP of the official instructional context 277 12.3 Function structure and RU structure of Extract 1 278 12.4 Function structure and RU structure of Extract 2 280 12.5 RU structure of Clarification element exemplified in Extract 3 282 12.6 Function and RU structure of Extract 5 284 12.7 RU analysis of Extract 6 285 12.8 GSP of local instructional context 289 12.9 RU structure of Extract 9 290 12.10 RU structure of Extract 10 291 12.11 RU analysis of Extract 11 292 12.12 RU analysis of messages 1–14 of Extract 12(a) 295 12.13 RU structure of Extract 12(b) 297 12.14 The structural element Practicum 297 12.15 Extract 12(a) amended 298 13.1 Place of cohesion and the textual metafunction rank matrix of SFL (reproduced from Halliday, 2009, p. 85) 303 13.2 Summary of cohesive devices (reproduced from Hasan, 1985, p. 82) 305 14.1 A child’s initial recording of some examples of language in use 344 14.2 Collaborative recording of oral language in use at school in a kindergarten class 345 14.3 The children’s example of dysfunctional wording in a command 347 14.4 Screenshot of ‘Mr Very Confused’s Recipe for Fairy Bread’ 349 14.5 An example of a child’s playful text to display his ability to identify Events 351 15.1 Support for deconstructing the Declaration of Independence passage 366 List of Tables

2.1 Logical analysis of Sonnet 30 44 2.2 Logical analysis of Sonnet 64 45 2.3 Logical analysis of Sonnet 60 46 2.4 Logical analysis of Sonnet 29 47 2.5 Logical analysis of Sonnet 65 48 3.1 Process types ascribed to Archie (from handout, Hasan, 2010) 75 3.2 Process types ascribed to death/dying (from handout, Hasan, 2010) 75 3.3 The close associations of time and death (from handout, Hasan, 2010) 76 6.1 Material and textual relations within one interaction: complex vs parallel simple texts (Hasan, 1999a, p. 267) 152 6.2 Properties of the sign types in Figure 6.2 (Hasan, 2014) 159 8.1 Categories of activities in the language events 199 9.1 Notions on distance with reference to the ordered typology of systems (adapted from Lam, 2010, p. 59) 214 10.1 The dimensions (forms of order) in language and their ordering principle from Halliday and Matthiessen (2004, p. 20) 235 10.2 Types of realization statements from Hasan (1996, p. 111) 240 13.1 Illustrative survey of work applying CHA 307 14.1 Transcription of written language in the child’s notes in Figure 14.1 344 14.2 Example of differences in wordings of commands in a recipe (transcribed from a smartboard screen) 349

ix Preface

It has been a pleasure and an honour to prepare this manuscript in recognition of one of the world’s leading linguists, Ruqaiya Hasan. From its inception, the aim of this book has been to offer a selection of work by close associates, previous students, and other scholars who have been taught or inspired by Ruqaiya, and who have actively engaged with her work and her ideas. Many years ago, just after finishing my BA (Hons) at Macquarie Univer- sity under Ruqaiya’s supervision, I (Wendy) was talking with a lecturer at Sydney University. When I mentioned that Ruqaiya had been my supervisor, that lecturer raised an eyebrow and said, ‘Really. That’s a scholar with a fertile mind.’ Indeed, Ruqaiya is such a scholar. Her incisive mind and deep thinking, along with her insistence on analytical integrity, consistency and academic rigour, have produced some of the most thought-provoking and ground-breaking insights into the relationship between language and society. Due to the complex subjects she deals with and her insistence on probing beyond the surface of an issue, much of her work still awaits development. However, the rich array of issues that she has raised, questions that she has asked, and analytical methods that she has proposed, are such that scholars seeking a satisfying and rewarding research path would do well to consider her work. The volume includes some of the key areas of research to which Ruqaiya Hasan has contributed. These include semantic networks, cohesive harmony, verbal art, literacy, and the concept of context. We are aware, however, that there are significant areas that are virtually untouched or only partially dealt with in the volume, including her engagement with the fields of pragmatics, sociolinguistics and sociology, and with issues to do with child language development. Some of these are mentioned in Chapter 16, which is an interview with Ruqaiya Hasan conducted by David Butt and Jennifer Yameng Liang. This chapter represents a personal and rather lively account of Ruqaiya’s views on a number of concerns within the field of linguistics. Nevertheless, although the collection can only provide a glimpse into some of Ruqaiya’s ideas, overall the book does give readers an indication of the scope of her insights and, we are sure, offers inspiration to many scholars and researchers, and will do for many years to come. We are grateful to all the scholars who have contributed to this collection. We have included in the volume a chapter entitled ‘In Honour of Ruqaiya Hasan: Notes from the Contributors’ in which each author has written a short description of their background or association with, and/or inspiration from, Ruqaiya.

x Preface xi

In particular, we wish to thank Ruqaiya for being dedicated, tenacious and ever willing to take the time to discuss her ideas, work through difﬁcult concepts, and foster academic integrity and rigour in those with whom she engages.

Wendy L. Bowcher Jennifer Yameng Liang December 2014

Addendum:

During the time we were proofing this book we heard the sad news that Ruqaiya Hasan had passed away. We know that we speak on behalf of all the contributors to this volume in expressing our deep sorrow that Ruqaiya was not able to see this book in print – a book that was to be presented to her to honour her intellectual generosity and mentorship. In announcing the sad news to the scholarly community, Geoff Williams made the following comment: ‘A wonderful life, an immense scholarly contribution, an extraordinary friend to so many people around the world.’ Ruqaiya’s contribution to the scientific field of linguistics is yet to be fully realized. While her written works attest to her extraordinary depth of insight, her incisive mind and critical thinking, the innumerable times that she spent critiquing, encouraging and otherwise lending her academic insights to individual researchers around the world have left an indelible mark that has enriched the academic community and which will continue to bear fruit. We will miss her dearly.

Wendy L. Bowcher Jennifer Yameng Liang June 2015 Acknowledgements

The editors would like to thank Zhang Zhenzhen, Chen Yuping and Chen Xia for their assistance during the preparation of the typescript. We would also like to thank Philip Tye for his careful copy-editing. Figures 11.3 and 11.4 have been reproduced by kind permission of Huashan Scenic Management Committee and the ofﬁce of the Ming Tombs Special Administrative District respectively. Thanks to John Whitworth for permission to use the poem ‘Little’.

xii Notes on the Contributors

Mariana Achugar is a Guggenheim Fellow and an Associate Professor of Hispanic Studies and SLA at Carnegie Mellon University in Pittsburgh, USA. Her research explores the role of discourse in processes of cultural reproduction and change, and also disciplinary literacy development in multilingual contexts. Her most recent work has appeared in Linguistics and Education, Discourse & Society, Critical Discourse Studies, and Text and Talk among other venues.

Tom Bartlett is Senior Lecturer in the Centre for Language and Communi- cation Research, Cardiff University, UK. His research interests are systemic functional linguistics (SFL), language and context, and discourse and par- ticipatory democracy. His recent publications include Systemic Functional Linguistics: Exploring Choice (co-edited), Choice in Language: Applications in Text Analysis (co-edited) and Hybrid Voices and Collaborative Change: Contex- tualising Positive Discourse Analysis.

Margaret Berry, now retired, was Reader in English Language at the Univer- sity of Nottingham, UK. She has published widely within the ﬁeld of SFL and wrote the classic ﬁrst introductory books to SFL: An Introduction to Systemic Linguistics (Vol. 1): Structures and Systems (1975) and An Introduction to Sys- temic Linguistics (Vol. 2): Levels and Links (1977). She has published articles on theoretical issues in SFL, exchange structure, theme and rheme, register variation, and the application of SFL to the teaching of English. She has lectured in China, Australia and Canada, as well as in Europe.

Wendy L. Bowcher is a Professor in the School of Foreign Languages at Sun Yat-sen University, China. Her research interests include multimodal discourse analysis of Japanese and English texts, context in SFL theory, language education, forensic linguistics, and English intonation. She is the editor of Multimodal Texts from around the World: Cultural and Linguistic Insights (2012), co-editor (with Terry D. Royce) of New Directions in the Analysis of Multimodal Discourse (2007) and co-editor (with Brad Smith) of Systemic Phonology: Recent Studies in English (2014). Wendy was instrumental in the formation of the Japan Association of Systemic Functional Linguistics (JASFL) in 1993 and served for several years as the Vice President of the Association. From 2011 to 2014 she was the Vice-Chair of the International Systemic Functional Lin- guistics Association (ISFLA) and served as Programme Chair of the ISFC held at Sun Yat-sen University in 2013.

xiii xiv Notes on the Contributors

David G. Butt is Associate Professor of Linguistics at Macquarie Uni- versity in Sydney, and was recently Director of the University Research Centre for Language in Social Life. Over the last decade and more, this Centre has conducted projects across communities and institutions for which functional linguistics provided signiﬁcant evidence about the management of change. Through the Centre he has been actively engaged with professionals in medicine (surgery and psychiatry), counselling, care for people with disabilities, intelligent systems design and brain sciences, cultural analysis (literature, theatre, world Englishes), complexity theory and ‘smart spaces’, Vygotskian approaches to education and training, ﬁnancial reporting, courtroom explanations and forensic evidence, media and jour- nalism, and child language development (in the traditions of Halliday). The Centre has also investigated the interrelations between linguistics, verbal art (especially poetry), philosophy and the arguments of natural sciences (especially biology, genetics and physics). The Centre has participated in educational projects in various cultures beyond Australia – Singapore, India, and especially with Australia’s close neighbours in Timor and in Indonesia.

Carmel Cloran (now retired) worked in educational settings before completing a BA and a PhD at Macquarie University, Sydney, under the watchful eye of Professor Ruqaiya Hasan. With Hasan, she explored (a) the parameters of context, and (b) the meaning-making practices of mothers interacting with their preschool children. She enjoyed a career teaching and researching linguistics at Macquarie, the National University of Singapore, the Regional Language Centre, Singapore, and the University of Wollongong, Australia. She has published a number of papers, the most recent being a comparison of the concept of rhetorical unit (proposed in her doctoral dissertation) with Bakhtin’s concept of chronotope (Functions of Language (2010) 17 (1): 29–70). She co-edited with David Butt and Geoff Williams, a collection of papers by Ruqaiya Hasan entitled Ways of Saying: Ways of Meaning (1996).

Andy Fung is a PhD candidate and a member of the health-care communication research team in the Department of English at the Hong Kong Polytechnic University. As inspired by Geoff Williams, he is interested in extending Ruqaiya Hasan’s semantic networks from the perspective of systemic functional Cantonese grammar, and examines the semantic and lexicogrammatical features of questions and answers in emergency communication. He also has an ongoing interest in multimodal discourse analysis and mobile apps communication as a tool in educational and tourism settings.

Michael A. K. Halliday was born in Leeds, England, in 1925. He studied classics (reluctantly) in secondary school, then trained in Chinese at the School of Oriental and African Studies for service with the British army in Notes on the Contributors xv the Second World War. He studied for three years in China (Peking Univer- sity, then Lingnan in Guangzhou), returning to England to teach Chinese at Cambridge University. Most of his working life was devoted to teaching and research in linguistics; he saw himself as a generalist, with special interest in grammar and phonology, and was committed to developing an ‘appliable’ linguistics which would be useful both to linguists and to others whose professional work required engagement with language. His major interests have been in language in education, the language of infancy and early childhood, the language of science, quantitative studies of language and the semantic functions of intonation and rhythm. He retired as Emeritus Professor of Lin- guistics, University of Sydney, at the end of 1987. His collected works, edited by Jonathan J. Webster, were published in ten volumes from 2002 to 2007, and an eleventh volume, Halliday in the 21st Century, was published in 2013.

Kristin M. Khoo is a PhD candidate in Linguistics with an Australian Post- graduate Award scholarship at Macquarie University, Sydney. She began her studies in linguistics at Macquarie in 2001, in combination with studies in psychology. Her doctoral research is focused on cohesion and cohesive harmony from an SFL linguistic perspective in the discourse of psychotherapy. Kristin’s research explores the role of linguistic evidence in clinical concepts of cohesion and coherence associated with self – concepts prominent in the Conversational Model of psychotherapy, as well as more broadly in discussions of brain, mind, consciousness, memory and relationship in mental health, neuroscience and related ﬁelds. Kristin has experience teaching in undergraduate linguistics and psychology courses since 2005, at Macquarie University, Sydney Institute of Business and Technology (Macquarie Univer- sity campus) and Open Universities Australia (Macquarie University). She also has experience in linguistic research through previous research assistance at the Centre for Language in Social Life at Macquarie University and participates in research collaboration through the Language in Social Life Psychotherapy Research network.

Marvin Lam is a Research Assistant Professor in the Department of English at the Hong Kong Polytechnic University. He is also an active member of the PolySystemic Research Group, an SFL working group with both an international and a regional focus in the Pearl River Delta, China, and Southeast Asia. His research focuses on SFL and applied linguistics in general, the modelling of interpersonal distance, interpersonal communication, text linguistics and discourse studies, language typology, language teaching and language across the curriculum. His work on the modelling of interpersonal distance and its objective measurement with the concept of ‘socio-semiotic distance’ has proved its signiﬁcance in yielding interpersonal information from linguistic analysis, and has been applied in various institutional contexts, such as the interpersonal relationship between patients xvi Notes on the Contributors and health-care practitioners in hospital emergency departments, therapists and clients in psychotherapy treatments, customers and customer service representatives in telephone call centre exchanges and between characters construed in novels and motion pictures. Together with Christian M. I. M. Matthiessen and Kazuhiro Teruya, Lam co-authored Key Terms in Systemic Functional Linguistics (2010).

Jennifer Yameng Liang is a lecturer in the School of Foreign Studies at the University of Science and Technology Beijing. She received her PhD degree from the School of Foreign Languages, Sun Yat-sen University. Her research interests centre on both theoretical explorations and practical applications of SFL, in particular the concepts of context and ecosocial environment and multimodal/multisemiotic studies. In her doctoral research, she conducted research on the dynamic meaning-making practices and processes of Chinese tourist site entry tickets. She has co-authored with Wendy Bowcher and published in Visual Communication, Semiotica and Social Semiotics.

Annabelle Lukin is a linguist in the Centre for Language in Social Life, at Macquarie University, Australia. She has drawn on many aspects of Ruqaiya Hasan’s work in SFL in her studies of literature, education, ideology in media discourse, translation, and discourse analysis. She curates the SFL linguists’ site on VIMEO, and is a contributor to Wikipedia entries on linguistics, and in particular, on topics and people relating to SFL.

Donna R. Miller holds the Chair of English Linguistics at the Depart- ment of Modern Languages, Literatures and Cultures of the University of Bologna, where she coordinates its English Language Studies Programme and heads the Department’s Centre for Linguistic-Cultural Studies (CeSLiC). Her research has largely focused, in an SFL perspective, on register analysis, particularly in institutional text types, her corpus-assisted investigations having speciﬁcally explored the grammar of evaluation in terms of appraisal systems. Currently her interests also extend to issues of World Englishes. No stranger to Ruqaiya Hasan’s framework for the study of verbal art, of late Miller has energetically taken up its defence, but has also been reﬂecting intensely on Jakobson’s potential place within it. See for example her 2010 essay, ‘The Hasanian framework for the study of “verbal art” revisited ...and reproposed’(Textus XXIII (1)), the 2012 ‘Slotting Jakobson into the social semiotic approach to “verbal art”: a modest proposal’ (in F. Dalziel et al. (eds) A Lifetime of English Studies: Essays in Honour of Carol Taylor Torsello, and 2013 ‘Another look at social semiotic stylistics: coupling Hasan’s “verbal art” framework with “the Mukaˇrovský–Jakobson theory” ’ (in C. Gouveia and M. Alexandre (eds), Languages, Metalanguages, Modalities, Cultures: Functional and Socio-Discursive Perspectives). Notes on the Contributors xvii

Alison Rotha Moore is a linguist with interests in functional linguistics, professional and institutional discourses, agency, identity and embodied action. Alison’s research mostly involves health discourse, but she has recently also begun working on discourses around animal welfare and her latest paper is ‘That could be me: identity and identiﬁcation in discourses about food, meat, and animal welfare’ (2014) in Linguistics and the Human Sci- ences. Alison has a PhD in Linguistics from Macquarie University and a Research Masters in Public Health from the University of Sydney. She con- venes the English Language and Linguistics Programme at the University of Wollongong.

Mary Schleppegrell is Professor of Education at the University of Michigan, Ann Arbor. Her research studies the role of language in learning with particular attention to the needs of English language learners. She draws on SFL to identify the demands of teaching and learning in different subject areas. She is the author of The Language of Schooling: a Functional Linguistics Perspec- tive, and co-author (with Zhihui Fang) of Reading in Secondary Content Areas: a Language-Based Approach. Her research has recently appeared in Language Learning, Linguistics and Education, Journal of English for Academic Purposes, Annual Review of Applied Linguistics, Journal of Adolescent & Adult Literacy and Australian Journal of Language and Literacy.

Rebekah Wegener is a linguistics and semiotics researcher at the Institut für Anglistik, Amerikanistik und Romanistik at RWTH Aachen University, Germany. She received her PhD from Macquarie University, Australia. Her research interests include theoretical and applied linguistics and communication in complex systems. She works on aspects of medical communication, modelling context for multimodal environments and behavioural interfaces for artiﬁcial intelligence. She is co-founder of Audaxi Pty Ltd, a company specializing in human and technical systems for higher education.

Geoff Williams completed his PhD in linguistics under Ruqaiya Hasan’s supervision at Macquarie University in 1995, studying the nature of semantic variation associated with speakers’ social positioning during joint book- reading in early childhood. He regards his research experience under Hasan’s supervision as changing the nature of his academic work. Geoff’s research also includes the ﬁrst systematic studies of children’s learning of SFL-derived grammatics, in collaboration with Joan Rothery and Ruth French. Geoff researched and taught at the University of Sydney for 30 years, initially in the Faculty of Education and later in the Department of English, Fac- ulty of Arts, where he introduced a range of new courses concerned with children’s language and literacy development. In 2005 he became Head of xviii Notes on the Contributors the Department of Language and Literacy Education, University of British Columbia, from which he retired as Emeritus Professor in 2010. He has held the position of Chair of both the International and the Australian Systemic Functional Linguistic Associations. He has contributed numerous chapters to SFL volumes and, with Ruqaiya Hasan, edited Literacy in Society (1996), and with Annabelle Lukin, The Development of Language: Functional Perspectives on Species and Individuals (2004). Part I Hasan’s Linguistics 1 The Ontogenesis of Rationality: Nigel Revisited M. A. K. Halliday

1.1 Introductory

It is exactly 50 years since Ruqaiya Hasan received her PhD degree from the University of Edinburgh, with a thesis in which she showed how much could be revealed by detailed lexicogrammatical analysis about a full-length prose novel in which every sentence was written in ordinary everyday language. This was at a time when few Anglophone linguists were paying any attention to literature; and among those who were, the prevailing view was that ‘literary language’ was marked out by deviation from the norms of everyday discourse. Since that time, a number of key motifs can be seen coursing through Ruqaiya’s work: the centrality of everyday language, and dialogic interaction, in life and especially in learning; registers of discourse, and their variation in structure and texture; the special nature of verbal art; rationality and inference; the networking of semantics and context. All these, in turn, can be thought of as aspects of one underlying pursuit: the study of meaning, and of reasoning, in the contexts of language use. Early in her career she worked alongside Basil Bernstein, Professor in the Sociology of Education at the University of London; her job was to analyse the stories made up by very young children as elicited in the course of his experiments into the role of language in transmitting the semiotic foundations of the culture. Later on, in her own groundbreaking programme of research at Macquarie University in the 1980s, she pursued this issue in depth, as revealed in the natural spontaneous exchanges between 3½-year- olds and their mothers. Papers such as ‘Questions as a mode of learning in everyday talk’ (1991), ‘Language in the processes of socialization: home and school’ (1988) and ‘The ontogenesis of ideology’ (1986) show how, from the ordinary exchanges going on inside the home, children build up a complex picture of the assumptions, and the values, that are naturalized in the world in which they live.

3 4 Society in Language, Language in Society

In this chapter I revisit some fragments of my original Nigel data from the early 1970s [see ‘Listening to Nigel’, included as a compact disc in vol- ume4ofmyCollected Works],1 picking up on a topic I had hoped to write about many years ago but never got around to: Nigel’s steps along the road to what the adult world he was growing up in would regard as coherent and acceptable reasoning. This too was a topic that Ruqaiya had explored in her own data: see ‘Reading picture reading’ (Hasan, 1987a) and ‘Rational- ity in everyday talk: from process to system’ (Hasan, 1992). I shall cite one or two passages from Nigel’s discourse which show how everyday linguistic interaction provides the means whereby the child learns to reason, about the eco-social environment and his own complex of relationships within it. These are part of the same story, of taking on and engaging with the human condition. The one really valuable study of Nigel’s earliest reasoning was that carried out by Joy Phillips, first as an MA course assignment, then developed and expanded into an MA honours thesis [‘The development of comparisons and contrasts in young children’s language’, University of Sydney, 1985]. Phillips’ principal source was the Nigel data; she also wrote up a study of Nigel’s development of modality and hypothetical meaning [Working Papers in Linguistics 3, 1986]. In her thesis, Phillips worked with four semantic categories, two of comparison (sameness, similarity) and two of contrast (difference, opposition), pointing out that each is typically commented on only where ‘there is an element of’ the other: things are compared only in the context of difference, contrasted only in the context of similarity. For example (1; 10), mummy hair like railway line, that tree got leaf on but that tree got no leaf on, that very hot (pan) ...that very hot (‘is the handle hot?’).2 The first is a straightforward comparison, made because mummy’s hair is not, in fact, a railway line; the second and third show internal contrast, one declarative with a switch of polarity and the conjunction but, the other with a switch of mood from declarative to interrogative. (The ‘interrogative’ is realized by intonation. Nigel already was using the inverted ‘form’ of interrogative at that time, but not to signal a question; the meaning was informative, ‘I’m telling you something you wouldn’t otherwise know’.) Comparing and contrasting are basic strategies in reasoning, and all such reasoning depends on being able to generalize: to ‘refer’ using terms that are ‘common’ (class names) as well as those which are ‘proper’ (individual names). Construing classes is in fact a condition of entry into the mother tongue; and children seem to take this step at about the same time as they take their first steps in walking, which is also when they can see objects from every angle under their own control. Nigel was soon naming interesting things that he saw when he was taken for walks in the neighbourhood; he would enumerate them on his return, such as (1; 6; 1) ‘ducks, sticks, holes and buses’ (gaaugaau tikutiku loulou baba). Repeating the name with full tonic prominence (both times with falling tone) was probably his way of saying M. A. K. Halliday 5

‘more than one’. Phillips notes the use of ‘another’ from 1; 6, and ‘more’ and ‘two’ from 1; 7; effectively at this time Nigel had a number system of ‘one/ two/a lot’ (or singular/dual/plural). By 1; 11 he could use the number ‘two’ in comparing quantity: (1; 11; 6) I can see more light than two. But as soon as you start naming classes, you face problems of classification. Nigel was already struggling with these at the age of 1; 7, and used the dual ‘two’ as a frame for sorting them out. There was no problem with two books – he could hold one in each hand, and look from one to the other; and likewise with two apples, two nuts. But sometimes there was no name for the common class, or if there was he did not know it. Sometimes he simply used one of the names to cover both, like two brush for a brush and a comb. A toy engine and a toy hammer he construed as two hammer, though with some sign of uncertainty; but contemplating an engine in one hand and a bus in the other, he first tried two engines, was clearly not convinced, and finally gave up on it altogether: two ...two chuffa ...two ...two .... These were all in the last ten days of 1; 7. He had now been listening to talk for his life of one year and a half, and talking himself for much of the past nine months, though not yet in the adult mode. But he now engaged firmly with the mother tongue, introducing in the first half of 1; 8 the critical functional distinction into mathetic and pragmatic which is a common feature of children’s transition from protolanguage to language. Pragmatic utterances were spoken on a rising tone, and they demanded a response, at first in the form of goods and services (‘give to me’, ‘do for me’) but increasingly, towards the end of his second year, in the form of information (‘tell me’). Mathetic utterances were spoken on a falling tone and demanded no response; they functioned as the construal of experience, ‘that’s the way things are’. Utterances which named or compared classes were of course mathetic. This did not mean that they were not addressed to anyone; they usually were, because the interlocutor was the one who was sharing or had shared the experience, and could therefore confirm or deny the proposition if he or she chose to do so. Nigel continued naming things, especially less familiar things, over the next few months; but such utterances became less and less prominent as his resources for talking about them expanded. This happened very rapidly; by 1; 9 he was saying things like Anna make noise grass (‘Anna was making a noise with (a blade of) grass’), letter fall out mummy book, and (a pragmatic one) dada get knife take skin off apple. When he did assign something to a class, it was likely to be as part of some more complex observation; for example, at 1; 11; 4, That a high chair. But that is not a high chair; that a low chair (with contrastive tonic prominence on low). Here one class is being recognized to include two subclasses. He does not distinguish, of course, between high in high chair, which is a Classifier (there is a class of high chairs) and low in low chair, which is an Epithet (there is no class of low chairs); both fall within the class chair. But by this time, new class names are incorporated into Nigel’s discourse without being brought into attention. 6 Society in Language, Language in Society

1.2 Working things out

I had often been struck by the contrast between what experiments had shown that children could and could not say, or understand, when tested at particular ages, and what Nigel actually did say, and understand, in the course of his everyday life. His mother had been conversing with him from a very early age, while Nigel listened attentively. I was simply out of a job and so had the chance to make detailed notes of what I was able to observe. A child – like all human beings – if put into an unnatural situation will tend to behave in an unnatural way; he may fail to produce a passive, or to understand a relative clause, when his performance is being monitored and assessed, long after he has been using these patterns with ﬂuency in his own unselfconscious conversation. I ought to make it clear, perhaps, that at no time in Nigel’s early life did either I or his mother try to elicit from him any reaction or response. Here is a little snatch of Nigel’s conversation from age 2; 4:3

The lady had to go out of the car to pick the dog ùp, because she thought the dog was lòst; but she wàsn’t.

The reasoning is fully explicit. But Nigel has been reasoning already for quite a long time; only, you could not hear it. The meaning has not been explicitly worded. Nevertheless you may be able to follow the semantic trail. At 1; 11, Nigel heard his mother and me planning a visit to the aquarium. We were talking to each other, not to him; he was doing something at the other side of the room; but he must have been paying attention. Then I heard him say thoughtfully to himself, ‘we not going to see a rao ...vopa ...there will be some water’. He had gathered that we were going somewhere like a zoo; he knew that a zoo is where you go to see lions – but instead of a lion we were going to look at fishes. Fishes live in water; so where we were going there must be water for them to live in. I knew that rao was ‘lion’ and vopa was ‘fish’; I also knew how these words had developed. I knew much of Nigel’s prior experience; I knew his mastery of language at the time (it was unlikely he had heard the word aquarium,orif he had he would have filtered it out as unknowable); and I knew the immediate context of what he was saying – not being addressed, but in earshot. So I was able to follow his reasoning and notice what inferences he was making. He was not talking to either of us; but he was thinking something out for himself. Here is another example from around the same age (1; 11). We were in a train, on our way to visit some friends for the day. Nigel was looking out of the window and talking about what he saw. ‘This not an underground train’, he said. ‘Ooh, there’s a bi-i-ig crane! There’s another railway line there; we’re not going on that railway line. It not lion’, he corrected himself with a growl. M. A. K. Halliday 7

‘It line ...Fast weel train. Our train going now. But it not say whoosh ...but it not say whoosh ...whoosh ...whoosh’, he repeated several times over. ‘This train not say whoosh’, very quietly to himself. His use of weel train for ‘diesel train’, and the word play with line and lion, were both familiar, having occurred many times in Nigel’s discourse. But there were other features that stood out as noteworthy here. When Nigel says ‘there’s another railway line there; we’re not going on that railway line’, he is rehearsing something he had worked out for himself some time before: that you cannot actually see the line that your train is running on, so if you see a railway line from inside the train it must be a different one. So far so good. But the repeated ‘whoosh’ (which he pronounced whoo’, without the final sibilant) was addressing a separate – though in fact related – problem. About a week earlier, Nigel had called me over to where he was looking at one of his books. ‘Daddy come and look at that train has gone’, he said. ‘It already gone whoosh’. It was a picture of a train which had just sped through a station without stopping; ‘whoosh’ was the noise it made if you were standing on the platform watching it go by – something with which he was by now thoroughly familiar. ‘That train already gone whoosh’, he explained. ‘That train has already gone whoosh.’ But that raised a rather puzzling problem. Nigel was now inside a train, and it was obviously moving very fast. So why did it not make the same familiar noise? – why did it not say ‘whoosh’? Presumably the question was simply set to one side, as being too difficult to answer; indeed it is hard to see how Nigel could have answered it, other than by construing it as a general principle, that that is the way things happen. Both these occasions are what we might call monologic moments: the child is talking to himself, but in the context of extended dialogue through which he has built up a robust meaning potential – robust already in its control of the semantics of the mother tongue, but robust also, and critically, in that it is capable of being expanded to an indefinite extent by the child’s own further experience. At the time of these two episodes, Nigel had been living, and actively listening at least during waking hours, for close upon two years. Let us look at two other occasions from about three months before and after, one from 1; 8 and one from 2; 3. At 1; 8 we had taken him to the London Zoo, where there was one section where small children could mingle among the animals and touch them. Nigel was stroking a goat, while clutching in his other hand a plastic lid from a drinking cup. The goat started nibbling at the lid; the keeper saw this and came to take it away, explaining to Nigel that the goat shouldn’t eat the lid – it wasn’t good for it. Back home some hours later, Nigel said to me ‘try eat lid’. I asked him what tried to eat the lid. He repeated ‘try eat lid’, so I asked the question again. ‘goat’, he said; then went on ‘man said no ...goat try eat lid ...man said no’. 8 Society in Language, Language in Society

That evening, as his mother was putting him to bed, he told her the story: ‘goat try eat lid ...man said no’. ‘Why did the man say no?’ his mother asked him. ‘goat shouldn’t eat lid ...[shaking his head] good for it’. His mother repeated this: ‘The goat shouldn’t eat the lid; it’s not good for it’. At this point, Nigel retold the entire story: ‘goat try eat lid ...man said no ...goat shouldn’t eat lid ...[shaking his head] good for it’. This was a coherent narrative; it was also, as I remarked at the time (1975; see Collected Works, vol. 4, ch. 7, p. 186), highly cohesive: ‘there is the repetition of the words goat and lid, the anaphoric it, and the marked tonic on shouldn’t showing that the following eat lid is to be interpreted as “Given” ’. Nigel repeated the story verbatim, many times in the course of the next few months; it may have served as a kind of template for construing complex narrative sequences. Fast forward to 2; 3: Nigel had a scrapbook into which he stuck his favourite pictures cut out from magazines. He had learnt to turn the picture over, put gum on the back, and then turn it right side up again before sticking it in. ‘You did put the gum on the second side,’ he said proudly, ‘but not on the underground train picture’. You at that stage meant ‘I’; it was how he referred to himself. He then went on to formulate this as a general principle: ‘you can put the gum on the back of the fast electric underground train picture, but not on the back of the fast electric underground train picture’. Once you had turned the paper over, the original picture on the front had now become the back of the picture. The intonation now totally conformed to the adult pattern: ﬁrst, tone 13, with fall on back and rise on picture; then tone 1, with fall on back, the rest now having become entirely Given.4 Nigel had stuck in one of his train pictures upside down. I looked at him inquiringly. ‘Did you stick it wrong way up because it doesn’t stick that way’, he explained. ‘You stuck it wrong way up because it doesn’t stick that way.’ I was still looking puzzled, so he went on. ‘No the train is not wrong way up’, he explained. ‘It’s the picture that’s wrong way up. The picture won’t fall off the scrapbook.’ Some explanation is needed of Nigel’s grammar at this time. As already noted, he refers to himself as you; and also, he makes a clear distinction between declarative and interrogative – though not the same distinction as is made in adult English. It will be seen that these two features are related. The declarative meant ‘I’m telling you something you already know’; it was a rehearsal of some shared experience. The interrogative meant ‘I’m telling you something you don’t know’; this was introduced as he came to realize that language could be used in this function – that is, as an alternative to shared experience, as a surrogate form of sharing. In this instance, he could see that I did not know why he had stuck the picture in upside down; so he told me the reason – in the interrogative. But when he repeated it, I did know, because he had just told me; so he switched over into the declarative. The reasoning was: the picture would not ﬁt in to the space where I wanted it when it was right way up, so I turned it around. It now looks as if the train M. A. K. Halliday 9 is going to fall off the rails; but it will not, because the train is really the right way up. It is the picture that is wrong way up; but that will not make the picture fall off the scrapbook. There was nothing remarkable about this exchange; it was just a piece of ordinary everyday conversation.

1.3 Reasoning about language: Ruqaiya Hasan’s quest for meaning

These last examples were not monologic moments; they were moments of typical dialogue. The goat incident shows clearly the joint construction of narrative, with the interlocutor promoting by asking questions; not just ‘what happened next?’ (that is, simply ‘continue’), but ‘who?’, ‘what?’ and even ‘why?’, which Nigel not only understood but also answered appropriately, and by 2; 3 he could infer that his father was asking for an explanation. The questions were ‘real’ questions. They might be prompts, like those his parents asked as he was constructing the goat story; but they were not quizzing him, getting him to perform as a way of testing what he knew, and hence they constituted genuine challenges to his developing meaning potential. Ruqaiya’s work on the semantics of question and answer showed how by 3½ to 4 years old the children were using the various systems of questioning very much in the ways that they had themselves been questioned in the course of natural informal dialogue in the home. In the Nigel data we can track this through from the beginning of the child’s entry into the mother tongue. Ruqaiya had arrived at the University of Edinburgh back in the year 1960. She arrived there direct from Pakistan, with an MA in English literature from Government College Lahore, which was part of the University of the Punjab. There she had studied with a number of scholarly and enlightened teachers; fortunately, in her time at the School of Applied Linguistics in Edinburgh she was also going to meet with some scholarly and enlightened teachers, notably Ian Catford, Angus McIntosh and David Abercrombie. Although she came armoured with a strong resistance to ‘grammar’, this was a personal response to earlier experience at school; Ruqaiya felt no trace of the antipathy that has so often arisen between ‘lang’ and ‘lit’, with elaborate defence being built up against any ‘scientific’ approach to language. On the contrary, she found no difficulty in coordinating the binocular vision, from linguistics and literary studies, into a coherent image of language as verbal art (see her Linguistics, Language, and Verbal Art, 1985a). This was wholly in keeping with the approach to learning that she encountered in Edinburgh; the atmosphere was open, friendly and collaborative. Ruqaiya went on to do her PhD comparing the writings of two contemporary English novelists, William Golding and Angus Wilson; the question being asked at the time was, what can be said about the language of a full- length prose novel where none of the sentences is in any way remarkable 10 Society in Language, Language in Society in its grammar? – Ruqaiya’s thesis was, to my knowledge, the first serious venture towards an answer. Verbal art, in the sense of the Russian term slovesnosty,asusedby Mukaˇrovský, has remained as a recurrent motif throughout Ruqaiya’s work, revisited from time to time in different contexts. It seemed clear to her that there was no such thing as a distinct ‘literary variety’ of a language; yet she was convinced that there was more to be said about how language works in a literary text than the rather oracular metaphors often favoured in critical appraisals of the author’s style. A comprehensive theory of language ought to serve as a tool with which to think about this question. Meanwhile, however, Ruqaiya went to work with the Nuffield Foreign Languages Teaching Materials Project, under the direction of Sam Spicer at the University of Leeds. Her task there was that of talking with children, selected year by year from ages 7 to 13, in small groups in their school, and recording and analysing their conversation. This was part of a project for producing materials for teaching foreign languages in primary school; it was designed to ensure that the materials kept in harmony with, but did not run ahead of, the children’s level of performance in their mother tongue. Like so many good educational projects, this one was abandoned just as it became clear how successful it was; ostensibly because it was too costly, but more because it ran ahead of current notions about, and understanding of, children’s language and learning. But this brief period when Ruqaiya was focusing attention on how children talk in interactive situations was valuable experience when she came to design and carry out her own research into young children’s spontaneous spoken language. This work really started when Ruqaiya joined Basil Bernstein’s research team at the Sociological Research Unit in the University of London Institute of Education. Bernstein was, in Ruqaiya’s view (and also in mine), one of the leading social theorists of the twentieth century; he ran a major programme throughout the 1960s investigating the phenomenon of educational failure as it was then becoming an issue in Britain, especially in the industrial inner cities. In one of Bernstein’s experiments, young children were given a cast of soft toys and asked to tell a bedtime story to the teddy bear about the boy and girl, the sailor, and the little dog; Ruqaiya had the task of analysing the recorded texts. Bernstein wanted to find out how the social order, with its fracturing along socio-economic fault lines, was maintained and transmitted in the culture. He was aware that the essential channel for the transmission was everyday language; he formulated the problem in terms of a disjunc- tion between two ‘socio-linguistic coding orientations’, the ‘elaborated’ and the ‘restricted’. To succeed in school, a child had to have access to the elaborated code, which was the medium of educational knowledge; but some children, especially of the working class, had seldom encountered this kind of language and found it virtually impenetrable. Bernstein wanted to characterize the difference between the two codes, and suggested various syntactic M. A. K. Halliday 11 features that might be involved; but he was not a linguist, and was not able to build his ideas into any coherent hypothesis. Ruqaiya would be able to show why. Before joining Bernstein’s research unit Ruqaiya had worked with me in the Nuffield/Schools Council Programme in Linguistics and English Teach- ing at University College London. (She had to resign from there when we got married.) She had started working on cohesion, in what later became our joint publication Cohesion in English (Halliday and Hasan, 1976). This gave her another effective tool for text analysis. It also helped to bring out more clearly the relation between lexicogrammar and semantics, as distinct but inseparable strata within the content plane of language. After the book was published, Ruqaiya continued to build on the framework of cohesion and developed the concept of ‘cohesive harmony’, showing that the overall texture of a text was dependent not just on the density of cohesive ties but on their resonance with other lexicogrammatical features, such as the selections of process type and participant roles in transitivity. Cohesive patterns both strengthened and were strengthened by the unfolding semantic motifs that persisted through distinct but overlapping moments of the text. While working with Bernstein, from 1967 to 1970, Ruqaiya had come to realize what were two fundamental problems that he faced. One was the nature of the data that Bernstein had assembled. His experiments had been thoughtfully designed and professionally carried out; but the language was still elicited rather than natural. People behave in a different way in experimental situations; especially in their language behaviour, and especially if they are children. In a sensitive enquiry such as this, it was essential to build on naturally occurring data. The second basic problem was the nature of the linguistic analysis. Ruqaiya realized that the variation in ‘code’ was an orientation to different meaning styles, and had to be investigated at the stratum of semantics, taking the lexicogrammar as the realization of semantic choices. The first problem posed a challenge to language technology, and to the logistics of research design; the second posed a challenge to linguistic theory and the extent and accuracy of the description (see Hasan, 1973). Ruqaiya took up an appointment at Macquarie University in 1976, and within a few years had embarked on a major research project which would engage with both these issues. By this time the technology for recording speech had advanced to the stage where it was possible to record natural conversation in everyday situations in the home, without impinging on the movement, or even the awareness, of those taking part. And the resources for analysing such discourse in terms of an adequate descriptive linguistic model had also significantly advanced; here Ruqaiya herself had played an active part. Ruqaiya had been one of a group of postgraduate students in Edinburgh, and subsequently a member of the project at University College London, working with the original ‘scale-and-category grammar’; she was closely 12 Society in Language, Language in Society associated with it as it evolved into systemic functional linguistics, and was herself a party to its evolution – it was she who pressed hardest for including the ‘functional’ in its name. For Ruqaiya it was first and foremost a theory of meaning; her inclination has always been towards substance rather than form. So when she embarked on her own research project – actually a cluster of related projects – she was able to set up semantic systems on a firm lexicogrammatical foundation. She also explored, in two important papers (1985b, 1987b), the notion of lexis as delicate grammar, showing that a lexical item could be described as a bundle of features each of which was located in a network deriving from some general grammatical system (for example, of process types in transitivity). The unifying aim of her research projects was that of re-examining Bernstein’s hypothesis, about social continuity and change, and the place of language in transmitting the patterns of the culture. But Bernstein’s researches had been carried out in England, in a society which had always worn its class structure on its sleeve; whereas Ruqaiya’s were to be carried out in Australia, in various locations in and around Sydney. Most Australians have seen themselves as a classless, egalitarian society, in explicit contrast to their perceived images of England; it was essential not to build any assumptions about social hierarchy into the design of the research. A sample population had to be selected so as to be able to reveal any patterns that might be there; but the linguistic variables themselves, and the statistical methods adopted, had to be such that any patterns could only emerge out of the data; they could not be artefacts of the way the data were managed and interpreted. What was Ruqaiya looking out for? – Semantic variation: variation in the way the meaning potential of the language was deployed in a given situational context. Given a set of everyday situations in which mothers were interacting with their children (who were all aged 3½–4), like playing a game, or preparing a meal, or helping the child to dress, it might turn out that all mothers would draw on more or less the same reservoir of meanings; alternatively, different mothers might show significant differences in their use of semantic resources. If they did, was the variation randomly distributed among different individuals, or did it correlate with some non- linguistic factor present in the sample population? The computer program contained no information about the individuals taking part – they were just 24 mother–child dyads; so any such correlation might turn out to be of interest.5 Out of six hours of dialogue from each of the 24 dyads, Ruqaiya and her research associate Carmel Cloran selected a total of 20,000 messages, and analysed them manually in terms of systems of semantic options. The results were subjected to a form of cluster analysis known as ‘Principal Com- ponents’, identifying which clusters of semantic features accounted for the greatest amount of variation in the data. The results are now well known – or M. A. K. Halliday 13 if not, they should be. There was considerable variation in the choice of meanings exchanged between mother and child in their day-to-day encounters in the home; and two clusters of semantic features stood out as chiefly responsible. When these were plotted against the 24 dyads, one turned out to correlate with the sex of the child: mothers of girls talked to their child significantly differently from mothers of boys. But the other, statistically more powerful correlation was with the social class of the family, defined as the relative workplace autonomy of the main breadwinner: mothers in the HAP (higher autonomy profession) families talked significantly differently from mothers in the LAP (lower autonomy profession) families. And comparable differences were found in the talk of the children, who by this age were already (as I put it elsewhere) fully paid-up members of the social bond. Ruqaiya and Carmel reported their findings in a number of articles, bringing out the critical role of language – spontaneous speech in everyday contexts in the family – in maintaining and transmitting the underlying ideology of the culture. This was a remarkable finding in relation to the goals of the research (Hasan, 2009a). But what they had done in passing, so to speak, is to my mind no less noteworthy. Ruqaiya had developed a semantic analysis within the framework of systemic functional theory, with networks of systems including entry condition, options and their realization in lexicogrammatical terms. Carmel Cloran had carried out a statistically robust piece of research, using as data carefully observed and recorded (not elicited) natural spoken language. Together they had provided a model for the scientific study of language, facing a problem that was posed not by linguists but by colleagues in another discipline: a sociological problem that could be addressed by investigating natural language in its natural eco-social environment. The essential concept underpinning all this work was that of semantic variation (see Hasan, Collected Works, vol. 2, 2009a). The principle of variation in language is that of variation against a higher stratum constant: phonetic variation with phonology remaining constant, phonological variation with the lexicogrammar remaining constant (same wording, different sounding), lexicogrammatical variation with the semantics remaining constant (same meaning, different wording). So if there is semantic variation, where is the higher-level constant? Where the meaning is different, what is it that remains the same? Ruqaiya had always taken seriously Firth’s (e.g. 1957) conception of the ‘exterior relations’ of language: his view that the ‘context of situation’, as first formulated by Malinowski, could be brought within the scope of a general linguistic theory. Her work on semantic variation showed why it had to be: the higher-level constant must be provided by the context. Bernstein’s ‘socio-linguistic coding orientation’ was the choice among variant meanings as functional in the same context of situation; Ruqaiya had given substance 14 Society in Language, Language in Society to this insight. In the more immediate context of her mother-and-child dialogue, the patterns of question and answer showed considerable variation, but all within the same framework, of using the occasion to facilitate the child’s future entry into and survival in the existing socio-economic order. On a wider canvas, one and the same context would be taken by some mothers as an occasion for experiential instruction and by others as an occasion for interpersonal management. This means, as Ruqaiya has always maintained, that one needed to be able to model the context as one stratum in the total description of language. The linguistic study of context has been a motif running throughout her work, taken up in many of her publications from the early 1970s onwards (see especially 1981, 1985c, 1999, 2001, 2009b). Three issues figure prominently: one, the delimitation of ‘context’; two, its manifestation, and three, its representation in systemic terms.

1. The context of situation of a text is delimited by its value along the three parameters of field, tenor and mode; these taken together specify the ‘contextual configuration’. Since these are realized by the semantic choices in the text, this will cover just those features of the context which are relevant to the text’s production and understanding. Ruqaiya refers to this as the ‘relevant context’ and distinguishes it from the ‘material situational setting’, accidental features of the environment which have no bearing on the text itself (such features may intrude, of course, but if they do, they thereby become relevant). 2. The context is manifested in the choice of register, the functionally defined semantic variety of the text. Ruqaiya describes this in structural terms as its ‘generalised structure potential’; this characterizes a broad family of register varieties, such that any particular variant (say, purchasing an umbrella, as a more specific instance of a service encounter) will display one of the possible actual structures derivable from the general formula. She drew a clear distinction between the structure of a text and its ‘texture’ (Hasan, 1985c). Both are essential features of the unity of a text; but texture is a measure of the semantic relations between one part of the text and another, including but not limited to the relations of cohesion. Texture is what gives a text its recognizable quality of coherence, independently of its structural make-up. 3. Is it possible to systemize the stratum of context – that is, to represent it in systems of mutually exclusive options, and organize these into system networks? Various scholars, including Martin (1992), Bowcher (2013) and Butt (2004), have proposed contextual system networks; Ruqaiya has discussed these as well as offering her own proposals in a recent paper (Hasan, 2014) on the contextual parameters of field, tenor and mode and their relation to her concept of ARC (action, relation and contact) as the underlying variables M. A. K. Halliday 15 embodied in social practice. Extending the paradigmatic perspective from the core strata of language to semantics (options in meaning) and through these to context (options in social practice) has been perhaps the underlying motive pervading much of her work over the past two decades.

1.4 Learning to reason with language: construing time and cause

At 2 years old, Nigel had many of the basic lexicogrammatical resources of his language in place, such that his meaning potential could now expand along with his experience, including of course his linguistic experience, the ongoing exchange of meanings with other people. This combination of the material and the semiotic, each serving as environment of the other, is what challenges and nudges forward a child’s ability to reason things out and find order in the flux of goings-on around him. Nigel’s struggle with the ‘whoosh’ phenomenon shows something of the work that he has to do in order to reach what he himself regards as understanding. We cannot point to any specific features of the grammar that are critical in constructing a framework for reasoning with; every region of the network is brought into service. Here I will just pick out two features which illustrate Nigel’s development at around this age: time, as construed in the grammatical system of tense, and cause and condition, as construed in the grammar of the clause complex.6 Until he was about 21 months, verbs consisted of a lexical element only, with no marking for tense or aspect; for example,

1; 8; 25 big nòise (Mother: who made a big noise?) drìll make big noise 1; 8; 25 train go cròss ...go tòwn ...no more tràin ...time go hòme

So Mummy open window could mean either ‘(I saw) Mummy open the window’ (falling tone: mathetic) or ‘I want Mummy to open the window’ (rising tone: pragmatic). The functional distinction between get and got had already been established:

1; 8; 12 dada got scrambled ègg ...mummy get fór you scrambled egg

– the ﬁrst mathetic the second pragmatic; then at around 1; 9 Nigel began to use two marked tense forms, the simple past, and the secondary present in -ing, in contrast with the unmarked form: that stick broke, ﬂy climbing Mummy’s chair, Pauline went on two fast chuffa [trains]. The contrast appears clearly in a sequence such as:

1; 9; 16 that train gòing ...that train staying hère ...people want get òff that train 16 Society in Language, Language in Society while the simple form of the verb remains frequent in the pragmatic, as in dada get knife take skin off apple, signalled as always by the rising tone. But there were still phonetic constraints: it was only during 1; 10 that Nigel made an audible distinction between present and past tenses of regular ‘weak’ verbs; by this time the simple future had been added, and also the negative forms:

1; 10; 11 that train won’t go any mòre ...that train doesn’t say ànything 1; 10; 25 [to himself] yes you can have some more méat ...but it will be a bit hòt 1; 11; 13 thàt’s where you walked on the wáll ...càn’t walk on the wall from thére

Meanwhile, as so often happens in early language development, an instance occurs which seems to be ahead of its time, yet it anticipates other things to come. This was the ﬁrst appearance of a compound tense:

1; 9; 24 have céllo [I want the record sleeve with the picture of a cello] (Grandmother: does he know what a cello looks like? Father: I don’t think he’s ever seen a cello.) have seen cello in pàrk

The past in present have seen is appropriate in the context (if factually mis- taken!); it contradicts his father’s I don’t think he’s ever seen, but at the same time is not directly modelled on it – the relation between the two is in fact quite complex. Another early instance of a compound tense was one of past in past:

1; 10; 3 that blue pin had got lòst ...that white pin had lòst where had got lost may have been a repeat of something he had heard before, perhaps said by his mother; but it was not in the talk that had just preceded, and his own attempt to repeat it did not quite work out. The first compound tense to appear regularly in Nigel’s speech was of course the present in present (the so-called ‘present continuous’), since this is the unmarked variant of ‘present’ in the context of a material process. At first this is simply the -ing form of the verb, as in that train going above; the full finite form was however clearly audible in:

1; 9; 26 you’re not having stewed àpple ...you’re having blàckcurrant where the Finite is fused with the Subject you (meaning as always ‘I’). The -ing form also appears with other ﬁnite verbs, as in:

1; 10; 0 first Daddy finish talking then go in párk M. A. K. Halliday 17 where the first ...then sets up a clear temporal relation between the two hoped-for events. Shortly after this, Nigel introduced the when ...then sequence, construing the time relation as a hypotactic clause nexus:

1; 10; 10 when music fínish then I sing you dùck song [= ‘you sing me’] 1; 10; 28 when you ﬁnish your tea thèn we play rao [‘when I ﬁnish my tea’]

The then clause typically refers to the future; by 1; 11 these usually take the future tense with will:

1; 11; 11 when I’ve finished my téa there’ll be a lot of tiny tiny leàves in the cup of tea [‘when you’ve finished your tea’, ‘in the teacup’] 1; 11; 16 when Daddy finished drawing a railway line with that pen and thèn I will take out a green pen for you to dràw [‘you will take out a green pen for me to draw’]

But the future may also occur in a narrative of future time:

1; 11; 24 that tree has yellow leaves ...the leaves will fall óff and there will be no leàves on

Other temporal relations, and other more complex tenses, began to appear in Nigel’s third year of life:

2; 2; 28 want Daddy to give you a toothpick while you’re rushing away with his things and disturbing him 2; 3; 7 when you were looking out of the window of the white and blue trainˇ [present in past], you saw some sparks [past] which the other train had left behìnd [past in past]

This now extended to include a tense of three terms:

2; 3; 0 (Father: It’s been snowing.) like it had been snowing in the big pàrk [present in past in past]

Sometimes the tense construction goes awry:

2; 5; 24 are you going to get into your cháir [‘I’m going to’: future in present] ...did you went to get into your cháir [future in past]

This may seem obscure, although it is actually quite understandable in terms of Nigel’s system at the time. The you meant ‘I’ and the interrogative meant ‘I’m telling you’; so ‘I’m going to get into my chair’. The second clause meant ‘you see? I was going to get into my chair – and I did’. did you went to was 18 Society in Language, Language in Society his construction of future (went to) in past (did), the equivalent of adult was going to. The implication of this is typically ‘was going to, but didn’t’; here however the meaning is ‘was going to – and did’, comparable to an adult saying I always was going to do that. Since ‘did you went to’ is a form of wording which Nigel could not have heard used by anyone else, it shows clearly that he had understood the serial principle that governs the English tense system. By the same token, it shows how important it is, as Ruqaiya has always maintained, to trace the origin and development of the child’s capacity for meaning, which is so critical to success in education. In the construal of time, even quite closely related languages can differ quite signiﬁcantly, in how they organize it and how they distribute it between the grammar and the lexis; the English tense system, which in modern linguistic terms would probably qualify as ‘extreme’, is a resource which, once the child has cracked the code, is rather handy for thinking with. The compound tenses help to construe not only temporal relations but also relations of condition and of cause. Again Nigel provided a trailer, an advance notice of what was about to come:

1; 10; 23 if you were standing on the platform,ˇ that train will go whòosh ...if you were standing on the railwayˇ line, that train will go boòmp ...train will go boòmp

Reasoning of this kind became frequent from about age 2, with both paratactic and hypotactic clause nexuses:

2; 0; 4 that Russian train didn’t go fast enòugh ...it didn’t say whòosh so they taken it awày 2; 0; 12 your clock is gòing nów ...but if you pull the wire then the clock will nòt go r-r ...then the clock will stòp 2; 0; 25 that nòt a good way to get dówn because ...otherwise [‘if you do get down that way’] you will bump your head on the tàble

Here Nigel hesitated at because, not sure how to proceed; otherwise was his formulation of ‘if that was, or is, the case’, for which there is no conjunction in English; compare:

2; 1; 16 you mustn’t take the wheelsˇ off that train because ... otherwise the wheels will get bròken 2; 3; 25 they don’t want you to eat their grapes all ùp, otherwise they will be very sòrry

The same reasoning is also brought in to account for consequences which are or would be from Nigel’s point of view undesirable: M. A. K. Halliday 19

2; 1; 17 have that cráne ...in Mérry-go-round ...well you should have put that puzzle bàck, thèn you could have had that crane 2; 2; 15 if you make it fall on the ﬂoorˇ how will Daddy be able to cùt it?

Both purpose and cause are grammaticalized by age 2; 5:

2; 5; 21 we had to put some òintment on it, some stìcking plaster on it, so that it would get bètter 2; 5; 22 it was not shut properly because it was òutside ...but if it goes ˇın I think it’s shut properly ...it’s shut properly now

Sometimes Nigel had used a clause complex to turn a narrative of recalled events into a general hypothesis about the future:

2; 0; 31 if you take the red toothbrush for a walk in the park, and you will throw it and it broke, then Daddy will pick that little bit up, and then Daddy will put that little bit in his pocket again to mend with sellotape

Mending something is of course making it new again. Nigel is looking for his plasticine, and assumes his father must have put it away:

2; 6; – why did Daddy put the brown plasticine away before breàk- fast? (Father: I didn’t.) it wasn’t crumblyˇ ...the oldˇ plasticine was crumbly ...we have to wait till the old plasticine gets new again and thèn we can use it to make a diesel train

There will always be problems with what is the opposite of old:

2; 4; 16 (Father: Yes that was a very old lion.) but the otherˇ one was a nèw lion

Thus does the lexis sabotage the grammar’s attempt to domesticate the passing of time. We have seen from the beginning of Nigel’s breakthrough into the mother tongue at about 1; 7½ that he has been using language both to act with and to think with. Indeed he himself made this quite clear with the distinction between the pragmatic and the mathetic: the pragmatic, spoken on a rising tone, was language as action: it meant ‘somebody do something!’; while the mathetic, spoken on a falling tone, meant ‘that’s how things are’. Every utterance, with one class of exceptions, accorded clearly with this pattern; the exception was the set of utterances which were demands for music, which had not the rising but the falling tone – because here (and 20 Society in Language, Language in Society only here) the mood was realized gesturally, by beating time with the right arm, the one gesture that Nigel had carried over from the repertory of his protolanguage. That demands for music reverted to the falling tone showed that the falling tone was the unmarked option. That could simply be an artefact of the phonetics: the falling tone makes less demands on energy. But it might suggest that the mathetic is the default condition for the child’s construction of meaning. The mathetic is the child construing his own experience, giving meaning to the world he finds around him and inside him. You need this construal, of course, in order to make demands of other people; once you progress beyond the basic ‘I want’, Nigel’s early have it!, have that!, you have to specify what it is you want given to you or done for you. Eventually, the pragmatic will evolve into two different kinds of demand, the demand for goods-&-services (imperative) and the demand for information (interrogative); by then it will have become grammaticalized within the mood system of the child’s mother tongue. Just as the pragmatic function cannot be realized without the backing of the mathetic, so also the mathetic cannot be successful without a flavour of the pragmatic: thinking, making sense of one’s experience, needs interaction with other people – at least gaining their attention, but better their engagement in dialogue. As the child completes the transition to the mother tongue, this simple opposition of mathetic/pragmatic resolves itself into the more complex systems of the adult language. The systemic opposition between pragmatic and mathetic utterances evolves into the elaborated systems of mood, which modulate the forms of interaction between speaker and attendant others. The construing of the content, as elements configured into processes, process types, and logical relations between processes, evolves into transitivity and all the manifold resources for making ideational meaning. (They are not evolving, of course; they are developing; but it feels like witnessing evolution, when you notice it happening.) Along the way, there is always room for the imagination, and for imaginative lines of inquiry. Ruqaiya’s researches into the discourse of 3½–4-year-old children demonstrated beautifully both the interactive nature of thinking/learning processes and the variety of semantic strategies, both ideational and interpersonal, which children (and their mothers) invoke as these processes are taking place. Here is Nigel at 3; 5, just coming into this age band:

Wednesday Tuesday Thursday Friday Mònday ...Monday Tuesday Wednesday Thursday Frìday ...Why does Friday come before Sàturday? – (Father: Well I don’t know. I suppose that’s where they put it.) – Yes but ˘I put it in the middle of Wèdnesday. – (Father: Is that where you put Friday?) M. A. K. Halliday 21

–Yes...[looking up at the sky] Is it way up in the clouds?ˇ – (Father: You mean that’s where the days are? That’s a nice idea.) – [seriously] Yes but àre they way up in the clóuds?

One of Ruqaiya’s mentors has always been Benjamin Lee Whorf. I hope Whorf was way up in the clouds, listening in.

1.5 Postscript

I kept my diary of Nigel’s language up to age 2½, noting only some interesting occasions over the following 12 months. Even then I never found time to give proper attention to the last four months of my record; I am extremely grateful to Joy Phillips for all the work she did, which I have drawn on in writing this chapter. Most of all, of course, I am grateful to Nigel for being such a lively and talkative child; and that, in turn, is owing to the encouragement of his mother, who treated him as a communicating being from the moment when he was born. Time and cause remain as problems for us throughout life; I will conclude with a few examples from Nigel’s subsequent conversations:

5; 5 If you started from Chicago instead of Glenview, would you get to Minneapolis later? Cos it’s further. — (Father: No, you’d get there at the same time. You’d just start earlier.) — No, you wouldn’t ...but ...[unable to work it out] 6; 4 How old are you? — (Father: Fifty.) — Then you weren’t born when the Blackheath Hill track was taken up. When you are eighty, how old will I be? — (Father: Thirty-six.) — Thirty-six. Ooh! When you’re eighty, then you will have been born before the track was taken up. 7; 0 How do you see what happened long agò, before you were bòrn? — (Father: You read about it in books?) — Nó; use a mìcroscope to look back. — (Father: How do you do that?) — Wèll. If you’re in a car,ˇ or in an observationˇ coach, you look báck, and then you see what happened befòre. And you need a mìcroscope to see what happened long ago,ˇ because it’s very far awày. 7; 10 [making a complicated space vehicle with a construction toy] In the future,ˇ when I inventˇ these gadgets, then in the future futureˇ they’ll be dumped on rùbbish heaps. 22 Society in Language, Language in Society

Notes

1. For an account of Nigel’s early language development see Halliday (2004). The transcript of the Nigel data (‘Listening to Nigel’, University of Sydney Department of Linguistics, 1984) is reproduced on compact disc inside the back cover of that volume. 2. The formula ‘years; months; days’ in language development studies represents the age of the child; thus if the child is nine months and twelve days old, this will appear as 0; 9; 12. If only two numbers appear, these refer just to the year and month. 3. The simpliﬁed tone symbols used in the present chapter are falling as shown on the second syllable of agò, rising as shown in báck, and falling–rising as shown in the second syllable of ago.ˇ The placement shows the location of tonic prominence (‘primary stress’). 4. For the analysis of intonation see Halliday and Matthiessen (2014, ch. 1); or, for a more detailed account, M.A. K. Halliday and William S. Greaves, Intonation in the Grammar of English [London: Equinox 2008]. Nigel’s falling tone is tone 1. His rising tone begins as the fusion of tones 2 (high rise) and 3 (low rise); and his fall– rise as the fusion of tone 2 (high fall–rise), tone 4 (falling–rising) and tone 13 (fall plus low rise). Nigel separates out these variants during 1; 11–2; 1. 5. The 24 dyads were selected according to (1) the profession of the main breadwinner in the family, and (2) the sex of the child; so 12 ‘higher autonomy profession’ (HAP) families and 12 ‘lower autonomy’ (LAP), each group equally divided between those with a boy and those with a girl. 6. For the ‘serial’ analysis of the English tense system, see Halliday and Matthiessen (2014, ch. 6); for the analysis of the clause complex, Halliday and Matthiesssen (2014, ch. 7).

References

Bowcher, W. L. (2013) ‘Material action as choice in ﬁeld’. In L. Fontaine, T. Bartlett and G. O’Grady (eds) Systemic Functional Linguistics: Exploring Choice (Cambridge: Cambridge University Press), pp. 318–41. Butt, D. G. (2004) Parameters of Context: on Establishing Similarities and Dissimilarities between Social Processes (Sydney: Centre for Language in Social Life, Macquarie University). Firth, J. R. (1957) ‘A synopsis of linguistic theory, 1930–1955’. In J. R. Firth et al., Stud- ies in Linguistic Analysis (Oxford: Basil Blackwell) (Special Volume of the Philological Society). Halliday, M. A. K. (1975) Learning How to Mean: Explorations in the Development of Language (London: Edward Arnold) (Explorations in Language Study). Halliday, M. A. K. (2004) The Collected Works of M. A. K. Halliday,Vol.4:The Language of Early Childhood (London and New York: Continuum). Halliday, M. A. K. and Greaves, W. S. (2008) Intonation in the Grammar of English (London: Equinox). Halliday, M. A. K. and Hasan, R. (1976) Cohesion in English (London: Longman). Halliday, M. A. K. and Matthiessen, C. M. I. M. (2014) Halliday’s Introduction to Functional Grammar (London and New York: Routledge). Hasan, R. (1973) ‘Code, register, and social dialect’. In B. Bernstein (ed.) Class, Codes and Control, Vol. II: Applied Studies towards a Sociology of Language (London: Routledge and Kegan Paul), pp. 253–92. M. A. K. Halliday 23

Hasan, R. (1981) ‘What’s going on: a dynamic view of context in language’. In J. E. Copeland and P. W. Davis (eds) The Seventh LACUS Forum (Columbia: Hornbeam Press), pp. 106–21. Hasan, R. (1985a) Linguistics, Language and Verbal Art (Geelong, Vic.: Deakin University Press). Hasan, R. (1985b) ‘Lending and borrowing: from grammar to lexis’. In J. J. Clark (ed.) The Cultivated Australian (Festschrift presented to Arthur Delbridge) (Hamburg: Helmut Buske). Hasan, R. (1985c) Part B in M. A. K. Halliday and R. Hasan (eds) Language, Context and Text: Aspects of Language in a Social-Semiotic Perspective (Geelong, Vic.: Deakin University Press). Hasan, R. (1986) ‘The ontogenesis of ideology: an interpretation of mother–child talk’. In T. Threadgold, E. A. Grosz, G. Kress and M. A. K. Halliday (eds) Semiotics, Ideology, Language (Sydney: Sydney Association for Studies in Society and Culture), pp. 125–46. Hasan, R. (1987a) ‘Reading picture reading’. In The Proceedings of the 13th Annual Conference of the Australian Reading Association, July 1987, Sydney. Hasan, R. (1987b) ‘The grammarian’s dream: lexis as most delicate grammar’. In M. A. K. Halliday and R. P. Fawcett (eds) New Developments in Systemic Linguistics, Vol. 1: Theory and Description (London: Pinter), pp. 184–212. Hasan, R. (1988) ‘Language in the processes of socialisation: home and school’. In J. Oldenburg, T. van Leeuwen and L. Gerot (eds) Language and Socialisation: Home and School (Macquarie University, School of English and Linguistics), pp. 36–96. Hasan, R. (1991) ‘Questions as a mode of learning in everyday talk’. In T. Lê and M. McCausland (eds) Language Education: Interaction and Development (Launceston: University of Tasmania), pp. 231–68. Hasan, R. (1992) ‘Rationality in everyday talk: from process to system’. In J. Svartvik (ed.) Directions in Corpus Linguistics, Proceedings of Nobel Symposium 82, Stockholm (Berlin: de Gruyter). Reprinted in Hasan (2009a), pp. 313–56. Hasan, R. (1999) ‘Speaking with reference to context’. In M. Ghadessy (ed.) Text and Context in Functional Linguistics (Amsterdam and Philadelphia: John Benjamins), pp. 219–328. Hasan, R. (2001) ‘Wherefore context? The place of context in the system and process of language’. In S. Ren, W. Guthrie and I. W. R. Fong (eds) Grammar and Dis- course: Proceedings of the International Conference on Discourse Analysis (Macau: Macau University), pp. 1–21. Hasan, R. (2009a) The Collected Works of Ruqaiya Hasan,Vol.2:Semantic Variation: Meaning in Society and in Sociolinguistics, edited by J. J. Webster (London: Equinox). Hasan, R. (2009b) ‘The place of context in a systemic functional model’. In M. A. K. Halliday and J. J. Webster (eds) Continuum Companion to Systemic Functional Linguis- tics (London and New York: Continuum), pp. 166–89. Hasan, R. (2014) ‘Towards a paradigmatic description of context: systems, metafunctions and semantics’. Functional Linguistics 1(9). doi:10.1186/s40554-014-0009-y Martin, J. R. (1992) English Text: System and Structure (Philadelphia and Amsterdam: John Benjamins). Phillips, J. (1985) ‘The development of comparisons and contrasts in young children’s language’. MA honours thesis, University of Sydney. Phillips, J. (1986) ‘The development of modality and hypothetical meaning: Nigel 1; 7½–2; 7½’. Working Papers in Linguistics 3: 3–20. University of Sydney Linguistics Department. 2 ‘Construe My Meaning’: Performance, Poetry and Semiotic Distance David G. Butt

2.1 Exordium: ‘Construe my meaning, and wrest not my method’

These words, quoted from the Elizabethan madrigal by Giles Farnaby (1560– 1640), evoke for me many of the dimensions of the thought and linguistic technique of Ruqaiya Hasan (e.g. Hasan, 2007). The primary point of Hasan’s emphasis is that a linguist should work so that all that is performed in linguistic analysis demonstrates its relevance to meaning. Second, in evaluating linguistic proposals, one needs to be assiduous in thinking through the implications of the theory – what the theory carries by way of its own meanings. Throughout Hasan’s semantic odyssey, there are then at least two levels of meaning for construe: it applies to the cultural and linguistic meaning under investigation through wording; and to the theoretical meanings that we bring with our modes of enquiry, and which are inherent in the terms of representation that we employ. With respect to the latter, ‘wrest not my method’ is also relevant: namely, do not take my method from me, or turn it to different purposes (more archaic). When I reﬂect on the line above, I am also reminded of Hasan’s regard for the power of Elizabethan creativity and the problem of the object in the analysis of verbal art. The latter was a starting point for her career in linguistics. And, below, I argue for the aptness of this entry into semantic exploration. Furthermore, the cultural character of Elizabethan England reminds us that the richness and complexity of a culture, and its power to reﬂect on the values that structure the living of life, are not the province of any one culture, nor are they inherent in the progress of modernity. When a realm of between 3 and 5 million people (in England 1550–1650)

24 David G. Butt 25 can produce the bank of literature that it did, then there is more to culture and meaning than modernity; and this ‘more’ needs defending. These resonances, which I will explain more directly in what follows, pick up the echoes from the experience of being taught by Hasan, and in my turn hearing the echoes of her teachers – both linguists (especially Michael Halliday) and empathic language specialists with deep commitment to literary analysis, for instance Angus McIntosh (1963), who was the Head of the Department of English Language at Edinburgh University (see the Foreword of Hasan, 2011, for an account of this period). The multiplying meanings of ‘construe’ bring us deep into a science of values. It is with such an endeavour – the investigation of values – that I associate Hasan’s forensic, linguistic gaze in linguistics. There are uses of ‘construe’ which apply to the interpretation of a text, the making sense of a cultural act of meaning, or just the more generalized knowing how to carry on from a certain point in a culture (to echo Wittgenstein’s characterization of what it means to ‘know’ a language). There is ‘construe’ as in the self-examining application of linguistic ideas (techniques and units) to the description of such texts and acts. There is also the more Elizabethan valeur, which emphasized the process of taking meanings to and from another language (so much superior to the use of ‘trans-late’, which emphasized the carrying across of a quantum of meaning rather than any complex struggle with interpretive possibilities). This interpretive complexity in ‘construal’ was always close by in Hasan’s disquisitions and reflections on meaning: ‘Can I express that English form of definiteness in Urdu? Where is it signalled, that is, given the absence of a definite article? If the person walking in the door is thematized, she is definite or known anaphorically ...If not thematized, is she non-definite?’ From this kind of intense enquiry, we have the systematic connection of meaningful patterns from culture to morpheme in the study ‘Ways of Saying: Ways of Meaning’ (1984a [1996a]). There is also a more difficult notion of ‘construe’ that pertains to how to conduct, and where to direct, one’s linguistic efforts. This relates to the rigour mentioned already; but it takes in what is yet to be achieved in linguistics, where the subject might go with sound theoretical navigation. It is the ‘irrealis’ in the academic discipline, something read off from the drift of the discipline (Butt, 1983). And this is, in my reading of Hasan’s meanings, a projection that preoccupies and animates her thinking, and a potential that explains the earnest and vigorous way in which she engages with alternative proposals from outside, and from within, systemic functional linguistics (henceforth SFL). But here I return to the social construction of value, and the study of verbal art, as it was such study that brought Hasan into linguistics; and it was in relation to problems raised by verbal art that she presented certain of her initial, original proposals concerning text and context. 26 Society in Language, Language in Society

2.2 Performance: ‘What’s aught, but as ’tis valued?’: Troilus 2.2.53

Shakespeare’s play Troilus and Cressida (T&C) is a disturbing and iconoclastic experience: so many of the positive and heroic topoi of traditional drama are given an absurd or malignant rendering in the context of the characters’ actions and words. All conventional values are nullified: young love is brought about through pandering and voyeurism; heroes are by turns petulant and self-serving; Troilus gives up his own love without contest after his grandiloquent arguments defending Helen and the actions of Paris; Hector is counting the cost of Helen (‘she is not worth what she doth cost the keeping’); and Achilles and his myrmidons murder Hector. In a retrospective on ‘The Age of Shakespeare’ (2004), Kermode comments that T&C is ‘in some respects linguistically the most difficult of the plays’ (2004, p. 122), in particular with respect to ‘complex debate’. It is even conjectured that the intended audience may well have been originally lawyers at the Inns of Court (2004, p. 127). The recurring motif of this rhetorical drama is value: how is a value recognized? How weighed against a contending value? How quantified against an individual reputation? And how pursued against the interests of the population that suffers the consequence of heroic evaluations? The BBC production of the play, directed by Jonathan Miller, supports this debunking of the heroic characters – the Greek kings wander their camp with wine glasses in hand; dread Priam is a mumbling wreck; Hector goes on about spongy bowels; and Troilus argues that value is not inherent in a thing but derives from how one decides to value it (viz. What’s aught, but as ’tis valued?). Kermode has also noted that sitting through the scene of Troilus witnessing Cressida make love to her new Greek master, to whom she has been relinquished, is ‘one of Shakespeare’s most painful’ (2004, p. 127). Nothing of traditional value survives the system of semantic relations that is the play. Hasan’s approach to, and account of, verbal art helps us make sense of what is going on here, as well as in the wide spectrum of forms in which language carries a primarily aesthetic purpose. At this point in my account, I submit a construal of her theory, as she schematizes it in a diagram (Figure 2.1). One needs to keep in mind that in verbal art we are dealing with ‘language all the way’ – it is not realistic to abstract, as some critics have, an artistic or aesthetic system as ‘parasitic’ on language. Verbal art is a form of language that needs explanation in terms of system and function just as any other use of text requires a representation in the interpretation of culture (Butt, 2007). Nevertheless, as Figure 2.1 suggests and other writings of Hasan elaborate, there are special conditions which apply to aesthetic motivation in text, and there is a form of ‘double articulation’ in the process overall. David G. Butt 27

Theme ← the semiotic system of verbal art Symbolic articulation Verbalization Semantics Lexicogrammar ← the semiotic system of language Phonology

Figure 2.1 Verbal art and language (Hasan, 1985, based on Hasan, 1979)

These special conditions include the fact that the author is not directly subject to contextual pressures in the way that the business of the world typically exerts its pragmatic demands. The topics and treatment of experience in verbal art are organizational choices in which the totality of the work all comes to be a metaphor – a construct which brings a ‘theme’ or ‘thesis’ into realization. Such a thesis concerning experience is more likely to be emergent rather than intended or previously enunciated; writers find the meanings by contesting with the demands of form and, less typically, by deviating from the demands of conventional forms. There is also a paradox- ical vector in the character of verbal art: even within the strictures of artistic forms, the effects of cumulative choices tend to establish uniqueness, and identifiable style. This is in contrast to the standardization that we expect in the contextually driven registers of quotidian community affairs. The language of verbal art is at the instantial pole of the cline of instantiation: namely, it is drawing on the ‘system’ of all texts in the language, but valued for the fact that it achieves a distinctive semantic ensemble from this system, an ensemble that can be recognized by its community to extend our ‘stock of reality’ (Blackmur, 1957 and Simpson, 1984, p. 73) – a valued vision of ‘what it means to be’. Such uniqueness would be alarming in other registers of our ‘collective consciousness’ – regularities and expectations guide us like a social compass for transacting with the world. Consider a legal contract or a manner of greeting or a medical discussion: these are cases when even clichéd formulations are assuring and required. The circumstances drive the situation causally, as Russian Formalists urged. Verbal art is different. The plasticity of language is at full stretch. The choices are directed to some uncertain final goal; it is heading somewhere towards a semantic denouement more latent than explicit before the formation of the work. While the Russian Formalists called this ‘teleological’ or end directed rather than ‘causal’, this is precisely where Hasan was willing to concede, ironically, some element of ‘divine madness’ in the creative choices that produce the level of ‘symbolic articulation’ (see Figure 2.1). An author may achieve a distinctiveness of text at one or more levels of ‘verbalization’. This could involve a novelty of plot, of tropes or figures of speech, of lexicogrammatical ‘pyrotechnics’, or of phonological patterning (or of all these strata). At one end of a spectrum of generic novelty, one 28 Society in Language, Language in Society can see the relatively strong expectation of formulaic (pre-fashioned) units of meaning in the creative ‘splicing together’ of traditional pieces in bardic cultures (e.g. Anglo-Saxon verse; medieval poetry and the troubadour traditions; or, for example, the Persian couplet forms in Firdowsi’s Shah-nameh). In some traditions, the act of the bard is given a particular place in the system of verbs: viz. ‘wrixlan’ (Beowulf ll.867–74) for weaving in Old English; or even the ‘Homerizing’ of classical Greek. Apparently, poets of bardic Ireland could be fined for infelicities of metrical performance, as well as rewarded for their virtuoso skills. The dynamic principles in formulaic verse were brought to light through the different contributions of Sievers, of Parry, and of Lord (see a useful overview of formulaic poetic process: Edwards, 1983). On the other hand, twentieth-century literature in Europe and America has accom- modated to generic and metaphoric novelty along with the acceptance, and even prioritizing, of individuated, confessional versions of subjectivity (e.g. Joyce’s Ulysses; and Lowell’s poems: Life Studies). These trends in verbal art can require the reader or listener to work hard for principles of cohesion and thematic unity. Thematic unity is sometimes ostentatiously eschewed. Between these poles – along the scales of tradition and inven- tion – there are strategies of semantic dissonance by which writers create a simulacrum of disorder and ‘anti-thesis’: existential dissonance or subjective ‘disaggregation’ (a term from the French psychiatrist, Janet, 1924). There is also the middle ground of plain statement exemplified by the American poet Robert Frost, to choose but one case. He creates semantic strangeness out of lexicogrammatical plainness. There is something of an inverse in the often anthologized works of the Welsh poet, Dylan Thomas: despite the initial shock of his challenging syntax (e.g. Freeman, 1975), Thomas’s poems often enunciate recognizably traditional topoi – nostalgia over the passing of youth; the solidarity of small communities under threat; or the striving against death and the dying of the light. At least in my experience of her discussions, Hasan tended to favour the investigation of the powerful plain style over verbal ‘pyrotechnics’ (Hasan p.c.). Examples of the ‘plain style’ from Australian literature include the work of the poet Les Murray and the novels of David Malouf, both subjects of Hasan’s stylistics. I am not in a position to comment on the stylistic traditions within which Hasan grew up in India and Pakistan, although Hasan often made profound asides on how different values were observable at work in the traditions: a case in point was the difference in attitude between European ‘mutability’ and Indian observations on living that reached the fullness of its cycle. Yet, generalizations on a dominant stratum in an artist’s style are only one, more overt, aspect of the wider investigation demanded by verbal style. For example, as a colleague has shown me, by applying Hasan’s stylistic methods to a more adequate selection of Dylan Thomas’s oeuvre, the ‘cohesion’ and ‘semantic innovation’ in Thomas’s poetry become insistent: Thomas explored a number of strategies for ‘semantic progression’ across his poetry; David G. Butt 29 that is to say, such progressions venture far beyond the density of verbal connections, the techniques, and the traditional, ‘public’ themes of those most often anthologized poems (see Ryan, 2013). The diagram on strata in verbal art (Figure 2.1) leads on to other insights about literature. In some cases, these insights were working principles (although often only implicit) in the language and literature training once shared by many who studied literature at universities. But they were not so clearly articulated and so systematically observed as they have been in Hasan’s work. These include:

(a) treating ‘semiotic distance’ as a problem requiring explicit, methodical analysis: for instance, by analysing the cultural distance between the context internal to the work and the milieu in which the work is ultimately read (context of creation in relation to context of reception); (b) establishing any claims of ‘consistency of foregrounding’ in the author’s choices through a forensic and quantitative argument rather than through just illustrative instances; (c) showing how the text under consideration must draw on the ‘meaning potential’ that a community shared at the same time that it can act as a mode for extending our sense in that potential; and demonstrating the profound plasticity of text – that language could, in its plasticity, serve all kinds of masters, to a great variety of semantic ends: it could be transforming; it could be revealing and extending of life choices; or it could be turned to values disastrous for a society – viz. the pursuit of nobility through gratuitous acts of arms (one possible dimension of T&C).

Consequently, Hasan’s stylistics (which took solid shape in her PhD work on four novels in 1964) was never swayed by the narrow focus of the Ohmann hypothesis of the generative era (with its fascination with optional transfor- mations as an index of mind style (1964, 1971)), or of the later Ohmann ‘speech act’ suggestion (as followed by Levin (1976) by prefixing lyric poems with: ‘I imagine a world in which ...’). With such partial and banal proposals, it is little wonder that so many literature specialists abandoned linguistic methods of the MIT school of stylistics for the smoke and suavity of Derrida’s ineffables (viz. grammatology over grammar). Nor was Hasan subject to the later fashionable extremes of some Reception theorists, who diminished or even nullified the author’s role in creating literary value: as if all texts were so similar in their organizing principles that they could be treated as equal in deserving our time and analytical attention. By contrast with these ‘en courant’ postures, Hasan’s questions were simply: what does the text mean? How does it mean what it means? And, why is the text valued as it is? And with that concept of ‘value’, we come back to Troilus in T&C. The whole play can be viewed as a complex system of semantic relations (a 30 Society in Language, Language in Society vehicle of plot; of tropes; and of wordings) in which the ways and words of the world offer no space to the flowering of goodness. The rational suc- cumbs to the rhetorical; acts of cupidity or cruelty are apparent but not declared; and neither victors nor vanquished bring credit upon themselves in this supposedly elevated, mythic, ‘theatre’ of war. Language, whether from the voyeurism of Lord Pandarus or in the sulk of Achilles, makes us wince. The world is ignoble and consequently malignant. It may be recognizable for us today as the world of sycophants and prejudicial advisors that we associate with tabloid-driven foreign policies and lobbyists, those who renominalize a catastrophic invasion as ‘Operation Iraqi Freedom’ and ‘Operation New Dawn’. The same issues were lampooned by the Athenian playwright Aristophanes, for instance in his portrayal of the populist, violent demagoguery of Cleon and Cleon’s promotion of ‘sycophants’ – originally, in Classical Greek, applied to those who would create ‘trumped up’ litigation against any prominent Athenian who stood against Cleon’s commercial and war-waging policies. One might claim however, that, in Shakespeare’s negative rendering of these ‘events’ from the Fall of Troy, there is another form of inversion. We can be driven to declare that there is more to humanity than is realized here in the play. The ‘absurd’ in the play builds up; it becomes unrelenting, venal, and uncomfortable and unfunny. And in becoming unrelenting, it can provoke a counterargument in the viewer, a cry akin to the line from Sonnet 65 and its ‘O fearful meditation’: ‘How with this rage shall beauty hold a plea ...’ But this is merely to suggest, not to analyse. Hasan makes a strong distinction between ‘explication de texte’ and stylistic analysis. Explication may be revealing – it may assist us to approximate to what a text means, and even offer some of the cultural context within which a text is valued; but it is not addressing the core of the three questions above if it shares no verbal evidence about what we are claiming to be an achievement in verbal art (see Lukin and Webster, 2005). The debates on human values and over humans as merchandise are echoed in key speeches in T&C, as we shall see below. But demonstrating this is not a straightforward process. A lexical survey through the much praised search engine The Works of the Bard ([email protected]) shows that the play does not stand out within the oeuvre of 38 plays for its use of the words ‘value’ and ‘valued’ (only the three instances of the exchange between Hector and Troilus: Act 2 Sc. 2). In fact, the statistics in T&C over the statistics for other plays and in relation to any other play are not supportive of my claims until the lexical item ‘worth’ is brought into the network of lexical connections. As we include ‘worth’ in the concordance net, the weight of evidence begins to incline towards the theme I have proposed: the play has 17 instances of ‘worth’ (out of 204 across the total oeuvre of Shakespeare, plays and poems). Measure for Measure, with its themes of justice and value, David G. Butt 31 has only 6; and Pericles, Prince of Tyre, is the closest play to T&C with 12 instances of ‘worth’. It does appear that the interpretive ‘logic’ of a theme on ‘value’ is also supported by the fact that The Merchant of Venice leads all the plays in relation to the use of the word ‘value’ (seven instances). The Merchant of Venice deals with another case of ‘weighing up’ of a human in barter, and also has seven instances of ‘worth’. When we search for cognate terms, however, the evidence for this semantic preoccupation begins to grow, in quantity and through the grammatical interactions or crossings of lexical connections. The terms cognate with the meanings of ‘worth/worthiness’ and ‘cost’ do change the weight of evidence in favour of my earlier claims, but not in a way that is unequivocal. When I switch to the SYSconcordance offered in SYSamanuensis (as developed by Wu and Matthiessen), my power to interrogate the lexical probabilities of the text is amplified dramatically (appropriate for the study of a play). Due to the power to gather in the interacting lexis in one operation – to gather numerous associative choices which make up a field of relevance – I am able to estimate the outcomes of lexical chain analysis, and the usefulness of capturing chain interaction as a method of quantifying ‘cohesive harmony’ (Hasan, 1968, 1984b; Cloran, 1995; Butt, 1985; Butt et al., 2010; Lukin, 2003; Khoo, 2013, see also Khoo, this volume). The findings for the 210 lines (c.208 clauses) of Act 2 Sc. 2 begin to emerge quantitatively as set out in the following:

1. There are 72 tokens that have some direct bearing on the semantic field of value, merchandise; revenue; cost; theft; property; and expense. These are treated as distinct from the complementary tokens of what is dear to a person as honour – there are 57 tokens that entwine with the c.70 cost/value/trade related terms. Given that cost is pitched against honour in the rhetorical battle between Hector and Priam on the one side (she is not worth what she doth cost the keeping), and Troilus and Paris on the other (we will be canonized), the separation of mercantile expense from loss of honour helps to reveal the dialectic in textual organization. When combined, however, there are at least 137 tokens building the topic of ‘value’ – that is, around 137 words in a shared ideational domain. Think of cost and dear as possible synonyms, and then as antonyms in this play, as in dear of ‘worldly expense’ against dear of ‘human feeling’. 2. Comparison of the cost/value/trade similarity chain with the other thematic or topical consistencies both supports and modifies my original claims about what is the preoccupation in the scene, and ultimately across the play: the division of the scene into lexical ‘phases’ (e.g. as in Gregory and Malcolm: see Malcolm, 2010, pp. 4–24) suggests c.10 phase shifts. These are marked off by an apparent change in lexical proportions – for instance, a mix of five topical sets might shift to a 32 Society in Language, Language in Society

focus on just two or three. This ﬁrst occurs, for instance, around clauses 9/10 after the ‘dread father ...’, King Priam, poses the question to Hector as to whether Helen should be returned to the Greeks. From clause 10 to 25, Hector and Troilus lay out the primary opposition of the text – that between value and cost on the one hand, and name, merit, honour on the other. Less obtrusive are the cumulative references to a chain of words referring to semiotic and rhetorical acts: question; reason; denies; mean;andwise. Clause 10 is the crossover, where the absurd references to digestion, spongy bowels, and toucheth my particular introduce the bodily and sexual terms which come in clusters across the phases of the scene, and which manage to render absurd the sententiousness of the adjacent arguments. 3. In terms of such phases, or signiﬁcant clusters, the ten subdivisions proposed here are:

A: the dithering king (clauses 1–9); B: the rhetorical agonists state their position (10–25); C: a third party (Helenus) demands reasons, and Troilus rebutts reasons for bodily lustihood and manhood, and also honour, all of which brings Hector and Troilus back into the explicit focus on cost, and worth, and where value resides (26–55); D: Troilus: his hypothetical argument (Today, I take a wife ...) and the irony of his covert, actual wife (56–80); E: Troilus on what was cried when Paris set out (81–100); F: The sudden and actual Cry Troyans of Cassandra (101–128); G: Hector opposes Troilus (Cassandra’s high strains of divination as opposed to brain sick raptures) as Priam challenges Paris (129–157); H: Paris’s response (158–170); I: Hector’s reprise of the whole argument, and his decision to pursue our several and joint dignities (171–192); J: Troilus’s coda and Hector’s transition to plot against the Greeks (193–208).

4. Early lines provide ironic portents of the later events, although some of these are unobtrusive or to be appreciated retrospectively: for instance, while Hector denies fear of the Greeks and refers to the spongy bowels of women who cry out what is to come, Cassandra is the one who deﬁantly and repeatedly cries out; and she is in no doubt as to what is to come. Her brain sick rapture (Troilus) resonates with the seizing (fair rape) of Helen: a play on different kinds of seizure (see also Jakobson on Shakespeare, 1987, p. 202). 5. Figure 2.2 gives a schematic view of the way these tokens accumulate logogenetically – that is, as the scene unfolds. Hasan’s technique of setting out referential Identity Chains against those that maintain sense David G. Butt 33

relations but not identity – that is, Similarity Chains – can provide an iconic and a quantitative chart of the texture (see Khoo, this volume, for a detailed treatment of ‘cohesive harmony’; Hasan, 1984b). The creation of lexical chains and the interaction of these chains (as measured by ‘re- echoed’ grammatical roles) was a technique ﬁrst envisaged by Hasan in her analyses of children’s accounts of events in the 1960s. It changed the power of cohesion measures up to a level that gives realistic measures of coherence, hence ‘Cohesive Harmony’. The setting out of chains and capturing their interactions are reliable ways of representing semantic oppositions between lexical items (e.g. cost vs honour) and the ‘crossovers’ of thematic tokens in nominal groups (e.g. fair rape; ransacked queen). Here I am only showing the way the lexis is recruited by Shakespeare to shift the semantic balance in the unfolding ‘argument’ of the play. The ﬁgure shows, for example, the strength of the emergence of words concerning persuasion: sayings, reasons, wisdom, knowledge, denial, confession ...as well as the ‘crying’ out of Cassandra. It also shows how bodily references are bunched to undermine the weight of argument (viz. King Priam’s in hot digestion of this cormorant war), or to provide sexual innuendo (from Hector: what toucheth my particular to Troilus on Paris’s voyage to Helen: He touched all the desired ports ...).

20 18 16 14 12 10 8 6 Number of tokens 4 2 0

10 (A) 25 (B) 80 (D) 55 (C) 100 (E) 128 (F) 157 (G) 170 (H) 192 (I) 208 (J) Clause number (phase) value/cost speech/reasons honour/personal merit body parts/processes

Figure 2.2 A schematic view of the logogenetic accumulation of tokens in Act 2 Scene 2ofTroilus and Cressida 34 Society in Language, Language in Society

If you enter http://www.fashionsofspeaking.com/construe-my-meaning. html, you will be able to access a selection of slides which illustrate steps of cohesion analysis and which, through colour coding, give some sense of how the different strands of lexical connectedness contribute to a cohesive texture. The latent power of this network of connectedness comes from the way ‘similars’, and tokens not so similar, multiply at a rate not easily tracked by a listener or reader, although sensed for their semantic consequences. One aspect of Shakespeare’s verbal artistry is the way such semantic resonances can be brought so effectively into the ‘symbolic articulation’, with thematically cohesive words worked into different permutations and concentrations from early to late in a play. This may have been an authorial virtuosity of the era, for instance learned from predecessors like John Lyly (e.g. his use of scratched faces and broken mirrors in the masque Campaspe). But with respect to extent, complexity, and variability of purpose, Shakespeare’s wordings bring cohesion into a higher realm of artistic tool power, with the cost–honour opposition of T&C being one example of how we experience a semantic dialectic through the allocation of the lexis: we experience thematic ‘dialexis’. The grammatical patterning of Act 2 Sc. 2 has particular relevance to the dialectic structure of the rhetorical struggle between the characters, as well as the shifts of perspective that produce the ﬂuid irony of the play for the audience. This is particularly the case at the lexical nub of the cost/value/trade tokens: the deﬁnitional debate between Hector and Troilus (clauses 48–55), which precedes Troilus setting off on his hypothetical exemplum (I take today a wife ...).

HECTOR. (48) Brother, she is not worth [[what she doth cost The keeping.] . Voc. Carrier Attribute [[ Identifying: i.e. she doth cost X ]] Relational: Attributive TROILUS. (49) What’s aught (50) but as ’tis valued? (Anything IS (only) [[as it is valued]]) Reversible Identifying: ‘(Only)[[as it is valued]] IS anything’. Consequently: A thing’s identity IS its value – or paradoxically: A thing’s quiddity IS (represented by) its valuation Value Token (‘is represented by’: thus encoding)

Among the complexities here is that there is a grammatical ‘pun’: rather than ﬁnding that we can conceive of value as a thing’s role or place in the scheme David G. Butt 35 of things (see Halliday and Matthiessen, 2014, ch. 5), and that this role will be made manifest through some available token of what is more abstract in the value, the reverse is suggested by Troilus – namely, that anything is manifest to us, is made real, only by the valuation which it carries. This is a deep semiotic absolutism which Hector contradicts with great animus.

HECTOR (51) But value dwells not in particular will: Carrier Attribute Relational: Attributive: dwells not is ontological = ‘is not’ (52) It holds his estimate and dignity Carrier Attribute (with the two lexical themes of cost and honour) Relational: Attributive: holds is closest to ‘has x and y’. (53) As well wherein ’tis precious of itself As in the prizer. (It is (as) precious of itself as ( ...) in the prizer.) Carrier Attribute [[ ...it is ... ]] (54) ’Tis mad idolatry [[To make the service greater than the god;]] It [[to make ...god]] is mad idolatry Carrier: (not Id.) Attribute (55) And the will dotes [[that is attributive To [[what infectiously itself affects]], Without some image of th’ affected merit.]]

Here the attribution is in the embedding – (gloss: The will loves foolishly (dotes) if it attributes quality that infects the thing idolised, that is, without some vision of the actual merit of the thing so affected): (55) And the will [[that is attributive To [[what infectiously itself affects]], Without some image of th’ affected merit]] DOTES. The following discussion takes up some implications of the lexical ‘stitch- ing’ and lexical phases of T&C before moving over to the syntagmatic potential of latent patterning in the background effects of clause complexing, with illustrations from Shakespeare’s sonnets (in section 2.3). The linguistic regularities in a play or in any plot-based narration take on their value within the polyphony of events and of the characters to whom the responsibilities of ‘utterance’ are allocated. This is to point out that, beside the deeper theses that may be derived from the tracking of ‘subliminal patterning’ (Jakobson, 1987, ch. 14), the ‘symbolic articulation’ of Hasan’s schema also applies to symbolic articulation in the literal sense: who says what to whom and under what circumstances. We must reflect on the meanings in the specifics of action. With this in mind, we can consider how the ironies of Troilus’s moral exemplars and figures of speech diminish his own status in the very act of their enunciation. We know the historical 36 Society in Language, Language in Society outcome; we see that a community will pay for the canonization of nobility that Troilus’s adventurism extols. The problem of this ‘problem play’ crystallizes in the hypothetical problem Troilus poses to Priam, to Hector, and to the Court in order to argue for the retention of Helen. The day after his sexual contact and lover’s contract with Cressida, he launches in with a rhetorical flourish. The heuristic, hypothetical tone he adopts in his central speech must be turned back on him by the audience as they witness the later action of the plot: Cressida is given over to the Greeks, uncontested, like a wordless chattel in the bargaining between noblemen and warriors. Yet the irony is gentle as Troilus begins in Act 2 Sc. 2, with our thought that the words are not hypothetical for us as we know what Priam, Hector, and the rest of the Court do not: that he has just taken ‘a wife’ for whom he had expressed great ardour.

I take today a wife, and my election Is led on in the conduct of my will; My will enkindled by mine eyes and ears – Two traded pilots ‘twixt the dangerous shores Of will and judgement – how may I avoid, Although my will distaste what it elected, The wife I chose? There can be no evasion To blench from this and to stand ﬁrm by honour. We turn not back the silks upon the merchant When we have soiled them; nor the remainder viands We do not throw in unrespective sieve Because we now are full ...

What to make then of the dissonance between his avowals and his action, between all heroic reputations and the sordid or small-minded behaviour, between pathos and bathos, and between tragedy and farce? As a linguist challenged by the play, I cannot but relate the revaluing of all values in the play as a literary precursor to two modernist influences: firstly, on the one hand, to Nietzsche’s declared project – the revaluation of human values in a world without God. Shakespeare’s Troy and the Greek encampment are represented as a world in need of a moral compass, and without authentic motivations for action. On the other hand, there is the semiotic and systemic idea put forward by Saussure (1978 [1916]): namely, that in a semiotic system all the terms of the system are relational – they exist only because of the differences they make in relation to the totality of signs in the system. Actions and words have value, not because of the way we see them, but in the way they stand with respect to other acts, words, and meaningful behaviours. There is no ontological anchor beyond an intense reflection on what Hasan refers to simply as ‘the living of life’, much as Wittgenstein also concluded through his investigation of ‘philosophical David G. Butt 37 grammar’ – value is made through community life forms: lebensformen (Glock, 1996, pp. 124–9) ...end of story! A powerful demonstration of this relation between patterns of life and the community grammatical system is Hasan’s essay (mentioned above), ‘Ways of Saying: Ways of Meaning’ (originally presented in 1975, and published in 1984: see Hasan, 1984a [1996a]). In this extended comparison between the semantic style of English in Britain and Urdu in Pakistan, Hasan shows how expectations one can have about social structure and particular roles in Pakistan could supply references which, as a consequence, take semantic pressure off the grammatical function of an explicit subject (which one would expect in English). A similar form of cultural pattern supports the use of what might be analogous to the use in English of the restricted exophoric, or the definite article (you know THE one). The patterns of cultural order work to motivate, or diminish motivation, for the explicitness in the code. This becomes especially relevant when we come to section 2.4 below where we briefly consider the reception of Shakespeare in English-speaking communities which simultaneously draw on traditions of non-British and non-European cultures. It is useful to add, at this point of apparent linguistic and cultural relativism, that there are questions now arising in various domains of neu- rosciences in which debates about ‘value’ run down to the most ancient and supposedly most basic elements of human and evolutionary systems: where is the engine for the value systems? What is the biological primitive of value? (Damasio, 2012, p. 47). Such questions engage with the legacies of our ancient, emotive, and motor-based subcortical systems (see section 2.4 below). Setting out from Saussure’s linguistic concept of ‘valeur’, since this is a benchmark for the relationalism behind Hasan’s ideas, we can envisage the whole play as a metaphoric system: T&C. Shakespeare takes the acme of heroic signification and calibrates it against a sordid display of human strivings and motivations, all buttressed by logically plausible (if irksome) rhetorical strategies. These pit nobility against trade and mercantilism. As Troilus remonstrates:

If you’ll avouch ’twas wisdom Paris went – As you must needs, for you all cried ‘Go, go’; If you’ll confess he brought home worthy prize – As you must needs, for you all clapped your hands And cried ‘Inestimable!’; why do you now The issue of your proper wisdoms rate, And do a deed that Fortune never did, Beggar the estimation which you prized Richer than sea and land? O, theft most base, That we have stolen what we do fear to keep! (Act 2 Sc. 2, ll. 84–93) 38 Society in Language, Language in Society

There are then various reasons why, as Hasan urges, the analysis of literature is such a useful preparation for the development of studies in linguistic theory (Hasan, 2011, foreword). Commonly in literature, two units with morphological and semantic afﬁnities (from the perspective of the total linguistic system) may take on contrastive ‘valeurs’. Such marked values are motivated by the themes instantial to the emergent metaphoric goals of the ‘drift’ in authorial choices. This is to say, similar values at the level of system may be turned to quite contrastive purposes through the construction of action, imagery, or other regularities in the system instantiated by the text. For example, from T&C, the notions of merchant and thief are cited at crucial points in the exchanges between Troilus and all other characters. The terms are not simple in the relations of this text, however. Merchant can be both positively marked or unmarked in its role in the imagery (as in the merchant and silks cited by Troilus in the ﬁrst passage of Act 2 Sc. 2 above); it is also used as a derogatory appellation in the way kings are turned to merchants bartering for Helen’s return (Act 2 Sc. 2, ll. 82–3). In the system of the play, the words thief, stolen, and rape are not negative terms, even in the environment of pearls, merchants, women and queens. The notion of something stolen is inverted in the debates between Hector and Troilus and Paris – the lexis for what Troilus is seeking to defend is explained as a theft that is dishonourable if the stolen item be returned:

...O, theft most base, That we have stolen what we do fear to keep! But thieves unworthy of a thing so stolen, That in their country did them that disgrace We fear to warrant in our native place!

The problem is not the theft, but not defending what one has stolen from another country since, as Hector ultimately concedes:

...’tis a cause that hath no mean dependence Upon our joint and several dignities.

Whereupon Troilus responds:

Why, there you touched upon the life of our design: ... She is a theme of honour and renown ...in time to come canonize us;

Along with the arguments of Paris – that he would have the soil of her fair rape/Wiped off in honourable keeping her (clause 159) – the inversions are cumulative: theft is justified; honour and our several dignities make necessary David G. Butt 39 the retention of that which is stolen; the act will lead to canonization (both saintly and literary); and, despite a law in each well-ordered nation / to curb those raging appetites that are / Most disobedient and refractory, Hector concedes to his sprightly brethren (Troilus and Paris), rather than that they bend to his more mature counsel. At points, the blending of semantic types shows how far there is an idiom to the play – as when Troilus rejects the cries of Cassandra: Her brain-sick raptures cannot distaste the goodness of a quarrel (clause 134). There is a further, awkward irony for the audience here in the historical vindication of Troilus’s claim to canonization. He and Hector truly are ‘saints’ of literary renown – through the Iliad, and Chaucer’s Troilus and Criseyde, the characters have become the topoi of a continental culture, not just of English literature. The concepts of value and mercantilism are brought into an apparent resolution by a standard trope, namely that promised glory is a rich advantage that Hector could not lose for the wide world’s revenue. The sonnets often use this trope – that the reputation or fame bestowed by verse will outlast the mutability of even brass or stone, and the blast of natural forces. Such a response to any fearful meditation on time is not satisfying for a viewer in this play, however. The protagonists have been ostentatious and self-absorbed in their aspirations and certainly unmindful of Every tithe-soul ’mongst many thousand dismes who has been lost since the first sword was drawn about this question (2.2. ll. 18–25). The disconcerting effects of claims to noble aspirations and glory are a part of a ‘consistency of foregrounding’ that extends from the semantic dissonance between speech and character; between argument and action; and between the ideas fused in images (tropes and enthymemes) and schemes (lexicogrammatical structures at the rank of clause and group). Some dis- parities of argument and action I have addressed. But the dissonances at the smaller scale (of group and word) are an accumulating aggravation to the sense of value or tragedy in the play. The use of soiled silks and wip- ing off the soil of the fair rape of the ransacked queen; the collocation of sex and viands (meat), and the numerous ‘double entendres’ of this and other scenes, all go to undermine the claims of perpetuity through literature that are a topos of Shakespeare’s work and of the Elizabethan milieu. The ‘symbolic articulation’ – the full mobilization of semiotic resources – is towards farce, not tragedy. The ‘entelechy’ of Aristotle’s fourth form of cause – the emergent goal directedness of material – is towards the absurd, and possibly to the end of cynicism (as suggested in the Jonathan Miller directed version (1981), part of the BBC’s The Shakespeare Collection 1978–85. In presentation of this chapter, I have shown Act 2 Sc. 2 from this intricately directed production). But it is not cynicism, in my view at least, that a clever and sophisticated young woman like Cressida (see her repartee with Pandarus at their first meeting in the play), when given over to the Greeks peremp- torily and without any rhetorical effort from her supposedly ardent lover, 40 Society in Language, Language in Society should work out some new ‘modus vivendi’ under the admiration of her new masters, especially with the particular protection of the ambassador Diomede. There may be a suggestion in the play, then, of the quotidian reality beneath titanic events – the petulant Achilles; the wily Ulysses; the mediocrity of kings. Is this so different from what historians have revealed of the petulance and struggle for kudos between the protagonists (or ‘alpha males’) of modern warfare and political leadership: Montgomery, Rommel, Patton, and Eisenhower; Churchill and de Gaulle; McArthur and Truman; Zhukov and Eisenhower? According to Roman historians Livy (XXXV.14) and Appian, there was a meeting in Ephesus between Hannibal (by then in exile from Carthage) and Scipio Africanus, the Roman who burned and razed Carthage in 202 BCE. The two generals debated who were the greatest generals of all time. Both agreed on the primacy of Alexander of Macedon, and Hannibal won a battle of wits over second and third places. Such historical anecdotes remind me that consequential actions and decisions are typically made within a welter of envy, bitter argument, and spiteful com- petition; and that it may be that Shakespeare’s Ilium is not so unreservedly farcical, but an experiment in values amid contests of ambition, ‘dignitas’, and rhetoric. From this reflection, I will move on to one further act of ‘double-seeing’ (Bateson, 1980, ch. III), one that further complicates the play’s interpretation: by reworking the episode around Troilus and Cressida, Shakespeare inevitably animates the myth anew. He extends its regenerative scope or ‘meaning potential’ (to invoke Halliday’s perspective on a semiotic system). Shakespeare’s symbolic virtuosity taunts the views of those who preceded him, and challenges those who might take up the theme thereafter. Like military generals, artists may experience what Harold Bloom (1973) characterized as the ‘anxiety of influence’: artists work in the long shadow cast by their predecessors, as well as in anticipation of later generations of virtuoso writers.

2.3 Poetry: the silent legislation of thought through language

While the ensemble effects of the ‘symbolic articulation’ in the play need to be pursued across the tropes, lexis, and grammatical cruces of the text, as well as across the quasi-pragmatics of dramatic action and character, there are also the less overt resources of meaning which work as latent patterns in verbal art. These include choices we appear to ‘take for granted’ as part of a necessary, or natural, background to saying anything at all. They become then among the least consciously controlled of linguistic regularities; and they arise where the effects of patterns are extremely plastic in their semantic consequences, that is where arrangements in the lexicogrammar are little attached to speciﬁc semantic effects, or are little associated, probabilistically, with speciﬁc textual effects. David G. Butt 41

In relation to latent patterns, we might begin with the many ways that the phonological minutiae of a line can be distributed and varied against the arguments and semantic progression in a text (Cléirigh, 1998). The variations and alignments of phonology can be generalized into the vague collective notion of ‘cadence’. Alternatively, we can take up the individual patterns of pitch, metrics, and rhythm (prosodies) through the phonological rank scale of tone group, foot, syllable, and phoneme. (This is not, in SFL, the Daniel Jones universal phoneme, but is better deﬁned as the smallest relevant unit of phonological order for the language under investigation: as Halliday has demonstrated in Mandarin, this is better aligned with what in English would be a syllable. See Halliday, 2005, ch. 6.) If I now shift my discussion to Shakespeare’s sonnets, such unconscious patterns of organization, whether of lexicogrammar or phonology, accrue greater responsibility in the artistry of verbal art. And I will turn to illustrate the meaning potential of certain of these resources. I will set out from the case of alignments between

1. The conventions of the sonnet, including the formal requirements of lines, rhyme, and stanzaic organization (all of which are subject to conscious deliberation and commitment); 2. The semantic structures of argument, petition, and other rhetorical moves (which may be self-consciously drawn from traditions of literary and rhetorical style); and 3. The clause complexing, and some implications of the clausal arrangements for tone groups.

My discussion for this chapter will need to draw attention to the demands of a sonnet, that is, the alignments of (1) and (2). But it will focus on the clause complex in (3), and offer suggestions about the significance of alignments with tone groups. One consequence of the ‘backgrounded’ status of clause complexing and tone groups in language is their low visibility as choices at all. Slight manipu- lations of their forms are rarely addressed as central meaning-making options (and are sometimes overlooked even by linguists!). Certainly, there is little idea of fixity between options of construction, say of clause complexing, and aesthetic effects. There is a much discussed exception to this claim – projecting clauses in reporting speech are crucial to narratology. Both linguists and literary specialists have considered an array of forms that purport to grade evidentiality, subjectivity, focalization, and shifts in authorial distance. Still, Voloshinov (1929 [1986]) set the ‘implication of utterance’ as a challenge that structural linguistics would never adequately meet; and it is still chimerical and very much work in progress for stylistics and linguistics. Before proceeding to some examples of sonnet organization, the idea of linguistic ‘plasticity’ needs further explanation. Hasan herself emphasizes 42 Society in Language, Language in Society the way language can be servant to so many purposes, thematic and ideological; language is not inherently good, bad, or neutral (2014). Realiz- ing linguistic meaning commits a speaker or writer to choices, and certain of these will appear most responsible for the uniqueness of a text. The Russian Formalist critics gave such a factor the title of ‘dominant’ – the system(s) most responsible for the individuated character of the text. But all spoken texts need to build from fundamental resources of tone groups and clause complexes. And small variations within the selections of intonation or taxis between clauses work like an unobtrusive but insistent ‘tide’ in the logogenesis of any text. In verbal art and rhetorically heightened, emotionally charged texts, these effects take on a ‘dominant’ role. Few analysts allocate sufficient attention to this cumulative orchestration, with its possibilities for pacing (accumulation; arrested progress; dynamic tensions and their release into cascades of qualifying elaboration), its opportunities for congruence (e.g. between the boundaries of tone group and clause), for counterpoint (e.g. non-alignments between units of grammar and of information/intonation); and for ‘making strange’, for instance by extended complexity, sometimes as an analogue of psychological intensity; or by paratactic aggregation, which can build up a text through a form of linguistic ‘pointillism’. Recent work by Webster (in Halliday and Webster, 2014) on a debate concerning hypotaxis and parataxis in Obama’s presidential speeches breaks new ground in how clause complexing might be monitored in logogenesis. This also adds to the scale and discrimination that may be achieved by new applications of earlier work between taxis in grammar and rhetorical structure by Matthiessen on clause complexing and rhetorical structure (e.g. 2002) and Mann, Matthiessen and Thompson (1992); it adds to work by Martin on ‘intricacy’ (1992); by Tuckwell on ‘emplexity’ (1999); and analyses by Butt, using Hasan’s ideas of ‘symbolic articulation’ and ‘consistency of foregrounding’ on clause complexing and the ‘existential fabric’ of texture (1985, 1988). The contribution of SFL linguists to the interplay between clauses and tone groups can be tracked through the work of Halliday (1985), Halliday and Greaves (2008), as well as, for example, El-Menoufy (1969), Davies (1992, 1993), Tench (1992), Vonwiller and Cleirigh (1997), Webster (1998), O’Grady (2013), and Bowcher and Smith (individually, and in the edited volume 2014). It is with respect to the unconsciously elaborated resources of taxis and alignment that an actor is most empowered in the role of interpreter: the writer of a play, a poem, a political speech, or any public utterance, can set the wording but not its delivery, can compose the phrase but not its acoustic ‘phrasing’. The Greek word ‘hypocrisis’ referred to the ‘orator’s delivery’, and directs me to suggest that if language cuts us ‘to the quick’ – to our living core – the human embodiment implied by the way language is offered to us, through the interface of semantic and phonological rhythms, will carry our degree of conviction. David G. Butt 43

Classical rhetorical training, and its medieval and Renaissance forms, did include being alive to the force of allocating information to paratactic as against hypotactic structure (e.g. Lanham, 1991, p. 87). My focus will now turn to a small selection of sonnets in order to demonstrate alignments (and non-alignments) between the lines of organization set out above, namely: (1) the units of generic convention – the expectations the reader may have of the sonnet form; (2) the argument structure or rhetorical progression of the sonnet (remembering that the Italian form of the sonnet appears to have been derived from the dialectic of a legal advocate); and (3) the way that the two levels of generic and semantic organization ((1) as to (2)) are laid out against the equalizing or subordinating options of clause complexing (1 : 2 : 3). Comments on alignments with a fourth level – prosodic patterns – cannot be detailed in addition here, though it is exactly such variation (congruence or counterpoint) which affords us the commonplace wisdom that every reading is a distinct interpretation. Firth (1957) reminds us that ‘meaning is made at all levels’, that is, as the ‘product’ of ‘patterns and processes’. Let us turn then to some ‘canonical’ cases. By investigating a small sample of Shakespeare’s sequence of 154 sonnets, a sample taken up in an enquiry into the motif of ‘mutability’ – viz. sonnets 23, 29, 30, 60, 64, and 65 – we can consider a number of theses about the sonnet form and the subliminal effects of latent patterns of clausal tectonics. The ﬁrst resource to note might be called semantic intensity or mod- ulation – namely, the rate and regularity at which units of information challenge the reader for integration into the forward movement of the texture. For example, parataxis creates an aggregate effect that can be – or must be – managed in the sequence at which it comes to the reader. Hypotactic accumulation involves other choices (see Table 2.1). Information can be pri- oritized in dependent clauses with the main messages held over, thereby giving the hypotactic clause a delayed value or status (i.e. creating a tension: ‘How will this be relevant?’).

Sonnet 30

(1) When to the sessions of sweet silent thought I summon up remembrance of things past ||, PROBLEM (APORIA) (2) I sigh the lack of many a thing [[I sought]] ||, (3) And with old woes new wail my dear time’s waste: || (4) Then can I drown an eye( [[unused to ﬂow]]) CUMULATIVE COMPLAINT For precious friends [[hid in death’s dateless night]],|| (5) And weep afresh love’s long since cancell’d woe,|| 44 Society in Language, Language in Society

(6) And moan the expense of many a vanish’d sight|| (7) Then can I grieve at grievances foregone || RESULT (8) And heavily from woe to woe tell o’er The sad account of fore-bemoaned moan, || (9) Which I new pay|| (10) as if not paid before || (11) But << (12) if the while I think on thee>>, REVERSAL dear friend, All losses are restored || (13) and sorrows end.||

Table 2.1 Logical analysis of Sonnet 30

1α [[ ]] 2 3[[ ]] [[ ]] 4 5 1 2α 3α 4 1β 2βα 3β 2ββ Problem Complaints Result Reversal

Sonnet 30 offers us a ‘benchmark’ for the work done through the Shakespearean form. It begins at a Dilemma (When ...|| I sigh ...); it proceeds through an accumulation of Complaints (clauses 2, 3[[ ]][[ ]], 4, 5); and arrives after the first 8 lines (2 quatrains) to a ‘volta’ or ‘turn’ in the semantics. In the last 6 lines, we have a Result in 4 clauses (Then can I ...|| And heavily ...|| Which I ...||As if not ...). And, in the final couplet, the malaise of the first 12 lines is not resolved, but opposed by a Reversal of Feeling (‘But if the while I think on thee, dear friend, / All losses are restored, and sorrows end.’).

Sonnet 64

(1) When I have seen || (2) by Time’s fell hand defaced The rich proud cost of outworn buried age;|| (3) When ...|| (4) sometime lofty towers ...<>...down-razed || (5) And (E ...) brass eternal slave to mortal rage;|| (6) When I have seen || (7) the hungry ocean gain Advantage on the kingdom of the shore,|| (8) And the ﬁrm soil win of the watery main, || (9) Increasing store with loss and loss with store;|| (10) When I have seen such interchange of state,|| (11) Or state itself confounded,|| (12) to decay;|| David G. Butt 45

(13) Ruin hath taught me thus || (14) to ruminate: || (15) That Time will come || (16) and take my love away. (17) This thought is as a death,|| (18) which cannot choose || (19) But weep || (20) to have that [[ which it fears || to lose.]]||

Table 2.2 Logical analysis of Sonnet 64

αα β1αβ2αβ3αβ4αβ5αβ1 β2α β1ββ2β1 [E]β2β2 β3β1 β3β2αβ4βα β5β1 β5β2 β2β[[ || ]] β3β2β1 β4ββ

By contrast, Sonnet 64 arrives at a Paradox, in this case an Aporia: a state in which the resources of humans run out. By having love, the poet is subject to the thought which ‘is as a death’, and so weeps ‘to have that [[which it fears|| to lose]]’. The clausal architecture is more complex (20 clauses over 4 depths, with 18 clauses subordinate) (see Table 2.2). There are then two main clauses – ‘Ruin hath taught me thus to ruminate ...’ (line 11) and ‘This thought ...’ (line 13). The further three depths of hypotaxis do not involve embedding up to the final words – the final nominal group: ‘that [[which it fears || to lose]] ||.’ Rhetorically, the first eight lines set out an aggregate of observations (‘When I have seen ...’ ×3): the first are political, then global. After the ‘turn’, the principle behind the observations is ‘seen’: the ‘interchange of state’ and its personal consequences. This in its turn leads to the Paradox – a form of personal principle that anyone in love cannot escape.

Sonnet 60

(1)Like as the waves make towards the pebbled shore,|| (2)So do our minutes hasten to their end;|| (3)Each changing place with that [[which goes before,]] || (4)In sequent toil all forwards do contend.|| (5)Nativity, once in the main of light, Crawls to maturity, || (6)wherewith being crown’d, || (7)Crooked elipses ’gainst his glory fight,|| (8)And Time [[ that gave ]] doth now his gift confound. || (9)Time doth transfix the flourish [[set on youth ]] || 46 Society in Language, Language in Society

(10)And delves the parallels in beauty’s brow, || (11)Feeds on the rarities of nature’s truth, || (12)And nothing stands || (13)but for his scythe to mow: || (14)And yet to times in hope my verse shall stand, || (15)Praising thy worth, despite his cruel hand. ||

Table 2.3 Logical analysis of Sonnet 60

1α 21α 2<< >> 1[[]]234αα 1β11β2[[ ]] 1β<< >> 4ββ

Sonnet 60 involves an Analogy (waves: minutes) and then builds a quiet, paratactic aggregate of cases involving the relentless haste of Time (see Table 2.3). Against this three-quatrain accumulation, the couplet comes as a humble, Adversative suggestion about what ‘shall stand’. Though the argument is held back to the couplet, it is plausible since it is an expression of hope, not of ostentation. Sonnets 23 and 29 further display the range of complexity between rhetorical and clausal constructions, but also might be thought of as relatively unequivocal rhetorical ‘postures’. In 23 (not set out below) the Analogies with an ‘unperfect actor’ or with ‘some ﬁerce thing’ are employed in a carefully qualiﬁed Apology for what the lover has forgot ‘to say’. Qualifying information interrupts the progress in the sonnet. The similitudes lead on to Petitions: asking the lover to accept the books as eloquence and to ‘hear with eyes’. By contrast, in Sonnet 29, the cumulative structure is employed up to the Volta; and then the last six lines incorporate a Reversal of the woe through analogy and hyperbole (see Table 2.4).

Sonnet 29

(1)When, in disgrace with fortune and men’s eyes, I all alone beweep my outcast state || (2)And trouble deaf heaven with my bootless cries || (3)And look upon myself || (4)and curse my fate,|| (5)Wishing me like to one more rich in hope, || (6)Featured like him, || (7)like him with friends possess’d, || (8)Desiring this man’s art and that man’s scope,|| (9)With [[what I most enjoy]] contented least; || (10)Yet in these thoughts myself almost despising, || David G. Butt 47

(11)Haply I think on thee, || (12)and then my state, << (13)Like to the lark at break of day arising>> From sullen earth, sings hymns at heaven’s gate; || (14)For thy sweet love [[remember’d]] such wealth brings || (15)That then I scorn || (16)to change my state with kings. ||

Table 2.4 Logical analysis of Sonnet 29

1234 5α 6α 7α[[ ]] 4β 4β24β3α 4β4α 5β 6β 7βα 4β3β 45β4β 7ββ

The power of Sonnet 65 demonstrates the effect of a ‘paratactic insistence’ in the response to the two disturbing questions that open the poem. The Interrogative is set aside with the arresting or punctiliar effect of the Apos- trophe: ‘O fearful meditation!’ This is one more variant on the linguistic resource under examination – the dynamic effects of tightening, releas- ing, and punctuating the flow of rhetorical moves. The opening quatrain builds to the question around time and change: ‘How with this rage shall beauty hold a plea, ...’. The second quatrain begins with a question which is then qualified by three subordinate clauses at one depth. The Problem is then restated in three paratactic clauses, and answered by the parataxis of the final couplet. This is to say the strategy of accumulation, so often the way of building in the first eight lines, is adopted in the final six. In fact, the poem is dominated by steady steps of parataxis since the hypotaxis of the opening lines involves two main clauses with five clauses at the same first depth of dependence (thereby retaining some additive impression overall) (see Table 2.5). This is also prepared for us in the group structure of line 1.

Sonnet 65

(1)Since brass, nor stone, nor earth, nor boundless sea, But sad mortality o’er-sways their power, || (2)How with this rage shall beauty hold a plea, || (3)Whose action is no stronger than a ﬂower?|| (4)O, how shall summer’s honey breath hold out Against the wreckful siege of battering days,|| (5)When rocks impregnable are not so stout,|| (6)Nor gates of steel [E] so strong,|| (7)but Time decays?|| 48 Society in Language, Language in Society

(8)O fearful meditation! || (9)where, alack, Shall Time’s best jewel from Time’s chest lie hid? || (10)Or what strong hand can hold his swift foot back? || (11)Or who his spoil of beauty can forbid? || (12)O, none, unless this miracle have might,|| (13)That in black ink my love may still shine bright. ||

Table 2.5 Logical analysis of Sonnet 65

αα 111212 β1 β2 β1 β2[E] β3

The total aggregation weighs against the couplet: there is a dramatic tension for the reader as to whether the poet can turn around the relentless build-up of the ‘wrackful siege’ and ‘batt’ring’ of Time by any ﬁnal, brief counterpro- posal. The answer comes in Paradox: ‘O none’, unless the rage is opposed by the miracle of ‘might’ in the power of ‘black’ to ‘shine bright’ – in the form of ‘ink’. The sonnet has a clear ﬁgural outline, like a legal or philosophical dialectic. One side builds a case, and this is opposed by a singular possibility of last resort. The dialectic ends in the human condition: to oppose ‘decay’ with an act of meaning.

2.4 Reprise

The clause complexing of a text may seem to be the least paradigmatic of the arrangements from which a speaker can choose (for the reasons suggested above). But I am arguing that such arrangements constitute a profound, even visceral, dimension of the embodiment of language. The flow, the halting, the interpolating, the qualifying, and the varying of grammatical depths available to the impassioned speaker, along with the acoustic rhythms that enunciate the realizations of wording, are the most direct extension of unconscious bodily systems into the otherwise conventional architecture of natural language. This connectedness of experiential ‘flow’ with semantic process can be pursued in the recent syntheses of evolutionary knowledge in the work of Panksepp and Biven (2012) and Porges (2011) and in the Conversational Model of Psychotherapy as elaborated by Meares (2012) and Butt et al. (2012). They provide a platform from which I suggest that the unequivocal biological implications of linguistic sounds also extend to these semantic units. The clause-to-clause patterns provide the congruence and counterpoint of the texture which is the heart of human communication. In Halliday’s methodical mapping of taxis against semantic David G. Butt 49 relations of apposition, expansion, qualification/extension, and projection, we have another way of understanding the insight of the great psychologist William James (1842–1910), namely that our experience is made up of ‘flights and perchings’: points of static conceptual recognition versus rapid transitions. While such a contrast might be likened to the difference between entity and process (and perhaps following from these, between ‘noun-ness’ and ‘verb-ness’ respectively), it also captures the reciprocation from Given over to New; Theme to Rheme; and from paratactic aggregates to hypotactic dependencies. This reciprocation assists in constructing and realizing the rhythm of empathic intersubjectivity of psychotherapy (see Korner, 2014, on physiological measures and the semantics of the Conversational Model of Psychotherapy). So, while the Saussurean line of arbitrariness runs between the segmen- tal phonology and the lexicogrammar, there is an impact from the prosodic characteristics of sound and their interface in the complexing of units in the grammar and semantics that has the power of what the semiotician C. S. Peirce (1839–1914) called ‘indexical signs’ – signs that have a natural dimension, a dimension that connects the expression of the higher orders of cultural meaning, whether in poetry or ‘flights’ of passionate discourse, to the body. Verbal art takes up these dimensions as a resource.

2.5 Semiotic distance and cultural context of reception: Indian, Chinese, and Japanese adaptations of Shakespeare

In analysing the differences between the context of creation and the context of reception, Hasan has emphasized an issue which has a clear initial outline, but has subtle, even covert, reactances for the interpretation of text. In verbal art, the subtleties double up due to the ‘double articulation’ of the system and the primary and secondary contexts. With an internal context (the context created by the terms of the composition) and an external context (the potentially strong directive of social and artistic expectations on the composer), there are two forms of distance to be considered: the distance between the milieu of the composition and that which is the lived experience of the interpreter; and the distance between the conventions and world view of the composer and the culture of the interpreter. The double articulation (see Figure 2.1) means that any choice on any level of verbal construction may be motivated from the ‘terms and conditions’ of the arti- fice alone, or from the conditions that pertained to the act of artistry that led to the emergent meanings of the text – viz. Shakespeare may have truly been moved to persuade a young man to marry and pass on his image to future generations (sonnets 1–17); or John Lyly may have bent his art to ensure his place at court, that is over and above his place in any putative canon of artists. More direct connection to social context can be seen in the poem of Sir Thomas Wyatt (1502–42), ‘Whoso list to hunt’ – a dangerous 50 Society in Language, Language in Society love poem that finishes with the Latin injunction Noli me tangere (It is for- bidden to touch me) since the woman addressed was Anne Boleyn, who was to become the queen to Henry VIII. There is no absolute boundary between intrinsic and extrinsic determinations of the form of a work. English novels began with the nascent sociology and narrative of Defoe’s Journal of the Plague Year; and returned to a quasi-sociological report in Mailer’s Armies of the Night. The spectrum of possible motivations and choices in our aesthetic responses to language is complicated; but, as Hasan has shown, this does not mean that the complexity cannot be clarified systemically and methodically, even if never to be utterly corrigible. Further, according to Hasan, if we are to claim a teacher’s role with respect to literature and verbal art, we need to be able to address the central claim of the activity: namely, ‘wherein lies the artistry?’ If we do not commit this question to rational and rigorous examination, we are in danger of being left with historical enquiries and hegemonies of ‘taste’ and sensibility. Consider T. S. Eliot’s negative judgements of Milton, of Blake, on metrics, and his speculative claims on the ‘dissociation of sensibility’: all narrow, tendentious evaluations of the poetry of others (see for example Steele, 1990 on Eliot’s claims about metre). Jakobson emphasized this issue in 1928, when he noted that instead of a science of poetics, we have created a plethora of cottage industries (O’Toole and Shukman, 1978). Furthermore, he characterized the literary studies of his era as a murder scene where police take in all the contingent factors, even arrest the chance passers-by, but overlook the perpetrator. The aim of this chapter has been to communicate the kind of approach that Hasan encouraged and created in her students and colleagues, namely: go to – engage with – the artistry of verbal art; and do not teach through the hegemonies of taste. Part of these directives is to be methodical about semiotic and semantic differences: examine the cultural parameters and grammatical systems through which a work is interpreted, as well as those in which it may have been constructed. Many communities have engaged with Shakespeare. There are those who must adapt his plays linguistically and contextually, as in the case of Kurosawa’s monumental adaptation of Lear to the context of warlords in Ran (1985), and the earlier reworking of Macbeth, Throne of Blood (1957). More recent, there is the phenomenon in England of directors from Indian communities taking up the plays of Shakespeare with the gusto of those who feel that they have a background that is closer to the values and contexts within which the plays are set: the loyalties assumed; the legal and cultural interpretations of responsibilities and order in family or society (see the list of Asian Adaptations and Critical Approaches to Shakespeare listed separately in the References). This is a domain for which Hasan offered clear theoretical clarifications in the 1970s, both in relation to verbal art and educational David G. Butt 51 systems (Hasan, 2011, chs 8, 9, and 3, originally published in 1976, 1978, and 1985 respectively).

2.6 Coda

Latent patterning is critical in the double articulation of verbal art – whether it be the subtle tides of lexical domains, running together or in opposition, or the syntagmatic variation of semantic pacing through complexing of clauses and groups. These resources, not so easy to monitor consciously in the ‘time of the text’ (logogenesis), are signiﬁcant contributors to what Hasan refers to as ‘symbolic articulation’ in verbal art, the mobilization of the linguistic system for a ‘consistency of foregrounding’ which moves us towards a thesis or theme constellation (Hasan, 1971, 1985). Through such consistencies and their mapping onto sound at one pole and onto our social and material experiences at the other, we can experience the embodiment of cultural differences as well as the emotions around new potential in the construction of the self. By extending patterns of text tied to ‘the living of life’, verbal art renews our potential to mean, our opportunity to renegotiate our semiotic ‘contract’ with the shifting values we call reality.

Acknowledgement

To my nephew, David McIlwain, I would like to give special thanks for his learned advice, debate, knowledge of Japanese and Korean, and his collaboration with me in matters of poetry, culture, and Shakespeare.

References

Bateson, G. (1980) Mind and Nature: a Necessary Unity (Waukegan, Ill.: Fontana). Blackmur, R. P. (1957) Form and Value in Modern Poetry (New York: Doubleday Anchor). Bloom, H. (1973) The Anxiety of Inﬂuence: a Theory of Poetry (Oxford: Oxford University Press). Bowcher, W. and Smith, B. (eds) (2014) Systemic Phonology: Recent Studies in English (Shefﬁeld: Equinox). Butt, D. (1983) ‘Semantic “drift” in verbal art’. Australian Review of Applied Linguistics 6 (1): 34–48. Butt, D. (1985) Talking and Thinking: the Patterns of Behaviour (Geelong, Vic.: Deakin University Press). Butt, D. (1988) ‘Randomness, order and the latent patterning of text’. In D. Birch and M. O’Toole (eds) Functions of Style (London: Pinter), pp. 74–97. Butt, D. G. (2007) ‘Thought experiments in verbal art: examples from modernism’. In D. R. Miller and M. Turci (eds) Language and Verbal Art Revisited: Linguistic Approaches to the Study of Literature (London, Equinox), pp. 68–96. 52 Society in Language, Language in Society

Butt, D. G., Moore, A. R., Henderson-Brooks, C., Meares, R. and Haliburn, Joan (2010) ‘Dissociation, relatedness and “cohesive harmony”: a linguistic measure of degrees of “fragmentation” ’. Linguistics and the Human Sciences 3 (3): 263–93. Butt, D., Moore, A. and Henderson-Brooks, C. (2012) ‘Discourse correlates of therapeutic method and patient progress’. In R. Meares (ed.) Borderline Personality Disorder Treated by the Conversational Model: a Clinician’s Manual (New York: Norton Interpersonal Neurobiology Books), pp. 267–90. Cléirigh, C. (1998) ‘A selectionist model of the genesis of phonic texture’. Unpublished doctoral dissertation, University of Sydney. Cloran, C. (1995) ‘Deﬁning and relating text segments: subject and theme in discourse’. In R. Hasan and P. Fries (eds) On Subject and Theme: a Discourse Functional Perspective (Amsterdam: John Benjamins), pp. 361–403. Damasio, A. (2012) Self Comes to Mind: Constructing the Conscious Brain (London: Vintage Books). Davies, M. (1992) ‘Prosodic cohesion in a systemic perspective: Philip Larkin reading “Toads Revisited” ’. In P. Tench (ed.) Studies in Systemic Phonology (London: Pinter), pp. 255–62. Davies, M. (1993) ‘Theme, tonality and lineation in Shakespeare’s “Sonnets” ’. Paper given at the Second European Society for the Study of English Conference, Bordeaux, 4–8 September. Edwards, C. L. (1983) ‘The Parry–Lord theory meets operational structuralism’. The Journal of American Folklore 96 (380): 151–69. El-Menoufy, A. M. E. S. (1969) ‘A study of the role of intonation in the grammar of English’. Unpublished PhD thesis, University of London. Firth, J. R. (1957) Papers in Linguistics: 1934–1951 (Oxford: Oxford University Press). Freeman, D. (1975) ‘The strategy of fusion: Dylan Thomas’s syntax’. In R. Fowler (ed.) Style and Structure in Literature: Essays in the New Stylistics (Oxford: Blackwell), pp. 19–39. Glock, H.-J. (1996) A Wittgenstein Dictionary (Oxford: Blackwell). Halliday, M. A. K. (1985) Spoken and Written Language (Geelong: Deakin University Press). Halliday, M.A. K. (2005) ‘A systemic interpretation of Peking syllables’. In J. J. Webster (ed.) The Collected Works of M. A. K. Halliday,Vol.8:Studies in Chinese Language (London and New York: Continuum), pp. 294–320. Halliday, M. A. K. and Greaves, W. (2008) Intonation in the Grammar of English (London and New York: Equinox). Halliday, M. A. K. and Matthiessen, C. M. I. M. (2014) Halliday’s Introduction to Functional Grammar (4th edn) (London and New York: Routledge). Halliday, M. A. K. and Webster, J. J. (2014) ‘Visualizing the architeXture of the text: analyzing Obama’s ﬁrst inaugural address and Nixon’s second inaugural address’. In M. A. K. Halliday and J. J. Webster (eds) Text Linguistics: the How and Why of Meaning (London: Equinox), pp. 272–366. Hasan, R. (1964) ‘A linguistic study of contrasting features in the style of two contemporary English prose writers’. Unpublished PhD thesis, University of Edinburgh. Hasan, R. (1968) ‘Linguistics and the study of literature’. Etudes de Linguistique Appliquée 5. Hasan, R. (1971) ‘Rime and reason in literature’. In S. Chatman (ed.) Literary Style: a Symposium (London: Oxford University Press), pp. 299–329. Hasan, R. (1975) ‘The place of stylistics in the study of verbal art’. In H. Ringbom (ed.) Style and Text: Studies Presented to Nils Erik Enkvist (Stockholm: Skriptor), pp. 49–62. David G. Butt 53

Hasan, R. (1979) ‘Workshop report No. 6: language in the study of literature’. In M. A. K. Halliday (ed.) Working Conference on Language in Education: Report to Participants (Sydney: University of Sydney). Hasan, R. (1984a [1996a]) ‘Ways of saying: ways of meaning’. In R. Fawcett, M. A. K. Halliday, S. Lamb and A. Makkai (eds) The Semiotics of Culture and Lan- guage,Vol.1:Language as Social Semiotic (London: Pinter), pp. 105–62. Reprinted in C. Cloran, D. Butt and G. Williams (eds) Ways of Saying: Ways of Meaning (London: Cassell), pp. 191–242. Hasan, R. (1984b) ‘Coherence and cohesive harmony’. In J. Flood (ed.) Understanding Reading Comprehension (Newark, Del.: IRA), pp. 181–221. Hasan, R. (1985) Linguistics, Language and Verbal Art (Geelong, Vic.: Deakin University Press). Hasan, R. (1988) ‘The analysis of one poem: theoretical issues in practice’. In D. Birch and M. O’Toole (eds) Functions of Style (London, Pinter). Hasan, R. (2007) ‘Private pleasure, public discourse: reﬂections on engaging with literature’. In D. R. Miller and M. Turci (eds) Language and Verbal Art Revisited: Linguistic Approaches to the Study of Literature (London, Equinox), pp. 41–67. Hasan, R. (2011) Selected Works of Ruqaiya Hasan on Applied Linguistics (Beijing: Foreign Language Teaching and Research Press). Hasan, R. (2014) ‘Towards a paradigmatic description of context: systems, metafunctions and semantics’. Functional Linguistics 1(2). Jakobson, R. (ed.) (1987) Language in Literature (Cambridge, Mass.: The Belknap Press of Harvard University Press). Janet, P. (1924) Principles of Psychotherapy (London: George Allen and Unwin). Kermode, F. (2004) The Age of Shakespeare (New York: Phoenix). Khoo, K. M. (2013) ‘Cohesive harmony: exploring a measure of cohesion and coherence in psychotherapy’. Unpublished honours thesis, Macquarie University, Sydney. Korner, A. (2014) ‘Analogical ﬁt: dynamic relatedness in the psychotherapeutic setting’. Unpublished PhD thesis, Macquarie University. Lanham, R. A. (1991) A Handlist of Rhetorical Terms (2nd edn) (Berkeley and London: University of California Press). Levin, S. R. (1976) ‘Concerning what kind of speech act a poem is’. In T. van Dijk (ed.) Pragmatics of Language and Literature (North Holland Studies in Theoretical Poetics, Vol. 2) (Amsterdam: North Holland Publishing Co). Lukin, A. (2003). ‘Examining poetry: a corpus based enquiry into literary criticism’. Unpublished PhD thesis, Department of Linguistics, Macquarie University. Lukin, A. and Webster, J. (2005) ‘SFL and the study of literature’. In R. Hasan, C. M. I. M. Matthiessen and J. Webster (eds) Continuing Discourse on Language (Vol. 1) (London: Equinox), pp. 413–56. McIntosh, A. (1963) ‘ “As you like it”: a grammatical clue to character’. Review of English Literature 4: 68–81. Malcolm, K. (2010) Phasal Analysis: Analysing Discourse through Communication Linguis- tics (London and New York: Continuum). Mann, W. C., Matthiessen, C. M. I. M. and Thompson, S. A. (1992) ‘Rhetorical structure theory and text analysis’. In W. C. Mann and S. A. Thompson (eds) Dis- course Description: Diverse Analyses of a Fund-Raising Text (Amsterdam: Benjamins), pp. 39–39. Martin, J. R. (1992) English Text: Systems and Structure (Amsterdam: Benjamins). 54 Society in Language, Language in Society

Matthiessen, C. M. I. M. (2002) ‘Combining clauses into clause complexes: a multifaceted view’. In J. Bybee and M. Noonan (eds) Complex Sentences in Grammar and Discourse: Essays in Honor of Sandra A. Thompson (Amsterdam: Benjamins), pp. 237–322. Meares, R. (2012) A Dissociation Model of Borderline Personality Disorder (Norton Series on Interpersonal Neurobiology) (New York and London: W.W. Norton & Company Inc.). O’Grady, G. (2013) ‘Choices in Tony’s talk: phonological paragraphing, information unit nexuses and the presentation of tone units’. In G. O’Grady, T. Bartlett and L. Fontaine (eds) Choice in Language: Applications in Text Analysis (London: Equinox), pp. 125–57. Ohmann, R. (1964) ‘Generative grammars and the concept of literary style’. Word 20: 423–39. Ohmann, R. (1971) ‘Speech, action, and style in literary style’. In S. Chatman (ed.) Literary Style (Oxford: Oxford University Press), pp. 241–54. O’Toole, L. M. and Shukman, A. (1978) Formalism: History, Comparison, Genre (Russian Poetics in Translation, Vol. 5) (Oxford: Holden Books Ltd). Panksepp, A. and Biven, L. (2012) The Archaeology of Mind: Neuroevolutionary Origins of Human Emotions (The Norton Series of Interpersonal Biology) (New York and London: W.W. Norton & Company). Porges, S. W. (2011) The Polyvagal Theory: Neurophysiological Foundations of Emotions, Attachment, Communication, and Self-Regulation (The Norton Series of Interpersonal Biology) (New York and London: W. W. Norton & Company). Ryan, L. (2013) ‘A progressive line of movement in every poem: semantic progression in the poetry of Dylan Thomas’. PhD thesis, Macquarie University. Saussure, F. de (1978 [1916]) Course in General Linguistics. Edited by C. Bally and A. Sechehaye in collaboration with A. Riedlinger (Glasgow: Fontana/Collins). Simpson, E. (1984) Poets in their Youth: a Memoir (London: Pan Books). Steele, T. (1990) Missing Measures: Modern Poetry and the Revolt against Meter (Fayetteville, Ark.: University of Arkansas Press). Tench, P. (1992) ‘From prosodic analysis to systemic phonology’. In P. Tench (ed.) Studies in Systemic Phonology (London and New York: Pinter), pp. 1–17. Tuckwell, K. R. (1999) ‘The grammar of desire: complexity, metaphor and signiﬁcation in Lacan’. BA (Hons) thesis, Macquarie University. Voloshinov, V. N. (1929 [1986]) Marxism and the Philosophy of Language (Harvard University Press). Vonwiller, J. and Cléirigh, C. (1997) ‘Concinnity: the cohesive co-ordination of tonality and grammatical structure’. Unpublished paper. Webster, J. (1998) ‘The poet’s language: foregrounding in Edwin Thumboo’s “Gods can die”?’ World Englishes 17 (3): 359–68.

Selection of Asian adaptations

Angoor (India, 1982) (Bollywood adaptation of The Comedy of Errors). Kaliyattam (India, 1997) (adaptation of Othello), director Jayaraaj. Kannaki (India, 2002) (adaptation of Antony and Cleopatra). Maqbool (India, 2004) (adaptation of Macbeth), director Vishal Bharadwaj. Omkara (India, 2006) (adaptation of Othello), director Vishal Bharadwaj. Ran (Japan, 1985) (adaptation of King Lear), director Akira Kurosawa. David G. Butt 55

Shakespeare Must Die (Thailand, 2005): http://asiancorrespondent.com/79666/thai- macbeth-movie-banned/ The Bad Sleep Well (aka Warui yatsu hodo yoku nemuru) (Japan, 1960) (adaptation of Hamlet), director Akira Kurosawa. The Banquet (China, 2006) (adaptation of Hamlet), director Feng Xiaogang. See: http: //news.xinhuanet.com/english/2006-09/05/content_5050440.htm Throne of Blood (aka Cobweb Castle or Kumonosu-jo) (Japan, 1957) (adaptation of Macbeth), director Akira Kurosawa.

Selection of criticism based on Asian perspectives

Ashizu, Kaori, ‘Kurosawa’s Hamlet?’ Shakespeare Studies No. 33 (1995): 71–99: www. s-sj.org/pdfs/St-Asizu.pdf :‘ ’ 3.6, 15 (2007): http://www.pu- kumamoto.ac.jp/∼tosho/file/pdf/kbs/3/bunsai3-3.pdf Coghlan, A. ‘Globe to Globe: A Midsummer Night’s Dream, Shakespeare’s Globe’, theartsdesk.com (1 May 2012): http://www.theartsdesk.com/theatre/globe-globe- midsummer-nights-dream-shakespeares-globe French, C. (1993) ‘Adaptation as interpretation: three versions of King Lear. Prison in theatre: confinement as a microcosm of our society’. See: http://preserve.lehigh. edu/etd/154 Gallimore, D. ‘Speaking Shakespeare in Japanese: some contemporary exponents’, in D. Kennedy and Y. L. Lan (eds) (2010) Shakespeare in Asia: Contemporary Performance (Cambridge University Press), pp. 42–56: http://www.social-sciences.brookes.ac.uk/ research/centres/ejrc/downloads/speaking_shakespeare.pdf Huang, A. (2003) ‘The tragic and the Chinese subject’. Stanford Journal of East Asian Affairs 3(1): 55–68: www.stanford.edu/group/sjeaa/journal3/china4.pdf Ick, Judy Celine, ‘Performing Shakespeare in colonial Southeast Asia’. The Asian Scholar 4 (e-journal): http://www.asianscholarship.org/asf/ejourn/articles/ Judy%20C%20Ick2.pdf Lee, Adele, ‘Year of Shakespeare: Coriolanus’, blog post (25 May 2012): http:// bloggingshakespeare.com/year-of-shakespeare-coriolanus Mohanty, Sangeeta, ‘The Indian response to Hamlet: Shakespeare’s reception in India and a study of Hamlet in Sanskrit poetics’, Dissertation, Basel University (2010): edoc.unibas.ch/1168/1/finalversionJan2010.pdf Nosheen, Iqbal, ‘Much ado about Delhi: RSC’s Indian Shakespeare. Chastity, courtship, arranged marriages’, The Guardian (1 August 2012): http://www.guardian. co.uk/culture/2012/aug/01/much-ado-rsc-indian-shakespeare Seoul Stages: ‘Some more Shakespeare in Korea’, blog post (20 October 2011): http:// seoulstages.wordpress.com/2011/10/30/some-more-shakespeare-in-korea/ Sharma, Yojana, ‘Shakespeare with extra spice’, TES (27 April 2007): http://www.tes. co.uk/article.aspx?storycode=2376259 Verma, Jatinder, ‘What the migrant saw’, The Guardian (10 January 2008): http://www. guardian.co.uk/stage/2008/jan/10/theatre1 Wu, Hui, ‘To seek revenge or to forgive: two Chinese films about Hamlet’, Shakespeare et l’Orient. 2009: http://www.societefrancaiseshakespeare.org/document.php?id= 1517 Part II Verbal Art 3 Jakobson’s Place in Hasan’s Social Semiotic Stylistics: ‘Pervasive Parallelism’ as Symbolic Articulation of Theme Donna R. Miller

3.1 Introduction

The foremost aim of this chapter is to provide further evidence for my proposal concerning the rightful place of Jakobson’s theory of ‘pervasive parallelism’ (1966, p. 423) within Hasan’s social semiotic stylistics (henceforth SSS) framework. Further substantiation is required to make the case more convincing, if Hasan’s (privately communicated) reservations are to be successfully countered.1 Section 3.2 begins by making plain and justifying the chapter’s premises: unqualified advocacy of Hasan’s analytical model of ‘verbal art’ and the theory behind it and a firm, not least because analytically and pedagogically tested, conviction that the model is unique, but also uniquely valid, due both to its holistic, coherent and systematic nature and to her (but also Halliday’s) view that literature is ‘different’, indeed ‘special’. This has become a bold notion which goes against the grain of what mainstream stylisticians (e.g. Simpson, 2004) have been preaching more doggedly over time, i.e. that literature is merely another text type, needing none other than the sundry tools brought to the analysis of any register. Subsequently, in section 3.3, attention turns to the model (into which I am arguing Jakobson should be ‘slotted’) itself. For those readers who are not sufficiently familiar with the particulars of Hasan’s ‘double- articulation’ framework, this will be sketched and its essential notions briefly described. Having laid this groundwork, in section 3.4 I address ‘the Mukaˇrovský– Jakobson theory’ (Fowler, 1986), delineating Jakobson’s contribution to the thesis and its relevance for SSS. Attempting to ally the misgivings of Hasan concerning the relevancy of Jakobson’s thought for her framework, I then set

59 60 Society in Language, Language in Society forth correspondences between her approach and what Jakobson theorized as the empirical linguistic evidence of his ‘poetic function’, i.e. grammatical parallelism (1960, henceforth GP). I argue that what he calls ‘pervasive parallelism’ functions just as Hasan says that significant patterning in verbal art must, i.e. as a consistent and motivated foregrounding device which symbolically articulates the theme of the literature text. Section 3.5 is dedicated to bolstering this claim. Choosing texts that confirm that the model is ‘maximally applicable to the genre [i.e. to literature], irrespective of variations in time, sub-genre, and the critic’s response’ (Hasan, 1985, p. 90), I offer analyses of foregrounded patternings in terms of pervasive parallelism in a nineteenth-, twentieth- and twenty-first-century poem: Matthew Arnold’s renowned lyric, ‘Dover Beach’; Siegfried Sassoon’s cheer- ily execrating social poem, ‘Does it Matter?’ and John Whitworth’s ‘Little’, a poignant contemporary elegy that has been closely examined by Hasan herself.2 In reproposing her analysis, my aim is not to dispute or surpass her findings, but simply to demonstrate that they can be effectively revisited à la Jakobson. In closing, my now reinforced if modest proposal for reconciling the two approaches is reiterated.

3.2 The centrality of language in verbal art

Hasan is convinced that literature is a special kind of text. It is indeed this belief which brought about her twofold model for its analysis. Her own crucial contributions to register theory in systemic functional linguistics (henceforth SFL) led to her demarcation of verbal art as unquestionably a kind of language use in a particular social context, but not simply a register on a par with any other. It is different largely because its context–language connection is fraught with complexities that other registers are, quite simply, not heir to (1975, p. 54; 2007, pp. 22–3); thus it requires a different methodological take. This fundamental belief rubs against the grain of claims by an ever larger number of stylisticians that literature is merely another text type, needing none other than the assorted tools brought to the analysis of any register. Such claims are publicized as a language-based move to debunk the age- old ‘lit crit’ veneration of literature (e.g. Simpson, 2004, pp. 98–9), which is indubitably an admirable ambition. To do that, however, one need not deny literature’s distinctive nature. And, clearly, neither does promoting the specialness of literature mean envisioning the language of literature as being located outside the boundaries of the linguistic system. I will begin with Hasan’s thought on the tools one exploits for the analysis of any text vs those required for the analysis of literature, as this is one of the key points on which she is often misunderstood. On one hand, central to her thought is that Donna R. Miller 61

the starting point for the description of literature is identical to that elsewhere [ ...] these descriptive categories are applicable to all uses of language, irrespective of where they occur; the semantic values assigned to them in the clauses in which they occurred is also a constant. There can be no dispute that these patterns of language are not the prerogative of literature; wherever they occur, their initial analysis is the same. (Hasan, 1985, p. 92, my emphasis)

So then, at the level of its language, literature is to be analysed with the identical tools and in exactly the same manner as any other text type. But the words I have italicized in the above segment – ‘starting point’ and ‘initial’ – plainly tell us that there is more to the analysis of literature than there is to that of other text types. In what follows I address why this should be so, and then how. In 1964, and again in 1971 and 1985, Hasan writes:

It is not that there is art [DM: somewhere ‘out there’, so to speak], and the job of language is simply to express it; rather it is that, if there is art, it is because of how language functions in the text [ ...] in verbal art the role of language is central. Here language is not as clothing to the body; it is the body. (Hasan, 1985, p. 91)

In the same year Halliday makes an analogous assertion: ‘Literature is language for its own sake: the only use of language, perhaps, where the aim is to use language’ (Halliday et al., 1964, p. 245, my emphasis). In 1982, he formu- lates the celebrated quip: ‘the paradox of “poetic” language [is] that there is no such thing [ ...] but we can all recognize it when we see it’ (Halliday, 1982, p. 134, my emphasis). But what is it that we recognize – what can the distinguishing characteristic of such language be said to be? It is not its ‘deviance’ – that 1960s’ notion that has enjoyed such enduring acclaim – for, as Hasan points out, nonconformity to a ‘norm’ is not a characteristic of literature alone since metaphorical ways of saying and meaning may be a feature of any register, including spontaneous face-to-face communication. Besides, ‘Today’s deviation is tomorrow’s norm’ (Hasan, 1985, p. 92). What she prescribes is a shift of focus, one that will abandon the misguided search for the language of literature and focus instead on language in literature, on its function in the text, on just what makes it ‘verbal’ art. And she also advances a method for distinguishing between what is verbal art and what, most likely, is not. In so doing, Hasan has constantly acknowledged her debt to Mukaˇrovský’s (1977, 1978) theory of foregrounding, which she interprets in terms of ‘contrast’, or better, signiﬁcant contrast, i.e. foregrounding that ‘counts’. And what counts, for Hasan, is, again Mukaˇrovskian, ‘consistency’ and ‘motivation’ (Mukaˇrovský, 1964), meaning that the articulated features 62 Society in Language, Language in Society are continually working, in a noticeably meaningful way, towards the construction of some theme. In sum:

If the patterning [DM: foregrounding] of patterns is consistently utilised for a second-order semiosis [ ...] then the text in question is a LITERATURE TEXT. If, however, such a role is not played by the patternings, then we have a LITERARY TEXT. The recognition of this distinction is important, not least because the techniques for the study and evaluation of the two are not identical. (Hasan, 1985, p. 101, emphasis in the original)

This reasoning is what led Hasan to propose her model of ‘double articulation’, which will be sketched below. But before turning to that model and further ﬂeshing out these notions, there is another aspect of the ‘specialness’ principle that needs glossing. Let us begin with a quote from Fowler, that herald of linguistic stylistics who, however, thought very differently on this point:

No plausible essentialist or intrinsic deﬁnition of literature has been or is likely to be devised. For my purpose, no such theory is necessary. What literature is, can be stated empirically, within the realm of sociolinguistic fact. It is an open set of texts, of great formal diversity, recognized by a culture as possessing certain institutional values and performing certain functions. (Fowler, 1981, p. 81)

The conflict is obvious, as Hasan’s definition is exquisitely essentialist – although her social semiotic perspective clearly rules out undervaluing the importance of culture and community. Indeed, she says that ‘the literature text [ ...] embodies precisely the kind of “truths” that most communities are deeply concerned with’ (Hasan, 1985, p. 100). But she does not invest the power for deciding what is or is not literature in the community, whether it be that of the time and place of the text’s creation or at a further semiotic social distance (Hasan, 2007, p. 34). This is not to say that the social does not impinge on verbal art: indeed ‘perhaps the most critical part it plays is in the shaping of the ideological orientations of those who write and those who read literature’ (Hasan, 2007, p. 25). And a text’s endurance as art will hinge on the value which is awarded it by successive generations of readers: ‘The challenge for the creator of verbal art is that the symbolically articulated theme has to be capable of striking a chord in the reader over substantial distances in time and space’ (Hasan, 2007, p. 25). So then, if literature is not simply whatever a specific community considers it to be, what is it? It is text whose meanings articulate – successfully and lastingly, but most importantly ‘symbolically’ – something akin to a generalization on the nature of human social existence. Verbal art must in fact Donna R. Miller 63 be what Butt describes as ‘the central problem-solving activity in the culture [ ...] show[ing] a broad concern for the community’s deepest problems, particularly those concerns which continue unresolved or which need to be renegotiated with each generation’ (Butt, 1996, p. 86). The Equinox blurb to Hasan (to appear) echoes all this flawlessly:3

Language, as Whorf said, is the best show man puts on. And perhaps the most amazing performance in this repertory is verbal art, which, thanks to the power of ordinary language, peoples the world with beings who, although they do not exist, hold a many-angled mirror to human life. They live, creating histories which possess deeper reality than our own real existence in society: through these histories is distilled human experience, made potent as an extended metaphor for the essential human condition. The deepest level of meaning, the themes in verbal art, relate to human social existence – its dilemmas and its delights, thus bearing witness to its socio-semiotic origins. All of this is achieved through a patterning of the patterns of language.

The crucial role of those ‘patternings of the patterns of language’ are indeed what has led me to take up certain aspects of Jakobson’s work and to argue that SSS must engage with them. But to better argue my case, an outline of the double-articulation framework is needed.

3.3 Hasan’s twofold framework for the study of verbal art

The ‘double-articulation’ model aims at providing a satisfying account of the distinctive functional role that language patterns in literature play in construing a special kind of social meaning exchange. ‘Functional’ here means the way language may act to shape patterns which are meaningful to the degree that they articulate an artistic theme. Figure 3.1 provides a visual representation of the framework and its orders of abstraction. Let us take the levels in their bottom-up order.

3.3.1 The ﬁrst level: the semiotic system of language For SFL, this is the multiple coding system which is the valid entry point for the analysis of any register. The system itself is made up of levels, or strata, each of which is realized, i.e. becomes accessible to us, through the one below it. In short, semantics, or meanings, are realized in lexicogrammar, or wordings, which become accessible to us in phonology, or soundings, or, in written text, in graphology, in written symbols. Above the system – at its source as it were, though not displayed in Figure 3.1 – is the context of situation which activates the meanings that will most likely be chosen from the total meaning potential at the text maker’s disposal, according to the features of the variables of ﬁeld, tenor and mode. In relation to that is the 64 Society in Language, Language in Society

Semiotic system of verbal art

THEME

SYMBOLIC ARTICULATION

VERBALIZATION SEMANTICS

LEXICOGRAMMAR

PHONOLOGY

Semiotic system of language

Figure 3.1 The overlapping semiotic systems (based on Hasan, 1985, p. 99) global context of culture in which the social languaging activity always takes place (cf. Halliday and Hasan, 1985, pp. 3–49).

3.3.2 The second level: the semiotic system of verbal art The first stratum of this second system is labelled ‘verbalization’. As is evident, this stratum is the point at which the two semiotic systems make contact, where the system of verbal art concretely links up to the system of the language of the text. Moreover, verbalization contains the whole of the first system; it equals it. Hence the broken line in Figure 3.1. The highest stratum of the second order of semiosis is the ‘theme’, after the conventional sense of the word in literary studies, i.e. in reference to the ‘deepest’ meaning of any verbal artefact (Hasan, 1975, p. 56). For Hasan, such ‘deepness’ means deeper than any of the meanings that the analyst is capable of discovering at the first semiotic level.4 For Hasan, theme is ‘what the text is about when dissociated from the particularities of that text’ (Hasan, 1985, p. 97). As we have already seen, it is very close to a generalization which can be viewed as a hypothesis about some aspect of the human condition. As Hasan intriguingly observes, this may be just what Aristotle had ‘in mind when he declared that art is truer than history’ (Hasan, 1985, p. 97). It is this second order of semiosis that is the critical criterion of the literature text. Without this ‘theme’ and its ‘symbolic articulation’, there is, for Hasan, no verbal art (Hasan, 1985, p. 100): a characterization that is categor- ically ‘intrinsic’, pace Fowler’s disbelief in the existence of, or need for, such a definition. Donna R. Miller 65

The middle stratum is called symbolic articulation, and this is where it all happens. But what? Here too an accessibility relationship obtains between the strata: symbolic articulation makes the theme accessible to us – just as, in the semiotic system of language, the lexicogrammar makes the semantics, i.e. the stratum ‘above’ it, accessible to us. Looked at from ‘below’, symbolic articulation is the place where the basic first-order meanings are ‘added to’, or ‘expanded upon’, or ‘heightened’, or ‘enriched’, or ‘deepened’, whatever metaphor one prefers, and are made ‘art’. How it does this is through foregrounding (or the patterning of patterns), the process by which the meanings resulting from the analysis of the semiotic system of language are symbolically turned into signs, for the purpose, the aesthetic motive, or ‘artistic intention’ (Mukaˇrovský, 1977) of expressing a theme – a deeper meaning. Let us now revisit some notions introduced above, in an attempt to better comprehend just what symbolic articulation consists in. As with most stylisticians, Hasan’s account of foregrounding is largely derived from Mukaˇrovský (1964, pp. 17–30). In her view, foregrounding is ‘contrast’. How- ever, the analyst can only be concerned with contrast which is significant. But just what does ‘significant’ mean? And how can the analyst tell when contrast is significant? Although there is no simple map that unproblemati- cally plots the road to certainty, there are modi operandi that can help us find our way. Recall that ‘significant foregrounding’ is foregrounding that ‘counts’. And what counts, for Hasan, is not occurrence or non-occurrence of linguistic mechanisms in and of themselves but rather ‘consistency’ and ‘motivation’. If there is motivation, it means that the features are articulated in a way that is working towards, unto, the construction of some significantly deeper meaning, or theme. And, if foregrounding is noticeably meaningful, then such foregrounding will also prove to be ‘consistent’, and in two ways, both of which contribute to making foregrounding noticeable.5 Such consistency is a question of: (1) the stability of its semantic direction, and (2) the stability of its textual location. Stability of semantic direction means that ‘the meanings which are being highlighted by the foregrounded patterns converge toward the same direction’ (Hasan, 1985, p. 95). In practice, it is the very semantic tension between foreground and background that is responsible for creating this kind of consistency. Without the background’s own consistency there would be no foreground. The second mode of consistency, the stability of textual location, refers not to predictable, fixed, material locations in the text (Hasan, 1985, p. 95), but merely to the tendency of significant patterns of foregrounding to take place in textually significant places. 66 Society in Language, Language in Society

But it is time to address the crux of this chapter. I have said above that Hasan has continuously acknowledged her debt to Mukaˇrovský. I now suggest, however, that she has stopped short of paying due attention to Jakobson, to whom I now turn.

3.4 The Mukaˇrovský–Jakobson theory

To my knowledge this theory is a brainchild of Fowler’s, coined in his rightly noting the theoretical intersection of Mukaˇrovský’s foregrounding and Jakobson’s parallelism: ‘For both of these writers, literary language draws the reader’s attention to its own artiﬁces of construction’ (1986, p. 73). But Fowler expressly censures their take on the functions and motives of foregrounding and explicitly sets it up against his own. Jakobson’s thought emerges in his seminal modelling of the communicative factors and corresponding functions of language, where he states that ‘The set (Einstellung) toward the MESSAGE as such, focus on the message for its own sake, is the POETIC function of language’(1960, p. 356, emphasis in the original). Admittedly, he had no functional lexicogrammar with which to substantiate the claim, but simply provided examples based on traditional grammar and phonology. Admittedly too, nowhere that I know of does he explicitly speak of the role of foregrounding in articulating the aesthetically motivated theme of the work, but the analogies between Jakobson’s insights and those of Mukaˇrovský are rightly remarked on by Fowler and are too many and essential to be ignored. Summarized below are the ideas that are most relevant to my reasoning.

3.4.1 Jakobson, the poetic function and GP A point that many stylisticians lose sight of, but one that is absolutely essential to Jakobson’s hypothesis, is that what he calls the poetic function has an ‘empirical linguistic criterion’, and that is GP, what he means by the much quoted but decidedly less understood assertion: ‘The poetic function projects the principle of equivalence from the axis of selection into the axis of combination’ (Jakobson, 1960, p. 358). In SFL terms, this means moving from a system to a structure perspective, from the choice axis to the chain axis. GP involves the recurring reiteration of equivalent grammatical units which can be of any or all types on the rank scale (from phoneme and metrical pattern, up through word, group, phrase and clause). However, the true, and most profound, signiﬁcance of the notion of parallelism for Jakobson – following up on Hopkins’ 1865 insight – is that GP calls forth a corresponding recurrence of meaning, of sense. Accordingly, GP is, at one and the same time, semantic parallelism. This is a vital hypothesis, and one which also tallies well with the ﬁnding that parallelism, as a structural cohesive device that tends to realize textual Donna R. Miller 67 meanings, typically works in a way that goes considerably beyond what is warranted cohesively. As Martin (1992, p. 386) puts it:

Grammatical parallelism [ ...] is exploited rhetorically in contexts where strictly speaking it is not needed to realise the meaning at hand. The result is a ‘surfeit’ of cohesive harmony.

Thus it becomes redundant and annuls its own textual function. At the same time, however, it perfectly illustrates the (typically literary) process of ‘de-automatisation’ (Halliday, 1982; cf. note 4), one ‘whereby a particular linguistic stratum makes meaning which is not predicted by its context’ (Martin, 1992, p. 386). Martin (1992, p. 386) contends that GP ‘is a concept fundamental to the interpretation of verbal art’. But how? Let us ﬁrst assume that cohesion is not the primary purpose of the (over)use of the resource, and that such an excessive use of GP in verbal art is therefore presumably ‘motivated’, i.e. has been chosen for the purpose of instantiating deeper meanings. In this case the question seems to become: to what experiential and interpersonal end(s) can such a markedly exploited surfeit of these ‘gorgeous grammatical tropes and ﬁgures’ (Jakobson, 1960, p. 375), or ‘syntactic imagery’ (Halliday, 1973, p. 121) be? What I am proposing is that the function of such parallelism must be likened to that of Hasan’s patterning of patterns, her foregrounding that counts – the motivated consistency of semantic direction for the symbolic articulation of verbal art, what its art resides in. But Hasan has objections, which I would attempt to put a happy end to.

3.4.2 Tackling Hasan’s reservations Hasan has various qualms concerning Jakobson’s ideas, some of which are easier to dispel than others. Firstly, she takes issue (1985, p. 99) with what she feels is Jakobson’s (and the Prague School linguists’ generally) too restrictive notion of the aesthetic function of verbal art, seeing such functions as being wider. But Jakobson clearly and carefully stresses that, on one hand, ‘the linguistic study of the poetic function must overstep the limits of poetry and, on the other hand, the linguistic scrutiny of poetry cannot limit itself to the poetic function’ (1960, p. 357). The ﬁrst tenet means that not all texts that display GP are to be considered literature texts; the second, that the poetic is not the sole function of literature. Jakobson’s six communicative functions are never seen as being mutually exclusive. Literature texts are said to have a dominant poetic function but are not conﬁned to it. They can also be strongly linked to his emotive one (lyric poetry), the conative one (literature with a social message), a referential one (epic poetry), and so on. In any case, Jakobson hypothesized an overlap of functions in any one text as being the norm. 68 Society in Language, Language in Society

A more serious, if not unrelated, objection that Hasan raises has to do with her critical distinction between verbal art vs verbal art manquée. She mistrusts the extent to which Jakobson is equally ﬁrm on this division.

Whenever the question arises as to which patternings of language are important in verbal art – imageries or metaphors or parallelism – my response has always been: depends on the work you are analyzing. If there is a work (purporting to be an instance of verbal art) where we do ﬁnd a lot of linguistic patternings including parallelism, but most of it does not have the function of articulating some theme(s), the questions arise: (i) is the work to be seen as verbal art?; (ii) is it to be seen as verbal art manquée? I believe these considerations are important – there has to be some difference between pretty language as in belles lettres and verbal art. (Hasan, personal communication, 10 January, 2011)

But Jakobson would agree with this distinction, and its importance. As already indicated above, he believed that the poetic function cannot ‘be arbitrarily confined to the domain of poetry’ (1960, p. 359), that it is to be found even outside poetry proper, where the function is not primary, i.e. in, to use Hasan’s expression, ‘verbal art manquée’. In these other varieties of text (among them, mnemonic texts, advertising, medieval law and Sanskrit theses!) the poetic function does not have ‘the coercing, the determining role it carries in poetry’ (1960, p. 359). So then, although Jakobson did not see the quality of ‘literariness’ as being exclusive to literature, he unmistakably saw literariness as being the distinguishing feature of literary language, part and parcel of its self-reflexivity, its awareness of itself as a (special) medium. As such then, I submit that proper ‘literariness’, for Jakobson, is inseparable from the poetic function, and its empirical evidence, that is, GP. Perhaps most difficult to argue against is an added reason that Hasan contests the central role of GP in her model; she does not see it as foregrounding in and of itself, but rather as merely one of a myriad variety of patterning mechanisms which may be there in a text, or not. She says, ‘If imagery, metaphor, repetition, parallelism – any patterning of patterns – is significant to the work then it will be there’ (personal communication, 10 January, 2011). In response, I bring in Jakobson’s notion of ‘pervasive parallelism’ (1966), which corroborates even further the Mukaˇrovskian/Hasanian emphasis on the foregrounding function of contrasting patterns. While any form of parallelism involves a choice of invariants and variables and works on the principle that the stricter the distribution of the former, the greater the discernibility and effectiveness of the latter, pervasive parallelism is wider-ranging. As Jakobson explains (1966, p. 423), it

inevitably activates all the levels of language – the distinctive features, inherent and prosodic, the morphologic and syntactic categories and forms, the lexical units and their semantic classes in both their Donna R. Miller 69

convergences and divergences acquire an autonomous poetic value. This focusing upon phonological, grammatical, and semantic structures in their multiform interplay does not remain conﬁned to the limits of parallel lines but expands throughout their distribution within the entire context; therefore the grammar of parallelistic pieces becomes particularly signiﬁcant.

And it becomes particularly signiﬁcant, I maintain, for the symbolic articulation of theme. My point is that pervasive parallelism is not merely one among a myriad of patterning mechanisms at work foregrounding the theme(s) in verbal art; rather, it can be likened to the process of foregrounding itself. But now to offer what will, hopefully, be persuasive evidence in support of my argument.

3.5 Pervasive parallelism as symbolic articulation of theme: further illustration

3.5.1 Matthew Arnold’s Dover Beach (1867) This celebrated nineteenth-century lyric has been chosen as exemplary of a tension between background and foreground, or between invariants and variants, as a model, in a word, of contrasting wordings/meanings.

1 The sea is calm tonight, 2 The tide is full, the moon lies fair 3 Upon the straits; on the French coast the light 4 Gleams and is gone; the cliffs of England stand, 5 Glimmering and vast, out in the tranquil bay. 6 Come to the window, sweet is the night air!

7 Only, from the long line of spray 8 Where the sea meets the moon-blanched land, 9 Listen! you hear the grating roar 10 Of pebbles which the waves draw back, and ﬂing, 11 At their return, up the high strand, 12 Begin, and cease, and then again begin, 13 With tremulous cadence slow, and bring 14 The eternal note of sadness in.

15 Sophocles long ago 16 Heard it on the Aegean, and it brought 17 Into his mind the turbid ebb and ﬂow 18 Of human misery; we 19 Find also in the sound a thought, 20 Hearing it by this distant northern sea. 70 Society in Language, Language in Society

21 The Sea of Faith 22 Was once, too, at the full, and round earth’s shore 23 Lay like the folds of a bright girdle furled. 24 But now I only hear 25 Its melancholy, long, withdrawing roar, 26 Retreating, to the breath 27 Of the night wind, down the vast edges drear 28 And naked shingles of the world.

29 Ah, love, let us be true 30 To one another! for the world, which seems 31 To lie before us like a land of dreams, 32 So various, so beautiful, so new, 33 Hath really neither joy, nor love, nor light, 34 Nor certitude, nor peace, nor help for pain; 35 And we are here as on a darkling plain 36 Swept with confused alarms of struggle and ﬂight, 37 Where ignorant armies clash by night.

Though they also function appreciably with other linguistic mechanisms in the text to mean, its contrasting patterns of transitivity are essential to its deeper meanings. The principal process types are relational, material and mental. Relational processes dominate stanza 1, where they are generally attributive, with grammatically parallel inanimate ‘natural’ Carriers, most in Thematic position, and having positively evaluated Attributes. They immediately set a peaceful scene, a background that is subsequently exploded: ‘The sea’ is ‘calm’; ‘The tide’/‘full’; ‘the moon’/‘fair’;6 ‘the cliffs of England’/‘glimmering and vast’, located ‘out in the [expressly said to be] tranquil bay’; ‘sweet is the night air’. Relationals reappear again in stanza 4 in describing what another, this time metaphorical sea, ‘The Sea of Faith’, ‘was once’ – akin to the tide in stanza 1, ‘at the full’. It is said to have once circumstantially ‘lay’, figuratively ‘like the folds of a bright girdle’ ‘round the earth’s shore’. A Google search reveals that the image has challenged critics since its construal. Co-textually it can be read as having once been happily and intensely all-pervading. Relationals – attributive, circumstantial and possessive – figure again strongly in the fifth and final stanza: in the Speaker’s plea to his ‘love’ to ‘let us be true to one another’ and in the reason offered for the entreaty: ‘the world, which seems to lie [ ...] Hath really neither joy, nor [ ...] And we are here [ ...]’. A significant semantic location I suggest, replete with GP, adding importantly to the poem’s construal of the opposition between animate ‘us’ vs ‘the world’. Donna R. Miller 71

Material processes in the poem are cumulative in their number and force (in an experiential gradatio, as it were). Their Actors are mostly inanimate. In stanza 1, a ‘light gleams’ and then ‘is gone’, and the Poetic voice commands a companion to ‘Come to the window’ (to the sweet night air). In stanza 2, ‘the sea meets the moon-blanched land’. At this point, inanimate nature’s gentle motion is transformed, becomes foregroundedly tangible, even fierce and disturbing, shattering the stillness thus far represented with the ‘grating roar’ made by the pebbles that ‘Waves draw back, and fling’, ‘bring[ing]/The eternal note of sadness in’. This same motion is reiterated again in stanza 4 as, however, the ‘Sea of Faith’s unidirectional ‘melancholy, long, withdrawing roar/Retreating’, heard by the ‘I’ of the poem (explicit just this once). The image of the pebbles in stanza 2 is replaced with the even more harsh and troubling extended NG (nominal group) ‘naked shingles of the world’, part of the Location to which the ‘roar’ retreats. In stanza 5, an Actorless material process – ‘Swept’ – having ‘we’ as Goal and working with the violent ‘clash’ of those ‘ignorant armies’, amplifies the chaotic confusion ‘we’ are caught up in. Patternings of mental processes function as the means to the perception of the sounds of sorrow and of menacing inanimate Phenomena that turn into destructive Actors. They are almost invariably the same lexical verb: ‘hear’ – four times – once immediately following the more active, near mental, behavioural process ‘Listen!’. The semantic location of the perception is concentrated in the central point of the poem, in which Sophocles functions as Sensor and his ‘mind’ as Location of his realization of ‘the turbid ebb and flow of human misery’ that the sound (of the ebb and flow on the shingle beach, albeit unspecified as such), as Initiator, ‘brings’. A second, explicitly corresponding, experiential metaphor, ‘we find also in the sound a thought’, reiterates the Initiator function of the sound, this time making ‘we’ think. Causation as variant – and ‘sound’ as its Initiator – are thus also foregrounded. The discovered thought is the already seen opaque description of the former fullness and all-encompassing brightness of the ‘Sea of Faith’. It need not unduly trouble us what it is that makes it retreat, though knowledge of the then sociocultural context7 has led critics to speak plausibly of fading religious certainties in the new age of (particularly Darwinian) science – none of which participates grammatically in the poem however. The final stanza instantiates all we need to know: the parallel positive but illusory attributes of ‘the world’, intensified by thrice reiterated ‘so’, which are belied by the negated things Possessed of that ‘world’, foregrounded by their parallel coordination:

[ ...] for the world, which seems To lie before us like a land of dreams, So various, so beautiful, so new, 72 Society in Language, Language in Society

Hath really neither joy, nor love, nor light Nor certitude, nor peace, nor help for pain

The din and darkness of the ﬁnal three lines are all that remain. Space precludes the appraisal analysis (Martin and White, 2005) that would reveal how pervasively parallel positive and negative evaluation skil- fully functions to reinforce the contrasting experiential meanings. A word will only be said about foregrounded imperative Mood. There are three instances: ‘Come’ in stanza 1; ‘Listen!’, in stanza 2, and collaborative ‘let us be true’ in stanza 5. The process types are those examined above: material, (near) mental and relational. The ﬁrst functions to further emphasize the calm visual scene outside ‘the window’; the second to emphasize the onset of the negative symbolic sounds; the third points up both the call for a bond and its quality: a time-honoured faithfulness which alone, it is implied, might provide the human ‘joy’, ‘love’, ‘light’, ‘certitude’, ‘peace’ and ‘help for pain’ that the world, as construed in the poem, no longer can. The theme is indeed timeless, linked explicitly by the Speaker to the classical past, with ‘Sophocles long ago [ ...] on the Aegean’, but also linkable to any time, past, present and future, in which a loss of former beliefs and values has the power to deeply distress and dismay.

3.5.2 Siegfried Sassoon’s Does it Matter? (1918) A more different poem than Matthew’s is hard to imagine, yet this poem too articulates its biting, scornful theme in and by an adept use of GP:

1 Does it matter?—losing your legs? ... 2 For people will always be kind, 3 And you need not show that you mind 4 When the others come in after hunting 5 To gobble their mufﬁns and eggs.

6 Does it matter ?—losing your sight? ... 7 There’s such splendid work for the blind; 8 And people will always be kind, 9 As you sit on the terrace remembering 10 And turning your face to the light.

11 Do they matter?—those dreams from the pit? 12 You can drink and forget and be glad, 13 And people won’t say that you’re mad; 14 For they’ll know you’ve fought for your country 15 And no one will worry a bit. Donna R. Miller 73

It may as well be said straightaway that the most significant contrast of the poem is that of its overall ironic surface meaning, enhanced by form, and what is actually being meant. The title and thrice reiterated (in the first line of each stanza) interrogative asks a yes–no question to which no sane person would answer ‘no’, and yet the dominant semantic direction construes reasons why ‘it’ does not/should not matter at all. The almost carelessly appended post-posed subjects – twice parallel Act + possessive deictic + Thing, also lexically identical except for the lost physical possessions, ‘legs’ and ‘sight’, and the third stanza’s psychological after-effect of combat in the First World War’s trenches, ‘those dreams from the pit’ – are cus- tomarily held to ‘matter’ very much indeed.8 This slippage between surface and ‘real’ meaning actually begins in the domain of verse, with an apparently tensionless, reassuring, quasi-limerick rhythm and rhyme scheme: a fixed ABBCA pattern, in basically anapestic trimeter with a variant: an ‘extra’ unstressed iamb at the end of the fourth line of each stanza. The pervasive parallelism of words highlights the central, conflicting, human and grammatical participant chains: ‘people’ and ‘you’, in constant tension through the poem. A key reiterated group is possessive deictic + Thing: ‘your sight’ and ‘your legs’, as noted, but also ‘terrace’, ‘face’, ‘country’: the (ex)possessions of the protagonist. Foregrounded as grammatical variants are the Acts, not only of ‘losing’, but also of ‘hunting’, ‘remembering’ and ‘turning’, only the second of which is a vigorous material action and, as such, is something only ‘the others’ can do. Besides ‘hunting’ and ‘gobbl[ing]’, ‘people are’, notably twice reiterated (facile and low-cost) ‘kind’, fully parallel in lines 2 and 8. Moreover, they are negated Sayers (line 13) and Sensers (line 15), in terms of the future, and/or perhaps of their determination to ignore the uncomfortable costs of war. They are predicted Sensers solely of a (then still generally unquestioned as noble) behaviour, evinced in line 14’s Fact of having ‘fought for your country’. They act, and are, nothing special; they neither feel, nor perceive, nor understand the maimed soldier’s loss. He himself is construed as potential Actor, Behaver, Senser and Carrier, but never Sayer. In stanza 1, line 3, he is actually advised not to complain (be a bore); in stanza 2, line 7, he is flippantly told of there being ‘splendid work for the blind’, and in stanza 3, line 12, of the possibility/ability, in the face of ‘those dreams’, to ‘drink and forget and be glad’. All he effectively does in the poem, however, is to ‘sit [ ...] remembering’ and ‘turning [his] face to the light’ (lines 9 and 10). Research into the context of creation of the poem reveals the widespread lack of social attention to the shell-shocked and mutilated victims of the trenches, and provides extratextual evidence of the cause of the poem’s decidedly ante litteram denunciation. The poet’s grim social censure had already largely emerged, however, from analysis of the consistently motivated pervasive parallelism in the text, pointing up how its apparently increasing nonchalance is at the same time being slowly eroded, line by line. 74 Society in Language, Language in Society

3.5.3 John Whitworth’s Little (2010) The text of Whitworth’s delicately wrought elegy is divided into numbered ranking clauses and its rhyme scheme indicated (from handout, Hasan, 2010).

//1 When Archie died //2 the year was dying too. a //3 Late loitering leaves were drifting to the ground, b //4 A time when dying has a lot to do, a And does it with a dry, susurrant sound. b

//5 Some say //6 when Archie died //7 it was not much – c A boy who did not walk or talk. //8 But he d Did love to smile and laugh and look and touch. c //9 He did do that. //10 He did it constantly. d

//11 So small, so weak, so fragile, //12 yet << //13 when death, e That most ingenious and practiced thief, f Unlocked the house //14 and stooped //15 and stopped the e breath,>> He left a strong sufﬁciency of grief. f

//16 Now is a winter and a summer since, g //17 And now the days of dying are come again – h A year since Archie died, our Little Prince, g A year of rain and sun, and sun and rain. h

The analysis Hasan performed was much more extensive than there is space to attend to here, so focus is chiefly on one aspect that strikingly demonstrates pervasive parallelism as symbolically articulating the poem’s theme: the tension between the experiential meanings construed through transitivity and tense patternings. Tables 3.1 and 3.2, summarizing these, are also Hasan’s impeccable handiwork. The relevant findings concerning the two unequal Doers are evident in the tables. The human participant, Archie, is already lifeless in the poem, having – thrice reiterated – ‘died’. Among the process types of his invariably past ‘doings’ are none construing action as such: he ‘did not walk’ or ‘talk’, but he had been wont ‘to smile and laugh and look and touch’. The parallelism of the four coordinated behavioural processes foregrounds their significance and in clauses 9 and 10 the partially reiterated ‘He did do that. He did it constantly’ markedly insist on the value of that gentle habitual behaviour that ‘Some say [ ...] was not much’. Likewise, the thematically prominent and grammatically parallel intensified attributes ascribed to Archie – ‘So small, so weak, so fragile’ – function to Donna R. Miller 75

Table 3.1 Process types ascribed to Archie (from handout, Hasan, 2010)

Clause Archie: -er role Main verb Process type Tense

1 Archie died Behavioural Past 6 Archie died Behavioural Past 7 A boy [[who did not walk Behavioural Past (did not) talk Behavioural Past 8 he did love to smile Behavioural Past and laugh Behavioural Past and look Behavioural Past and touch Behavioural Past 9 He did do that Behavioural Past 10 He did it Behavioural Past 11 (he) (was) so small, so Ascriptive Past weak, so fragile 12 He left a strong sufﬁciency Material Past of grief 17 Archie, our died Behavioural Past Little Prince

Table 3.2 Process types ascribed to death/dying (from handout, Hasan, 2010)

Clause Death: -er role Main verb Process type Tense

[4a] dying has Possessive Present (dying) does it Material Present 13 death ...thief unlocked the house Material Past (death ...thief) stooped Material Past (death ...thief) stopped the breath Material Past 17 the days of dying are come again Material Present

lay emphasis on his defencelessness. In spite of this, however, (‘yet’, clause 12), he had the power to leave ‘a strong sufficiency of grief’, an understated mourning, which is adroitly pre-prepared and amplified by the parallelisms remarked above and functions to position the reader to feel that ‘grief’ as well. Alliteration does its part too, here with the reiterated /s/ of that experi- entially low-profile ‘strong sufficiency’. Without arguing for that so inexact science, sound symbolism, the prominent frequency of phoneme /s/ (24 times) in this short poem does not seem insignificant; it occurs 11 times in stanza 3 – the location of Archie’s death – alone. Might it be said to function to enact, repeatedly but especially in a highly significant semantic location, that explicitly described ‘dry, sussurant sound’ (stanza 1) with which dying is said to do what it has to do? At the very least such reiteration further highlights the happening, as the temporal clause’s (13) abrupt suspension 76 Society in Language, Language in Society of the adversative clause (12) in the first line of the stanza (‘yet when’) may also be said to do. Thus it is within a dependent clause that ‘death’, per- sonified as ‘That most ingenious and practiced thief’, makes its downplayed entrance. The again coordinated doings are three but contrast significantly with Archie’s; these are material and increasingly menacing, robber-like; two are fatally Goal-oriented. Death’s material actions in the poem are performed in a specified moment in the past, but quasi-synonymous ‘dying’ has a more extensive role, one that is natural, cyclical, punctiliar and unremitting. The poem opens with Archie’s death couched in the dying year; it closes, symmetrically, with the return of the ‘days of dying’, explicitly linked to the anniversary of Archie’s death, commemorated, as it were, by the natural cycle of time and what time recurrently brings. Grammatically the final stanza is made prominent by the stately parallelisms of the final two lines’ reiterated ‘year’ and inverted GP: ‘rain and sun and sun and rain’. But Hasan (2010) had analysed these connections too. She merely neglected to observe the fundamental role of pervasive parallelism in construing them. As Table 3.3 (from handout, Hasan, 2010) shows, this time- worn cliché is, as Hasan put it, ‘rejuvenated’ by the poem in an ‘almost

Table 3.3 The close associations of time and death (from handout, Hasan, 2010)

Clause Time and death The relation of death to time

1 When Archie died Identifying a moment of time The year was dying too. Describing a moment of time 3 A time when dying has a lot to do Identifying a moment of time 5 when Archie died Identifying an event in time 9 ...11 when death ...unlocked the house, Identifying a moment of time and stooped and stopped the breath 13 now is a winter and a summer since The parts of year identify the time (death stopped the breath, and when death ‘unlocked the house Archie died) and stooped and stopped the breath’ leaving Archie dead 14 now the days of dying are come Time and death are in it together: again they come and their coming marks the part identiﬁed by Archie’s death 15 A year since Archie died, our Little Prince, 16 A year of rain and sun, and sun No matter how the annual cycle and rain runs Donna R. Miller 77 covert way’. The patterning she traces is crucial to the symbolic articulation of the subtly evoked theme of grief implanted by a loved one’s untimely death. ‘Time and death are in it together’, she says, and indeed the deeper theme that this truth bonds us to is our invoked common mortality: ‘death’, with time, will ‘stoop and stop the breath’ of us all.

3.6 Wrapping up

Following a brief excursus into Hasan’s ideas on the ‘specialness’ of the literature text and her analytical model of double articulation, this chapter has further argued my case for the inclusion of GP in the social semiotic approach to verbal art, offering new and hopefully persuasive substantiation for the proposition. I believe that my albeit partial examination of pervasive parallelism in the three poems above has credibly clarified which reiterated grammatical instantiations are operating, which patternings of meanings such choices are constructing and, ultimately, the higher-order theme that these work together to articulate. It is my extensively tested conviction that where there is consistent and motivated foregrounding, symbolically articulating a literature text’s theme, chances are that GP has an essential role in its construction. My proposal, however, is no more than a modest one, since coupling Jakobson’s insight with the tenets of SSS is, I submit, wholly unproblematic. Theoretically, and practically, it fits; indeed, it is already there. After all, for Jakobson parallelism is the supreme empirical linguistic criterion of the poetic function (1960, p. 358) which, recall, focuses on the corresponding factor of the ‘message’ itself, the instantiated shape of its meanings. Between this proposition and Hasan’s conviction that in verbal art ‘language is not as clothing to the body; it is the body’ (1985, p. 91), there is indisputable affinity. Surely it is time to recognize this.

Notes

1. Miller (2010b, 2012, 2013) proposes an analysis of the opening and closing of Eliot’s ‘Ash Wednesday’. But other ‘evidence’ includes: Miller’s (2007) illustration of GP as symbolic articulation in three poems from different phases of D. H. Lawrence’s oeuvre; Miller’s (2010a) analysis of Blake’s ‘The Garden of Love’, using Taylor Torsello’s exploration of the poem as well (see e.g. 1992, pp. 47–53). Though they have undergone considerable revision, more than a few of the points made in sections 3.2–3.4 were also discussed in Miller (2012, 2013). 2. The analysis was part of a workshop I collaborated with on ‘Semiotic Stylistics and Verbal Art’ at the LinC Summer School and Workshop 2010 in Cardiff. 3. https://www.equinoxpub.com/home/verbal-art-social-semiotic-perpectivethe -collected-works-ruqaiya-hasan-vol-7-ruqaiya-hasan-edited-jonathan-j-webster/, accessed 11 June 2015, my emphasis. 78 Society in Language, Language in Society

4. It should be pointed out that in 1973, already engaging with Hasan’s work, Halliday interprets Mukaˇrovský’s foregrounding in terms of motivated ‘prominence’ of grammatical features and also proposes twin levels of meaning, both grammatically realized, but one ‘underlying’ or ‘deeper’ than the ﬁrst, ‘immediate’ level, which is also glossed as ‘subject matter’ (Halliday, 1973, pp. 118–20, et passim). The deeper semantic meanings are said, crucially, ‘to serve a vision of things [ ...]. The vision provides the motivation for their prominence.’ It will be primarily Hasan’s task in the course of the 1970s to better delineate this ‘vision’. In Halliday (1982) the term ‘foregrounding’ is replaced by ‘de-automatisation’, with a view to underlining the distinct if complementary roles of these two semantic levels in a systemic functional perspective. As he puts it:

what is in question is not simply prominence but rather the partial freeing of the lower-level systems from the control of the semantics so that they become domains of choice in their own right. In terms of systemic theory the de- automatization of the grammar means that grammatical choices are not simply determined from above: there is selection and pre-selection. Hence the wording becomes a quasi-independent semiotic mode through which the meanings of the work can be projected. (1982, p. 136, my emphasis)

5. Mukaˇrovský too claimed the consistency of foregrounding (1964, pp. 17–30). 6. An interesting intertext is ‘Look Down, Fair Moon’, from Walt Whitman’s collection of Civil War poetry, Drum-Taps, published in 1865, and so, perhaps, known to Arnold. 7. The ‘context of creation’ of verbal art is a level of analysis in Hasan’s model I have not disambiguated for lack of space. Its function is mainly to corroborate and/or further disambiguate the theme as formulated with the twofold model of semiosis. 8. I interpret the semantics of ‘losing’ as depriving, or causing to not have, though here the Depriver is not pinpointed but rather only implied as warfare itself. I read ‘matter’ as relational attributive, being ‘important’, but it could well be read as an experiential metaphor for a mental process (worry you), especially in this text.

References

Butt, D. (1996) ‘Literature, culture and the classroom: the aesthetic function in our information era’. In J. E. James (ed.) The Language–Culture Connection (Singapore: SEAMO Regional Language Centre), pp. 86–106. Fowler, R. (1981) Literature as Social Discourse: the Practice of Linguistic Criticism (London: Batsford). Fowler, R. (1986) Linguistic Criticism (Oxford: Oxford University Press). Halliday, M. A. K. (1973) ‘Linguistic function and literary style: an enquiry into the language of William Golding’s The Inheritors’. In M. A. K. Halliday, Explorations in the Functions of Language (London: Edward Arnold), pp. 103–40. Halliday, M. A. K. (1982) ‘The de-automatization of grammar: from Priestley’s “An Inspector Calls” ’. In J. Anderson (ed.) Language Form and Linguistic Variation (Amsterdam: John Benjamins), pp. 132–58. Halliday, M. A. K. and Hasan, R. (1985) Language, Context, and Text: Aspects of Language in a Social-Semiotic Perspective (Geelong, Vic.: Deakin University Press). Halliday, M. A. K., MacIntosh, A. and Strevens, P. (1964) The Linguistic Sciences and Language Teaching (London: Longman). Donna R. Miller 79

Hasan, R. (1964) ‘A linguistic study of contrasting linguistic features in the style of two contemporary English prose writers’. Unpublished PhD thesis, University of Edinburgh. Hasan, R. (1971) ‘Rime and reason in literature’. In S. Chatman (ed.) Literary Style: a Symposium (London: Oxford University Press), pp. 299–329. Hasan, R. (1975) ‘The place of stylistics in the study of verbal art’. In H. Ringborn (ed.) Style and Text: Studies Presented to Nils Erik Enviste (Stockholm: Skriptor), pp. 49–62. Hasan, R. (1985) Linguistics, Language and Verbal Art (Geelong, Vic.: Deakin University Press). Hasan, R. (2007) ‘Private pleasure, public discourse: reﬂections on engaging with literature’. In D. R. Miller and M. Turci (eds) Language and Verbal Art Revisited: Linguistic Approaches to the Study of Literature (London: Equinox), pp. 41–67. Hasan, R. (2010) Unpublished analysis of John Whitworth’s Little, presented in the seminar on ‘Semiotic Stylistics and Verbal Art’ at the LinC Summer School and Workshop in Cardiff, 14–16 September. Hasan, R. (forthcoming) The Collected Works of Ruqaiya Hasan (Vol. 7): Verbal Art: a Social Semiotic Perspective. Edited by J. J. Webster (London: Equinox). Jakobson, R. (1960) ‘Closing statement: linguistics and poetics’. In T. A. Sebeok (ed.) Style in Language (Cambridge, Mass.: MIT Press), pp. 350–77. Jakobson, R. (1966) ‘Grammatical parallelism and its Russian facet’. Language 42 (2): 399–429. Martin, J. R. (1992) English Text: System and Structure (Amsterdam and Philadelphia: John Benjamins). Martin, J. R. and White, P. R. R. (2005) The Language of Evaluation: Appraisal in English (Houndmills, Basingstoke: Palgrave Macmillan). Miller, D. R. (2007) ‘Construing the “primitive” primitively: grammatical parallelism as patterning and positioning strategy in D. H. Lawrence’. In D. R. Miller and M. Turci (eds) Language and Verbal Art Revisited: Linguistic Approaches to the Study of Literature (London: Equinox), pp. 41–67. Miller, D. R. (2010a) ‘The Hasanian framework for the study of “verbal art” revisited ...and reproposed’. Textus 23(1): 71–94. Miller, D. R. (2010b) ‘The “double-articulation” of verbal art: defying Ockham’s Razor?’ Paper presented at the 22nd ESFL Conference and Workshop, Koper, Slovenia, 9–12 July 2010. Miller, D. R. (2010c) ‘Notes in the margins of the NON social semiotic stylistic approach to literature.’ Lecture given introducing a workshop with R. Hasan on Semiotic Stylistics and Verbal Art at the LinC Summer School and Workshop, Cardiff, 14–16 September 2010. Miller, D. R. (2012) ‘Slotting Jakobson into the social semiotic approach to “verbal art”: a modest proposal’. In F. Dalziel, S. Gesuato and M. T. Musacchio (eds) A Life- time of English Studies: Essays in Honour of Carol Taylor Torsello (Padua: Il Poligrafo), pp. 215–26. Miller, D. R. (2013) ‘Another look at social semiotic stylistics: coupling Hasan’s “verbal art” framework with “the Mukaˇrovský–Jakobson theory” ’. In C. A. M. Gouveia and M. F. Alexandre (eds) Languages, Metalanguages, Modalities, Cultures: Functional and Socio-Discursive Perspectives (Lisbon: BonD), pp. 121–40. Miller, D. R. and Turci, M. (eds) (2007) Language and Verbal Art Revisited: Linguistic Approaches to the Study of Literature (London: Equinox). 80 Society in Language, Language in Society

Mukaˇrovský, J. (1964) ‘Standard language and poetic language’. In Garvin P. (ed. and trans.) A Prague School Reader on Aesthetics, Literary Structure and Style (Washington, DC: Georgetown University Press), pp. 17–30. Mukaˇrovský, J. (1977) The Word and Verbal Art: Selected Essays by Jan Mukaˇrovský. Edited and translated by J. Burbank and P. Steiner (London: Yale University Press). Mukaˇrovský, J. (1978) Structure, Sign and Function: Selected Essays by Jan Mukaˇrovský. Edited and translated by J. Burbank and P. Steiner (London, Yale University Press). Simpson, P. (2004) Stylistics (London: Routledge). Taylor Torsello, C. (1992) Linguistica Sistemica e Educazione Linguistica (Padua: Unipress). Part III Semantic Networks 4 Can Semantic Networks Capture Intra- and Inter-Registerial Variation? Palliative Care Discourse Interrogates Hasan’s Message Semantics Alison Rotha Moore

4.1 Introduction

Perhaps the most important line of argument in Ruqaiya Hasan’s work is the idea that the semantic stratum of language can be modelled paradigmatically, through the tool which has been used to model other strata, namely, the system network (Hasan, 1996a, 2009b). As Hasan and Cloran put it (2009[1990], p. 95):

Since the principle of paradigmatic organisation applies to all levels of language, it is reasonable to suppose that the facts at the semantic level can also be represented as systems of interlocking choices.

Hasan’s model of ‘message semantics’ (1983, 1989, 1992a, b, 2009c) is prin- cipally known as a tool for studying semantic variation on the basis of speakers’ social position, but this chapter explores the contribution of message semantics – as a paradigmatic account of the organization of meaning – to the study of register variation, or the way that language varies according to its situational use. I argue that the framework is robust and that it has been underutilized in register studies, illustrating my argument with analyses of palliative care discourse and its relation to other registers. An early statement by Hasan about the need for modelling semantic options on a separate stratum from the lexicogrammar, and its importance for the study of register, appears in Hasan (1973, p. 273):

It has been too readily assumed that the easiest and most valid form of describing the linguistic characteristics of registers is to state the frequency or likelihood of individual patterns or of their combinations.

83 84 Society in Language, Language in Society

I would suggest that it might be advantageous to specify the characteristics of given registers by reference to some high-level semantic component.

In that paper, first prepared for a conference in 1969, Hasan illustrates the idea of the ‘high-level semantic component’ by contrasting utterances such as It certainly is lovely but it’s expensive and It is expensive but it’s unique ..., exploring why it is that these are not two ways of saying almost the same thing, as they might at first appear to suggest, given their shared and similar lexical and grammatical items (expensive, lovely, unique, but, it ...). Rather, one sentence denigrates the artefact in question, whereas the other promotes its desirability, and this antithesis constitutes one key semantic component of the register concerned. Hasan goes on to say that ‘no item-inventory [= lexicogrammatical item inventory, AM] could handle such features for the simple reason that they are not the property of individual items but of items of often different levels in combination’ (1973, p. 274). By the mid 1980s, Hasan had developed an ambitious, network-driven account of crucial contrasts in the semantics of English known as ‘message semantics’. By the late 1980s she had published innovative and controversial accounts of how middle- and working-class mothers used quite distinct ‘fashions of meaning’ in essentially the same context of control, based on this approach (e.g. Hasan, 1989). Central to Hasan’s vision of a paradigmatic semantics are two claims. Firstly, she suggests that the semantic systems of a language are contextually open and must be modelled that way – in other words, it is not the best kind of linguistics to draw up one semantics at a time, register by register or context by context;1 rather a map of the semantics of any language must incorporate registerial variation within that language, and must systematize it (Hasan, 1996a; Hasan et al., 2007). Secondly, for any proposed paradigmatic account of the semantics of a particular language to be viable it must be statable – like models of other strata – in terms of system networks with realization statements for each term in the system (Hasan, 1989, 1992a, b, 1996a; Hasan et al., 2007; cf. Halliday, 1973). If Hasan’s first claim holds, there would need to be evidence that such networks function to identify and explain relevant intra- and inter-registerial variation. Certainly the networks developed by Hasan have proved to be useful in describing and explaining highly significant intra-registerial phenomena, since they have generated important findings about semantic variation in the context of maternal control and in related pedagogical contexts. However, most register studies in SFL2 and nearly all such studies using Hasan’s networks have concentrated on intra-registerial variation only, or compare only very close contexts. Coming to the second claim, Hasan and her colleagues have developed and tested realization statements for the networks. Hasan views her semantic Alison Rotha Moore 85 networks as potentially language-exhaustive, but says that the networks as they stand do not reach this ideal, because they stop short in delicacy – at a point that permits examination of the questions central to her project (Hasan, 1989). While the networks and the empirical work based on them must be seen as a major development in functional linguistics, it remains important to treat the language-exhaustiveness of the networks as itself an empirical question, requiring testing across a wide variety of linguistic contexts, just as one might test a grammar across different contexts. So, to summarize, the proper evaluation of Hasan’s claims, and of the viabil- ity of her semantic approach for register, has been held up because there are still too few registers described using message semantics to make robust inter-registerial comparisons. In order to fill in some of this missing picture, the present chapter reports on a case study of message semantic analysis in the context of palliative care – specifically, end-of-life (EOL) discussions between patients and their doctors – and compares these findings with studies of other contexts that have used Hasan’s networks, focusing on the network for questions (Hasan, 1983, 2009b). Comparing results between such broadly different contexts suggests that Hasan’s semantic networks are indeed useful for mapping inter- registerial difference and similarity, in addition to intra-registerial variation. One thing that the network analysis usefully brings out is the way that palliative care consultation as a discourse is dominated by the semantics of ‘incipience’, ‘implicitness’ and ‘individuation’, as I hope to show below. From the point of view of enhancing clinical communication, an awareness of these patterns allows researchers (and, potentially, clinicians) to track intra-registerial variation in palliative care, such as whether the topic of ‘end of life’ is made available for discussion or not in different consultations, and if so, then to what extent, and via which semantic pathways, EOL discussion gets developed. These analyses can then contribute to research on how well different communication styles meet the needs of different patients, families and health workers. From the point of view of linguistics, such intra-registerial variation is already valuable data, but patterns of inter-registerial variation are of particular interest, because unexpected similarities and disjuncts between different contexts are revealed, which give insight into language speciation quite broadly, including interfaces between register, code and sociolect. In the case of palliative care discourse, the crucial concern with individuation is also known to be a hallmark of middle-class consciousness in English-speaking societies, as Hasan and colleagues found in their foundational work with semantic networks (Hasan, 1989, 1992a, b; Cloran, 1989; Williams, 1995; Hasan et al., 2007). It will be important to consider what this implies for the equitable delivery of health care services across different social class groups. The observations reported here provide additional support for Hasan’s claim that a networked semantics has the capacity to ‘systematize’ registerial 86 Society in Language, Language in Society variation, and they help justify her demand for contextual openness in semantic networks. As I hope to show below using the palliative care discussion, Hasan’s approach counts as ‘systematizing’ register because it allows us to motivate and explain semantic features and orientations that are shared by distinct contexts, in terms of relations between linguistic strata – or in other words, in terms of meaning in context. Without such contextual openness in descriptive tools, distinct contexts cannot be compared and register as a whole cannot be modelled. On Hasan’s second claim, researchers using semantic networks in new contexts of application have invariably needed to make some adjustments or extensions to the network, for example increasing the delicacy in the networks (Hasan et al., 2007). However in my testing of Hasan’s networks in palliative care and other medical discourse it appears there may be more substantive issues involved. One specific tension is that in order to ‘net in’ all relevant instances of what appears to be a single important semantic phenomenon, it seems necessary to modify Hasan’s realization rules, in some cases perhaps allowing ‘context-specific’ realizations as we will see below. Such tensions provoke queries about whether it is possible at the semantic level to exhaustively specify realization relations between strata, along with exponence relations between terms in the system and their structural ‘out- puts’. Before turning to the palliative care case study, a brief explanation and summary of relevant findings using semantic networks are in order.

4.2 Key ﬁndings from message semantics research

As Hasan and colleagues (2007, p. 717) report, semantic networks have been used successfully to investigate what they call ‘fashions of meaning’ in specific contexts. The original context was the context of ‘maternal control’, namely the day-to-day conversations between mothers and preschool children that shape young children’s consciousness. An analysis of approximately 22,000 messages using semantic networks yielded a robust pattern of variation at the semantic level correlating primarily with speakers’ social location, but also with the children’s sex (e.g. Hasan, 1989, 1992a, b; Cloran, 1989, 2000; Williams, 1995, 2001, 2005). Although message semantics is much broader in scope than speech function alone, many of the prominent studies using message semantics have focused on the analyses of the interpersonal function, especially through Hasan’s DEMAND INFORMATION network within her interpersonal semantic system of RELATION ENACTMENT, and I focus on it here because it is relevant to the palliative care case study. A particular finding that concerns us is that, although middle- and working-class mothers did not differ significantly in terms of whether they asked their children to confirm information or supply it, middle-class mothers tended to use certain semantic features, including [prefaced], which foregrounds point of view (discussed in detail below); [related], which modifies and links messages in terms of relevance; Alison Rotha Moore 87 and [non-assumptive], which avoids construing things as obviously normal or valued (also discussed below). These features are illustrated in an exchange between Kristy, aged 4, and her mother (Hasan, 2009c, p. 254), in particular turn 3.

1 Mother: I’m going out with some of the ladies because Sue is leaving. 2 Kristy: Mm 3 Mother: Did you know that they are going to leave? 4 Kristy: No 5 Mother: They’ve been building a house. 6 Kristy: Mm 7 Mother: Well, they haven’t been building it, somebody else has been building it for them.

Turn 3 contains one message, exemplifying the feature [demand information: confirm: ask] – in other words it is a ‘confirm question’. This message also exemplifies the feature [+ prefaced: subjective: ...child] in that Sue’s leaving is presented as something that it is important for Kristy to have an awareness of through the wording did you know, and construes the question of knowing as one that is up to Kristy – she has privileged access. The feature [assumptive] is also illustrated here by its absence. If Kristy’s mother had said Didn’t you know they are going to leave? this would be [+ assumptive] in that it implies that Kristy ought to know, or that such knowledge is obvious, however in this case the question is non-assumptive. In Hasan’s analyses (based on the statistical method of principal components analysis) these features clustered with some others forming a syndrome of features that was used more frequently by middle-class parents than by working-class parents, and also used more frequently with girl children than with boys (Hasan, 1989, 1992a, b, 2009c). Additionally, turn 1 illustrates the feature [supplementing/related],3 since the second message in turn 1, because Sue is leaving, specifies the reason for ‘going out’ in the first message. As Hasan points out, the various features used by the middle-class mothers can be explained in terms of the principle of ‘individuation’, according to which ‘each person is a unique being and their beliefs and opinions are inac- cessible to others without conversational mediation’ (Hasan, 2009c, p. 261). By contrast, working-class mothers, who tended to avoid prefacing and supplementary messages, and tended to use assumptive questions, subscribe to ‘the principle of naturalised reflexivity, acting as if most things can be taken for granted between persons who share the contexts of living with each other’. Hasan also stresses that these are two different orientations to interaction where ‘neither is better or worse; each is maintained at some expense’ (Hasan, 2009c, p. 263). What the message semantic analysis of maternal care allowed Hasan to demonstrate was that clusters of semantic features, when seen as ensembles or syndromes, can constitute a ‘sociolinguistic variable’ which operates 88 Society in Language, Language in Society within a specific context and distinguishes class groups (Hasan, 1989). A second phase of Hasan’s project focusing on kindergarten children talking with their teachers represented an additional, albeit closely related, context investigated with the same semantic networks. Taken together (e.g. Hasan, 2009f), Hasan’s studies demonstrate the kind of variation suggested by Bernstein (1971) and explain in part its mechanism. Countering Labov (especially – according to Hasan – Weiner and Labov, 1983), Hasan argues that variation occurs at every level including the semantic, and that semantic variation is not merely indexical but constitutive of social identity (Hasan, 2009a, d). One quite different social process studied using message semantics is political media. In a study of Australian broadcast media coverage of the Second Iraq War one key finding was that, surprisingly, in TV interviews with leading politicians and commentators covering the initiation of the war, journalists did not ask the kind of question that required these experts to explain why the country had in fact gone to war (Lukin, 2012). Rather, there was a striking use of questions such as the one put to the then Prime Minister, Did you imagine that you would be seeing the kinds of images of innocent victims now emerging ...?4 The selection of the feature [confirm] from Hasan’s DEMAND INFORMATION network, together with the selection [mental] from the CLASSI- FICATION network, along with the feature [prefaced] and other features, make the focus point of the question the mental state of the PM in the lead-up to the invasion of Iraq. The question is one which seeks not so much to know ‘what the world is like; it is rather an enquiry about someone’s ...“mental representation of that world” ’ (Hasan, 2009b, p. 250). For Lukin (2012), message semantic analyses of Australian and international media suggest the inadequate social functioning of the fourth estate. Other contexts studied with semantic networks include joint book-reading with young children (Williams, 1995, 2001, 2005), constructing ‘offers’ in ordering pizza (Matthiessen et al., 2005), court discourse (Maley and Fahey, 1991), political media (Lukin, 2012) and surgical teamwork (Moore, in press; Moore’s case study in Lukin et al., 2011). See Hasan et al. (2007) for brief descriptions of other contexts in which message semantic analyses have been productive. In each case, the achievement of the network-based analyses has been to organize multiple attributes of speakers’ meanings into coherent patterns that can be seen as motivated by features of the context of situation and the context of culture studied. I now turn to the case study on palliative care.

4.3 Semantics and variation in palliative care discourse

In this section I report on a study of spoken interaction between palliative care doctors and patients with advanced cancer by way of testing Hasan’s semantic networks for their ability to capture and organize crucial variation in this repercussive context. Note that consultations often included Alison Rotha Moore 89 a family member or members. Goals of the study included mapping the typical-actual linguistic practices found in this important social process, identifying linguistic variation that might be associated with contextual phasing, and identifying linguistic variation that might be associated with gradations in the quality of health care delivery, including the successful discussion of EOL issues. The approach set out by Hasan of linking paradigmatic accounts of context, meaning and wording provided the core principle of our analysis, and particular use was made of her message semantics model.

4.4 Implicitness, incipience and individuation in palliative care discourse

Before discussing the text analyses, some contextual background on palliative care is in order. One of the many challenging things for practitioners in the palliative care context is that, if they are going to follow current consensus models of good practice, doctors (and other health care professionals) have to balance a number of seemingly competing directives on how to approach EOL discussions (e.g. see Clayton et al., 2007b). Doc- tors are obliged to provide patients with opportunities to discuss EOL issues, by raising EOL topics and facilitating talk about them. But they must not force patients to talk about EOL issues if their patients do not wish to discuss them – even though it is now well documented in the literature that discussing EOL is associated with more realistic patient expectations and less aggressive medical care, which are in turn associated with better quality of life for patients and better caregiver adjustment after bereavement (Clayton et al., 2007b; Wright et al., 2008). This brings us immediately to the concepts of implicitness, incipience and individuation, which, I argue, are core semantic characteristics of palliative care discourse, possibly in the sense of what Hasan calls an ‘over-all tone’ that may pertain to whole texts or whole discourses at a particular historical moment (Hasan, 2009e), or possibly in her sense of ‘formative motif’ (2009b). As I hope to show below, the linguistic resources that bring such a sense of implicitness, incipience and individuation into being through doctor–patient talk can be identified – at least in part – as corresponding to specific selections of features from Hasan’s semantic networks, and these in turn can be seen as motivated by a contextual configuration that prioritizes patient agency and autonomy in a particular way and orients to local immediate and longer-term goals (Hasan, 2009g) for both individual patients and the developing professional identity of palliative care (cf. Semino et al., 2014). These concepts also turn out to be important in identifying registerial features that palliative care discourse shares with quite distinct contexts, based on shared configurations of contextual and semantic features. 90 Society in Language, Language in Society

At the request of a team of palliative care and medical communication researchers, a small study using linguistic approaches was built into the research design of a larger study on palliative care consultations. The larger study was a randomized controlled trial (RCT) design which ultimately showed that patients who were provided with a question-prompt booklet had more EOL discussion with their palliative care physicians than those in the control group (Clayton et al., 2007a). The linguistic sub-study concentrated on how EOL issues emerged as topical, and how discussion developed (or not), looking in particular at the role of doctors’ questions. In the linguistic sub-study, my focus was on how to model different ways of framing questions – in particular, how doctors’ framings might facilitate (or hinder) an overall strategy of enabling EOL talk. Although it must be stressed that interactants always co-construct communicative events (Drew and Heritage, 1992), doctors’ questions were of particular interest because they are the profession’s primary ‘entry point’ into offering patients options to discuss EOL issues. The health communication literature acknowledges this, but support for doctors’ reflective practice on question strategies is held back, because its models of how questions vary are largely limited to describing questions as either open or closed. A more comprehensive analysis of how questions vary could be useful for professional development and research in this area.5 Since it claims to be an ‘exhaustive’ semantic network, Hasan’s DEMAND INFORMATION network could be expected to provide a comprehensive set of distinctions in question framing. And since it is designed to work in conjunction with semantic-level analysis of experiential, textual and logical features, and relate all such results to accounts at the strata of context and lexicogrammar, it should be able to capture any variation in question framing that influences the way that EOL discussions unfold in palliative care. In other words, one test of semantic networks is that they display ‘contextual significance’, and the palliative care context is a good place to test for significance. The DEMAND INFORMATION network was therefore our primary analytical tool (Hasan, 1996a, 2009a) and other aspects of Hasan’s message semantics were drawn upon.6 For the linguistic analysis a subsample of 46 consultations across 6 doctors was examined. These included consultations with a question-prompt booklet and those where no booklet was given. The consultations ranged from not raising EOL issues at all, to extensive discussions about EOL issues. Data excerpts shown below were selected for their suitability in illustrating transi- tional points where EOL discussion emerged, or points where it might have developed but did not.7

4.5 Exemplifying EOL talk and how it develops

Extract 1 below shows the kinds of question and answer patterns associated with EOL discussion. In all examples, D = doctor, P = patient, K = kin Alison Rotha Moore 91

(a family member, partner, or close friend of the patient). Arrows in left-hand margins mark item(s) under focus. Underlining marks questions or features of questions under scrutiny that might otherwise be unclear. The numbers in the leftmost column are turn numbers. Where relevant, message numbers are indicated after an underscore.

Extract 1 – Transcript 101, turns 135–175 135 D Are you ﬁnding any problems with nighttime? 136 P No I just ...lay there and wait for daytime to come. ⇒ 137_1 D Do you? 137_2 A lot of patients tell me that nighttimes are sometimes very difﬁcult 137_3 especially if you’re not sleeping; 137_4 like you said you’re only sleeping for 4–5 hours. 138 P Yes I can go to bed at 10 o’clock and wake up at 2 and then I just lay awake there. 139 D What are you thinking about? 140 K Listen to the radio most of the time. ⇒ 141 D Really? 142 P Just listen to talk back radio, what’s happening, and think how much longer and all these normal questions – things go though your mind I guess. 143 D Yes sure and that is the quiet time when your mind thinks about these sorts of things. Do you ever get fearful? 144 P No I don’t get fearful I just worry about leaving the kids behind! That’s the thing that worries me the most. 145 D OK which is the natural feeling isn’t it? Shows that you’re very close. 146 P Oh we are. 147 D Do you ever think about the pain and the breathlessness and worry about that side of things? 148 P No because I think the main thing I think about is not waking up in the morning and having the kids – I call them kids but they come down and I’m no longer there. That sort of worries me a bit. 149 D OK have you spoken about that? 150 K Yes we all know. 151 D Is that something that worries you Keith [Kin]? [13 turns omitted] 165 K Yes well like Dad said, if he could stay at home for as long as possible, we’ll do that. I’ve already told him, I’ve told the other nurses and that that I’m prepared to do everything at home and ...if he loses bodily functions well that’s part of life ... 92 Society in Language, Language in Society

166 D Sure. 167 P And I’ve got no worries whatsoever doing that. 168 D Yes but that doesn’t necessarily always happen. People just assume that’s going to happen. 169 K Well we were told 6 months ago that’s what’s going to happen – he’s going to lose all his bodily functions and become a vegetable and just waste away. 170 D Yes – no, no – ⇒ 171 K One of the ladies in here said a couple of weeks ago that he could be like he is now. 172 D Absolutely! 173 K And could stay that way. 174 D Absolutely. 175 P That made me feel a lot better actually.

4.6 Transition points in EOL talk

The excerpt above starts after 134 turns at talk, covering mostly medication and routine symptom updates. Arguably, the point at which the discussion starts to instantiate ‘EOL talk’ is at turn 142 where the patient says that the issue of ‘how much longer’ goes through his mind – implying how much longer will I live? This implicit reference has cataphoric cohesive ties with other indirect references to death in turns 144 leaving the kids behind and 148 not waking up in the morning. The interaction then expands to include the patient’s son’s thoughts, as seen in the implicit we all know at turn 150. The ensuing EOL discussion, which cannot be shown in entirety for reasons of space, covers a total of 63 turns, and deals with a number of logistic issues, as well as some quite intense discussion of how people feel about these issues and how they manage them. Crucially, the discussion develops in a way that allows the doctor to correct a misapprehension that the patient and patient’s family had arrived with, namely that loss of function towards the end of life is inevitable (see turns 165–175). In the next section I consider how the doctor’s questions seem to open up the space for EOL discussion to develop, using Hasan’s semantic networks to characterize strategic choices made.

4.7 Strategic use of questions

The ﬁrst thing to notice is that most of the doctor’s questions are a kind of polar interrogative or a ‘closed question’, such as Are you ﬁnding any problems with nighttime? in turn 135. Yet the doctor does not come across as pushy, nor does this stretch of the consultation have the kind of one-way directive manner that often occurs during the part of a consultation that is Alison Rotha Moore 93 focused on checking physical symptoms. The second thing to notice about turn 135 is that a kind of gate leading to EOL discussion seems to be opened by this particular ‘closed question’ from the doctor, in conjunction with the way the doctor responds in turn 137 to the immediate answer given by the patient in 136. Asking questions is of course nearly always part of a complex communicative project, and it may be more common for questions to have a ‘multi-unit design’ rather than occurring as a single interrogative clause (Linell et al., 2003; cf. Berry, 1981; Heritage and Greatbatch, 1991; Martin, 1992; Williams, 1995).

4.8 Analysing question framing as options in semantic networks

Under Hasan’s approach, as in SFL more generally, in order to think about how this seemingly innocuous couple of turns at 135–137 work, we need to recall the key concepts of function and choice. Linguistic systems can be seen as sets of choices which enable people to shape their interactions as they unfold across three primary functions – changing the field of experience, the interpersonal relation and the textual organization8 as the interaction unfolds. At any point in a professional consultation or any other dialogue, speakers are faced with a multitude of choices (though usually not deliberate choices)9 in how to put together their next message. One of the most central choices is how the speaker is going to position their addressee(s). From the interpersonal point of view, when one asks a question one is acting out a social role of demanding some information from an interactant, whereas when one makes a statement one is putting oneself in the role of the giver of information and, simultaneously, putting the addressee in the role of the receiver of information for a certain time. At this very basic ordering of interpersonal choice, the remaining alternatives are to demand goods and services (with a command, including suggestions, requests, etc.) or offer (see Halliday and Matthiessen, 2014). Along with these primary choices in speech function, there are more ‘delicate’ choices as well, and Hasan’s system networks map out one view of these as a system of RELATION ENACTMENT,10 which includes options in the system DEMAND INFORMATION. Hasan’s model is a claim about how different patterns of wording are systematically related to different kinds of interpersonal meanings within questions, though not in a one-to-one fashion (Hasan, 1996a; cf. Harris, 1984; Labov and Fanshel, 1977; Tsui, 1992). To illustrate the message semantic approach and its usefulness, consider the following five ‘versions’ of turn 135 from palliative care transcript 101, each of which selects slightly different interpersonal meanings. The version shown first, as turn 135a, is the actual question used by the doctor in Consultation 101. 94 Society in Language, Language in Society

135a Are you ﬁnding any problems with nighttime? 135b Do you have any problems with nighttime? 135c What problems do you have at nighttime? 135d What are the most problematic things for you at nighttime? 135e What about nighttime – are there things that worry you?

Each of these versions of the question is a way of trying to get the patient to identify and describe problems that he might have regarding nighttime, but each puts a different kind of spin on the interpersonal relations between the addressee (at this point the patient) and the speaker (here the doctor) and how these roles might change over the subsequent turns.

4.9 Conﬁrm questions and apprize questions

The question that opens our excerpt from Consultation 101, Are you finding any problems with nighttime? displays Hasan’s feature [confirm], located at the most primary distinction in Hasan’s DEMAND INFORMATION network (see Figure 4.1). This question construes the existence of nighttime problems as a matter that is not yet settled between the doctor and the patient, and construes the patient as in a position to settle the matter. The feature [confirm] nets in all such questions that seek confirmation about some proposition. The question in turn 137 also selects the feature [prefaced] which is discussed in more detail below. The contrasting feature at this initial fork in the DEMAND INFORMATION network is [apprize], as exemplified in turn 139, What are you thinking about? and also in the hypothetical turn 135c above, What problems do you have at nighttime? In this example, nighttime problems are represented as a given – as if already known by both parties to exist. Here the doctor’s role as ques- tioner is set up as apprizing some missing element of a proposition (which problems?) rather than confirming or disconfirming a proposition. The semantic feature [apprize] is realized by preselecting mood as interrogative non-polar. One thing that such an [apprize] question would do is allocate the patient, for their next move, the role of specifying which problems were occurring for them at night. This is not to say that the patient’s next move could not be to deny that problems exist, but in that case a dispreferred response would be required (Levinson, 1983). One characteristic of [confirm] questions is that they often open two conversational doors at once, leaving the addressee with the opportunity to decide which way to move. When a speaker says ‘Have you got any questions? or ‘Are you finding any problems with nighttime?’ the addressee may orient to the grammatical (polar interrogative) form and just say ‘no’, meaning there are no problems. Alternatively they can report a problem or problems, in which case the ‘Yes, I do have problems’ is taken as understood. 95 alternative 2 simple 1 f – prompt + prompt e constrained 2 constrained open 1 assumptive 1 assumptive 2 non-assumptive e f contingents actants g event 1 event 4d d global 1 tentative 2 tentative reassure 1 particular 3 check 2 check probe 2 ask 1 c b enquire 2 verify 1 verify explain 1 explain specify 2 confirm G apprize H Hasan’s ‘Demand Information’ network (based on Hasan, 2013, p. 289, see also Hasan, 1996a and 2009b for earlier versions; ... entry condition in each version is ‘demand information’) Figure 4.1 96 Society in Language, Language in Society

These kinds of questions have elsewhere been described as incorporating a precondition: in logical form something like ‘please tell me whether you have a question and if so what’. In the palliative care context, this property makes [confirm] questions valuable for opening up the space to talk about EOL without pushing people into it, which is important for facilitating discussion about difficult topics. Interestingly, in the excerpt from Consultation 101 above, the patient tends to answer [ask] questions with a ‘no’ but then gives some information that implies ‘yes’. This apparent contradiction does not stop EOL discussion from developing, but in fact appears facilitative. Arguably therefore, the choice of [confirm] framing to begin this episode from Consultation 101 helps achieve the topical move into EOL issues in this case. I am not suggesting that [confirm] questions are ‘better’ questions than [apprize] questions, or that this choice exhausts the strategic semantic footwork available. The initial point is that choices in question framing have a range of semantic consequences which are subtle, which go further than the open/closed distinction, and which can be described in a systematic way, beginning with [confirm] versus [apprize]. Further distinctions in questions from Hasan’s model also seem to be important in shaping discourse about EOL, along with choices from experiential, textual and logical networks. There is insufficient space to detail all relevant distinctions here but in the sections below I outline some and illustrate their function.

4.10 Different types of conﬁrm questions: options ask, check, probe and reassure

Although [confirm] questions are very often realized as polar interrogatives, as in the first example, the semantic feature [confirm] can be realized by one of several kinds of indicative clauses or a clause complex. The more delicate options in [confirm] are realized by distinct lexicogrammatical features or groups of features. In fact, if a speaker selects [confirm], they must further select either [verify] or [enquire], thus [confirm] is an analytical category only – it cannot be instantiated without selecting further features.11 The same is true for Hasan’s next level of delicacy. If a speaker selects [enquire] they must select either [ask], by using a polar interrogative, or [check] by using a declarative with a Tone 2.12 Examples include:

[confirm: enquire: ask] Are you finding any problems with nighttime? (101_135) [confirm: enquire: check] So she’s not sleeping through the night? (2307_486)

If a speaker takes the [verify] pathway, they will select either [reassure], which is realized by a declarative with a reversed tag, or [probe], which is realized by Alison Rotha Moore 97 a declarative with a constant tag, according to Hasan’s speciﬁed realization rules.

[conﬁrm: verify: reassure] It’s hard, isn’t it? (1907_66) [conﬁrm: verify: probe] I take it that is your husband, is it? (2308_39)

4.11 The crucial role of probe and its agnates in EOL discussion

In the palliative care data [ask] and [reassure] questions occur frequently, and [check] questions are not uncommon. Question types described by Hasan as [probe] (with the realization declarative: tagged: constant) are rare in the transcripts, but items with an arguably similar function occur regularly and are exempliﬁed by turns 137 and 142 in Consultation 101.

136 P No I just ...lay there and wait for daytime to come. ⇒ 137 D Do you? [conﬁrm: verify: elliptical probe]13

In Hasan’s clearly articulated accounts of the terms in the systems and the possible realizations for those terms (see e.g. Hasan et al., 2007) ellipsis sometimes plays a role in identifying options. If we retrieve the elliptical elements in the message in turn 137 this question equates to Do you just lay there and wait for daytime to come? Since this question directly follows the patient asserting this fact, the question at 137 cannot be heard as merely a request to supply missing information – its function is not simply to [ask]. I suggest it could be descriptively treated as a kind of [probe] where the initial clause is elliptical and only the tag is specified, as a result of its place in the interactive and topical sequence.14 Like Hasan’s original [probe] such elliptical probes build in flexibility to the next speaker’s options for responding, while maintaining the focus on the information provided in the interlocutor’s last turn. In Consultation 101 at this point the doctor gives a short description of some of the problems that patients typically have, as a form of validation, giving some form of encouragement to the patient to further discuss EOL. Notice, at turn 137, that the doctor does not merely accept the patient’s disconfirmation of the idea of nighttime problems and move on to the next topic. The strategy the doctor follows here can be seen as one choice among several, where others might simply move on, or perhaps use a more direct, challenging line of questioning, such as Don’t you think laying there wait- ing for day is a problem?, using a question with the features [confirm: verify: ask; assumptive]. In comparison with such a challenging strategy the elliptical [probe] question can be seen as a way of creating an opportunity or invitation for the patient to step into the ‘problem’ space without being 98 Society in Language, Language in Society pushed – keeping that gate open a little longer. Very shortly after, a similar probe-like question is chosen by this doctor, in turn 141.

139 D What are you thinking about? [apprize: specify] 140 K Listen to the radio most of the time. ⇒ 141 D Really? [conﬁrm: verify: elliptical probe] 142 P Just listen to talk back radio, what’s happening and think how much longer and all these normal questions – things go through your mind I guess.

This time, the gate has been held open long enough and with enough sense of optional invitation. The patient steps gingerly into this space of EOL issues, which is incipient at ﬁrst as shown above, but becomes quite explicit, even including the very need to discuss EOL issues. In one sense the doctor’s turn 137 illustrates question types which may hold up topical progression of a conversation, but arguably the message is [progressive] not [punctuative] in terms of realization rules (it has a Predicator, although elliptical) and it is quite apparent that its inclusion changes the direction of the discussion. The question at turn 141 is perhaps closer to a [punctuative] message but it too meets requirements for [progressive] and in fact its grammatical alter ego (the clause) selects for Finite, Predicator and Polarity at minimum. One noteworthy feature of Consultation 101 is that there are three interlocutors, and these probe or probe-like questions from the doctor serve at least in part to calibrate the views of the three interactants. Therefore without ellipsis turn 141 might incorporate a third person subject or a second person subject/addressee: Does he/do you really listen to the radio most of the time? To my mind this is a good reason to err on the side of reconsidering additional terms or additional realizations from other registers that might modify or amplify the original message semantic proposal, since there is no reason to privilege dialogues of two interactants only in grammar or semantics – especially in spoken language. If the above probe-like questions (realized as tag only, and modal adjunct only, respectively) seem to be agnate to each other and in turn to Hasan’s original [probe] questions, then there are also some other items in this agnation series, which in turn appear to shade topologically into systems for prefacing and supplementing messages (projection/expansion as clausal relations). Consider the following.

Extract 2 – Consultation 2406, Turns 268–279 268 D Do you ﬁnd the effort of eating an issue? [conﬁrm: enquire: ask; prefaced] 269 P Yes it is at times. Alison Rotha Moore 99

270 D Yeah, mm yeah Ok. 271 P It’s painful there. That’s probably what turns you off eating I guess?15 272 D Mm, mm. We’ll come back to that. [turns omitted] 278 D You’ve lost a bit of weight, is that right? 279 P Yeah since um, since I was ﬁrst diagnosed. I was 74 kilos before I had the lung out.

In each case, a place can be found within Hasan’s system. For turn 271 in Consultation 2406, I guess could be treated as a prefacing element, since there is no stipulation in the realization rules that the projecting clause needs to precede the projected clause in a [prefaced] message. In turn 278 the doctor’s use of is that right could be treated as a separate [check] question, following a message of the type [give information]. But these examples are very close in function to the [probe] question with its declarative + polarity- consistent question tag, and they productively contrast with the reversed tag [reassure] and its variations, such as You’ve lost weight, isn’t that right? It seems unhelpful not to be able to net these into a general grouping, for various reasons. One important reason is suggested by the distribution of conﬁrm-type questions across the palliative care corpus: some individuals seem to use [reassure] as described by Hasan, whereas others seem to favour features such as is that right to do similar interpersonal work. In other words they are variants of some feature. Our picture of the variation within and between register domains such as palliative care will have important inaccuracies if we do not net in such agnation relations around question types. As well as the semantic functions of the question types described above, there are interesting patterns in our data that involve the way that questions are sequenced and nuanced, which Hasan’s system serves to bring out (although patterns such as this are exactly the kind of phenomenon that could go unnoticed without sufﬁcient capturing of agnate features). For example, from the interpersonal perspective, Extract 1 is organized around an iterative sequence which can be seen to occur quite commonly across the palliative care corpus, although it is not possible from the data analysed so far to make any strong statements about obligatory and optional moves in this register. In the excerpt shown, the doctor begins a topic with an [ask]∧[verify] sequence, followed by an [apprize: specify]∧[verify] sequence, then another [ask]∧[verify] sequence, then [ask], [ask], [ask]. In Consultation 2406, a similar pattern appears, if we net in ‘is that right’ and other forms that arguably function as a kind of [verify]. If we overlay such patterns onto a mapping of lexical sets and/or process types16 we get an even richer view of the ‘motivated selection’ occurring 100 Society in Language, Language in Society

(see Jakobson, 1987; Butt et al., 2013, 2014). In Consultation 101, we see that the series of [ask] questions described above sets up a kind of graduation (Martin and White, 2005) of mental activity from turn 135 to turn 141: have problems – get fearful – think of pain – worry about that side of things. This kind of graduation17 is a common pattern across the corpus, interacting with delicate question types, to create a semantic drift (Butt, 1983) that adds up to the higher-level semantic components I have described as incipience and implicitness – although the unit in question that selects [incipient] as distinct from, say, [abrupt] may be a unit ‘larger’ than the message. A further feature that can only be briefly mentioned here for lack of space is the frequent selection of options from various networks that have some measure of indefiniteness. These include the deictic ‘any’ in any problems in turn 137; the general process, and unspecified time and place, in the phenomenon the patient thinks about in turn 142, namely what’s happening; and the obscured process also in turn 142 how much longer; along with hyponymic relations that are invoked regarding such unspecified items, as in these sorts of things and that side of things. It becomes understood that ‘these sorts’ and ‘that side’ of things are things pertaining to the EOL.

4.12 When opportunities for EOL talk are not taken up by the patient

In this section I give two additional short extracts to illustrate some of the intra-register variation within this highly localized (Sydney, Australia; mid 2000s) domain of clinical interaction, and to help support my argument that this register shares some important variability with the variability in contexts of maternal control. Or perhaps more importantly, what the material here displays is the interconnectedness between the domain of palliative care and the domain of maternal control – that Hasan’s networked account of semantics helps to show how the patterns at work in these very different domains are really parts of an overall system of contrasts – the registerial repertoire of (Australian) English. I begin with a section from Consulta- tion 105 in which the doctor’s attempt or offer to move into EOL topics is declined by the patient.

Extract 3 – Consultation 105, turns 90–95 90 D Do you worry about the future? [enquire: ask] 91 P No. 92 D OK. 93_1 K We can’t discuss it. 93_2 We never talk about it. Alison Rotha Moore 101

94_1 D You never talk about it? [enquire: check; assumptive; non- prefaced] 94_2 Right, and you’re happy not to talk about it. [check; non- assumpt; prefaced] 94_3 You don’t want to talk about it, Paul? [check; assumpt; prefaced] 95 P No I’m quite happy not to talk about it.

In turn 90 the doctor’s question is of the type [confirm: ask] and is similar in many ways to the question Do you ever get fearful? from Consultation 101. Unlike the patient in Consultation 101, this patient gives a simple ‘no’, completing their answer. In this case the patient’s partner/carer (K) elaborates with We never talk about it, which, although it seems like a way of closing off that topic, actually provides a small aperture for the doctor to explore this couple’s perspective on talking about EOL issues. At this point the doctor switches from [ask] questions to a series of [check] questions, additionally manipulating the feature [assumptive] from the DEMAND INFORMATION and features from the PREFACING network, to tease out the ‘mind’ of the patient and their kin. The first [check] question, addressed to the carer, You never talk about it?, takes the value [+assumptive] by combining negative polarity in the Finite, with the feature [enquire].18 As discussed with reference to Hasan’s mother–child discourse studies, this feature implicitly codes speaker expectation and values. In this case the [assumptive] feature implies that it is unusual (or potentially problematic) not to discuss EOL issues. The doctor’s subsequent question drops the value [assumptive]19 but takes the feature [+prefaced], And you’re happy not to talk about it? The final question in this sequence is addressed back to the patient, and includes both [+assumptive] and [+prefaced]: You don’t want to talk about it Paul?

4.13 A note on prefaced questions

As with Hasan’s ﬁndings on mother–child language introduced earlier, the feature [prefaced] makes someone’s viewpoint the focus point of the question. Hasan calls this a ‘prefaced’ message because, semantically, it adds a kind of ‘point of view preface’ to the underlying non-prefaced message.

x You never talk about it? x You don’t talk about either? You’re happy not to talk about it? You don’t want to talk about it?

preface + underlying msg 102 Society in Language, Language in Society

Note that the ﬁnal extract from Consultation 105 displays the feature [prefaced: subjective: addressee (patient/carer/both); experiential: idea: reaction], thus it is the patient’s and carer’s views that are made the choice point. Because it foregrounds point of view, the selection of [prefacing], both with and without [assumptive] here, helps to construe the possibility of three separate views (doctor, patient, carer), and perhaps even imply the idea that whether you talk about the future or not is less important than whether you have the same perspective as your partner about such talk. In addition, this strategic manipulation of [assumptive] and [prefacing] options can be seen as a pathway into achieving what turns out to be a remarkable but concise metadiscussion about EOL talk, by a couple who have agreed not to talk about it. For instance, Consultation 105 contains a comprehensive rationale about why EOL is not to be discussed, which demonstrates the couple’s philosophy about dealing with the unknown: see turns 99–101 below.

Extract 4 – Consultation 105, turns 99–101 99 K When we saw Professor X, he said everyone handles shock, you know differently. 100 D Differently, yeah. 101 K And he said, if you’ve got the attitude that you can take today, today – you wake up in the morning and enjoy the day, and wait and see what tomorrow’s got – it’s a good way if you can be like it.

Features from all four networks are of course involved in the subtle semantic drift achieved in this discussion, but they are beyond the scope of this chapter to consider.

4.14 When the doctor does not build on incipient EOL talk

A ﬁnal extract is provided to illustrate the case where EOL discussion is not raised in any individuated way. The comparison of this extract with others shown suggests that the semantic variation captured through message semantics appears to be related to the presence or absence of EOL discussion. It should be noted that in the following consultation the doctor uses fewer conﬁrm questions, in particular fewer [check], [reassure] and [probe] questions, and fewer questions overall than the other consultations shown in this chapter. The consultation is however very long, and the doctor very attentive, with many elliptical [ask] questions posed, such as Any other concerns? The feature [prefaced] is used, but what is interesting in this consult – and arguably criterial – is that this uses his own viewpoint as the mediating perspective in [prefaced] messages, rather than the patient’s or carer’s viewpoint Alison Rotha Moore 103 as demonstrated in Consultations 101 and 105, and seen elsewhere in the corpus. Untaken opportunities to develop the patient’s contributions into discussion about EOL topics, including how much longer she is likely to live, and fears or concerns about that, occur in Consultation 2307 at turns 353, 355, 409 and 493.

Extract 5 – Consultation 2307, turns 350–494 (some turns omitted) 350 D I’m very pleased the way things are going. Did you have any questions at all about any of the medicines, anything at all? ⇒ 351 P Oh I am just concerned about the usage of morphine. 352 D OK what are your concerns with that? ⇒ 353 P Oh I normally hear that when people are on morphine they are at the end of the road. 354 D That’s the most common thing that people say. ⇒ 355 P How far at the end of the rope am I? 356 D OK. 357 P ((laughter)) 358 D The thing with morphine, is that what you’ve just said is a very common thought but it is totally inaccurate. We have patients on morphine for 5 or 6 years, ok, and we use morphine for 2 particular reasons. One is obviously pain ...All I can do is to re-assure you that in your situation it is very very safe. And your concerns are also very normal. ----- [turns omitted] 407 P I used to work at a nursing home. 408 D Yes. ⇒ 409 P Patients who had been given morphine, they they die shortly after. 410 D Well the reason they are usually given morphine is they’re about to die anyway, OK because they’re suffering. 413 P Mmmm. 414 D You know that they have pain. They have pneumonia ah and a lot of patients in the end-stage or the terminal phase are suffering. So we give them morphine to ease their suffering. ----- [turns omitted] 491 D Any other concerns? 492 P No. ⇒ 493 K Um my concern is her sleeping, ah wakefulness. And waking up in the middle of the sleep. 104 Society in Language, Language in Society

494 D So she’s not sleeping through the night ...Most people will wake up maybe once through the night and go straight back to sleep. And with the problems that you have, such as a bit of shortness of breath or a bit of a cough. The fact that you wake up doesn’t concern me too much. I’d be more concerned if you woke up ...

This doctor responds energetically to patient and kin questions, but in a different way from others displayed above, giving rich detailed information about the role of morphine which is relevant but may not have been [adequate] as a reply to the patient’s question how far at the end of the rope am I? Assuming that the patient in 2307 is really wanting to talk about her own ‘time left’, this has largely failed. And in contrast with the discussion in 105 about what ‘never gets discussed’, when it is the doctor who does not step through a possible ‘open gate’ for potential EOL talk, there is generally no metadiscussion about it, and this may be a problematic asymmetry.

4.15 The role of semantic variation in styles of palliative care practice

While the kind of message semantic features selected by a doctor cannot directly or necessarily produce or block EOL discussion, there does seem to be a predisposing relation of some kind between these phenomena. For instance, if we consider the doctor’s discussion of EOL in 2307, in terms of its generalizing semantics (what everyone says and what normally happens), in terms of professional ethics (what a palliative care doctor does and does not do, that is, relieve suffering, but not shorten life), and professional judgement (what the doctor, not the patient, is most and least concerned about), then taken together these features are consistent with each other. They are also semantically consistent with the message semantic features observed in his discourse, such as his selection of self-oriented [prefacing]20 and his non-selection of [probe] questions, which would hand topical development and relevance (but not next topic selection) to the other party. At the same time, the other doctors discussed in this chapter, from consultations 101, 105 and elsewhere, are also using consistent semantic orientations when they select other-oriented [prefacing], [probe] and other [verify] questions, and tease EOL discussion out of minute apertures. As in Hasan’s study of the discourse of maternal control, the contrast appears largely to be around the construal of individuation (cf. Martin, 2010). To put it another way, individuation is a site of intra-registerial variation for both contexts. But palliative care discussions are distinct in terms of their implicitness and incipience: even the less individuating dialogues Alison Rotha Moore 105 display tokens such as thinking about things, all of this stuff and what happens from here, along with sequences such as Do you have any concerns ...anything at all ...about what’s happening ...cards that are to play.21 In other words, inter- registerial difference is also indicated by the message semantic analysis of palliative care. A number of issues arise from this analysis, most of which can only be touched on in this chapter. An important one is that, if individualized concerns are on the agenda of the patient, then the semantics of individuation, incipience and implicitness are interdependent. As Hasan points out, different ideologies give rise to different evaluations of strategies such as prefacing, individuation and implicit speaking. Prefacing for instance can be taken as bolstering authority, or as a feature that is crucial to the construction of point of view (Hasan, 2009b; Cloran, 1989). Implicitness can be taken as respectful of the other’s right not to be confronted with unwelcome topics, which links it with individuation. These values in turn must be seen as related to, perhaps manifesting, features of the context of Western culture more generally, in which death and dying remain taboo (Exley, 2004; Seale, 1998). The creation of a profes- sionalized discourse that individuates patients appears to be one response to the awkwardness of death and dying, but one which also incorporates such awkwardness in a ‘routinized’ semantic orientation – or register (Linell and Bredmar, 2007; Sandén et al., 2001). Walter (1991) suggests that modern societies – as societies – handle and process death very well, but the individuals who are dying or bereaved become ‘uniquely isolated, lepers even, because they highlight the Achilles heel of the modern individual’ (Walter, 1991, quoted in Exley, 2004, p. 112). In other words, the model developed by Hasan brings out the ‘solidary’ nature of relations between wording and meaning (Hjelmslev, 1961) as well as the solidary nature of relations between meaning and context. It makes sense that prefacing should occur in environments of quite a different type in cultures where individuation is considered to be important, whether at home reading books in the early years of one’s life or in the doctor’s rooms discussing its end. The exact nature of these solidary relations requires more empirical and theoretical work, around questions such as the following:

• Is the predisposing relation between those semantic features selected and the extent of EOL talk taken up a predictive or a realizational relationship? • Do the three features isolated above (individuation, incipience and implicitness) constitute a kind of ‘badge’ of palliative care, or possibly of a female style of palliative care practice, or of a middle class style of palliative care practice? 106 Society in Language, Language in Society

• What are the implications for patients who are not middle class and/or Western, if access to the meaning systems that organize palliative care discussions are differentially distributed, based on patients’ social and cultural positioning?

This last point is particularly important and suggests a role for ‘variationist’ research across health discourse more generally (Williams, 2014). Although they are separate theoretical constructs, register cannot be corralled from code, as pointed out by Bernstein (1971) and as shown to be crucial to equity across a range of contexts, whether studied using message semantics or other analytical tools (e.g. Martin, 2012; Christie et al., 1991). Health status and health care provision are, like education, notoriously better for the middle class and resistant to equity interventions. Palliative care, like other specialisms, has produced research and resources to help meet the needs of patients from linguistically diverse backgrounds, but class-based coding variation of this kind does not appear to have got onto the clinical research radar and would require a different mix of strategies from the usual translation and interpretation approaches (Williams, 2003; Moore and Grossman, 2003).

4.16 Implications for the Hasanian message semantics project

At the outset of this chapter I claimed that Hasan’s semantic approach to register analysis deserves fuller exploration, testing across multiple registers, and most likely some revisions to accommodate the insights returned. I have explored the applicability of Hasan’s semantic options for the characterization and explication of palliative care discourse, particularly around the issue of whether and how EOL issues are developed in consultations between palliative care specialists, patients and carers. Without modification, Hasan’s contextually open network has provided very good coverage in bringing out relevant consistencies and patterns of divergence in the palliative care corpus. But to be maximally useful in this context I believe it would need further revision. From my reading, Hasan has not said whether the requirement that semantic networks be ‘contextually open’ means that both the terms in the system and their realizations must be non-variant for all contexts. Although it seems theoretically problematic, one possibility that seems empirically worth exploring is that the terms in the system might be the same for all contexts, but the realizations are at least somewhat context-specific. In the present chapter I have illustrated this idea with possible additional realizations for options arising from the feature [confirm] within the system of DEMAND INFORMATION. Across the palliative care corpus similar gaps have been identified in the network’s coverage of items, for instance items which have functions agnate to the features [apprize] and [constrained]. There are also commonly used Alison Rotha Moore 107 patterns, which might be considered variations in the way [prefaced] and [assumptive] are realized. These gaps are to be discussed in a later paper. In previous work on surgery, I suggested a similar idea, namely that in the context of surgical teamwork, the realizational possibilities for what counts as a demand for goods and services might be expanded, relative to other contexts (Moore, in press, and Moore’s worked example in Lukin et al., 2011). In a similar vein, Matthiessen et al. (2005) consider the extraordinary variability in the realization patterns for offers in the context of pizza ordering/marketing by telephone. While not using Hasan’s message semantic networks as an analytical frame, some other studies argue for and demonstrate accounts in which semantic-level features mediate the contextual variation observed in a similar way, such as O’Donnell’s (2001) study on the semantic feature ‘realis’ and tense in different news registers in English, and Caffarel (1992) on tense and time in French. Although appealing for empirical work (for the same reasons that context- specific networks appeal), if we allow realization statements to be variable by register, this raises theoretical difficulties and implications. One issue is that this way of thinking suggests that Halliday’s notion of context as the ‘input’ for the semantics, and semantics as the ‘input’ for the grammar (Halliday, 1972 [2003], p. 331), might be stating things too unilaterally. If the realizations of a given semantic feature vary according to context, then what counts as ‘meaning x’ depends on grammar in context. This suggests that semantic features might be better described as the product or output of the grammar in context, rather than the input to the grammar. Such a description of realization relations might be a better fit with the kind of data seen in this chapter. And it might fit the purposes of a model that eschews a ‘one- to-one’ relationship between semantics and grammar – see for instance the need to ‘net in’ dependencies between distal systems in message semantics, as in the case of counting certain expressions as [probe] questions in DEMAND INFORMATION, given appropriate selections from CONTINUATION. It is important to understand that Hasan’s networks do not represent isomorphic mapping between two strata, since for isomorphic mapping the two whole systems would need to be wired in the same way, converging and diverging at the same points. Giving a realization statement for each term is not one-to-one system equivalence. But having said that, the options in Hasan’s semantics of English do stay quite close to the options in Halliday’s grammatics of English – offering only a few ways of bundling up patterns of meaning that are distributed grammatically across different parts of the system. While this is deliberate and well founded from the perspective of ‘solidary relations’ I feel that the message semantic categories sometimes run out of tool power for mediating or calibrating grammatical and contextual analyses at some important places. One of those places for me has been the handling of agency at the semantic level. Possibilities do not seem to be available for unifying represented and enacted agentivity, and especially for 108 Society in Language, Language in Society tracking agentive options in the construal of semiotic action (Moore, 2004, 2005). In this sense the message semantic project appears to have yielded a different system of high-level semantic components from the type that Hasan called for in 1969 (Hasan, 1973). It should be noted that Hasan has also proposed the concept of ‘formative motif’ to handle ‘solidary’ combinations of message semantic features (2009b), but such motifs are not generally networked. Despite their ingenuity, it also seems difficult at this point to imagine how Hasan’s system of contextually open networks could produce the robust tristratal descriptions achieved in her non-networked accounts of specific contexts, which for me are best exemplified by her specification of the crucial and associated semantic components in the realization of the contextual move ‘Placement’ in nursery tales (Hasan, 1996b). But it is not that a networked approach cannot in principle handle such dispersed realizations of a unified semantic drift. The ‘prefacing’ network within Hasan’s message semantics seems to my mind to offer such a tool. It handles the comprehensive integration of different ways that the grammar produces a relatively consistent semantic effect, and therefore appears to be operating at about the semantic ‘level’ suggested in Hasan’s paper ‘Code, register and social dialect’ (1973). Additionally, the message semantic networks should be understood as part of a larger project to develop a rank scale for the semantics (see Hasan, 2013; Hasan et al., 2007), so it would be wrong to dismiss the idea of a contextually open networked semantics on the basis of the mapping of only one rank. After all, as Hasan points out (2009b, p. 65):

Surely there is a constant dialogue between theory and practice: their development is interdependent; descriptive means grow in the service of solving linguistic problems. To wait for the means to ﬁrst develop is like holding back a child from taking part in linguistic interactions until s/he has ‘mastered her/his mother tongue’.

4.17 Concluding remarks

This chapter has described Hasan’s network-based model of message semantics and located it within Hasan’s career as the central manifestation of her concern with developing functional linguistics into a workable and powerful tool for contributing to discourse analysis and sociolinguistics. The networked tool produced by Hasan is a remarkable achievement as a schematic for relating wording to meaning, and meaning to the living of life and the differentially distributed constraints and opportunities for human flourish- ing. As an empirical body of work Hasan’s own findings using this have enormous ongoing social and theoretical significance. The linguistic study of palliative care reported here in small part has been conducted on a much smaller scale of ingenuity and energy (and funding!). Alison Rotha Moore 109

It has not involved exhaustive semantic analysis or employed statistical tests, but it has provided initial evidence that semantic variation occurs within the context of palliative care and is likely to be a factor in how and why some people get a better palliative care experience than others. Although it identiﬁes some problem areas, it also helps justify Hasan’s demand for contextual openness in semantic networks. The study provides further support for Hasan’s claim that a networked semantics has the capacity to ‘systematize’ registerial variation. Hasan’s approach allows us to motivate and explain semantic features and orientations that are shared by, or which mark out, distinct contexts. These motivations and explanations can be expressed in terms of relations between linguistic strata – or in other words, in terms of meaning in context. With- out such contextual openness in descriptive tools, distinct contexts cannot be compared and register as a whole cannot be modelled.

Notes

1. Hasan does not rule out context-specific modelling but appears to see it as something one does when one cannot do contextually open modelling (Hasan et al., 2007). 2. Exceptions include the work of Matthiessen and colleagues (Matthiessen, 1993; Matthiessen et al., 2008 and others) and computational approaches aligned with SFL such as Teich (2003), Steiner (2008), Neumann (2013), etc. There are also ‘fellow traveller’ accounts such as those of Biber. 3. The name of this feature (and several others) has changed in Hasan’s model over time. 4. The rest of the question is omitted – see Lukin (2012). 5. Moore, Tuckwell, Butow, Tattersall and Clayton are completing further articles on this material. Other professions could benefit from better question typologies, for example police (Hall, 2008). 6. Transcripts were also analysed for process type, agency, voice, mood and theme (Halliday and Matthiessen, 2014). 7. According to the medical researchers, our data comprised the following three groups: ‘EOL issues not raised’ n = 9, ‘EOL issues raised but not discussed’ n = 5 and ‘EOL issues raised and discussed’ n = 32. Total 46. However, the linguistic researchers ended up undertaking their own analysis of whether EOL issues were discussed. A finding from our research is that there is relatively poor inter-rater reliability (between disciplines) on this variable. 8. Hasan’s message semantics has no internal grouping of experiential and logical meanings, hence there are four major networked semantic systems, corresponding to the experiential, logical, interpersonal and textual metafunctions. 9. ‘Choice’ here means the contrasts themselves, not the process of selection. But see Butt et al. (2013) and other chapters in Fontaine et al. (2013) for discussion of the complex notion of ‘choice’ in functional linguistics. 10. An earlier term Hasan used for this system was ‘role allocation’ (Hasan, 1996a). 11. This is also true for grammatical categories, for example ‘indicative’ in Halliday’s system (Halliday and Matthiessen, 2014). 110 Society in Language, Language in Society

12. For intonation, see Halliday and Greaves (2008); for grammatical terms (declarative, interrogative, mood, etc.), see Halliday and Matthiessen (2014) or earlier editions. 13. Note the term ‘elliptical probe’ is mine. This is a non-Hasanian realization of a [probe]. 14. By way of precedent, a tag is itself a kind of elliptical interrogative. Additional realization criteria could be specified using Hasan’s networks for classification (experiential meaning) and continuation (which pertains to the logical function in Hasan’s model, but deals with some phenomena modelled under ‘exchange’ elsewhere, notably Martin, 1992). There is no space to develop this point or alternative accounts, such as adding another option in [verify], or taking a ‘punctuative’ interpretation. 15. Note that this is a question from the patient. 16. This should be able to be handled using Hasan’s ‘classification’ network. 17. Patterns of graduation such as ‘problems – concerns – worries’ were observed by Kathryn Tuckwell and are the subject of a separate joint paper in preparation. 18. Hasan has ‘wired’ the networks to favour semantic contrasts over grammatical ones. Thus although [ask] and [check] questions are realized through different mood selections, they are internally grouped in the Demand Information network under the feature [enquire] partly because [ask] and [check] can each lead to the option [assumptive], whereas the options which are grammatically more similar to [check], namely [reassure] and [probe], do not constitute an environment from which [assumptive] can be chosen (see Hasan, 1996a). This is one reason why claims that Hasan’s semantic networks are in a one-to-one relationship with grammatical networks are not well founded. 19. All messages discussed in this chapter are [−assumptive] unless otherwise indicated. Some are [+prefaced] without being annotated in this chapter as having that feature. 20. I use ‘self-oriented [prefacing]’ as a short cut – it is not a term Hasan uses. 21. This sequence is from Consultation 2308, and will be discussed in a separate paper.

References

Bernstein, B. (1971) Class, Codes and Control,Vol.I:Theoretical Studies towards a Sociology of Language (London: Routledge & Kegan Paul). Berry, M. (1981) ‘Systemic linguistics and discourse analysis: a multi-layered approach to exchange structure’. In M. Coulthard and M. Montgomery (eds) Studies in Discourse Analysis (London: Routledge & Kegan Paul), pp. 120–45. Butt, D. G. (1983) ‘Semantic “drift” in verbal art’. Australian Review of Applied Linguistics 6 (1): 38–48. Butt, D. G., Henderson-Brooks, C., Moore, A., Meares, R., Haliburn, J., Korner, A. and Eyal, R. (2014) ‘Challenges for stylistic method: from literary enquiry to mental health treatments – when two methods are at one’. In Y. Fang and J. J. Webster (eds) Developing Systemic Functional Linguistics: Theory and Application (London: Equinox), pp. 298–322. Butt, D., Moore, A. and Tuckwell, K. (2013) ‘The teleological illusion in linguistic “drift”: choice and purpose in semantic evolution’. In L. Fontaine, T. Bartlett and G. O’Grady (eds) Choice: Critical Considerations in Systemic Functional Linguistics (Cambridge: Cambridge University Press), pp. 55–79. Alison Rotha Moore 111

Caffarel, A. (1992) ‘Interacting between a generalized tense semantics and register- speciﬁc semantic tense systems: a bi-stratal exploration of the semantics of French tense’. Language Sciences 14 (4): 385–418. Christie, F., Devlin, P., Freebody, P., Luke, A., Martin, J. R., Threadgold, T. and Walton, C. (1991) Teaching Critical Social Literacy: a Project of National Signiﬁ- cance on the Preservice Preparation of Teachers for Teaching English Literacy (Canberra: DEET). Clayton, J. M., Butow, P., Tattersall, M. H. N., Devine, R. J., Simpson, J., Aggarwal, G., Clark, K., Currow, D., Elliott, L., Lacey, J., Lee, P. J. and Noel, M. (2007a) ‘Randomized controlled trial of a prompt list to help advanced cancer patients and their caregivers to ask questions about prognosis and end-of-life care’. Journal of Clinical Oncology 25 (6): 715–23. Clayton, J. M., Hancock, K. M., Butow, P. N., Tattersall, M. H. N and Currow, D. C., with expert panel including Moore, A. and Tuckwell, K. (2007b) ‘Clinical practice guidelines for communicating prognosis and end-of-life issues with adults in the advanced stages of a life-limiting illness, and their caregivers’. Medical Journal of Australia 186 (12 Suppl.): S77, S79, S83–108. Cloran, C. (1989) ‘Learning through language: the social construction of gender’. In R. Hasan and J. R. Martin (eds) Language Development: Learning Language, Learning Cul- ture. Meaning and Choice in Language: Studies for Michael Halliday (New Jersey: Ablex), pp. 111–51. Cloran, C. (2000) ‘Socio-semantic variation: different wordings, different meanings’. In L. Unsworth (ed.) Researching Language in Schools and Communities: Functional Linguistic Perspectives (London: Cassell), pp. 152–83. Drew, P. and Heritage, J. (eds) (1992) Talk at Work: Interaction in Institutional Settings (Cambridge: Cambridge University Press). Exley, C. (2004) ‘Review article: the sociology of dying, death and bereavement’. Sociology of Health & Illness 26 (1): 110–22. Fontaine, L., Bartlett, T. and O’Grady, G. (eds) (2013) Systemic Functional Linguistics: Exploring Choice (Cambridge: Cambridge University Press). Hall, P. (2008) ‘Policespeak’. In J. Gibbons and M. T. Turell (eds) Dimensions of Forensic Linguistics (Amsterdam: John Benjamins), pp. 67–94. Halliday, M. A. K. (1972 [2003]) ‘Towards a sociological semantics’. In J. J. Webster (ed.) Collected Works of M. A. K. Halliday,Vol.3:On Language and Linguistics (London and New York: Continuum), pp. 323–54. Halliday, M. A. K. (1973) Explorations in the Functions of Language (London: Edward Arnold). Halliday, M. A. K. and Greaves, W. (2008) Intonation in the Grammar of English (London: Equinox). Halliday, M. A. K. and Matthiessen, C. M. I. M (2014) Introduction to Functional Grammar (4th edn) (London: Routledge). Harris, S. (1984) ‘Questions as a mode of control in magistrates’ courts’. International Journal of the Sociology of Language 49: 5–28. Hasan, R. (1973) ‘Code, register and social dialect’. In B. Bernstein (ed.) Class, Codes and Control, Vol. II: Applied Studies towards a Sociology of Language (London: Routledge & Kegan Paul), pp. 253–92. Hasan, R. (1983) ‘Message semantics networks and coding manual’. Unpublished mimeo, Department of Linguistics, Macquarie University. Hasan, R. (1989) ‘Semantic variation and sociolinguistics’. Australian Journal of Linguistics 9 (2): 221–75. 112 Society in Language, Language in Society

Hasan, R. (1992a) ‘Rationality in everyday talk: from process to system’. In J. Svartvik (ed.) Directions in Corpus Linguistics, Proceedings of Nobel Symposium 82, Stockholm (Berlin: de Gruyter). Reprinted in Hasan (2009b), pp. 313–56. Hasan, R. (1992b) ‘Meaning in sociolinguistic theory’. In K. Bolton and H. Kwok (eds) Sociolinguistics Today: International Perspectives (London: Routledge). Reprinted in R. Hasan (2009b), pp. 275–312. Hasan, R. (1996a) ‘Semantic networks: a tool for the analysis of meaning’. In C. Cloran, D. Butt and G. Williams (eds) Ways of Saying: Ways of Meaning: Selected Papers of Ruqaiya Hasan (London: Cassell), pp. 104–31. Hasan, R. (1996b) ‘The nursery tale as genre’. In C. Cloran, D. Butt and G. Williams (eds) Ways of Saying: Ways of Meaning: Selected Papers of Ruqaiya Hasan (London: Cassell), pp. 51–72. Hasan, R. (2009a) ‘Meaning in sociolinguistic theory’. In The Collected Works of Ruqaiya Hasan,Vol.2:Semantic Variation: Meaning in Society and in Sociolinguistics. Edited by J. J. Webster (London: Equinox), pp. 275–312. Hasan, R. (2009b) The Collected Works of Ruqaiya Hasan, Vol. 2: Semantic Variation: Meaning in Society and in Sociolinguistics. Edited by J. J. Webster (London: Equinox). Hasan, R. (2009c) ‘Questions as a mode of learning in everyday talk’. In The Col- lected Works of Ruqaiya Hasan,Vol.2:Semantic Variation: Meaning in Society and in Sociolinguistics. Edited by J. J. Webster (London: Equinox), pp. 231–68. Hasan, R. (2009d) ‘On semantic variation’. In The Collected Works of Ruqaiya Hasan, Vol. 2: Semantic Variation: Meaning in Society and in Sociolinguistics. Edited by J. J. Webster (London: Equinox), pp. 45–76. Hasan, R. (2009e) ‘Wanted: a theory for integrated sociolinguistics’. In The Col- lected Works of Ruqaiya Hasan,Vol.2:Semantic Variation: Meaning in Society and in Sociolinguistics. Edited by J. J. Webster (London: Equinox), pp. 5–40. Hasan, R. (2009f) ‘Language in the processes of socialization: home and school’. In The Collected Works of Ruqaiya Hasan,Vol.2:Semantic Variation: Meaning in Society and in Sociolinguistics. Edited by J. J. Webster (London: Equinox), pp. 123–83. Hasan, R. (2009g) ‘The place of context in a systemic functional model’. In M. A. K. Halliday and J. J. Webster (eds) Continuum Companion to Systemic Functional Linguis- tics (London and New York: Continuum), pp. 166–89. Hasan, R. (2013) ‘Choice taken in the context of realization’. In L. Fontaine, T. Bartlett and G. O’Grady (eds) Systemic Functional Linguistics: Exploring Choice (Cambridge: Cambridge University Press), pp. 269–99. Hasan, R. and Cloran, C. (2009 [1990]) ‘A sociolinguistic interpretation of everyday talk between mothers and children’. In The Collected Works of Ruqaiya Hasan, Vol. 2: Semantic Variation: Meaning in Society and in Sociolinguistics. Edited by J. J. Webster (London: Equinox), pp. 41–72. Originally published in condensed form in M. A. K. Halliday, J. Gibbons and H. Nichols (eds) Learning, Keeping and Using Language (Vol. 1) (Amsterdam: John Benjamins), pp. 67–99. Hasan, R., Cloran, C., Williams, G. and Lukin, A. (2007) ‘Semantic networks: the description of meaning in SFL’. In R. Hasan, C. M. I. M. Matthiessen and J. J. Webster (eds) Continuing Discourse on Language,Vol.2:A Functional Perspective (London: Equinox), pp. 697–738. Heritage, J. and Greatbatch, D. (1991) ‘On the institutional character of institutional talk: the case of news interviews’. In D. Boden and D. Zimmerman (eds) Talk and Social Structure: Studies in Ethnomethodology and Conversation Analysis (Cambridge: Polity Press), pp. 93–137. Alison Rotha Moore 113

Hjelmslev, L. (1961) Prolegomena to a Theory of Language (Madison: University of Wisconsin Press). Jakobson, R. (ed.) (1987) Language and Literature (Cambridge, Mass.: Belknap Press). Labov, W. and Fanshel, D. (1977) Therapeutic Discourse: Psychotherapy as Conversation (New York: Academic Press). Levinson, S. (1983) Pragmatics (Cambridge: Cambridge University Press). Linell, P. and Bredmar, M. (2007) ‘Reconstructing topical sensitivity: aspects of face work in talks between midwives and expectant mothers’. In T. van Dijk (ed.) Discourse Studies (Vol. 3) (London: Sage), pp. 418–44. Linell, P., Hofvendahlm, J. and Lindholm, C. (2003) ‘Multi-unit questions in institutional interactions: sequential organizations and communicative function’. Text 23: 539–71. Lukin, A. (2012) ‘Hasan’s semantic networks as a tool in discourse analysis’. In J. Knox (ed.) To Boldly Proceed: Papers from the 39th International Systemic Functional Congress (Sydney: ISFC Organizing Committee), pp. 141–6. Lukin, A., Moore, A., Herke, M., Wegener, R. and Wu, C. (2011) ‘Halliday’s model of register revisited and explored’. Linguistics and the Human Sciences 4 (2): 187–243. Maley, Y. and Fahey, R. (1991) ‘Presenting the evidence: constructions of reality in court’. International Journal for the Semiotics of Law 4 (1): 3–17. Martin, J. R. (1992) English Text: System and Structure (Amsterdam: Benjamins). Martin, J. R. (2010) ‘Semantic variation: modelling realization, instantiation and individuation in social semiosis’. In M. Bednarek and J. R. Martin (eds) New Discourse on Language: Functional Perspectives on Multimodality, Identity, and Afﬁliation (London: Continuum), pp. 1–34. Martin, J. R. (2012) Collected Works of J. R. Martin,Vol.6:CDA/PDA. Edited by Wang Zhenhua (Shanghai: Shanghai Jiao Tong University Press). Martin, J. R. and White, P. (2005) The Language of Evaluation: Appraisal in English (New York: Palgrave Macmillan). Matthiessen, C. M. I. M. (1993) ‘Register in the round’. In M. Ghadessy (ed.) Register Analysis: Theory and Practice (London: Pinter), pp. 221–92. Matthiessen, C. M. I. M., Lukin, A., Butt, D. G, Cleirigh, C. and Nesbitt, C. (2005) ‘A case study of multistratal analysis’. Australian Review of Applied Linguistics S (19): 123–50. Matthiessen, C. M. I. M., Teruya, K. and Wu, C. (2008) ‘Multilingual studies as a multi-dimensional space of interconnected language studies’. In J. J. Webster (ed.) Meaning in Context: Implementing Intelligent Applications of Language Studies (London: Continuum), pp. 146–220. Moore, A. (2004) ‘The discursive construction of treatment decisions in the management of HIV disease’. Unpublished PhD thesis, Macquarie University. Moore, A. (2005) ‘Modelling agency in HIV decision-making’. Australian Review of Applied Linguistics S (19): 103–22. Moore, A. R. (in press) ‘Lovers, wrestlers, surgeons: a contextually sensitive approach to modelling body alignment and interpersonal engagement in surgical teams’. In S. White and J. Cartmill (eds) Communication in Surgical Practice (London: Equinox). Moore, A. and Grossman, J. (2003) ‘Commentary (Assignments of meaning in epi- demiology by Miles Little)’. In J. M. Little, I. Kerridge, C. Jordens and E. J. Sayers (eds) Restoring Humane Values to Medicine: a Miles Little Reader (Annandale: Desert Pea Press), pp. 99–102. Neumann, S. (2013) Contrastive Register Variation: a Quantitative Approach to the Comparison of English and German (Berlin: de Gruyter Mouton). 114 Society in Language, Language in Society

O’Donnell, M. (2001) ‘Editorials vs front page news: differences in realis and tense across newspaper genres’. Proceedings of the 2nd International Contrastive Linguistics Conference, 25–27 October (Compostela, Spain). Sandén, I., Linell, P., Starkhammar, H. and Larsson, U. S. (2001) ‘Routinization and sensitivity: interaction in oncological follow-up consultations’. Health 5 (2): 139–63. Seale, C. (1998) Constructing Death: the Sociology of Dying and Bereavement (Cambridge: Cambridge University Press). Semino, E., Demjen, Z. and Koller, V. (2014) ‘ “Good” and “bad” deaths: narratives and professional identities in interviews with hospice manager’. Discourse Studies 16 (5): 667–85. Steiner, E. (2008) ‘Explicitation: towards an empirical and corpus-based methodology’. In J. J. Webster (ed.) Meaning in Context: Implementing Intelligent Applications of Language Studies (London: Continuum), pp. 235–78. Teich, E. (2003) Cross-Linguistic Variation in System and Text: a Methodology for the Investigation of Translations and Comparable Texts (Berlin: Mouton de Gruyter). Tsui, A. (1992) ‘A functional description of questions’. In J. Sinclair and M. Coulthard (eds) Advances in Spoken Discourse Analysis (London: Routledge), pp. 89–110. Walter, T. (1991) ‘Modern death: taboo or not taboo?’ Sociology 25 (2): 293–310. Weiner, J. and Labov, W. (1983) ‘Constraints on the agentless passive’. Journal of Linguistics 19 (1): 29–58. Williams, G. (1995) ‘Joint book-reading and literacy pedagogy: a socio-semantic examination’. Unpublished doctoral thesis, Macquarie University. Williams, G. (2001) ‘Literacy pedagogy prior to schooling: relations between social positioning and semantic variation’. In A. Morais, I. Neves, B. Davies and H. Daniels (eds) Towards a Sociology of Pedagogy: the Contribution of Basil Bernstein to Research (New York: Peter Lang), pp. 17–45. Williams, G. (2003) ‘The determinants of health: structure, context and agency’. Sociology of Health and Illness 25 (3): 131–54. Williams, G. (2005) ‘Semantic variation’. In R. Hasan, C. Matthiessen and J. Webster (eds) Continuing Discourse on Language: a Functional Perspective (Vol. 1) (London: Equinox), pp. 457–80. Williams, G. (2014). ‘Semantic variation theory as appliable linguistics in healthcare communication research’. Presentation at Sydney SFL Friday seminars, 28 March. Wright, A., Zhang, B., Ray, A., Mack, J., Trice, E., Balboni, T., Mitchell, S., Jackson, V., Block, S., Maciejewski, P. and Prigerson, H. (2008) ‘Associations between end- of-life discussions, patient mental health, medical care near death, and caregiver bereavement adjustment’. JAMA 300 (14): 1665–73. 5 Hasan’s Semantic Networks Revisited: a Cantonese Systemic Functional Approach1 Andy Fung

5.1 Introduction

In systemic functional linguistics (SFL), Ruqaiya Hasan and her colleagues have placed much emphasis on the stratum of semantics. Such emphasis reﬂects her view that the study of meaning is indeed a central aspect of SFL, and theoretically shares the same status as the study of lexicogrammar. For Hasan, semantics and lexicogrammar are ‘two sides of the same coin’ (Hasan et al., 2007, p. 699). That is to say, without lexicogrammar, meaning could not be manifested in our human social practices; and by the same token, without meaning, lexicogrammar would have no reason for existence in human language (Hasan et al., 2007; Hasan, 2010). Systemic functional semantics, in Halliday’s (2009, p. 82) words, is considered to be a separate stratum in the model of language — the ‘linguistic inter-level to context’ (Matthiessen, 1993, p. 227) (see e.g. also Halliday, 2009; Hasan et al., 2007; Hasan, 2010). There are various points of departure concerning semantics in SFL literature: (1) stylistic analysis, particularly the study of literary text (e.g. Butt and Lukin, 2009), (2) computational linguistics, with a particular focus on text generation parsing (e.g. Mann, 1984; Mann and Matthiessen, 1985; Matthiessen, 1988; Halliday and Matthiessen, 1999), (3) sociological linguistics and semantic variation (e.g. Halliday, 1973; Turner, 1973; Hasan, 1983, 1996; Hasan and Cloran, 1990 [2009]) and (4) the collaborative and interactive exchange of dialogue (e.g. Halliday, 1984; Martin, 1992; Matthiessen, 1995; Eggins, 1990; Eggins and Slade, 2004). This chapter falls under the third perspective, ‘sociological linguistics and semantic variation’ and concerns the notion of ‘semantic networks’, with a particular focus on Hasan’s (1983) contextually open semantic networks. The networks have been used in such diverse areas as the study of semantic variation in carer–child discourse (e.g. Hasan, 1983; Hasan and Cloran, 1990; Williams, 1995), courtroom discourse (Maley and Fahey, 1991), classroom

115 116 Society in Language, Language in Society interaction (Wake, 2006; Chu, 2011), police interviews (Hall, 2004), service encounters (Hasan et al., 2007), surgical interactions (Lukin et al., 2011) and media interviews (Lukin, 2012). This large and developing body of literature powerfully demonstrates the centrality of semantic networks in discourse analysis. However, with the exception of only a few research projects (e.g. Wong, 2009), most research into semantic networks has been conducted using English texts. This chapter thus aims at demonstrating that Hasan’s semantic networks are extendable to languages other than English. Here, I report on a research project aimed at extending Hasan’s (1983) contextually open semantic networks to Cantonese. I ﬁrst situate Hasan’s semantic networks in the framework of SFL then review the speciﬁc concepts of these networks. Building on systemic functional typology research, I argue that Hasan’s semantic networks could be extended into languages other than English. The data set for this chapter comprises several doctor–patient interactions in Hong Kong. Focusing on questions within the data set, I develop a fragment of semantic networks and their respective lexicogrammatical realization statements. This chapter ends with an illustration of the research implications for the use of semantic networks in discourse analysis.

5.2 Situating Hasan’s semantic networks in SFL

To understand how Hasan’s semantic networks are conceptualized, one must first understand the architecture of language from the perspective of SFL theory. SFL recognizes six semiotic dimensions: hierarchy of language, cline of instantiation, metafunction, rank scale, delicacy and axis (see Halliday and Matthiessen, 2013, ch. 1). Briefly, adult language in SFL2 is regarded as a ‘multi-stratal’ semiotic system,3 with an internal organization of semantics, lexicogrammar and phonology and phonetics. As asserted by Hasan (2013, p. 276), there is ‘no adult language with less than four strata’. In addition, there is one stratum above it, which serves to illustrate the environment where the text is functioning, viz. context. This highest stratum, together with the language-internal strata, forms the hierarchy of stratification. Hasan’s concept of ‘semantic network’, as the name suggests, is located at the level of semantics, enabling us to understand why and how something is said. As Hasan (2009, p. 170, my emphasis) writes,

To put it simply, to explain why anyone says anything one must appeal to the context which exerts pressure on the speaker’s choice of meaning; and to explain why these patterns of wordings appear rather than any other, one must appeal to the meanings which, being relevant to the context, activated those wordings: semantics is thus an interface between context and linguistic form.

Viewing the strata vertically, there is a dialectic relation of realization functioning across context, semantics and lexicogrammar. When looking from Andy Fung 117 above, the contextual choices activate the semantic choices, which in turn further activate the lexicogrammatical ones. When looking from below, the lexicogrammatical choices construe the semantic choices, and in turn further construe the contextual conﬁguration (Hasan, 2009). Important in this ‘activation–construal dialectic’ (Hasan, 2009, p. 170) is that it is not of ‘deter- ministic relationship’ but a matter of tendency (Bowcher, 2010, p. 73, see also Butt and Wegener, 2007; Bowcher, 2010, 2014; Halliday and Hasan, 1985; Hasan, 1995, 2009, etc.). As asserted by Bowcher (2014, p. 180), it is

a realisational-assemblage in the sense that it is a simultaneous patterning of events at different levels of abstraction, not causal in nature but dialectic, where features specific to each level of abstraction show a specific configuration, re-sorted by the simultaneous and interdependent process of realisation.

The non-determining, dialectic realization between context and semantics thus implies that semantic options might be well interpreted as a correlation with the contextual conﬁguration (CC) (Halliday and Hasan, 1985; Hasan, 1995, 1999, 2009, etc.), rather than being linked to individual contextual parameters. Viewing each stratum horizontally, context, semantics and lexicogrammar extend along the cline of instantiation, extending from the instance pole to the potential pole. Since semantic networks are a ‘hypothesis about patterns of meaning’ (Halliday, 1973 [2003], p. 327) aiming at specifying the total meaning potential, they are located at the potential pole of the cline of instantiation. As meaning potentials resonate with contextual demands, semantic networks are thus metafunctionally organized, resulting in four simultaneous networks. Features of each network are identiﬁed as semantic options, which are organized along the syntagmatic and paradigmatic axes. Viewing along the syntagmatic axis, each semiotic option acts as the entry condition of other more delicate options. The total set of selected semantic options is termed a ‘selection expression’ (e.g. Halliday, 1973; Hasan, 1996, Hasan et al., 2007). When viewing along the paradigmatic axis, each semantic option is grammaticalized, known as lexicogrammatical realization.

5.3 Hasan’s contextually open semantic networks in English

The incentive of developing semantic networks is not new in SFL literature. The first chapter in the history of semantic networks opened in the 1970s (see Turner, 1973; Halliday, 1973 [2003]). Halliday (1973 [2003]) develops his semantic networks based on sociological and linguistic concerns, asserting that the verbal behaviour of a speaker could be described in both social and grammatical systems since they are of the same phenomenal realm. How- ever, these two descriptions could not be simply related because the social system is ‘wholly outside language’ and the grammatical system is ‘wholly 118 Society in Language, Language in Society within language’ (Halliday, 1973 [2003], p. 347). To reconcile these descriptions, Halliday proposes the notion of semantic network, for semantics is the stratum intermediate between the social system and the grammatical system (Halliday, 1973 [2003], p. 345). In this sense, semantic network is defined as a ‘hypothesis about patterns of meaning’, in which meanings are represented in terms of ‘paradigmatic relations on the semantic stratum’ (Halliday, 1973 [2003], p. 327). Significant developments in semantic networks were witnessed in the 1980s, fuelled by Hasan’s strong research needs when investigating the meaning productions in mother–child talk. This research pressure motivated Hasan to construct her own semantic networks in 1983. Rather than following Turner’s and Halliday’s context-specific, sociological approach, Hasan (1983, 1996, 2009, 2010, 2013) conceptualized her semantic networks in a distinctive manner, particularly in terms of open context, and rank scales of semantics and metafunctions.4

5.3.1 Semantic networks and open context A unique point of departure of Hasan’s semantic networks is that they are contextually open, which could be related to her conceptualization of ‘context’. According to Hasan (1996), building a contextually open system network is advantageous for it is able to deal with ‘within context variation’ (Hasan, 1996, p. 115). By distinguishing the terms ‘specific context’ and ‘unique context’, Hasan suggests that a specific context is not just a unique context. For example, there could be two systemically unique but related contexts that belong to one specific context. A contextually open network could thus generate a ‘multiplicity of agnate contexts’ for one specific context (Hasan, 1996, pp. 114–15). Another justification for open context is ‘context permeability’. Hasan (1973, 1995) has shown that contexts are permeable, suggesting that distinctive context-specific semantic networks will eventually become permeable as the semantic specific descriptions are developed (Hasan, 1996, p. 115). In other words, the ‘actualized’ semantic choices in turn construe a wide range of contexts. Based on these conceptualizations, Hasan suggests that her networks are ‘representations of the potential’ of a language (Hasan, 1996, p. 115).

5.3.2 Semantic networks and rank Hasan (1996) proposes a four-unit rank scale in studying semantics, moving from the largest to the smallest: text ∼ rhetorical unit5 ∼ message ∼ text radical.6 Among the four scales, Hasan suggests that message serves as the point of origin of semantic networks. According to Hasan, ‘the message is the smallest semantic unit which is capable of realising an element of the structure of a text’ (Hasan, 1996, p. 117). It is realized as a ranking clause at the lexicogrammatical level, except if the clause is a projecting one. In Hasan’s view, a message with a projecting clause is a selection of the feature [prefaced], a semantic option which enables us access to the speaker’s viewpoints. Andy Fung 119

One strong implication of establishing the unit of semantic networks is that meaning analyses are distinguished from other semiotic systems but solidly founded on grammar.

5.3.3 Semantic networks and metafunction Contra Turner’s (1973) and Halliday’s (1973 [2003]) networks, Hasan incorporates Halliday’s metafunctional hypothesis into her semantic networks so that they are metafunctionally organized. Hasan (1989, p. 244) ﬁrst outlined the four simultaneous vectors of a message as follows:

(1) system of interpersonal meanings, for example options in message function (questioning, informing, commanding, etc.); options in personal evaluation, point of view, etc.; (2) systems of experiential meaning, for example the ascription of actional, evolutional etc. roles, identiﬁcation, deﬁnition; construction of time, etc.; (3) systems of logical meaning, for example cause, condition, and meta- textual relations, etc.; (4) systems of textual meanings, for example options in topic maintenance, topic changes, etc.

Hasan (1996) further labels these four vectors of a message, terming ROLE ALLOCATION7 as the system of interpersonal meanings, CLASSIFICATION as the system of experiential meaning, AMPLIFICATION as logical meaning and CONTINUATION as textual meaning. It should be added immediately that not all messages are metafunctionally organized. Hasan further introduces a basic distinction, where message is either [punctuative] or [progressive]. It is only the semantic option [progressive] which serves as the entry condition to these metafunctionally regulated networks. Each of these networks in turn develops into its individual system networks, resulting in different sets of semantic options. A strong consequence of metafunctionally regulated semantic networks is that subtle meaning differences can be captured since each message undergoes multiple metafunctional interpretations.

5.4 Research motivation

Though language speciﬁcity has not been discussed in Hasan’s previous works, one can infer from the following quote concerning her networks that they are literally language-speciﬁc:

...displaying semantic options available to speakers of English for attempting to ﬁnd out some information from someone. (Hasan and Cloran, 1990, p. 80, my emphasis) 120 Society in Language, Language in Society

Given the fact that Hasan’s semantic networks are well situated within the dimensions of systemic functional theory, a general theory which enables us to describe all languages in a systemic functional manner, I argue that Hasan’s semantic network could be redeveloped in languages other than English. Indeed, functional linguists not only concentrate on English but have attempted to globalize Halliday’s functional theory to languages other than English, an area of investigation known as systemic functional typology research/multilingual research (Matthiessen, 2009, p. 18). The past decade has witnessed a signiﬁcant expansion in research concerning systemic functional typology. In 2004, an edited volume entitled Language Typology: a Functional Perspective was introduced by Caffarel, Martin and Matthiessen, a development fuelled by workshops held at the University of Sydney in 1996 and 1999 focusing on systemic functional descriptions of various languages (Caffarel et al., 2004, p. xi). This pioneer volume explores clausal grammars of eight languages, including French (Caffarel, 2004), German (Steiner and Teich, 2004), Tagalog (Martin, 2004), Vietnamese (Thai, 2004), Mandarin Chinese (Halliday and McDonald, 2004), Japanese (Teruya, 2004), Telugu (Prakasam, 2004) and Pitjantjatjara (Rose, 2004), covering seven language families (Caffarel et al., 2004, p. xi). Here, I argue that multilingual research and Hasan’s semantic networks stand in a relationship of mutualism. On the one hand, Hasan’s semantic networks can contribute a systematic description on semantics, a stratum which has not yet been of great focus in systemic functional typology research/multilingual research. Existing typological research tends to focus on the lexicogrammar, with few attempts to focus on the semantic stratum (e.g. McDonald, 1998; Li, 2003, 2007 for Chinese). Yet, the point of departure of these few attempts lies on the exchange structures of dialogues. Meanwhile, the robust descriptions of lexicogrammar, which are the solid foundation of multilingual research, can help explore how semiotic options are realized lexicogrammatically. Lexicogrammatical realization is a key concern in doing semantic networks since it is hypothesized that ‘a difference in linguistic “form” equals a difference in meaning’ (Hasan, 2010, p. 283). The extensive lexicogrammatical systems generated from these typological research studies thus enable us to grammaticalize the meanings explicitly. Given these research implications, studies embracing these two areas of Hasan’s semantic networks and systemic functional typology are scarce. In my foray into the literature, there is only one initial attempt by Wong (2009), who has extended Hasan’s network of punctuative message to Cantonese, with a particular focus on the particles in a computer-mediated communication (CMC) context. The semantic network of progressive messages is currently lacking. To continue the development of Cantonese semantic networks, the remainder of this chapter attempts to extend the semantic network of progressive messages in Cantonese, with a particular focus on the networks for ‘asking questions’. Andy Fung 121

5.5 Guiding principles in extending Hasan’s semantic network

Prior to presenting the semantic networks of ‘asking questions’ in Cantonese, it is necessary to illustrate some ways to extend the English networks to Cantonese. The experience gained from modelling lexicogrammatical patterns in language typology research sheds light on the network extension. As highlighted by Caffarel et al. (2004, p. 7), when accounting for the lexicogrammatical systems in languages other than English, one needs to treat the language concerned ‘in its own right’ rather than being ‘anglocentric’. Thus, a related question is: in what ways can we extend the semantic networks so that they are appropriate to a language in its own right? To facilitate the extension of semantic networks, I draw on Halliday’s (1992 [2003]) work. Halliday suggests a set of principles, which, in his view, correspond closely to the general position scientists adopt so that the language descriptions will be objective and scientiﬁcally valid, and analogous to ‘doing science’ (Lemke, 1990). Though the focus of Halliday’s principles is based on ‘doing grammar’ (Halliday, 1992 [2003], p. 211), his ideas, in my view, are crucial and should be considered when doing semantics, particularly when extending the semantic networks from English to Cantonese. Building on Halliday’s principles in doing grammar, a set of guiding principles is proposed as follows:

1. Semantic options that are used in the analysis of meaning are general concepts, which help us to explain the semantic phenomena. They are not ‘endowed with a spurious reality of their own’ (Halliday, 1992 [2003], p. 200). 2. The semantic options are of two kinds: theoretical and descriptive. The- oretical semantic options are ‘general to all languages’ while descriptive semantic ones are ‘language specific’ and are ‘redefined in the case of each language’ (Halliday, 1992 [2003], p. 201). 3. Both theoretical and descriptive semantic options are defined ‘in relation one to another’ but not individually (Halliday, 1992 [2003], p. 202). 4. All the postulated descriptive semantic options are validated by taking a trinocular perspective: (i) from above, (ii) from around, and (iii) from below (Halliday, 1992 [2003], p. 204). Thus, the validity of the semantic networks would be confirmed if the semantic options construe the recognizable contextual features, and if they construe the lexicogrammatical patterns and if the semantic options relate to each other systematically. 5. When doing a comparative description, one should acknowledge the fact that the existing semantic options in Hasan’s semantic network are only available to English speakers. Though they could be set as the point of departure in modelling the semantic networks in Cantonese, 122 Society in Language, Language in Society

considerations should be given to whether there are comparable meanings in the language under description. 6. In naming the semantic options, it is sensible to reuse the existing terms. However, it is important to note that such names have to be reinter- preted, deﬁned and explained within the language under description (Halliday, 1992 [2003], p. 207).

So far, this chapter has illustrated how Hasan’s English semantic network could be extended through systemic functional typology research. In the remaining section, I demonstrate the construction of Cantonese semantic networks based on these guiding principles.

5.6 Cantonese semantic networks: an initial attempt

The main texts that I use to illustrate the extension of the semantic networks are categorized under the heading ‘patients’ journeys’, which were collected in one local emergency department in a Hong Kong hospital. These texts are of a larger study conducted by a healthcare research team at the Department of English at the Hong Kong Polytechnic University8 (Matthiessen, 2013). Doctor–patient communication is of focus for it is complex and highly stressed in nature. Developing a linguistic resource in analysing this social practice will shed light on what types of meanings are at risk, or in Butt and Wegener’s (2007, p. 589) words, the ‘more probable selections’, thus contributing to our understanding of this register as well as medical safety.

5.6.1 Primary systemic options in the system message When considering the ‘point of origin’ of semantic networks, Hasan deﬁnes the semantic unit ‘message’ as the entry condition, arguing for a basic distinction between [progressive] and [punctuative]. Hasan interprets the former as the messages entailing the exchange of propositions or proposals, which move the discourse of the text forward. It serves as the entry condition of the metafunctionally regulated systems, viz. RELATION ENACTMENT,9 CLASSIFICATION, AMPLIFICATION and CONTINUATION. According to Hasan, [progressive] messages in English are realized by a [major] clause, i.e. a clause in which there is a Predicator. They are described as follows:

[progressive] 1) preselect option [major]; 2) insert element Predicator; 3) preselect (an instance of) verbal group at Predicator (Hasan, 1989 [2009], p. 245)

Messages which fail to move the discourse forward are termed [punctuative], covering two main functions, i.e. locutionary guidance and expressive guidance (Hasan, 1996, p. 118). Punctuative messages of the former function Andy Fung 123 to ‘guide the flow of interaction’ (Hasan, 1996, p. 118), the phenomenon which is referred to as ‘conversation management devices’ (Ventola, 1979), or ‘pragmatic particles’, ‘presequences’ and ‘channel repairs’ in the literature of pragmatics (Levinson, 1983). Punctuative messages of this type reveal the textual orientation since they often punctuate the discourse into stages. Typical examples include I see, that’s it, pardon, right,etc. Punctuative messages of the expressive guidance type orient to the interpersonal metafunction, signalling the affect of the speakers towards the listeners. Messages of this type are concerned with speech role allocation management and the expression of emotions such as surprise, joy or dis- gust (Hasan, 1996, p. 119). In English, [punctuative] messages are realized in [minor] clauses or clauses with formulaic meanings like how are you?, bless you!, oh shit!, great! etc. The choice of [punctuative] may therefore be more delicately classified as either [text-oriented] or [person-oriented], where the former concerns locutionary guidance and latter expressive guidance (see Wong, 2009, p. 90). Following Hasan’s proposition, I define the point of origin of Cantonese semantic networks as ‘message’, the smallest semantic unit in Cantonese which is ‘capable of realising an element of the structure of a text’ (Hasan, 1996, p. 117). It is realized as a ranking and non-projected clause in the lexicogrammatical stratum. If this view is valid, the next question is to determine whether messages in Cantonese are capable of ‘furthering’ the text. Dialogue 1 illustrates the primary systemic options in the system of message of Cantonese.

Dialogue 1: Doctor–Crystal dyad Turn S Msg ID Messages

154 D: 1 Crystal Crystal ‘Crystal’

155 P: 2 hai6 hai6 hai6 hai6 yes yes yes yes ‘Ye–ye–ye–yes.’

156 D: 3 nei5hou2 hello ‘Hello’

D: 4 jau5 mou3 joek6mat6man5gam2 aa3? Have not-have drug allergy PRT Predicator ‘Do you have any drug allergies?’ 124 Society in Language, Language in Society

157 P: 5 e6 ... jau5 aa3, Ah have PRT Predicator ‘Ah ...yeah,’

P: 6 jau5 zek3 tau4wan4 go2di1 have CL dizzy that type’ Predicator ‘there’s one type for dizziness, that type’

158 D: 7 dim2 man5gam2 gaa3? How allergic PRT Predicator ‘How allergic is it?’

159 P: 8 e6 ... ngo5 sik6 zo2 keoi5 ne1, ah I eat ASP it PRT Predicator ‘Ah ...I take this’

P: 9 wui2, e6, go3 zeoi2 me2 gaa3. Will ah CL mouth twist PRT Predicator ‘my mouth would, ah, twist to one side.’

P: 10 e6 ... ni1 zek3. Ah this CL ‘Ah, this one.’

160 D: 11 zeoi2 me2? Mouth twist Predicator ‘Your mouth twists?’

161 D: 12 me2 maai4 jat1 bin6 aa4? Twist ASP one side PRT? Predicator ‘Twists to one side, doesn’t it?’

162 P: 13 hai6 aa3 hai6 aa3! Yes PRT yes PRT ‘Right–Right’

D: 14 o2 o2 o2. Uh uh uh ‘Uh–uh–uh.’ Andy Fung 125

D: 15 hou2 laa1! OK PRT ‘OK!’

162 D: 16 gin3 me1 si6 aa3? see what matter PRT? Predicator ‘What’s the matter?’

Dialogue 1 is a typical example of the opening phase of a doctor–patient consultation in an emergency department where the doctor and patient first encounter each other. As shown in this dialogue, Cantonese messages function to either move the discourse forward or serve as locutionary and expressive guidance. To initiate the consultation, the doctor strategically starts with [punctuative: person-oriented: address], as in message 1 e.g. ‘Crys- tal’. The effect of this preselection is to establish the social roles between doctor and patient. Under the emergency context, it is only the healthcare practitioners who have the authority to check the patient’s identity, which is the doctor in this case. In this sense, the institutional roles of ‘doctor’ and ‘patient’ are established. Having established the social roles, the doctor does not go straight to the history-taking phase but starts with a [punctuative: person-oriented: greeting] message 3 ‘nei5hou2’ (Hello), where the effect is to establish the social relation with the patient. Here, the patient is not perceived as a subject of a bio-psychosocial model but rather as ‘patient-as- person’, suggesting the patient is the centre of the consultation. Following these messages is a series of [progressive] messages forming question–answer sequences. Here, the doctor and the patient co-construct the consultation by ‘furthering’ the discourse forward, where the doctor asks questions (e.g. messages 4, 7, 11, 12) and the patient gives answers (e.g. messages 5, 6, 8, 9, 13). As the discourse unfolds, the doctor selects [punctuative: text-oriented: maintain] messages right after the patient’s reply, as in message 14 ‘oh-oh- oh’ (Uh-uh-uh), signalling that the doctor is attentive to the patient’s talk. Message 15 ‘hou2laa1!’ (OK!), by contrast, selects the feature [punctuative: text-oriented: frame], where the effect is to punctuate the flow of discourse. Here, the doctor signals that he terminates the existing topic, e.g. ‘drug allergy’ and is about to start a new one. It should be added immediately that though Hasan’s terms are reused here, they are interpreted only in a Cantonese sense. As a result, the lexicogrammatical realization statements of these options must be re- examined. Take [progressive] messages as an example. A review of all the Cantonese [progressive] messages identified in the data suggests that the predicators vary, ranging from verbal group, adjective or preposition. This finding is in line with Tam’s (2004) study on systemic functional Cantonese 126 Society in Language, Language in Society

Text-oriented...

Punctuative PROGRESSIVENESS Message Person-oriented... Progressive...

Figure 5.1 Primary options of message semantics in Cantonese grammar where the predicator of a clause may be realized by verbs, adjectives, adverbs or a preposition. Realizations for Cantonese [progressive] and [punctuative] messages are summarized as follows:

[progressive] (1) preselect option [major]; (2) insert element Predicator; (3) preselect (an instance of) verbal group, adjective or preposition as Predicator. [punctuative] preselect option [minor];

The semantic options for PROGRESSIVENESS are summarized in Figure 5.1.

5.6.2 Semantic network of RELATION ENACTMENT in Cantonese In Hasan’s message semantics, the semantic network speciﬁes choices related to the four metafunctions; each message is analysed in terms of options under experiential, interpersonal, textual and logical meanings. Interper- sonal meanings are concerned with a ‘speaker’s subjectivity and the enactment of speaker–addressee relations in interaction’ (Hasan, 1992 [2009], p. 281). These have been usefully explored by Hasan (e.g. 1989 [2009], 1996, 1991 [2009], 1992 [2009]) and Hasan and Cloran (1990 [2009]), who have proposed a number of sub-networks relevant to the enactment relationships between interactants. These are the systems within the system of RELATION ENACTMENT. According to Hasan, the system of RELATION ENACTMENT comprises two simultaneous systems with options either [give] versus [demand], as well as [information] versus [goods-and-services], resulting in four semantic categories: ‘statement’ [give; information], ‘question’ [demand; information]; ‘command’ [demand; goods-and-services] and ‘offer’ [give; goods-and-services]. These semantic categories are in turn realized lexicogrammatically in ‘mood and modality’ (Hasan et al., 2007, p. 712). On the assumption that Cantonese can be well described within the metafunctional framework, the next step is to examine whether the options [give], [demand], [information] and [goods-and-services] exist in Cantonese. The message function of each message in Cantonese can be deduced by the probe question ‘What kind of contextual pressure does a speaker act on the addressee discursively though the speech role and exchange of commodity?’ Andy Fung 127

Dialogue 1 illustrates most of the subsystems of RELATION ENACTMENT. For example, through messages 4, 7, 11 and 12, the doctor obtains medical information from the patient. Here, the doctor takes on the role of information seeker; and thus these messages select the feature [demand; information]. By selecting this feature, the patient is positioned as information provider and is invited to provide verbal responses, as in messages 5, 6, 8, 9, 10 and 13. These messages thus select the feature [give; information]. The patient also takes on the role of goods-and-services provider where she presents the allergic drug to the doctor in message 10. In this sense, the feature [give; goods-and-services] is selected. Thus, from the perspective of speech role and commodity of exchange we can see that RELATION ENACTMENT does exist in Cantonese. The semantic options for RELATION ENACTMENT are summarized in Figure 5.2.

AMPLIFICATION ...

CLASSIFICATION ... Statement Give ...

Progressive Question Demand ...

RELATION ENACTMENT

Command ... Information

Goods-and- Offer ... CONTINUATION services ...

Figure 5.2 Systemic options of RELATION ENACTMENT in Cantonese

Let me further elaborate the system of RELATION ENACTMENT by focusing on ‘question’ or [demand; information] in terms of the selection expression. Building on Hasan’s (Hasan and Cloran, 1990 [2009], pp. 96–9) definition, the semantic network of Cantonese questions can be defined as a systemic network illustrating semantic options available to Cantonese speakers who attempt to find out some information from someone.

5.6.3 Semantic options in asking yes/no questions in Cantonese When considering the feature [demand; information] in English conversation, Hasan argues for a basic distinction regarding the type of information being solicited, that is, between the yes–no answer or specific information. Hasan (1991 [2009], p. 243) interprets the former feature as [confirm], referring it to the questions which aim to elicit ‘a yes/no response’, and the 128 Society in Language, Language in Society latter as [apprize] which attempts to seek some ‘specific element of information’. In English, the option [confirm] is realized lexicogrammatically as a polar clause, with [apprize] as a non-polar clause (Hasan, 2010, p. 293). Like English, Cantonese shares such a kind of distinction. Consider dialogue 1 again. Message 4 ‘jau5 mou3 joek6mat6man5gam2 aa3?’(Do you have a drug allergy?) selects the feature [confirm]. Here, the doctor projects a context hoping to elicit a yes/no answer; either jau5 (have)ormou3 (not have) is adequate to answer the question. By contrast, message 7 ‘dim2 man5gam2 gaa3’(How allergic is it?) selects the semantic feature [apprize]. As shown in messages 5 and 6, the patient has responded that she has a drug allergy. The option [confirm] is not at issue since drug allergy has already been established as a common ground between the doctor and patient, or more specifically, the negotiated ‘topic’. It is only the kind of drug allergy that has not been specified. To fill this cognitive gap, the doctor selects the semantic option [apprize] projecting a context signalling an absence of information and an attempt to determine the drug allergy. Here, the patient is encouraged to expand the unspecified topic. Had the patient responded with a yes/no answer, a communication breakdown would have resulted. In determining whether a question selects the semantic option [confirm] or [apprize], we can test with the prompt ‘What type of answer is presupposed by the enquirer?’ It should be added immediately that these semantic options only serve as the interpretation of the state of desire of the enquirer; and one cannot determine them directly by considering the responses of the listeners. In some cases, responses may be delayed, irrelevant, indirect or even absent (cf. semantic network of answers in Williams, 1995). The semantic option [confirm] in turn acts as the entry condition for other meaningful choices. Let us consider dialogue 2, the interaction following the last utterance of dialogue 1 in which the doctor initiates a new topic by asking the question ‘gin3me1si6aa3?’ (What’s the matter?).

Dialogue 2: Doctor–Crystal dyad

Turn S Msg ID Messages

163 P: 17 e6 ... ngo5 gam1 ciu4 cat1 dim2 zung1 hei2 san1 ne1 ah I this morning seven o’clock wake up PRT Predicator ‘Ah ...when I got up at seven o’clock this morning,’

P: 18 zau6 tau4 wan4 laa3, at once dizzy PRT Predicator ‘I was dizzy’

P: 19 go3 jan4 ne1 dam4dam4zyun2 gam2joeng2 CL person PRT spinning like this Predicator ‘I was spinning and such.’ Andy Fung 129

164 D: 20 m6, Mm ‘Mm’

D: 21 hou2 aa3 good PRT ‘Okay.’

D: 22 cat1 dim2 zung1 hei2, = =hai6 mai2 aa3? seven o’clock start ==right-not-right PRT Predicator ‘At seven o’clock you==u start dizzy, right?’

165 P: 23 = =hei2 –hei2 cong4 wake up bed Predicator ‘==Out–out of bed.’

166 D: 24 = = jau5 mou5 bat1sing2jan4si6 go2 di1 aa3? ==have- not-have unconsciousness like that PRT Predicator ‘==Did you pass out and such?’

167 P: 25 keoi5 ... zik1hai6 dou1 cing1sing2 it ... means still awake Predicator ‘It ...I mean, still awake’

P: 26 daan6hai6 [27] zau6 zong1haa2 zong1haa2 but at once dipped dipped Predicator ‘I dipped and dipped.’

P: 27 hang4 hei2 soeng6 lai4 ne1 walk ASP up come PRT Predicator ‘when I was walking’

Here, the patient justifies her reason for admission as shown in messages 17, 18 and 19. Interestingly, the scope of information in the answer goes beyond the point of enquiry of the question. The patient elaborates the answer by detailing when the dizziness occurs and the extent of the dizziness, resulting in three new information foci: (1) she feels dizzy, (2) she wakes up at seven o’clock and (3) the dizziness is so serious that she feels as though she is spinning. Based on the information solicited, the doctor infers that the patient develops her dizziness around seven o’clock in the morning but not the day before or so. To validate the inference, the doctor selects the feature [confirm] by asking a yes/no question. In Cantonese, the different ways of construing a yes/no question reflect the speaker’s presupposition with respect to the answers, whether he is pre- supposing an agreement (either positive or negative) or a neutral answer 130 Society in Language, Language in Society

(Matthews and Yip, 2011, p. 359). To capture this subtle distinction, I term the former as the semantic feature [verify] and the latter as [enquire]. When the semantic option [verify] is selected in Cantonese, the speaker is explicitly asking the addressee for a response which he expects. Lexicogrammatically, it is realized as [indicative: declarative: tagged], where the clause is organized into two parts: a declarative and a tag with a juncture in between. The inference being asserted lies in the declarative, and verification is sought through the tag. Contra English, question tags in Cantonese do not respect the polarity of the clause and are realized in particles like ho2 and ha2 or lexical phrases such as hai5-m4-hai5, hai5-mai2, hai5-maa3, ha2-maa3, ha2-waa3 meaning ‘right-not-right’ which is similar to isn’t it, don’t you in English (Luke, 1990; Wu, 1990, 1996; Tam, 2004; Matthews and Yip, 2011). As shown in message 22, the doctor here is explicitly asking the patient for a response which he expects; specifically an agreement on the truth of the inference embedded in the declarative. In other words, the semantic feature of [verify] could be glossed as ‘This is what I inferred. Please confirm that I am right.’ Had the doctor selected the feature [enquire], the presupposition held by the speaker would have been altered. In that case, the speaker would have no assumption or preconception of the truth of the inference; that is, ‘the chance of a positive or negative answer in the mind of the question, is just 50 : 50’ (Tang, 1972, p. 58). The lexicogrammatical realization statement of [verify] is thus summarized as follows:

[verify] major: indicative: declarative: tagged (i) insert tag (ii) tag preselects (1) particles ho2, ha2,OR (2) lexicalized phrases hai5-m4-hai5, hai5-maa3 and hai5-mai2 (iii) tag ∧ #

In English, Hasan argues an attitudinal distinction between the features [ask] and [check], where the former is ‘the most neutral way of attempting to elicit a yes/no response’ (Hasan, 2009 [1991], p. 246). By neutral, Hasan means that it is attitudinally unmarked. In other words, questions preselecting [ask], in Hasan’s view, could be regarded as straight questions in the sense that there is no hidden or implied meanings, or ‘no string attached’ in Hasan’s (2010, p. 293) words. Lexicogrammatically, the feature [ask] is realized by a polar interrogative and tone 2, signalling a neutral attitude (Halliday and Greaves, 2008, p. 116). By contrast, the feature [check] indicates the speakers’ attitudes towards the proposition as well as the listeners (Hasan, 2010, p. 293). Questions preselecting [check] are thus attitudinally marked, whether that may be Andy Fung 131 surprise, disappointment, enthusiasm, etc. These attitudinal meanings are realized by pitch movements, or more specifically, the choices of the system of KEY in English (cf. Halliday and Matthiessen, 2013; Halliday and Greaves, 2008, etc.). For example, the feature [check] is realized as untagged declarative and tone 2 which signals a sense of challenge towards the compliant listeners (Halliday and Greaves, 2008, p. 113). Hasan (1989 [2009], p. 211) paraphrases the [check] question as ‘I have some reservations against conceding that this is a true state of affairs: tell me what you think.’ The Cantonese semantic features [ask] and [check], however, are conceptualized in a slightly different manner.10 Rather than making an attitudinal distinction for options [ask] and [check], I define the semantic feature [ask] in Cantonese as the unbiased way of eliciting information. When the semantic option [ask] is selected, the enquirer projects a context declaring his uncertainty concerning the polarity of the thesis that is at issue. To remove this uncertainty, the enquirer simply makes a query by selecting a neutral polarity, creating a rhetorical effect of being ‘unbiased’ (Halliday and McDonald, 2004, p. 335). That is, the enquirer has ‘no built-in point of departure’ (Halliday and McDonald, 2004, p. 335). Questions selecting [ask] could be paraphrased as ‘I don’t know if it is the case or not. Please let me know’ and the compliant listener is given an impression that he or she could feel free to affirm or negate the question. In Cantonese, the neutral polarity of a question is maintained through various means. One way is to select the ‘A-not-A’ interrogative marker, by placing the negative marker between the first Predicator and the reduplicated Predicator (e.g. Wu, 1990, 1996; Tam, 2004; Cheung, 2007; Matthews and Yip, 2011). Both message 4 in dialogue 1 and message 24 in dialogue 2 select the semantic option [ask] and are realized by clauses selecting the mood choice [indicative: interrogative: polar: non-particle: A-not-A]. In some cases, the negative marker and the reduplicated Predicator fuse together. Take message 4 as an example. The negative marker ‘mut6’(not) fuses and contracts with the reduplicated predicator ‘jau5’(have) to form the lexical item ‘mou3’ (not-have). The lexicogrammatical realization of the option [ask] is proposed as follows:

[ask] major: indicative: interrogative: polar: non-particle (1) insert negative marker (2) preselect A-not-A interrogative marker (3) A conﬂates with Predicator

The neutral polarity of the semantic option [ask] may also be realized by clauses selecting the mood choices [indicative: interrogative: polar: particle]. These particles are known as ‘question particles’ or ‘interrogative particles’ in traditional Cantonese grammar (Kwok, 1984; Wu, 1990, 1996; Matthews and Yip, 2011; Leung, 2005, etc.). However, not all particles can realize the 132 Society in Language, Language in Society semantic option of [ask]. Typical particles realizing the semantic option [ask] include maa3, aa1maa3 and me6. Consider Examples 1 and 2:11

Example 1 ∗ D: nei5 gin3 hou2 di1 maa3 you feel good a bit PRT Predicator ‘Do you feel better now?’

In Example 1, the doctor would like to know if the patient feels better or not. Instead of constructing the question with the A-not-A marker, the enquirer simply adds the particle maa3 at the end, which in turn trans- forms the declarative into a question. Indeed, the particle maa3 could be regarded as a combination of the negative marker m4 and the particle aa3 resulting from negative-ﬁnal particle contraction (e.g. Wu, 1990, 1996; Leung, 2005). In this sense, the polarity of the whole question is still neutral. The second lexicogrammatical realization of the option [ask] is proposed as follows:

[ask] major: indicative: interrogative: polar: particle (1) insert particle (2) particle preselect maa3, aa1maa3 etc. (3) particle ∧ #

While the former lexicogrammatical realizations denote a neutral aspect in Cantonese, the other type of lexicogrammatical realization of the semantic option [ask] has a realizing clause which selects the mood choice [indicative: interrogative: polar: non-particle: negator], denoting a perfective aspect. In Cantonese, mei6 means ‘not yet’, which implies a negative meaning. By putting mei6 at the end of the utterance, the polarity of the whole question is maintained neutral. Consider Example 2:

Example 2 ∗ D: nei5 sik6 zo2 joek6 mei6 aa3 you take ASP medicine not yet PRT Pre -dicator “Have you taken the pill yet?”

Here, the doctor would like to know if the patient has taken the pill yet. To highlight the perfective aspect, the doctor again selects the features [ask]. Unlike the previous example, the neutral polarity of the question is realized by the negator mei6, suggesting the perfective aspect of the activities. Had Andy Fung 133 the doctor realized the question with A-not-A or maa3, the utterance would have become ‘nei5 sik6-m4-sik6 joek aa3’or‘nei5sik6joek maa3’, meaning ‘Would you like to take the pill’, which is an offer instead of a question. The third lexicogrammatical realization of the option [ask] is proposed as follows:

[ask] major: indicative: interrogative: polar: non-particle (1) insert negator (2) negator preselect mei6 (3) negator ∧ #

The contrasting option of [ask] is the semantic option [check], which is defined as an attempt ‘to draw an attestation from the addressee’ (Hasan, 1991 [2009], p. 246). Similar to the feature [ask], the enquirer projects a context signalling his uncertainty. However, the uncertainty projected lies in the validity of his inference of what is heard. To ascertain the validity, he simply makes a statement selecting either positive or negative polarity. The selection of polarity creates a rhetorical effect of ‘biasness’ since the enquirer makes his built-in point of departure explicit (Halliday and McDonald, 2004, p. 335) and the respondent is given an impression that he or she could only confirm or deny the validity. Similar to [check] in English, [check] in Cantonese is ‘attitudinally marked’. When the option [check] is selected, the enquirer may express his strong sense of reservation or challenge to the hearer, which is often marked by rising intonation on the final syllable. For example, in message 11, the intonation of the final syllable me2 (twist)ismarkedbya rising tail, suggesting that the doctor has some ‘reservations against conceding that what is heard is valid’ (Leung, 2005, p. 81), i.e. mouth twisting is a kind of drug allergy. Message 11 could be paraphrased as ‘I do have reservations about what you said. Do you mean this? Please tell me if I am right or not.’ Presumably, the doctor recognizes that the patient does not share the same understanding of ‘drug allergy’. To ascertain a common ground is established, the doctor makes another [check] question by making his inference explicit ‘me2 maai4 jat1 bin6’ (twist to one side). Here, the doctor selects a falling intonation, enforcing a certainty of his inference. Here, message 12 could be paraphrased as ‘This is what I infer or I heard. Please tell me if I am right or wrong.’

5.7 Discussion and implications

It is impossible within the scope of a chapter to develop a full account of the semantic network of asking questions in Cantonese, detailing each semantic option as well as the respective lexicogrammatical realization statements. Yet, the attempt so far is sufﬁcient to point to the fact that Hasan’s semantic networks can be redeveloped for languages other than English, as shown in Figure 5.3. 134 Society in Language, Language in Society

Give

Verify Demand Confirm Ask Question Enquire

Apprize... Information Check

Goods-and-services

Figure 5.3 A tentative semantic network of asking questions in Cantonese

A strong implication of developing a Cantonese semantic network is that it could complement the current descriptions of healthcare communication research in Hong Kong. While the major mode of doctor–patient communication is conducted in Cantonese, the analyses of current linguistic research studies are mostly based on English translations (e.g. Cheng et al., 2013). An ever-present danger in analysing translated text concerns the quality of the multilingual translation. While some would kindly claim that the translation quality can be maintained if the target texts are checked by native speakers, judging the quality of translation is not an easy task. As Halliday (2001, p. 14) has remarked,

It is notoriously difﬁcult to say why, or even whether, something is a good translation, since this must depend on a complex variety of different factors that are constantly shifting in their relationship one to another.

Identifying and minimizing ‘semiotic risks’ in healthcare settings have long been the key objectives of quantitative healthcare communication research (Matthiessen, 2013, p. 458). Yet, linguistic research in which analyses are based on translated texts is fraught with problems at the level of meaning. That is, it further puts meanings at risk because translation is indeed a semiotic act and risks changes in meaning. The concept of translation shift has been well discussed in Matthiessen (2001).12 My own perception is that in doing linguistic studies in a clinical context, with a particular focus on the meaning-making strategies, we cannot afford meaning loss or distortion since semiotic risks are highly relevant to patient safety. SFL, being an ‘appliable linguistics’ (Halliday, 1985, p. 7), is problem-oriented and ‘it is designed to assist towards identifying and tackling problems that arise from outside itself’ (Halliday, 2009, p. 61). Under this premise, the proposed networks, which are well situated in the SFL tradition, could enable us to fully explore how meanings are construed and realized in the healthcare context. Andy Fung 135

5.8 Conclusion

In this chapter, I have revisited the notion of semantic networks, a fruitful research tool for the analysis of meaning. I have also explored how systemic typology research could inform Hasan’s semantic networks. This study thus contributes a new front of development in the research into and application of Hasan’s semantic networks, which has not yet been readily taken up by systemic functional researchers. The past decades have witnessed a strong momentum in systemic functional typology research, with a particular focus on compiling metafunctional proﬁles in the lexicogrammar (see Caffarel et al., 2004). With the experience gained in these studies, I believe it is time for us to divert our focus from lexicogrammar to the stratum above, given its important status: as an interface between context and lexicogrammar. As I remarked earlier, Hasan’s semantic networks and language typology research exist in a relation of mutualism. The extension of Hasan’s semantic networks demonstrated in this chapter also contributes a new dimension in studying ‘systemic functional Cantonese semantics’, an area of investigation which is currently lacking. No doubt much remains to be done in specifying and organizing the meanings in Cantonese. Preliminary and indicative as the presented networks are, it is hoped that this chapter has provided a small step in the work of systemic functional semantics, as well as in the discourse of doctor–patient communication.

Notes

1. This chapter is an elaboration of the conference presentation at the 40th Interna- tional Systemic Functional Congress, held at Sun Yat-sen University, 15–19 July 2013. I am very grateful to Geoff Williams and Ruqaiya Hasan for their intellectual discussions on semantic networks, and Wendy Bowcher who revised and gave me valuable feedback on earlier drafts of this chapter. 2. In SFL, stratiﬁcation occurs within the development from protolanguage to adult language, a process known as ontogenosis (see Halliday, 1973; Halliday and Matthiessen, 2013, ch. 1, etc.). By ‘protolanguage’, it means child language, which is perceived as a combination of two strata planes (i.e. content and expression). During the ontogenetic process, these two planes are elaborated to multi-stratal semiotic systems. 3. While adult language is generally perceived as ‘tri-stratal’, treating phonology and phonetics as one expression stratum (e.g. Caffarel et al., 2004, p. 10), I follow Hasan’s idea that adult language is a ‘multi-stratal’ semiotic system. ‘Multi- stratal’ means the content plane of protolanguage is divided into semantics and lexicogrammar, and the expression plane is developed into phonology and phonetics, resulting in four distinct strata (Hasan, 2009, 2013; see also Halliday and Greaves, 2008; Halliday and Matthiessen, 2013, etc.). 4. Lack of space precludes a full account of Hasan’s semantic network. The following discussions only cover the most distinctive features of Hasan’s account of 136 Society in Language, Language in Society

semantic networks. For a complete overview of the conceptualization of Hasan’s semantic networks, see Hasan (1996, 2013) and Hasan et al. (2007). 5. The notion of rhetorical unit (RU) is pioneered by Cloran (1994), with a particular focus on the degree of (de-)contextualization of messages in discourse. Since a text is made up of one or more RUs, rhetorical unit analysis can thus realize the contextual conﬁguration of a register (Cloran, 1994, p. 128). See Cloran (1999) and Lukin (2010) for illustrations. 6. In Hasan’s (2013) recent discussion, text radical is termed as seme. 7. The term ROLE ALLOCATION is used in Hasan’s previous works (e.g. Hasan, 1996, 2009; Hasan et al., 2007). However, as stated in Hasan’s most recent work, the term ROLE ALLOCATION is replaced by RELATION ENACTMENT for it is the most inclusive term indicating the interpersonal orientation of a message (Hasan, 2013). 8. See http://www.engl.polyu.edu.hk/hcc/ 9. Though ROLE ALLOCATION has been used in Hasan’s previous works (see note 6), following Hasan’s recent (2013) proposition, I will use the term RELATION ENACTMENT in this chapter. 10. While Hasan argues that the basic distinction of [check] and [ask] lies in the attitudinal distinction, I believe this distinction is less valid in Cantonese. In Cantonese, an [ask] question, the most neutral way in soliciting information, is always attitudinally marked. When there is an absence of clause-ﬁnal particle, it denotes a sense of rudeness. By the same token, when particle such as maa3 is present, it represents a sense of politeness, functioning as tone-softening. In this sense, attitudinal distinction appears less valid in differentiating [ask] and [check] and further examinations are required. 11. Examples 1 and 2 are not found in the data. 12. Translation shift is well discussed in Matthiessen (2001). For example, Matthiessen has illustrated that even the metafunctions of the source text and the target text are preserved; considerable translation shifts could occur across and within ranks, in structure and even register.

References

Bowcher, W. L. (2010) ‘The history and theoretical development of context of situation in Systemic Functional Linguistics’. Annual Review of Functional Linguistics 2: 64–93. Bowcher, W. L. (2014) ‘Issues in developing unified systems for contextual field and mode’. Functions of Language 21 (2): 176–209. Butt, D. G. and Lukin, A. (2009) ‘Stylistic analysis: construing aesthetic organisation’. In M. A. K. Halliday and J. J. Webster (eds) Continuum Companion to Systemic Functional Linguistics (London: Continuum), pp. 190–215. Butt, D. G. and Wegener, R. (2007) ‘The work of concepts: context and metafunction in the Systemic Functional model’. In R. Hasan, C. Matthiessen and J. J. Webster (eds) Continuing Discourse on Language,Vol.2:a Functional Perspective (London: Equinox), pp. 589–618. Caffarel, A. (2004) ‘Metafunctional profile of the grammar of French’. In A. Caffarel, J. R. Martin and C. M. I. M. Matthiessen (eds) Language Typology: a Functional Perspective (Amsterdam and Philadelphia: John Benjamins), pp. 77–138. Caffarel, A., Martin, J. R. and Matthiessen, C. M. I. M. (2004) Language Typology: a Functional Perspective (Amsterdam and Philadelphia: John Benjamins). Andy Fung 137

Cheng, W., Chung, W. Y. J. and Wong, M. C. (2013) ‘Gender communication: nurse– client interactions in the mobile integrative health centre’. Linguistic Research 12: 139–57. Cheung, H. N. S. (2007) (A Grammar of Cantonese as Spoken in Hong Kong) (Hong Kong: Chinese University of Hong Kong). Chu, P. Y. (2011) ‘Picture book reading in a new arrival context: a multimodal perspective on teaching reading’. Unpublished PhD thesis, University of Adelaide. Cloran, C. (1994) Rhetorical Units and Decontextualisation: an Enquiry into Some Relations of Context, Meaning and Grammar. Monographs in Systemic Linguistic Linguistics, No. 6 (Nottingham: Department of English Studies, University of Nottingham). Cloran, C. (1999) ‘Context, material situation and text’. In M. Ghadessy (ed.) Text and Context in Functional Linguistics: Systemic Perspectives (Amsterdam and Philadelphia: John Benjamins), pp. 177–217. Eggins, S. (1990) ‘Keeping the conversation going: a systemic-functional analysis of conversational structure in casual sustained talk’. Unpublished PhD thesis, University of Sydney. Eggins, S. and Slade, D. (2004) Analyzing Casual Conversation (London: Equinox). Hall, P. (2004) ‘Prone to distortion? Undue reliance on unreliable records in NSW Police Service formal interview model’. In J. Gibbons, V. Prakasam, K. V. Tirumalesh and H. Nagrajan (eds) Language in the Law (New Delhi: Orient Longman Private Limited), pp. 44–81. Halliday, M. A. K. (1973) Explorations in the Functions of Language (London: Edward Arnold). Halliday, M. A. K. (1973 [2003]) ‘Towards a sociological semantics’. In J. J. Webster (ed.) Collected Works of M. A. K. Halliday,Vol.3:On Language and Linguistics (London and New York: Continuum), pp. 323–54. Halliday, M. A. K. (1984) ‘Language as code and language as behaviour: a systemic- functional interpretation of the nature and ontogenesis of dialogue’. In R. P. Fawcett, M. A. K. Halliday, S. Lamb and A. Makkai (eds) The Semiotics of Culture and Language, Vol. 1: Language as Social Semiotic (London and New York: Pinter), pp. 3–35. Halliday, M. A. K. (1985) ‘Systemic background’. In J. D. Benson and W. S. Greaves (eds) Systemic Perspectives on Discourse,Vol.1:Selected Theoretical Papers from the 9th International Systemic Workshop (Norwood, NJ: Ablex Publishing Corporation), pp. 1–15. Halliday, M. A. K. (1992 [2003]) ‘Systemic grammar and the concept of a “science of language” ’. In J. J. Webster (ed.) Collected Works of M. A. K. Halliday,Vol.3:On Language and Linguistics (London and New York: Continuum), pp. 199–212. Halliday, M. A. K. (2001) ‘Towards a theory of good translation’. In E. Steiner and C. Yallop (eds) Exploring Translation and Multilingual Text Production: Beyond Content (Berlin and New York: Mouton de Gruyter), pp. 13–18. Halliday, M. A. K. (2009) ‘Methods – techniques – problems’. In M. A. K. Halliday and J. J. Webster (eds) Continuum Companion to Systemic Functional Linguistics (London: Continuum), pp. 59–86. Halliday, M. A. K and Greaves, W. S. (2008) Intonation in the Grammar of English (London: Equinox). Halliday, M. A. K. and Hasan, R. (1985) Language, Context and Text: Aspects of Language in a Social Semiotic Perspective (Geelong, Vic.: Deakin University Press). Halliday, M. A. K. and McDonald, E. (2004) ‘Metafunctional proﬁle of the grammar of Chinese’. In A. Caffarel, J. R. Martin and C. M. I. M. Matthiessen (eds) Language 138 Society in Language, Language in Society

Typology: a Functional Perspective (Amsterdam and Philadelphia: John Benjamins), pp. 305–96. Halliday, M. A. K. and Matthiessen, C. M. I. M. (1999) Construing Experience through Meaning: a Language-Based Approach to Cognition (New York: Continuum). Halliday, M. A. K. and Matthiessen, C. M. I. M. (2013) An Introduction to Functional Grammar (4th edn) (London: Routledge). Hasan, R. (1973) ‘Code, register and social dialect’. In B. Bernstein (ed.) Class, Codes and Control, Vol. II: Applied Studies towards a Sociology of Language (London: Routledge & Kegan Paul), pp. 253–92. Hasan, R. (1983) ‘A semantic network for the analysis of messages in everyday talk between mothers and their children’. Mimeo, Macquarie University. Hasan, R. (1989 [2009]) ‘Semantic variation and sociolinguistics’. Australian Journal of Linguistics 9: 221–75. Reprinted in Hasan (2009), pp. 180–230. Hasan, R. (1991 [2009]) ‘Questions as a mode of learning in everyday talk’. In Thao Le and M. McCausland (eds) Language Education: Interaction and Development: Proceed- ings of the Interactional Conference (Faculty of Education, University of Tasmania), pp. 70–119. Reprinted in Hasan (2009), pp. 231–68. Hasan, R. (1992 [2009]) ‘Meaning in sociolinguistic theory’. In K. Bolton and H. Kwok (eds) Sociolinguistics Today: International Perspectives (London: Routledge), pp. 80–119. Reprinted in Hasan (2009), pp. 271–308. Hasan, R. (1995) ‘The conception of context in text’. In P. H. Fries and M. Gregory (eds) Discourse in Society: Systemic Functional Perspectives (Norwood, NJ: Ablex), pp. 183–296. Hasan, R. (1996) ‘Semantic networks: a tool for the analysis of meaning’. In C. Cloran, D. Butt and W. Williams (eds) Ways of Saying: Ways of Meaning (London: Cassell), pp. 104–31. Hasan, R. (1999) Speaking with reference to context. In M. Ghadessy (ed.) Text and Context in Functional Linguistics (Amsterdam and Philadelphia: John Benjamins), pp. 219–328. Hasan, R. (2009) The Collected Works of Ruqaiya Hasan,Vol.2:Semantic Variation: Meaning in Society and in Sociolinguistics. Edited by J. J. Webster (London: Equinox). Hasan, R. (2010) ‘The meaning of “not” is not in “not” ’. In A. Mahboob and N. K. Knight (eds) Appliable Linguistics: Texts, Contexts and Meanings (London: Continuum), pp. 267–305. Hasan, R. (2013) ‘Choice, system, realization: describing language as meaning potential’. In L. Fontaine, T. Bartlett and G. O’Grady (eds) Systemic Functional Linguistics: Exploring Choice (Cambridge: Cambridge University Press), pp. 269–99. Hasan, R. and Cloran, C. (1990 [2009]) ‘A sociolinguistic interpretation of everyday talk between mothers and children’. In M. A. K. Halliday, J. Gibbons and H. Nichols (eds) Learning, Keeping and Using Language (Vol. 1) (Amsterdam and Philadelphia: John Benjamins), pp. 67–99. Reprinted in Hasan (2009), pp. 75–118. Hasan, R., Cloran, C., Williams, G. and Lukin, A. (2007) ‘Semantic networks: the description of meaning in SFL’. In R. Hasan, C. M. I. M. Matthiessen and J. J. Webster (eds) Continuing Discourse on Language,Vol.2:A Functional Perspective (London: Equinox), pp. 697–738. Kwok, H. (1984) Sentence Particles in Cantonese (Hong Kong: Centre of Asian Studies, University of Hong Kong). Lemke, J. L. (1990) Talking Science: Language, Learning, and Values (Language and Educational Processes) (Norwood, NJ: Ablex). Levinson, S. C. (1983) Pragmatics (Cambridge, New York: Cambridge University Press). Andy Fung 139

Leung, C. S. (2005) A Study of the Utterance Particles in Cantonese as Spoken in Hong Kong (Hong Kong: Language Information Sciences Research Centre, City University of Hong Kong). Li, E. S. E. (2003) ‘A text-based study of the grammar of Chinese from a systemic functional approach’. Unpublished PhD thesis, Macquarie University. Li, E. S. E. (2007) Systemic Functional Grammar of Chinese: a Text-based Analysis (London: Continuum). Luke, K. K. (1990) Utterance Particles in Cantonese Conversation (Amsterdam and Philadelphia: John Benjamins). Lukin, A. (2010) ‘ “News” and “register”: a preliminary investigation’. In A. Mahboob and N. Knight (eds) Appliable Linguistics (London: Continuum), pp. 92–113. Lukin, A. (2012) ‘Hasan’s semantic networks as a tool in discourse analysis’. In J. Knox (ed.) To Boldly Proceed: Papers from the 39th International Systemic Functional Congress (Sydney: ISFC Organizing Committee. ISBN 978-0-646-58257-3), pp. 141–6. Lukin, A., Moore, A, Herke, M., Wu, C. and Wegener, R. (2011) ‘Halliday’s model of register revisited and explored’. Linguistics and the Human Sciences 4 (2): 187–213. McDonald, E. (1998) ‘Clause and verbal group systems in Chinese: a text-based functional approach’. PhD thesis, Macquarie University. Maley, Y. and Fahey, R. (1991) ‘Presenting the evidence: constructions of reality in the court’. International Journal for the Science of Law 4 (1): 3–17. Mann, W. (1984) ‘A linguistic overview of the Nigel text generation grammar’. In A. Manning, P. Martin and K. McCalla (eds) The Tenth LACUS Forum 1983 (Columbia, SC: Hornbeam Press), pp. 255–65. Mann, W. and Matthiessen, C. M. I. M. (1985) ‘Demonstration of the Nigel text generation computer program’. In J. Benson and W. S. Greaves (eds) Systemic Perspectives on Discourse: Selected Papers from the 9th International Systemic Workshop (Vol. 1) (Norwood, NJ: Ablex), pp. 50–83. Martin, J. R. (1992) English Text: System and Structure (Amsterdam and Philadelphia: John Benjamins). Martin, J. R. (2004) ‘Metafunctional proﬁle of the grammar of Tagalog’. In A. Caffarel, J. R. Martin and C. M. I. M. Matthiessen (eds) Language Typology: a Functional Perspective (Amsterdam and Philadelphia: John Benjamins). pp. 255–304. Matthews, S. and Yip, V. (2011) Cantonese: a Comprehensive Grammar (London: Routledge). Matthiessen, C. M. I. M. (1988) ‘Representational issues in systemic functional grammar’. In J. D. Benson and W. S. Greaves (eds) Systemic Perspectives on Discourse: Selected Papers from the 9th International Systemic Workshop (Vol. 1) (Norwood, NJ: Ablex), pp. 136–75. Matthiessen, C. M. I. M. (1993) ‘Register in the round: diversity in a uniﬁed theory of register analysis’. In M. Ghadessy (ed.) Register Analysis: Theory and Practice (London: Pinter), pp. 221–92. Matthiessen, C. M. I. M. (1995) Lexicogrammatical Cartography: English Systems (Tokyo: International Language Sciences Publishers). Matthiessen, C. M. I. M. (2001) ‘The environments of translation’. In E. Steiner and C. Yallop (eds) Exploring Translation and Multilingual Text Production: Beyond Content (Berlin: Mouton de Gruyter), pp. 41–124. Matthiessen, C. M. I. M. (2009) ‘Ideas and new directions’. In M. A. K. Halliday and J. J. Webster (eds) Continuum Companion to Systemic Functional Linguistics (London: Continuum), pp. 12–58. 140 Society in Language, Language in Society

Matthiessen, C. M. I. M. (2013) ‘Applying systemic functional linguistics in healthcare contexts’. Text and Talk 33 (4–5): 437–66. Prakasam, V. (2004) ‘Metafunctional profile of the grammar of Telugu’. In A. Caffarel, J. R. Martin and C. M. I. M. Matthiessen (eds) Language Typology: a Functional Perspective (Amsterdam and Philadelphia: John Benjamins), pp. 433–78. Rose, D. (2004) ‘A metafunctional profile of the grammar of Pitjantjajara’. In A. Caffarel, J. R. Martin and C. M. I. M. Matthiessen (eds) Language Typology: a Functional Perspective (Amsterdam and Philadelphia: John Benjamins), pp. 479–536. Steiner, E. and Teich, E. (2004) ‘Metafunctional profile of the grammar of German’. In A. Caffarel, J. R. Martin and C. M. I. M. Matthiessen (eds) Language Typology: a Functional Perspective (Amsterdam and Philadelphia: John Benjamins), pp. 139–84. Tam, H. S. (2004) ‘A systemic functional interpretation of Cantonese clause grammar’. Unpublished PhD thesis, University of Sydney. Tang, T. C. (1972) A Case Grammar of Spoken Chinese (Taipei: Hai-Guo Book Co). Teruya, K. (2004) ‘Metafunctional profile of the grammar of Japanese’. In A. Caffarel, J. R. Martin and C. M. I. M. Matthiessen (eds) Language Typology: a Functional Perspective (Amsterdam and Philadelphia: John Benjamins), pp. 185–254. Thai, M. D. (2004) ‘Metafunctional profile of the grammar of Vietnamese’. In A. Caffarel, J. R. Martin and C. M. I. M. Matthiessen (eds) Language Typology: a Functional Perspective (Amsterdam and Philadelphia: John Benjamins), pp. 397–432. Turner, G. J. (1973) ‘Social class and children’s language of control at age five and age seven’. In B. Bernstein (ed.) Class, Codes and Control, Vol. II: Applied Studies towards a Sociology of Language (London: Routledge & Kegan Paul), pp. 121–79. Ventola, E. M. (1979) ‘The structure of casual conversation in English’. Journal of Pragmatics 3 (3–4): 267–98. Wake, B. J. (2006) ‘Dialogic learning in tutorial talk: a case study of semiotic mediation as a learning resource for second language international students’. Unpublished PhD thesis, University of Adelaide. Williams, G. (1995) ‘Joint book-reading and literacy pedagogy: a sociosemantic examination’. Doctoral dissertation, Macquarie University. Wong, Y. T. (2009) ‘The linguistic function of Cantonese discourse particles in the English medium online chat of Cantonese speakers’. MA dissertation, University of Wollongong. Wu, K. Y. (1990) ‘A linguistic study of interrogation in Cantonese: comparisons with English’. Unpublished MPhil thesis, University of Hong Kong. Wu, K. Y. (1996) ‘A functional classification of questions in Cantonese’. Journal of Chinese Linguistics 24 (2): 355–90. Part IV Context of Situation 6 Language and Society, Context and Text: the Contributions of Ruqaiya Hasan Annabelle Lukin

6.1 Introduction

The principle that language is responsive to its context of situation is, to borrow an unfortunate metaphor, deep in the ‘DNA’ of Halliday’s systemic functional linguistics (SFL). From Halliday’s early writings on grammar, the concept has been at work: ﬁrst in a relatively non-technical sense, then in a mostly Firthian way, before Halliday articulated a conception of context much more his own, albeit with a very strong echo of Malinowski’s ideas (Halliday et al., 2007). From Malinowski came not only the crucial early conceptualization of ‘context of situation’ and ‘context of culture’ but also, and just as importantly, an intuitive feel that function must be encapsulated in the form and structure of language. For someone who was not actually a linguist, this was a remarkable observation; other later linguists have resolutely failed to notice the symbiosis between form, meaning and context that was, so to speak, right under their noses. For Halliday, Malinowski’s observation was the kernel of his metafunctional hypothesis, to which his conception of context is inextricably tied. What humans do with language in our social lives, and how language is itself organized, are two sides of the same social– semiotic coin. As Halliday is fond of saying, language is the way it is because of the functions it has evolved to serve. Hasan has taken this fundamental principle from Halliday, and asked what follows from it. Although persuaded to Halliday’s view of context as a tripartite semiotic structure of ﬁeld, tenor and mode, whose dimensions are recapitulated in the internal organization of language as ‘metafunctions’, Hasan has gone on to wonder how adequate Halliday’s account is for the task of describing the cultural heights and breadths to which language has taken its context-construing power. She has framed the issue in deceptively simple terms: ‘given that speaking is done with reference to the contexts of social living, what, if anything, does this signify for the relations of language

143 144 Society in Language, Language in Society and culture?’ (Hasan, 1999a, p. 219). In pursuing an answer to this question, Hasan has shown, more than any other linguist within or outside the Hallidayan paradigm, that unravelling the web of interconnections between language and context is ‘a formidable endeavour’, which entails relating language to the ‘ceaseless ﬂux of human life and activity’ (Spencer, 1978, pp. v–vi). In the trajectory of her career, Hasan has considered what a conception of context requires if it is to describe language embedded in a ‘here-and-now’ context of material action as well as language which construes imaginary and virtual worlds. She has explored the relationship of text-in-context to culture and social structure. She has explained how culture, via context of situation, is ‘in’ text, in its structure and its texture. She has argued why context is central not only to the process of language, but ultimately, to the very emergence of language, concluding that there can be no language without context (Hasan, 2001a, p. 8). She has shown that there can be no viable linguistics without the study of parole, and that parole cannot be studied without a theory of context (Hasan, 2009a, p. 168). For Hasan, a robust theory of context is necessary simply because ‘understanding language as a resource for social action in society is what SFL is really about’ (Hasan, 2005a, p. 69). As she argues,

The work of language is always and only done in society ...hence the importance of asking where, why, and how people use language and what is indicated from this fact. (Hasan, 2005a, p. 56, emphasis in original)

Thus, Hasan has extended the Hallidayan conception of the relations of text to context, and language to society, at the same time that she has made clearer its profound significance – clarifying the place of context in linguistic theory, and its critical role in both the evolution and ontogenesis of language. For almost 40 years, she has been arguing that linguistics needs ‘a model of language in which context is a well-defined category’ and not ‘a model which distorts facts by attempting to create clear-cut boundaries where only fuzzy edges exist’ (Hasan, 1978, p. 241). A theoretical notion of context is the basis of understanding where language comes from, how it is learned, what it does in and for human societies, why it is the way it is, and how and why it changes over time. Its conceptual reach is, therefore, profound – there is no part of linguistic theory and description it does not touch. This means that variations to its conceptual delimitation, or changing its place with respect to other concepts in a theory, cannot be merely terminological differences (Martin, 2012). As Wittgenstein remarked, ‘An unsuitable type of expression is a sure means of remaining in a state of confusion’ (Wittgenstein, 1953, p. 109 cited in Hasan, 2000, p. 44). In this short chapter, I cannot hope to do more than briefly enumerate the scope of Hasan’s work on context, and suggest the descriptive and theoretical Annabelle Lukin 145 power it has brought to SFL. In this process I hope to show not only the conceptual ‘content’ of the theory of context from Hasan’s perspective, but also the precepts that drive her work: a focus on meaning in society as the brief of linguistics, a dislike of dualisms, a striving for genuinely interdisciplinary approaches to some of society’s big questions, and, perhaps above all, a belief that academic progress comes from reading the work of relevant scholars deeply and without prejudice. Though often expansive in her critiques of other scholars’ writings, her appraisals are always based on her own detailed reading of the scholar in question (see, for instance, her readings of Bourdieu, Derrida, Labov and Chomsky, in e.g. Hasan, 1988, 1989, 1999b, 2005b). She sees both strengths and weaknesses in scholars she admires (e.g. Saussure, Malinowski, Firth, Whorf, Bernstein, Martin). And she does not suspend her critical gaze when looking at the work of her colleagues or, indeed, her friends.

6.2 Is language social, or biological ...or both?

Hasan has made what might seem an extraordinary claim to some linguists: there can be no language without context. The category of context is required in linguistics, she argues, because ‘human languages as we know them cannot come about without context’ (Hasan, 2001a, p. 6). In defending this position, Hasan turns to observe the world of the human neonate, to examine context in ‘early semiosis’, or what she describes as ‘proxemic semiosis’, the semiosis prior to protolanguage where the child’s body is enlisted as the ‘expressive resource’. Drawing on studies of child language development (Bateson, Bullowa, Brazelton and Trevarthen) and from neuroscience (Edelman, Dennett, Greenﬁeld),1 she argues that the biological basis of language is not a universal grammar, but an instinct for interaction. The human infant is ‘genetically other-oriented’. Interaction needs meaning, and so the human infant is biologically predisposed ‘to cope with the basic content–expression principle’, also described as ‘the conjunction of two dissimilars’ (Hasan, 2001a, p. 5), the principle on which all semiotic systems are based (Hasan, 2001a). In these and other statements, Hasan abandons ‘the age old polarisation between the biological and the cultural, accepting instead a view put forward by Vygotsky (1978) that the opposition between the biogenetic and the cultural is false: language, rooted in human biology, grows by cultural intervention’ (Hasan, 1999a, p. 220). Cultural intervention is an unfolding process, not something delivered to infants in one fell swoop. And, as Malinowski intimated, and Halliday theorized, culture is the systemic potential behind a community’s patterns of living. It is the reservoir of potential ways of being, doing and saying, shaped by ‘innumerable small momenta’ to use Whorf’s term (Whorf, 1956, p. 151). The child’s access to these cultural patterns comes via instances of interaction with already acculturated, and socially located, members of his/her ‘speech fellowship’ 146 Society in Language, Language in Society

(Firth, 1957). Thus, context acts as a kind of porous membrane through which biology and culture mutually work in the emergence of language, both ontogenetically, and phylogenetically. With her characteristic curiosity, Hasan has asked how it is that the child breaks into the adult meaning system, a system in which, as Saussure established, signs are arbitrary,2 and meaning depends not only on reference, but on internal sign relations, or ‘valeur’. In other words, the meaning of a word is a function of its co-text, and the other words with which it has semantic affinities. And although ‘ “valeur” determines the scope of “signification”, valeur is not where the baby’s mind initially makes contact with meaning’ (Hasan, 2005a, pp. 62–3): the child, as Hasan notes, neither learns language through dictionaries, nor through reflection on the internal organization of language. The arbitrary nature of the sign makes the child dependent on experience – interaction in social contexts – to be able to learn the language(s) of its loved ones. But to do so, infants require contexts of a certain kind, described by Malinowski as those in which ‘the pragmatic relevance of words is greatest’ (Malinowski, 1935, p. 52). Calibrating insights from both Saussure and Malinowski, Hasan argues that such contexts are crucial in orientating children to the social and semiotic nature of language; they provide an environment where the correspondence between a sign and what it stands for is ‘immediately observable’ (Hasan, 1985a, p. 26). Through mediation in social contexts of this kind, the child, piece by piece, and via the orientation to what is relevant to their loved ones, enters the adult meaning system. In this process, Hasan argues, ‘signification (as the term is used by Saussure, 1966) could initially take effect; and it was here that the ineffa- ble valeur of the linguistic sign became at once palpable and significant’ (Hasan, 1995, pp. 185–6). Thus, context provides a mitigating condition which compensates for the arbitrary nature of the sign. And contexts of a certain kind provide the ‘bootstrapping’ mechanism through which language, which must be learned, can be learned.

6.3 Is context in parole only, or does it shape langue?

If there can be no language without context, then there can be no linguistics without a conception of that on which the very emergence of human language depends. Yet linguists continue either to ignore the social locus of language, or to relegate it to the function of demystifying the bric-a-brac of parole. In these latter forms of linguistic inquiry, context is ‘an a-theoretical appendage which functions as a disambiguator of ambiguous sentences, a trouble shooter for inadequate theories of language’ (Hasan, 1999a, p. 220). In these theories, context, ‘in the true Derridean fashion’, acts only as ‘a supplement for a lack of the theory itself’ (Hasan, 2001a, p. 3). When context is ‘viewed simply as the frame for some focal communicative event’, a crucial problem remains unresolved, which is ‘the problem of the theoretical Annabelle Lukin 147 status of context itself’ (Hasan, 2001a, p. 3). Modern linguistics lacks a robust account of context because it has relegated context to the margins, treating it as ‘a category whose place ...is just to illuminate the nature of parole, not of langue’, with the implication that the system ‘has a being outside the context of the social conditions of human existence’ (Hasan, 2001a, p. 3). But, as Hasan asks, ‘if context has no part in the genesis of language as resource for the exchange of meaning, how come it possesses such wide efficacy in revealing the nature of that same language when it occurs in actual uses of a decidedly varied kind?’ (Hasan, 2001a, p. 3). Logically, context cannot shape parole without also shaping langue. Context must be as central to the emergence, maintenance and change of the system, as it is in calibrating wording and meaning in one specific context of situation. Context is necessary to the study of parole because it was essential in the formation of the linguistic system. Thus, the nature of the activity (field) and the relations amongst the participants of the activity (tenor), engendered in and maintained through those cultural activities, exercise pressure on the semantic evolution of language, as do changing modes of contact between participants in the activity (Hasan, 2005a, p. 67). Hasan also shows the systemic character of language is the way it is because it is a corollary of the systemic nature of all joint social action; it is in one sense like all other semiotic systems. To explain, Hasan again turns to the neonate, arguing that proxemic semiosis is the foundation for all semiotic systems. In the interactions between caregivers and newborns, some initiating action of a baby and/or an adult precipitates

the convergence of attention and acts as the occasion for creating contact, and where there is contact, there will be relation. And, of course, action and relation are two sides of the same construct: action without relation is strictly speaking impossible, while the quality of joint action produces what we see as some relation. (Hasan, 2001a, p. 7)

All semiotic action, Hasan proposes, is based on three conditions: ACTION, RELATION and CONTACT. Halliday’s terms ‘ﬁeld’, ‘tenor’ and ‘mode’ are more speciﬁc characterizations of these general parameters, in that they articulate the linguistically construed dimensions of context.

6.4 How should ‘context’ be described?

The fundamental outline of the description of context – as field, tenor and mode – first appeared in Halliday et al. (1968) (Halliday et al., 2007).3 These parameters are briefly enumerated and exemplified in Halliday (1977) and Halliday and Hasan (1985). But Halliday has not otherwise attempted to elaborate the systems entailed in the parameters of context, nor the potential features of such, though the principle of delicacy of description with regard 148 Society in Language, Language in Society to the parameters of context was articulated in the 1964 paper. In Hasan’s work, however, there is a consistent preoccupation, from at least the early 1970s, with the problem of how context is to be characterized, first in the face of a linguistics that dismissed context as irrelevant, then in contrast with linguistic theories that treated the concept as necessary to greater and lesser extents, but largely self-evident. A crucial step was taken in Hasan (1973), when she articulated the distinction between ‘relevant context’ and the ‘material situational setting’ (MSS).4 This allowed the possibility to consider the relations of relevant context to the MSS. Hasan argues ‘the greater the reliance on the shared situational setting ...the less, and the less explicitly, will the language of the text encapsulate its context’ (Hasan, 1999a, p. 231). As she has theorized context, Hasan has theorized its corollary: text. The two concepts are so intertwined that ‘neither concept can be enunciated without the other’ (Hasan, 1985b, p. 53). Text is characterized by unity of two types: structure and texture. Both are a function of the ‘contextual configuration’, Hasan’s term for a particular convergence of values in field, tenor and mode recognized as some kind of social process by some community. ‘A’ register is the expression/construal of a contextual configuration; all values are implicated in the text’s generic structure. But what is recognized as ‘the same’ register can display structural variation, which is why Hasan proposed ‘generic structure potential’ (GSP), a statement of obligatory and optional elements, and their possible order, pertaining to the structures observed in texts instantiating a given register. Hasan’s description relates ‘culture, meaning and situation’ (see Figure 6.1).

Culture

Semiotic potential Significant situational values

Semantic potential All possible values of FTM

Genre-specific One calibration semantic potential of values of FTM

Figure 6.1 Culture, meaning and situation (Hasan, 1985b) Note:F= ﬁeld; T = tenor; M = mode. Annabelle Lukin 149

But the values in the contextual configuration do not only activate the register’s structure. Since everything in discourse is ‘beholden’ to the contextual configuration (Hasan, 1985b, 2004a), the texture of texts is also the manifestation of values in field, tenor and mode. Texture is, however, the more delicate manifestation of a contextual configuration. Hasan writes: ‘the facts of texture construe the very detailed aspects of the situation in which the text came to life’; and that situation type is ‘the motivating force of texture’ (Hasan, 1985b, p. 115). To explore the workings of texture in texts, Hasan developed the method known as Cohesive Harmony (Hasan, 1984; Hasan, 1985b). Cohesive Harmony maps texture by encompassing chain development (chains both of identity and similarity) and chain interaction, bringing together the relations on which these two vectors depend. Cohesive Harmony has been widely applied across many registers, including studies of children’s stories, mother–child dialogues, academic writing, writing by primary-school students, media discourse, psychotherapeutic interaction, literary criticism, and the language of aphasia (Khoo, this volume). Hasan has drawn on Cohesive Harmony to model relations between contexts unfolding in the same material location (see section 6.5). Hasan has also taken on the task of developing descriptions of the parameters of context. Proto-systems for tenor and mode appear in Hasan (1985b), though discussion of the content of these vectors appears as early as Hasan (1973). She has not only argued the need for explicit, paradigmatic descriptions of the contextual variables (Hasan, 1999a, 2004a), she has produced tentative networks (Hasan, 1999a, 2009a) showing how the ‘contextualisation systems’ can be modelled (see Figure 6.2; and see Bowcher, 2011 for a comparison of two networks for field by Hasan with the network for field in Butt, 2003). These models act ‘as a device for placing grids on language use, thereby introducing system into parole’ (Hasan, 2014, p. 14). The systems and details of the network may be disputed. But this is Hasan’s point: only a model of this kind offers the explicitness that allows scholars to debate and test hypotheses about the nature of field, tenor and mode systems. Hasan has sought not only to elaborate the vectors – their systems and features – but also the principles of processes of ‘contextualization’. This has included, for instance, questions pertaining to the relations between the parameters. She has argued, for instance, that the choices in one parameter do not ‘determine’ or ‘predict’ all the choices in the other two, ‘otherwise we would not have needed to recognise three separate parameters’ (Hasan, 1999a, p. 245). But the choices across the three parameters display ‘ “mutual prehension”: the echoes of a choice in one are found to some extent in the choices of the other’ (Hasan, 1999a, p. 245). She has also wondered about the question of ‘default dependency’, whereby certain selections in context by default entail particular selections in the other parameters. She notes that default dependency relations are very infrequent at the level of grammar. Their appearance in modelling context may indicate that ‘the organisation 150 Society in Language, Language in Society

Present MATERIAL ACTION Deferred Non-present Absent Install

Ancillary... Manage Maintain Instruct Create... Repair Practical Plan

Cooperative VERBAL Relation based ACTION Conflictual Reflection based...

First order

Reproductive Constitutive Second order Metadiscursive Commenting... Informing Reason Conceptual Describing Field Class

Specialized Event SPHERE OF ACTION Institutional Congruent Quotidian Tale Individuated Metaphorical Simple Fable Inventing Complex Story Stop Novella ITERATION Independent Narrating Go Self Aligned Other Integrated Personal Episode Sequence Recounting Immediate Distant Communal Narrow focus Wide focus

Figure 6.2 A ‘rudimentary’ network for ﬁeld (Hasan, 1999a) of the level of context may differ from that of semantics and/or lexicogrammar in subtle ways’, a consequence, she suggests, of the fact that while context may be construed in language, ‘it can never be dissociated from the material and institutional aspects of a culture’ (Hasan, 1999a, p. 245). While arguing for the simultaneity of all parameters of context, Hasan gives tenor primacy. She argues, for instance, that ‘most variation in language is directly related to the contextual parameter of relation’ (Hasan, 2001a, p. 13). It is, after all, a speaker’s ‘ideological stance that furnishes the principles through which a social context is seen for what it is for that member, both from the point of view of the recognition of the occasion and from the point of view of participation in it – the modes of being, doing and saying that are ideologically legitimate’ (Hasan, 2001a, p. 15). Annabelle Lukin 151

6.5 What if there is more than one context unfolding at the same time?

From her studies of mother–child interactions across two distinct social locations, Hasan has produced a number of significant insights for the understanding of text–context relations, including the characterization of the conditions under which one can argue that across the same speakers in the same location, more than one context is unfolding. Hasan observes that under certain conditions, the principle of ‘contextual/registerial consistency’, which underpinned Halliday and Hasan’s definition of text as ‘a passage of discourse which is coherent ...with respect to the context of situation, and therefore consistent in register’ (Halliday and Hasan, 1976, p. 23) does not hold. In the right conditions, speakers can ‘reclassify [ ...] the discursive situation’ (Hasan, 2001b, p. 63). Hasan has, to a certain degree of delicacy, specified the conditions for discursive reclassification. ‘Contextual consistency’ is a feature of contexts in which the channel is [graphic], and the addressee is [virtual]; or if the channel is [phonic] and the medium [spoken], then the nature of the social activity would have the feature [institutionalized]; that is, it would be a spoken, face-to-face interaction which is to some degree ritualized, and where social distance is [maximal] (Hasan, 1999a). Contexts with these features are inhospitable to discursive reclassification (see below). But the dialogic mode where social distance is [minimal], and speakers are held constant – a contextual configuration in which people spend much of their time – affords the opportunity for the construal of ‘complex texts’ which display the principle of ‘contextual integration’. To make sense of these conditions, Hasan proposes the distinction between a DOM- INANT CONTEXT, construed in a PRIMARY TEXT, and either DEPENDENT or COLLABORATIVE CONTEXTS, which are realized by a SUBTEXT.5 Dependent and collaborative contexts bear a functional relation to the dominant context: a dependent context acts to facilitate a dominant context, while a collaborative text supports the dominant context through its ‘tone-setting’ function; it contributes to the primary text by ‘managing the affective tone of whatever is going on’ (Hasan, 1999a, p. 255). Evidence that these texts play a role in relation to the dominant context is found in cohesive relations. Thus, these kinds of subtexts are integrated into the primary text because they are ‘cohesively multiply related to the primary text’ (Hasan, 1999a, p. 256). In the absence of such relations, a co-located text is INDEPENDENT of the primary text. An independent text is ‘materially co-located’ while a primary text and its subtexts are ‘textually integrated’ (Hasan, 1999a, p. 262). Table 6.1 summarizes the attributes of complex versus parallel simple texts, that is, texts materially co-located but which are entirely unrelated. Hasan’s discussion of con/textual integration is neither another term for ‘genre combination’ nor for ‘hybridity’, terms which both imply ‘an unfortunate 152 Society in Language, Language in Society

Table 6.1 Material and textual relations within one interaction: complex vs parallel simple texts (Hasan, 1999a, p. 267)

Relations within complex Relations between parallel texts simple texts

Material: Same location; same interactant Same location; same interactant (= co-location) Principle: Integration of different registers Consistency of register within each text Structural: Subtexts and primary text Each text a separate structural integrated identity Contextual: Dependency or collaboration Each context independent of the other Functional: Facilitation or tone setting One text has no function in the other(s) Texture: Substantial textural unity Textural unity absent or present; chain sharing, chain marginal; chain sharing interaction minimal: no interaction

reiﬁcation of the process of register, as if what is happening is simply a co-location or a fusion of two (or more) recognisable already existing objects’ (Hasan, 2000, p. 44).

6.6 How does context of situation relate to context of culture?

Figure 6.3 recapitulates the standard presentation of the relations of context of situation to context of culture in Hallidayan theory. The relation is one of instantiation. Hasan interprets this relation to mean that ‘every mention of situation implicates culture ...behind every situation lies the context of culture’ (Hasan, 1999a, p. 315); and that ‘situation must be “aware” of culture’ (Hasan, 2014, p. 14). What does this mean in practice? Hasan’s studies of children and mothers interacting in everyday situations is perhaps the most extended, empirical examination in SFL of how culture permeates the instance. It is, at the same time, an example of what truly interdisciplinary research looks like. The full significance of this work requires a reading of at least the first two volumes of Hasan’s Collected Works. Summarized briefly here, the most ordinary of daily interactions between mothers and children reveal the social structure at work, in the way that Bernstein predicted, but could not show:

The experience of the child is transformed by the learning generated by his [sic] own, apparently, voluntary acts of speech. The social structure becomes ...the substratum of the child’s experience through the manifold consequences of the linguistic process ...every time the child speaks or Annabelle Lukin 153

listens, the social structure is reinforced in him and his social identity shaped. The social structure becomes the child’s psychological reality through the shaping of his acts of speech. (Bernstein, 1971, p. 144)

In the minutiae of everyday life, a child’s attention is recruited towards certain ‘orders of relevance’. Hasan argues that ‘the primary function of semiotic mediation in the life of young humans is to enable the inter- nalisation of cultural designs essential to everyday living’ (Hasan, 2009b, p. 436). The acts of meaning through which the internalization of cultural designs takes place vary. The variation is semantic, and pertains to differences with regard to (no less than) one’s sense of self and one’s relations to a wider world. One key social distinction which Hasan has shown to correlate with degree of autonomy at work for the family breadwinner is whether everyday contexts are treated as strongly or weakly ‘classified’ (Bernstein, 1990). Children of higher-autonomy professionals grow up in families where the boundaries of more contexts of situation are more permeable (that is, weakly classified) than for children of lower-autonomy professionals. What from a distance appears to be the same social situation – eating, bathing, playing, bedtime – turns out to vary in the most consequential of fashions (Williams, 1999). Bernstein calls this form of variation ‘coding orientation’ (Bernstein, 1990). Hasan has not only shown what coding orientation looks like in exhaustive detail – that is, how meanings vary by social location of the mother–child dyad – but she also shows this variation to be statistically significant. While the superficial features which overtly distinguish class, such as ‘variation in accent and certain kinds of ritualistic grammatical patterns’, have been of interest to some in sociolinguistics, these forms,

Instantiation SYSTEM INSTANCE

SOCIETY Society Context of situation R (cultural (situation e a domain) type) l i z a t i o n (register) (text type) Language Parole LANGUAGE as system as text

Figure 6.3 Realization and Instantiation in Halliday’s linguistic model (Hasan, 2009c, p. 12, based on Halliday, 1999) 154 Society in Language, Language in Society

Hasan shows, are irrelevant to coding orientation. Code, Hasan has written, is ‘purely variation in perception and production of meaning’ (Hasan, 2011, p. xxxvii). This variation, Hasan argues, requires explanation from the ‘context of culture’, that is, from how society is organized. For instance, I described in section 6.5 Hasan’s account of the kind of contexts which afford the opportunity for discursive reclassiﬁcation. But this opportunity is not equally taken up by all speakers. In some families, the talk associated with making and having a meal is seamlessly interwoven with dissertations drawn from ‘the knowledge base of a community’ (Hasan, 2001b, p. 70). This form of contextual variation correlated with the distinction between families in which the main breadwinner was a high- autonomy professional, and those in which the main breadwinner had very little personal discretion over his or her working schedule. Discourse in society, Hasan observes, is not homogeneous. The crucial distinctions ‘arise from subjects’ ...social positioning’ where their ‘orders of relevance are shaped, their sense of what is legitimate formed and psychic defences built and turned into second nature’ (Hasan, 2004b, p. 37). This ‘does not mean they are engraved in stone’ but that changing ‘at this level is a major enterprise’ (Hasan, 2004b, p. 37). Figure 6.4 is Hasan’s presentation of her model of the semiotic exchange from the perspective of Bernstein (Hasan, 2005c). It locates text, and other meaningful exchanges, as the tip of a breathtakingly expansive iceberg. The diagram also illustrates why Hasan gives primacy to tenor.

6.7 What does ‘decontextualized’ mean?

The notion that a text can be ‘decontextualized’ is taken for granted by many scholars, in the sense of treating the notion of ‘decontextualization’ as so self-evident it does not require definition or explanation. Hasan treats nothing as self-evident. Theory, Hasan notes, ‘models the object of enquiry’ (Hasan, 2009c, p. 35). Concepts are involved in the mutual delimitation of each other: she has argued, for instance, that ‘text’ cannot be defined without ‘context’, because ‘each is instrumental in the precise definition and development of the other’ (Hasan, 1999a, p. 223). Hasan has intimated the potential survival advantage of decontextualization in the evolution of Homo sapiens: ‘in the battle for survival, somewhere in their history the species homo sapiens chose to use social semiosis rather than bodily change for effecting adjustment to their environment’ (Hasan, 2001b, p. 48). But her focus has been on the ontogenesis of decontextualized language, and its survival advantages in the jungles of schooling. Hasan distinguishes ‘actual’ and ‘virtual’ contexts, a distinction which is not the same as that between language that is ‘context embedded’ and language that is ‘decontextualized’. An ‘actual’ context can be ‘immediate’ or ‘displaced’ – in both cases the relationship of the context to the speaker is ‘sens-ible’ [sic], that is, ‘rooted in 155

Division of labour

Distribution of Distribution of material capital esoteric knowledge <1>

Access to material Access to symbolic resources resources

Class relations

Distribution of power Principles of control <2> Classification Framing

Coding orientations

Sense of interactive contexts Recognition criteria Realization principles <3>

Variant Legitimate interpretations participation within of context context

Sense of relevant meanings

<4> Habitual ways of being doing and saying in specific interactive events

Activity Relations Contact modes

Semiotic exchange

Figure 6.4 Class, codes and communication: an interpretation (Hasan, 2005c) 156 Society in Language, Language in Society experience that is essentially sensuous’. But Hasan draws a crucial distinction between these two types of ‘decontextualized’ contexts. A displaced context ‘re-creates an actual context that actually existed and that is now spatio-temporally displaced ...[it] is as actual as an immediate context in which I use my language as an ancillary tool for negotiating the performance of some ongoing physical action such as helping someone wash-up’ (Hasan, 2001b, pp. 54–7). ‘Virtual’ contexts are ‘non-material and removed from situational realities, they ...cannot be directly and physically experienced: they are intellig-ible, not sens-ible. They reside only in a conceptual universe, and consist simply of the said and the imagined’; it is ‘a terrain navigated by the intellect alone’ (Hasan, 2001b, pp. 54–7). Both ‘displaced’ and ‘virtual’ contexts can be described as ‘decontextualized’, if the term means at a distance from a material situation setting. But, as Hasan argues, there are qualitative differences between these kinds of decontextualization; here Hasan refers to Cloran’s ‘cline of de/contextualisation’ (Cloran, 1994). She describes Cloran’s discussion of the concepts ‘context dependent’ and ‘context independent’ as the most careful and detailed in SFL (Hasan, 2001b). Hasan has also shown the acquisition of a facility with decontextualized discourse is not uniform. Instead, ‘ ...the establishment of disembedded talk is highly selective’ and arises where the coding orientation of the familial environment is characterized by relatively weak classiﬁcation and framing values (Hasan, 2001b, p. 62). Put in less abstract terms, the families who socialize their children to see disembedded talk as natural treat everyday contexts, like playing, eating and bathing, as permeable, and, therefore, as sites where talk of different kinds is welcome (see section 6.6). It is in this potential for context permeability that children of middle-class families can experience the kind of ‘invisible semiotic mediation’ which is ‘likely to be far more signiﬁcant in the life of these children than the actual elements of knowledge acquired through explicit semiotic mediation in the early years’ (Hasan, 2009b, p. 450). For children of families of this social location,

the experience of receiving sustained explicit information in emotionally supportive environments develops in these children an orientation toward decontextualized knowledge. It becomes an aspect of their mental disposition, colouring their mental activities. (Hasan, 2009b, p. 450)

Why would everyday activities of bathing and feeding be the ideal circumstances for the ontogenesis of decontextualized language practice? Hasan notes first of all how unnatural decontextualized forms of discourse are. But when they form a seamless part of caregiving to a child, in which the child’s physical needs are being attended to, when the interpersonal relation is one of cooperation and not conflict, under these conditions ‘bodily satisfaction Annabelle Lukin 157 is thus combined with the satisfaction of companionship’ (Hasan, 2001b). Thus, the child is physically attended to, and at the same time ‘they are being attended to intellectually and semiotically’ (Hasan, 2001b). The activity of the intellect becomes associated with the satisfaction of bodily needs (Hasan, 2001b). There is perhaps no more profound a way to naturalize a child into the forms of decontextualized language so prized in schooling. In fact Hasan argues that by contrast to the structure of schooling, this form of ‘home- schooling’ is far better organized for the ontogenesis of decontextualized discourse than schools can be. Based on a small sample of data, from the children Hasan studied both as preschoolers then after they began at school, Hasan argues that the classification and framing values of the contexts of schooling in fact mitigate against developing in children an orientation to decontextualized language:

With their stronger framing of the discourse, with their great emphasis on disembedded meanings, their distancing from the personal concerns of the children themselves, it seems very unlikely that schools provide the best environment for learning how to use such language for children who do not already possess this expertise to some extent before they enter the school. (Hasan, 2001b, p. 74)

6.8 Context in the context of other semiotic systems

Hasan has, on numerous public occasions, stated that as a linguist, her concerns are with the workings of language in social context. But she has not ignored the very obvious fact that language operates in the environment of other semiotic codes, and often directly in consort with them (see Bowcher and Liang, Wegener, this volume). In the search for a scientific conception of context, Hasan has noted that all semiotic systems must share some general characteristics as a condition for their being able to meet the human biological foundations for semiosis, described as both a predisposition for sociality, and an inherited capacity for semiosis, more specifically as the capacity to conjoin ‘two dissimilars’ (Hasan, 2001a, p. 5). As noted in section 6.2, this biological foundation requires interaction for its expression and development. It requires, in other words, context, which at some level of abstraction must be the same for all semiotic systems. Hasan has proposed, then, that the components of a context – irrespective of the modality recruited to meaning making – are ACTION, RELATION and CONTACT (discussed briefly in section 6.3). These vectors constitute ‘a condition for the occurrence of any joint social activity irrespective of the semiotic system employed for creating and maintaining contact’ (Hasan, 2001a, p. 7). Hasan has noted that ‘the human body is an amazing system for engaging in multimodal semiosis’ (Hasan, 1999a, p. 241), though it imposes the requirement of co-presence of interactants for this potential to be actualized. As a neonate, co-presence is a 158 Society in Language, Language in Society

Sign types

Natural Artefactual

Somatic Exo-somatic

Linguistic Kinetic Iconic Symbolic Indexical

Universal Specific Communal Personal

Figure 6.5 Hasan’s typology of sign types (Hasan, 2014, Figure 6.1, p.112) constant condition of experience. In the absence of any linguistic capacities, the child engages the ‘proxemic modality’ as its expressive resource. This early proxemic engagement is necessary for the ‘epigenesis of all human semiotic systems’ (Hasan, 2001a, p. 5). Hasan has, naturally, also noted the ‘distinct representational capacities’ of different semiotic codes, and the ‘partiality of a code to certain kinds of relevance’ (Hasan, 1996a, p. 38). Hasan (2014) presents perhaps the most comprehensive description of sign types in semiotics (see Figure 6.5 and Table 6.2). By proposing this typology of signs, and elaborating the key characteristics of sign types, Hasan has some basis for her claim that linguistic signs differ from all other sign types with respect to ‘the intensity of their mutual relationships’ (Hasan, 2014, p. 9) – ‘the behaviour of the “wholly arbitrary” linguistic sign is intensely systemic; without its systemic relations the linguistic sign would not be a sign of anything’ (Hasan, 2014, p. 11). Echoing Saussure – ‘Signs that are wholly arbitrary realize better than the others the ideal of the semiological process’ (Saussure, 1966, p. 68) – Hasan argues that linguistic signs have a degree of systemicity not reproduced in any other sign system, the members of which do not appear to impinge on the life of the other signs of the same code. Comparing the sign types, she argues that ‘none of the non-linguistic sign types have so far got to what Saussure called “the first principles” in linguistic research: their description is presented largely as an inventory of “isolated individuals” ’ (Hasan, 2014, p. 11). Moreover, she wonders whether non-linguistic sign types are ‘capable of being scientifically studied, if the term means more than simply describing the empirically observed phenomena objectively’ (Hasan, 2014, p. 11, emphasis in the original). 159 visitor → rain order to leave → → → play time justice happy ‘cut here’ → → → Front door bell School bell [general category; various] Dark clouds [general category; various] [see next two entries] [general category; various] No No InherentNo ‘Go away!’ No Scales – No – + Similarity No Arbitrary Arbitrary Arbitrary Arbitrary – Physical logic No – – Yes Yes Yes Yes Yes Human Matter S-maker S-medium User-volition Sd/sr-relation Systemic Example Properties of the sign types in Figure 6.2 (Hasan, 2014) Iconic Personal Human Matter IndexicalCommunal Human Human Matter Matter SpecificLinguisticExo-somatic Human Human Human User User body Matter body Uncertain Yes Uncertain Not inherent [culturally variable gesture] SomaticKineticUniversal Human Human Human User body User User body body – Uncertain Uncertain Uncertain Uncertain Not Not inherent inherent [general Smile category; various] Table 6.2 S-type Natural Nature Natural force None Artefactual Human Material – Symbolic Human Matter Yes Similarity 160 Society in Language, Language in Society

6.9 What is ‘context’ in the context of literature?

The ‘problem’ of context in Hasan’s thinking arose in her early research on the language of literature (see Butt, Miller, this volume). This was a question that ‘haunted’ Hasan throughout her doctoral research: ‘if literature is a variety of register, what is its context like?’ (Hasan, 2011, p. xxiv). Her answer has been to propose a ‘multiple contextual framework’ (Hasan, 2011, p. xxvi) for genres of verbal art. The literary text has inner and outer contexts. The outer contexts are those of production, which ‘locate [ ] the artist in his/her cultural ways of being, doing and saying’, and the context of reception, aligned with ‘the reader’s cultural modes of being, doing and saying’ (Hasan, 2011, p. xxvi). But there is also the text’s ‘relevant context’, described by Hasan as the contexts construed by the artist’s use of language. It is these contexts which ultimately give shape to ‘the deepest level of meanings’ in that work, which are ‘voiced by the text’s semiotic organisation playing a central role in cumulatively realizing the theme’s artistic articulation’ (Hasan, 2011, p. xxvi). With respect to the inner context of the text, there are a number of dimensions to be explored; tenor, for instance, is a vector relevant to the internal relations of characters in a text. It is also relevant to the description of reader–narrator relations, which can be one of faux intimacy, in which the ‘dear reader’ is directly addressed, or one of maximal social distance. With regards to field, there are also two levels: the level of that which can be subjected to paraphrase, and that of the social action of the text, which is the telling or showing of some deeper meaning pertaining to human experience. Hasan argues that ‘the art of the artist lies precisely in the skill with which s/he is able to calibrate the two levels of articulation of field, so that the events, episodes, claims of the first level succeed in articulating the deepest meaning – the theme of the text’ (Hasan, 1996b, p. 51). Mode, similarly, has two orders: on one level, literature is monologic. Internally, it involves the recreation of often many ‘as-if registers’ (Hasan, 1996b, p. 53).

6.10 When do individuals have the freedom to (re-)negotiate the context?

Hasan has described the quixotic character of the linguistic system, which both acts, through inertia, as a force for conformity, at the same time that individuals have certain degrees of freedom. This turns out not to be so para- doxical. The individual acts semiotically through shared linguistic resources. The conventionality of the sign prohibits individual freedom in choices of signiﬁeds. Children play with this arbitrariness. My daughter and her friend have an invented language (Quibinish, the language of creatures known as Quibins), a small set of signs which will shortly wither and die. The child must submit to the sign conventions of her community; his/her life depends on it. Thus, the creation of the individual is a social and cultural process. Annabelle Lukin 161

As Vološinov argued, individual consciousness is ‘not the architect of the ideological superstructure, but only a tenant lodging in the social ediﬁce of ideological signs’ (Vološinov, 1973, p. 13). Do individuals, then, have no freedom to negotiate the parameters of a context? Is all the power in the system, and none in the instance? In a linguistic theory that privileges neither the system nor the instance, this question becomes empirical. Hasan has argued:

A hypothesis about the relationship between text and context must provide a systematic account of both variant norms and individual autonomy not because failure to do so is reactionary, but because these concepts are essential to our understanding of how contexts, behaviours and cultures are related to each other and how any change ever occurs in any of these. (Hasan, 1996a, pp. 44–5)

The freedom to negotiate contexts is not equally available to all at all times. In what environments do individuals have freedom to negotiate the nature of the context? One essential condition is that the mode is dialogic. Sec- ondly, Hasan notes that if there is to be (re-)negotiation of a context, then that implies that one party to the interaction ‘could not have predicted the nature of the context with a significant probability of being correct without the benefit of access to what the other party is saying’ (Hasan, 1996a, p. 45). Thus the setting of the interaction must be largely immaterial to its nature, it must be ‘interpretively neutral’ (Hasan, 1996a, p. 46), so that the context is built out of the unfolding discourse, rather than preceding it. By contrast, individual autonomy is at its lowest when context precedes the text, that is, where the unfolding text conforms to conventional agreement about the nature of the social process entailed in the business to hand. This vector of contextual variation is partly described by Hasan’s cline at one end of which is the feature ‘institutionalized’ and at the other ‘individuated’ (Hasan, 1981).6 Contexts which are highly institutionalized are not hospitable to individual renegotiation. Where the MSS ‘displays a series of convergent codes, then the probability of the individual negotiation of context is appreciably reduced’ (Hasan, 1996a, p. 47). When first proposed by Hasan, this vector was considered to be a feature of both tenor and field. In Hasan (1999a), this distinction, from nearly 20 years prior in her work, became a system of field which she called ‘sphere of action’. On a related trajectory to this discussion on the exercise of an individual’s individuality is Hasan’s inquiry into the problem of ‘goal orientation’. The category has been fundamental to the emergence of the ‘Sydney School’ of genre studies, since genre has been, over some years, defined as ‘staged, goal- oriented social process realised in register’ (Martin, 1992, p. 505). Arguing for a reconceptualization of the category ‘goal’, and noting a long philosophical discourse on the problem of ‘goal’, Hasan (1999a) makes a number of points: 162 Society in Language, Language in Society the degree of goal awareness is variable; to have a goal is not necessarily to achieve it; in some activities there may be an array of goals; that knowing fully the outcome of a social process that we engage in may be less likely than knowing why we are so engaged; that engagement in a social process cannot be captured in a singular notion like ‘goal’ because it is ‘a continuous struggle on the part of the speaker(s) to calibrate their perceived goals with the perceived design of the social process so that the outcome matches the goal’ (Hasan, 1999a, pp. 234–7).

6.11 Conclusion

Standing back and observing Hasan’s theoretical cut and thrust, one can come to understand what a theory of context is like. The dimensions of her capacity for scholarship are perhaps illustrated by her claim that the ‘real context for the conceptualisation of the category of context in linguistics’ is ‘a trinity no one member of which can exist with the other two’. The trinity, echoed in the title of the first volume of her Collected Works, is ‘society, semiosis and the brain’ (Hasan, 2001a, p. 17). Reading Hasan on context provides a wealth of conceptual tools for the task of analysing the ‘ceaseless flux of human life and activity’, but it is also a masterclass in what the activity of theorizing in linguistics entails. This is what is also to be learned – the invisible curriculum – in studying Hasan’s model of context. While she privileges tenor in the contextual configuration, and her lived experience is relevant to her development as a scholar, Hasan is assiduously objective in her argumentation. I make this point because at times her academic argumentation has been accused of being personalizing (see e.g. Robbins’ (1999) response to Hasan’s article ‘The disempowerment game: Bourdieu on language’). Hav- ing known Hasan personally, what makes this kind of claim amusing is that her scholarship is characterized by what she once described to me as a personality defect: her inability to do anything other than say what she really thinks on a topic, regardless of its consequences for tenor. Hasan’s ideas are not simply and easily recontextualized. Her 1999 paper (‘Speaking with reference to context’) I have been reading now for 15 years. Getting inside her views takes time, and there is much about modern academia that neither values nor enables deep and careful reading. But the notion of context is central to Halliday’s SFL, and Hasan has done no less than shown the way to making the concept scientific. This is only one of her many contributions to a linguistics for the twenty-first century.

Notes

1. Space prevents me citing the research outside of the ﬁeld of linguistics that Hasan draws on for making claims about language seen from other disciplines. 2. Halliday has noted that the line of arbitrariness holds only between the strata of lexicogrammar and expression; the relations of other stata in the linguistic system are ‘natural’. Annabelle Lukin 163

3. The paper has three authors, but was largely written by Halliday. See endnote to paper published in Vol. 10 of Halliday’s Collected Works. 4. Bowcher argues that the precursor for the notion of MSS is found in Halliday (1957) (Bowcher, 2011). 5. In Hasan and Butt (2011), the term ‘sub-text’ is attributed to Cross (1991). 6. Some notion of ‘cline of individuation’ has now appeared in the work of Hasan’s colleagues, such as Martin and Matthiessen (see entry in Matthiessen et al., 2010). Hasan’s conception of this cline pre-dates all the work mentioned in Matthiessen et al. (2010); and she gives this cline a particular location in the architecture of SFL.

References

Bernstein, B. (1971) Class, Codes and Control,Vol.I:Theoretical Studies towards a Sociology of Language (London and Boston: Routledge & Kegan Paul). Bernstein, B. (1990) Class, Codes and Control,Vol.IV:The Structuring of Pedagogic Discourse (London: Routledge). Bowcher, W. (2011) ‘The history and theoretical development of context of situation in Systemic Functional Linguistics’. In C. Chang and Y. Chen (eds) Context in Sys- temic Functional Linguistics (Beijing: Foreign Language Teaching and Research Press), pp. 64–93. Butt, D. (2003) ‘Parameters of context’. Unpublished mimeo, Macquarie University. Cloran, C. (1994) Rhetorical Units and Decontextualisation: an Enquiry into Some Rela- tions of Context, Meaning and Grammar (Nottingham: Department of English Studies, University of Nottingham). Cross, M. (1991) ‘Choice in text: a systemic approach to computer modelling of variant text production’. Unpublished PhD thesis, Macquarie University. Firth, J. R. (1957) Papers in Linguistics: 1934–1951 (Oxford: Oxford University Press). Halliday, M. A. K. (1957) ‘Some aspects of systematic description and comparison in grammatical analysis’. Studies in Linguistic Analysis (Special Volume of the Philological Society) (Oxford: Basil Blackwell), pp. 54–67. Halliday, M. A. K. (1977) ‘Text as semantic choice in social contexts’. In T. A. van Dijk and J. Petoﬁ (eds) Grammars and Descriptions: Studies in Text Theory and Text Analysis (Berlin: Walter de Gruyter), pp. 176–225. Halliday, M. A. K. (1999) ‘The notion of “context” in language education’. In M. Ghadessy (ed.) Text and Context in Functional Linguistics (Amsterdam: Benjamins), pp. 1–24. Halliday, M. A. K. and Hasan, R. (1976) Cohesion in English (London: Longman). Halliday, M. A. K. and Hasan, R. (1985) Language, Context and Text: Aspects of Language in a Social-Semiotic Perspective (Geelong: Deakin University Press). Halliday, M. A. K., McIntosh, A. and Strevens, P. (2007) ‘The users and uses of language’. In J. J. Webster (ed.) The Collected Works of M. A. K. Halliday,Vol. 10: Language and Society (London and New York: Continuum), pp. 5–37. Hasan, R. (1973) ‘Code, register and social dialect’. In B. Bernstein (ed.) Class, Codes and Control, Vol. II: Applied Studies Towards a Sociology of Language (London: Routledge & Kegan Paul), pp. 253–92. Hasan, R. (1978) ‘Text in the systemic-functional model’. In W. Dressler (ed.) Current Trends in Text Linguistics (Berlin: de Gruyter), pp. 228–46. Hasan, R. (1981) ‘What’s going on?: a dynamic view of context in language’. In J. E. Copeland and P. W. Davis (eds) The Seventh LACUS Forum1980 (Columbia: Hornbeam Press), pp. 106–21. 164 Society in Language, Language in Society

Hasan, R. (1984) ‘Coherence and cohesive harmony’. In J. Flood (ed.) Understanding Reading Comprehension (Newark, Del.: IRA), pp. 181–221. Hasan, R. (1985a) ‘Meaning, context and text: ﬁfty years after Malinowski’. In J. Benson and W. Greaves (eds) Systemic Perspectives on Discourse (Vol. 1) (Norwood, NJ: Ablex), pp. 16–47. Hasan, R. (1985b) Part B In M. A. K. Halliday and R. Hasan Language, Context and Text: Aspects of Language in a Social Semiotic Perspective (Geelong, Vic.: Deakin University Press), pp. 52–121. Hasan, R. (1988) ‘Directions from structuralism’. In N. Fabb, D. Attridge, A. Durant and C. MacCabe (eds) The Linguistics of Writing: Arguments between Language and Literature (Manchester: Manchester University Press), pp. 103–22. Hasan, R. (1989) ‘Semantic variation and sociolinguistics’. Australian Journal of Linguistics 9: 221–75. Hasan, R. (1995) ‘The conception of context in text’. In P. H. Fries and M. Gregory (eds) Discourse in Society: Systemic Functional Perspectives. Meaning and Choice in Language: Studies for Michael Halliday (Norwood, NJ: Ablex), pp. 183–283. Hasan, R. (1996a) ‘What’s going on: a dynamic view of context in language’. In C. Cloran, D. Butt and G. Williams (eds) Ways of Saying: Ways of Meaning (London: Cassell), pp. 37–50. Hasan, R. (1996b) ‘On teaching literature across cultural differences’. In J. James (ed.) The Language–Culture Connection (Singapore: SEAMEO), pp. 34–63. Hasan, R. (1999a) ‘Speaking with reference to context’. In M. Ghadessy (ed.) Text and Context in Functional Linguistics: Systemic Perspectives (Amsterdam/Philadelphia: John Benjamins), pp. 219–328. Hasan, R. (1999b) ‘The disempowerment game: Bourdieu and language in literacy’. Linguistics in Education 10 (1): 25–7. Hasan, R. (2000) ‘The uses of talk’. In S. Sarangi and M. Coulthard (eds) Discourse and Social Life (London: Longman), pp. 28–47. Hasan, R. (2001a) ‘Wherefore context?: the place of context in the system and process of language’. In S. Ren, W. Guthrie and I. W. R. Fong (eds) Grammar and Discourse: Proceedings of the International Conference on Discourse Analysis (Macau: University of Macau), pp. 1–21. Hasan, R. (2001b) ‘The ontogenesis of decontextualised language: some achievements of classiﬁcation and framing’. In A. Morais, I. Neves, B. Davies and H. Baillie (eds) Towards a Sociology of Pedagogy: the Contribution of Basil Bernstein to Research (New York and Oxford: Peter Lang), pp. 47–79. Hasan, R. (2004a) ‘Analysing discursive variation’. In L. Young and C. Harrison (eds) Systemic Functional Linguistics and Critical Discourse Analysis: Studies in Social Change (London and New York: Continuum), pp. 15–52. Hasan, R. (2004b) ‘The concept of semiotic mediation: perspectives from Bernstein’s sociology’. In J. Muller and A. Morais (eds) Researching with Bernstein, Working with Berstein (London: Routledge), pp. 30–43. Hasan, R. (2005a) ‘Language and society in a systemic functional perspective’. In R. Hasan, C. M. I. M. Matthiessen and J. J. Webster (eds) Continuing Discourse on Language: a Functional Perspective (Vol. 1) (London: Equinox), pp. 55–80. Hasan, R. (2005b) ‘Semiotic mediation and three exotropic theories: Vygotsky, Halliday and Bernstein’. In J. J. Webster (ed.) The Collected Works of Ruqaiya Hasan, Vol. 1: Language, Society and Consciousness (London: Equinox), pp. 130–56. Hasan, R. (2005c) ‘Society, language and mind: the metadialogism of Basil Bernstein’s theory’. In J. J. Webster (ed.) The Collected Works of Ruqaiya Hasan,Vol.1:Language, Society and Consciousness (London: Equinox), pp. 51–72. Annabelle Lukin 165

Hasan, R. (2009a) ‘The place of context in a systemic functional model’. In M. A. K. Halliday and J. J. Webster (eds) Continuum Companion to Systemic Functional Linguis- tics (London and New York: Continuum), pp. 166–89. Hasan, R. (2009b) ‘The world in words: semiotic mediation, tenor and ideology’. In J. J. Webster (ed.) The Collected Works of Ruqaiya Hasan,Vol.2:Semantic Variation: Meaning in Society and Sociolinguistics (London: Equinox), pp. 433–54. Hasan, R. (2009c) ‘Wanted: an integrated theory for sociolinguistics’. In J. J. Webster (ed.) The Collected Works of Ruqaiya Hasan, Vol. 2: Semantic Variation: Meaning in Society and Sociolinguistics (London: Equinox), pp. 5–40. Hasan, R. (2011) ‘A timeless journey: on the past and future of present knowledge’. In Selected Works of Ruqaiya Hasan on Applied Linguistics (Beijing: Foreign Language Teaching and Research Press), pp. xiv–xliii. Hasan, R. (2014) ‘Linguistic sign and the science of linguistics: the foundations of appliability’. In Y. Fang and J. J. Webster (eds) Developing Systemic Functional Linguistics: Theory and Application (London: Equinox), pp. 1–30. Hasan, R. and Butt, D. G. (2011) ‘Forms of discourse, forms of knowledge: reading Bernstein’. In J. J. Webster (ed.) The Collected Works of Ruqaiya Hasan,Vol.3: Language and Education (London: Equinox), pp. 99–165. Malinowski, B. (1935) Coral Gardens and Their Magic (2) (London: Allen and Unwin). Martin, J. R. (1992) English Text: Systems and Structure (Amsterdam: Benjamins). Martin, J. R. (2012) The Collected Works of J. R. Martin,Vol.3:Genre Studies. Edited by Z. H. Wang (Shanghai: Shanghai Jiaotong University Press). Matthiessen, C. M. I. M., Teruya, K. and Lam, M. (2010) Key Terms in Systemic Functional Linguistics (London and New York: Continuum). Robbins, D. (1999) ‘Bourdieu on language and linguistics: a response to R. Hasan’s “The disempowerment game: Bourdieu on language in literacy” ’. Linguistics and Education 10 (4): 425–40. Saussure, F. de (1966) Course in General Linguistics (New York: McGraw-Hill). Spencer, J. (1978) ‘General editor’s preface’. In M. Gregory and S. Carroll (eds) Lan- guage and Situation: Language Varieties and Their Social Contexts (London, Henley and Boston: Routledge & Kegan Paul), pp. v–vii. Vološinov, V. N. (1973) Marxism and the Philosophy of Language (New York: Seminar Press Inc.). Vygotsky, L. (1978) Mind and Society (Cambridge, Mass.: Harvard University Press). Whorf, B. (1956) Language, Thought, and Reality: Selected Writings (Cambridge, Mass.: MIT Press). Williams, G. (1999) ‘The pedagogic device and the production of pedagogic discourse: a case study in early literacy education’. In F. Christie (ed.) Pedagogy and the Shaping of Consciousness: Linguistic and Social Processes (London: Cassell), pp. 88–122. Wittgenstein, L. (1953) Philosophical Investigations. Translated by G. E. M. Anscombe (New York: Macmillan). 7 Multiscalar Modelling of Context: Some Questions Raised by the Category of Mode Tom Bartlett

7.1 Introduction1

The idea that a concept of context is essential for the accurate interpretation of linguistic acts is commonplace in both lay language (just enter taken out of context in your search engine) and in sociolinguistic research (see Duranti and Goodwin, 1992). However, what exactly counts as context and how it affects the meaning of an utterance is not quite so clear. In his novel Death at Intervals, José Saramago (2008, p. 58) suggests in his usual playful way that we all ‘know’ in an everyday sense what context means; yet, as he also points out, the term exists in close proximity to other everyday terms, such as background, while the precise scope of context itself is hard to deﬁne, given the depth and diversity of information it potentially comprises:

Lovers of concision, laconicism and economy of language will doubt- less be asking, if the idea is such a simple one, why did we need all this wafﬂe simply to arrive, at last, at the critical point. The answer is equally simple, and we will give it using a current and very trendy term ...and that term is context. Now everyone knows what we mean by context, but there could have been doubts had we rather dully used that dreadful archaism background, which is, moreover, not entirely faithful to the truth, given that context gives not only the background, but all the innumerable other grounds that exist between the subject observed and the line of the horizon. It would have been better to call it a framework.

It is for this reason that, when we move away from everyday usage to the development of a scientiﬁcally adequate conception, Hasan (2001, p. 1) draws on Lenin’s metaphor to refer to context as ‘the aching tooth’, that troublesome sore spot to which the inquisitive tongue of the linguist cannot

166 Tom Bartlett 167 help but return again and again. And it is a topic to which Hasan herself has returned on many occasions over several decades as she has set about the task of mapping the variables of context and their relation to language within a unifying and systematic framework. A key theoretical move in Hasan’s work in this area has been her reﬁne- ment of the scope of context, from its amorphous everyday usage to a linguistic conceptualization of ‘relevant context’ (1995, p. 219) as ‘that part of the extralinguistic situation which is illuminated by language-in-use, by the language component of the speech event, the other name for which is text’. At the same time, however, Hasan has always been mindful of the relationship between the relevant context and ‘the larger canvas of human time and space’ (2001, p. 13), which would seem to relate to Saramago’s ‘innumerable other grounds that exist between the subject observed and the line of the horizon’. With regard to this wider canvas, Hasan (2009a, p. 36), in her call for a ‘theory of integrated sociolinguistics’, has been forthright in her evaluation of the limitations of existing work within systemic functional linguistics (SFL):

...for a theory that introduces itself as a social semiotic one, it [SFL] is woefully neglectful of specifically sociolinguistic issues; its only substantial contribution is in the field of discourse analysis where it offers a framework for the analysis of social context as well as for that of discourse. However, in both cases the emphasis is classificatory and concerned with the description of linguistic phenomena rather than sociolinguistic ones; the social enters only somewhat superficially, especially in the description of the vectors of context, making no reference to any sociological framework, and often confusing the description of a phenomenon with its production.

Related criticisms have come from outside SFL. Most relevant to Hasan’s concept of context as that which is made relevant by the language-in-use is Blommaert’s (2005, p. 35, emphasis in original) critique of Fairclough’s use of SFL within critical discourse analysis (CDA):

The emphasis on linguistic analysis implies an emphasis on available discourse, discourse which is there. There is no way in which we can linguistically investigate discourses that are absent, even if such an analysis would tell us an enormous amount about the conditions under which discourses are being produced (by whom? when? for what purpose?) and circulated (who has access to them and who doesn’t?). It also means that discourse analysis starts from the moment that there is linguistically encoded discourse, bypassing the ways in which society operates on language users and inﬂuences what they can accomplish in language long before they open their mouths, so to speak. It means that analysis stops 168 Society in Language, Language in Society

as soon as the discourse has been produced, while ...a lot happens to language users long after they have shut their mouths.

Hasan is, as we have seen, aware of such limitations, and has drawn, in particular, on the work of Bernstein, to suggest a larger architecture for the description of both language-in-use and the conditions for and repercus- sions of that use (Hasan, 2001, p. 15). Following from Hasan’s lead, in this chapter I make some observations about the form such a wider architecture might take and, in particular, the relationship between the context made relevant through the language-in-use and broader sociocultural factors. Specifically, this means questioning what exactly is meant by the term ‘relevant’ and a consideration of the different ways in which the relevance of contextual features is observable through the ‘traces’ they leave on the texts produced. It goes without saying that it is well beyond the scope of a single chapter to map out such an architecture. In Bartlett (forthcoming), I attempt to set out the most significant features in systematic form; in the present chapter I focus on a particularly ‘aching tooth’, the relationship between the written/spoken opposition, presented in various ways in the most significant systemic descriptions of context, and the linguistic features that are said to ‘construe’ these contextual features. This will also entail a consideration of what is meant by ‘construal’, a key term in SFL as a multistratal theory of language as a social semiotic. Bartlett (2013) considers the place of those ‘latent’ but unrealized discourses that were available in the context but which have not been made available as text.

7.2 First- and second-order context

A good place to begin this discussion of the architecture of SFL and its relation to and relevance for ‘a theory of integrated sociolinguistics’ is the SFL distinction between ﬁrst- and second-order contexts:

First-order context is a semiotic model of social processes (modelled as first-order field) and social roles and relations (modelled as first-order tenor). Second-order context is a semiotic model of linguistic and other semiotic processes in terms of second-order field (the domain of experience created by semiosis), second-order tenor (the speech roles and relations created by semiosis), and mode (which is inherently second- order: the role played by language and other semiotic systems in context). (Matthiessen et al., 2010, p. 77)

In this formulation,2 second-order context would appear to relate to what Hasan refers to as the relevant context, while ﬁrst-order context relates to the broader canvas (though things may not turn out to be so clear-cut).3 This Tom Bartlett 169 would appear to be a useful distinction, but it has been largely absent from the literature since Halliday’s original formulation4 and in what is probably the most accessible overview of SFL’s position on the relationship between context and language (Halliday and Hasan, 1989, p. 12), ﬁrst- and second- order context appear to be grouped together within a general category of context and in a rather non-uniform way across the three subcategories of Field, Tenor and Mode:

The FIELD OF DISCOURSE refers to what is happening, to the nature of the social action that is taking place: what is it that the participants are engaged in, in which language ﬁgures as some essential component?

The TENOR OF DISCOURSE refers to who is taking part, to the nature of the participants, their statuses and roles: what kinds of role relationship obtain among the participants, including permanent and temporary relationships of one kind or another, both the types of speech role that they are taking on in the dialogue and the cluster of socially signiﬁcant relationships in which they are involved?

The MODE OF DISCOURSE refers to what part language is playing, what it is that the participants are expecting language to do for them in the situation: the symbolic organisation of the text, the status that it has, and its function in the context, including the channel (is it spoken or written or some combination of the two?) and also the rhetorical mode, what is being achieved by the text in terms of such categories as persuasive, expository, didactic, and the like.

In the above definitions we see that Field is limited to those activities in which ‘language figures as some essential part’, to the exclusion of first-order features that are co-present and which, while not part of those activities themselves, may have some bearing upon them. Tenor, however, includes both the second-order semiotically created relationships between participants and first-order statuses and roles, whether these are reflected semiotically in the current event or not. Mode is a little more complicated as it can be argued (pace Matthiessen et al., above) that channel is either first or second order – the nature of the channel is realized in tandem with the act of speaking. But is it really ‘created by semiosis’? The rhetorical Mode is also a difficult case in that there is a rhetorical function (first order?) behind language acts before these acts are realized as text (second order). Each of these individual points will be taken up in the course of this chapter, but the more general thrust of my argument is that, while it may be acceptable to include both first- and second-order features within a single category of ‘context’, it is still necessary, given the concerns expressed by Hasan and others above, to see them as at different levels of abstraction within this category. This entails that there is a different set of relations between the levels from those that hold within 170 Society in Language, Language in Society them, between features at the same level of abstraction. Between levels there is a relationship of motivated selection, while within levels there is a relationship of configurational alignment and combined realizational capacity. In discussing these distinctions, I shall refer, following SFL theory, to context as a whole as a stratum, but, following recent work in sociolinguistics (e.g. Blommaert, 2005, pp. 133ff.), to the different levels within context itself as ‘scales’. The general grouping of Field, Tenor and Mode has remained more or less consistent in the five decades since its first appearance, despite the fact that they were produced before significant developments in the SFL architecture were introduced and notwithstanding Hasan’s considerable work in clarifying the content of these variables and the occasional movement of one feature to a different category that this has entailed. I would say that there are two principal reasons for this relative inertia: firstly, that SFL has been theoretically concerned with finding broad tendencies within the population as a whole and with regularities in variation; and, secondly, as a consequence of this, that the potential incongruity between first- and second-order contexts has been hidden by the fact that much applied work in SFL has focused on congruent behaviour (i.e. doing the expected thing) in (relatively) reified communicative contexts, as with the Bishop of Woolwich ‘doing being’ the Bishop of Woolwich (Halliday and Hasan, 1989, pp. 12–14). Such an approach has also been the hallmark in SFL work on genres and their relation to teaching.5 An alternative possibility is that the tension between first- and second-order features within the categories of Field, Tenor and Mode was always intended to be an object of focus in SFL analysis, in which case my argument for a more explicit and fuller recognition of the differences between the two orders and the relationships between them still stands. Unfortunately, in much SFL work the distinction is blurred, with congruent behaviour often treated as the sole logical outcome of the situational variables. It might be worth adding here that, in essence, SFL is a linguistic theory, the primary aim of which is to describe the properties of language as a social semiotic as opposed to a formal system, while the primary goal of sociolinguistics is to describe and explain variations in performance. This is not to say that SFL has ignored performance, far from it; in SFL, unlike most other linguistic theories, parole and langue are seen as two perspectives on the same system and the role of parole in shaping langue is a central functional tenet. Therefore, while SFL and other sociolinguistic approaches may differ in perspective, Hasan has consistently argued that these two perspectives should be congruent (as opposed to merely complementary) and that it is therefore a necessary endeavour to develop a unified global architecture that informs both theoretical and applied work. With regard to the SFL focus on tendencies, while this is a worthwhile sociological goal in its own right, it leaves open the question of what we Tom Bartlett 171 do with the exceptions to the posited rule other than labelling them as exceptions. In cases such as Blommaert’s (2005, pp. 78–123) analysis of gate- keeping interviews with asylum seekers, discussed below, it would surely be of benefit to all involved to understand the reasons behind the differences between expectations and performance. These are not just stylistic oddi- ties, but deep-rooted cultural differences that represent variations in voice (Hymes, 1996; Bernstein, 2000; Blommaert, 2005, pp. 70–8; Bartlett, 2012), the dialectic link between ways of speaking and social organization. It is a question worth pursuing, therefore, whether Hasan’s (Halliday and Hasan, 1989, pp. 97–116) concept of contextual configuration (CC), the interplay between Field, Tenor and Mode, is a sufficient tool in accounting for such differences in language use. If used superficially, there is a danger that such an approach will amount to little more than the mere classification and description that Hasan seeks to avoid (2009a, p. 36, above). There is a need, therefore, to move beyond listing the interacting contextual variables which correlate with individual language features and to seek explanatory inputs from the wider fields of anthropology and sociology, as Hasan herself (1995, pp. 267–8) suggests will be necessary. The above observations raise several points worthy of further discussion: (i) that the long-standing groupings of Field, Tenor and Mode may still be necessary but are no longer adequate in the current SFL framework; (ii) that they are potentially unsuitable for discussing individual case studies; and (iii) that they are potentially unsuitable in considerations of non-congruent behaviour or non-reified communicative events, many of which, such as the challenging of hegemonic norms (Bartlett, 2012) and the concept of superdiversity (Blommaert, 2013), are ongoing concerns in current CDA and sociolinguistics. In the following sections I consider these points with specific reference to the contextual variable of Mode and the place of the written/spoken opposition as a feature within it.

7.3 Written and spoken language as variables within the contextual category of Mode

As suggested above, there has been much criticism of the SFL treatment of context from within the wider disciplines of discourse analysis and sociolinguistics. Perhaps the most strident of these has been van Dijk (2008, pp. 28–55), who devotes an entire chapter to a sustained and harsh critique of the SFL approach. His most severe criticism is reserved for Mode, of which he says (with reference to Halliday’s 1978 deﬁnition) (2008, p. 43):

...‘mode’ – originally deﬁned as the role language plays in the situation – here includes a totally heterogeneous list of notions such as spoken mode, dialogue, reference to situation, textual cohesion (objects, processes), furthering child’s actions, as well as orientation to task. 172 Society in Language, Language in Society

While some of van Dijk’s critiques are a little under-informed and over- enthusiastic with regard to the complexities of the SFL literature, I think it is necessary to examine his stance in terms of the need outlined above to distinguish different scales within those features lumped together under ‘Mode’. In what follows I will take up this criticism but limit myself here to a consideration of ‘written/spoken’ as a contextual variable within Mode and to distinctions in ﬁrst- and second-order context (though this would be only the least delicate opposition in one of several necessary networks). The spoken/written contrast appears in Hasan’s systems for mode of discourse (Halliday and Hasan, 1989, pp. 58–9), or rather it appears twice, in order to account for genres such as speeches which are written in order to be spoken. One system, LANGUAGE CHANNEL, represents the phonic vs graphic distinction, while a second, LANGUAGE MEDIUM, is a scalar system which ranges from ‘most spoken’ to ‘most written’. This raises several questions: most notably, what exactly is meant by ‘most spoken’ and ‘most written’? Secondly, in terms of the phonic/graphic opposition, is the relation between the options in these systems and the features in what should be the stratum below, the semantics, that which is posited within standard SFL theory? Thirdly, does the phonic/graphic opposition relate to semantic features in the same way as other features posited as being within Mode? Fourthly, if the relation is not in accordance with SFL theory, what alternative system of oppositions can we put forward that is? Fifthly, what is the relationship between any alternative system of oppositions and those features that are listed within the current category of Mode, and in particular the phonic/graphic opposition? I address these last four questions in turn. In this way we can subject the phonic/graphic opposition and its relation to text to an examination in terms of the trinocular perspective advocated by Halliday before returning to my ﬁrst question concerning the least written/most written cline.

7.3.1 Is the relation between the phonic/graphic opposition and the features in what should be the stratum below, the semantics, that which is posited within standard SFL theory? Hasan (2001, p. 12) refers to the relationship between a context and the semantic features of a text as one of activation and construal: the context activates the features of text and conversely the semantics construe the context. This bidirectional relationship in which neither context nor semantics is given primacy is also a key principle in modern sociolinguistics (see Duranti and Goodwin’s seminal 1992 collection). In this section I ﬁrst consider the concept of construal to see if this relationship holds between the phonic/graphic opposition and linguistic features and then consider the concept of activation and discuss whether the two terms are indeed truly in a converse relationship. Tom Bartlett 173

Matthiessen et al. (2010, pp. 76, 206) agree with Hasan when they define the relation between adjacent strata as one of construal/realization: the higher stratum is construed or realized by the lower. However, across the SFL literature the construal/realization relation between text and context is defined in a variety of ways: context is what is illuminated by text (Hasan, 1995, p. 219; 2009b, pp. 176–7); context is what is made relevant by text (1995, p. 219); and context is what leaves a trace on the text (2009b, p. 177). I argue below that these, along with Hasan’s further claim that context can be read off the text alone (1995, p. 228), prove troublesome and potentially contradictory. We can, however, derive a definition of construal through an examination of the relationship between the strata of lexicogrammar and semantics. Here we see that combinations of abstract lexicogrammatical categories work together to create a higher or less abstract order of meaning, the semantic stratum. Thus the combination of declarative plus reverse-polarity tag at the lexicogrammatical stratum construes the move ‘reassure’ at the semantic stratum (Hasan, 1996, p. 123). This to me seems a good theoretical definition: features at one stratum work together in a synergetic but non-arbitrary way to realize a less abstract type of meaning in the stratum above. But is this the relationship that holds between the phonic/graphic opposition posited at the stratum of context and the semantic features that are said to correlate with them? Do variations in features such as thematicity/periodicity, reference and cohesion determine whether a text is phonic or graphic in the same way as an amalgam of lexicogrammatical and phonetic features determines whether a message is a ‘reassure’ move? Clearly not, as these same semantic features can be combined in the same way within both phonic and graphic texts. Taking another perspective, the bidirectional relationship between text and context entails that changes in context will result in changes in the text and vice versa. However, while it may be empirically the case that changes between the phonic and graphic channel result in changes in the semantics, it is manifestly not the case that changes in the semantics alter the channel through which the text is produced. I find it hard to maintain, therefore, that there is a specifically construal relation, as defined above, between the phonic/graphic opposition as a feature of context and the stratum below, that of the semantics. This is not to say, of course, that we can ignore the phonic/graphic opposition as a feature: the question is where to locate it in the architecture on the basis of the relations holding between it and other features. This brings us to the question of activation, and here we can say that changes in the phonic/graphic opposition do activate a different range of possibilities at the semantic stratum (at least in terms of percentages): certain features of textual semantics are more common in the written channel than the spoken and vice versa. It thus appears that construal and activation are not in a true converse relationship but instead are features of different 174 Society in Language, Language in Society scales of relationships between text and context in its widest sense, a point I develop below.

7.3.2 Does the phonic/graphic opposition relate to semantic features in the same way as other features posited as being within Mode? In discussing the concept of bidirectionality above I stated that changes in the context should entail changes in the text and vice versa. Looking at features of Field and Tenor, we can see how this operates. Firstly, for Tenor, if two speakers use semantic features such as first names, informal talk and equality in turn-taking, then they are construing a context of −power, +solidarity irrespective of the more reified first-order relations that hold between them (and see Watts, 2003, for a discussion of the congruence or otherwise between expectations of politeness and the features in specific texts). Now of course, we might want to ask why this Tenor here and now?, or, conversely, what is the effect of this Tenor in both long and short term?, and these are questions that necessitate a fuller understanding of the context at various scales. Similarly, as Hasan (1995, p. 255) illustrates, if a couple are discussing politics while doing the washing up, then it is politics that is construed as the context, not washing up – unless the topic switches for some reason to that activity.6 Of course, we might want to say that the washing-up is a relevant feature even when not mentioned: the couple might always talk politics at that time of night, perhaps inspired by the traditional allocation of domestic labour within British families, and these are again questions to be answered by considering different scales of context. My point here, however, is that in both my examples we see a truly converse relationship between features of the context (−power, +solidarity of Tenor in the first example and the Field of politics in the second) and the linguistic features that realize them. This relationship was not seen to hold for the phonic/graphic opposition. An analysis ‘from alongside’ would therefore militate against the phonic/graphic opposition being in the same relation to the semantics as are other variables of context such as the activity within Field, and power and solidarity relations within Tenor. And within the category of Mode itself we can see a similar distinction between CHANNEL and LANGUAGE ROLE, with its opposing features of ancillary and constitutive (Hasan, 1985, p. 58). If a text is textually constitutive, i.e. it construes a context without recourse to additional material action, then this will be apparent in the features of the text itself even if the participants and processes construed are in fact co-present. Similarly, if a text explicitly construes material features as if they were co-present, then this will be apparent in the features of the text itself (e.g. by the use of semantic features such as exophoric deixis), even if the processes and participants referred to are not in fact co-present.7 In all three cases then (a displaced Field, a temporary Tenor and non- congruent deixis), we see that the contextual features are defined as those Tom Bartlett 175 which are realized by the text itself and that the text and the context are therefore in a truly converse relationship of construal. This is not the case with the phonic/graphic opposition as described, where the context is defined according to features of the material setting irrespective of the linguistic features of the text which are said to construe that context.

7.3.3 If the relation is not in accordance with SFL theory, what alternative system of oppositions can we put forward that is? A central point of my argument above has been that we need to distinguish between what I shall refer to as the ‘sctx’,8 the context that is directly construed in and by language, and the material, interpersonal and sociocultural environment. My term ‘sctx’ is roughly analogous to Halliday’s ‘second- order context’ and what Berry (this volume) calls the ‘via-text context’, while my term ‘environment’ is roughly analogous to the first-order context and what Berry calls the ‘pre-text context’. There is a need to introduce these separate terms, at least as placeholders, however. Sctx refers purely to the context that is construed in the semantics while Berry’s via-text context potentially refers to the effect of the text on the wider context (pers. comm.) and, as observed in note 2, there are areas of disagreement as to what Halliday’s second-order context refers to. Similarly, my term ‘environment’ refers to anything that exists outside the text itself but which might affect it in any way, from the time of day to the sociocultural history of the speakers’ ethnic groups. I am not sure if either Berry’s or Halliday’s terms are as inclusive.9 With regard to Hasan’s categories, it might appear that my sctx is just another term for what she labels the relevant context, while my environment relates to Hasan and Cloran’s term ‘material situational setting’ (MSS; Cloran, 1999, p. 178), but there are differences. As stated above, sctx only includes those features that are directly realized in the semantics in a converse relationship of construal, whereas Hasan’s conception of relevant context includes a much wider and more obliquely related set of variables, such as age and gender. Sctx is not seen as an alternative to relevant context, therefore, but as a means of distinguishing between what I see as ‘direct construal’ and the many other relevance relations that can hold between the linguistic and extra-linguistic worlds and which belong (at different scales) within a wider conception of context. Conversely, Hasan’s MSS is a much narrower concept than my environment, being largely restricted to the material here and now of the setting. In Bartlett (forthcoming) I set out a schema for more delicate distinctions in both sctx and environment, but the basic distinction can be made either by placing the environment at a stratum above the sctx or by recognizing different scales between both the sctx and the environment and between different features of the environment10 within a general category of context. In this way we make a clear distinction between, for example, the long-term social roles of speakers (environment) and the Tenor relations they might 176 Society in Language, Language in Society wish to construe (sctx). Such a distinction forces us to ask, for both congruent and non-congruent behaviour, why this sctx in this environment? How do generic conventions arise and under what conditions are they broken? It also highlights distinctions between reified and non-reified relationships between language and situation. In new social environments, for example, the generic expectations of different speakers might be in tension or new, hybrid genres might be emerging (Bartlett, 2012, 2013). Whether we see sctx and environment as in different strata or at different scales within a single stratum of context, rescaling the phonic/graphic opposition to a position above the sctx leaves a gap in the model as there is now no corresponding textual element within the sctx. However, following the logic of construal set out above, we can ask what is construed through the amalgamation of textual features at the semantic stratum across whole texts? My answer to this would be textuality, or text type (which would seem to chime with what van Dijk (2008) says about Mode being ‘originally defined as the role language plays in the situation’). To illustrate the contextual role of textuality we can draw on more work carried out by Hasan, in conjunction with Cloran, into the concept of Rhetorical Units (RUs) and the relations between them (Cloran, 2010). RUs are defined as stretches of text within which the semantic features of central entity (CE – realized in the lexicogrammar as the Subject) and event orientation (EO – realized in the lexicogrammar as the Finite) are held constant. Thus a Commentary is a stretch of text in which the CEs are all interlocutors in the event and the EO is concurrent with the event. Commentaries are at one end of a spectrum of contextualized and decontextualized language,11 with Conjectures and Generalizations at the other (see Cloran, 2010 for a full schema). There are various links between RUs, realized by Theme–Rheme relationships between the last message of one RU and the first of the next. One RU can be embedded within another, when the content of a previous RU is picked up in the Theme of the initial message of the new RU. Alternatively, one RU is an expansion of the previous one when the content is picked up in the Rheme of the new RU.12 Finally, RUs are discontinuous when the content picked up on comes from a non- adjacent RU. When no textual connection can be found between one RU and a previous one it is claimed that we have a new text. It can be seen, therefore, that we can characterize the textuality of a whole text in terms of the RUs that realize it and the relations between them. This begs three questions: as multi-message elements, at what rank do we place RUs? Are RUs purely textual? And how do we classify whole texts in terms of the RU structure? In terms of rank (including text for the purposes of argument), it could be argued that RUs are equivalent to clauses/messages in experiential and interpersonal terms, given that it makes little sense to talk of textuality in terms of single clauses/messages. Alternatively, it could be argued that Tom Bartlett 177 they occupy a unit of text intermediate between clause/message and text. This is the position taken by Hasan and Cloran (e.g. Cloran et al., 2007), though they take the view that RUs combine features from all three metafunctions, which leads to the second question. I see a problem with Hasan and Cloran’s categorization as it is based, in part, on the idea that a defining feature of RUs, the CE, conflates unmarked Theme (a textual feature), with a participant (an experiential feature) and the Subject (an interpersonal personal feature). However, it is not the CE’s function as Subject that gives it relevance within the RU architecture, nor is the referent itself of concern, as is shown by the fact that RUs continue even when the reference of the CE changes. Of the three conflated functions, it is only the thematic quality of the CE that is relevant, and this, in terms of its deictic characteristics. As both Theme and deixis are textual elements, we can say the CE is a textual element both in terms of selection and of categorization. But what of the EO? Halliday places tense within the ideational metafunction because of the logical chaining of features of complex tenses in English. However, I would say that logical relations hold in textual and interpersonal as well as ideational meaning (though this is an argument for another day). More importantly, tense is a deictic category, the temporal correlate of spatial deixis, which is the defining feature of the CE. Temporal and spatial deixis are between them, therefore, the defining features of an RU. An alternative view of tense might also define it as experiential as it relates to our representations of the world. By this view we state the tense of an event according to when it happened and not for any textual reasons. I would question this assumption, positing instead that when we talk we are motivated by two simultaneous factors: the subject matter (experiential orientation) and the relationship of the talk to the here and now as, for example, a Recount, a Commentary or a Generalization (textual orientation). Neither of these two orientations has automatic primacy. For example, when we are asked to recount what we did last night, we start with a past time frame and then fill it with the relevant subject matter. In this way textuality can be seen as the function of the language in context (chiming again with Mode being ‘originally defined as the role language plays in the situation’). Turning to my third question, it follows from everything that I have been arguing so far that, while the configuration of RUs across texts will be responsive to changes in the non-linguistic environment, the categorization of these configurations as the textual component of the sctx has to be based entirely on the semiotic features displayed. This is work that has yet to be done. However, my feeling is that, just as interpersonal and experiential meanings shift across texts, it will be more appropriate to define textual units at various levels between the single RU and the text as a whole, and then to account for shifts across the text as recalibrations of the context (Bartlett, 2012, ch. 6). 178 Society in Language, Language in Society

7.3.4 What is the relationship between any alternative system of oppositions (sctx, textuality) and those features that are listed within the current category of Mode, and in particular the phonic/graphic opposition? Many first-order features of the environment appear in the text as sctx, as when people make reference to their material surroundings. In such cases it seems appropriate to say that the text construes these features as relevant. However, as argued above, the construal/activation relationship is not a true converse relationship, as there are other features of the environment that leave a trace upon the text, in that they affect what is construed within it, or alter the probabilities of construal, but which do not figure directly in the text itself and are not, therefore, construed as part of the sctx. Activa- tion is therefore a wider-ranging concept than construal and is often in an oblique relationship, via features that are directly construed, with the sctx itself. Hasan’s glossing of context as being that which is illuminated by the text or that which can be read off the text alone would therefore appear to cover direct construal but not more oblique activations. The notion of leaving traces on the text, however, covers both direct construal and oblique activation. As I have argued elsewhere (Bartlett, 2013), when a Scotsman refuses to allow a redhead to first-foot him, the would-be visitor’s red hair does indeed ‘leave a trace’ on the text but it would be difficult to claim that this physical feature of the context is either illuminated by the text or can be ‘read off’ it alone. What I am arguing, then, is that there are different relationships between the various features of context, as currently defined, and the sctx as the amalgam of textual semantics. These relationships occur across different scales within the broader concept of context and it is necessary to make these distinctions theoretically explicit. In these terms there are several distinct types of relationship holding between the phonic/graphic opposition and the sctx. Firstly, I would suggest that the use of a graphic or phonic medium serves as a constraint on the types of sctxs that are construed, or the ways in which they are construed, owing to physical differences such as the type of contact between interlocutors, the speed of response between messages, and the ability to refer to co-present features of the environment without naming them. These are features that appear in one contextual system of Mode or another. As constraints, however, they are at a higher and less abstract scale within the environment than the sctx in that they are first-order features as opposed to second-order ones. Moreover, physical constraint is not the only relationship involved: another possibility is that in specific environments, the classroom, for example (see Berry, 2013a, b), there is a social preference for certain types of text, or textualities, construed by specific semantic features, a question I will discuss in more depth when I return to my first question. Here I would just like to make the point that social preferences are no more in a construal relationship with texts than are constraints. Rather, they are both first-order Tom Bartlett 179 phenomena that in some way limit or guide the semiotic choices we make when construing a context and are therefore at a higher scale than the construed context itself. My overall argument in the last few sections has been that the phonic/graphic opposition only affects alternations in the semantics through the mediating concept of sctx and that, while the relationship between the sctx and the semantics is one of direct construal, there are various types of relationship between the sctx and features of the environment. These include direct construal, but also such relations as constraint and preference. Within the general concept of ‘context’ as currently defined, therefore, there is a difference in scale both between the sctx and the environment and between different features of the environment. Taking this argument a step further in reference to the phonic/graphic opposition, I would say that these and other ‘modes’ such as images (i.e. at a scale above sctx) each have their own resources which are utilized in specific ways, as well as specific constraints and affordances that activate differences in sctx and semantics across all three metafunctions. This is an argument also made by van Leeuwen (1999, p. 190), who states that:

different semiotic modes have different metafunctional conﬁgurations, and ...these metafunctional conﬁgurations are neither universal, nor a function of the intrinsic nature of the medium, but cultural, a result of the uses to which the semiotic modes have been put and the values that have been attached to them.

In this way different modes are comparable to different codes, a term which encompasses both different languages and different uses of ‘the same language’ by different social groups (Bernstein, 2000), and Mode-switching is as far-reaching in its effects across the metafunctions as is code-switching.13

7.3.5 What is meant by ‘most spoken’ and ‘most written’? Returning to the first of my questions above, what can we make of the terms ‘most spoken’ and ‘most written’ which form a cline under the label MEDIUM in Hasan’s (1985, p. 85) system for Mode? I believe this was introduced as a simultaneous system alongside CHANNEL to account for language that is written to be spoken, which is therefore purely neither written nor spoken. At best, however, these terms are circular (or merely descriptivist with no motivational explanation provided): the context is most written when it displays semantic features defined as most written and semantic features are most written when they occur in such a context. Bowcher (2014) makes a similar point with regard to Butt’s inclusion of Medium and Channel as simultaneous features of context within Mode, with Medium defined as ‘the pattern of the textual organisation adopted when “organisation” pertains to all those reactances of “spoken-ness” and “written-ness” ’ (Butt, 2004, p. 7). Bowcher (2014, p. 202), however, sees such patterning as relating ‘more 180 Society in Language, Language in Society to the realization of Mode than the “realisation-activating contextual features” ’ of Mode as context. Bowcher goes on to make a point along the lines of my argument here when she says (Bowcher, 2014, p. 202) that ‘Butt admits that his view of context is relatively flexible in the sense that the line between the non-linguistic (context) and the linguistic (semantics) is not as defined as it is for Hasan (Butt, pers. comm.).’ A more contentious interpretation would be that Medium in both Butt and Hasan’s systems refers not to reactances per se but to expectations of written and spoken language. Questions then arise as to whose expectations these are and for what reasons and with what effects. These are tremendously important considerations in applied work, particularly in culturally diverse settings (in terms of any relevant variables such as class, ethnicity, etc.). Blommaert (2005, pp. 78–123), for example, describes how asylum seekers in Belgium produce written applications which are not in accordance with the expectations of the gatekeepers and asylum is often refused. While the tensions include factors beyond the textual features of the applications, elements of textuality such as deixis and cohesion (and even the use of pictures) are crucial. In cases such as these we have to question the rationale behind categorizing aspects of these written texts, such as exophoric deixis, as contextually at the ‘least written’ end of the cline of MEDIUM, as well as the potential consequences. Surely it would make much more sense, theoretically and practically, to describe the sctxs of the texts produced in terms of their textuality and to look at the tensions between this and the textuality expected by the gatekeepers and to relate both of these to wider environmental features at multiple scales. Why, pragmatically and politically, do the gatekeepers expect certain textualities? What function do they expect the text to fulfil? And, similarly, what is the intended function of the texts for the asylum seekers and what are the cultural reasons behind their textual choices? In what social contexts did they learn to use their linguistic resources in this manner (cf. Hasan, 2009a)? Such questions are related to Silverstein’s (2003) work on ‘indexical orders’ and Blommaert’s (2005) extension of this to include the concept of ‘orders of indexicality’. In my terms, how do the sctxs construed by certain social groups index their group membership and what is the sociocultur- ally determined order of preference for these within specific intercultural settings? Beyond this, what are the conditions of possibility (linguistic and social) for the order of indexicality to be perturbed in favour of marginalized groups (see Bartlett, 2013, on the concept of ‘perturbation potential’)?

7.4 Conclusions

To conclude, throughout this chapter I have questioned the theoretical and practical suitability of including both first- and second-order features of Tom Bartlett 181 context within a single category of ‘context’ and have argued both for a primary distinction between the sctx and the environment and for a recognition of the different scales operating within the environment. Many, though not all, of the specifics I have looked at have been discussed by Hasan in her voluminous considerations of the relations within context and between context and text. In particular, the horizontal relations captured in her concept of contextual configuration (CC) are essential in showing that features of context do not relate in isolation to features of texts but through processes of alignment. Furthermore, Hasan (2001) has considered the need for vertical relations in mapping the relations between all the variables that are in some way relevant to the construal of texts in a systematic manner. As Hasan acknowledges, however, SFL still has a lot of work to do in this area. This chapter has sought to bring out some of the complexity that Hasan points to but which is, I believe, currently obscured in the current SFL modelling of text and context. This is not seen as an end in itself, but as one small step in the ongoing mapping of the horizontal and vertical dimensions of context and the different relations that hold between these dimensions and between texts, the features of which have meaning only as they relate to the contexts they construe.

Notes

1. My thanks to Margaret Berry and Annabelle Lukin, with whom I discussed many of the ideas in this chapter. Once differences in terminology reflecting our different theoretical perspectives are overcome, it will be seen that there is a significant amount of congruence between the arguments in Berry’s chapter and my own, though they approach the question from different ends of the theory–analysis dialectic. 2. I would agree with Annabelle Lukin, who points out (pers. comm.) that the distinctions here do not seem to correspond to Halliday’s original distinctions. I would also agree with Margaret Berry (pers. comm.), who sees discrepancies in the original distinctions themselves. 3. Hasan (1985, p. 100) does use the related term ‘second-order semiosis’ in relation to verbal art. 4. Matthiessen (2009, pp. 31ff.) and Halliday and Matthiessen (1999, section 9.3) are exceptions. 5. This is not to say that I take issue with the teaching of mainstream genres to minority groups in this fashion, rather that I see the framework as limited with regard to other situations. 6. Hasan (1995, p. 255) expands upon the potential in which silent (or latent, Bartlett, 2013) features of the material situational setting (MSS, which she distinguishes from the ‘relevant context’) might intrude into ongoing contexts. 7. This is often the case with children’s writing. 8. This was originally an acronym for the semiotic context, though second-order, situational or SFL context would also be possible. However, as none of these terms would be unproblematic and as several already have specific meanings in the literature I find it easier to work with an abstract label. As stated above, sctx is 182 Society in Language, Language in Society

similar to several other theoretical categories, and if it turns out that there is little new in the term, then perhaps sctx can be decoded as superﬂuous. 9. The relationship between environment and context of culture is a topic for another day. 10. I would argue that it is the failure to recognize this that has led to problems in the SFL literature where there is disagreement and change not just in what systems are needed in a particular contextual category but even in which category existing systems belong (e.g. Hasan, 1999, on the placement of the ancillary/constitutive opposition in either Field or Mode). 11. Along with Actions. Commentaries are informative while Actions are regulatory. 12. There is room therefore for a further subcategorization of RUs depending, for example, on whether the content picked up on is in the Theme or Rheme of the previous RU. My colleague Gerard O’Grady and I are working on such subcategorization, also taking into account the role of intonation. 13. It should be noted in passing that Mode-switching might entail a difference in the way in which certain semantic differences are realized rather than altering the systemic choices available.

References

Bartlett, T. (2012) Hybrid Voice and Institutional Change: Contextualising Positive Dis- course Analysis (Abingdon and New York: Routledge). Bartlett, T. (2013) ‘ “I’ll manage the context”: context, environment and the potential for institutional change’. In L. Fontaine, T. Bartlett and G. O’Grady (eds) Systemic Functional Linguistics: Exploring Choice (Cambridge: Cambridge University Press), pp. 346–64. Bartlett, T. (forthcoming) ‘Context in systemic functional linguistics’. In T. Bartlett and G. O’Grady (eds) The Routledge Handbook of Systemic Functional Linguistics (Abingdon and New York: Routledge). Bernstein, B. (2000) (rev. edn [1996]) Pedagogy, Symbolic Control and Identity: Theory, Research, Critique (Lanham, Boulder, New York and Oxford: Rowman and Littlefield Publishers Inc.). Berry, M. (2013a) ‘Contentful and contentlight Subject Themes in informal spoken English and formal written English’. In G. O’Grady, T. Bartlett and L. Fontaine (eds) Choice in Language: Applications in Text Analysis (London: Equinox), pp. 243–68. Berry, M. (2013b) ‘Towards a study of the differences between formal written English and informal spoken English’. In L. Fontaine, T. Bartlett and G. O’Grady (eds) Systemic Functional Linguistics: Exploring Choice (Cambridge: Cambridge University Press), pp. 365–83. Blommaert, J. (2005) Discourse: a Critical Introduction (Cambridge: Cambridge Univer- sity Press). Blommaert, J. (2013) Ethnography, Superdiversity and Linguistic Landscapes: Chronicles of Complexity (Bristol, Buffalo and Toronto: Multilingual Matters). Bowcher, W. L. (2014) ‘Issues in developing unified systems for contextual field and mode’. Functions of Language 21 (2): 176–209. Butt, D. (2004) ‘Parameters of context: on establishing the similarities and differences between social processes’. Unpublished mimeo, Macquarie University, Sydney. Cloran, C. (1999) ‘Context, material setting and text’. In M. Ghadessy (ed.) Text and Context in Functional Linguistics (Amsterdam: Benjamins), pp. 177–217. Tom Bartlett 183

Cloran, C. (2010) ‘Rhetorical unit analysis and Bakhtin’s chronotope’. Functions of Language 17 (1): 29–70. Cloran, C., Stuart-Smith, V. and Young, L. (2007) ‘Models of discourse’. In R. Hasan, C. M. I. M. Matthiessen and J. J. Webster (eds) Continuing Discourse on Language (Vol. 2) (London and Oakville: Equinox), pp. 647–70. Duranti, A. and Goodwin, C. (1992) Rethinking Context: Language as an Interactive Phenomenon (Cambridge: Cambridge University Press). Halliday, M. A. K. (1978) Language as Social Semiotic: the Social Interpretation of Language and Meaning (Baltimore: University Park Press). Halliday, M. A. K. and Hasan, R. (1989) Language, Context and Text (2nd edn) (Oxford: Oxford University Press). Halliday, M. A. K and Matthiessen, C. M. I. M. (1999) Construing Experience through Meaning (London and New York: Continuum). Hasan, R. (1985) Linguistics, Language and Verbal Art (Geelong: Deakin University Press). Hasan, R. (1995) ‘The conception of context in text’. In P. Fries and M. Gregory (eds) Discourse in Society: Systemic Functional Perspectives. Meaning and Choice in Language: Studies for Michael Halliday (Westport, Conn. and London: Ablex), pp. 183–283. Hasan, R. (1996) ‘Semantic networks: a tool for the analysis of meaning’. In C. Cloran, D. Butt and G. Williams (eds) Ways of Saying: Ways of Meaning: Selected Papers of Ruqaiya Hasan (London: Cassell), pp. 104–31. Hasan, R. (1999) ‘Speaking with reference to context’. In M. Ghadessy (ed.) Text and Context in Functional Linguistics (Amsterdam: Benjamins), pp. 219–328. Hasan, R. (2001) ‘Wherefore context? The place of context in the system and process of language’. In S. Ren, W. Guthrie and I. W. R. Fong (eds) Grammar and Discourse: Proceedings of the International Conference on Discourse Analysis (Macau: University of Macau), pp. 1–21. Hasan, R. (2009a) ‘Wanted: a theory for integrated sociolinguistics’. In J. J. Webster (ed.) The Collected Works of Ruqaiya Hasan,Vol.2:Sematic Variation: Meaning in Society and Meaning in Sociolinguistics (London and Oakville: Equinox), pp. 1–40. Hasan, R. (2009b) ‘The place of context in a Systemic Functional model’. In M. A. K. Halliday and J. J. Webster (eds) Continuum Companion to Systemic Functional Linguis- tics (London: Continuum), pp. 166–89. Hymes, D. (1996) Ethnography, Linguistics, Narrative Inequality: Towards an Understand- ing of Voice (London and Bristol, Pa: Taylor and Francis). Matthiessen, C. M. I. M. (2009) ‘Ideas and new directions’. In M. A. K. Halliday and J. J. Webster (eds) Continuum Companion to Systemic Functional Linguistics (London: Continuum), pp. 12–58. Matthiessen, C. M. I. M., Teruya, K. and Lam, M. (2010) Key Terms in Systemic Functional Linguistics (London: Continuum). Saramago, J. (2008) Death at Intervals. Translated by Margaret Jull Costa (London: Vintage). Silverstein, M. (2008) ‘Indexical order and the dialectics of sociolinguistic life’. Language and Communication 23: 193–229. Van Dijk, T. A. (2008) Discourse and Context: a Sociocognitive Approach (Cambridge: Cambridge University Press). Van Leeuwen, T. (1999) Speech, Music, Sound (Basingstoke: Macmillan). Watts, R. J. (2003) Politeness (Cambridge: Cambridge University Press). 8 On Describing Contexts of Situation1 Margaret Berry

8.1 Introduction

8.1.1 General purpose My general purpose in linguistics is to provide information which I hope will be helpful to teachers of English. Schoolteachers I have worked with have said that they see themselves as (a) extending the range of choices open to their students and (b) helping the students to make the choices appropriately. I have written elsewhere (Berry, 2013a, b) about the importance of the SFL notion of ‘choice’ for work with teachers. The present chapter is relevant to (b). ‘Appropriately’ presumably means appropriately in relation to context. I hope to work towards an answer to the question: what contextual features do adult native speakers and writers respond to when they speak and write appropriately? These would seem to be the contextual features that it would be most worth teachers alerting their students to, so that the students might speak and write appropriately also. This should not be taken to imply that I am expecting teachers to impose norms on their students. Norms can be taught descriptively as well as pre- scriptively. It is always open to speakers and writers to go against norms. But as I have said (Berry, 2013b), going against a norm because you have chosen to do so is a very different matter from going against a norm because you do not know what the norm is.

8.1.2 More immediate purpose The aim indicated above is a long-term one. My more immediate purpose is to engage with Hilary Hillier and Geoff Thompson in a project designed to investigate how two particular people (Hilary and her close friend Ruth) use language in different ways in different contexts of situation (Berry et al., 2014). This is relevant to my general purpose as, effectively, one of the questions we are asking is: what contextual features are Hilary and Ruth responding to when they speak and write appropriately? Our procedure is as follows: (1) Hilary has provided the data, this consisting of (a) texts ranging

184 Margaret Berry 185 from conversations through letters to a book and PhD theses and (b) her participant’s account of the contexts which gave rise to the texts (for details and discussion, see Hillier, 2013);2 (2) Hilary and I are now attempting to characterize the contexts in such a way as to bring out the differences in contextual features between them; (3) Geoff and I will then analyse the texts to see if the contextual features we have identified seem to have influenced the semantic choices that Hilary and Ruth have made. The general hypothesis is that each of the contextual features we have identified will be found to ‘have left a trace’ (Hasan, 2009, p. 176) in the texts in that it does seem to have affected the semantic choices. If this turns out not to be the case, we shall of course have to revise our system networks.

8.1.3 The importance of Hasan’s work in this connection My role at present is to read up on systemic functional linguistic (henceforth SFL) theoretical work on context, to see if it does or does not relate to our project. Hasan has I think got further than anyone else in SFL with work on context, so her work is the obvious place to begin. The ﬁrst thing anyone else needs to do if they want to say anything about context is to position themselves in relation to Hasan’s work. As I show in the following sections, I ﬁnd myself in agreement with a very great deal that Hasan says, although I am also aware of certain points of difference. After general theoretical discussion of Hasan’s views (sections 8.2–8.4), I attempt an analysis of the contexts of situation of our data, again drawing on Hasan’s work for certain details of the analysis.3

8.2 Points of agreement with Hasan

1. Hasan distinguishes between material situational setting (MSS) and relevant context (e.g. Hasan, 2009, pp. 177–8). Not all aspects of the immediate situation of a language event are necessarily relevant to the language event itself. I am not sure I am going to be drawing the distinction in quite the same way as Hasan, but I certainly agree that a distinction of this kind is necessary. This is discussed further below. 2. I also agree with Hasan that the so-called ‘rhetorical modes’, tradition- ally in SFL regarded as part of Mode – things like explaining, deﬁning, generalizing, reporting, narrating, etc. – are better regarded as part of Field. There will not be room in this chapter to further discuss this, but I accept Hasan’s view, for the reasons she gives (Hasan, 1999, pp. 270–1 and 281–2). 3. I like Hasan’s notion of a contextual conﬁguration (CC) – that is, the set of contextual features that relate to a particular language event (e.g. Hasan, 1999, p. 232). 4. However, I also agree with Hasan that a CC may not remain constant throughout a language event, and that it is necessary to specify which 186 Society in Language, Language in Society

contextual features do remain constant and which change (Hasan, 1999, pp. 225–6 and 246–7). I distinguish between the starting contextual configuration for a language event and modified contextual configurations which are the results of changes. Hopefully, in course of time, it may be possible to arrive at an assessment of what types of contextual feature are most likely to change, in what types of language event. 5. I particularly like Hasan’s 1999 paper, as she is working with real texts, and drawing out the differences between the contexts of the texts. In the case of her third text, she draws out the differences between the contexts of different sections of the text. We hope to emulate her in this. 6. I also agree with Hasan that descriptions of instances of context must go hand in hand with work on the system of context (Hasan, 2009, p. 175). Hasan writes of ‘the potential of context’. In our project we see ourselves as working towards an account of context potential to set alongside accounts of meaning potential. Hopefully if enough of us engage in empirical accounts of particular instances of context, it will be possible to detect a general pattern. Bowcher (2014) is already attempting to integrate proposals from Butt’s work, Hasan’s and her own.4 7. I emphatically agree with Hasan that we should be attempting to draw system networks for context. As Hasan points out (2009, p. 185), if one merely describes a context in ordinary words, the description cannot be challenged by anyone who does not have the text and its context immediately in front of them. But ‘Systemizing the possible relevant features of context makes the claims explicit ...and raises the options to the conscious level as an object under description. Thus the description can become a focus of discussion, and objections can be made ...’. Hasan acknowledges a paper by Bowcher (2007), which she regards as a ‘valuable critique’ of her 1999 paper (Hasan, 2009, note 24). In Hasan’s view, this kind of thing – explicit claims followed by criticism followed by revision of claims – is precisely what ought to be happening if real progress is to be made. I totally agree with this. (So do please feel free to challenge the system networks that I shall be presenting later in this chapter!)

8.3 Points of difference from Hasan

My starting questions are different from Hasan’s, and the texts I am working with are very different from hers. So it is not surprising if my approach differs from hers in some ways, at least in emphasis. Two points seem to me to be particularly important, these leading to a third.

1. Hasan distinguishes between interactants’ views of context and analysts’ views, but seems more concerned with the analysts’ views (e.g. Hasan, 1999, p. 240). Where she does consider interactants’ views, she seems more interested in the views of language receivers than in those of language producers (e.g. Hasan, 1999, p. 238). My focus, on the other hand, is on language Margaret Berry 187 producers, on the speakers and writers. I have said above that I want to discover what contextual features adult native speakers and writers respond to. In other words, I am concerned with how the speakers and writers construe contexts and how they respond in the semantic choices they make to the contexts they have construed.5 2. The second point here needs a bit of introduction. Halliday (1985/1989, p. 5) said: ‘I am going to talk about context first, for the reason that, in real life, contexts precede texts. The situation is prior to the discourse that relates to it.’ Hasan (1999, p. 221, emphasis in original) challenges the view that ‘context is always already there even before any speaking has been done’. She acknowledges that there is some truth in this, but draws attention to ‘the power of language to create context’. From then on she seems to be almost entirely concerned with aspects of the context that have been created by language, to the exclusion of aspects of the context that have preceded the texts. Cloran (1999, p. 180) takes an even more extreme line. She says: ‘Prior to language there is no context.’ So who do we believe – Halliday or Cloran? Hasan, from what she says on p. 221, sounds as if she is going to be somewhere in the middle. But in practice she is not. She is much closer to Cloran. I certainly do want to take a middle line. I want to say that contexts, relevant contexts, at least in part, precede texts, but that there is a potentiality for the interactants, through their language, to change or at least modify their contexts. I want to investigate both the PRE-TEXT RELEVANT CONTEXTUAL FEATURES and the VIA-TEXT RELEVANT CONTEXTUAL FEATURES. I want to see if they have influenced Hilary’s and Ruth’s semantic choices. To give examples of PRE-TEXT and VIA-TEXT features: when Ruth phoned Hilary, as soon as the telephone rang and Hilary picked it up, this put in place the contextual feature ‘spoken’ (in the sense that communication was to be by voice, rather than by marks on the page/computer screen, etc.) (see Figure 8.1) and also the contextual feature ‘not co-present’ (in the sense that Hilary and Ruth were not in each other’s presence).6 These features were not created by the language Hilary and Ruth used. They were in place before Hilary or Ruth said anything at all. So I would regard them as PRE-TEXT features. The position with regard to social distance features is rather more complicated. If we take the letters in our data, those between Hilary and Ruth are between close friends, while Hilary’s letters to her publishers are to people she has never met. If we take ‘never met’ to be a PRE-TEXT feature, we might expect these letters to display maximum social distance. But Hilary saw herself as writing to her publishers in a friendly and informal manner. In effect, VIA her TEXTS Hilary reduces the social distance. We need to take into account both the never-met-ness and the friendliness leading to reduced social distance if we are to explain the semantic choices that Hilary has made. Hilary’s letters to her publishers are friendly and informal, but they are still not as friendly and informal as her letters to Ruth. She has 188 Society in Language, Language in Society introduced the VIA-TEXT features of friendliness and informality, but seems to have been constrained in the degree to which she can do this by PRE-TEXT features such as never-met-ness. I say more about this below. 3. A third point, which links the first two: Hasan writes of analysis being post hoc (Hasan, 1999, p. 236). She says (2009, p. 177): ‘relevant context refers to that frame of consistency which is illuminated by the language of the text’, which seems to imply that one does not know what the relevant context is until one has seen the text. That is OK if one is thinking from the point of view of the analyst. But one would not want to say, would one, that speakers and writers do not know what kind of context they are in until they see what they have said, that they do not begin to construe the relevant context until after they have produced their texts? I would suggest that, at the outset of engaging in a language event, speakers and writers (subconsciously or consciously) abstract from the MSS contextual features that they consider relevant and that in the unmarked case they then adopt the associated register. As the language event proceeds, they may through their language introduce other contextual features, which have the effect of modifying or changing the starting contextual configuration. So my interest in PRE-TEXT features follows from my concern with speakers and writers. These features may not come before the text for the analyst, but I am suggesting they do for speakers and writers. This is not to devalue post hoc analysis. Indeed we shall be engaging in post hoc analysis in our project (see section 8.1.2 above). But there would seem to be a difference between post hoc analysis which assumes that the context itself is post hoc, and analysis which aims to discover whether particular semantic options have been activated by contextual features which appeared to be important to Hilary and Ruth in advance of their texts.

8.4 Further theoretical discussion

Further discussion is necessary of some of the points raised above.

8.4.1 Material situational setting vs relevant context I said above that I agreed with Hasan that some distinction such as material situational setting vs relevant context was necessary, but I left open whether I would be drawing the distinction in precisely the same way as Hasan. There seem to be two oppositions involved here: the opposition between material and non-material and the opposition between potentially relevant and actually relevant. (This depends to some extent on exactly what is meant by ‘material’.) For me, the second opposition is more important than the ﬁrst. I do not see any great difference between material and non-material features of the situation from this point of view. Both can be potentially relevant, but actually relevant in some language events, not others. For instance, at Margaret Berry 189 a meeting there will probably be physical things such as chairs in the situational setting. These may or may not be relevant to the language that ensues. At one committee meeting I attend, chairs have to be moved around at the start of the meeting and there is a good deal of discussion of this and of who is going to sit where. At other meetings I have attended, chairs have simply been taken for granted and there has been no reference to them whatsoever. The same can be true of non-material features of the situational setting such as relationships. Bartlett (2012) discusses meetings in which relationships relevant outside the meetings are set aside in favour of the role relationships of the meeting. Henry refers to his son as ‘Mr Chairman’. Walter refers to close friends as ‘Miss Hilda’ and ‘Mr Bramley’. I assume then that the material situational setting – I would prefer to drop the word ‘material’ and simply call it ‘situational setting’ – includes all features both material and non-material that are potentially relevant to the language event. I assume that speakers and writers abstract from the situational setting those features that they deem to be actually relevant.

8.4.2 The system networks Also above, I agreed with Hasan that we should be trying to draw system networks for context. But what would such system networks represent? As always with system networks in SFL, I assume that each system represents a choice, a choice between options, the options being represented by the terms in the system. But what kinds of choice? I suggested (Berry, 2013b) that speakers and writers would need to make two kinds of choice with regard to context: choices about how they would construe the context in which they found themselves; choices about whether they wanted to accept the context as they had construed it or whether they wanted to try to change the context. The PRE-TEXT features I have been discussing would seem to be relevant to the ﬁrst of these kinds of choice, the construal choices, choices as to which of the features of the situational setting are actually relevant to the language event. Is the context to be spoken (communication to be by voice) or written (communication to be by marks on page/computer screen, etc.)? Are the interactants co-present? Are the interactants close friends or is it the case that they have never met? Very often such choices would be made subconsciously, by adult native speakers at least, although language learners might need to have their levels of consciousness raised. (See Berry, 2013b, section 2.2, for discussion of a cline of consciousness with regard to choices.) The VIA-TEXT features on the other hand would seem to relate to the second kind of choice. Does Hilary accept the maximum social distance that might be expected to follow from the fact that she has never met her publishers? Or does she try VIA the language she uses to reduce the social distance? 190 Society in Language, Language in Society

In the system networks later in the chapter, I assume that the PRE-TEXT systems relate to construal choices, while the VIA-TEXT systems relate to more active choices on the part of the interactants.

8.4.3 A point of origin for the networks A system network needs a point of origin, that is, an indication of when/how often its choices become available. (See Halliday, 2013, for discussion of point of origin.) In lexicogrammar a network may have the clause as its point of origin, indicating that the choices become available every time there is a new clause, or the nominal group, for example, for those choices that relate to the nominal group. In this chapter I have been using the term language event. I assume that a language event is a contextual unit, a contextual unit which centres on a text and which is realized by a text, text itself being a semantic unit. A language event can be characterized in terms of Field, Tenor and Mode, as I shall show below. I assume that a language event is an instance of a genre. In our present data set we have 17 language events: two conversations, six letters between Hilary and Ruth, six letters from Hilary to her publishers, one textbook and two PhD theses. I assume that a language event consists of stages, stage for me also being a contextual unit. (Both Hasan (e.g. 2009, p. 186) and Martin (e.g. 1992, p. 505) recognize genre as a staged phenomenon, although their approaches differ in other respects.) I would suggest that the language event is the most suitable point of origin for the networks I present in this chapter. The choices in the networks become available every time there is a new language event. There will be no room in this chapter to present networks that relate to the smaller unit, the stage.

8.4.4 The hypothetical nature of system networks It is important to recognize that a system network is always hypothetical (see Berry, 2013b, pp. 379–80). The networks here represent our hypotheses about the contextual features which Hilary and Ruth have construed as relevant to the language events and which have accordingly inﬂuenced their semantic choices. We shall not know whether or not our hypotheses are supported until we have carried out the linguistic analysis in stage 3 of our project. In order to test hypotheses about relationships between contextual features and semantic features, it is essential that the contextual features be established independently of the semantic features, and vice versa. This is an advantage of the PRE-TEXT features, which can be established without even looking at the texts. The VIA-TEXT features are slightly more problematic. However, I have based the VIA-TEXT features proposed here on the informant judgements of Hilary and Ruth rather than on text analysis. Margaret Berry 191

8.5 The contexts of our project

I will now apply all this to Hilary’s contexts, under the traditional SFL headings of Mode, Tenor and Field, beginning in each case with the PRE-TEXT features, and then moving to the VIA-TEXT features. I shall be concentrating on those features which distinguish between our contexts, rather than trying to include everything which is at all relevant. As I have indicated, we are eventually hoping to match differences in semantic choices with differences in contexts. Some of the contextual features most discussed in SFL do not in fact help us to distinguish between our contexts. For instance, there is no material action in any of our language events. In all cases the language is fully constitutive of the relevant social process. So although I am noting this here, I am not including the feature ‘constitutive’ in my network, or any of the associated features. They do not distinguish, because our language events are all the same from these points of view.7

8.5.1 Mode Hasan (1999, p. 232) glosses Mode as ‘the nature of contact’. (Her glosses seem to me to apply equally well here even though unlike her I am concerned with PRE-TEXT features.) Probably the most basic distinction under this heading is that between ‘spoken’ (in the sense that contact is to be by voice) and ‘written’ (in that contact is to be by marks on the page/computer screen, etc.) (see Figure 8.1). This distinguishes our Phone context and our Face-to-face context from all our other contexts.8

Co-present

Not co-present Spoken Spontaneous

Planned

Handwritten LANGUAGE EVENT Written Typewriter-written Computer-written

Immediate Interactive Delayed Not interactive

Figure 8.1 Language event: Mode systems 192 Society in Language, Language in Society

The Phone context and the Face-to-face contexts themselves differ in two ways: in the obvious way that in the Face-to-face context the interactants are ‘co-present’ but that in the Phone context the interactants are ‘not co- present’; and also in that the Phone conversation was spontaneous, but that the Face-to-face conversation was planned. The Phone conversation was accidentally recorded, but Hilary set up the Face-to-face conversation in order to be able to compare it with the Phone conversation. (The Face- to-face conversation is the only one of our texts that is not totally naturally occurring. All the other texts are naturally occurring in the sense that they would have occurred even if no one had been going to analyse them, and at the time they occurred, there was no thought of their being used as data for analysis.) As far as the written contexts are concerned, Hilary has noted whether the writing was by hand, by typewriter or by computer (Hillier, 2013). A quick look at the texts suggests that we are unlikely to ﬁnd semantic differences which correlate with these contextual distinctions. However, we will not know for certain until we have carried out systematic analysis. I have also included in the Mode network the contextual features ‘interactive’ and ‘not interactive’. Hilary has drawn attention to the two-way shared process nature of the spoken contexts. Hasan includes the features ‘active process sharing’ and ‘no process sharing’ under Mode in her tables (e.g. 1999, p. 244). I would like to suggest that the letters are also potentially interactive. One does not usually write a letter unless one is either replying to a letter or expecting a reply. (Not a letter to a speciﬁc person, that is, as opposed for instance to a letter to a newspaper.) However, the letters differ from the conversations in that responses are ‘delayed’ rather than ‘immediate’. (Halliday once said that the whole of linguistics was about saying that things were alike but different.9 Maybe a slight exaggeration, but certainly system networks enable one to show that things are alike but different.) The book and the theses I would regard as ‘not interactive’. It is of course possible that such texts may generate correspondence, but there is no inherent expectation of this at the outset of the writing. I have not found any examples of VIA-TEXT features under the heading of Mode. Ruth, in the Face-to-face conversation, evidently feels that there is too much monologue in proportion to dialogue and she comments on the fact. But she does not bring about any real change by this. The pattern continues as before.

8.5.2 Tenor Hasan (1999, p. 232) glosses Tenor as ‘the nature of social relation’. Bowcher (1999, p. 151), citing Hasan (1981), regards social relation as consisting of social distance and degree of hierarchy. She includes a Hasan network for Tenor, which is basically divided into these two sections (Bowcher, 1999, p. 153). My network follows this division (see Figure 8.2). 193

Secondary addressees

No secondary addressees

Frequent interaction

Infrequent interaction One-to-one Wide range of interaction

Narrow range of interaction

Accept social distance

SOCIAL Reduce DISTANCE Change social distance Increase

One-to-many

LANGUAGE EVENT

EQUAL GENERAL KNOWLEDGE Speaker/writer = PRIMARY KNOWER

NOT EQUAL Speaker/writer = SECONDARY KNOWER

Speaker 1 = primary knower

Speaker 2 = primary knower Variable Continue SPECIFIC KNOWLEDGE Change

Speaker/writer = primary knower Not variable Speaker/writer = secondary knower

Figure 8.2 Language event: Tenor systems 194 Society in Language, Language in Society

Social distance Bowcher (1999, p. 151), again following Hasan, says that social distance ‘has to do with the range and frequency of contact between interactants’. However, before discussing these matters, I suggest we need to consider something else. The first system in the social distance section of my Tenor network is that between ‘one-to-one’ relations and ‘one-to-many’ relations. The conversations and the letters are ‘one-to-one’, while the book and the theses are ‘one-to-many’. It does not seem relevant to talk about range and frequency of contact between Hilary/Ruth and the readers of their book/theses, many of whom they will not even have heard of. Hasan (1999, p. 238) calls such readers ‘virtual addressees’. So we need to distinguish between ‘one-to-one’ and ‘one-to-many’ before we consider range and frequency of contact. I have accordingly made ‘one-to-one’ an entry condition for the other social distance systems. There is one other thing to discuss before range and frequency. There is a slight complication over the ‘one-to-one’/‘one-to-many’ distinction. In one of Ruth’s letters to Hilary, where she is giving Hilary permission to use extracts from their conversations in her book, although she addresses the letter to Hilary, she is clearly expecting the letter to be read also by anyone else who needs to know about the permission. I have catered for this by including a system between ‘secondary addressees’ and ‘no secondary addressees’. Now we do come to range and frequency of contact. I have put in two systems, one between ‘frequent interaction’ and ‘infrequent interaction’, the other between ‘wide range of interaction’ and ‘narrow range of interaction’. Both of these distinguish between the conversations and letters between Hilary and Ruth on the one hand and the letters from Hilary to her publishers on the other. Hilary and Ruth have interacted frequently, not in the sense of interacting every day, but in the sense that they have engaged in many interactions over a long time. This compares with Hilary and her publishers’ representatives, who have not interacted frequently. The publishers’ representatives have in fact kept changing, so after a few interactions Hilary has found herself interacting with someone else. The range of contact has also been greater between Hilary and Ruth than between Hilary and her publishers. Hilary and Ruth have met face to face, telephoned each other and exchanged letters. Letters were the only form of communication between Hilary and her publishers. She has never met the publishers’ representatives to whom she is writing in the letters of our data, nor has she had telephone conversations with them. These are the facts of the case before the texts begin. They should add up to the social distance between Hilary and Ruth being at near minimum, and the social distance between Hilary and her publishers being at near maximum. However, the interactants do not have to accept the degree of social Margaret Berry 195 distance indicated by the PRE-TEXT facts. Hilary felt that she was writing in a friendly and informal manner to her publishers. On the other hand, Ruth, in a covering note with the permission letter mentioned above, indicated that the permission letter was more formally written than was normal between herself and Hilary. We are now talking about VIA-TEXT contextual features. Hilary and Ruth through the language they use are able to vary the degree of social distance from that of the PRE-TEXT facts. I have included systems in my Tenor network to cater for this. The features in these systems are in italics, to distinguish them from the PRE-TEXT features. First there is a system offering a choice between ‘accept social distance’ and ‘change social distance’ and then, if ‘change social distance’ is chosen, there is a further choice between ‘reduce social distance’ (as when Hilary decides to write informally to her publishers) and ‘increase social distance’ (as when Ruth writes more formally than usual to Hilary in the permission letter). I said above that, if we are to explain the linguistic choices that Hilary makes, we need to take into account both the PRE-TEXT features and the VIA-TEXT features. For instance, in her letters to her publishers, Hilary writes such things as ‘You’re right’ and ‘I’m a bit puzzled’. In a letter to a newspaper recently, she wrote ‘I am particularly concerned’. Presumably the contractions in the letters to the publishers are an indication that through her language, i.e. VIA her TEXT, she is reducing the social distance to a degree lower than would have been predicted by the PRE-TEXT facts. How- ever, we do not find in the letters to the publishers some of the linguistic features of the letters between Hilary and Ruth – ‘Must dash’, ‘Promise!’, ‘Hope to see you soon’. (Exophoric ellipsis of the Subject.) In the letters to the publishers, Hilary has not reduced the degree of social distance to the level which is normal between herself and Ruth. She is evidently in part constrained by the PRE-TEXT features, though not completely bound by them. (I am jumping ahead a bit here. We have not yet carried out our detailed linguistic analysis. The linguistic examples in the last paragraph were culled from a quick reading of the letters. I have included them here simply to illustrate the kinds of realizations I would expect to find of the contextual features I am discussing. I must emphasize that my proposals for VIA-TEXT features were based on Hilary’s and Ruth’s own comments on what they were trying to achieve prior to any linguistic analysis – see section 8.4.4 above. The detailed linguistic analysis will hopefully show whether their intentions did systematically lead to the selection of particular linguistic options.)

Hierarchical relations What is meant by ‘hierarchical relations’? Hasan, in the network cited by Bowcher (1999, p. 153), distinguishes different types of hierarchical 196 Society in Language, Language in Society relationship, but her distinctions – e.g. male/female, elder/younger – are not relevant to our data. It would appear that different types of hierarchical relationship are relevant to different types of language event. The hierarchical relationships most relevant to our present data would seem to be knowledge relationships. In Berry (1987) I distinguished two types of knowledge relation: general knowledge relations and specific knowledge relations. I was discussing three particular hierarchical relationships: Teacher–Student, Doctor–Patient, and Committee Chair–Committee Mem- ber. The three are alike in that in each case one of the interactants is assumed to be of higher status than the other(s). But I suggested that they were different in the knowledge relations that existed between the members of the pairs. (Another case of things being alike but different – see above.) This is where the distinction between general knowledge relations and specific knowledge relations became relevant. The distinction is perhaps easiest to see in relation to doctor–patient interviews. Presumably the doctor can be regarded as generally more medically knowledgeable than the patient. So I called the doctor the PRIMARY KNOWER (capitals) and the patient the SECONDARY KNOWER (capitals). But in doctor–patient interviews usually quite a lot of time is spent discussing the patient’s symptoms. And who knows most about the patient’s symptoms? The patient surely. The doctor may be the more generally knowledgeable, but the patient knows most about the specific details of the information under discussion. I called the patient ‘primary knower’ (lower case) and the doctor ‘secondary knower’ (lower case). The doctor is PRIMARY KNOWER, being generally more knowledgeable, but the patient is primary knower (lower case), knowing more about the specific details of the information being exchanged.10 A teacher will normally be both PRIMARY KNOWER and primary knower (though see Berry, 1987, p. 60). A committee chair will not necessarily be either. These distinctions are relevant to both our academic language events and our conversations. They are helpful in discussing Hilary’s book and Hilary’s thesis. To take the lower-case relationship first, Hilary’s book and Hilary’s thesis are alike in this. In both, Hilary is writing about her own investigations and the results of her own investigations. She certainly knows most about the specific details of the information. So she is lower-case primary knower. But what about the capitals relationship? In Berry et al. (2014), Hilary describes the intended audience of her book as ‘students’, but the intended audience of her thesis as ‘a PhD examining board’. So in her book she is presumably adopting the role of teacher with its associated role of PRIMARY KNOWER. But if the examining board are assumed to be generally more knowledgeable – that certainly seems to be the trend of Hilary’s remarks – then for the thesis Hilary has cast herself in the role of SEC- ONDARY KNOWER, with her audience as PRIMARY KNOWERS. For the book she is primary knower + PRIMARY KNOWER, but for the thesis she is primary knower + SECONDARY KNOWER. It will be interesting when we get to Margaret Berry 197 the detailed linguistic analysis to see if this has made any difference to the semantic options she has chosen. The general knowledge relationships in our data remain constant throughout the language events. The general knowledge relationships that hold at the outset of a language event remain through the event. For this reason I am regarding them as PRE-TEXT features. This is sometimes the case with specific knowledge relationships, but not always. Certainly Hilary is primary knower throughout her book and her thesis. But in the conversations, and to some extent the letters, the specific knowledge relationships change. For instance, in the Phone conversation, the early part of the conversation is all about Ruth’s job, Ruth’s travel arrangements, Ruth’s future plans, so all through this stretch of conversation Ruth is the primary knower and Hilary the secondary knower. But part way through, the topics shift to matters relevant to Hilary, so she becomes primary knower with Ruth as secondary knower. Again, at risk of anticipat- ing what we are going to be looking at in our detailed linguistic analysis, this shows in such things as which of them asks questions, which introduces tag questions, which makes statements that are heard as requests for confirmation. What brings about such changes? Again, pending our detailed linguistic analysis, the changes do seem to be brought about by things the interactants do with their language. For instance, the change just referred to in the Phone conversation seems to be brought about by Ruth asking Hilary ‘How’s your research going?’ Ruth, as a good conversationalist, recognizes that the talk has been mainly about her affairs and switches the conversation to things of relevance to Hilary. After a while on ‘Hilary topics’, Hilary asks Ruth a question which returns the primary knower role to Ruth. Since the changes in specific knowledge relationships seem to be brought about by things the interactants do with their language, I am regarding these as VIA-TEXT features. How is all this represented in my Tenor network? The first system under the General Knowledge heading is between ‘EQUAL’ and ‘NOT EQUAL’. I am assuming that in all the interactions between Hilary and Ruth – the conversations and the letters – Hilary and Ruth can be regarded as equally knowledgeable from a general knowledgeability point of view. They first met when they were both research students working on their theses, and the friendship has developed along those lines. As I have already indicated, for both the book and the theses, I would regard the general knowledge relationships between writer and readers to be ‘NOT EQUAL’, but ‘NOT EQUAL’in different ways. In the book the writer is PRIMARY KNOWER, but in the theses the writer is SECONDARY KNOWER. The first system in my Tenor network under the Specific Knowledge heading is the PRE-TEXT opposition of ‘variable’ and ‘not variable’. The conversations are variable from this point of view, the book and the theses not variable. Then ‘variable’ leads to a VIA-TEXT choice of which speaker is to be 198 Society in Language, Language in Society primary knower and so this is shown in italics. This system has an iterative option attached to it as the interchange of roles can occur more than once in a language event. Most of the distinctive features I have found under Tenor have been PRE- TEXT features, but there does seem to be some evidence in favour of Hasan’s view that interactants can through their language create, or at least change, contexts.

8.5.3 Field Hasan (1999, p. 232) glosses Field as ‘the nature of social activity’. Halliday (1978, pp. 142–3) writes of ‘the social action: that which is “going on” ...including “subject-matter” as one special aspect’. My proposals for distinctive features under the headings of Mode and Tenor were tentative. My proposals for distinctive features under Field are going to be extremely tentative. At least under Mode and Tenor it was possible to arrive at some measure of intersubjective agreement between researchers. And at least it was possible to arrive at distinctive features which looked as if they might have some relevance beyond the bounds of our particular project. But in Field there can be as many interpretations of what is ‘going on’ as there are researchers. And each time one turns to a new set of data, one ﬁnds oneself wanting to say something different about the relevant features of Field. However, I will plunge in and make some proposals, if only to start a discussion.

Purpose of language event From a look at the Bowcher (2014) unified network for Field, in which she draws on work by Hasan and Butt as well as her own, it would seem that the one of her systems most likely to be relevant to our data is that between ‘specialized’ and ‘quotidian’ (i.e. everyday). However, even that one is difficult to apply to our language events. One might expect that conversations between friends would be classifiable as ‘everyday’. But Hilary and Ruth, in both their conversations, are talking about their work. Does that make the conversations ‘everyday’ or ‘specialized’? Bowcher herself warns that analyses in terms of her unified network will ‘not [be] very revealing about the situation’ as they are at the least delicate end of the system networks. In Berry et al. (2014), Hilary proposed a form of classification which seems to be related to the ‘everyday’/‘specialized’ system, but which does seem to fit our data. She classified the letters between herself and Ruth as ‘personal’, ‘professional’ or ‘semi-professional’. This related to the purpose of the letters. A letter could have a personal purpose, or a professional purpose, or a mixture of personal and professional purposes. This received intersubjective agreement at least to the extent that the other two researchers were happy to Margaret Berry 199 work with the scheme and agreed with the way in which Hilary had applied it to the letters. Hilary has now extended the classification to cover all our language events. The Phone conversation has a personal purpose, the Face-to-face conversation has both a personal purpose and a professional purpose, the letters to publishers have a professional purpose, as have the book and the theses.

The nature of the social activity What happens if we try to answer Halliday’s question ‘What is going on?’ in our language events? What is the nature of the social activity? As for myself I get two distinct sets of answers. One of these has to do with what one might think of as the content of the activity; the other has to do with what the activity has to do with the relation between the interactants. The former to some extent resembles Hasan’s ‘reﬂection-based’ activity, the latter to some extent resembles Hasan’s ‘relation-based’ activity (Hasan, 1999, pp. 288–92). However, my categories are not exactly the same as Hasan’s. For one thing she puts hers together in a system, which means that they are in an either–or relation to each other. My categories are in a both–and relation. I am saying that each of our language events has an activity under each of my two headings. Since my categories are not exactly the same as Hasan’s, I had better give them different names. I call them ‘content-based activity’ and ‘people-based activity’. (I am carefully avoiding using terms such as ideational, experiential and interpersonal, as I do not want to confuse the context with the semantics.) Table 8.1 shows how I would assign activities under my two headings to each of our language events. The one language event in our data that does not ﬁt easily into Table 8.1 is the letter in which Ruth gives Hilary permission to use extracts from their

Table 8.1 Categories of activities in the language events

Language event Content-based activity People-based activity

Phone conversation Exchanging news and Maintaining social relationship views Face-to-face Exchanging news and Maintaining social relationship conversation views Personal letters Exchanging news and Maintaining social relationship views Semi-professional Negotiating permission Maintaining social relationship letters Professional letters Negotiating contract, Establishing social relationship copyright, etc. Book Informing Introducing people to ﬁeld Theses Informing Demonstrating to examiners competence in ﬁeld 200 Society in Language, Language in Society conversations in her book. It is part of the sequence of letters negotiating the permission, but it is not really concerned either with maintaining the relationship with Hilary or with establishing a relationship with the secondary addressees. The other language event that needs special mention is the Face- to-face conversation, which has a third kind of activity in addition to the two in Table 8.1: that of recording the conversation for later analysis and discussion. (I do not feel obliged to make room for this in Table 8.1 as it is part of the planned rather than naturally occurring nature of this language event.) I consulted Hilary about this table (Table 8.1) and she raised only one objection. Originally I had described the content-based activity of the book and the theses as ‘reporting’. Hilary felt that this carried connotations of a particular kind of language event, so I changed to the more neutral ‘informing’.

Towards a system network for Field Assuming that all this does at least begin to characterize the Fields of our language events, how do I represent this in a system network? The MK1 version here (Figure 8.3) does not get much further than the tables, simply listing the possibilities. The MK2 version (Figure 8.4) attempts a certain amount of factorization, in such a way as to continue the ‘alike but different’ motif. For instance, the maintaining relationship and establishing relationship activities seem to have something in common. And the introducing to field and demonstrating competence in the field activities seem to have something in common (field here in its normal academic sense). So I have grouped the first pair under ‘social activity’ and the second pair under ‘academic activity’. But there is a difference within each pair, so I have given each pair its own system to reflect the differences. Similarly exchanging news and views seems in some way like informing – both have to do with the transfer of knowledge/opinion. These seem very different from negotiating, which expects some action to follow. (The action is not material action, but it may be worth looking again at the SFL literature on material action to see if any analogies suggest themselves.) In my 1981 papers (Berry, 1981a, b) on exchange structure, I distinguished between exchanges which were about knowledge and exchanges which were negotiating action. Maybe a similar distinction is needed on a larger scale to apply to whole language events. In the network, I have initially grouped ‘exchanging news and views’ with ‘informing’ under ‘transferring knowledge’, to distinguish them from ‘negotiating action’. However, they then need to be differentiated themselves: ‘exchanging’ is two-way knowledge transfer, while ‘informing’ is only one-way. The Purpose section of the network, in the MK2 version, is represented as two simultaneous systems (a more elegant way of doing it!). This allows Margaret Berry 201

Personal PURPOSE Semi-professional

Professional

CONTENT- Exchanging news and views BASED ACTIVITY Negotiating

LANGUAGE EVENT Informing

Maintaining social relation

PEOPLE- Establishing social relation BASED ACTIVITY Introducing people to field

Demonstrating own competence in field

Figure 8.3 Language event: Field systems MK1

‘personal purpose’ to occur on its own, ‘professional purpose’ to occur on its own, or ‘personal purpose’ and ‘professional purpose’ to occur together. It also allows for language events which have neither a personal nor a professional purpose. There are no such language events among our data, but it is not difﬁcult to think of such language events elsewhere. For instance, one would not want to say that a buying stamps language event (e.g. Hasan, 1999, pp. 226–7) has a personal purpose or that it has a professional purpose, at least not in the sense that these terms are being used here.

PRE-TEXT or VIA-TEXT features? All the contextual features I have so far discussed under Field remain, in our data, constant throughout their respective language events. Indeed, they seem to be inherent in the particular types of language event from the outset. For this reason, I am regarding them as PRE-TEXT features. However, there are things that change. For instance, in the book and the theses, the activity of informing seems to have sub-activities and, as each work progresses, it moves from one sub-activity to another. Hilary has selected three extracts from each thesis for us to analyse in detail. In each 202 Society in Language, Language in Society

+ personal purpose − personal purpose PURPOSE + professional purpose − professional purpose

Exchanging CONTENT- Transferring knowledge BASED Informing ACTIVITY

LANGUAGE Negotiating action EVENT

Maintaining relationship Social activity PEOPLE- Establishing relationship BASED ACTIVITY Introducing people to field

Academic activity Demonstrating own competence in field

Figure 8.4 Language event: Field systems MK2

case, the ﬁrst extract is a general description of the data, the second describes the participants in the data, the third presents conclusions. She has selected two passages from the book, one of which describes the data and the other of which contributes to the presentation of conclusions. Are these sub-activities created by the language? Hilary’s authorial view is No – she had the sections of her book and thesis in mind from the beginning. Headings and subheadings can signal moves from one sub- activity to another, and the content of each section can realize its particular sub-activity, but there is a difference between realizing something and creating it. At this point we seem to have moved from discussing whole language events to discussing stages of language events. As I said earlier, there is not room in this chapter to fully discuss stages or the system networks that apply to them. Margaret Berry 203

8.6 Checkable criteria

Hasan (2009, p. 179) writes of the vagueness of descriptions of context, ‘the absence of “checkable” criteria, and the reliance on “common sense” ’. In order to remedy this situation, we need ‘contextualization system networks’ which ‘actually systemize the realization-activating contextual features and attempt to relate context to wording via meaning which acts as the interface between the two’ (Hasan, 2009, pp. 181–2). At present in our project we are still at the ‘common sense’ stage (albeit framing our distinctions in potentially theoretical terms). However, as I have indicated, we plan to engage in detailed linguistic analysis, whereby we do hope to relate context to wording via meaning. If we can show that there is a correlation between the contextual features we have identiﬁed and the meaning options Hilary and Ruth have chosen, these being realized by lexicogrammatical features, this will be a check on our contextual features. If we cannot ﬁnd such a correlation, we shall need to think again. We shall certainly be considering both the PRE-TEXT and VIA-TEXT features that I have been discussing in this chapter.

Notes

1. I am grateful to the following for discussing with me points relevant to this chapter: Hilary Hillier, Geoff Thompson, Tom Bartlett, Chris Butler, Ruqaiya Hasan, Sarah Mukherjee and Jeff Wilkinson; also to the editors of this volume, and to the organizers and members of the audience of the European Systemic Functional Linguistics Conference and Workshop, Coventry, July 2013, where an early version of this chapter was presented. Of course I alone am responsible for any errors or misrepresentations. 2. I am of course very grateful to Hilary and Ruth for providing the data, and to Palgrave Macmillan for allowing us to include in the study extracts from Hillier (2004). 3. It will also be important to consider the work of Martin (e.g. 1992, ch. 7). 4. Bartlett (2013) also uses the term ‘context potential’, but I think he is using it in a slightly different sense, a sense in which it is not precisely analogous to the term meaning potential. As I understand him, he sees the environment of each language event as constituting a context potential, rather than an account of context potential being an account of all the contextual possibilities, the contextual resources, of a whole language or of a language community. 5. I am using the term ‘interactant’ where Hillier (2013) uses the term ‘participant’. I avoid the term ‘participant’ in this sense as the word is a technical term in transitivity. If one is not careful in SFL, one can ﬁnd oneself using the term in two different senses in the same paragraph. 6. What happened was slightly more complicated than this suggests, in that initially Hilary’s son answered the phone, so there was a slight delay between the phone ringing and Hilary herself picking up the receiver. 7. There is in the middle of the Face-to-face conversation some activity whereby the recording tape is turned over, but it would seem perverse to regard this as part of the main conversation rather than as an interruption to the main conversation. 204 Society in Language, Language in Society

8. In Berry (2013a, b), I drew further distinctions under the headings of MEDIUM and CHANNEL, but these do not act as distinctive features for our present set of contexts, so again I am not including them here. 9. In a talk to the Nottingham Linguistic Circle many years ago. 10. In Berry (1981a) I related the primary knower/secondary knower distinction to exchange structure. In Berry (1981b), I also discussed primary actor and secondary actor.

References

Bartlett, T. (2012) Hybrid Voices and Collaborative Change: Contextualising Positive Discourse Analysis (London and New York: Routledge). Bartlett, T. (2013) ‘ “I’ll manage the context”: context, environment and the potential for institutional change’. In L. Fontaine, T. Bartlett and G. O’Grady (eds) Systemic Functional Linguistics: Exploring Choice (Cambridge: Cambridge University Press), pp. 346–64. Berry, M. (1981a) ‘Systemic linguistics and discourse analysis: a multi-layered approach to exchange structure’. In M. Coulthard and M. Montgomery (eds) Studies in Discourse Analysis (London: Routledge & Kegan Paul), pp. 120–45. Berry, M. (1981b) ‘Towards layers of exchange structure for directive exchanges’. Network 2: 23–32. Berry, M. (1987) ‘Is teacher an unanalysed concept?’ In M. A. K. Halliday and R. Fawcett (eds) New Developments in Systemic Linguistics,Vol.1:Theory and Descrip- tion (London: Frances Pinter), pp. 41–63. Berry, M. (2013a) ‘Contentful and contentlight subject themes in informal spoken English and formal written English’. In G. O’ Grady, T. Bartlett and L. Fontaine (eds) Choice in Language: Applications in Text Analysis (London: Equinox), pp. 243–68. Berry, M. (2013b) ‘Towards a study of the differences between formal written English and informal spoken English’. In L. Fontaine, T. Bartlett and G. O’Grady (eds) Systemic Functional Linguistics: Exploring Choice (Cambridge: Cambridge University Press), pp. 365–83. Berry, M., Thompson, G. and Hillier, H. (2014) ‘Theme and variations’. In M. de los Ángeles Gómez González, Francisco José Ruiz de Mendoza Ibáñez and F. Gonzálvez García (eds) Theory and Practice in Functional-Cognitive Space (Amsterdam and Philadelphia: John Benjamins), pp. 107–26. Bowcher, W. L. (1999) ‘Investigating institutionalisation in context’. In M. Ghadessy (ed.) Text and Context in Functional Linguistics (Amsterdam and Philadelphia: John Benjamins), pp. 141–76. Bowcher, W. L. (2007) ‘Field and multimodal texts’. In R. Hasan, C. M. I. M. Matthiessen and J. J. Webster (eds) Continuing Discourse on Language: a Functional Perspective (Vol. 2) (London: Equinox), pp. 619–46. Bowcher, W. L. (2014) ‘Issues in developing uniﬁed systems for contextual ﬁeld and mode’. Functions of Language 21 (2): 176–209. Cloran, C. (1999) ‘Context, material situation and text’. In M. Ghadessy (ed.) Text and Context in Functional Linguistics (Amsterdam and Philadelphia: John Benjamins), pp. 177–217. Fontaine, L., Bartlett, T. and O’Grady, G. (eds) (2013) Systemic Functional Linguistics: Exploring Choice (Cambridge: Cambridge University Press). Halliday, M. A. K. (1978) Language as Social Semiotic: the Social Interpretation of Language and Meaning (London: Edward Arnold). Margaret Berry 205

Halliday, M. A. K. (1985/1989) ‘Context of situation’. In M. A. K. Halliday and R. Hasan Language, Context and Text: Aspects of Language in a Social-Semiotic Perspective (Deakin: Deakin University Press. Reprinted 1989 by Oxford University Press), pp. 3–14. Halliday, M. A. K. (2013) ‘Meaning as choice’. In L. Fontaine, T. Bartlett and G. O’Grady (eds) Systemic Functional Linguistics: Exploring Choice (Cambridge: Cambridge University Press), pp. 15–36. Hasan, R. (1981) ‘What’s going on: a dynamic view of context in language’. In J. E. Copeland and P. W. Davis (eds) The Seventh LACUS Forum 1980 (Columbia, SC: Hornbeam), pp. 106–21. Hasan, R. (1999) ‘Speaking with reference to context’. In M. Ghadessy (ed.) Text and Context in Functional Linguistics (Amsterdam: Benjamins), pp. 219–328. Hasan, R. (2009) ‘The place of context in a systemic functional model’. In M. A. K. Halliday and J. J. Webster (eds) Continuum Companion to Systemic Functional Linguis- tics (London: Continuum), pp. 166–89. Hillier, H. (2004) Analysing Real Texts: Research Studies in Modern English Language (Basingstoke and New York: Palgrave Macmillan). Hillier, H. (2013) ‘On describing contexts of situation: a participant’s view’, paper presented to the European Systemic Functional Linguistics Conference and Workshop, Coventry University, UK, July 2013. Extended abstract and tables to appear in the online proceedings of the conference. Martin, J. R. (1992) English Text: System and Structure (Amsterdam and Philadelphia: John Benjamins). 9 Interfacing Field with Tenor: Hasan’s Notion of Personal Distance Marvin Lam

9.1 Hasan’s notion of personal distance

In her discussion on register, Hasan (1973, pp. 271ff.) explains that register is ‘characterized by reference to some syntactic, lexical or phonological patterns’, and such characterization is motivated by the following five factors: subject matter of the discourse, situation type of the discourse, participant roles within discourse, mode of discourse and medium of discourse. The discussion then elaborates the relations between these factors and certain linguistic patterns. On the factor of participant roles within discourse, Hasan (1973, p. 277) points out that individuals can be related by several ‘socially defined positional roles’ but these role relations are not equally relevant to a particular social activity, emphasizing the necessity of qualifying the term ‘participant roles’ as ‘participant roles within discourse’. The different degrees of relevancy of role relations to a social activity are likely to be reflected by certain linguistic patterns emerging from the language used by the individuals in that social activity, hence exhibiting registerial variation. Hasan provides an example of a social activity of classroom interaction in which the teacher happens to be the mother of a student. The role relation that is more relevant would be teacher–student instead of mother– daughter, which would probably be reflected in certain linguistic patterns of the language use of the teacher/mother and the student/daughter. In the same discussion, Hasan also emphasizes that there is an attribute of the participants of a discourse subsumed in their role relations, namely ‘personal distance’ (Hasan, 1973, 1978, 1981). Personal distance underlines the distinction between formality–familiarity and ‘in any given instance the personal distance factor may override the consequences of the publicly recognised roles’ (Hasan, 1973, p. 277). Hasan provides an example of a father and a son interacting differently in different situation types with different degrees of personal distance. With the possibility of having personal distance overriding the publicly recognized role relation(s), in the example

206 Marvin Lam 207 of the classroom interaction mentioned above, the factor of personal distance resulting from being a mother and daughter may ‘interfere’ with the more institutionally recognized role relation of teacher–student in terms of the linguistic patterns emerging from their language use. This relation between personal distance (together with participant roles) and language use in terms of lexicogrammar is further exemplified in the discussion (Hasan, 1973, pp. 278–9). Reduced personal distance can be related to the use of exophorics as there is less need for disambiguation. Hasan also suggests a correlation between greater personal distance and the higher likelihood of having the semantic components of tentativeness and uncertainty, which in turn are mostly realized by various lexicogrammatical resources of assessments (such as modality and modal assessments as described in Halliday and Matthiessen, 2014, pp. 176ff.). Apart from English, Hasan (1996, pp. 229ff.) offers a description of the restricted exophoric ellipsis in Urdu, a Central Indo-Aryan language, and relates the extent of ellipsis to the implicitness of language use and furthermore to personal relations. This shows that such phenomena are present in different languages albeit manifest in ways that are culturally tied. Relevant to Hasan’s notion of personal distance and its relation with language use, there is a rich body of work in linguistics of similar interests with multi-origins and references to sociology. This multidisciplinary nature will be further discussed in section 9.4 of this chapter. Semin (2007) situ- ates linguistic studies on personal distance within relationship management, stemming from Goffman’s (1955) discussion on facework and further illustrated by studies such as those on politeness (Brown and Levinson, 1987). Some of the many questions that Semin raises include whether there are ‘linguistic strategies that are implicitly used in the service of regulating social [personal] distance’ and whether it is ‘possible that the way we use language has slippages conveying distance or proximity implicitly’ (Semin, 2007, p. 390). Hasan’s notion of personal distance can be regarded as an affirmative response to Semin’s questions. Numerous scholars have detailed the relations between certain linguistic features and the regulation of personal distance. For instance, Brown and Gilman (1960) distinguish familiar from polite forms of personal pronouns in French, Turkish and German, which can be respectively related to the ‘shortening’ and ‘lengthening’ of personal distance. In speech accommodation theory (Giles, 1973), ‘convergence’, that is, the use of linguistic strategies by individuals to adapt to each other’s speech in terms of speech rates, pauses and utterance length, pronunciation, and so on, which leads to traits of positive feelings, such as friendliness and warmth (Coupland, 1985), can be regarded as ‘shortening’ personal distance; while ‘divergence’ can be considered as personal distancing. Rubini and Kruglanski (1997) investigate the relation between different predicate types – adopting Semin and Fiedler’s (1988) linguistic category model – and the perception of 208 Society in Language, Language in Society interpersonal rapport in question–answer situations. Questions with more concrete predicate types (such as descriptive action verbs) are said to shorten personal distance, as they are believed to trigger subjective thinking in the answering; while questions of more abstract type (such as state verbs) are associated with personal distancing because they are prone to logical objective focus. Edwards (1978) argues that teachers and students do not have equal access to the negotiation of personal distance. This view stems from Hymes’ (1977) argument that the nature and outcome of a relationship are conditioned by the means of communication available for that relationship in the context. In a classroom context, the conventions and setting provide teachers with more resources in regulating personal distance than the students, reflected by the larger variety of teachers’ language use than the students’ in shifting between friendly and authoritative tones and between issuing direct and indirect commands, and teachers’ authority in announcing different stages of the lesson. These differences in the access of resources to regulate personal distance connote a distanced relation. These studies reveal certain properties of linguistic work on personal distance: that the primary focus is on language – which is a natural choice of focus in the discipline – to identify relevant linguistic behaviours. This may account for the fact that in much of the relevant linguistic work, ‘personal distance’ is argued to be less vigorously defined, as it is variously defined as ‘traits of positive feelings, such as friendliness and warmth’, ‘intimacy, solidarity and equality’, or ‘feeling of friendliness’. In order to obtain a comprehensive account of the relations between personal distance and language use, a socio-semiotic approach to linguistic studies is necessary as the understanding of personal distance and language behaviours are intrinsic to each other. The better understanding of the personal relationship among, for example, the teacher/mother and student/daughter mentioned earlier can help understanding of their language behaviours in the situation, which in turn enhances our understanding of how personal distance operates in the situation. With this consideration, a socio-semiotic account of language, such as that of systemic functional linguistics (Halliday, 1975), is necessary and this socio-semiotic approach makes Hasan’s notion of personal distance unique from other linguistic work on personal distance as outlined above. Her notion offers a brief but comprehensive account on how personal distance can be understood with reference to participant roles and in relation to different uses of language. Furthermore, Hasan highlights one important fact that, instead of being static, personal distance is dynamically ‘determined by reference to the range and frequency of their [individuals’] interaction’ (1981, pp. 47–8). This dynamic nature of personal distance is further elaborated in Hasan’s discussion on the relations between contextual configuration and text structure, specifically on the tenor of discourse (Halliday and Hasan, 1985, pp. 56–7). Marvin Lam 209

Personal distance is related to what she calls a ‘biography of interactions’, and such distance is considered as a continuum with two end points being ‘maximal’ and ‘minimal’. A thin biography of interactions entails maximal personal distance, as ‘the persons involved know each other through infrequent encounters only in the capacity of the agent of some one insti- tutionalised activity and in the dyadic status that correlates with the agent role’ (Halliday and Hasan, 1985, p. 57); and minimal personal distance is obtained with a rich biography of interactions. Hasan provides an example of a day tourist buying some fruit from a vendor for the first time. The personal distance is likely to be maximal, as the biography of interactions only involves this single interaction of buying/selling fruit, so the tourist only knows the vendor by the institutional role of a vendor. In contrast, if the shopper has been buying fruit from the vendor over a number of years, or knows the vendor in some other capacities, such as belonging to the same club, as neighbours or relatives, their biographies would be rich in terms of range and frequency of interactions and their personal distance therefore minimal. The construction of biography in terms of range and frequency of interactions makes Hasan’s notion of personal distance complete, as such construction can be a means to measure personal distance and track its development between individuals. This raises an interesting issue concerning the three parameters of context, namely ‘field’, ‘tenor’ and ‘mode’ (Halliday and Hasan, 1985). Personal distance is clearly tenor-related. This is reflected by the fact that Hasan’s discussion on personal distance mentioned above (Halliday and Hasan, 1985, p. 57) is situated in the discussion on tenor of discourse. Besides, personal distance is centrally concerned with familiarity (also called ‘contact’) as ‘the more minimal the social [personal] distance, the greater the degree of familiarity between the carriers of the roles’ (Halliday and Hasan, 1985, p. 57). And FAMILIARITY, together with INSTITUTIONAL ROLE, POWER, AFFECT, SPEECH ROLE and VALUATION, are the systems in the description of tenor as role relationships (Matthiessen et al., 2010, p. 217). However, the measurement of this tenor-related notion of personal distance through the construction of a biography of interactions is a field-related endeavour, as it involves the categorization and identification of social activities so as to recognize their range and frequencies. Hasan’s correlation between the tenor-related notion of personal distance and the field-related endeavour of constructing a biography of interactions shows that contextual parameters are not discrete from each other, thus reflecting the relations between these parameters and the fuzzy nature of boundaries in natural language (Halliday, 1995). The following section will therefore switch to a discussion on Hasan’s contribution in describing the field of discourse and its significance in measuring personal distance and the tracking of its development. 210 Society in Language, Language in Society

9.2 Hasan’s description of the ﬁeld of discourse

Hasan (1999, p. 311) discusses in detail ‘field of discourse’ and includes a FIELD system network, which is further modified in Hasan (2009). The system network described in Hasan (1999) is adopted in this chapter because Hasan (1999) provides more descriptive details of the network than Hasan (2009). As this volume pays tribute to Hasan’s contributions, it would seem straightforward to adopt her description of field for the construction of a biography of interactions to measure and track the development of personal distance. However, to faithfully credit her merits, one needs to address the uniqueness of her description of field of discourse in relation to descriptions with similar concerns. With this purpose, Butt’s (2003) description of field and the functional text typology proposed by Matthiessen et al. (2008) are included in the current discussion. Butt’s ‘Parameters of context’ (2003) offers one of the most extensive descriptions of context in SFL literature in the form of elaborate system networks of field, tenor and mode, which are derived from earlier work by Halliday and Hasan (e.g. Halliday, 1975; Hasan, 1979, 1999). Butt’s FIELD system network has four simultaneous systems at the first degree of delicacy, namely SPHERE OF ACTION, MATERIAL ACTION, ACTION WITH SYMBOLS and GOAL ORIENTATION. With these four simultaneous systems as the domains of contrast, there are up to seven degrees of delicacy. As stated in Butt (2003, pp. 1ff.), this extensive FIELD network is largely developed in relation to theories of behavioural modelling, which can be considered as an attempt to model complex human behaviour envisioned in Butt (2000a) and is largely based on semantics (Butt, 2000b). This focus on semantics is reflected in the description of features in this FIELD system network. For example, in the system SPHERE OF ACTION, there is a more delicate system SPECIALIZED with the choices [Natural] and [Irrealis], which are correspondingly glossed as ‘phenomena involved rely on sensory report or evidence’ and ‘unactualised but intelligible, that is, a “conceptualisation” only’ (Butt, 2003, p. 26). These features are concerned with the nature of meaning negotiated in the situational context, reflecting the purpose of developing this network as modelling complex behaviour semantically, and so are not intended to be explicitly related ‘down’ to the lexicogrammatical stratum of meaning. The functional text typology proposed in Matthiessen et al. (2008) and detailed in Halliday and Matthiessen (2014, pp. 33ff.) is a typology of text types according to systematic criteria and based on field and mode values of situational context in accordance with Jean Ure’s (n.d.) account. In terms of the field value, texts are typologized according to different socio-semiotic activities, that is, activities attaining social value or social order (that is, socio-) and meaning making (that is, semiotic). Socio-semiotic activities can be categorized as expounding, reporting, recreating, sharing, doing, recom- mending, enabling and exploring, which are interrelated and technically Marvin Lam 211 defined. Together with the mode values of spoken vs written and monologic vs dialogic, the text typology is graphically organized as a discursive ‘wheel’ with text types as shading into one another in a continuous, multi- dimensional space at different positions (according to field) and of different concentric rings (according to mode). This model provides a holistic description of possible fields of discourse in terms of socio-semiotic processes, providing a comprehensive grid for different text types to be related to each other in terms of different field and mode values. Adopting this typology would be a powerful way to understand how social activities (identified as socio-semiotic processes) are related to each other in the construction of the biography of interactions. For instance, the typology is adopted in measuring and tracking the development of personal distance between patients and healthcare practitioners through patients’ journeys through hospital emergency departments (Lam, in preparation). In relation to the descriptions of the field of discourse proposed in Butt (2003) and Matthiessen et al. (2008), the uniqueness of Hasan’s (1999) description is shown by the following example. In describing the feature [install] in the FIELD system network, Hasan emphasizes that ‘only certain categories of semantic and lexicogrammatical choices are relevant to the performance of such verbal action: for example, in a text of the type that instructs how to install some device, semantically the event will be of the doing type, which will be realized lexicogrammatically by material processes’ (Hasan, 1999, p. 286, emphasis in italics as in original, bold added). This description of the feature [install] reflects a general property of Hasan’s description of field: that system features are clearly associated with certain lexicogrammatical choices. Further, Hasan’s description of field is in the form of a system network, ensuring the categorization of field in different degrees of detail. With the concern of both broad and detailed categoriza- tions of field for the preferred description, the concept of ‘degree of delicacy’ (Halliday, 1961) makes system networks an ideal format of the description of field for the current purpose of constructing a biography of interactions: the less delicate features in the FIELD system network reflect a broad categorization of field in describing the range of social activities; and more delicate features reflect a more detailed categorization. In brief, the range of social activities can be reflected by the spread of the different features in the FIELD system network being realized, which are in turn associated with certain lexicogrammatical properties which emerge from the language use of interactants; and the frequencies of social activities can be reflected by how often features in the FIELD system network are realized. Regarding the realization of features in the FIELD system network, Hasan emphasizes that

we can only specify the general categories of meanings/wordings that the given contexts will activate, such as doing events and material processes: 212 Society in Language, Language in Society

it does not enable the identification of the specific domain of doing or the specific lexical taxonomies the members of which might be expected to realize those doings lexicogrammatically. (1999, p. 287, emphasis in original)

This argument is understood here as being that the lexicogrammatical features associated with features in the FIELD system network should be ‘general’, that is, referring to linguistic properties that are exhibited by, or relate to, clauses distributed across a substantial portion of the interaction instead of individual clauses, to ‘identify the specific lexicogrammatical [ ...] domains at risk in the realization of specific choices from the systems of field’, as an approach to better understand the ‘genre specific semantic potential’ (Hasan, 1999, p. 287) and ‘generic structure potential’ (Hasan, 1978) in lexicogrammatical terms. To briefly summarize the discussion so far, Hasan’s notion of personal distance (Hasan, 1973, 1978, 1981; Halliday and Hasan, 1985) is unique not only in terms of the socio-semiotic approach of studying language by relating participant roles with different uses of language, but also in providing a methodology of measuring and tracking the development of personal distance through lexicogrammatical analysis of the realization of features of the field of discourse as described in Hasan (1999). With the SFL account of language as a semiotic resource, Hasan’s notion of personal distance and its measurement can be ‘positioned’ by referring to the dimensions of stratification and metafunction in the overall ‘architecture of language’ (Halliday and Matthiessen, 2014, ch. 1). With respect to the dimension of stratification (Halliday and Matthiessen, 2014, pp. 24–7), Hasan’s earlier (1973) discussion on personal distance mentioned at the beginning of this chapter has shown the possibility of relating personal distance, which is a feature on the contextual stratum, to features on the lexicogrammatical stratum, such as through the use of exophorics. Such relations between the contextual and lexicogrammatical strata have also been pointed out above as the unique feature of Hasan’s (1999) description of the field of discourse, making lexicogrammatical analysis a possible point of entry for the identification and categorization of fields for the purpose of constructing a biography of interactions (Halliday and Hasan, 1985). Having the point of entry on the lexicogrammatical stratum can take advantage of the fact that this stratum is of a lower level of abstraction to guarantee the robustness of the methodology, and the relatively rich and comprehensive descriptions available at this stratum, such as Halliday and Matthiessen (2014), developed since the 1980s (Matthiessen and Halliday, 2009, Foreword). With respect to the dimension of metafunction (Halliday and Matthiessen, 2014, pp. 30–1), as mentioned at the end of the previous section, constructing a biography of interactions adopting the descriptions in Hasan (1999) is a field-related endeavour, which concerns the ideational Marvin Lam 213 metafunction; whereas personal distance is a tenor-related notion, concerned with the interpersonal metafunction. The measurement and tracking of the development of personal distance through the construction of a biography of interactions therefore constitute a cross-metafunctional endeavour.

9.3 A semiotic deﬁnition and measurement of personal distance

With the attempt to detail the theoretical construct of Hasan’s notion of personal distance and its measurement presented in the previous sections, it would perhaps be timely to situate Hasan’s contribution to the wider context beyond linguistics to further clarify its properties. As briefly mentioned at the beginning of this chapter, linguistic work on personal distance often sees references to sociology, in which Bogardus’ definition is one of the most widely adopted. He coins the term ‘social distance’ and defines it as ‘degree of sympathetic understanding’, which ‘refers to feeling reactions of a favourable responsive type’ and ‘knowledge of a person which also leads to favourable responsive behaviour’, and ‘includes social nearness or social farness or any degree of distance between the extremes’ (Bogardus, 1959, p. 7). Bogardus (1925) argues that the origin of social distance is from one’s generalization of knowledge about a social/ethic group, and social distance can be reflected by one’s ‘social contact range’, that is, the range of encounters one would willingly admit members of other social/ethnic groups to be in, from close kinship such as marriage to excluding someone from one’s country. These reflect the general focus on society in sociology, and reflect how social distance is measured in numerous studies, such as those on occupations (Wilkinson, 1929), religions (Bogardus, 1933), between majority and minority groups (McAllister and Moore, 1991), tourism (Thyne and Lawson, 2001) and prison communities (Pabjan, 2005). One might see Hasan’s notion of personal distance as resembling the sociological view on social distance, especially in terms of the maximal– minimal continuum and the ‘extremes’ as stated by Bogardus, or relating biography of interactions with social contact range, such that the richness of a biography of interactions may mean a wider range of social contact. How- ever, they are different conceptions. The term ‘personal distance’ is used in Hasan (1973), and the term ‘social distance’ is used in her later discussions (Hasan, 1978, 1981; Halliday and Hasan, 1985). To differentiate the sociological notion of distance from Hasan’s, this chapter uses the term ‘personal distance’ to refer to Hasan’s notion, and the term ‘social distance’ to refer to the sociological one. Hasan’s notion of personal distance and the sociological notion of social distance are closely related as concepts on different orders of systems (cf. Matthiessen and Halliday, 2009, p. 13). Proposed by Halliday and developed by Matthiessen, the ordered typology of systems arranges systems 214 Society in Language, Language in Society

Table 9.1 Notions on distance with reference to the ordered typology of systems (adapted from Lam, 2010, p. 59)

Order of system Distance involved Basis of measurement

First Physical distance Spatial locations of entities Second Biological distance Biological beings of organisms Third Social distance Social identities of peoples Fourth Semiotic distance Social activities in which individuals participate

operating in different phenomenal realms in terms of their complexity, from physical, biological, social to semiotic systems; and each higher order of system inherits the properties of the immediate lower-order system, adding new characteristic properties (Matthiessen et al., 2010, pp. 152–5). Table 9.1 outlines the different notions on distance in the four orders and their corresponding basis of measurement. The sociological notion of social distance is considered to be in the third order (that is, social); while Hasan’s notion of personal distance is considered to be in the fourth order (that is, semiotic). Such distinction is foregrounded by the term ‘socio- semiotic distance’ proposed in Lam (2010) to refer to Hasan’s notion. These four notions of distance are not separate. As stated above, a higher-order system inherits the properties of the immediate lower-order systems. For instance, semiotic/personal distance is realized through the socially constructed meaning, which is ‘biologically activated and exchanged through physical channels’ (Halliday, 2003, p. 2). Relating semiotic/personal distance and social distance, so as to recognize how Hasan’s notion of personal distance is different but related to the sociological notion of social distance, can be considered as part of what Bernstein (1971, p. 118) regarded as ‘perhaps one of the most important events that has taken place in scientiﬁc endeavour in the twentieth century’, as ‘the convergence of both the natural and social sciences upon the study of linguistic aspects of communication’. This further underpins the importance of Hasan’s socio-semiotic approach in examining personal distance by relating participant roles with different uses of language, as mentioned at various points in this chapter.

9.4 Measuring and tracking the development of personal distance

The second half of this chapter turns to discussing the practical implementation of Hasan’s notion in measuring and tracking the development of personal distance between interactants, and illustrates its significance in relating participant roles and their use of language. This section will first Marvin Lam 215 illustrate how features in the FIELD system network (Hasan, 1999, reproduced as the leftmost part in Figure 9.1) can be realized through lexicogrammatical analysis to construct a biography of interactions. In Hasan’s FIELD system network, there are four simultaneous systems at the first degree of delicacy, namely MATERIAL ACTION, VERBAL ACTION, SPHERE OF ACTION and ITERATION.MATERIAL ACTION and VERBAL ACTION describe the ‘two faces’ of field conceptualized as ‘doing with different degrees of speaking’ (Hasan, 1999, p. 274, italics in original). As a linguist, Hasan treats verbal action as more central than material action, as her discussion states that non-linguistic human action, as described in MATERIAL ACTION, ‘impinges in some way on linguistic action – on the choices in speaking and interpreting’, and if there is no use of language present (that is, verbal action), then ‘there can be no call for a field of discourse’ (Hasan, 1999, pp. 274–5, emphasis in the original). With Hasan’s emphasis on the central role verbal action plays, this section will focus on a few features in VERBAL ACTION, while the importance of non-linguistic action will not be ignored. As in the illustration of the realization will show, the lexicogrammatical analysis involved is not only done on the language used by the individuals (that is, their verbal action), but the analysis of a record – very often in the form of language – of their non-linguistic action is also necessary. This is demonstrated in Lam (2010) in constructing biographies of interactions between the protagonists of a novel, which is treated as a record of social activities as if they happened in the real world. The lexicogrammatical analysis covers the dialogue between the characters under study (that is, their verbal action), as well as the omniscient voice in narrating the non-linguistic action (that is, the material action), referred to as ‘narration’. Examples of analysis in this chapter are quoted from this work, with which the methodology was applied to tracking the development of the personal relationship between the protagonists Sophie Neveu and Robert Langdon in Dan Brown’s The Da Vinci Code. Referring to the leftmost part of Figure 9.1, in VERBAL ACTION,atthe first degree of delicacy (that is, the second degree of delicacy with respect to FIELD), there are two choices of features: [ancillary] and [constitutive]. Ancillary verbal action is where language acts as an additional resource, playing an assisting role in a social activity (Hasan, 1999, p. 276); whereas constitutive verbal action, as the term suggests, is where language plays a central role in the social activity and the material action is irrelevant to the verbal action (Hasan, 1999, p. 277). As pointed out by Hasan (1999, p. 281), this distinction between ancillary and constitutive verbal action seems to relate to the role language plays in the interaction, and one might argue this as a mode-related issue (that is, relevant to the textual contextual parameter instead of the ideational one). However, this distinction, as pointed out by Hasan in the same discussion, addresses the fundamental nature of the action in question: action that can only be done verbally (that 216 Society in Language, Language in Society is, constitutive verbal action) or action that can be done without language (when verbal action is ancillary), and distinguishing the nature of the action is a field-related issue. This distinction can be seen as interfacing field with mode, as these parameters are not discrete from each other. This distinction between [ancillary] and [constitutive] can correspondingly be realized by the presence and absence of relations exhibited by the ideational choices (in terms of Processes, Participants and Circumstances) in the language used by individuals and in the record of their non-linguistic actions, that is, with or without cohesive chains and cohesive harmony (Halliday and Hasan, 1985, p. 84; Cloran, 1999, pp. 189–96). Extract 1 shows a small fragment of a social activity with ancillary verbal action, with Partici- pants in Sophie’s language and in the narration of her non-linguistic actions being coreferential (Halliday and Hasan, 1976, p. 3), and shown by underlining in the text. To contrast, Extract 2 does not show any cohesive chains or cohesive harmony between the individuals’ language use and the record of their non-linguistic action, thus realizing the feature [constitutive].

Extract 1: Realizing ancillary verbal action Narration: She grabbed a thick bar of soap, placed the tracking dot on top of it, and used her thumb to push the disk down hard into the bar. As the disk sank into the soft surface, she pinched the hole closed, ﬁrmly embedding the device in the bar. Handing the bar to Langdon, Sophie retrieved a heavy, cylindrical trash can from under the sinks. Before Langdon could protest, Sophie ran at the window, holding the can before her like a battering ram. Driving the bottom of the trash can into the center of the window, she shattered the glass. Alarms erupted overhead at earsplitting decibel levels. Sophie: Give me the soap!

Extract 2: Realizing constitutive verbal action with the feature [reﬂection based] Narration: Ahead, two topless teenage girls shot smoldering gazes into the taxi. Beyond them, a well-oiled black man in a G-string turned and ﬂexed his buttocks. Beside him, a gorgeous blond woman lifted her miniskirt to reveal that she was not, in fact, a woman. Langdon turned his gaze back inside the cab and took a deep breath. Sophie: Tell me about the Priory of Sion, Narration: Langdon nodded, [ ...] the brotherhood’s history spanned more than a millennium ...an astonishing chronicle of secrets, blackmail, betrayal, and even brutal torture at the hands of an angry Pope. Marvin Lam 217

Robert: The Priory of Sion, was founded in Jerusalem in 1099 by a French king named Godefroi de Bouillon, immediately after he had conquered the city. Narration: Sophie nodded, her eyes riveted on him. Robert: King Godefroi was allegedly the possessor of a powerful secret – a secret that had been in his family since the time of Christ.

Being the entry condition, the feature [constitutive] leads to the options [practical] and [conceptual]. Practical constitutive verbal actions are those which construe ‘the outline of some material action’ of which ‘the future occurrence of some material action is always on the cards’ (Hasan, 1999, p. 283); while the conceptual ones ‘do not call for any physical action’ (Hasan, 1999, p. 288). This distinction can be realized by whether there are material clauses with (1) the processes in future tense (Halliday and Matthiessen, 2014, pp. 399–400), (2) modal assessment of probability or inclination (Halliday and Matthiessen, 2014, pp.176ff.), (3) imperative mood (Halliday and Matthiessen, 2014, pp. 139ff.) and/or (4) temporal circumstance(s) referring to the future in an individual’s language use. Extract 3 is a fragment of a social activity with practical constitutive verbal action with the realizational lexicogrammatical features double-underlined.

Extract 3: Realizing practical constitutive verbal action with the feature [instruct] Robert: I’ll come. Sophie: No! We don’t know how long the Grand Gallery will stay empty. You have to go.Go. Now. I’ll see you at the embassy, Mr. Langdon. Robert: I’ll meet you there on one condition, Sophie: What’s that? Robert: That you stop calling me Mr. Langdon. Sophie: Good luck, Robert.

At the third level of delicacy in VERBAL ACTION (that is, the fourth level of delicacy in FIELD), [practical] is the entry condition to the choices of [instruct] and [plan]; whereas [conceptual] is the entry condition to three simultaneous systems: the first one with choices of [relation based] and [reflection based], the second one with choices of [first order] and [second order], and the third one with choices of [informing] and [narrating]. With the constraint of space, only the features [instruct], [plan], [relation based] and [reflection based] will be illustrated. A practical constitutive verbal action with the feature [instruct] ‘produces something of a resource for an addressee: the speaker’s wordings construe a 218 Society in Language, Language in Society virtual design of some deferred material action, specifying the various stages through which it may be performed’, and ‘[instruct] is to enable the bringing about of such action as and when the addressee might need/desire to do so’; and [plan] is defined ‘in the sense that it is oriented towards some future physical action of the interactants’ (Hasan, 1999, p. 282, my emphases). These definitions of the features [instruct] and [plan] appear to exhibit a dichotomy such that the material actions construed through the verbal actions are carried out with the unison of the speaker and the addressee (that is, [plan]) or without such unison (that is, [instruct]). This distinction can be lexicogrammatically realized by the nature of the Participants in the individual’s language use: if in general the Participants of the clauses construe the speakers OR the addressees, then the feature [instruct] is realized; if both speakers AND addressees are construed as Participants, then the feature [plan] is realized. Extract 3 would then further realize the feature [instruct], whereas Extract 4 would realize the feature [plan], with the realizational lexicogrammatical features in italics.

Extract 4: Realizing planning practical constitutive verbal action Sophie: Okay, we have to get off the road. We need a safe place to figure out what that key opens. Robert: How about my hosts at the American University of Paris? Sophie: Too obvious. Fache will check with them. Robert: You must know people. You live here. Sophie: Fache will run my phone and e-mail records, talk to my coworkers. My contacts are compromised, and finding a hotel is no good because they all require identification. Robert: Let’s call the embassy. I can explain the situation and have the embassy send someone to meet us somewhere. Sophie: Meet us? Robert, you’re dreaming. [ ...] How much cash do you have? Robert: A hundred dollars. A few euro. Why? Sophie: Credit cards? Robert: Of course. Sophie: We lost them for the time being, but we won’t last another five minutes if we stay in this car.

Referring to the first of the three simultaneous systems with which the entry condition is [conceptual], a verbal action with the feature [relation based] (in contrast to [reflection based]) is stated by Hasan as the most ‘potent in the enactment of social relations’ (1999, p. 288). While the distinction between [ancillary] and [constitutive] can be considered as interfacing field and mode as discussed above, the feature [reflection based] can be argued to be interfacing field and tenor. In contrast with [relation based], the feature [reflection based] refers to verbal actions that only ‘produce semiotic Marvin Lam 219 constructs’ (Hasan, 1999, p. 289). Therefore, the distinction between the features [relation based] and [reflection based] can be made according to the presence and absence of the enactment of interpersonal meaning in the verbal actions, which, according to Hasan (1999, p. 289), can be realized by ‘the (interpersonal) lexicogrammar of mood, modality, and modification which construe those meanings’. Lam (2010) suggests a more straightforward lexical analysis would be sufficient in distinguishing [relation based] from [reflection based]: the use of vocatives (Halliday and Matthiessen, 2014, pp. 159–60), first/second-person pronouns and possessive pronouns, indicating that individuals are addressing each other and are taking the interpersonal relationship into account in the negotiation of meaning, as indicated in Extract 5 by the dotted underlines. Extract 5 can be contrasted with Extract 2, which is a fragment of a social activity with reflection-based conceptual constitutive verbal action.

Extract 5: Realizing relation-based conceptual constitutive verbal action ... Robert: Sophie,...... listen. The letters P.S . do they mean anything else to you?.... Anything at all? Sophie: I. told you,.... the initials mean Princess Sophie. Robert: I. know, but did you.... ever see them anywhere else? Did your grandfather ever use P.S. in any other way? As a monogram, or maybe on stationery or a personal item? The initials, have you.... seen them? Sophie: Yes, I. saw the initials P.S. once. When .I was very young. Robert: Where? Sophie: On something very important to him. Robert: Sophie,...... this is crucial. Can you.... tell me... if the initials appeared with a symbol? A ﬂeur-de-lis? ... Sophie: But how could ....you possibly know that! Robert: I’m. fairly certain your grandfather was a member of a secret society. A very old covert brotherhood.

So far this section has outlined how certain features in Hasan’s FIELD system network can be realized through a lexicogrammatical analysis of individuals’ language use and the record of their non-linguistic actions. Figure 9.1 displays the analysis conducted in Lam (2010) as the accumulative realization of ﬁeld in various social activities participated in by Sophie and Robert. Thirty-six social activities are lined up horizontally in chronological order, with the features realized in the FIELD system network of each social activity listed vertically. This is a visual depiction of the biography of interactions between Sophie and Robert. The personal distance between Sophie and Robert is reduced along the course of the social activities in which they co-engage, as both the 220 1 4 8 2 22 19 20 10 10 10 11 13 16 4 8 10 0 15 16 1 1 9 8 (Lam, 2010) 0 1 15 16 17 18 19 17 18 19 20 34 4 11 12 13 14 15 16 17 18 19 20 21 22 15 16 24 25 26 27 28 29 30 31 32 33 34 35 36 Total 5 6 7 89 23 891 22 7 6 78 13 9101112 21 The Da Vinci Code 012314 8 9 10111213 20 7 7 8 9 10 11 12 13 14 56 17 18 19 5 12 56 567 8910 15 16 34 4 4 5 6 7 8 9 10 1 1234 3 Social activities co-engaged only by Sophie Neveu and Robert Langdon in Dan Brown's the Da Vinci Code and Robert Sophie Neveu Langdon in Dan Brown's by Social activities co-engaged only 12 3 4567 2 3 4 5 6 12 1234 12 12 3 1 2 3 1234567891011121314 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1 12 3 4 12 1 2 3 4 5 6 7 Instruct Plan Relation based First order Second order Informing Narrating Reflection based Deferred Absent Practical Conceptual Non- Present Ancillary Specialized Quotidian Present Constitutive Sophie Neveu and Robert Langdon’s biography of interactions in Dan Brown’s ACTION ACTION ACTION VERBAL MATERIAL SPHERE OF ITERATION Field Figure 9.1 Marvin Lam 221 variety and frequencies of social activities in which they co-engage increase and their biography of interactions becomes richer. This is a highly predictable development of their personal distance because the enrichment of biography of interactions is entailed with the co-engagement in social activities. Apart from this straightforward understanding of the development of personal distance, with reference to the distinctions of features in the FIELD system network, a biography of interactions can also elicit information regarding the characteristics of the involved personal relationship in relation to the nature of the social activities in which individuals participate.

9.5 Characterization of personal relationship

As elaborated in the previous section, the realization of features in the FIELD system network concerns the distinctions of certain properties of social activities. Apart from formulating the biography of interactions to show a reduction of personal distance, these distinctions can further serve as a source to characterize the personal relationship resulting from the involved individuals’ biographies of interactions. The characterization of personal relationships can be done with respect to the frequencies at which certain features of the FIELD system network are realized. With the degree of delicacy of the FIELD system network attained by the current discussion, there are three aspects in characterizing personal relationship, as discussed in the following paragraphs. A more detailed characterization of personal relationships is possible when more features in FIELD are taken into consideration, which is demonstrated in Lam (2010). If the realization of the features [ancillary] and [practical] outnumbers that of [constitutive], then the personal relationship resides more in the material realm than in the semiotic realm because the social activities are more concerned with material actions (either at the ‘here and now’ of the current social activity, that is [ancillary], or in the future as [practical]); if the realization of [constitutive] outnumbers that of [ancillary] and [practical], then the familiarity resides more in the semiotic realm based on the semiotic entities construed by the conceptual verbal actions. If the realization of [instruct] outnumbers that of [plan], then the personal relationship is subject to variation with respect to individuals’ points of view because the individuals do not have the same participating roles in the material actions construed through the verbal actions, and they do not have equal status in the negotiation of these material actions through the verbal actions. If the realization is otherwise, then the personal relationship tends to be less differentiated with regards to individuals’ points of view. If the realization of [relation based] outnumbers that of [reflection based], then the personal relationship is based on how individuals relate to each other instead of the content exchanged in the social activities; the opposite 222 Society in Language, Language in Society will result if the realization of [reflection based] outnumbers that of [relation based]. With these three characterizations and referring to the frequencies of the realization of the features indicated on the right-hand side of Figure 9.1, the personal relationship of Sophie and Robert can be characterized as follows: the relationship resides more in the semiotic realm than the material realm ([ancillary] and [practical]: 15; [constitutive]: 21); and the relationship is based slightly more on how Sophie and Robert relate to each other than on the contents of the social activities ([relation based]: 13; [reflection based]: 8). One can also suggest the relationship would probably be different with respect to Sophie and Robert’s points of view ([instruct]: 1; [plan]: 4), but the frequencies may not be high enough to justify this characteristic. Such characterization can also be done with a developmental perspective. Figure 9.2 is another visual presentation of Sophie and Robert’s biography of interactions, with the diagram at the top showing the accumulative frequencies of realizing [ancillary] and [practical] versus [constitutive] and the bottom diagram showing those for [reflection based] and [relation based]. Referring to the top diagram, the personal relationship is balanced to be in both material and semiotic realms from the beginning until social activity 20. Beyond that, the relationship shifts to being more in the semiotic realm. Referring to the bottom diagram, the relationship is concerned more with how Sophie and Robert relate to each other up to social activity 16. And beyond that, the relationship shifts to be more concerned with the content of the social activities. With these developmental trends of the personal relationship, the middle portion of the course of the social activities appears to be special because both trends undergo some changes. Such detailed tracking of the development of relationship is significant in its own right. Moreover, the development of the personal relationship coincides with the change of language used by the individuals. This is discussed in the following section to conclude the chapter by returning to Hasan’s notion of personal distance, i.e. the understanding of personal distance and language behaviours are intrinsic to each other.

9.6 Hasan’s notion of personal distance revisited

With the realization of features in Hasan’s FIELD system network in relation to the construction of a biography of interactions discussed in section 9.4, and the characterization, both in general and in developmental terms, of personal relationship discussed in the previous section, this concluding section revisits Hasan’s notion discussed at the beginning of this chapter by relating such measurement and tracking of the development of personal distance with other language features which emerge in the individuals’ use of language. Three phenomena are discussed below to exemplify the relation. 223 [Constitutive] [Reflection based] [Relation based] [Ancillary] and [practical] 123456789101112131415161718192021222324252627282930313233343536 123456789101112131415161718192021222324252627282930313233343536 8 6 4 2 0 5 0 12 10 25 20 15 10 Accumulative frequencies of [ancillary] and [practical] vs [constitutive] (top) and [relation based] vs [reﬂection based] (bottom) (Lam, 2010) Figure 9.2 224 Society in Language, Language in Society

As pointed out at the end of the previous section, the developmental trends of Sophie and Robert’s relationship (depicted in Figure 9.2) indicate that the middle portion of the course of the social activities is more significant than other portions. This period also sees the change of how Sophie and Robert respond to each other’s initiations. According to Halliday and Matthiessen (2014, p. 137), responses can be categorized as ‘expected’ or ‘dis- cretionary’. In the earlier portion of the course of the social activities, Sophie accumulatively gives more expected responses to Robert’s initiations than the other way round; Robert then ‘catches up’ during the period in question to give a comparable amount of expected responses to Sophie; beyond that period, Robert gives more expected responses to Sophie than the other way round. As indicated in Figure 9.1 and the bottom diagram in Figure 9.2, social activity 6 constitutes the first relation-based conceptual constitutive verbal action, and is a point of enrichment in the biography of interactions in terms of the range of social activities. This is also the first in a series of social activities in which Robert contributes more than Sophie in terms of the number of clauses uttered; and the first in which Sophie makes more ‘demanding’ moves than ‘giving’ moves. In other words, before this point of enrichment of the biography, Sophie is ‘more talkative’ than Robert and makes more statements than questions and commands. Beyond this point of enrichment, Sophie and Robert ‘swap’ in terms of the properties in their language use. As shown in Figure 9.1, the four social activities with the feature [plan] distribute quite evenly across the course of social activities (that is, social activities 3, 13, 26 and 36), serving as good signposts of any changes that occur along the development of the interpersonal relationship. With respect to these four social activities, Robert constantly shows an increase in propor- tional contribution to the verbal actions in the social activities in terms of the number of clauses he utters. One still needs to clarify the nature of such relations between the development of personal distance, as reflected by the biography of interactions, and the change of language used by individuals. However, the present discussion argues that it is not coincidental for the development of personal distance to be corresponding to these language changes. And the endeavours discussed and illustrated in this chapter, informed by Hasan’s notion of personal distance and its measurement, provide a legitimate way to interrogate more data to clarify the nature of the relation between personal distance and the language used of the involved individuals.

References

Bernstein, B. (1971) Class, Codes and Control,Vol.I:Theoretical Studies towards a Sociology of Language (London: Routledge & Kegan Paul). Marvin Lam 225

Bogardus, E. S. (1925) ‘Social distance and its origins’. Journal of Applied Sociology 9: 216–26. Bogardus, E. S. (1933) ‘A social distance scale’. Sociology and Social Research 17: 265–71. Bogardus, E. S. (1959) Social Distance (Los Angeles, Calif.: Antioch Press). Brown, R. and Gilman, A. (1960) ‘The pronouns of power and solidarity’. In T. A. Sebeok (ed.) Style in Language (Cambridge, Mass.: The MIT Press), pp. 253–76. Brown, P. and Levinson, S. C. (1987) Politeness: Some Universals in Language Usage (Cambridge: Cambridge University Press). Butt, D. G. (2000a) ‘The meaning of a network: linguistic networks for modelling complex behaviour’. In D. G. Butt and C. M. I. M. Matthiessen (eds) The Mean- ing Potential of Language: Mapping Meaning Systemically (Linguistics Department, Macquarie University, Australia, mimeo). Butt, D. G. (2000b) ‘Modelling behaviour semantically’. Manuscript presented in Systemic Functional Linguistics Research Seminar on 20 October 2000, Sydney University. Butt, D. G. (2003) ‘Parameters of context: on establishing the similarities and differences between social processes’. Unpublished mimeo, Macquarie University, Sydney. Cloran, C. (1999) ‘Context, material situation and text’. In M. Ghadessy (ed.) Text and Context in Functional Linguistics (Amsterdam and Philadelphia: John Benjamins), pp. 177–217. Coupland, N. (1985) ‘Hark, hark and lark: social motivations for phonological style shifting’. Language and Communication 5 (3): 153–71. Edwards, A. D. (1978) ‘Analysing classroom talk’. In P. French and M. MacLure (eds) Adult–Child Conversation (London: Croom Helm), pp. 291–308. Giles, H. (1973) ‘Accent mobility: a model and some data’. Anthropological Linguistics 15 (2): 87–105. Goffman, E. (1955) ‘On face-work: an analysis of ritual elements in social interaction’. Psychiatry 18: 213–31. Halliday, M. A. K. (1961) ‘Categories of the theory of grammar’. Word 17 (3): 241–92. Halliday, M. A. K. (1975) ‘Language as social semiotic: towards a general sociolinguistic theory’. In A. Makkai and V. B. Makkai (eds) The First LACUS Forum (Columbia, SC: Hornbeam Press), pp. 17–46. Halliday, M. A. K. (1995) ‘On language in relation to fuzzy logic and intelligent computing’. Proceedings of 1995 IEEE International Conference on Fuzzy Systems (Piscataway, NJ.: IEEE). Halliday, M. A. K. (2003) The Collected Works of M. A. K. Halliday,Vol.3:On Language and Linguistics. Edited by J. J. Webster (London and New York: Continuum). Halliday, M. A. K. and Hasan, R. (1976) Cohesion in English (London and New York: Longman). Halliday, M. A. K. and Hasan, R. (1985) Language, Context and Text: Aspects of Language in a Social-Semiotic Perspective (Geelong, Vic.: Deakin University Press). Halliday, M. A. K. and Matthiessen, C. M. I. M. (2014) Halliday’s Introduction to Functional Grammar (Oxford and New York: Routledge). Hasan, R. (1973) ‘Code, register and social dialect’. In B. Bernstein (ed.) Class, Codes and Control, Vol. II: Applied Studies towards a Sociology of Language (London: Routledge & Kegan Paul), pp. 253–92. Hasan, R. (1978) ‘Text in the systemic-functional model’. In W. U. Dressler (ed.) Current Trends in Textlinguistics (Berlin and New York: Walter de Gruyter), pp. 228–46. 226 Society in Language, Language in Society

Hasan, R. (1979) ‘On the notion of text’. In J. F. Petofi (ed.) Text versus Sentence: Basic Questions of Text Linguistics. Papers in Text Linguistics 20 (Hamburg: Buske), pp. 369–90. Hasan, R. (1981) ‘What’s going on: a dynamic view of context in language’. In J. E. Copeland and P. W. David (eds) The Seventh LACUS Forum (Columbia: Hornbeam Press), pp. 106–21. Hasan, R. (1996) ‘Ways of saying: ways of meaning’. In C. Cloran, D. Butt and G. Williams (eds) Ways of Saying: Ways of Meaning: Selected Papers of Ruqaiya Hasan (London: Cassell), pp. 191–242. Hasan, R. (1999) ‘Speaking with reference to context’. In M. Ghadessy (ed.) Text and Context in Functional Linguistics (Amsterdam and Philadelphia: John Benjamins), pp. 219–328. Hasan, R. (2009) ‘The place of context in a systemic functional model’. In M. A. K. Halliday and J. J. Webster (eds) Continuum Companion to Systemic Functional Linguis- tics (London and New York: Continuum), pp. 166–89. Hymes, D. (1977) Foundations in Sociolinguistics (London: Tavistock). Lam, M. (2010) ‘Socio-semiotic distance and its measurement in narrative discourse’. Unpublished PhD thesis, City University of Hong Kong. McAllister, I. and Moore, R. (1991) ‘Social distance among Australian ethnic groups’. Sociology and Social Research 75 (2): 95–100. Matthiessen, C. M. I. M. and Halliday, M. A. K. (2009) Systemic Functional Grammar: A First Step into the Theory (Beijing: Higher Education Press). Matthiessen, C. M. I. M., Teruya, K. and Wu, C. Z. (2008) ‘Multilingual studies as a multi-dimensional space of interconnected language studies’. In J. J. Webster (ed.) Meaning in Context: Implementing Intelligent Applications of Language Studies (London and New York: Continuum), pp. 146–221. Matthiessen, C. M. I. M., Teruya, K. and Lam, M. (2010) Key Terms in Systemic Functional Linguistics (London and New York: Continuum). Pabjan, B. (2005) ‘Measuring the social relations: social distance in social structure – a study of prison community’. Acta Physica Polonica B 36 (8): 2559–74. Rubini, M. and Kruglanski, A. W. (1997) ‘Brief encounters ending in estrange- ment: motivated language use and interpersonal rapport in the question–answer paradigm’. Journal of Personality and Social Psychology 72 (5): 1047–60. Semin, G. R. (2007) ‘Linguistic markers of social distance and proximity’. In K. Fiedler (ed.) Social Communication (New York and Hove: Psychology Press), pp. 389–408. Semin, G. R. and Fiedler, K. (1988) ‘The cognitive functions of linguistic categories in describing persons: social cognition and language’. Journal of Personality and Social Psychology 54 (4): 558–68. Thyne, M. A. and Lawson, R. (2001) ‘The design of a social distance scale to be used in the context of tourism’. In P. M. Tidwell and T. E. Muller (eds) AP – Asia Pacific Advances in Consumer Research (Vol. 4) (Provo, Utah: Association for Consumer Research), pp. 102–7. Ure, J. (ed.) (not dated) ‘Text types classified by situational factors’. Unpublished manuscript. Wilkinson, F. (1929) ‘Social distance between occupations’. Sociology and Social Research 13: 234–44. 10 Studying Language in Society and Society through Language: Context and Multimodal Communication Rebekah Wegener

10.1 Introduction

Hasan describes context in the following way: ‘context of situation as construed by discourse is a tripartite entity, each component of which is always active in the production of a text’ (Hasan, 2004, p. 21). This description of context of situation as an entity may be due to a number of reasons. For instance, there is a language-driven semantic drift towards reification because once we start to talk about ‘context’ we come to refer to it as an entity since all our language pushes us in that direction. There is also a discourse focus on context, in the sense that ‘context of situation is construed by discourse’ (Hasan, 2004, p. 21). And there is a bounded notion of discourse, in that a text view of discourse tends towards strong boundaries that make it more likely for context to be seen as an entity. Hasan extends her argument concerning the construal of context by discourse by making it very clear that the focus is discourse alone by suggesting that: ‘The qualifier of discourse is a reminder that unlike the “cultural activity theory” associated with the Russian, especially (neo-)Vygotskian literature (Engeström et al., 1999), “context theory” was not intended to apply to all kinds of social action, being designed specifically with discourse in mind’ (Hasan, 2004, p. 21, emphasis in the original). Thus, Hasan makes a distinction between the social action and those aspects of social action which relate specifically to discourse, or are ‘construed by discourse’. The sharp distinction between social action and discourse poses some interesting issues that must be addressed:

(a) Context of situation, as Hasan models it, includes other modalities as part of the context rather than as the discourse. In the wider context of systemic functional linguistics, this means that modalities are present on the expression plane as separate modes of meaning and represented in the contextual plane as part of the context for discourse.

227 228 Society in Language, Language in Society

(b) In applying Hasan’s model of context to multimodal communication (which includes most communicative events – even written ones), Hasan’s view of context restricts valid application to spoken and written language only. This restriction causes some problems for modelling alternate forms of communication such as communication through challenging behaviour, augmented communication or computer-mediated communication of some forms, which do not take a spoken or written form.

In this chapter I discuss the strengths of Hasan’s model of context for dealing with the boundary challenge and some of the challenges that arise when we attempt to extend Hasan’s model to multimodal communication.

10.2 The problem of context

Discourses are not produced in a vacuum, but in contexts which both shape and are shaped by the ongoing interaction. (Butler, 2003, p. 486)

When I ﬁrst read Hasan’s (1999) analysis of a mother and child eating food together, I was struck both by my instant familiarity with the context of the text and my own reaction to Hasan’s choice of boundaries for the text. I set out to see if I could record similar texts, not in terms of matching participants, but in terms of a similarity of fuzzy boundaries, time pressures, competing interests and multimodal interaction between a caregiver and those in their care. Texts 1 and 2 below are a tiny selection from many weeks of recording, but they illustrate the challenges of multimodal interaction and context. For many years I showed these texts to my ﬁrst-year students and asked them to reconstruct the context of the texts by considering where the texts might have come from, who the participants might be, their characteristics, as well as what the short- and long-term goals of the text might be and what evidence might be found for all of these aspects. Without fail they were able to identify at least some of these aspects of the context of the text, but there were certain aspects that they simply could not recover.

Text 1 C: Mum, I’ll be home by about 8pm A: I presume you’ll want food C: Yeah, I’m gonna call in and see Sally though ....I’m gonna go there about six tonight A: how come he’s coming out tonight? C: aaw I don’t know well he might not..but um I’ve got just end up I mean I don’t care no um he’ll just whinge and go ‘aw the trafﬁc’ like that but..but I’ve got to call in and see the others anyway. Rebekah Wegener 229

A: ah right C: because well Megan wants me to read her thesis and it’s Sally’s birthday, well it was her birthday on Friday, but she’s having a barbeque and I thought I’ve got to pick up her thesis so. A: you don’t really have the time to read it now do you? C: aw yeah A: What’s her thesis on? C: learning A: What do you mean? C: it’s on music and film like] A: [sorry, what was the first part? C: it’s on cued response in music] A: [aw right] C: [in film] A: [get dressed!

Text 2 A: (Yawns and sighs). Now what do you want on your sandwiches B: Peanut butter A: Just peanut butter? B: yeah. A: Aw I’m so sore (talk based radio can now be heard in the background) B: mm A: Um now ...you look very groggy dear ...aw ...yawn, it’s a bit rough having to get up on the ﬁrst day, it’s only going to get worse ...you better get up C C: yeah A: Can I go back to bed? Do you think anyone will notice? (laughter) .... alright, there’s the sandwiches, let’s see if I can rustle something up in the way of biscuits ..... we haven’t got anything ...I’m getting cold ...we always, well we don’t always, we used to have three hundred million ...yeah and he wanted the buses ...(laughter) ...yeah funny. B: look at the light. A: mmhm just a sec pet, what was that? B: look at all the light on the mountains A: Yeah that’s the sun reﬂecting off the windows, isn’t it beautiful C: aw ...I’m so tired A: What did you say C darling? C: I’m so tired B: nope A: So am I C: can I go back to bed? 230 Society in Language, Language in Society

B: yeah maybe A: You could stay home if you needed, if you wanted C: Not really ...aw look at the light; at the sun A: ...mmm. B: yeah I love those. Can I have two? A: Listen, I haven’t done any lunch for you do you want to be like me and rat around and see] C: [Yeah, can I take a K-time bar? A: a what? C: A K-time bar A: Ah yeah of course.. Shower! B: Yeah ok C: don’t forget your lunch.. I’ll put it in your bag B: uh huh ta

Annotation key: . = second long pause [ = overlapping discourse () = sounds not able to be transcribed

As outsiders confronted with these texts, we are perhaps initially struck by their individuality. They may seem strange, lacking cohesion and, at times, even logic. Indeed the strangeness of the texts and the potential for mis- understanding highlight the importance of context and multimodality1 to meaning making. Given time, most readers, like my students, will identify threads of familiarity, threads which allow the observer to identify what it is that is going on in these texts. These threads will weave together into questions like, what activity is taking place, who are the participants involved, and how are they connected to each other, what environment or setting are they in and for what purposes are they using language? Through asking questions such as these we are able to habituate to our environment, to predict what is likely to occur next and what behaviour this is likely to require from ourselves and those around us, both human and non-human. Understanding the importance of these questions and the ways in which they interact is central to the concept of context. It takes more than a collection or organization of lexical items for a text to make sense. It is necessary to have some further information about the environment of that text. Context can be expressed as everything that may be necessary for a text to make sense or the way that a connection is made with the culture. It is the seam between the social or cultural fabric and the fabric of language. A text is understandable or recognizable as a text because of this connection. To illustrate this point, let us consider Text 1. How do we know what Rebekah Wegener 231 it is that is going on here? We might begin by considering what the participants are doing. It appears from the text that the primary activity at this point is talk; talk about visiting friends, getting lifts, getting food, and theses. However, there are several indications that this might not be the only activity taking place. The text ends with a directive from A for C to ‘get dressed’. This suggests that there might be other activities going on as well that take priority over the talk about theses, visits and lifts home. This is further reinforced by the fact that A at one point requests that C repeat a section of the discourse, saying ‘sorry what was the first part’, suggesting that her entire focus may not be on the talk at hand, but rather some other activity. If we consider the directive to ‘get dressed’, it would appear that this other activity might have something to do with getting ready and possibly in the morning since it involves getting dressed (although one might get dressed at any time). Thus, the short-term goal of the activity might be to get away, possibly in the morning. A longer-term goal might be to plan the rest of the day and an even longer- term goal, given the willingness to discuss C’s apparent interests, might be the maintenance of close social relations. Of course we cannot be sure, but combined with the opening line ‘I’ll be home by about 8pm’ we can assume that the discourse takes place sometime at the start of the day. So, we might expect that the other activity is some activity to do with getting ready in the morning, such as washing, dressing, preparing food or packing bags for the day. In fact, A is preparing lunches and C is eating breakfast and this talk takes place in the kitchen, though we cannot establish any of this from the spoken (and transcribed) text. The aspects of the context outlined above are referred to within SFL as the field (Halliday and Hasan, 1985). The field of discourse refers to ‘the nature of social activity relevant to speaking’ (Hasan, 1999, p. 232). The field of discourse also tells us something about the participants. The fact that C refers to her friend’s thesis tells us that she is likely to be university educated, and thus at least in her twenties. Of course we have no information available to determine C’s gender in this snippet and we cannot be sure that C is female at all. The relationship between C and A is established at the start of the text when C refers to A as ‘Mum’. Thus, the agentive relation is mother to child, with the social relation being familial and asymmetrical in terms of power. The role asymmetry here is modified by the relative ages and level of education of the participants. Certainly, when C is discussing the topic of her friend’s thesis she is in control since her mother has no knowledge of the topic. However, we see that the primary context reasserts the familial power relations with A giving C an order to go and ‘get dressed’. Thus, the role asymmetry is only partially field dependent. The social distance between C and A is minimal; they are after all mother and daughter, and this is reflected in their discourse. Their language is that 232 Society in Language, Language in Society of people who share many contexts in common and have a long discourse history. For example, the outsider might wonder at the fact that A’s response to C’s statement that she is going to her friend’s house is ‘How come he’s coming out tonight?’ This does not flow logically from the discourse presented. It is the leap in logic that close network members make. What is involved in this leap is as follows: C would normally get a lift home with her mother, if she is going to her friend’s that means she is not getting a lift with her mother, which means she is getting a lift with her boyfriend on a night when he would not typically visit, hence A’s question. This complex reasoning reflects the nature of the relationship between the two participants and their long history in common (see Chapter 9 this volume). These aspects of context are referred to in SFL as the tenor of discourse. The tenor of discourse refers to the ‘nature of social relations relevant to speaking’ (Hasan, 1999, p. 232). We might also be interested in the role of language in this environment. Parts of this text are context independent. This textual independence has two aspects to it, namely, that the text might occur anywhere, and that it is abstract. Both these aspects mean that the text is, to a large degree, not grounded in the physical environment or what Hasan (1999) refers to as the material situational setting (MSS). This lack of grounding in the MSS makes it difficult to decide what role language plays. It is hard to decide whether or not language is the activity or is helping along another activity. Certainly, without the last line, ‘get dressed’, it would be hard to establish that any other activity is in fact taking place. Obviously from the disjointed structure, grammar, repairs and restarts this is spoken discourse, but the conversation might very easily be taking place in the car, on the phone or anywhere else and indeed it might have been written to be performed. The features of context discussed here are generally referred to as mode. Mode relates to ‘the nature of contact for the conduct of speaking’ (Hasan, 1999, p. 232). Typically we find that the MSS dominates in determining what the primary activity of a context is, but this is largely a matter of boundaries. As is often the case with research, the text presented here has been cropped from a much longer recording of the family, and if I had cropped the text to exclude the final line we would not be able to determine key aspects of the field, tenor or mode and our analysis of the context would have looked quite different. This illustrates the very real problem of determining and arguing for boundaries in analysis and the very real impact that our decisions about boundaries will have on our analysis.

10.3 The unbounded nature of context

Once context of situation asserted itself as necessary to the elucidation of the meaning of words, it would be clear that, unless a people had a ﬁxed and ﬁnite set of things to say in their daily activities, then the number Rebekah Wegener 233

of things they said and the contexts they said them in would be both inﬁnite and indeterminate. (Wolf, 1989, p. 262)

Wolf’s concern is echoed by Cook (1990), whose perspective on context focused on the potential that context analysis had to become a task in ‘transcribing infinity’. When viewed from the perspective of its place in daily social life, the dynamic flow of social process creates a sense of infinity and boundlessness that can be disconcerting. The problem is largely a theoretical problem and the price we pay for abstraction. As Hasan points out, ‘the price one pays for the power gained from abstraction is to have to define data in such a manner that it does not include everything that may be going on, for in nature there are no clear cut, given boundaries’ (1995, p. 187, emphasis in original), and it is necessary to create artificial boundaries to work with while at the same time bearing in mind the importance of the ongoing social process. These are the means by which choices within a theoretically defined contextual system are made meaningful. Hasan has been instrumental in addressing the question of context and language. In fact, Hasan has, by being rigorous in defining boundaries, made the problem of context tractable. The processes for defining the boundaries are by no means arbitrary, neither, however, are they without difficulty. Essentially it is possible to draw the boundaries anywhere depending, in part, on the functional motivation of the task at hand. This is possible because it is not so much that we need the same boundary to be drawn for every analysis, as it is that we need a clear boundary with reasons for our selection of this boundary, and this makes context the means for selection of a motivated piece of interaction from an ongoing flow of discourse. A text, as well as being located in the here and now, is both reflective and predictive, thus it is tied to what has gone before and makes contact with what is to follow. A text’s existence in the here and now only makes sense because of what has happened in the past and what is planned for the future. If we consider Texts 1 and 2, the artificial boundaries created by selecting a section from the ongoing flow of family discourse mean that many aspects of the context are not recoverable from the texts as presented, and because of the relationship between context and meaning, our understanding of the texts is impaired. As a researcher, I only know this because I collected information about these texts in many different ways. I interviewed the members, I videoed their interaction and recorded their verbal interaction. This rich diversity of data allowed me to triangulate my findings, but to do this I needed to define context in such a way that it was possible to apply to the social action of research itself in order that I could place my different forms of data in relationship to each other. Defining a concept however is never a neutral process. 234 Society in Language, Language in Society

10.4 Deﬁning context and the problem of reiﬁcation

It is perhaps common to all analysis that to study something is to make it a ‘thing’. In referring to context of situation as a ‘tripartite entity’ (Hasan, 2004, p. 21), the implication is that context of situation (as construed by discourse) is an entity and an independent ‘thing’. There are three distinct, yet interconnected, features that push towards reiﬁcation in this case:

(a) Language-driven semantic drift towards reiﬁcation: the representation of context as an entity is in part unavoidable. Once we start to talk about ‘context’, we must, at some point, come to refer to it as an entity since all our language pushes us in that direction. (b) The text-centric or unimodal focus on context: as Hasan (2004, p. 21) says, her view is, ‘context of situation as construed by discourse’. Elsewhere Hasan refers to this relation as a dialectic; however here we get the sense that Hasan’s view is very much one that is built through discourse and reﬂects a text-centric focus (Hasan, 2004). (c) A bounded notion of discourse: a text-centric view of discourse tends towards strong boundaries that make it more likely context becomes seen as an entity.

Although all three of these explanations are connected, it is the text-centric or unimodal perspective that I wish to consider in more detail, primarily because of the impact of this view on the study of natural spoken language and the study of questions that are social in nature.

10.5 Unimodality, multimodality and context: issues for theory and application

Whether or not the motivation is analytical or a construal of reality, in solving the boundary or infinity problem of context, Hasan (2004, p. 21) makes a very clear distinction between social action as a holistic venture and those aspects of social action which relate specifically to discourse, or are ‘construed by discourse’. Hasan argues that: ‘The qualifier of discourse is a reminder that unlike the “cultural activity theory” associated with the Russian, especially (neo-)Vygotskian literature (Engeström et al., 1999), “context theory” was not intended to apply to all kinds of social action, being designed specifically with discourse in mind’ (2004, p. 21). What then are the implications for theory and application of such a view of language? At a theoretical level, context of situation, as Hasan models it, includes the other modalities as part of the context rather than as the discourse. In the wider context of SFL, within which Hasan’s model sits, this means that modalities are present on the expression plane as separate modes Rebekah Wegener 235 of meaning and represented in the contextual plane as part of the context for discourse under mode. As Bowcher (2007, p. 626) suggests, there are definite theoretical consequences to this view of mode.

10.6 Issues for theory: stratiﬁcation and context

The SFL dimensions of language are set out by Halliday and Matthiessen (2004) and reproduced in Figure 10.1. Of these dimensions, it is stratiﬁcation that is most relevant to the issue of multimodality and context. According to Halliday and Matthiessen (2004), stratiﬁcation does not include context (it is not a level of language), yet, context stands in a realization relationship to semantics, which is a level of language. Semantics, according to Halliday (1974) and also to Halliday and Matthiessen (2004), is a realization of patterning of context (see Table 10.1). When we come to an analysis of context, however, it may be that we need to look at the issues of dimensionality afresh. If we are attempting to account for context, not language, but context, our object of analysis has changed and hence, some may argue, so have our dimensions. I want to point out here that I am raising these as options and not necessarily proposing them as solutions. I am in no way convinced that it makes sense to separate out context as a distinct object; on the contrary, I am more inclined to suspect that treating context as an object in its own right leaves the theoretical door open to making context very artefactual as I mention above. It could be that one of the strengths of the SFL approach is that context is not an object in its own right. Rather, within SFL, context is typically seen to be aspects of the culture as they are relevant to meaning; making meaning is very much the centre of attention in any consideration of context. To treat context as an object means considering all of context and this is a concept

Table 10.1 The dimensions (forms of order) in language and their ordering principle from Halliday and Matthiessen (2004, p. 20)

Dimension Principle Orders

Structure Rank Clause ∼ group or phrase ∼ word (syntagmatic order) ∼ morpheme System (paradigmatic Delicacy Grammar ∼ lexis [lexicogrammar] order) Stratification Realization Semantics ∼ lexicogrammar ∼ phonology ∼ phonetics Instantiation Instantiation Potential ∼ sub-potential or instance type ∼ instance Metafunction Metafunction Ideational [logical ∼ experiential] ∼ interpersonal ∼ textual 236 Society in Language, Language in Society that is too vast for an adequate account. If context is treated as the object however, it is necessary to reconsider the dimensions, since the object of analysis has changed. This raises the further question: if we can make context a separate object, can each stratum be considered as an object in its own right? This makes more sense for semantics and context, since they are distinct from the other strata in their nature, being the only two which have a dual usage. If language (i.e. linguistics), rather than meaning (i.e. semiotics), is the object, then we need to consider if our analysis is in some ways at risk of becoming oriented towards structure because language, unlike meaning, lends itself to a focus on the organizational and structural issues. A focus on language as our object of analysis also separates out the different orders of meaning in such a way that their relationship to each other loses focus. If meaning (i.e. semiotics) is our focus then it is important for us to accommodate multiple modes of meaning making. SFL uses a stratified model of language that incorporates the levels of the expression plane (including sound systems – phonetics and phonology, gesture, pixels, etc.),2 lexicogrammar (lexis/-grammar – or wording and structure), semantics (the meaning system) and context (culture and situation – elements of the social structure as they pertain to meaning) (Halliday and Matthiessen, 2004). Other theories have treated the stratification as a singular split between sounds and meaning, thus making the representation of language essentially mono-stratal (Lepschy, 1970). Halliday and Matthiessen (2004) suggest that this mono-stratal representation is an ontogenetic relic in the sense that, while the child’s protolanguage, having no grammar, does resemble this mono-stratal configuration, where the sound represents a meaning directly (or, more accurately, bi-stratal, having a content and expression stratum), the adult language is more complex.3 Within the adult system, both the content and expression plane become stratified (Halliday and Matthiessen, 2004). Thus, we as adults have access to a meaning potential which is able to ‘expand, more or less indefinitely’ (Halliday and Matthiessen, 2004, p. 24), allowing us to variously construe our experience and our social relations. In this way, uniqueness is built in, although it is a very socially oriented uniqueness (hence original rather than individual), since to be shared it must be coded (Hasan, 1999). Halliday and Matthiessen (2004) suggest that it is not only the content plane which is stratified, but the expression plane as well. Thus, where content is split into semantics and lexicogrammar, the expression plane divides into phonology and phonetics. The reason for this they suggest is for ‘sep- arating the organizing function from the function of interfacing with the environment’ (Halliday and Matthiessen, 2004, p. 25). If this is the case, then is it also possible that further stratification is possible at the stratum of context? Do we have an organizing function which is separate from the function of interfacing with the socio-material environment? If so, then we have Rebekah Wegener 237 a stratum which makes contact with the socio-material world (the socio- material setting) and one which organizes this (context). In many respects this is how context of culture and context of situation come to be treated, although the distinction here is apparently one of instantiation (see below). According to Halliday and Matthiessen (2004, p. 26), ‘the stratified linguistic system is “embedded” in context (cf. Halliday, 1978; Halliday and Hasan, 1985; Martin, 1992)’. This may be represented as in Figure 10.1, which is a nested diagram showing the stratification of the linguistic system and its embedding in context. Figure 10.1 represents the nested nature of language, showing what Halliday and Matthiessen (2004) refer to as the phylogenetic function motivating the division between content and expression and the strata of semantics, lexicogrammar, phonology and phonetics. What is also indicated here is that context is a stratum within language as well. Although Halliday (1978) and Halliday and Hasan (1985) both suggest that context is in fact outside language, this distinction is not apparent from this diagram. Certainly context is treated as a stratum within SFL regardless of statements suggesting otherwise. Halliday and Matthiessen (2004, p. 26), for example, suggest that ‘a language is a series of redundancies by which we link our eco-social environment to non-random disturbances in the air (sound waves)’. This construal of the relationship would seem to suggest that there is a directional movement from the culture on one side to the biological on the other. The biological and the contextual are logically part of the same plane as indeed the nested circle diagram represents, and we can refer to this as a contextual plane, where language meets biology on one hand and society on the other. This perspective is akin to taking the representation of stratification

CONTEXT

Content: Semantics

Content: Lexicogrammar

Expression: Phonology

Expression: Phonetics

Figure 10.1 Stratification: reproduced from Halliday and Matthiessen (2004, p. 25) 238 Society in Language, Language in Society as nested circles and stretching it such that context and biology meet. If we were to pull context out in both directions, so from under phonetics and above semantics, we would see biology emerge below phonetics. The particular structure of the nesting hides the biological because the focus is on the socio-material, this being the particular orientation of the theory. If we then take the two ends, context on one side and biology on the other, and represent them as joined together we end up with a perspective as is represented in Figure 10.2. The stretching of the strata in this way is like looking at language from the point of view of context. This is largely a matter of perspective. If we stand with our backs to biology and move towards social context, asking questions about society and social man, then the social context looms large and the expression plane and hence biological context become less significant, producing the representation as seen in Figure 10.1. Movement in this direction emphasizes the special status that the lexicogrammar has since it is the only stratum which has a semiotic stratum above and below it. This suggests it is important in providing language with the power that it has (Halliday, 1974). Turning towards biology on the other hand, the expression plane and hence the biological context are emphasized. Represented in Figure 10.2 is the content plane (lexicogrammar and semantics) and the expression plane (phonology and phonetics). This is set against the functional motivations for the stratification of these two

Organizing Organizing function (Linguistic sign) function Meaning

Lexicogrammar Phonology

Interface Interface Semantics Phonetics function function

Sociocultural/ Physical/ material biological Stratification on the context Meaner context contextual plane Grounding (Person) Grounding (below double lines) function function is the domain of other disciplines e.g. Content Expression sociology, plane plane psychology, anthropology

Figure 10.2 Stratification and the contextual plane representing the union of the social and biological Rebekah Wegener 239 planes; the interface function (semantics and phonetics) and the organizing function (lexicogrammar and phonology) (Halliday and Matthiessen, 2004). Since stratification is possible at the expression plane and the content plane there may be reason to suspect that it is also possible at the contextual plane. So for example, we might see biological context as akin to the expression plane and social context as akin to the content plane with similar functional motivations for stratification in each, albeit with person or the meaner as the central element rather than the linguistic sign. However, I would argue that any detailed consideration of this stratification lies outside the domain of linguistics since, as Halliday (1978) suggests, at context we are dealing with structural statements which represent the most delicate distinctions of others. As linguists, we are involved in research questions which extend beyond the linguistic realm to questions which are primarily located in the contextual plane, and this is particularly true given the focus of SFL on questions which involve looking at language in society and society in language. Although part of the method of addressing such questions will necessarily involve language, given the relation of language to culture (Hasan, 1999), they are not essentially linguistic questions, meaning that they are not within the parameters of linguistic theory, or, more specifically, while a linguist is often called on to bring their linguistic expertise to bear on non- linguistic questions, exploring the non-linguistic is not the job of a linguist and theorizing the non-linguistic is not the job of a linguistic theory. What is necessary is to theorize the relationship of the contextual to the linguistic or at least of how we go outside the linguistic. We also see in the development of a contextual plane the appearance of the linguistic sign between the content and expression plane and the meaner between the socio-material and the biological context. These two related concepts, although intimately connected to stratification, are of a different order. The meaner or personality emerges from the union of the biological and the social since these do not exist independently of each other and neither are they independent of language. As Halliday (1978, p. 14) suggests,

by virtue of his participation in a group the individual is no longer simply a biological specimen of humanity – he is a person. Again language is the essential element in the process, since it is largely the linguistic interchange with the group that determines the status of the individuals and shapes them as persons.

Context is represented in Figure 10.2 as a plane that includes both the biological and the social. Although Halliday (1978) is very explicit about context being outside language and of a different order, typically context has been seen as a stratum within a stratificational account of language. This 240 Society in Language, Language in Society distinction is represented in the diagram referenced where context is seen in relation to, but distinct from, the stratificational organization of the linguistic system. The confusion here has arisen largely, I would argue, because of realization and the fact that there is a realization relation between semantics and context. Halliday (1978, p. 79) explains that the semantic system is a realization of a social semiotic, and while suggesting that there are ‘many different ways of going outside language’ the semantic system is a realization of what the speaker ‘can do’. Realization, a concept central in the systemic paradigm, is a question on which there has been much debate and what one might term ‘theoretical dynamism’ (see e.g. Halliday, 1992; Hasan, 1995, 1996; Lemke, 1984; Matthiessen, 1995; Martin, 1992; Fawcett, 1988; Cloran, 1994; Berry, 1977; Hasan and Fries, 1995; Butt, 1996). Initially discussed independently of stratification, realization has in Halliday and Matthiessen (2004, p. 20) come to be represented as the central organizing principle behind stratification. Hasan (1996, pp. 110–12) discusses the notion that realization may be seen as being both intra- and inter-stratal, with Hasan’s focus in this discussion being on the realization chain.4 Her ideas are summarized in Table 10.2. Realization for Halliday (1992) and Halliday and Matthiessen (2004) is prototypically inter-stratal. Thus, since there is a realization relation between semantics and a social semiotic (Halliday, 1978, p. 79), and realization is inter-stratal, then that social semiotic must be a separate stratum. Halliday is careful here to call it a social semiotic and not context. Having made a sharp distinction between context and the linguistic system, it is crucial to recognize that a linguistic configuration is meaningful because (and only because) it realizes a recognizable and construable configuration of socio-material relations. So the distinction between the contextual and the linguistic is very much an artificial one and one which we make for analytical reasons.

Table 10.2 Types of realization statements from Hasan (1996, p. 111)

Function

Structuring (a) Insert Intra-stratal: relates facts of different order (b) Expand at the same strata, i.e. intra-stratal relations (c) Order Layering (a) Conﬂate Metafunctional: relates facts pertaining to different metafunctions, i.e. calibrates metafunctions Pre-selecting (a) Rank Intra- and inter-stratal: relates facts on one (b) Stratum rank to those on another, i.e. inter-stratal relations Rebekah Wegener 241

10.7 Issues in application: multimodality and context

When we come to application, the problem is a little more of our own making. Clearly, the model as set out by Hasan is not intended for multimodal communication. Yet it is testament to the strength of Hasan’s model (and the wider SFL model in which it sits) that it is not only possible, but fruitful, to use a model intended for essentially unimodal communication to apply to multimodal communication. In applying Hasan’s model of context to situations where communication is multimodal in nature (as most communicative events are), this view of context restricts context of situation to discourse alone. Not only does this restriction cause some problems for modelling alternate forms of communication such as communication through challenging behaviour, augmented or computer-mediated communication, it also biases our understanding, interpretation and analysis of what is going on in a situation. Multimodality has become something of a buzzword, but of course, most texts are multimodal. Technological changes have increased the availabil- ity of texts of different modality and allowed for them to be combined in new ways for a wider audience. These changes have made multimodal texts more accessible for analysis, although they have also changed the assumptions that many models have made about language. For instance, as Bowcher (2007, pp. 627–8) points out, where Hasan (1999) suggests that written texts will be monologic in nature by default, changes in technology mean that this relationship can no longer be a presupposition. What makes analysis of multimodal texts more challenging is that they will often not display the contextual trace in the way that single mode mon- umental5 texts might and this is perhaps a result of the models used for analysis than a feature of the discourse (Bowcher, 2007, 2013; Bartlett, 2013). Bowcher (2007), in particular, displays these limitations through a study of how Hasan’s (1999) Field networks might be applied to multimodal texts. Most human discourse is multimodal and this practice of coding information in multiple ways has the advantage that multi-coding is a fail-safe. If interactants miss a message in one modality they will likely receive the message through another modality if it is coded in multiple ways. Admittedly, this will not be an identical message since the different modalities do not necessarily share the same meaning potential. However, although the meanings created are not identical, at least some modalities will cut across linguistic barriers and other barriers to communication such that, for example, we do not need to speak the same language to understand that someone is upset, happy, angry or sad. As Hasan (1981) suggests, the functional nature of a code predisposes it to certain meanings. Multimodality also has other beneﬁts for humans because it frees up space in meaning making. Just as the modalities can be used to send the same message, they can also be split up and exploited to send different messages. 242 Society in Language, Language in Society

Such multiple messages through different modalities is a common technique in many children’s books. One story will be told through the pictures and another, sometimes contradictory, story will be told through the written words. Such basic rhetorical devices as sarcasm and irony often rely heavily on our ability to use different modalities to convey different meanings and our abilities to read these in others. Indeed, inability to read different messages being conveyed through different modalities forms the basis of a number of disabilities and disorders (see for example a discussion of the link between multimodal communication and early dementia diagnosis in Kipps et al., 2009). Similarly, it is possible to use multimodality to create space to do other things with talk. Consider the example of the family getting ready to leave in the morning in Text 1. Although I (deliberately) only use a single modality transcript here in the recordings of the family, the mother is negotiating lunch preparation with different family members through gesture at the same time as she is carrying out a conversation about the weather, sunlight, theses or travel plans with other family members. This modality multitasking is of course not limited to families and is common in situations where there is a temporal constraint or requirement for multiple tasks to be performed in parallel. The multimodality in Text 1 that is not visible is made very apparent in Text 2 because one of the participants chooses to respond verbally to the non-verbal exchanges. At the end of Text 2 we begin to see responses from B that are incongruous with the rest of the turns from the other two participants. In the exchange produced below, the highlighted responses from participant B are not responses to the spoken discourse, but to a parallel gestural conversation with participant A. In turn 2, participant B rejects the offer of a banana for his lunch, while in turn 5 participant B accepts the offer of savoury biscuits and in turn 9 very readily accepts the offer of chocolate biscuits and requests two for his lunch. Further non-verbal communication in this interaction takes place at the same time as A and C communicate verbally, but what stands out in this section is the verbal responses which mean that A is listening and responding to two participants at the same time in two different modes. The non-verbal communication is not part of the context of a spoken text, but is actually part of the text itself. It is a multi-participant and multimodal text.

Text 2 excerpt: 1C: I’m so tired 2B: nope 3A: So am I 4C: can I go back to bed? 5B: yeah maybe Rebekah Wegener 243

6A: You could stay home if you needed, if you wanted 7C: Not really ...aw look at the light; at the sun 8A: ...mmm. 9B: yeah I love those. Can I have two? Annotation key: 1–9 = Turn A–C = Participant

The same modality multitasking can be seen in surgical theatres (see Moore, in press), emergency rooms or air traffic control. Often surgical teams will talk about seemingly unrelated topics at the same time as they are negotiating complex surgery visually. Such splitting of the semiotic load is more likely to be seen in situations where people are engaged in a shared task with a common goal. Just as language can be social network specific, gestural and visual communication can become very specialized. Complicated codes are developed such that onlookers, as outsiders to the code, often do not know what is going on. Such instances emphasize the importance of other modalities and the material situational setting to meaning making. We are only just discovering the importance of some of these aspects to thinking and meaning making. The analysis of multimodal context is certainly very much more difficult by comparison to a single modality text. The inclusion of the possibility for multiple stories or the further nuancing of a story through other modalities makes analysing such situations very difficult. Take the example of a television news report. There are multiple contexts and multiple reports taking place in such a situation. While as an audience we receive the report as a unified whole framed by the anchor, it is difficult to know how to analyse such a situation. As viewers, we see the whole social process as news reporting and we have certain expectations about how such a process should unfold and what it should contain (although it is possible to argue that these expectations are changing). However, when we consider an individual instance of a news report we see that it is made up of many different interactions. The newsreader is engaged with viewers and with the field reporter. The field reporter is engaged with the people they are interviewing or reporting on and the people they are reporting on are engaged with those around them. Each of these relations represents different fields, tenor relations and potentially different modes, despite the fact that we receive the news story as a whole. Potentially, even the information from each of these interactions is quite different (cf. Lukin, 2014). To a certain extent this can also happen with monumental texts. There are multiple contexts embedded within the one text and it is necessary to analyse all of these to get a complete picture of what is going on. Moore and 244 Society in Language, Language in Society

Wegener (2010) approach this problem in pharmaceutical texts by assessing each context in turn to build a richer picture of the expectations that a reader might have. Moore and Wegener (2010) support this analysis with field studies of how readers reacted to the text and the understandings that they drew from the texts. To return to the relation between theory and application, if we take Halliday’s (1992, p. 61) perspective of looking at anything from the point of view of how it means, we are in the position of viewing all behaviour as potentially meaning bearing, and indeed, all artefacts and even the environment itself (Fawcett, 1990). These different modes of meaning making are all used by humans in their interactions, yet the broad-scale use of different modes by humans does not mean that the different modes have the same meaning potential. As Hasan suggests, ‘despite overlaps, what can be said through the verbal code is not coextensive with what can be said through the gazing code or the gesture code or the code of dress’ (Hasan, 1981, p. 107). Each code carries distinct representational capacities which relate to the means of interaction. So, for example, language has the capacity to tran- scend the here and now with the consequence that its temporal qualities are almost boundless. By comparison, gesture or gaze need some sort of temporal proximity even if this is mediated by technology (for example, technologies such as video-enabled mobile technology have significantly enhanced the communicative potential for many sign language users). These modes, gaze in particular, are heavily oriented towards interpersonal meanings (Hasan, 1981). To see the significance of gesture for interpersonal meaning we need only consider the attempts at iconic representation of aspects of gesture in the form of emoticons. Variability in semantic potential is not new, after all; Bernstein’s studies showed that different individuals do not share the same meaning potential. The distinction between these two, however, is that one refers to individuals using a mode and the other to the modality itself. Not having equal access to the full range of meanings in a mode is distinct from the mode itself having a limited potential. Individuals may not have the same access to the mode, but the mode has the same potential whether we access it or not. Here the situation is that the modes themselves do not have the same potential (cf. Kress, 2003; Kress and van Leeuwen, 2001). If this is the case, and certainly it would seem to be, then certain ideas and concepts will be limited to certain modalities; for example, it is not possible to do abstract concepts, or what Hasan (2001) calls decontextualized language, through gestures that co-occur with speech. The implication of this limit on the meaning-making potential for different modalities is that they will have a different semantic stratum. They will also have different reactances in the contextual stratum, and may not even have a distinct organizing stratum6 (cf. Bowcher and Liang, this volume). Rebekah Wegener 245

10.8 Future focus

Ultimately, when we deal with context we are dealing with language in use. As Halliday (1978) has argued, humans are inherently social and inclined to interact. By virtue of the way humans are, this interaction is inherently multimodal, and of the modalities it is perhaps the visual that is the most dominant. Changes in technology have opened up new opportunities for interaction but more importantly they have opened up new ways to study interaction, as it becomes easier to capture what were once ephemeral interactions. As we develop new ways to capture different sorts of data it becomes more important to be able to analyse the different ways that these modalities interact with each other. Without having been designed or necessarily intended for use on other modalities, SFL and the models which sit within it, such as Hasan’s, provide strong foundations for bringing together the different modalities. This is primarily because they locate their work within a wider frame that views language as a social semiotic. Although currently SFL is a model that locates the different modalities both at the level of context and at the expression plane, this is perhaps not a bad thing since it is a better representation of how the modalities work together in actual interaction. However, to apply this model it becomes necessary to reconcile the dual representation for the purposes of analysis.

Notes

1. Multimodality has had a varied history with respect to definitions and many of the linguistic definitions are somewhat at odds with definitions elsewhere in the sciences. I use the term here to refer to discourse where ‘at least one communica- tor uses at least two different sensory interfaces for sign production or reception’ (Mehler et al., 2012, p. 3). In the case of the texts used, the communication recorded was both verbal and behavioural, thus drawing on the aural and visual sensory systems. 2. The expression plane poses some interesting problems for multimodality and these have been considered in the work of Moore (2004, in press). I take the view that the different modalities may combine in contextually defined and complex ways to form a stratified ‘system’ of meaning making, and that while each modality independently may not have a distinct stratal arrangement that corresponds to the organization of language, within a given context these may work together to create meaning, thus a stratified system. 3. ‘Adult’ refers in this usage to the language system not to the age at which we make use of this system since, as Halliday suggests, ‘children grow up and their language grows up with them. By the age of two and a half or even earlier, the child has mastered the adult language “system”, the framework is all there. He will spend the rest of his childhood – the rest of his life even – mastering the adult “language” ’ (Halliday, 1978, p. 27). 4. The distinction between realization as inter- and intra-stratal relates to slight variations in understanding of realization and the two views are entirely compatible 246 Society in Language, Language in Society

with each other. Thus, while realization is ‘prototypically’ inter-stratal, ‘we often use the term to refer to any move which constitutes a link in the realizational chain, even one that does not by itself cross a stratal boundary (for example, features realized as structures)’ (Halliday, 1992, p. 352). Hence realization is a short- hand for the entire chain which may be both inter- and intra-stratal as can be seen in Table 10.2. The central point for our discussion is that realization is inherent in a stratified representation. 5. Monumental texts have a physicality, defined boundaries, and a fixed and recognizable state in a way that monuments might be considered to have these qualities. The term borrows from the idea that the monument enshrines a specific point in time and it captures the difference between, for instance, a novel or book and more ephemeral texts with less defined boundaries such as lecture notes or an online chat. The concept of a monumental text is distinct from the concepts of product and process and can be usefully combined with the distinction. 6. The extension of this line of reasoning takes us into the area of affordance research as Aronin (2014) has discussed in relation to applied linguistics.

References

Aronin, L. (2014) ‘The concept of affordances in applied linguistics and multilingualism’. In M. Pawlak and L. Aronin (eds) Essential Topics in Applied Linguis- tics and Multilingualism (Second Language Learning and Teaching) (Berlin: Springer), pp. 157–73. Bartlett, T. (2013) ‘I’ll manage the context: context, environment and the potential for institutional change’. In L. Fontaine, T. Bartlett, and G. O’Grady (eds) Choice: Critical Considerations in Systemic Functional Linguistics (Cambridge: Cambridge University Press), pp. 346–64. Berry, M. (1977) An Introduction to Systemic Linguistics 2: Levels and Links (London: B.T. Batsford Ltd). Bowcher, W. (2007) ‘Field in multimodal texts’. In R. Hasan, C. M. I. M. Matthiessen and J. J. Webster (eds) Continuing Discourse on Language (London: Equinox), pp. 619–46. Bowcher, W. (2013) ‘Material action as choice in ﬁeld’. In L. Fontaine, T. Bartlett and G. O’Grady (eds) Choice: Critical Considerations in Systemic Functional Linguistics (Cambridge: Cambridge University Press), pp. 318–41. Butler, C. (2003) Structure and Function: a Guide to Three Major Structural-Functional Theories. Part 2: From Clause to Discourse and Beyond (Amsterdam: John Benjamins). Butt, D. (1996) ‘Theories, maps and descriptions’. In R. Hasan, C. Cloran and D. Butt (eds) Functional Descriptions: Theory in Practice (Amsterdam: John Benjamins), pp. xv–xxxv. Cloran, C. (1994). Rhetorical Units and Decontextualisation: An Enquiry into Some Rela- tions of Context, Meaning and Grammar (Nottingham: Department of English Studies, University of Nottingham). Cook, G. (1990) ‘Transcribing inﬁnity: problems of context presentation’. Journal of Pragmatics 14: 1–24. Engeström, Y., Miettinen, R. and Punamäki, R.-L. (eds) (1999) Perspectives on Activity Theory. Learning in Doing: Social, Cognitive and Computational Perspective (Cambridge: Cambridge University Press). Rebekah Wegener 247

Fawcett, R. P. (1988) ‘Language generation as choice in social interaction’. In M. Zock and G. Sabah (eds) Advances in Natural Language Generation (Vol. 2) (London: Pinter Publishers), pp. 27–49. Fawcett, R. P. (1990) ‘The COMMUNAL Project: two years old and going well’. Network: News, Views and Reviews in Systemic Linguistics and Related Areas 13: 35–39. Halliday, M. A. K. (1974) ‘Michael Halliday’. In H. Parret (ed.) Discussing Language: Dialogues with Wallace L. Chafe, Noam Chomsky, Algirdas J. Greimas, M. A. K. Halliday, Peter Hartmann, George Lakoff, Sydney M. Lamb, André Martinet, James McCawley, Sebastian K. Saumjan, Jacques Bouveresse (The Hague: Mouton), pp. 81–120. Halliday, M. A. K. (1978) Language as Social Semiotic: the Social Interpretation of Language and Meaning (London: Edward Arnold). Halliday, M. A. K. (1992) ‘New ways of meaning: the challenge to applied linguistics’. In M. Putz (ed.) Thirty Years of Linguistic Evolution (Amsterdam: John Benjamins). Halliday, M. A. K. and Hasan, R. (1985) Language, Context, and Text: Aspects of Language in a Social-Semiotic Perspective (Geelong, Vic.: Deakin University Press). Halliday, M. A. K. and Matthiessen, C. M. I. M. (2004) An Introduction to Functional Grammar (3rd edn) (London: Edward Arnold). Hasan, R. (1981) ‘What’s going on: a dynamic view of context’. In J. E. Copeland and P. W. Davis (eds) The Seventh LACUS Forum (Columbia, SC: Hornbeam Press), pp. 106–21. Hasan, R. (1995) ‘The conception of context in text’. In P. H. Fries and M. Gregory (eds) Discourse in Society: Systemic Functional Perspectives, Meaning and Choice in Language: Studies for Michael Halliday (Norwood, NJ: Ablex), pp. 183–283. Hasan, R. (1996) ‘Semantic networks: a tool for the analysis of meaning’. In C. Cloran, D. G. Butt and G. Williams (eds) Ways of Saying: Ways of Meaning: Selected Papers of Ruqaiya Hasan (London: Cassell), pp. 104–32. Hasan, R. (1999) ‘Speaking with reference to context’. In M. Ghadessy (ed.) Text and Context in Functional Linguistics: Systemic Perspectives (Amsterdam: John Benjamins), pp. 219–328. Hasan, R. (2001) ‘Wherefore context? The place of context in the system and process of language’. In S. Ren, W. Guthrie and I. W. R. Fong (eds) Grammar and Discourse: Proceedings of the International Conference on Discourse Analysis (Macau: University of Macau Publication Centre), pp. 1–21. Hasan, R. (2004) ‘Analysing discursive variation’. In L. Young and C. Harrison (eds) Systemic Functional Linguistics and Critical Discourse Analysis: Studies in Social Change (London: Continuum), pp. 15–52. Hasan, R. and Fries, P. (eds) (1995) On Subject and Theme: a Discourse Functional Perspective (Amsterdam: John Benjamins). Kipps, C. M., Nestor, P. J., Acosta-Cabronero, R., Arnold, R. and Hodge, J. R. (2009) ‘Understanding social dysfunction in the behavioural variant of frontotemporal dementia: the role of emotion and sarcasm processing’. Brain: a Journal of Neurology 132: 592–603. Kress, G. (2003) Literacy in the New Media Age (London: Routledge). Kress, G. and van Leeuwen, T. (2001) Multimodal Discourse: the Modes and Media of Contemporary Communication (London: Arnold). Lemke, J. (1984) The Formal Analysis of Instruction and Action, Context and Meaning. Semiotics and Education (Victoria University, Toronto Semiotics Circle Monographs, Working Papers and Prepublications). Lepschy, G. (1970) A Survey of Structural Linguistics (London: Faber and Faber). 248 Society in Language, Language in Society

Lukin, A. (2014) ‘Creating a parallel universe: mode and the textual metafunction in the study of one news story’. In W. L. Bowcher and B. A. Smith (eds) Systemic Phonology: Recent Studies in English (London: Equinox), pp. 53–90. Martin, J. (1992) English Text: System and Structure (Amsterdam: John Benjamins). Matthiessen, C. M. I. M. (1995) Lexicogrammatical Cartography: English Systems (Tokyo, Taipei and Dallas: International Language Sciences Publishers). Mehler, A., Romary, L. and Gibbon, D. (2012) ‘Introduction: framing technical communication’. In A. Mehler and L. Romary (eds) Handbook of Technical Communication (Berlin and Boston: Walter de Gruyter), pp. 1–26. Moore, A. (2004) ‘The discursive construction of treatment decisions in the management of HIV disease’. Unpublished PhD thesis, Macquarie University. Moore, A. (in press) ‘Surgical teams in action: a contextually sensitive approach to modelling body alignment and interpersonal engagement’. In A. Baldry and E. Montagna (eds) Interdisciplinary Perspectives on Multimodality: Theory and Practice (Campobasso, Italy: Palladino). Moore, A. and Wegener, R. (2010) ‘Report on the language of Consumer Medicines Information Leaﬂets’. Draft project report. Wolf, G. (1989) ‘Malinowski’s “context of situation” ’. Language & Communication 9 (4): 259–67. Part V Structure and Texture: Two Kinds of Unity 11 GSP and Multimodal Texts Wendy L. Bowcher and Jennifer Yameng Liang

11.1 Introduction

Hasan has explored the problem of how to identify ‘the factors that are relevant to the construction of the outline of a discourse type’ (1994, p. 127). By outline, Hasan means ‘(a) what elements/parts must/may occur in each and every instance of a particular discourse type; and (b) how these various elements stand in relation to each other’ (Hasan, 1994, p. 128, italics in original). The name she gives to this kind of structure or outline is ‘GSP’ or generic structure potential. Formulating the GSP of a discourse type can provide a descriptive feature relevant to defining a specific register. This can then be used to compare and contrast that register with other registers, and potentially, to generate an instance of a register. When developing the concept of GSP, apart from making some brief comments on advertisements (Hasan, 1981), Hasan has typically focused on linguistic texts including spoken continuous dialogues and nursery tales (Hasan, 1994, 1996, 1999). She has argued that certain features of texts such as their propositional content, dialogic exchanges or identifiable speech acts, are incapable of shedding light on the nature of a discourse type in terms of its overall shape or structure (e.g. Hasan, 1996, p. 56). Rather, in order to gain some idea of the structure of a text, the analyst needs to appeal to the context in which the text is operating and with which the linguistic features of the text are in a realizational relation. This is because GSP is the configuration of features of the context of situation that are in realizational relation with those meanings of a text which relate to a text’s structure. GSP, as the name indicates, is a statement of potentiality. It offers a formula of those elements which are obligatory, optional and/or iterative, and the constraints on those elements in terms of their order of occurrence and/or their variability of occurrence. Thus, it is a statement about how an instantiation of a discourse type may be potentially structured and constructed.

251 252 Society in Language, Language in Society

Over the years, researchers in systemic functional linguistics (SFL) have extended the definition of text to include not only linguistic, but non- linguistic and/or multimodal phenomena. This extension of the definition has also seen an adaptation of many SFL concepts and analytical tools to encompass and analyse these ‘new’ texts. Multimodal texts, too, can be grouped according to text types or registers, such that certain texts categorized as an instantiation of a specific multimodal register bear similarities with other texts belonging to that register. It is not difficult to verify this at an a-theoretical level by just asking anyone to group a selection of texts into ‘advertisements’, ‘recipes’ or ‘knitting instructions’. Can we also extend the concept of GSP to multimodal texts and registers? The aim of this chapter is to explore this question by considering the concept of GSP as developed by Ruqaiya Hasan in terms of its applicability to multimodal texts, and specifically to page-based multimodal texts. That is, if GSP is a statement about the structure of a text type, can this be adapted to making statements about the structure of multimodal text types, and hence to comparing and contrasting the identifying features of one multimodal register with another? What are some of the issues involved in adapting this concept?

11.2 GSP and the unity of a text

According to Hasan, a text can be characterized in terms of two types of ‘unity’: ‘unity of structure’ and ‘unity of texture’ (Hasan, 1985, p. 52). These two types of unity are ‘separate phenomena’ (Hasan, 1985, p. 72). Unity of texture relates to the way meanings hang together in a text and is measured according to the degree of ‘cohesive harmony’. Cohesive harmony concerns the interaction of grammatical and lexical features of a text, and is essentially about textual coherence and within-text interaction; ‘The cohesive devices create texture because they establish relations of meaning’ (Hasan, 1985, p. 96; see also Khoo, this volume). Complementary to the unity of texture is the ‘unity of structure’, which is summed up in the term ‘generic structure potential’. GSP concerns the meaning of the structural elements of a particular genre in relation to its contextual features. It is an analytical abstraction in realizational relation with the semantic and lexicogrammatical features of a text. The term ‘genre’, according to Hasan, is (1) ‘a short form for the more elaborate phrase “genre-specific semantic potential” ’ and (2) the verbal expression of a contextual configuration (1985, p. 108). The structural potential of a text belonging to a particular genre contains a number of elements: obligatory and optional. These elements are organized in a certain way and bound by certain constraints. That is, the GSP ‘formula’ shows the optional and obligatory elements of a genre and the variability and order in Wendy L. Bowcher and Jennifer Yameng Liang 253 which they typically occur. Hasan defines the GSP in terms of the following capabilities:

1. It must specify all those elements of structure whose presence is obligatory, if the text is to be regarded as a complete instance of a given genre by the members of some sub-community; 2. In addition, it must enumerate all those elements whose presence is optional, so that the fact of their presence or absence, while affecting the actual structural shape of a particular text, does not affect that text’s generic status; 3. The GSP must also specify the obligatory and optional ordering of the elements vis-à-vis each other, including the possibility of iteration. (Hasan, 1996, p. 53)

A GSP is represented as a linear ‘formula’, which is exempliﬁed in Figure 11.1.

[(^) Initiating Event^] Sequent Event ^ Final Event [^(Finale)*(Moral)]

Figure 11.1 The GSP of the nursery tale (Hasan, 1996, p. 54)

Figure 11.1 shows the GSP of the nursery tale. The various notations in the formula bear specific meanings. The optional elements are framed by round brackets; in this GSP they are Placement, Finale and Moral. The rest of the elements are obligatory: Initiating Event, Sequent Event and Final Event. Placement and Initiating Event, and Finale and Moral are enclosed together by square brackets and this indicates their co-occurrence. The angled bracket around Placement suggests that it is likely for Placement to be included or interspersed in the Initiating Event. The caret sign ∧ is read as ‘is followed by’ and thus indicates the order of the elements, and an asterisk ∗ marks reversibility, here between Finale and Moral.1 That is, whereas the order between Finale and Moral is reversible, the order between other elements is fixed, with Placement (if there is any) followed by Initiating Event followed by Sequent Event followed by Final Event followed by either Finale (if there is any) or Moral (if there is any). In his discussion of Hasan’s GSP model, Lemke (1988, p. 161) remarks that the obligatory elements feature both the ‘heterogeneity’ and ‘complete- ness’ of a genre. The identification of the obligatoriness and optionality of elements along with their ordering ‘completes this picture’ (1988, p. 161). Yet, there is more to the picture of GSP representation than the specification of elements as well as their order vis-à-vis each other. According to Hasan, the GSP and the specific contextual configuration (CC) relevant to a register are intimately connected. That is, the CC ‘can be used 254 Society in Language, Language in Society for making certain kinds of predictions about text structure’ (Hasan, 1985, p. 56). These are:

1. What elements must occur; 2. What elements can occur; 3. Where must they occur; 4. Where can they occur; 5. How often can they occur. (Hasan, 1985, p. 56, emphasis in original)

Thus, context plays a key role in ascertaining what elements are optional and/or obligatory, what their order is and why, and whether certain elements are iterative or not. Structure can be regarded as a ‘set of relations on a linear sequence of units, such that criteria can be deﬁned for when the structure (a higher-level unit in a constituency hierarchy) has been completed (instanced, realised)’ (Lemke, 1988, p. 159). When examining the generic features in relation to context, the GSP statement is primarily concerned with a potential, a repertoire of possible instances of structural shape. Thus, it is ‘analogous to a system network’ (Hasan, 1996, 2014) in that choices are represented in the formula. An instantiation will exhibit a speciﬁc set of choices from the GSP. In applying the GSP model to multimodal genres, we are concerned with the potential set of choices available within a multimodal genre and how the structure of a multimodal genre may be represented.

11.3 Genre and multimodal texts

In considering the applicability of the concept of GSP to multimodal texts we review some previous work on the genre of multimodal texts. Of particular relevance is Bateman’s (2008) GeM (genre and multimodality) model. The GeM model places emphasis on the central role of genre in examining the meanings of multimodal artefacts. The importance attached to genre derives primarily from its capability of ‘theorising the range of possibilities open to documents’, which includes the degree to which the various resources deployed in different multimodal documents play a role in their make-up and in the construal of meaning (Bateman, 2008, p. 10). Bateman praises approaches to genre description and categorization that favour rigorous linguistic analyses such as those proposed by Swales (1990, 2004), Martin (1992) and the GSP model proposed by Hasan. Such approaches ‘allow predictions to be made’ and ‘reveal more genres than superﬁcially available by inspection or folk-labelling within a given discourse community’ (Bateman, 2008, p. 196). In linguistic texts, the GSP model emphasizes the kinds of constraints on variation. These constraints are ‘considered as a signiﬁcant symptom for an associated genre phase’ (Bateman, 2008, p. 194). With regard to Hasan’s Wendy L. Bowcher and Jennifer Yameng Liang 255 work on GSP, Bateman points out that in developing statements about generic structure, Hasan emphasizes ‘the motivation for distinguishing particular elements from others’, thus highlighting the evidential requirement for claiming the identity of the different stages in a genre (Bateman, 2008, p. 187). As for multimodal texts, Bateman (2008, p. 187) argues:

we also need to be able to find concretely identifiable empirical evidence to motivate particular structures and interpretations rather than others. Only then do we have a foundation sufficiently firm for further theory building.

In his GeM model, Bateman (2008, pp. 225–9) has firstly categorized four ‘regions’ of page-based multimodal documents. They are: (1) linear linguistic texts that follow the ‘rules of typology and print’, (2) photographs shown onto a 2D space (e.g. page and web page), (3) diagrams and charts whose meanings are realized by the co-deployment of elements such as space, connectors, etc. and (4) drawings that follow ‘principles and conventions of schematised graphic representation’. Depending on how boundaries are crossed and how these regions are combined, multimodal texts manifest different kinds of compositional properties. Bateman then proposes various levels of analysis: rhetorical structure, document structure and layout structure. Rhetorical structure refers to the way the argument of a text is structured, and is more of an abstract concept (Bateman, 2008, p. 94). Layout structure is ‘surface-oriented’ as it is ‘explicitly oriented towards the visual make-up of the page’ (Bateman, 2008, p. 122, emphasis in original). Document structure is a concept that can be applied to the multilayering of a multimodal semiotic artefact. For Bateman, understanding the way that a multimodal artefact is structured requires analysis at various levels, disclosing the range of elements deployed, and making statements about how these elements and layers are related. His framework for analysis includes both layout features and semantic content as well as generic, rhetorical and ‘navigational’ elements (Bateman, 2008). This comprehensive way of accessing the structure of a multimodal text seems closer than many ‘surface-oriented’ approaches to arriving at an understanding of how structure and meaning are connected in multimodal texts. The development of the GeM model draws from a variety of multimodal studies such as Waller (1987a, b), who proposes four categories that can be used to organize the typical typographic features: (1) the context of use (i.e. situations and products), (2) format and configuration (i.e. features of medium or materials, and other typographic characteristics), (3) treatment of verbal language and (4) treatment of visual elements (1987a, p. 290). Waller’s analysis is rather enlightening in terms of a consideration of multimodal modes as configurations motivated by certain contexts. In this regard, Bateman proposes the concept of page flow to refer to how verbal and non-verbal modes combine 256 Society in Language, Language in Society within a single page. The process of combination does not mean simple jux- taposition. Rather, it refers to a process where semiotic modes of different kinds closely interact with each other to form a ‘rhetorical unity’ that is con- ducive to the ‘communicative intentions of the document’ (Bateman, 2008, p. 176). This accords with Lemke’s (1998) idea of ‘multiplying meaning’, Royce’s (1999, 2007) framework for ‘intersemiotic complementarity’ which he proposes as a means of explaining coherence across modes, Baldry and Thibault’s (2006, p. 18, emphasis in original) ‘resource integration principle’ which ‘refers to the ways in which the selections from the different semiotic resource systems in multimodal texts relate to, and affect each other, in many complex ways across many different levels of organization’, and O’Halloran’s (1999) concept of ‘semiotic metaphor’. The GeM research conducted under the concept of page flow is in line with these studies in that they all share a certain consensus, that is, meanings in multimodal texts are construed through a configuration of modes of different expressive partialities. However, while there are currently automated ‘spatial constraints’ gleaned from observing and analysing various document types, the obligatory, optional and spatial constraints relevant to multimodal documents have not yet been fully worked out (Bateman, 2008, p. 223). Nevertheless, Hiippala, who applies Bateman’s GeM framework to the analysis of tourist brochures, notes that ‘medium’, in this case ‘the printed page, its method of bind- ing and fold geometry’, ‘determines the physical size and form of the artefact’ (2013, pp. 206–7). Hiippala’s study also makes important head- way into identifying certain patterns of organization within multimodal texts (in this case tourist brochures). He describes these patterns as types of ‘configurations’: those related to ‘text flow’ and those ‘used for integrat- ing graphic elements into the multimodal structure’ of his tourist brochures (Hiippala, 2013, p. 208). But there is still much that needs to be considered in relation to adapting the concept of GSP to multimodal texts as we shall see. In the following section, we discuss a key aspect of GSP – the order of elements. Relevant studies are presented in order to highlight some issues and problems that arise in determining the order of elements in a multimodal text.

11.4 Order of elements in a multimodal genre

Language and images feature distinctive demands and affordances. For instance, Kress (2010) points out that

speech happens in time ...so that sequence in time is one fundamental organizing principle and means for making meaning in this mode ...By contrast with speech, image is ‘displayed’ on a surface in (a Wendy L. Bowcher and Jennifer Yameng Liang 257

usually) framed space. All its elements are simultaneously present. (Kress, 2010, p. 81, emphasis in original)

In a similar vein, Stöckl points out that ‘language is a linear mode that calls for the successive integration of signs into phrases, whereas images are rather based on simultaneous and holistic gestalt-perception’ (2004, p. 17). Thus, while the order of elements for a linguistic genre is linear, which is relatively easy to identify, the order of elements in a multimodal genre is more complex and difﬁcult to determine. And, unlike written language texts which typically follow a linear reading path, multimodal page-based texts are structurally more complicated in that their reading paths and semantic ﬂow vary in relation to different combinations of semiotic modes. Of relevance to the order of multimodal elements on page-based multimodal texts is Kress and van Leeuwen’s (2006) work on the relation between the placement of elements and the compositional meanings construed. Compositional meanings are analysed from the point of view of three main systems: information value, salience and framing. These are explained in the following way:

(1) Information value. The placement of elements (participants and syntagms that relate them to each other and to the viewer) endows them with the speciﬁc informational values attached to the various ‘zones’ of the image: left and right, top and bottom, centre and margin. (2) Salience. The elements (participants as well as representational and interactive syntagms) are made to attract the viewer’s attention to different degrees, as realized by such factors as placement in the foreground or background, relative size, contrasts in tonal value (or colour), differences in sharpness, etc. (3) Framing. The presence or absence of framing devices (realized by elements which create dividing lines, or by actual frame lines) discon- nects or connects elements of the image, signifying that they belong or do not belong together in some sense. (Kress and van Leeuwen, 2006, p. 177)

Kress and van Leeuwen (2006) have shown that different semiotic resources are placed in different zones on a page and their reading orientations do not always match the linear style adopted when reading a piece of written text. Information value, for instance, results from the placement of elements in various ‘zones’, whether they are in the left or right, top or bottom, centre or margin zones. As far as reading orientation is concerned, in cultures where the reading path is from left to right and from top to bottom, the elements placed on the top left tend to be the first to be read and thus considered to be ‘old’ or ‘given’ information whereas those placed on the bottom right are considered ‘new’. However, even in cultures that follow a left-to-right and 258 Society in Language, Language in Society top-to-bottom reading orientation, we cannot say for sure that the first GSP element of a page-based multimodal text is always that which appears on the top left of the page. For one thing, a page within a book or magazine will likely have information positioned according to what may be first seen when the pages are turned (see Berazhny, 2012). An interesting point is raised by Lemke who points out that ‘no visual semiotic can be [linear], for all are at least two-dimensional and any one- dimensional sequence represented in two-dimensions can be accessed at any point at any time’ (1998, p. 95). This, naturally, applies to written texts, as typography is a visual display of language (see Thibault, 2007). In relation to scientific articles, Lemke goes on to say

Many scientists, after reading the title and abstract, may skip to the end to see an endnote telling who supported the research or to skim the references for familiar or unfamiliar citations. They may well read the tables or graphs ﬁrst, and then their captions, and only then the main text. (Lemke, 1998, pp. 95–6)

But just because a reader may flit from one part of the text to another does not mean there is no generalized ‘order’ of elements of the text. There is a ‘default’ order of such texts and this order can be argued on the grounds of semantic flow as well as layout. For instance, part of the reasons why a reader is able to flit from one part of a written text to another is that they are familiar with the relationship between the semantic flow and layout structure of such texts, in this case, an academic article. Thus, Lemke’s (1998) comment really seems to relate to the configuration of features of the visual semiotic of writing and the semantic structure of academic articles coming into play in motivating different readers’ choices. Written texts often display a variety of fonts and sizes and colours of fonts, and these typically play a role in emphasizing the ‘order’ of elements of certain texts. Further, even if one were to analyse only the linguistic aspect of a multimodal text, it would seem that the other ‘decorative’ aspects (van Leeuwen, 2011a) might affect the perception of the ‘linearity’ of such texts and hence of the order of elements of such texts, but that would not necessarily mean the semantic order of elements. As noted earlier, Bateman’s GeM framework, with its various levels of analysis, provides a way of at least determining the ‘rhetorical structure’ of such texts, which he argues may be separate from their layout structure. And Hiippala’s research into the relation between ‘configurations of text-flow’ and configurations of ‘image- text complexes’ (Hiippala, 2013, p. 208) suggests that determining ‘order’ and ‘constraints on the order’ of verbal elements in a multimodal text is problematic. GSP is a linear representation and this reflects the realization of meaning into a linear arrangement in a written text. The ‘visuality’ of a mode of Wendy L. Bowcher and Jennifer Yameng Liang 259 meaning is part of its ‘affordance’ – its material capabilities and features. It is also the expression form of the content – the grammar and meanings that are construing a specific context of situation. The system of salience, which is one of the parameters for examining multimodal compositional meanings, is considered to be useful for determining reading paths and helpful in determining the ‘order’ of importance of elements in a multimodal text. Kress and van Leeuwen posit that ‘regardless of where they are placed, salience can create a hierarchy of importance among the elements, selecting some as more important, more worthy of attention than others’, thus influencing reading paths (2006, pp. 201, 218). The characteristics of salience are realized by factors such as placement in the foreground/background, relative size, contrasts in tonal value (or colour), differences in sharpness, etc. But this principle is brought into question when several elements assume relatively the same degree of salience. Because of the relation between salience and perceived reading paths, it is not surprising that reading paths have been taken into account in discussions of the generic structure of multimodal texts. For instance, in developing a framework for analysing multimodal texts, O’Halloran (1999, pp. 322, 324) says

With multisemiotic texts, the most important stage is a step-by-step analysis of the text through the reading path determined by the choices within different semiotic codes. Signiﬁcantly, the use of multiple semiotic systems means that the potential exists for the construction of unique reading paths. Selections function within each system so that interactions between semiotics become the focal point at different stages. Underlying this method of analysis is the assumption that all forms of semiosis are read syntagmatically. In this manner, the semantic shifts occurring with movements between each semiotic may be documented and the impact of the joint constructions determined.

While O’Halloran’s analysis is highly informative with regard to the various features of the multisemiotic resources utilized in the text she analyses, the ‘reading path’ she proposes is largely interpersonally driven and somewhat debatable, and although there were similarities, it did not entirely reflect our own reading paths of her text. This finding in itself lends weight to her comment that ‘the potential exists for the construction of unique reading paths’ but not quite in the sense that she is stating here. Here, O’Halloran is suggesting that there is a reading path unique to each multisemiotic text, but we would suggest that there may be an array of reading paths for any multisemiotic text (recall Lemke’s (1998) comment about scientists reading scientific articles). There is really no way of gauging the potential reading paths without empirical investigations to reveal what reading paths readers take when encountering a multisemiotic text, and why. That is, while 260 Society in Language, Language in Society analysis of a multisemiotic text may indicate likely reading paths due to features such as gaze, relative prominence (salience), contrast, etc., these cannot necessarily ‘determine’ the reading path that each reader will take. Research by Holsanova et al. (2006) which compares a semiotic analysis of a newspaper with the results from eye-tracking individuals reading those same newspapers finds that while typical patterns of ‘image viewing’ do emerge, the design of a multimodal page, which reflects the creator’s view of what is salient and likely to attract most attention by the reader, may differ from what a reader appears to notice or may view as worthy of spending time on. Berazhny (2012) discusses the concept of reading paths in relation to multimodal magazine texts and some of the ways that an advertiser might manipulate the relation between the materiality of the magazine, the mechanics of ‘turning the page’ and the location of matter likely encountered first, with ease, with difficulty, or with ‘strained intent’. For instance, he notes that when reading a magazine, ‘the reader encounters the right side of the up-coming double page first, and only after that proceeds to the contents on the left side’ (Berazhny, 2012, p. 161). He finds that alcohol advertisers typically place a label used to warn of the dangers of alcohol consumption ‘either on the left or at the bottom of a page, i.e. outside immediate view’ (Berazhny, 2012, p. 161). Furthermore, through using eye-tracking technology to investigate readers’ actual reading paths of multimodal texts, Bateman (2008), for instance, finds that while some of Kress and van Leeuwen’s claims are supported, such as headlines and images attracting attention ‘as expected’, there is no defini- tive support for their claims overall. Bateman argues that ‘reading paths are also ...susceptible both to the knowledge that a reader has concerning how a visual image/page is to be interpreted and to the purpose the reader has in using the image/page’ (Bateman, 2008, p. 64; cf. Lemke, 1998). Indeed, Holsanova comments that ‘the way recipients perceive, understand and remember the content of the multimodal messages is guided partly by visually salient and partly by semantically relevant aspects’ (2012, p. 252). Of interest, too, is that certain visually depicted elements such as faces and eyes may more strongly attract viewers than other features, such as abstract shapes (see Bergh and Beelders, 2014), even though typical indices of salience in terms of size, colour, position, etc. of the depicted faces and eyes may be less significant than other elements of a text. Van Leeuwen observes that reading paths based on salience of elements in a multimodal text are problematic when there are ‘several elements that are more or less equal in salience’ (2005, p. 82). He further explains that although eye-tracking technology may provide an insight into what a reader sees as salient, and how a reader progresses through a text, such methods may not actually point to ‘the structuring of the text, but [to] the structuring of the reading (using) process’, although the elements of the text that readers select ‘must still have semiotic characteristics that will allow them Wendy L. Bowcher and Jennifer Yameng Liang 261 to function as ...specific stages’ of a genre (van Leeuwen, 2005, p. 84). He explains that:

...this reader-activated ‘reading path’ is only possible if the text has elements that can realise the communicative activities of ‘offering a welcome’, ‘providing a product choice’ and so on, and allows them to be read in the above order. In other words, the text becomes an environment facilitating a number of staged, goal-oriented reading processes. However, the text itself is no longer a staged, goal-oriented process. It is an environment for such processes, and must be analysed as a kind of map, a spatial structure allowing a number of trajectories, or as the layout of a building, a spatial structure designed to facilitate a range of speciﬁc activities. (van Leeuwen, 2005, p. 85)

In his study, van Leeuwen (2005) discusses how the concept of genre developed by systemic functional linguists can be modiﬁed when discussing the ordering of multimodal elements. Genres of linguistic texts have unique communicative functions, for example initiating a conversation, providing feedback. These communicative meanings are realized by a succession of communicative moves, which are termed ‘stages’ (cf. Martin, 1992). The concept of stage can be transferable to multimodal genre analysis with a few distinctive characteristics (van Leeuwen, 2005, pp. 76–81). Firstly, unlike linguistic texts, a given multimodal stage can be realized by alternative semiotic modes. Van Leeuwen argues that the choice between visual and verbal mode may depend on the medium or cultural factors. Thus, for a text such as a tourist site entry ticket, in providing additional descriptive meaning of the site, a photograph is more often favoured over an informative text, i.e. a picture is worth a thousand words (see Bowcher and Liang, 2013; Liang, 2013). Another characteristic of multimodal texts is that a given multimodal stage could be realized through more than one mode of expression. This is not unusual or surprising in itself, but how to account for this in a GSP statement is not necessarily straightforward. Van Leeuwen proposes two levels of multimodal genre analysis in order to bring to light the order and the environmental constraints on the order. The ﬁrst level involves a visual analysis to ‘study the environment of the staged, goal-oriented process, and the pathways it allows’ (van Leeuwen, 2005, p. 85). This methodology examines the environmental constraints on the possible reading paths. The second level, according to van Leeuwen, is observational and it involves an ethnographic analysis of the actualized reading paths and user/reader trajectory. A quantitative observation as such is useful to ‘bring out the degree to which different users follow the same trajectories, and hence the degree to which the spatial or quasi-spatial (“virtual”) composition of the environment constrains what can be done in it’ (van Leeuwen, 2005, p. 85). 262 Society in Language, Language in Society

Without extensive empirical investigations of reading paths and their relation to the arrangement of elements in multimodal texts, to the intentions of the designers of multimodal texts, and to the relevant features of the text’s contextual conﬁguration, it is difﬁcult to make any viable claims about reading paths in relation to the generic stages of a multimodal text, which, in any case, should be seen as semantic in nature. Furthermore, for page-based multimodal texts, eye-tracking equipment invariably requires a reader to read a computer screen in order to analyse eye movements and this changes the nature of the medium (cf. Bateman, 2008, p. 65; Berazhny, 2012, p. 160). Thus, one of the biggest issues that we see with regard to exploring the GSP of ‘display texts’ such as static multimodal texts is understanding the concept of order, how to account for ‘order and constraints on order’ of elements, and how to represent order in a formulaic expression. Unlike linguistic texts which are typically arranged linearly, page-based multimodal texts have their elements arranged in a non-linear way, so we need to ask whether investigating the generic structures for static (in this case, page- based) multimodal texts is a viable endeavour at all. Furthermore, no matter the reading path that viewers may take, should the ‘order’ of elements of a multimodal text be contingent on the reading path of the viewer? That is, is GSP tied more to the ‘abstract’ rhetorical meanings of the text than its material qualities (cf. Bateman, 2008), or are there other issues at stake? In the next section, applications of Hasan’s concept of GSP to page-based multimodal texts are considered with a particular emphasis on the issues of order and constraints on the order of elements.

11.5 Applications of GSP to multimodal texts

There have been only a few attempts to represent the GSP of multimodal static (page-based) texts. However, we would argue that these have failed to provide adequate problematization of the issue of ‘order’ and its relationship with meaning and structure. Cheong (2004) and Bortoluzzi (2010), for instance, have considered Hasan’s concept of GSP in relation to page-based multimodal texts. Cheong (2004, p. 176) claims that Hasan ‘establishes “Capture∧Focus∧Justification” as the generic structure for advertisements’. However, her interpretation of Hasan as ‘proposing’ or ‘establishing’ these elements as being in an ‘is followed by’ relation is debatable. Hasan does not present these elements in this format. That is, while Hasan does say for instance, ‘A second obligatory element of structure may be referred to as Focus’ (Hasan, 1996, p. 41), this does not necessarily imply ‘order’ of elements, but may, within the context of her discussion, suggest order of elements in her discussion. She suggests that these elements are ‘obligatory’, but not that they necessarily occur in the order as formulated by Cheong. One could appeal to ‘logic’ and say, ‘well, you need to “capture” a viewer’s attention before getting them to “focus” on what you are “selling” and Wendy L. Bowcher and Jennifer Yameng Liang 263 before setting out your “justification” for selling’. So, it is ‘logical’ to present these in this order. This accords with the view that, for instance, an ‘orientation’ is likely to precede a ‘complication’ in a narrative rather than the other way around. But whether they are in an ‘is followed by’ structural relation is difficult to determine definitively. Nevertheless, Cheong (2004) presents a possible GSP for advertisements as shown in Figure 11.2. Cheong points out that her research indicates that ‘Lead’ and ‘Emblem’ are the only obligatory elements. She uses Hasan’s GSP conventions – items within brackets represent the optional elements that may occur and those without brackets are obligatory elements – and proposes that Lead is the ‘first’ obligatory element ‘where the reader enters the page’ (Bohle, 1990, p. 36). The Lead, she says, is often realized through photographs salient in size, position, colour, etc. Her reason for proposing Lead as the first element in the GSP is that it is the immediate element that attracts and guides the reader into appreciating the whole advertisement. Lead, thus, is given first place in the linear GSP formula because of its ‘standout’ and ‘attract’ qualities which Cheong relates to her interpretation of the meaning it construes within the context.

Lead ∧ (Display) ∧ Emblem ∧ (Announcement) ∧ (Enhancer) ∧ (Tag) ∧ (Call-and-Visit Information)

Figure 11.2 GSP of multimodal print advertisements (Cheong, 2004, p. 164)

For Cheong, the second element that must occur is Emblem, which is the visually presented logo and/or the linguistically realized brand name. As ‘the stamp of authority’ of the product advertised, Emblem serves as a mark of the status and the identity of the product (Cheong, 2004, pp. 173–4). The criterion for its role as the second element does not appear to be related to ‘salience’ but to its function of marking status and identity. Thus it appears that the concept of ‘is followed by’ is conﬂated with ‘salience’, ‘identity’ and ‘obligatoriness’, suggesting somewhat of an inconsistency in the criteria for each feature of the formula. Without further empirical evidence, it is unclear as to whether the order that Cheong proposes can really be represented in a formula using the notation of ∧ (‘is followed by’). Bortoluzzi (2010) applies Cheong’s GSP representation to the analysis of printed advertisements that contain ‘hybrid’ texts. However, like Cheong, while she highlights that a GSP can express the optional and obligatory elements of a text and sets out the elements of an advertisement using the notation ∧ to indicate ‘is followed by’, she does not problematize this key feature of Hasan’s GSP model – the account of the order of occurrence and iteration within this order. Nevertheless, Bortoluzzi is critical of Cheong’s model in a number of ways. For instance, she claims that Cheong’s analysis tends to be too ‘clear-cut’ 264 Society in Language, Language in Society with respect to the categorization of GSP components as ‘mainly visual or mainly verbal and discussed as such in terms of their Ideational meaning’ (Bortoluzzi, 2010, p. 162). Bortoluzzi argues that ‘prototypical print advertisements ...are inherently both [visual and verbal] and subdivisions and analysis of the different components do not entirely capture the complexity of the whole composition in context and in its interaction with the readers/users’ (Bortoluzzi, 2010, p. 162). Further, in Bortoluzzi’s data, which consist of print adverts from Total’s corporate campaign in 2005, ‘Announce- ments and Emblem usually have such visual relevance that they can hardly be considered mere “linguistic components” [and] ...[Announcements] could be interpreted without reading any of the linguistic components apart from the Emblem’ (Bortoluzzi, 2010, p. 162). To sum up, while Cheong and Bortoluzzi propose a GSP for page-based multimodal texts using the formulaic conventions proposed by Hasan, their analyses reveal that their focus is largely on what their data reveal are obligatory or optional elements in the texts. Neither Cheong nor Bortoluzzi problematizes the issue of order of elements and constraints on this order in multimodal texts, nor on how best to represent this in a formulaic expression. Therefore, how to represent the order and the constraints on the order of multimodal structural properties remains a key issue that needs to be addressed. In the following section, we discuss this issue focusing on a speciﬁc type of multimodal text – tourist site entry tickets.

11.6 A GSP of tourist site entry tickets

11.6.1 Specifying the elements Hasan (1996, pp. 53–4) has pointed out that a GSP statement is ‘analogous to a system’ by virtue of the fact that it presents the potential choices open to the instantiation of a given text. The structure of an individual text ‘is just one possible instantiation of some particular path allowed by the GSP’ (Hasan, 1996, p. 53). A text deemed to be ‘complete’ must realize the obligatory elements of the GSP, and presumably in the order in which the GSP statement proposes them. Let us consider one type of multimodal text: a tourist site entry ticket. Based on a database of 320 paper-based entry tickets to tourist sites in mainland China2 from the 1980s to 2014, a brief analysis is presented here as part of a discussion of the GSP representation of multimodal texts and the issues involved. From the database, we ﬁnd that the tickets may have something printed on both sides of the page, but not all do. So laying aside, for the moment, the argument of what constitutes the ‘back’ of an entry ticket and its ‘front’ (see Bowcher and Liang, 2013 and Liang, 2013 for a discussion of this) we now focus on what we consider to be the ‘front’ of the tickets in the database. It is on this ‘front’ side that we ﬁnd features that function to ‘identify’ the tourist site to which the ticket can be used as a legitimate ‘entry’ Wendy L. Bowcher and Jennifer Yameng Liang 265

Figure 11.3 A tourist site entry ticket to Hua Shan (with ‘Enhancer’ circled) pass. Identification is realized by both language and image. In fact, across the database, we find that all of the images represent scenery or objects from the site, and these images are always accompanied by the title of the tourist site, which is often verbally or verbo-pictorially presented. However, in the ticket shown in Figure 11.3, a viewer who has never been to this site could barely tell from the photograph what this place is. The image seems rather generic, showing pine trees, mist, snow and the vague outline of steep mountains in the background. Here, the title of the site, Hua Shan (Hua Mountain) verbally printed on the left side of the image, serves as the ‘anchor’ of the image and ‘directs the reader through the signifieds of the image’ and ‘remote-controls him towards a meaning chosen in advance’ (Barthes, 1977, p. 40). In other words, the image and the title together identify the tourist site. Although it should be noted that it is possible to omit the image and simply have the title Hua Shan, all of the tickets in the database include an image (or images) with the title. Therefore, it would appear that based on our database, an obligatory element in the structural potential of tourist site entry tickets is ‘Identifier’, and this element is realized by the integration of specific verbal and visual choices. Another obligatory feature of the tourist site entry tickets is ‘Authorizer’. This element serves to identify the authenticity of the ticket as capable of being used as a licence to enter the tourist site. The element of Autho- rizer appears to be realized by a range of elements. One element is the serial number, which can be either/both the taxation serial number or/and the manufacturing serial number of the ticket. The taxation and manufacturing serial numbers actually signify two distinct authorities. On the one hand, the taxation serial number represents the authority that the local taxation bureau exerts on the tourist site in that the former grants the permission to allow the latter to participate in business activities.3 On the other hand, the manufacturing serial number may serve to authorize the 266 Society in Language, Language in Society tourist’s entry by assigning each ticket a one-of-a-kind number. Another element that exerts authority is the time stamp. This feature is shown in the image in Figure 11.3. The time stamp validates and also limits the visitors’ entry on a particular day or within a given time span. Thus, we find two obligatory elements realized by several features displayed on the ticket. The overall and distinct function of a tourist site entry ticket – its function as a licence to enter a tourist site – is enabled by a ticket’s configuration of features in relation to the specific site at which it is used. It is most likely for this reason that tourist site entry tickets display varying multimodal features, but realize few obligatory elements. There appears to be something akin to what Stöckl (2004) refers to as a ‘hierarchy of modes’ at work in a multimodal text, but not in quite the same way as Stöckl describes. That is, Stöckl discusses ‘core modes’ such as ‘image’ or ‘language’ from a textual/compositional point of view. Core modes are comprised of ‘medial variants’, such as ‘static’ or ‘dynamic’, and these in turn are comprised of sub-modes such as ‘colour’, ‘size’, ‘volume’, etc. While Stöckl’s description also applies to tourist site entry tickets, and no doubt to multimodal texts in general in that ‘the text is the locus where all modes, sub-modes and features are realised’ (Stöckl, 2004, p. 15), what we are suggesting in our description is that one overall function, or generic stage, may be a ‘core stage’ that can be comprised of one or more realizations depending on the circumstances of the artefact’s use. That is, one ticket may not have a date stamp, but that may be because it is unnecessary to have this kind of authorization in order to enter a specific site. The ticket may not have a serial number, but that form of authorization, too, may be construed in some other way on the ticket, or may be construed in a physical way such as tearing off a portion of the ticket as one enters the site. So different sites may require different forms of authorization that deem a ticket ‘authentic’ and hence eligible to be used as a licence to enter the site. The ticket is thus part of the paraphernalia connected with the macro process of entering/visiting a tourist site. One could make an analogy with a baseball. We could describe the various features of baseballs from a database of different baseballs, but their role is evident when they are ‘put to use’ within a social process, in this case, one of playing a game of baseball. Like a baseball, a tourist site entry ticket is the essential accoutrement of the social process of visiting a tourist site. Thus, by categorizing the features of the tickets in terms of their functional role in relation to the sites, we find that there is an array of features of the ticket, visual, verbal and numerical, serving to realize the obligatory elements of the genre. Across our database, we find that there are two optional elements. These are Enhancer and Attributor. The Enhancer refers to those feature(s) that add further information about the site. In Figure 11.3, the Enhancer is Wendy L. Bowcher and Jennifer Yameng Liang 267

Figure 11.4 A tourist site entry ticket with the Attributor circled circled. It is verbally realized and reads: , (the inherent qualities of the nation; the roots of the Chinese civilization). Attributor, on the other hand, indicates both the classiﬁcation of the ticket and points to the requisite identity of the holder of the ticket, for example tickets for adults/children/student/senior citizens, off-season tickets, tickets for package tours, etc. In Figure 11.4, the Attributor is circled. It reads , identifying this ticket as one that is for tourist groups. The attributes of the holder of the ticket need to ‘match’ the classiﬁcation of the ticket. That is, visitors holding this ticket should enter the site as a group and be accompanied by the tourist guide.

11.6.2 Representing order Up to this point we have been able to identify what is obligatory and what is optional in the tourist entry tickets. But we are still left with the issue of ‘order’ and how to account for this in a multimodal text. We have already noted that multimodal texts present a type of ‘environment’. Social roles, participants, attributes, etc. are ‘depicted’ simultaneously rather than in a linear arrangement (see Kress, 2010, Kress and van Leeuwen, 2006, and van Leeuwen, 2005 discussed in section 11.4). How meanings are depicted depends on a number of factors, including the affordance, or the materiality, of the medium. The sizes, shapes and material forms of the tourist site entry tickets in our database play a role in how much verbal text can be placed on the ticket, and how much detail an image contains (cf. Hiippala, 2013). Within our database, the types of materials used for tourist site entry tickets include paper (both hard and soft) and plastic. The size of tickets appears to be carefully chosen, with the length ranging from 15 to 20 cm and the 268 Society in Language, Language in Society width from 5 to 10 cm. Therefore, the ticket is convenient for the visitor to carry and also for the site officials, or other types of validating apparatus, to check. As Kress and van Leeuwen point out, new meanings are imposed on the artefacts by the ‘physical processes of articulation and the physical qualities of the material used’ (2001, p. 5). The physical characteristics of form enter the realm of meaning and have a bearing on the kinds of meanings the artefacts construe. One can also see a pattern in the arrangement of the linguistic and visual features. In the two tickets shown here (see Figures 11.3 and 11.4), we can see that both images are framed and their shape, which is a rectangular one, echoes the shape of the ticket. The name of the site is located on the left of the ticket in both cases, too. This is not always the case, but it is a common and prolific pattern of arrangement across the database. This arrangement, too, is largely determined by the affordance of the medium. However, we would argue that another key determiner of the arrangement of features on the tickets is aesthetics. And this is perhaps the most difficult of all to measure and to investigate, without having a clear understanding of where it derives from. We are not suggesting that investigating the reasons why some arrangement may be considered more aesthetically pleasing than another is an insurmountable problem, but it seems far from possible to make any reliable or intersubjectively verifiable claims about this at this point, and it is, in any case, outside the scope of this chapter. As already noted, the order of the multimodal features in a multimodal text is non-linear and, perhaps, to a considerable extent, dependent on ‘design’ and ‘aesthetics’ (see Kress and van Leeuwen, 2001; Stöckl, 2004). Nevertheless, as Stöckl points out,

All modes need to signal their internal structure as keys to the retrieval of portions or layers of meaning ...Syntax, then, in its widest sense is a core feature of any accumulation of signs, both with respect to the internal structure of one mode as well as with respect to the overall structure of a multimodal text. (2004, pp. 25–6, emphasis in original)

But the problem of how to interpret the ‘syntactic’ structure of a multimodal text is yet to be resolved. Another key issue in developing a GSP for a multimodal text concerns ‘realization’. Van Leeuwen, as we noted earlier, observes that a multimodal text is like an ‘environment’ wherein it is possible to have clusters of genres, and yet, as he also observes, the modes of meaning employed in a multimodal text may be ‘so utterly intertwined with one another that they no longer make sense on their own’ (van Leeuwen, 2011b, p. 176). Thus one problem in determining the GSP elements is understanding how they are realized. With reference to ‘realization’, Hasan makes the following statement: Wendy L. Bowcher and Jennifer Yameng Liang 269

I suggest that the essential attributes of ‘the structurally important units’ of any text type will have to be stated in semantic terms. But this in itself is not sufﬁcient ...For this strategy to be successful, we also need a model of language description, which can be used for making non-ambiguous statements about the realisation of the semantic attributes by reference to which the structurally important units of text types can be identiﬁed. Thus any discussion of the realisation of textual structures will involve at least three types of abstraction. (Hasan, 1996, p. 58)

The three types of abstraction that Hasan notes are the functional elements themselves, their ‘crucial semantic attributes’ and the ‘lexicogrammatical patterns capable of realising’ these semantic attributes (Hasan, 1996, p. 58). Further, the lexicogrammatical patterns may either explicitly or implicitly realize the semantic attributes. How elements of the GSP of a multimodal text are realized requires an understanding of the different units of analysis involved. For the semiotic system of language, we have a description of those units of analysis, and therefore, we can make more certain, unambiguous statements about the linguistic realization of the GSP features of a multimodal text. But a survey of much of the research devoted to understanding multimodality shows that there is still no agreed-upon description of the units of analysis, and this is problematic for describing the strata involved in realization. In the tourist site entry tickets themselves, one could argue that the image of the site plays a role in both the obligatory element of ‘Identiﬁer’ and the optional element of ‘Enhancer’. But then, maybe it is not the image as a whole, but features of the image that play different roles in these different GSP elements. Understanding how visual elements may realize GSP features explicitly or implicitly may also be important in understanding the realizational relationships involved in a multimodal text (see Hasan, 1996, pp. 58ff.). At present, without an understanding of the nature of the unit of analysis in visual elements such as pictures and other decorative features used in a multimodal text, the assignment of this or that feature as realizing this or that generic stage remains problematic and open to ambiguity (cf. Hasan, 1999, p. 58). Another issue that perhaps bears mention is that multimodal texts are tied to their context of creation in a different way than linguistic texts. That is, the design and aesthetic practices that are brought to bear on the arrangement of semiotic resources utilized within the text along with the speciﬁc environmental and social conditions in which the text will play a role, its ecosocial environment, are crucial to how the text is eventually realized (see Liang and Bowcher, in preparation). How features of the context of creation play a role in the context of situation of the text is, however, beyond our present theoretical and analytical capabilities. A tourist site entry ticket, as a multimodal text, is recognizable from other texts that may be used in that environment, such as tourist brochures. 270 Society in Language, Language in Society

We can make this judgement based on our general understanding of the repertoire of tourist-related texts, and from our participation within society at large. Thus, given the current theoretical and analytical limitations as noted in this chapter, we can, nevertheless, present a preliminary GSP of a tourist site entry ticket. This is presented in Figure 11.5.

TSET-GSP: Identiﬁer ∗ Authorizer ∗ (Attributor ∗ Enhancer)

Figure 11.5 A tentative GSP of the register of tourist site entry tickets (TSET)

The GSP formula is not able to state the elements in terms of an ‘is followed by’ relation, but it does state the obligatory and optional elements. However, what must be kept in mind is that not representing a speciﬁc ‘order’ does not mean that elements of the GSP occur randomly in any instantiation. Rather, the formula recognizes that we have a considerable distance to go before being able to represent ‘order’ of elements in a multimodal page-based text such as a tourist site entry ticket. It may be that ‘order’ in relation to multimodal phenomena needs to be viewed in a different way, perhaps as a prosody or as pulses of semiotic detail. Further, van Leeuwen’s notion that there are genres embedded within the ‘space’ of a multimodal text, suggests an interesting avenue of exploration of the formulation of a GSP of multimodal texts. Unfortunately, discussion and investigation of these ideas must be left to another time. The multimodal text GSP proposed in this chapter serves to represent a more ‘rhetorical’ model which features a conﬁguration of functional elements, albeit with some ambiguity in the way in which the non-linguistic and non-numerical features realize these functional elements. However, to resolve this ambiguity requires one to understand the units of analysis of the non-linguistic matter of the multimodal text in question.

11.7 Conclusion

This chapter began by posing the question of whether or not the concept of GSP could be adapted to account for the structure of multimodal texts. Our interest in exploring an answer to this question arises from the fact that GSP provides a statement of the obligatory and optional elements of the structure of a text and the order and constraints on the order of those elements, and is thus able to be used as a means of discriminating among various registers – showing which are related and which are not within registerial repertoires. Intuitively, we know that multimodal texts are ‘organized’ and display ‘structure’, and we can recognize different ‘types’ of multimodal texts, which indicates that there are indeed registers of multimodal texts. Therefore, to explore the registers of multimodal texts and their organization Wendy L. Bowcher and Jennifer Yameng Liang 271 through the lens of GSP seems a fruitful way to increase our understanding of the structural characters of various multimodal texts. Of crucial importance in adapting the concept of GSP, however, is acknowledging that texts are semantic constructs and as such they are in realizational relation with their context of situation. With a multimodal text, however, it seems easy to focus on the layout features of the text and to background the fact that the multimodal text is not simply made up of pictures and marks on the page but is a construct of meaning. Bateman (2008) also raises this concern and separates out the various ‘layers’ of a multimodal text – layout, rhetorical, navigational – in his attempt to arrive at an understanding of the structure of a multimodal text. Therefore, in an effort to interpret multimodal texts as semantic units, a major issue that has arisen in our discussion is that defining the GSP of a specific multimodal register requires understanding the realizational relationship between the elements of GSP and the semantic and modal conventions through which the elements are realized. To say that in some situations generic stages are multimodally construed, is not new. Many researchers have made this observation (e.g. van Leeuwen, 2005; Ventola, 1987). In discussing face-to-face interaction, Hasan observes that speaking may be ‘only one aspect ...and may not be easily intelligible in dissociation from the working of other modalities’ (Hasan, 1999, p. 241). Halliday et al. (1964, p. 92) also observe that ‘utterances often integrate with other non- linguistic activity into a single event’. However, notwithstanding various proposals regarding ‘grammars’ of non-linguistic semiotic systems (e.g. Kress and van Leeuwen, 2006; O’Toole, 2010), without an agreed-upon understanding and description of the units of analysis for semiotic systems other than language (a discussion too involved to include here, but see Bateman, 2008; Stöckl, 2004; Abousnnouga and Machin, 2010, for some discussion of this issue in relation to multimodal texts), it is difficult to make any statements about relationships between specific units of a multimodal text and the text’s GSP. For the same reason it is also difficult to arrive at a description of the ‘order’ of the GSP elements and the constraints on that order. Further, in applying the concept of GSP to multimodal texts, we would argue that we need to understand the way in which the text is operating and ‘responsive’ to its environment, as GSP is derived from the contextual configuration relevant to a specific register. For texts which are intimately connected with a social process, in the sense that they are a requirement for the fulfilment of that process, such as tourist site entry tickets, the whole multimodal text, its ‘communicative gestalt’ (Stöckl, 2004, p. 19), is ‘embedded’ in a ‘pragmatic environment’ (Hasan, 1996, p. 51). It is this environment, its ‘ecosocial environment’ (see Liang, 2013, for a discussion of this; see also Liang and Bowcher, in preparation) that bestows meaning to the text itself – its role, value, function and affordance. Until these kinds 272 Society in Language, Language in Society of issues are resolved, the full tool power of GSP in relation to multimodal texts remains effectively dormant.

Notes

1. Hasan uses a ‘raised dot’ rather than an asterisk to indicate reversibility between elements. 2. Mainland China does not include the Province of Taiwan, Hong Kong Special Administrative Region and Macau Special Administrative Region. 3. Under the current tourism system in China, the design, manufacturing and selling of entry tickets must be under the supervision of the local taxation bureau.

References

Abousnnouga, G. and Machin, D. (2010) ‘Analysing the language of war monuments’. Visual Communication 9 (2): 131–49. Baldry, A. and Thibault, P. (2006) Multimodal Transcription and Text Analysis: a Multimodal Toolkit and Coursebook with Associated On-Line Course (London: Equinox). Barthes, R. (1977) Image, Music, Text (London: Fontana). Bateman, J. (2008) Multimodality and Genre: a Foundation for the Systematic Analysis of Multimodal Documents (Basingstoke, Hampshire: Palgrave Macmillan). Berazhny, I. (2012) ‘Alcohol advertising in Aeroﬂot’. In Wendy L. Bowcher (ed.) Multimodal Texts from around the World: Cultural and Linguistic Insights (Basingstoke, Hampshire: Palgrave Macmillan), pp. 146–72. Bergh, L. and Beelders, T. (2014) ‘An eye-tracking account of reference points, cognitive affordance and multimodal metaphors’. In A. Maiorani and C. Christie (eds) Multimodal Epistemologies: Towards an Integrated Framework (London: Routledge), pp. 13–27. Bohle, R. (1990) Publication Design for Editors (New Jersey: Prentice-Hall). Bortoluzzi, M. (2010) ‘Energy and its double: a case-study in critical multimodal discourse analysis’. In E. A. Swain (ed.) Thresholds and Potentialities of Systemic Functional Linguistics: Applications to Other Disciplines, Specialised Discourses and Languages other than English (Trieste: Edizione Universitarie Trieste), pp. 158–81. Bowcher, W. L. and Liang, J. Y. (2013) ‘Chinese tourist site entry tickets: intersemiotic complementarity in an ecosocial environment’. Social Semiotics 23 (3): 385–408. Cheong, Y. Y. (2004) ‘The construal of ideational meaning in print advertisements’. In K. O’Halloran (ed.) Multimodal Discourse Analysis: Systemic-Functional Perspectives (London: Continuum), pp. 163–95. Halliday, M. A. K., McIntosh, A. and Strevens, P. (1964) The Linguistic Sciences and Language Teaching (London: Longman). Hasan, R. (1981) ‘What’s going on: a dynamic view of context’. In J. E. Copeland and P. W. Davis (eds) The Seventh LACUS Forum (Columbia, SC: Hornbeam Press), pp. 106–21. Hasan, R. (1985) Part B in M. A. K. Halliday and R. Hasan, Language, Context and Text: Aspects of Language in a Social Semiotic Perspective (Geelong, Vic.: Deakin University Press). Hasan, R. (1994) ‘Situation and the deﬁnition of genres’. In A. D. Grimshaw (ed.) What’s Going on Here? Complementary Studies of Professional Talk (Vol. 2 of the Multiple Analysis Project) (Norwood, NJ: Ablex), pp. 127–72. Wendy L. Bowcher and Jennifer Yameng Liang 273

Hasan, R. (1996) ‘The nursery tale as a genre’. In C. Cloran, D. Butt and G. Williams (eds) Ways of Saying: Ways of Meaning: Selected Papers of Ruqaiya Hasan (London: Cassell), pp. 51–72. Hasan, R. (1999) ‘Speaking with reference to context’. In M. Ghadessy (ed.) Text and Context in Functional Linguistics (Amsterdam: John Benjamins), pp. 219–328. Hasan, R. (2014) ‘Towards a paradigmatic description of context: systems, metafunctions and semantics’. Functional Linguistics 1: 9. doi: 10.1186/s40554-014-0009-y. Hiippala, T. (2013) ‘Modelling the structure of a multimodal artefact’. Unpublished PhD dissertation, Faculty of Arts, University of Helsinki, Finland. Holsanova, J. (2012) ‘New methods for studying visual communication and multimodal integration’. Visual Communication 11 (3): 251–7. Holsanova, J., Rahm, H. and Holmqvist, K. (2006) ‘Entry points and reading paths on newspaper spreads: comparing a semiotic analysis with eye-tracking measurements’. Visual Communication 5 (1): 65–93. Kress, G. (2010) Multimodality: a Social Semiotic Approach to Contemporary Communica- tion (London: Routledge). Kress, G. and van Leeuwen, T. (2001) Multimodal Discourse: the Modes and Media of Contemporary Communication (London/New York: Arnold/Oxford University Press). Kress, G. and van Leeuwen, T. (2006) Reading Images: the Grammar of Visual Design (London: Routledge). Lemke, J. (1988) ‘Text structure and text semantics’. In E. Steiner and R. Veltman (eds) Pragmatics, Discourse and Text: Some Systemically-Inspired Approaches (London: Pinter), pp. 158–70. Lemke, J. (1998) ‘Multiplying meaning: visual and verbal semiotics in scientiﬁc text’. In J. R. Martin and R. Veel (eds.) Reading Science: Critical and Functional Perspectives on Discourses of Science (London: Routledge), pp. 87–113. Liang, J. Y. (2013) ‘Analysing multisemiotic artefacts in an ecosocial environment: a focus on Chinese tourist site entry tickets’. Unpublished PhD thesis, Sun Yat-sen University, China. Liang, J. Y. and Bowcher, W. L. (in preparation) ‘From licences to mementos: the ecogenesis of tourist site entry tickets’. Martin, J. R. (1992) English Text: System and Structure (Amsterdam and Philadelphia: John Benjamins). O’Halloran, K. (1999) ‘Interdependence, interaction and metaphor in multisemiotic texts’. Social Semiotics 9 (3): 317–54. O’Toole, M. (2010) Language of Displayed Art (2nd edn) (London: Routledge). Royce, T. (1999) ‘Visual–verbal intersemiotic complementarity in The Economist magazine’. Unpublished PhD thesis, University of Reading. Royce, T. (2007) ‘Intersemiotic complementarity: a framework for multimodal discourse analysis’. In T. D. Royce and W. L. Bowcher (eds) New Directions in the Analysis of Multimodal Discourse (Mahwah, NJ.: Lawrence Erlbaum (now distributed through Routledge)), pp. 63–109. Stöckl, H. (2004) ‘In between modes: language and image in printed media’. In E. Ventola, C. Charles and M. Kaltenbacher (eds) Perspectives on Multimodality (Amsterdam and Philadelphia: John Benjamins), pp. 9–30. Swales, J. (1990) Genre Analysis: English in Academic and Research Settings (Cambridge: Cambridge University Press). 274 Society in Language, Language in Society

Swales, J. (2004) Research Genres: Explorations and Applications (Cambridge: Cambridge University Press). Thibault, P. (2007) ‘Writing, graphology and visual semiosis’. In T. D. Royce and W. L. Bowcher (eds) New Directions in the Analysis of Multimodal Discourse (Mahwah, NJ.: Lawrence Erlbaum (now distributed through Routledge)), pp. 111–45. van Leeuwen, T. (2005) ‘Multimodality, genre and design’. In S. Norris and R. Jones (eds) Discourse in Action: Introducing Mediated Discourse Analysis (London: Routledge), pp. 73–94. van Leeuwen, T. (2011a) ‘The semiotics of decoration’. In K. O’Halloran and B. Smith (eds) Multimodal Studies: Exploring Issues and Domains (London: Routledge), pp. 115–31. van Leeuwen, T. (2011b) ‘Rhythm and multimodal semiosis’. In S. Dreyfus, S. Hood and M. Stenglin (eds) Semiotic Margins: Meanings in Multimodalities (London: Con- tinuum), pp. 168–76. Ventola, E. (1987) The Structure of Social Interaction (London: Pinter). Waller, R. (1987a) ‘The typographical contribution to language: towards a model of typographic genres and their underlying structures’. Unpublished PhD dissertation, University of Reading, Reading, UK. Waller, R. (1987b) ‘Using typology to structure arguments: a critical analysis of some examples’. In D. Jonassen (ed.) The Technology of Text (Vol. 2) (Englewood Cliffs, NJ.: Educational Technology Publications), pp. 105–25. 12 Construing Instructional Contexts Carmel Cloran

12.1 Introduction

What is the semantic ‘shape’ of the instructional context construed by mothers in interaction with their preschool-age children, and how is it recognized? These questions describe, informally and succinctly, the issues addressed here. It may be said that in the life of the individual cultural initiate (the young child) all contexts are contexts for learning. The question may then be asked: what is meant by ‘instructional’ context and what is distinctive about it? The instructional context was postulated by Bernstein in the 1960s as one among four generalized contexts of primary socialization of the child in the family. Each of these four contexts, though interrelated, has its own distinctive semantic base. Bernstein defined the instructional context as that context in which ‘the child learns about the objective nature of objects and persons, and acquires skills of various kinds’ (1971, p. 206). In later writings Bernstein distinguished the ‘local’ instructional context of the home from the institutionalized or ‘official’ instructional context of the school, and in the process specified how these two sites of instruction are related to one another. In particular, he identified as a characteristic feature of the discourse of the ‘official’ instructional context the use of decontextualized language, i.e. language that is not directly related to any specific material base. Such decontextualized language use may also characterize the discourse of the ‘local’ instructional context. This occurs in those homes where the pedagogic practice – selection, transmission and evaluation of knowledge and skills – is consonant with that of the school, i.e. where ‘the “official pedagogic practice” dominates the “local pedagogic practice” ’ (Bernstein, 1990, p. 179). The semantic shape of a context refers to the structure of texts realizing a particular context – their generic structure potential (hereafter GSP) (Hasan, 1978). In this chapter a number of concepts developed by Hasan are applied to some instances of both the local instructional context as well as

275 276 Society in Language, Language in Society the official instructional context, i.e. home instruction and school instruction, revealing their similarities as well as their differences. A proposal for the structure potential of the official instructional context will be outlined in section 12.3.1 together with some discussion of its linguistic features. Section 12.3.2 will outline the structure potential of the local instructional context. As a preliminary to this discussion, the linguistic characterization of instructional discourse will be briefly outlined in section 12.2.

12.2 Linguistic features of instructional discourse

Taking Bernstein’s characterization of ‘official’ instructional discourse as a starting point in the investigation of the discourse of the instructional context in the home, transcripts of mother–child talk1 were scanned for decontextualized language use, applying a linguistic framework developed for the purpose (Cloran, 1994). Thus, an instructional context was operationalized as one in which some information was given by the ‘instructor’ (mother) which was generalized and thereby had the potential to be related to a non- specific non-local material base. Such discourse is, prototypically, constituted by a linguistic unit known as Generalization (Cloran, 1994). A Generaliza- tion is constituted by a message2 (Hasan, 1983) which is characterized by reference to: (a) a central entity (hereafter CE) which is class-exhaustive, and (b) an event or state which is habitual or timeless. Habituality or timelessness of a message’s event or state may combine with a CE which is non-class- exhaustive. Such a combination construes the rhetorical unit Account when the CE refers to a situationally (and therefore perceptually) absent object or person, and an Observation when the CE refers to a co-present object or (non-interacting) person. Central entity reference in the case of Account and Observation is specific so that language use is more directly related to some material base. On a cline these categories of rhetorical unit may be represented as shown in Figure 12.1. Study of the rhetorical unit3 structure of instructional contexts construed by mothers and children (and indeed by teachers and students) revealed other characteristic features of such discourse. These refer more to the function of rhetorical units within the context of the instructional texts rather

Observation Account Generalization

Direct Indirect

Relation to material base

Figure 12.1 Relation of RU classes to a material base Carmel Cloran 277 than to features of messages constituting rhetorical units. Informally, these functions may be described as expanding upon information in order to (i) clarify and/or define; (ii) provide reasons; and (iii) make concrete by reference to an instance of a generalized category. Indeed, the latter function has attained the status of an axiom within the official instructional context. More will be said about these functions below. Before considering the instructional context as it occurs in the home environment, the structure of instruction in the official pedagogical site – the school – will be examined. This examination will take as a point of departure a proposed structure potential for the genre of instruction within the official pedagogic context.

12.3 The structure of instruction

12.3.1 The ofﬁcial pedagogic context Within the ofﬁcial pedagogic context, instruction is, according to Bernstein, embedded within regulation. Leaving this fact aside for the moment, the school instructional context is likely to have a structure potential such as that shown in Figure 12.2.

Orientation ^Exposition (^Practicum)=Thesis Argument (Clarification))=(Recap)(Check)(Review)(Preview) (n)

(= indicates sub-elements of the adjacent superordinate term)

Figure 12.2 GSP of the ofﬁcial instructional context

In the following discussion the elements and sub-elements will be deﬁned and exempliﬁed using extracts of classroom talk. The initial element of the structure potential – Orientation – may be considered the point from which instruction proceeds. Within Orientation the discourse may look back to previous instruction thus serving to ‘contextualize’ what is to follow. The teacher may state such previous instruction in summary terms, giving in effect a Recapitulation of the relevant information that has gone before. Additionally (or alternatively) the teacher may Check students’ previous learning or their current state of knowledge via a question–answer series. A Review of the problem (or solution) may be stated and/or a Preview given in summary form of the activity that is to follow immediately. To use a grammatical analogy in terms of information distribution, this Orientation element(s) announces the Given. Thereafter begins the New – the Exposition which involves the presentation of new information. The sub-elements of the macro-element Orientation are represented as optional, though one such element is generally present. In Extract 1, three sub-elements are present. 278 Society in Language, Language in Society

This extract is taken from the transcript of a kindergarten lesson in which the teacher and children are to ‘read’ a picture:

Extract 1 1 Teacher Well do you remember we looked at this picture the other day?∗ 2 It was all about making things and sorting things 3 Chn Mmm 4 Teacher We said that some of the things could be at home 5 and some would be at preschool 6 We could even do those things at school ... 7 Well now, here is another picture 8 Now what do you think this is a picture about?

The actual structure of this Orientation element (in terms of its sub- elements) is shown in Figure 12.3 together with the rhetorical units (RU) which realize it.

Recapitulation In this example, messages 1–3 constitute the RU Recount, which realizes the element Recapitulation; messages 4–6 constitute the RU Conjecture, realizing the element Review, and message 7 constitutes the RU Commentary, realizing the element Preview. Message 8 – an Observation – begins the Exposition. The element Exposition has two sub-elements – Thesis and Argument. The central concept is put forward in the Thesis and explained in the Argu- ment(s). The explanatory arguments supporting the Thesis may be of three types: (i) definitional, (ii) exemplificatory and (iii) causal. Within definitional arguments a concept may be defined in terms of (a) its use or function;

RU 1 Teacher Well do you remember we looked at this picture the other day?* Recount 2 It was all about making things and sorting things 3 Chn Mmm Review 4 Teacher We said that some of the things could be at home .1 Conjecture 5 and some would be at preschool 6 We could even do those things at school... Preview 7 Well now, here is another picture .2 Commentary Exposition 8 Now what do you think this is a picture about? Observation

Figure 12.3 Function structure and RU structure of Extract 1 Carmel Cloran 279

(b) its inherent attributes; (c) its meaning or interpretation. Exempliﬁcatory arguments provide an illustrative example of the general concept and causal arguments give reasons for the state of affairs presented in the Thesis. The element Exposition is exempliﬁed in Extract 2 which is from a Year 7 geography lesson:

Extract 2 210 Teacher Now there’s some important rules about contour lines 211 Each line – or every so often you will find a line with a number on it 212 Now that is the number which indicates the height 213 and the height on this map that you’re all looking at is in metres, OK? 214 so for the line – you may see a line with 100 on it or 200 215 that means that that point or that line is 100 metres above sea level 216 which in ...which then enables us to say OK, well everything on that line, every point that that line passes through is also 100 metres above sea level 217 Another important thing that you need to remember when you’re looking at contour maps is that there is a specific distance between each line 218 and that specific distance is known as the contour interval, OK? 219 Now the contour interval refers to the distance between each line. 220 In other words the distance between one line and the next 221 and in the case of the map that you’re looking at, the contour interval is 40 metres 222 So from one line to the next there is a height variation of 40 metres

Figure 12.4 shows the actual structure of this Exposition element in terms of its sub-elements – ‘Thesis’ and ‘Argument’. As with Figure 12.3, the RU structure by which this element is realized is also shown. Message 210 construes the Expository sub-element Thesis and messages 211–222 construe arguments in support of the Thesis asserted. The Thesis presented in message 210 is that ‘important rules exist concerning contour lines’. This message constitutes the RU Generalization, i.e. Generalization is the RU which realizes the sub-element Thesis. This is typical, for the presentation of some central concept usually involves the assertion of the timeless state or activity of some class-exhaustive entity. It is the function 280 Society in Language, Language in Society

Thesis RU 210 Teacher Now there’s some important rules about contour lines Generalization Argument 1 211 Each line - or every so often you will find a line with a number .1 Prediction on it

212 Now that is the number which indicates the height .1.1 Generalization 213 and the height on this map that you’re all looking at is in .1.1.1 Observation metres, OK?

214 so for the line - you may see a line with 100 on it or 200 .1.1.1.1 Conjecture 215 that means that that point or that line is 100 metres .1.1.1.1.1 Observation above sea level

216 which in .. which then enables us to say OK, well everything on that line, every point that that line passes through is also 100 metres above sea level

Argument 2 217 Another important thing that you need to remember when you’re looking .2 Generalization at contour maps is that there is a specific distance between each line

218 and that specific distance is known as the contour interval, OK? 219 Now the contour interval refers to the distance between each line .. 220 In other words the distance between one line and the next 221 and in the case of the map that you’re looking at, the contour .2.1 Observation interval is 40 metres 222 So from one line to the next there is a height variation of 40 metres

Figure 12.4 Function structure and RU structure of Extract 2 of supporting arguments to explain and to make concrete this central concept. Thus, in the following arguments ‘important rules’ are instantiated, defined and exemplified. Messages 211–212 constituting the RU Prediction [Generalization] identify and define an ‘important rule’; an example of such a rule is then given in messages 213–216 which constitute the RU complex [Observation [Prediction [Observation]]]. The above discursive characterization of messages 211–216 as functioning as separate instantiating, defining and exemplifying arguments is debatable. An alternative would be to consider that this segment construes a single argument consisting of an instance which is defined and exemplified. What would constitute an argument for one or other alternative? From a linguistic point of view evidence from structure may be considered valid. Thus, for example, at the lexicogrammatical stratum, a clause containing a nominal element which is postmodified by an embedded clause is nevertheless considered to be a single clause, for example the clause realizing message 212: Now that is the number which indicates the height. On the basis of this kind of argument from structure, the second alternative given above is favoured: messages 211–216 constitute five RUs, each successive one of which is embedded within the preceding RU. In other words, messages 211–216 constituting the RU [Prediction Carmel Cloran 281

[Generalization [Observation [Prediction [Observation]]]]] are considered to construe Argument1. On the basis of this decision, messages 217–222 may be considered to construe Argument2. The function structure of Argument2 echoes Argument1: another ‘important rule’ is instantiated (message 217–218), defined (message 219–220) and exemplified (message 221–222). Note however, that, though the two arguments are identical in terms of their function structure, they vary in terms of the depth of the RU structure that actualizes the functional elements: Argument2 has less embedding of RUs, being realized by the RU structure [Generalization [Observation]]. That these two RUs realize a definition and exemplification is predictable: a definition tends to assert the timeless state or activity of a class-exhaustive entity and is therefore typically realized by the RU Generalization. The function of exemplification, on the other hand, is to make a general rule concrete. It is therefore typically realized by an RU whose CE is sensorily available, e.g. Commentary or, as is the case here, Observation. In summary terms, the elements of structure – Thesis and Argument– as exemplified in Extract 2, involve the enunciation of a universal which is then exempified by a specific instantiation of its application. The information presented in the Exposition may be clarified. This often involves a question and answer series. In the classroom, the teacher may invite the students to ask questions in order to clarify the information provided. Extract 3 which is taken from the same Year 7 geography lesson as the previous extract illustrates this element:

Extract 3 388 Teacher Hands up anyone who’s got any trouble with that 389 Do you all understand? 390 Joan, what’s wrong? 391 Child On the map it’s got some parts that are white 392 and they don’t have contour lines on them 393 What about those parts? 394 Teacher What you’ll ﬁnd there is where there are no contour lines, where it is white, the vegetation has been cleared, OK?

In message 388 of Extract 3 the teacher implicitly invites students to seek clariﬁcation. One student, Joan, responds to the invitation, noting in messages 391–392 what appears to be an anomaly in the example of the generalized instance that is before the class, and querying it. The element Clariﬁcation provides the teacher with the opportunity to discern the gaps in her presentation and to remedy the situation. The teacher in Extract 3 begins her answer to Joan’s question in message 394 and continues for another 30 messages which, for the sake of brevity, are not given here. 282 Society in Language, Language in Society

388 Teacher Hands up anyone who’s got any trouble with that Action

389 Do you all understand? Reflection

390 Joan, what’s wrong? Commentary 391 Child On the map it’s got some parts that are white .1 Observation 392 and they don’t have contour lines on them 393 What about those parts? 394 Teacher What you’ll find there is where there are no contour lines, .1.1 Prediction where it is white, the vegetation has been cleared, OK?

Figure 12.5 RU structure of Clariﬁcation element exempliﬁed in Extract 3

The discourse features of such clarificatory elements are typically concerned with the immediate, as the RU analysis of Extract 3 shows in Figure 12.5. The RUs Action, Reflection, Commentary and Observation all involve CEs which are in the immediate material environment; in the case of Extract 3 the students (Action, Reflection, Commentary and Prediction), and the map that they are studying form the CEs of the constituent messages. The element Clarification thus provides the novice with the opportunity to resolve any apparent discrepancies between the generalized information presented by the instructor in the Thesis and Argument(s) and the specific instantiation. Within the official instructional context, Exposition is typically followed by the element Practicum. This element involves the application of the concepts discussed in the form of some practical activity which may or may not involve a pen-and-paper task. In Extract 4 which exemplifies this element, no pen-and-paper task is involved:

Extract 4 269 Teacher Just to practise your grid references 270 and to get this idea of spot height into your head 271 the next one is 72 3 ...72 3, 0 5 5 ...723055grid reference ... 272 OK, ﬁnd that place on your map please 272 Child [? Excuse me] what [? did you say ﬁrst]? 273 Teacher Its on the board 274 Chn Walsh’s Pyramid 275 Teacher OK 276 Child and its 922 metres above sea level 277 Teacher OK, excellent Carmel Cloran 283

278 Child and there’s a trig station 279 Teacher Excellent 280 Very good

Through the element Practicum the teacher provides the students with the opportunity to practise the skills being developed or to record in summary form the information transmitted. Though not represented in the GSP of Figure 12.2, the element Exposition is iterative, i.e. more than a single exposition is possible within a lesson. This means that a reorienting element is likely to occur as, indeed, is Practicum. It is thus possible to view all macro-elements as iterative. Extract 5 illustrates a Practicum followed by a Reorientation:

Extract 5 196 Teacher Right, any of those that you don’t have on your list add them to your list please ...very quickly 197 Child Miss, what about [? ] 198 Teacher I have that down already (BUSY NOISE) ... 199 Child Do you have to write those things down? 200 Teacher Quickly, get those down very quickly ... 201 Now looking at your map 202 Once you’ve got those down I want everyone looking at their map ...thank you Robert 203 and on your map you’ll be able to notice these brown lines that we were just talking about which are called contour lines 204 Now, as I deﬁned them before 205 they are lines that join places of equal height above sea level 206 so in other words every place that that line passes through is the same height above sea level ...OK? 207 So that is very important that you are aware of what that deﬁnition is 208 Now ...by studying those contour lines 209 we are able to determine what the landform is

The elements Practicum and Reorientation exempliﬁed in Extract 5 together with their RU structure are shown in Figure 12.6. Note that the Reorientation sub-elements are embedded within a matrix RU which is that of Action. The Action RU typically construes regulative discourse. That elements of instructional discourse occur thus embedded within regulative discourse illustrates (albeit at a fairly low level) Bernstein’s claim that instructional discourse is embedded within regulative discourse. 284 Society in Language, Language in Society

Practicum 196 Teacher Right, any of those that you don’t have on your list add them to your list please .. Action very quickly 197 Child Miss, what about [? ] 198 Teacher I have that down already (BUSY NOISE) ... .1 Commentary 199 Child Do you have to write those things down? 200 Teacher Quickly, get those down very quickly .. Reorientation 201 Now looking at your map Action 202 Once you’ve got those down I want everyone looking at their map .. thank you Robert Recapitulation 203 and on your map you’ll be able to notice these brown lines that we .1Prediction were just talking about which are called contour lines Review 204 Now, as I defined them before 205 they are lines that join places of equal height above sea level .1.1 Observation 206 so in other words every place that that line passes through is the same height above sea level .. OK? 207 So that is very important that you are aware of what that .1.1.1 definition is Commentary Preview 208 Now .. by studying those contour lines .1.2 Reflection 209 we are able to determine what the landform is

Figure 12.6 Function and RU structure of Extract 5

With regard to Bernstein’s notion of instruction embedded within regulation, it is worth noting that the elements of the structure potential may be interrupted at any point while the teacher attends to disciplinary matters. Consider the Orientation element of the Year 7 geography lesson, presented here as Extract 6:

Extract 6 1 Teacher Right, we’re going on – 2Shush 3 We’re going on from, now today, when we looked at topographic maps and we looked at grid references 4 We tried to ﬁgure out how we can locate places on the surface of the earth 5 using a grid reference 6 Now it is important that you keep that in mind 7 A quick revision of that: what are the lines – 8 Hands up please 9 What are the lines that run up and down our map called? 10 Chn Eastings (CHORUS) 11 Teacher Hands up, thank you 12 Child Eastings Carmel Cloran 285

13 Teacher Eastings 14 Very good 15 Lines that run across our map are called what?∗ 16 Hands up 17 Chn Northings (CHORUS) 18 Child What an interesting [? ] 19 Teacher Right, we’ll stop that again 20 while [? I ] ask for hands up

In Figure 12.7, the RU analysis of Extract 6 clearly shows the embedding relation between instruction and regulation. It can be seen in Figure 12.7 that the regulative and instructional contexts are intimately intertwined though the discourse construing each is generally fairly distinctive. At times, however, the two seem to be almost indistinguishable and the discourse itself seems to be regulated as well as the students’ conduct. This situation occurs where the regulative context (in particular) is non-congruently realized, i.e. where the teacher controls the students’ behaviour not by a demand for services but by giving information as in messages 19–20 of Extract 6. Further, the regulation may be covert, for example when the teacher addresses a question concerning what is being presented to an individual who appears not to be attending. In the former

1 Teacher Right, we’re going on - Plan 2 Shush Action 3 We’re going on from, now today, when we looked at topographic maps and we looked at grid references 4 We tried to figure out how we can locate places on the .2 Recount surface of the earth 5 using a grid reference 6 Now it is important that you keep that in mind .2.1 Commentary

7 A quick revision of that: what are the lines - 8 Hands up please Action 9 What are the lines that run up and down our map called? Generalization 10 Chn Eastings (CHORUS) 11 Teacher Hands up, thank you Action 12 Child Eastings 13 Teacher Eastings 14 Very good 15 Lines that run across our map are called what?* 16 Hands up Action 17 Chn Northings (CHORUS) 18 C What an interesting [? ]

19 Teacher Right, we’ll stop that again Plan 20 while [? I ] ask for hands up

Figure 12.7 RU analysis of Extract 6 286 Society in Language, Language in Society case – where the regulative context is non-congruently realized – what distinguishes the discourse construing this context from the exchange of information typically realizing instructional discourse is the subject matter (the experiential domain). Within the classroom, regulation is used not only to control student behaviour so that instruction can occur; it is also used to construe instructional activities (e.g. the element Practicum) which facilitate the development of knowledge. Hasan (1999), drawing on Firth’s description of the features of context relevant to language, points out that in facilitative activities ‘the “relevant non-verbal objects” are likely to be such as the blackboard, the desk, the book; the relevant actions are likely to be opening, copying, writing down etc.’ Such an experiential domain is evidenced in Extract 5, messages 196–200. In Extract 6 by contrast, regulation is used to control the moment-to-moment conduct of the students. Though such control may occur anywhere in the lesson as the need arises, it seems to typically occur within the Orientation macro-element (e.g. Extract 6), including any iteration of this, i.e. Reorientation (e.g. Extract 5, messages 201–202). Once the Exposition has got under way, conduct seems to become self- regulating as the students’ interest is caught up in the information being transmitted.

12.3.2 The local pedagogic context The embedding of official instructional discourse within the regulatory context seems to be specific to the school environment and would seem to be a product of the one-to-many ‘instructor–instructed’ relationship. Within the home, instruction is likely to be embedded within any ongoing activity. In other words, instruction occurs ‘on the run’ as it were, in the course of the daily routine. To the extent that ongoing verbal activity – whether constitutive or ancillary – construes any of the other three generalized socialization contexts – the regulative, the imaginative or the interpersonal – then it may be said that the local instructional context is embedded within these4 (see also Cloran, 1999, for a discussion of such embedding). One of the central features of the local instructional context is that it tends to be unplanned; there is no predetermined curriculum that is more specific than what Hasan (e.g. 1988) terms the long-term goal orientation of the mother. Local instruction is generally motivated by two more immediate factors: (a) the fact that the child wants to know, in which case it typically arises from the child’s information-seeking; or (b) the fact that the child is considered to need to know. In this latter case, the instruction tends to be: (i) some kind of additional information arising from some context other than the instructional, as in Extract 8 below; or (ii) some kind of rationalization arising from the regulative context, i.e. from some kind of prescription or prohibition on the child’s behaviour. These points will be discussed further presently. Carmel Cloran 287

The fact that instruction in the home is unplanned is reﬂected in its structure potential. Unlike instruction in the school, there is, generally speaking, no Orientation element – no Recapitulation, Review, Check or Preview. Thus the context for home instruction is whatever it is that is going on – the construal of the regulative, interpersonal or imaginative contexts – which gives rise to the instruction. The only possible exception to this state of affairs is when an instructional context itself becomes the context for further instruction. In such circumstances the initial instructional context may conceivably be reviewed by the mother before proceeding to activate a new instructional context. It is not only the element Orientation that is absent from the structure potential of the local instructional context. Within the home, the element Practicum – the allocation of some practical task in order to apply the concepts transmitted in the Exposition – is unlikely to occur. If it does, then it is likely to be at the instigation of the child rather than the mother. This possibility will be illustrated presently. The only element of structure in common to both the local and the ofﬁcial instructional context is the macro-element Exposition. The consti- tution of this macro-element varies, however, between home and school. Whereas in the instructional context of the school this element consists of the sub-elements Thesis and Argument(s) as described above, in the home this element may best be described as consisting of the sub-elements Ask and Inform. Such a description captures perhaps the most important difference between instruction in the school versus the home, i.e. that in the latter environment the child instigates the instruction either directly via his/her verbal act of seeking information, or indirectly via his/her need (as perceived by the mother) to be given some information. This is not to suggest that in the instructional context of the school the teacher does not use the child’s sayings as the point from which to build new information. Clearly this does happen, though it is likely to be the child’s (wrong) answers to the teacher’s questions which evoke such new information, as in Extract 7:

Extract 7 176 Teacher Does anyone know the part of the hospital you go to 177 if you have an accident 178 and you’ve got to go there in a hurry? 179 Child Oh .... 180 Child [? ] 181 Teacher No 182 Child Oh ... 183 Child Ambulance 184 Teacher No 185 The ambulance takes you there sometimes 288 Society in Language, Language in Society

186 Child Theatre 187 Teacher Not the theatre 188 That’s where you have your operation, in the theatre

The information given by the teacher in messages 185 and 188 is provided to correct the ‘wrong’ information provided by the children in answer to her question (176–178). It is in this sense that the teacher may give information in response to the children’s sayings. This situation contrasts with the strategies of some mothers, as Extract 8 shows:

Extract 8 208 Stephen This is a boat that ﬂies 209 Mother A boat that ﬂies! 210 Stephen Yeah ... 211 Mother Is it a hydroplane? 212 Stephen Yes 213 Mother A hydroplane is a plane that can land on the water 214 Stephen Yes 215 Mother Or is it a hydrofoil? 216 Stephen No, 217 it’s a hydroplane ... 218 This is – that’s the water 219 and it saves people 220 Mother Does it? 221 Stephen Yes ... 222 every person that’s um that’s got drowned 223 Mother Oh ...there’s a helicopter that goes up and down the beaches in summer watching out for people 224 It’s called a rescue helicopter 225 Stephen Oh that’s [? mine] 226 Mother I think it mainly watches um for sharks 227 but it might also rescue people 228 if they’re a long way out from the beach and in trouble

In message 213 the mother provides unsolicited information in the form of the RU Generalization. This information was prompted by the child’s sayings in his construal of the imaginative context and, unlike the teacher’s messages 185 and 188 in Extract 7 it is not corrective in function; rather it is simply providing the child with additional information which serves to relate his fantasy world to the world of external reality. The mother’s messages 223–228 also serve this function. The structure potential for the local instructional context may therefore be described as in Figure 12.8. Carmel Cloran 289

Exposition = {(Ask ^) Inform (Extended)}

Figure 12.8 GSP of local instructional context

Indicated in the above GSP is the fact that the element(s) (Ask∧ ) Inform are iterative. This means that more than one Ask∧ Inform sequence may occur within a sequence of messages in which information related to a single topic is sought and given, i.e. an instructional episode. Extract 9 illustrates the iterative nature of Ask ∧ Inform.

Extract 9 175 Carol Mum, do pussy cats die 176 when people die? ... 177 Mother Do pussy cats what, love? 178 Carol Die 179 when people die? 180 Mother Well pussy cats die 181 when their time comes ... 182 Everything dies one day 183 Carol Do dogs?∗ 184 Do they one day? 185 Mother Do what? 186 Carol Do dogs die one day 187 when – 188 Mother Yes, dogs die too ... 189 Carol Do fruit die? 190 Mother Fruit dies, yes, in a different sort of way 191 Carol How? 192 Mother Well see how the fruit up there on the tree is green? 193 Carol Mm 194 Mother See how, down here, it’s gone all yellow and squashy and horrible? 195 Carol Ah 196 Mother That means it’s died 197 It’s – well we don’t say it’s died 198 We say it’s gone bad

In Extract 9, messages 183, 189 and 191 begin new Ask ∧ Inform sequences; however, messages 175–190 are regarded as construing a single instructional episode in which each new Ask ∧ Inform cycle is cohesively related to the previous one. This is clearly seen in the RU analysis of Extract 9 shown in Figure 12.9; the analysis also shows that this instructional episode is embedded within (arises out of) some other context. 290 Society in Language, Language in Society

175 Carol Mum, do pussy cats die .3 Generalization 176 when people die? .. 177 Mother Do pussy cats what, love? 178 Carol Die 179 when people die? 180 Mother Well pussy cats die 181 when their time comes .. 182 Everything dies one day 183 Carol Do dogs?* 184 Do they one day? 185 Mother Do what? 186 Carol Do dogs die one day 187 when - 188 Mother Yes, dogs die too ... 189 Carol Do fruit die? 190 Mother Fruit dies, yes, in a different sort of way 191 Carol How? 192 Mother Well see how the fruit up there on the tree is green? .3.1 Commentary 193 Carol Mm 194 Mother See how, down here, it’s gone all yellow and squashy and horrible? 195 Carol Ah 196 Mother That means it's died .3.1.1 Generalization 197 It’s - well we don’t say it’s died .3.1.1.1 Recount 198 We say it’s gone bad

Figure 12.9 RU structure of Extract 9

The characterization, in Figure 12.8, of the element Ask as optional indicates that information does not have to be sought in order for the instructional context to evolve (be activated). This point has already been made and illustrated by reference to Extract 8 above. In addition, when instruction arises within the regulative context, information which functions as a rationalization for the Direction may be unsolicited (as shown in Extract 11 below) though clearly it is often solicited by the child in the form of a Why question. The obligatory element in the local instructional context is the element Inform. This element may or may not be extended by expanding messages. If expansions are present, these provide (i) additional information; (ii) clarification; (iii) exemplification; or (iv) rationale. Such expansions are reminiscent of the functions of Arguments and Clarification within Exposition in the official pedagogic context and are exemplified in Extract 10:

Extract 10 130 Linda What does “ﬂu” mean? 131 Mother That’s when you get coughs and colds and sore throats 132 Lots of people get ﬂu now because of the cold weather, the wet weather Carmel Cloran 291

133 Linda Like you? 134 Mother Like I did, yes 135 Linda You’ve got ﬂu 136 Mother Yeah, just like Mummy 137 Linda You’ve got ﬂu haven’t you? 138 Mother Mm, Mummy gets it every year

Messages 130–131 constitute the RU Generalization: message 130 is an eliciting Generalization and 131 provides the generalized information sought in the child’s question. The mother adds to this information in message 132 and in messages 133–137 further information which exemplifies the Gen- eralization made in message 132 is cooperatively built up. This situation is more clearly revealed in an RU analysis of the extract. In the RU analysis shown in Figure 12.10 the context of this instructional episode is given, showing that instruction arose within an interpersonal context as a result of the occurrence of the item ‘flu’. As previously noted, the instructional episode begins at message 130 with the element Ask: the child’s question What does “flu” mean? This is followed by the element Inform (message 131) which is then expanded (messages 132–138). The expansions involve an Account RU with an embedded Com- mentary; these RUs function to exemplify the Generalization and to thereby make it more concrete and immediate for the child. The most immediate

117 Mother What did you do at school today? Recount 118 Linda Um - 119 Mother I didn’t notice Jane 120 Linda She wasn’t there 121 Mother Did she have a day off? 122 Linda Yeah 123 Mother Was she sick? 124 Linda Probly 125 Mother Probably? 126 Linda She’s sick like you probly .1 Report 127 Mother Maybe 128 Linda Probly she’s got something else 129 Mother Or flu or something

130 Linda What does ‘flu’ mean? Generalization 131 Mother That’s when you get coughs and colds and sore throats 132 Lots of people get flu now because of the cold weather, .1 Account the wet weather 133 Linda Like you? 134 Mother Like I did, yes 135 Linda You’ve got flu .1.1 Commentary 136 Mother Yeah, just like Mummy 137 Linda You’ve got flu haven’t you? 138 Mother Mm, Mummy gets it every year .1.1.1 Reflection

Figure 12.10 RU structure of Extract 10 292 Society in Language, Language in Society

RU – Commentary – is itself elaborated by a Reflection which adds yet further information. This particular instructional episode is reminiscent of instructional episodes in the classroom in that generalized information is exemplified: there is a progressive movement in the discourse from the general to the particular. Such a movement is also to be seen in Extract 9 as its RU structure shows in Figure 12.9. Instructional discourse involves such a movement from the general to specific instance typically when the evoking discourse is itself a generalization. In fact, the discourse may move in the opposite direction. This is seen in Extract 11 where instruction occurs in the context of regulation:

Extract 11 315 Mother I don’t think Nanna wants her blind cord chewed. 316 It’s ﬁlthy, darling, 317 very dirty. 318 All that dirt’s going into your mouth and down into your tummy 319 It’s really best not to ﬁll your tummy with dirt. 320 Would you like some mandarin to put in your tummy instead 321 Stephen Yes

Message 315 in Extract 11 construes the regulative context, albeit non- congruently; that it is a prohibition on his behaviour must be inferred by the child. The meaning – do/do not – is inferred by reference to the fact that some unsolicited information is being given and by assuming that there is usually some purpose in giving such information. If this inference is made, then messages 315–319 can be seen as a rationale – a reason for not behaving. The ﬁnal message of the Rationale – message 319 – provides the generalized information which enables the child to transfer the prohibition on ‘eating dirt’ to all occasions. Note that the instructional discourse provided by the Rationale proceeds from the speciﬁc instance – chewing of the dirty blind cord – to the general. This can be seen in the RU analysis of Extract 11 shown in Figure 12.11.

315 Mother I don’t think Nanna wants her blind cord chewed. Report 316 It’s filthy, darling, .1 Commentary 317 very dirty. 318 All that dirt’s going into your mouth and down into your tummy 319 It’s really best not to fill your tummy with dirt. .1.1 Generalization

320 Would you like some mandarin to put in your tummy instead Action 321 Stephen Yes

Figure 12.11 RU analysis of Extract 11 Carmel Cloran 293

12.4 Summary: Home versus school instruction

That local and official pedagogic contexts differ yet have certain similarities is brought out by the above descriptions. One major difference not addressed so far between instruction in the two environments is the degree of their institutionalization. Thus the school environment is an institutionalized learning environment in which the construal of the instructional context is assumed. Indeed, the physical environment is multiply semiotically coded (Hasan, 1981) for this purpose. In other words, children go to school to learn – to be instructed by people whose primary responsibility lies in this task of instructing. This amounts to saying that the official instructional context is planned. Attendance by both parties – instructor and instructed – in the place of instruction initiates the activation and construal of the instructional context. In other words, charged with the task of teaching and faced with a group of children the teacher has little choice but to teach. The home environment, by contrast, is a non-institutionalized learning environment. Here, the instructional context is activated by the mother usually (though not invariably) in response to a question from the child. Not every question prompts the activation by the mother of the instructional context; and not every mother seizes the opportunity provided by a question from the child to activate the instructional context – to assume the role of instructor. Mothers vary in this respect though the variation may be shown to be systematic, i.e. associated with social attributes of families. Sys- tematic variation occurs with respect to the kind of RU which realizes the structural element(s) (Ask and) Inform and, if Inform is expanded, the kind of RU which functions as exemplification, etc. It is to such variation that I now turn.

12.5 Variation in the construal of the local instructional context

An important variable in the context of instruction in the home is the kind of RU which realizes the element(s) (Ask and) Inform and, where Inform is expanded, the kind of RU which functions as clarification, exemplification, etc. In middle-class families the prototypical instance of decontextualized language use, i.e. Generalization, tends to occur with greater frequency than in working-class families (Cloran, 1994). Bearing in mind that the instructional context is often activated by a mother in response to her child’s question, it would seem that even at age 3.5 middle-class children are oriented to class-exhaustive entities and timelessness with respect to events/states. That this is the case is not surprising given the middle-class mother’s appropriation of the pedagogic practices of the school. This is not to say, however, that expository instruction as practised in the ‘official’ 294 Society in Language, Language in Society pedagogic environment and exemplified in the previous discussion does not occur in working-class families; Extract 10 is an example of the instructional context that occurred in a working-class family. I will examine a fairly lengthy extract from a working-class dyad in order to show how opportunities for the development of the instructional context having features of ‘official’ instructional discourse are/are not taken up by a working-class mother. In Extract 12(a) Karen seeks information from her mother:

Extract 12(a) 1 Karen What’s that? 2 Mother It’s a torch 3 but it doesn’t work 4 because it hasn’t got any batteries in it

If Extract 12(a) is considered to construe an instructional context despite the absence of class-exhaustive reference, then message 1 realizes the element Ask and messages 2–4 the element Inform (extended). The extension of the Inform element serves to provide additional information about a state of affairs involving the object of interest and a reason for this state of affairs. Karen’s mother chose to stop providing information at this point. Karen therefore must infer the principle ‘Torches need batteries in order to work.’ There is nothing in her subsequent sayings that indicate whether or not this inference was made. Rather, she goes on to demand the opportunity to examine the object more closely:

5 Karen Show me 6 Mother No, 7 just wait a minute

Karen’s demand in message 5 may be seen as a clarificatory kind of extension: the child seeks to clarify/verify for herself the information provided by the mother. This then is more like the element Clarification that may occur within the official pedagogic context. Rather than regarding this as an extension of the element Inform, perhaps it should be seen as a separate element as it is in school instruction. While on ‘hold’, Karen seeks further information related to the information seeking–giving of messages 1–4:

8 Karen Why doesn’t it work? ... 9 Mother Oh the batteries are stuck Carmel Cloran 295

The mother’s Inform, here, is not extended in any way, and again Karen must infer the principle. Again she goes on to seek clariﬁcation:

10 Karen Show me 11 Show me 12 Show me 13 Mother Don’t break it 14 Karen No

The RU structure of the two sequences which form the instructional episode so far is given in Figure 12.12. The RU constituted by messages 1–2 realizing the elements Ask ∧ Inform is an Observation: information is sought and given concerning the identity of an object in the material situational setting. The element Inform in the ﬁrst instructional sequence in this instructional episode (messages 1–7) is extended by (i) the RU Commentary (message 4) since the rhetorical features of this message concerns the current (non- habitual) states of a co-present object; and (ii) the RU Action by which the child demands to see and presumably examine the object herself, in order to clarify/verify the information provided by the mother (message 5). The second Ask ∧ Inform sequence (messages 8–9) is also an Observation. Again the child’s clariﬁcation-seeking (10–14) is realized by the RU Action.

1 Karen What’s that? Observation 2 Mother It’s a torch 3 but it doesn’t work 4 because it hasn’t got any batteries in it .1 Commentary

5 Karen Show me Action 6 Mother No, 7 just wait a minute

8 Karen Why doesn’t it work? .. Observation 9 Mother Oh the batteries are stuck

10 Karen Show me Action 11 Show me 12 Show me 13 Mother Don’t break it 14 Karen No

Figure 12.12 RU analysis of messages 1–14 of Extract 12(a) 296 Society in Language, Language in Society

Despite the expansion of information and the iteration, the instructional episode of Extract 12(a) does not depart from the here and now of the immediate situational setting but remains tied to it. Indeed, each occurrence of the clarificatory expansion is expressed by an RU which is even more closely based in the material here and now, i.e. the Action RU (messages 5–7; 10–18). It might be said then that, just as there is a progressive movement from the general to the specific within the discourse of those dyads where the pedagogic practice is more like that of the school, so too in Extract 12(a) there is such a movement in the discourse. The difference between the previous extracts and Extract 12(a) is that in the latter the general is less general and the specific more specific. This instructional episode continues but the nature of the instruction changes from expository to procedural:

Extract 12(b) 19 Mother Now we’ll do that up 20 and then we put that in there 21 That’s the light 22 Karen Yep and I’m gonna put – 23 Can I put this in? 24 Mother Give it here to me please for a minute 25 Karen I will do it 26 Mother Well give it here to me for a minute ... 27 There, that goes there 28 Karen Yeah ... 29 Mother Oops! 30 Karen Hold it there 31 Mother Put it on there ...like that 32 Karen Now let me do it all by my own now 33 Mother And then when it’s got batteries 34 you turn it on 35 and then [?light] comes out

The RU structure of this segment of the instructional episode is shown in Figure 12.13. It is worth noting that it is in this procedural segment of the instructional episode that a Generalization occurs. Finally, it was mentioned above that the structural element Practicum which often occurs in ofﬁcial instructional contexts may occur in the home environment. In message 36, the child, Karen, seeks permission to ‘put into practice’ the sequence of events mentioned in the Generalization (see Figure 12.14). Carmel Cloran 297

19 Mother Now we’ll do that up Plan 20 and then we put that in there .1 Reflection 21 That’s the light .1.1 Observation 22 Karen Yep and I’m gonna put - 23 Can I put this in? .1.1.1 Action 24 Mother Give it here to me please for a minute 25 Karen I will do it 26 Mother Well give it here to me for a minute .. .2 Action 27 There, that goes there .2.1 Observation 28 Karen Yeah .. 29 Mother Oops!

30 Karen Hold it there Action

31 Mother Put it on there .. like that Action 32 Karen Now let me do it all by my own now 33 Mother And then when it’s got batteries .1 Generalization 34 you turn it on 35 and then [?light] comes out

Figure 12.13 RU structure of Extract 12(b)

36 Karen [? Can] we did it? (LAUGHS) Action 37 Mother But it hasn’t got any batteries in .1 Commentary it. 38 It’s not working 39 Karen Well I can fix it? .1.1 Action 40 Mother It’s not broken .1.1.1 Commentary 41 it just needs new batteries .. doesn’t it? 42 That’s all it needs

Figure 12.14 The structural element Practicum

12.6 Concluding remarks

The question may be asked: what is the value of the foregoing analysis? Is there any practical application to be drawn from the descriptions in the previous sections? It is suggested that any description of phenomena 298 Society in Language, Language in Society at the semantic stratum has the potential to be readily adapted in practice. This is because the semantic stratum is closer to everyday informal understandings about language and linguistic events than is, say, the stratum of lexicogrammar: people think about linguistic events in terms of their meanings rather than their wordings. Thus, it would be possible to suggest to caregivers or teachers of young children how to construe the instructional context such that children become habituated to the kinds of meanings likely to be found therein. For example, given a dialogue such as that between Karen and her mother (Extract 12a), the locations at which the more decontextualized categories of language could be inserted might be pointed out, as in Figure 12.15 showing Extract 12(a) amended.

1 Karen What’s that? Observation 2 Mother It’s a torch 3 but it doesn’t work 4 because it hasn’t got any batteries in it .1 Commentary See, torches need batteries .1.1 Generalization

in order to work It might work .1.1.1 Conjecture if we put new batteries in it

Figure 12.15 Extract 12(a) amended

The discussion in this chapter provides an understanding and practical application of the linguistic features of the kind of decontextualized language construing instructional discourse as well as the structure and function of the contexts evoking it.

Notes

1. The transcripts of both mother–child and teacher–pupil interaction reproduced here are from Hasan’s corpus collected at Macquarie University in the 1980s. 2. Hasan (1983) proposed that the message is the basic unit at the semantic stratum. 3. A rhetorical unit is the name given to the intermediate linguistic unit located at the semantic stratum in the SFL model of language; the relation between message and rhetorical unit is considered to be one of constituency (Cloran, 1994, p.125). 4. It may be of interest to ascertain whether there is any variation among the dyads concerning which of these three contexts is most likely to ‘host’ the instructional context, and, if variation exists, what is its nature? Carmel Cloran 299

References

Bernstein, B. (1971) Class, Codes and Control,Vol.I:Theoretical Studies towards a Sociology of Language (London: Routledge). Bernstein, B. (1990) Class, Codes and Control,Vol.IV:The Structuring of Pedagogic Discourse (London: Routledge). Cloran, C. (1994) Rhetorical Unit and Decontextualisation: an Enquiry into Some Relations of Context, Meaning and Grammar. Monographs in Systemic Functional Linguistics, No. 6 (Nottingham: Department of English Studies, University of Nottingham). Cloran, C. (1999) ‘Contexts for learning’. In F. Christie (ed.) Pedagogy and the Shaping of Consciousness: Linguistic and Social Processes (London: Cassell), pp. 31–65. Hasan, R. (1978) ‘Text in the systemic functional model’. In W. Dressler (ed.) Current Trends in Textlinguistics (Berlin: Walter de Gruyter), pp. 228–46. Hasan, R. (1981) ‘What’s going on: a dynamic view of context in language’. In J. E. Copeland and P. W. Davis (eds) The Seventh LACUS Forum 1980 (Columbia, SC: Hornbeam), pp. 106–21. Hasan, R. (1983) ‘A semantic network for the analysis of message in everyday talk between mothers and their children’. Mimeo, Macquarie University, Sydney. Hasan, R. (1988) ‘Language in the processes of socialisation: home and school’. In L. Gerot, J. Oldenburg and T. van Leeuwen (eds) Language and Socialisation: Home and School – Proceedings from the Working Conference on Language in Education (Sydney: Macquarie University), pp. 36–96. Hasan, R. (1999) ‘Speaking with reference to context’. In M. Ghadessy (ed.) Text and Context in Functional Linguistics (Amsterdam: John Benjamins), pp. 219–328. 13 ‘Threads of Continuity’ and Interaction: Coherence, Texture and Cohesive Harmony Kristin M. Khoo

13.1 Introduction

The method of Cohesive Harmony Analysis (hereafter CHA) was originally developed by Hasan as a measure of coherence in text. Detailed in Hasan (1984), CHA was derived in pursuit of a research question concerning the degree of coherence in children’s stories and correlations with ‘social factors’ (Hasan, 1984, p. 181). CHA was further elaborated in Hasan (1985, p. 52), anchored within a discussion of text as ‘unity of structure’ and ‘unity of texture’ and the relation to context of situation. Cohesive harmony is seen through ‘threads of continuity’ (Hasan, 1985, p. 84) – identity and similarity chains constructed through cohesive relations – and the interaction of these chains, based on experiential transitivity relations, ‘echoed’ through the text. The visual patterning of the cohesive harmony can be displayed in chain interaction diagrams, though the method also involves a quantitative component through token counts and ratios. Since the development of CHA, the method has been applied in studies across registers: with variation in sample size, text length, mode characteristics (e.g. spoken, written, monologic, dialogic), study methodology (e.g. elicited, naturally occurring), and type of population (e.g. language disordered) (see section 13.4). There is also variation in the degree of consistency in application of the method, requiring an exploration of such differences – as modifications specific to register or population, for the purpose of particular research questions, or through difficulties encountered in applying the technique. Yet, even where difficulties have been documented, CHA has been viewed a method worth testing and refining, the value of the tool expounded in varied ways. Furthermore, although developed as a quantitative measure of coherence, CHA has also been positioned as a method for modelling attributes of texture; and increasingly, evidence from CHA has been used for research

300 Kristin M. Khoo 301 questions regarding texture, text structure and context. The opportunities for iconic representation of semantic continuities have been embraced, often without the further application of quantitative counts of coherence, for example in identifying semantic boundaries in text (see section 13.4.3). There has been a shift to view and utilize CHA as a tool for discourse analysis, not limited in scope to a measure of coherence alone. This chapter considers the following questions: what is cohesive harmony and CHA? How is cohesive harmony positioned within the theory of text, texture and cohesion in systemic functional linguistics (hereafter SFL)? What are the range of text types and populations to which CHA has been applied? What research questions are being addressed through CHA, and what types of evidence does CHA yield? What are the strengths of CHA and how does it need to be further developed? Through a discussion of these questions, this chapter will also draw attention to questions of linguistic method and development, such as the adaptation of method to register and research questions.

13.2 Texture, cohesion and coherence in text: the theory of cohesive harmony

A discussion of cohesive harmony requires an understanding of the concept of text according to Hasan, and the theoretical foundations of a social- semiotic perspective of SFL. A text is deﬁned in Halliday and Hasan’s (1976, p. 1) Cohesion in English as ‘a unit of language in use’, and therefore considered a semantic unit. A text can be viewed as both product and process:

It is a process in the sense of a continuous process of semantic choice, a movement through a network of meaning potential, with each set of choices constituting the environment for a further set. (Halliday and Hasan, 1985, p. 10)

A text has two types of unity – that of structure and texture (Halliday and Hasan, 1976; Hasan, 1985), both of which are related to the context of situation, though in different ways and at different levels of delicacy in system networks (Hasan, 1985, 2004, 2009). In Halliday and Hasan (1976), texture is positioned as the result of semantic conﬁgurations of register and cohesion, with a ‘unity of texture’ referring to the resources within the textual metafunction of systemic functional modelling (see Halliday and Matthiessen, 2014 for SFL model). Halliday argues that it is the textual function that ‘breathes life into language’ (Halliday, 2003[1976], p. 70), seen through the ‘enabling’ role of the textual with view to the ideational and interpersonal metafunctions (Halliday and Matthiessen, 2014, p. 30). Tex- ture is therefore related to the ‘property of connectedness’ (Hasan, 1979, p. 370): the resources that allow a text to cohere with itself and the context 302 Society in Language, Language in Society of situation. Expressed in the notion of texture is the property of ‘being a text’ (Halliday and Hasan, 1976, p. 2). One of the resources for texture that expresses this continuity ‘in order for the discourse to come to life as a text’ (Halliday and Hasan, 1976, p. 299) is cohesion: a semantic concept that refers to nonstructural relations that capture semantic links or the ‘relations of meaning’ in text (Halliday and Hasan, 1976, p. 4). Cohesion forms the basis for internal semantic bonding, each link expressed through lexical and grammatical choices; and in speech – also through intonation. Cohesive resources include reference, substitution and ellipsis, conjunction and lexical cohesion, seen through cohesive ties across grammatical structures (Halliday and Hasan, 1976). In Figure 13.1, the place of cohesion and the textual metafunction is located in a metafunction/rank matrix of SFL (Halliday, 2009, p. 85). The attribute of texture is foregrounded in Hasan’s discussion of coherence, where coherence refers to ‘the property of “unity”, of “hanging together” ’ (Hasan, 1984, p. 181). Texture, Hasan claims, is related to the perception of coherence of text (Hasan, 1985) and since ‘structure is a uni- formly integrative device’, it is in the nonstructural relations that coherence is pursued (Hasan, 1984, p. 184). CHA is one way of modelling attributes of texture (e.g. Hasan, n.d., 1985), and seen through the relation of texture and coherence, ‘cohesive harmony’ has come to refer to patterns of texture:

...a particular pattern of texture, to which I have referred as cohesive harmony. (Hasan, 1994, p. 138) ...it is cohesive harmony that gives ‘texture’ to a text, particularly where such patterns of continuity overlap. (Halliday and Webster, 2009, p. 241)

The theory of cohesive harmony offers a way into both the method of CHA and discussion of the opportunities that are afforded by the technique. As captured in the quote from Hasan below, the patterns of cohesive harmony are seen through ‘harmony in more than one respect’ (Hasan, 1985, p. 94). Following the summary of cohesive harmony below, the method of CHA will be detailed in the next section (13.3).

It is harmony in more than one respect: it brings together lexical and grammatical cohesive devices, subjecting them to semantic considerations of identity and similarity. This is as it should be; a text, after all, is not a unit of form but of meaning. Secondly it is harmony because it harmonises the output of two macro functions: the textual and the experiential. The output of the textual function are the chains and the interactions; the output of the experiential function at the rank of clause and group is what the interaction is built upon. Thus cohesive harmony is an account of how the two functions ﬁnd their expression in one signiﬁcant whole. (Hasan, 1985, p. 94) 303 (cohesive) COHESIVE RELATIONS REFERENCE ELLIPSIS/ SUBSTITUTION CONJUNCTION CONJUNCTION DETERMINATION THEME CULMINATION VOICE DEICTICITY CONJUNCTION MINOR MOOD (adjunct type) PERSON ATTITUDE MOOD MODALITY POLARITY FINITENESS VOICE TION) COMMENT (adjunct type) TRANSITIVITY (circumstance type) CLASSIFICATION EPITHESIS QUALIFICATION TRANSITIVITY (process type) MINOR (circumstance type) experiential EVENT TYPE ASPECT (non-finite) TENSE DERIVATION (DENOTATION)ACCENTUATION (CONNOTA- KEY INFORMATION MODIFICATION THING TYPE MODIFICATION QUALITY simplexes complexes ideational interpersonal textual INTERDEPENDENCY (parataxis/hypotaxis) and LOGICAL–SEMANTIC RELATION (expansion/projection) logical complexes (clause- phrase- group- word) unit info. complex ‘A “function/rank matrix” for the grammar of English, where (i) rows show rank and primary rank metafunction, and show (ii) columns show of English, where (i) rows class, the grammar for matrix” “function/rank ‘A system(s) in each cell” (iii) capitals show [nominal] [adverbial] metafunction Place of cohesion and the textual metafunction rank matrix of SFL (reproduced from Halliday, 2009, p. 85) rank [class] clause phrase [prepositional] group [verbal] word information unit Figure 13.1 304 Society in Language, Language in Society

13.3 The development and method of Cohesive Harmony Analysis

CHA was developed using spoken texts produced by 6 to 7-year-old children asked to tell a bedtime story to a teddy bear (Hasan, 1984, 1985). The sample included 80 texts, though only 5 analysed texts are available in Hasan (1984) and (1985) combined. In the development of CHA, cohesion and the effects of exophora and ambiguity were documented (cf. Hasan, 1984, 1985, n.d.), with cohesion observed to be a ‘necessary but not sufficient’ condition of coherence (Hasan, 1985, p. 94). Drawing on work from Halliday and Hasan (1976), cohesive chains are formed from componential cohesive devices – both lexical and grammatical – though chain construction was revised ‘so that they are seen neither as just lexical or just grammatical, but have a status by reference to their potential function in the text’ (Hasan, 1984, p. 211). The classification of devices and cohesive tie relations are seen in Figure 13.2, reproduced from Hasan (1985, p. 82). The construction of cohesive chains is based on semantic relations (the cohesive links) of co- reference, co-classification and co-extension. Chains formed on relations of co-reference are referred to as identity chains (ICs), and those formed on the basis of co-classification and/or co-extension through repetition or synonymy, antonymy, hyponymy and meronymy are referred to as similarity chains (SCs). Chains may also be constructed on both relations of identity and similarity, with patterns of complex chain formation, such as ‘bifurcation’ or ‘conjunction’ (Hasan, n.d., pp. 70, 77; 1984, p. 199). Chaining is based on a ‘lexical rendering’ of text, with lexical items referred to as tokens. Tokens that form chains are referred to as relevant tokens (RTs), and those that are not cohesive, or do not form chains, peripheral tokens (PTs). As can be seen from Figure 13.2, both grammatical and lexical cohesive devices are chain forming. Lexical cohesion may be based on either their relation to the linguistic system (general) or relation to the text (instantial).1 Through a series of hypothesis testing in preliminary analysis, Hasan (1984) found that chain development alone did not account for coherence, and therefore, a second layer of analysis was required. The move to CHA was to track relations of continuity along two axes (the syntagmatic and paradigmatic) – mapping both cohesive lexical chains through relations of identity and similarity, and chain interactions when ‘two or more members of a chain stand in the same grammatical relation to at least two other members’, therefore creating an ‘echo’ (Hasan, 1984, p. 211). Chain interactions are constituted by repetitions of grammatical relations in the experiential metafunction of Halliday’s SFL model (cf. Halliday and Matthiessen, 2014). Chain interaction in Hasan’s method includes connection of chains through repetition within group (e.g. Epithet–Thing) and within clause-level grammatical roles (e.g. Process–Goal). Tokens that interact are referred to as central tokens (CTs). Kristin M. Khoo 305

NON-STRUCTURAL COHESION COMPONENTIAL RELATIONS ORGANIC RELATIONS Device Typical Tie Relation

GRAMMATICAL A: Reference A: Conjunctives COHESIVE 1. Pronominals e.g. causal tie 2. Demonstratives concession tie . . . DEVICES co-reference 3. Definite article 4. Comparatives

B: Substitution and Ellipsis B: Adjacency pairs 1. Nominal e.g. Question (followed by) answer; offer (followed by) 2. Verbal co-classification 3. Clausal acceptance; order (followed by) compliance . . .

A: General Continuatives LEXICAL 1. Repetition (e.g. still, already . . . ) COHESIVE 2. Synonymy co-classification DEVICES or 3. Antonymy co-extension 4. Meronymy

B: Instantial 1. Equivalence co-reference or 2. Naming co-classification 3. Semblance

STRUCTURAL COHESION A: Parallelism B: Theme–Rheme development C: Given–New organization

Figure 13.2 Summary of cohesive devices (reproduced from Hasan, 1985, p. 82)

Hasan (1984) reports that the addition of chain interaction shows differences between texts that could not be captured with cohesive chains alone, and reports correlations with perceptual rankings of coherence. Hasan therefore hypothesizes that the degree of chain interaction would correlate with the degree of coherence in text: ‘the greater the cohesive harmony in text, the greater the text’s coherence’ (Hasan, 1984, p. 216). Furthermore, she proposes that the patterns of chain interaction could be interpreted both iconically and numerically, therefore providing a way of measuring textual coherence. The visual displays of the chain interactions allow for an iconic representation of the logogenetic patterning of cohesive harmony and qualitative descriptions of chain connection patterns, such as simultaneous or staggered interaction, or chain breaks (cf. Hasan, 1984). The ﬁnal move of CHA involves the calculation of within-text relations, and, if required, the ranking of texts and statistical correlations. There is, though, a difference between Hasan (1984) and Hasan (1985) in the hypotheses and deﬁnitions that constitute the method of CHA and 306 Society in Language, Language in Society which affect the calculations. In Hasan (1984, p. 217), the calculation that represents the measure or ‘degree of cohesive harmony’ is CTs as a percent- age of total tokens (TTs). This is the calculation Hasan found ‘consistently correlated with readers’ judgment on how the texts ranked vis a vis each other in coherence’ (Hasan, 1984, p. 218). It is also the number by which Hasan hypothesized the 50 per cent coherency level: ‘any text will be seen as coherent, in which the central tokens (CT) form at least 50 percent of the total tokens (TT)’ (Hasan, 1984, p. 218). This calculation has been taken as the cohesive harmony index (CHI) in many subsequent applications of the procedure. Alongside this measure, Hasan (1984) proposes two further hypotheses: if the CHI is equal, variation in coherence would be associated with gaps seen on the chain interaction diagrams; and the higher the ratio of CT to PT, the more coherent the text, all else being equal. In contrast, Hasan (1985) refers to cohesive harmony as the sum of the following three phenomena, presented as ordered principles:

A text is more likely to be coherent if: (i) the ratio of relevant to peripheral tokens is high, (ii) the ratio of central to non-central tokens is high, (iii) there are no or few breaks in the chain of interaction. (Hasan, 1985, p. 94)

Non-central tokens refer to relevant tokens that do not interact (Hasan, 1985, p. 93).

13.4 A survey of work applying Cohesive Harmony Analysis

Since the development of CHA, the method has been tested and utilized for a wide range of research questions and examples of text analysis – a selection of which will be drawn on in this chapter. Table 13.1 shows a snapshot of these studies, with groupings based on clusters of similarities such as register or population, for ease of reference.2 In Table 13.1 the number of texts has been included though the number of participants has not, and therefore studies that included a repeated measures design – either through longitudinal design (e.g. Rentel et al., 1983; Rentel, 1988; Otto, 1992; Armstrong, 1996; Fang and Cox, 1998), production of multiple texts or elicitation of texts in multiple contexts (e.g. Rentel et al., 1983; Pappas, 1985, 1987; Bottenberg et al., 1985; Armstrong, 1987a, b; Pappas and Brown, 1987; Cox et al., 1990; Spiegel and Fitzgerald, 1990; Otto, 1992; Yang, 1989) have not been indicated. The number also refers to the reported number of texts analysed, therefore the corpus from which the texts are drawn may be larger, and the presentation of results and diagrams in the papers, less. An important question, one that will be followed through this chapter, is whether the research question demanded a measure of coherence or an 307 e.g. CHI ∗ ∗ Yes Yes Yes Yes YesYes Yes Yes Yes Yes Elicited Quantities Yes Yes YesYes Yes Yes Yes Yes occurring Yes Yes Yes Yes Yes Yes Spoken-to- Yes Yes Yes Yes Yes Yes be-written (source) (source) 16 5(80 preliminary) 3 (78 prelim?)128 Yes216 Yes 54 Pappas (1985)Pappas (1987) 33 6 Yes Yes Yes Yes Yes Yes Parsons (1991a, b) Parsons (1996) Harvey (2002) 12 Reference No. of textsHasan (1984, Written1985) Spoken Dialogic Naturally Pappas and Brown (1987) Otto (1991)Otto (1992) 28 Fang and Cox (1998) 60? Rentel et al. (1983) Rentel (1988)Cox et al. (1990)Spiegel and UnclearFitzgerald 192 (1990) Yes Yes ‘Pretend reading’ Illustrative survey of work applying CHA Academic writing Table 13.1 Groupings (e.g. register, population similarities) Children’s stories Primary-school stories and reports 308 e.g. CHI Elicited Quantities occurring ∗ speakers ∗ Yes ∗ Yes 1Yes 2Yes Scott (2009, 2010)Lukin (2010) 3 1 Yes Yes* Multiple Yes Yang (1989)Lukin (2002)Butt 4 (1988) 4Fries (1985) 1Refer also to Otto (1991, 1992), YesPappas Yes (1987), 1Pappas and Brown (1987) above Yes (2005) Yes Yes Yes Yes Reference No. of texts Written Spoken Dialogic Naturally Lukin (2013)Garcia Marrugo (2012) 2 Short storyChildrens Björklund (1993) Martin (2003) 1 1 Yes Yes Literary criticism (Continued) News discourse Letters Demantik (2008) 4 Yes Yes Literature Poetry Butt (1984) 8TextbookMagazine text Yes Cloran et al. Fries (1992) 1 Yes Yes Table 13.1 Groupings (e.g. register, population similarities) 309 2 examples Yes Yes Yes YesYes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes (then spoken) as 4 extracts) 6024 + 56 Yes Yes Yes 3 Yes Yes Yes 18 Yes Yes Yes Yes may include additional e.g. visual/multimodal, more complex conditions. Hasan (1994) 1 Cloran (1999) 1 (presented Butt et al. (2010)Khoo (2013) 1 1 (2 prelim) Armstrong (1996)Bottenberg 96et al. (1985) (also control comparison) Hedberg and Westby (1993) Hedberg and Fink (1996) (also control comparison – Yesprimary aged stories) Yes Yes Yes Armstrong (1991) (also control comparison) Armstrong (1987a) Armstrong (1987b) ∗ Narrative stories Cohesive Harmony Index) = Dissertation defence Mother–child discourse Psychotherapy discourse Learning disabled discourse Aphasic discourse (CHI 310 Society in Language, Language in Society exploration of texture. In many instances, there is an alignment with coherence and the application of quantitative measures, though this is not always the case. For example, Butt et al. (2010) were concerned with questions of coherence and yet did not apply the coherence indices in full; and Scott (2009) did apply coherence measurements though a coherence measurement was not the purpose for which the method was applied. Following the work of Hasan, a large clustering of studies has focused on the application of CHA to children’s stories and texts, testing Hasan’s claims and applying the method to research questions in the development of children’s literacy. These have included studies in coherence across spoken and written conditions (e.g. Pappas, 1985); cohesion and coherence in children’s writing and the variation with content, quality and grade (Spiegel and Fitzgerald, 1990); comparison of reading performance with written text coherence (Cox et al., 1990); ‘pretend reading’ or story reconstruction (Otto, 1991, 1992; Pappas, 1987; Pappas and Brown, 1987) and longitudinal studies of children’s narratives and stories (Rentel et al., 1983; Rentel, 1988). CHA has allowed for both the reporting of statistical results of large numbers, such as the 216 texts in Rentel et al. (1983), and also for formation of further hypotheses resulting from findings, such as in Rentel et al. (1983) where ICs and SCs followed a different course of development in written texts – ICs dominating the earlier texts with an expansion of similarity relations only once a child was able to exploit the resources of reference. The method has also been used to compare populations where coherence may be ‘at risk’. In children’s literacy, for example, Hedberg and Fink (1996) compared ‘language-learning disabled’ children and their peers, reporting significant differences in cohesive harmony between groups, and also that these were differences rather than delays in chain construction and interactions. Parsons (1991a, b, 1996) and Harvey (2002) have also used CHA as a way of comparing coherence in academic writing of native and non-native speakers; and Demantik (2008) in the comparison of letter writing by native and non-native speakers. In aphasic discourse, Bottenberg et al. (1985) was the first to apply CHA, reporting fewer chain interactions and a greater variability in cohesive harmony to a control comparison, with significant differences across three conditions of oral narrative in the aphasic group. Armstrong (1987a, b) followed, pursuing the application of CHA to aphasic discourse and reporting on the linguistically significant 50 per cent criteria, the correlation of cohesive harmony with listener perception and the effects of context through six elicited text types. Armstrong (1996) also tested whether the CHI may be practical as a measure of performance, investigating the aspects of cohesion and cohesive harmony that may be most amenable to change (and therefore most likely to affect coherence) in a 12-month longitudinal study of aphasic patients post-stroke. In both Armstrong (1987a) and (1996), modifications were required for the method, with a discussion on pathological Kristin M. Khoo 311 phenomena such as paraphasia and perseveration and the effect on cohesive harmony ratios.3

13.4.1 Cohesive Harmony Analysis in clinical intervention In addition to CHA as a method for analysis and measurement of coherence in clinical research, CHA has also been proposed as a potential clinical tool to guide intervention, ‘the analysis being intricately interwoven with subsequent treatment goals’, for clinicians treating individuals with aphasia (Armstrong, 1991, p. 47) and individuals with language disorders (Hedberg and Westby, 1993). Armstrong (1991) has suggested that the construction of chains and interaction diagrams in therapy may provide a visual tool with which to work:

In treatment, then, it may be possible to introduce a discourse visually to a patient, providing him with component parts, i.e. the chains, while at the same time incorporating the concept of continuity and relationships between the words he is using to create a functional meaning. (Armstrong, 1991, p. 47)

Armstrong (1991, p. 46) also suggests that cohesive harmony diagrams are of particular benefit in populations where ‘normal coherence’ may not be achievable and clinical intervention is targeted to the strengths of an individual. In this way, patterns of cohesive harmony not only allow for identification of errors and difficulties, but offer a visualization of the textural resources for improvement of coherence. Hedberg and Westby (1993) have also promoted CHA as a tool for intervention, providing a clinical version of CHA in addition to the ‘standard version’ for use with language-disordered populations. Hedberg and Westby (1993) include in their publication a step- by-step description of the adapted clinical version with illustrative texts and a ‘Cohesive harmony coding sheet’ for analysis. The guide also comes with a coding scheme for non-cohesive devices, instructions on how to interpret and examples of suggested interventions based on CHA patterns. It is important though that Hedberg and Westby restrict application to the analysis of stories and refer to chains as ‘participant’ and ‘process’ chain categories, which will be discussed further in section 13.5.4.

13.4.2 Cohesive Harmony Analysis as a measure of coherence CHA began as a measure of ‘coherence’ in text, seen through a particular understanding of coherence, as noted earlier:

...coherence in a text is the property of hanging together (Hasan, 1984, p. 183) ...cohesive harmony is the lexicogrammatical reﬂex of the semantic fact of coherence. (Hasan, 1984, p. 218) 312 Society in Language, Language in Society

The method and theory of cohesive harmony are consistent with Halliday’s position that:

For a text to be coherent it must deploy the resources of cohesion in ways that are motivated by the register of which it is an instance; it must be semantically appropriate, with lexicogrammatical realizations to match (i.e. it must make sense); and it must have structure ...(Halliday, 1994, p. 339)

There are, however, arguments as to whether cohesion and connectedness can be turned into coherence, i.e. whether coherence is a text-inherent property, although it is beyond the scope of this chapter to review these arguments and the theory of sign and language from which these derive. Factors that affect the discussions of coherence may be readily acknowledged – the nature of phoric references, degree of ambiguity dependent on the extent of shared knowledge with the text creator, definitions of text and discourse – though any measure will be unable to account for all factors affecting ‘coherence’. In CHA, Hasan has proposed a relatively objective measure that correlates with the perception of coherence, and it is through an analysis of texture and specifically the components of CHA that this is achieved. Thus, since CHA has been employed as a coherence measure in many of the papers in Table 13.1, it is important that CHA does correlate with a listener/reader’s judgements of coherence. Hasan (1984, p. 218) reported that coherence rankings by readers consistently correlated with cohesive harmony scores. A number of studies have since tested this claim through experimentally constructed scales (e.g. Armstrong, 1987a, b),4 comparison with paper gradings/comments (e.g. Harvey, 2002) and through users of the technique reporting on its success (e.g. Martin, 2003). Rentel et al. (1983, p. 37) argue that the value of the CHA method is (1) the strong theoretical grounding of the coherence measure combined with (2) the means to reli- ably rank texts for coherence that are consistent with, but not reliant on, reader judgements. CHA has also been compared with other ‘coherence’ measures (e.g. Spiegel and Fitzgerald, 1990; Fang and Cox, 1998). Spiegel and Fitzgerald (1990) replicated a previous study that used Bamberg’s (1984) ‘multidimensional’ coherence construct (Fitzgerald and Spiegel, 1986, as cited by Spiegel and Fitzgerald, 1990) which revealed a different profile of results for CHA, reporting that cohesion and coherence in children’s writing varied according to text context, though not according to writing quality or grade level. Fang and Cox (1998) made comparison to a measure of ‘textual quality’ scoring on dimensions of organization, participant, process and tense, from which CHA was reported to be ‘a more consistent and reliable measure of textual Kristin M. Khoo 313 quality’ (Fang and Cox, 1998, p. 351).5 Pappas (1985) also found a different patterning of results to studies that used other analysis tools such as narrative form schemas that give prominence to events in text analysis. Hasan’s method has been utilized as a measure of coherence in many of the studies seen in Table 13.1. There is, though, a difference in the measurement of coherence and in examining how coherence is constructed in the creation of text (texture). A number of studies have also adopted CHA as a framework for the analysis of texture, and while these studies may improve the reliability and validity of the tool through the application to a wider variety of text types and registers, it is noted that studies such as these often do not apply the quantitative component of the tool, as seen in Table 13.1.

13.4.3 Cohesive Harmony Analysis as evidence for structure and context While the method of CHA looks to the patterning of non-structural cohesive devices for the coherence of text, this does not disregard the role of structure in the theory of text; Hasan (1985) is very clear that text is both unity of texture and structure. Furthermore, Hasan (1985, p. 99) proposed structure to be the link between texture and context (Hasan, 1985, p. 99). Despite this, there have been criticisms that CHA is an ‘incomplete analysis of the representational unity’ of a text (Friedman and Sulzby, 1987, p. 301), with implications or criticisms that the components of CHA are the only sources of unity or coherence. Friedman and Sulzby (1987), for example, raised the issue of ‘clausal order’ in text – that if a solution was presented in the midst of a problem, the coherence would be affected, though the order of clauses would not affect the token counts and consequently the quantitative measure in CHA. While structure is not directly accounted for within CHA, the move to examine texture and structure together has been taken up by a number of studies. One method, which allows for the tracking of clausal order and structural elements, is through combination with an analysis of generic structure potential (GSP).6 Pappas and Brown (1987) and Pappas (1987) compared ‘pretend’ readings of a children’s storybook to the source text with the patterns of texture – seen through ICs, SCs and interactions – examined within global structural elements. In this comparative method, additional token types were analysed including approximate, ambiguous and ﬁve types of extrapolated tokens. These further discriminations were made to indicate such information as, for instance, whether a token occurred in the same chain as the source text, though in the wrong structural element. Pappas (1987) reported on one case study where all instances of ambiguous tokens fell within two structural elements – the Placement and Initiating Event. Using this method of combined techniques to address a particular research focus, token counts were calculated per global structural element, though no coherence measurements were given. 314 Society in Language, Language in Society

There are, though, papers in which evidence for generic structure of text is seen in the CHA (e.g. Cloran et al., 2005). Hasan (1985, p. 115) reported that ‘cohesive chains display a close relationship to the structural movement of the text’ in particular text types, and since this, the extension of chain interaction patterns in a number of registers has begun to be explored. CHA has provided a way of showing variation and consistency within register, through the relation of texture and structure to context. The following two quotations explain this:

Assuming that a paradigmatic description of the potential of contextual components is viable, we may claim that the options towards the primary end of the network are realizationally related to the text’s structure while those towards the more delicate end would contribute to its texture. In other words, the relatively general nature of the contextual conﬁgura- tion is realized as structure; its more speciﬁc aspects which render it an ‘individual’ occasion for talk are realised as texture. (Hasan, 2004, p. 25) No two texts belonging to the same register are expected to be exactly alike ...As we move further to the right end of the network, the options lose their power; instead they become critical to the texture of the text, emphasizing its unique instantial nature ...It is elements of meaning such as these that will enter into texture, creating some kind of cohesive harmony pattern which will be unique to each text, though generalizations can be made about the text on the basis of the cohesive harmony patterns. (Hasan, 2009, p. 188)

Cloran (1999) provides evidence for structural boundaries through CHA with evidence of semantic shifts seen not only in the onset and disappear- ance of chains but in their interaction patterns in the logogenesis of discourse. This is consistent with Hasan’s claims that cohesive harmony is ‘the manifestation of the topical continuity of a text’ (Hasan, n.d., p. 71). The application of CHA to the recorded spontaneous interaction of mother–child dialogue shows the flexibility and value in the tool as the chain interaction relations were used to constitute evidence of semantic boundaries and contextual configurations, through representations of caregiving, instruction, casual conversation and control within the same material situational setting. A similar application of the tool is shown by Hasan herself, in the extent to which chain interaction could provide a realization of structural elements in the analysis of dialogic interaction of a dissertation defence (Hasan, 1994). Following from these studies, CHA has been employed as a framework for discourse analysis with a wider research focus than the measurement of coherence. Lukin (2013) has used CHA as linguistic evidence of texture and Kristin M. Khoo 315 structure for realizations of the context of situation through system network diagrams. Similarly, Scott (2009, 2010) uses CHA to explore consistency and variation in register, with comparison across texts considered evidence for shifts in contextual configuration. Garcia Marrugo (2012) has used CHA as a tool to identify the structural and textural features of organizational patterns of news reporting. In Lukin (2010), claims about register are based on evidence of CHA and Cloran’s Rhetorical Unit analysis, with cohesive harmony patterns providing evidence of how elements are connected. Results from these studies provide support for the value of CHA as a framework for analysis beyond the measure of coherence, based on theoretical foundations of SFL theory:

Situation type, at a high degree of speciﬁcity, is relevant to texture; you could see it as the motivating force of texture. But by the same token, the facts of texture construe the very detailed aspects of the situation in which the text came to life. (Halliday and Hasan, 1985, p. 115)

13.5 A focus on method and development emerging from applications of CHA: considerations and challenges

This section is directed to areas of method and analysis requiring discussion, further investigation, and development in the current method of CHA. What, then, is important to cohesive harmony, or needs to be included as a measure of coherence or texture? Can changes be integrated to the methodology, or mapped onto current methodology?

13.5.1 The quantitative and qualitative approach The emphasis of the quantitative component is clearly seen in Hasan (1984), where the approach was devised as a measure of coherence – one that could rank texts. However, as can be seen from Table 13.1, there are applications of CHA that only include the presentation of chain interaction diagrams or in-text discussion of cohesive harmony patterns, without the quantitative component. While an in-text description of CHA may appear to be at odds with the quantitative method first proposed by Hasan (1984, 1985), this has arisen in part through the move to a larger range of research questions including arguments about structure and context (e.g. Cloran, 1999; Lukin, 2010, 2013), though also in part due to the lack of a systematic method for resolving the difficulties that arise across text types, and the difficulties in visualization and presentation of results. There is a tension between capturing a quantitative and replicable approach – especially as numbers increase in length (number of clauses) and sample size (number of texts) – and the interpretation of these texts without reference to a full set of texts, diagrams or analytical decisions that may be specific to the 316 Society in Language, Language in Society research design, or instantial to the text. The quantitative orientation, is however, of importance where CHA is employed as a measure of coherence, and has been utilized, as shown in Table 13.1, in studies that require comparison of texts, or the communication of results to a research community that demands numerical evidence – even if tools are still evolving. A quantitative orientation also suggests consistency in application and reliability of CHA. In studies that have employed a quantitative orientation, issues such as ambiguity need to be resolved, since the ability to offer multiple interpretations is restricted to some degree. In Pappas (1985, p. 182) all possible options and potential interactions are considered for ambiguous items in children’s texts and credit given for possible interactions, modifying the computations to distinguish texts that include ambiguities from those that do not. Similarly, Hasan (1984) provides multiple calculations for comparison. In Butt et al. (2010) and Khoo (2013) there are particular issues with exophora and ambiguity, and though the identity or similarity of the semantic content is more important to the creation of texture and coherence than the content itself (Hasan, 1985), difficulty in resolving exophoric reference makes it hard to determine whether endophoric repetition is indeed a relation of identity, even when examining the chain patterns. It is then difficult to apply quantities without resolving ambiguities (cf. Hasan, 1985; Butt et al., 2010; Khoo, 2013). Even where a quantitative approach has been taken, there are many papers in which the chain interaction diagrams are required to support the counts, since the quantities have not necessarily captured certain features or differences in text (cf. Armstrong, 1996; Khoo, 2013). The chain interaction diagrams allow for a visual representation of the logogenetic patterning of connections and continuities of texts (when displayed in their full form). This is one of the strengths of the technique, which is especially valuable for identifying and mapping localizations in text that are not captured in or are dissonant with respect to the quantitative results (cf. Armstrong, 1996; Butt et al., 2010; Khoo, 2013). Diagrams allow for the mapping of chain length, and therefore topical consistency and change, chain interaction patterns through focal chains (a chain that interacts with a large number of chains), bifurcated or chain conjunction/dysjunction, simultaneous interaction (chains connected at the same time) or staggered interaction, and breaks in chain patterns. The patterning within chains allows for an iconic view of ‘bunchings’ of tokens, both relevant (chain forming) and central (interacting) that may be discussed in terms of the notion of tight and loose texture, as in Halliday and Hasan (1976). If CHA is not to be employed as a measure of coherence, or if there are other issues affecting the research methodology such as the application to dialogic text, the question of prioritizing the iconicity and visualization of the text through cohesive harmony diagrams, and the value added through a quantitative orientation and coherence scoring must be considered. Kristin M. Khoo 317

13.5.2 Cohesive Harmony Analysis and dialogic texts: considering psychotherapy discourse In applying a measure of coherence to dialogic text – especially in discourse where there is a question or interest in the coherence of one particular speaker – it is important to take into account that the numeric values of CHA represent the dialogic text as a coherent interaction, not reflective of individual speakers. The challenges of CHA and dialogic text will be briefly mentioned here in relation to my own research area in psychotherapy discourse. Research by Butt et al. (2010) and Khoo (2013) both refer to the conversational model of psychotherapy, where treatment is focused on restoration of the self through therapeutic conversation (cf. Meares, 2012a, b). Butt et al. (2010) first applied CHA to an instance of dissociation from within a psychotherapy session, in which a therapist may report an altered perception of coherence, fragmentation and connectedness. In exploring CHA as a tool to measure and visualize fragmentation, a high level of coherence was seen diagrammatically (though not with exact values numerically), and shown to attribute to the characteristic of ‘fusion’ in the instance. Similarly, in Khoo (2013) a high level of coherence in therapeutic interaction was shown to mask local disturbances in coherence. One of the areas raised in this research, and consistent with psychotherapy literature, is the impact of the therapeutic relationship and its bearing on textual realization of the patient and the overall coherence of the text.7 Looking at CHA and dialogic text more broadly – naturally occurring, spoken, dialogic interaction can be found in Hasan (1994), Cloran (1999), Butt et al. (2010) and Khoo (2013), and an elicited dialogic task in Armstrong (1987a). Of these, Khoo (2013) and Armstrong (1987a) apply token counts and coherence indices. Armstrong (1987a) analyses dialogic texts in two ways, as a measure of both speakers, and secondly, only including the clauses of an aphasic speaker. A similar approach is discussed in studies of mother– child dialogue (Hasan, 1989, p. 249) where there is a shift in emphasis to the individual semantic orientation within the dialogue. In my own research, however, I have only applied measures to the text as a co-constructed interaction (cf. Khoo, 2013). The benefits of a quantitative orientation and application of coherence measures, as seen in section 13.5.1, must be balanced with the research questions and the ability to effectively communicate numeric results. In relation to psychotherapy discourse and the nature of treatment, a text that is poorly connected and ‘incoherent’, though brought together in a significant point of interaction or change in pattern, may be judged a more valuable therapeutic text than one more consistently cohesive or coherent. It must also be clear how the discourse problem in research areas such as psychotherapy discourse for patients with borderline personality or disorders of self – as in Butt 318 Society in Language, Language in Society et al. (2010) and Khoo (2013) – is concerned with cohesion and coherence of the patient’s ‘text’ and the issue of coherence in relationship. The decision to include both patient and therapist turns within the scoring draws from the theory of cohesion and cohesive harmony, irrespective of monologue or dialogue. In removing the therapist’s contribution to the dialogue the cohesive relations and interactive patterns would not be visible since sematic relations are present across turns; nor would the scores be interpretable, even if the tokens were coded prior to removing the turns. Each meaning realized in text provides the context for the next, and therefore it would be misrepresentative to present the scores as reflective of a single speaker – especially in a supported context such as psychotherapy. The coherence scores are therefore reflective of the interaction and quantitative results need to be discussed in light of the shared role of the participants. There is scope, though, to see the contributions of patient and therapist through the support of the chain interaction diagrams and the shading of turns. While there may be an imbalance of clause numbers, for example the therapist with fewer, the role of the therapist may be crucial in maintaining texture and holding the chains together. Looking at the chain interaction diagrams, it is possible to examine where tokens cluster and chain interaction occurs, both within and across turns. Also, the order in which the contributions occur may be of interest – whether cohesion and echoing occur with therapist following the patient, or patient following therapist. This type of evidence may also reflect techniques that therapists use within sessions,8 with therapeutic techniques also coding as functional echoes in coding that increase the cohesive score of the text. A CHA of dialogic text such as therapeutic conversation can show how many tokens are integrated into the text through token counts and diagrams, and how tightly they are integrated (based on the relations of identity and similarity) in chain interaction diagrams. In the findings of a highly ‘coherent’ text by measure, prominence may also be given to tokens that are not central, and may be of particular interest in this context.

13.5.3 Visualization and representation of Cohesive Harmony Analysis While iconicity is one of the strengths of CHA, there is also development required in the areas of visualization and representation. It can be difficult to display chain interaction diagrams given the limited space available in journals and similar publications, with the problem compounding as the number or length of texts increase. In some papers where large numbers of texts are analysed, no diagrams are provided (cf. Spiegel and Fitzgerald, 1990; Hedberg and Fink, 1996; Fang and Cox, 1998). In others, selected chains or sections of diagrams are provided for in-text discussion (e.g. Lukin, 2010, 2013). The publication of diagrams in a compressed form can show an overall shape with the patterning of chains and their connections, though the tokens may be obscured. The value of this visual representation decreases Kristin M. Khoo 319 as chain interaction diagrams become more difficult to manage and display. For instance, as there are additional chain conjunctions and dysjunctions, this is difficult to manage in a two-dimensional space that displays the patterning within the context of logogenesis and interaction. This contributes to difficulty in representation of results, and consequently their communication. If it is not possible to display a large text logogenetically, then a researcher must ask: how can one extract the significance of a large diagram and model it so that it is communicated effectively to an audience? Scott (2009, 2010), for example, has constructed schematic diagrams to represent the interaction type (i.e. participant to process) and strength (i.e. number of interactions between chains). It is challenging though to represent additional variables such as the dialogic nature of text, or the development of multiple sense relations, multiple interaction types, and bifurcation (identity and similarity) that may exist within chains in schematic diagrams. Addi- tionally, summary tables may be appropriate only for shorter texts or specific purposes, and are not necessarily easily communicated to non-linguistic audiences. While doctoral theses have scope for the attachment of appendices, and in an electronic format to allow readers to zoom and navigate between overall shape, and view sections or tokens, this only manages one aspect of the logistical problems. The lack of ‘readable’ chain interaction diagrams relies on a standard application of method and consistency in decision-making, including across lexis categories which are not ‘precoded’. This is an area that may be considered especially important in a tool that is still developing and requires calibration across registerial and text-based issues. Further work on representing analysis, modelling complex information, and engagement with computational methods needs to be explored.

13.5.4 Consistency in applications of method Although the method of constructing chains and marking out interactions based on grammatical relations has been picked up through the literature, there are differences in analysis and coding across studies due to misinterpretation of the technique,9 changes in the procedure to address specific research questions, participants or texts, decisions related to complex or instantial sets, or register specific constraints. In moving from theory to practice, discussion related to research methodology needs to be communicated clearly – particularly where measurement of coherence and quantitative scores are applied, and consistency and replicability are important. In adopting a quantitative orientation, Hasan (1984) prioritizes methodological and analytical consistency. Many of the methodological or analytical decisions have effects on the interaction scores, which become more significant for larger texts. Although decisions may need to be specific to the discourse question addressed, or the register in which the linguist is working, the implications for the linguistic tool must 320 Society in Language, Language in Society also be considered. CHA was developed on a particular text type, and therefore certain complexities were not encountered in the original work as Hasan acknowledges (Hasan, 1984). And, as with many research tools, it is developed according to the purposes of the research question and then refined as hypotheses are tested and difficulties are communicated to the research community. Areas in which consistency is at issue across different scales include principles of chain formation, lexical rendering of texts, analysis of lexical cohesion, and spoken and written language. These are briefly described in the following paragraphs.

(a) Principles of chain formation One of the largest variations from Hasan’s method of CHA is in the principles of chain formation. Rather than forming chains on the basis of identity and similarity, in a cluster of papers chains are constructed and referred to as ‘participant’ chains and ‘process’ chains (e.g. Cox et al., 1990; Otto, 1991, 1992; Hedberg and Westby, 1993; Hedberg and Fink, 1996). This method of chaining, however, suggests that chains are constructed with a view to interaction – an experiential analysis formed into chains.10 This is not simply an ‘analytic’ decision relevant to the register, but a change in the theoretical foundation of the method.

(b) Lexical rendering of text In Hasan’s methodology, a text is lexically rendered for CHA (cf. Hasan, 1984, 1985) (see e.g. Cloran, 1999 for lexically rendered text as consistent with Hasan). However, there are variations in the consistency of lexical rendering, with consequences for both chain patterns and coherence scores. Parsons (1991a, b) reports that sometimes lexical rendering can just too dramatically alter a text. Martin (2003, p. 5) suggests that lexically rendering the text makes the analysis more effective ‘to make explicit all of the ellipsis and substitution so that points of interaction are not missed’, although Martin (1992, p. 419) also suggests lexical rendering of substitution and ellipsis, though not pronominal reference.11 Changes to elements of lexical rendering may also be specific to the register and research question. In Butt et al. (2010) and Khoo (2013) lexical items in ICs are preserved in the original lexical form for chain diagrams especially in regards to self-reference.12 Similarly, inclusion of ellipsis and expansion of responses may obscure the information explicitly encoded in the text, which may be critical to questions of self in psychotherapy discourse.13 A difficulty that may arise across register concerns lexical items that refer, either anaphorically or cataphorically, to a portion of text. These items, termed ‘extended reference’ in Halliday and Hasan (1976), Hasan specifically refers to in CHA: ‘their interpretation reinvokes the meaning relations of the Kristin M. Khoo 321 referred to sections, thus echoing patterns of the interaction’ (1994, p. 153). While such items do work cohesively, a lexically rendered token may refer to more than one chain. Harvey (2002) refers to one text in which ‘this’ was referring to the previous 17 clauses and if lexically rendered, joins 26 chains. Clearly this has ramifications for chain patterning and quantification. Furthermore, the span of text to which such items refer may not always be clear. For example, in Khoo (2013) it is not clear whether the lexical item ‘that’ in ‘if that makes sense’ was referring to the previous two or four clauses, and hence spanning three or six chains. If the item is left in its unrendered form – whether or not the extent of the deixis is clear – a decision needs to be made as to the placement: whether in one chain (likely misrepresentative) or as a peripheral token, especially if it is a non-resolvable phoric item (even though there may be a cohesive element).

(c) Analysis of lexical cohesion There are a number of issues regarding consistency that arise from questions pertaining to lexical cohesion, such as instantial lexis (Hasan, 1984, 1985) and the determination of tokens (e.g. Scott, 2009; Lukin, 2013; Khoo, 2013). Questions regarding ‘phrasal’ tokens (e.g. ‘fit as a fiddle’) or the degree to which lexical items should be separated or kept together as representa- tive of a particular identity (especially as more complex constructions with embedding occur) must be considered (cf. Khoo, 2013). As Hasan reports, the difficulties are not inherent in cohesion, though the extent to which there are challenges in the analysis varies across register.

These problems are troublesome and affect the analysis of lexical cohesion at every step; but it must be pointed out quite clearly that they are not problems arising from any particular view of cohesion. Rather, their roots lie in our conception of the form of language – and more particularly in our ideas about the relationship between grammar and lexis. (Hasan, 1984, p. 194)

These challenges in chain construction need further discussion, especially across the range of text types and the level of reliability that can be achieved (see also section 13.5.5, Chain formation).

(d) Spoken and written discourse Spoken and written language have ‘different ways of constructing complex meanings’ (Halliday and Matthiessen, 2014, p. 726), for instance in the degree of lexical density and grammatical intricacy. As a result, there are differences in the types of difﬁculties encountered in applying CHA to different text types. One such is consistency in treatment of embedding and grammatical metaphor.14 322 Society in Language, Language in Society

In relation to embedded clauses that are encountered in texts, there have been differences as to the inclusion and treatment of embedded structures in analysis across, and sometimes within, studies. In some instances, these decisions are seen through an examination of the cohesive chains, and in others (cf. Harvey, 2002; Scott, 2009; Khoo, 2013) a discussion of the cohesive relations within embedding, the treatment of different types of embedding and the effect on analysis is seen in the paper. The exclusion of embedded clauses from analysis may significantly affect the results of certain texts – see for example Butt (1984) in the analysis of verbal art where more than half of the lines of Wallace Steven’s ‘Load of the Sugar Cane’ are an embedded qualifier. If though, tokens within embedding are chained, then there can be difficulty in marking out interactions where an embedded clause functions as a participant. Whether the echoes captured within a downranked clause should be marked out as distinct, as suggested with group-level interactions (see section 13.5.5), has not been investigated. In applying CHA to spoken text, there is variation in analytical decisions regarding features such as repetition, incompleteness and abandoned structures; and the potential effects on token counts and coherence. These features were excluded from Hasan though they may be significant in other registers or research questions (e.g Armstrong, 1987a, b, 1996; Khoo, 2013). For instance, Armstrong notes:

I wished to demonstrate the essential meanings which were being conveyed by the speaker and not falsely inﬂate coherence simply by counting repetitions of the same item. And yet I also wanted to put these meanings in the perspective of the pathological context in which they occurred. (Armstrong, 1996, pp. 311–12)

13.5.5 Development of current methodology (a) Chain interactions A number of proposals have been made regarding CHA,15 some of which may improve the method beyond a specific register or text type. Both Lukin (2002) and Armstrong (1996) have noted a difference in the contribution of chain interaction occurring between groups (e.g. Epithet–Thing), and between experiential elements (e.g. participant–process). Currently, both types of interaction are calculated equally within the method, though it is suggested that they may warrant a different weighting. Armstrong (1996) in aphasic discourse, for instance, finds that a coherence score would be inflated through the ability of an individual to construct nominal groups that were cohesive.16 Also, in the coding of chain interaction, verbal projection requires further investigation due to the effects on coherence in specific texts.17 The question of delicacy also arises as more delicate distinctions could be made in regards Kristin M. Khoo 323 to types of mental processes or relational processes. The challenges here need further exploration especially in relation to the application of scoring.

(b) Delicacy in chain formation The level of delicacy required in chain formation also needs to be explored across registers. The separation and combination of chains are often commented upon, and CHA response papers such as Friedman and Sulzby (1987) raise the issue of ‘hyponomy’ as such a problem. Cloran (1999) does not include passionfruit within the food chain since it was important to the exchange, and even within this chain separates passionfruit tokens in diagrams, coding relations as specific, non-specific and general (Cloran, 1999, p. 195). Similarly, Lukin (2013) reports challenges in the constructions of chains, for example where repetition of a proper noun does not necessarily express the relation of co-reference. The formation of SCs as texts increase in length also raises issues as to when co-extension becomes a different field of meaning. As Butt asserts, ‘one can see how a network of semantic affinity may be established, a network which is below the threshold of chain interaction, but which contributes to the sense of meanings implicated across a text’ (1988, p. 91).

13.6 Extending Cohesive Harmony Analysis beyond current methodology

The ﬁnal move in this chapter is to consider whether CHA needs to, or can, be extended18 to include additional components. As can be seen from Table 13.1, structural cohesion19 and organic relations have not been included in CHA. As CHA is not a cohesion analysis, organic cohesion has been omitted, with the view that:

...if there is a high degree of chain interaction in a piece of language, then each individual message in such a piece will be paired/joined with others, whether this fact is indicated explicitly or not; the presence of cohesion by componential devices in general provides the background for organic cohesion. (Hasan, 1979, p. 380)

To include all cohesive ties in an analysis, a cohesion analysis would need to be complete. However, on numerous occasions it has been seen as ‘incomplete’ due to such omissions from the method (e.g. Martin, 2003), despite the reasoning set out by Hasan.20 Chain interaction in CHA is currently completed through experiential meanings21 and though chains/strings have been combined with other relations (for a summary see Martin, 1992, p. 393), these interactions have replaced the experiential function in the respective analysis method. 324 Society in Language, Language in Society

Scope for mapping additional components onto CHA methodology includes the highlighting of Theme (Butt, pers. comm.) and mapping of intonation reflecting cohesion. Currently being pursued in relation to psychotherapy discourse is the mapping of Given and New, and the potential for mapping thematic and informational prominence onto chain and CHA diagrams.22 While there are potential avenues for extension, it must be noted that CHA will always be an ‘incomplete’ analysis, and Hasan has never claimed to capture all resources of texture of coherence. As Fine argues in response to criticisms of cohesion analysis: ‘The danger is in seeing a narrow focus on a particular type of analysis as the final step, rather than a stage of a more comprehensive approach within a broad theoretical framework’ (Fine, 1995, p. 28). CHA, as can be seen from this chapter, is a flexible tool that provides focus that is not so narrow. Cloran et al. (2005) compare CHA as a method of discourse analysis with other methods,23 among others that have made comparative analyses (e.g. Bottenberg et al., 1985). CHA provides a unique contribution to evidence from text, and can be combined with analyses where needed to capture the variables that are required of a research question. Butt (1984, 1988), for instance, provides an example of such combinations of methods in relation to the analysis of poetry.

13.7 Conclusion

This chapter is but an overview of the contribution of Hasan’s work on cohesive harmony and texture seen through the method of CHA. Hasan’s work contributes beyond a body of literature speciﬁcally on CHA, to theory and research applications concerning text and context. This is a tool from which there is much opportunity for evidence, not only in measurement in coherence, but in patterns of cohesive harmony and the relation of patterns to structure and context. A valuable method for addressing a wide range of research questions, the evidence can be considered within the theoretical framework on which it was developed, and the perspectives evolving within a systemic functional approach to language. There is also much to be learned through discussions of linguistic methodology and application across register, text type and population as well as in explorations of visualization and representation of results. As a measure of coherence, an analysis of texture and as a discourse analysis tool, CHA appears to be a method worth pursuing.

Acknowledgements

Thanks go to Annabelle Lukin and David Butt for helpful comments on this chapter. Kristin M. Khoo 325

This chapter was written while the author was receiving an Australian Postgraduate Award scholarship for PhD research.

Notes

1. Refer to Hasan (1985) for a discussion of coding instantial lexical cohesion. 2. Note that no claims are being made that texts within the clusters belong to the same ‘register’ or have the same contextual configuration. 3. One example discussed in Armstrong (1996) was contradiction in discourse, where such instances would be chained as antonyms and increase coherence scores. 4. Armstrong (1987a, b) reports that in aphasic discourse CHI scores correlate with listener ratings on a four-point scale of how much ‘sense’ they make of the text. 5. One note that must be made in reference to CHA literature is differences in the implications of terms, such as ‘writing quality’. There are at times conflations between terms. ‘Cohesive harmony’ and ‘cohesion’ may be considered measures of textual quality (e.g. Fang and Cox, 1998); ‘cohesive harmony’ may be referred to as ‘cohesion’ (such as if an additional coherence measure is included); or ‘coherence’ may be referred to as the ‘perceptual score’ given by readers. 6. Fries (1985) examined lexical sets and process type, considering the patterns within the structural elements of the text. Fries (1985) does not directly discuss the method in terms of CHA or chain interaction, though Fries (2004, p. 43) suggests that the paper was ‘an attempt to relate cohesive harmony ...to the generic structure of a story’. 7. Comparison of conversation or dialogue across register is also required to provide a clearer indication of the extent of differences for psychotherapy discourse. A cohesion study of aphasic patients and controls in monologic and dialogic conditions also suggests differences for supported interactions such as therapy contexts (Armstrong et al., 2011). 8. Examples of therapeutic techniques in the Conversational Model include coupling, amplification and representation, which aim to align with the patient and extend the meaning. A therapist may use the exact wording of the patient, altering the tone of voice or inflection. These techniques may contribute to the ‘shared’ feeling of the relationship, as well as trying to ‘expand the conversation’ at a level of delicacy that does not enhance disconnection (e.g. such as asking a question) (Meares, 2012b). 9. An example of misinterpretation is Yang’s (1989) principles of chain interaction. Yang (1989, p. 237) refers to two criteria for chain interaction, the first consistent with Hasan’s definition, and the second as ‘There are places where they [chains] touch together’. This second criterion appears to be chain conjunction, rather than chain interaction under Hasan’s definitions. 10. When applying the criteria of sense relations, a chain may look like a ‘process’ or ‘participant’ chain, though this is not because they are a ‘mental process’ chain for example, but because of the relations as set out in Figure 13.2. Note also that one of the consequences of chains with no grammatical constraints is that there may be instances in which a chain interacts with itself (especially in the case of a bifurcated chain, i.e. where the chain would be considered ‘representing primarily only one chain’, with a central component and split chains (Hasan, n.d., p. 89)). 326 Society in Language, Language in Society

11. This was suggested in relation to a difficulty with chain overlap in application of CHA, which he proposed was due to a precedence given to ICs (Martin, 1992). 12. The lexical and grammatical form of reference is especially important in this register (see Meares, 2012b for discussion) and therefore important for the psychotherapist audience to be able to read off in a diagram. 13. An example of expansion from Hedberg and Fink (1996, p. 75): ‘she said yes/we are going to the campgrounds with the family/’ (restoration in slanted brackets). Although efforts can be made to mark out ellipsis in chain diagrams (e.g. with ∧CAPS), the potential effects on interaction scores cannot be marked. 14. Scott (2009) encounters grammatical metaphor in the form of nominalization in news texts, and codes two realizations differently, though treats the different experiential relation as an echo of interaction (i.e. where one functions as actor– process–goal, and one as deictic–thing–qualifier). 15. Parsons (1991a, b), for example, proposes a concept of ‘significant chains’ and a distinction between significant and non-significant CTs based on an observed relationship between chain length and coherence. There is, though, a limit to which the effect applies (Parsons, 1996). 16. Armstrong (1996, p. 341) also reported on the ‘depth of chain interaction’, that chain interactions did not increase over time though the number of chains entering into interaction did, and also non-productive interaction (Armstrong, 1987a, p. 146) to which ‘the higher the proportion of non-productive interaction to the productive interaction, the less meaningful the discourse in proportion to the signs produced by the speaker’. It must be noted, though, that both of these findings arise from a particular population. 17. For example, Armstrong (1996) only codes verbal processes as repetitions if the projection was also a repetition, i.e. and then I said, well then I said would not be counted as interacting chains (Armstrong, 1996, p. 319). 18. Fries (2004) considers Lemke’s thematic formations to be an extension of CHA. In Lemke’s method, semantic relations are tracked across texts, though are abstracted away from the instances to construe the thematic formations (cf. Lemke, 1995). 19. Hoey (1991) explores what he terms a ‘complementary perspective’ to CHA, where he attempts to track structural cohesion that may be seen through parallelism in sentence structure. 20. At present, chains constructed of conjunctive relations would impact the scoring since interaction is based on experiential transitivity relations, and therefore tokens may code as RTs but not CTs. 21. Hasan (1984, pp. 212–13) suggests that it may be possible to complete chain interaction based on interpersonal relations, though this has not yet been attempted. The inclusion of the interpersonal through other means – e.g. in a few relevant instances Scott (2009, p. 151) included chains explicitly realizing evaluation and appraisal – has tentatively been explored. 22. Language in Social Life Psychotherapy Research Network, including Butt, D., Meares, R., Moore, A., Henderson-Brooks, C., Smith, B., Khoo, K., Hadfield, G. 23. Cloran et al. (2005) compare analysis of the same text using methods of CHA, rhetorical structure theory, rhetorical unit analysis and phasal analysis.

References

Armstrong, E. (1987a) ‘A cohesion analysis of aphasic discourse’. Unpublished MA (Hons) thesis, Macquarie University. Kristin M. Khoo 327

Armstrong, E. (1987b) ‘Cohesive harmony in aphasic discourse and its signiﬁcance to listener perception of coherence’. Clinical Aphasiology 17: 210–15. Armstrong, E. (1991) ‘The potential of cohesion analysis in the analysis and treatment of aphasic discourse’. Clinical Linguistics and Phonetics 5 (1): 39–51. Armstrong, E. (1996) ‘A grammatical analysis of aphasic discourse: changes in meaning-making over time’. Unpublished PhD thesis, Macquarie University. Armstrong, E., Ciccone, N., Godecke, E. and Kok, B. (2011) ‘Monologues and dialogues in aphasia: some initial comparisons’. Aphasiology 25 (11): 1347–71. Bamberg, B. (1984) ‘Assessing coherence: a reanalysis of essays written for the national assessment of educational progress, 1969–1979’. Research in the Teaching of English 18 (3): 305–19. Björklund, M. (1993) Narrative Strategies in Cechov’s The Steppe: Cohesion, Grounding and Point of View (Abo: Abo Akademi University Press). Bottenberg, D., Lemme, M. and Hedberg, N. (1985) ‘Analysis of oral narratives of normal and aphasic adults’. Clinical Aphasiology 15: 241–7. Butt, D. (1984) ‘To be without a description of to be: the relationship between lexicogrammar and theme in the poetry of Wallace Stevens’. Unpublished PhD thesis, Macquarie University. Butt, D. (1988) ‘Randomness, order and the latent patterning of text’. In D. Birch and M. O’Toole (eds) Functions of Style (London: Pinter), pp. 74–97. Butt, D. G., Moore, A. R., Henderson-Brooks, C., Meares, R. and Haliburn, J. (2010) ‘Dissociation, relatedness, and “cohesive harmony”: a linguistic measure of degrees of “fragmentation”?’ Linguistics and the Human Sciences 3 (3): 263–94. Cloran, C. (1999) ‘Context, material situation and text’. In M. Ghadessy (ed.) Text and Context in Functional Linguistics (Amsterdam and Philadelphia: John Benjamins Publishing), pp. 117–217. Cloran, C., Stuart-Smith, V. and Young, L. (2005) ‘Models of discourse’. In R. Hasan, C. M. I. M. Matthiessen and J. J. Webster (eds) Continuing Discourse on Language: a Functional Perspective (Vol. 2) (London and Oakville: Equinox), pp. 647–70. Cox, B. E., Shanahan, T. and Sulzby, E. (1990) ‘Good and poor readers’ use of cohesion in writing’. Reading Research Quarterly 25: 47–65. Demantik, M. (2008) ‘A cohesive harmony of reader’s letters column in “Time” and “The Jakarta Post” ’. Leksika 2 (2): 32–44. Fang, Z. and Cox, B. E. (1998) ‘Cohesive harmony and textual quality: an empirical investigation’. National Reading Conference Yearbook 47: 345–53. Fine, J. (1995) ‘Towards understanding and studying cohesion in schizophrenic speech’. Applied Psycholinguistics 16 (1): 25–41. Friedman, L. B. and Sulzby, E. (1987) ‘Cohesive harmony analysis: issues of text pragmatics and macrostructure’. In J. E. Readance (ed.) Research in Literacy: Merging Perspectives (Vol. 36th Yearbook) (Rochester, NY: National Reading Conference), pp. 297–305. Fries, P. H. (1985) ‘How does a story mean what it does? A partial answer’. In J. D. Benson and W. S. Greaves (eds) Systemic Perspectives on Discourse (Vol. 1) (Norwood, NJ: Ablex), pp. 295–321. Fries, P. H. (1992) ‘Lexico-grammatical patterns and the interpretation of texts’. Discourse Processes 15 (1): 73–91. Fries, P. H. (2004) ‘What makes a text coherent?’ In D. Banks (ed.) Text and Texture: Systemic Functional Viewpoints on the Nature and Structure of Text (Paris: L’Harmattan), pp. 9–50. Garcia Marrugo, A. I. (2012) ‘The texture of ideology: demonstrating bias in the respre- sentation of the internal conﬂict in the Columbian press’. Unpublished PhD thesis, Macquarie University. 328 Society in Language, Language in Society

Halliday, M. A. K. (1994) An Introduction to Functional Grammar (2nd edn) (London: Edward Arnold). Halliday, M. A. K. (2003 [1976]) ‘Early language learning: a sociolinguistic approach’. In J. J. Webster (ed.) The Collected Works of M. A. K. Halliday, Vol. 2: The Language of Early Childhood (London and New York: Continuum), pp. 23–81. Halliday, M. A. K. (2009) ‘Methods – techniques – problems’. In M. A. K. Halliday and J. J. Webster (eds) Continuum Companion to Systemic Functional Linguistics (London and New York: Continuum), pp. 59–86. Halliday, M. A. K. and Hasan, R. (1976) Cohesion in English (London: Longman). Halliday, M. A. K. and Hasan, R. (1985) Language, Context, and Text: Aspects of Language in a Social-semiotic Perspective (Oxford: Oxford University Press). Halliday, M. A. K. and Matthiessen, C. M. I. M. (2014) Halliday’s Introduction to Functional Grammar (4th edn) (London and New York: Routledge). Halliday, M. A. K. and Webster, J. J. (eds) (2009) Continuum Companion to Systemic Functional Linguistics (London and New York: Continuum). Harvey, E. (2002) ‘Coherence and cohesive harmony in essays written by Thai- speaking ESL students’. Unpublished Master of Applied Linguistics thesis, Macquarie University. Hasan, R. (n.d.) ‘Cohesive categories’. Unpublished manuscript. Hasan, R. (1979) ‘On the notion of text’. In J. S. Petöﬁ (ed.) Text vs Sentence: Basic Questions of Text Linguistics Second Part (Hamburg: Helmut Buske Verlag), pp. 369–90. Hasan, R. (1984) ‘Coherence and cohesive harmony’. In J. Flood (ed.) Understanding Reading Comprehension: Cognition, Language and the Structure of Prose (Newark, Del.: International Reading Association), pp. 181–219. Hasan, R. (1985) Part B in M. A. K. Halliday and R. Hasan (eds) Language, Context and Text: Aspects of Language in a Social-semiotic Perspective (Geelong, Vic.: Deakin University Press). Hasan, R. (1989) ‘Semantic variation and sociolinguistics’. Australian Journal of Linguistics 9: 221–75. Hasan, R. (1994) ‘Situation and the deﬁnition of genres’. In A. D. Grimshaw (ed.) What’s Going on Here: Complementary Studies of Professional Talk (Norwood, NJ: Ablex), pp. 127–67. Hasan, R. (2004) ‘Analysing discursive variation’. In L. Young and C. Harrison (eds) Systemic Functional Linguistics and Critical Discourse Analysis: Studies in Social Change (London and New York: Continuum), pp. 15–52. Hasan, R. (2009) ‘The place of context in a systemic functional model’. In M. A. K. Halliday and J. J. Webster (eds) Continuum Companion to Systemic Functional Linguis- tics (London and New York: Continuum), pp. 166–89. Hedberg, N. L. and Fink, R. J. (1996) ‘Cohesive harmony in the written stories of elementary children’. Reading and Writing: An Interdisciplinary Journal 8: 73–86. Hedberg, N. L. and Westby, C. E. (1993) Analyzing Storytelling Skills: Theory to Practice (Texas: Communication Skill Builders). Hoey, M. (1991) ‘Another perspective on coherence and cohesive harmony’. In E. Ventola (ed.) Functional and Systemic Linguistics: Approaches and Uses (Berlin and New York: Mouton de Gruyter), pp. 385–414. Khoo, K. M. (2013) ‘Cohesive harmony: exploring a measure of cohesion and coherence in psychotherapy’. Unpublished Honours thesis, Macquarie University. Lemke, J. L. (1995) ‘Intertextuality and text semantics’. In P. H. Fries and M. Gregory (eds) Discourse in Society: Systemic Functional Perspectives Meaning and Choice in Language: Studies for Michael Halliday (Norwood, NJ: Ablex), pp. 85–114. Kristin M. Khoo 329

Lukin, A. (2002) ‘Examining poetry: a corpus based enquiry into literary criticism’. Unpublished PhD thesis, Macquarie University. Lukin, A. (2010) ‘ “News” and “register”: a preliminary investigation’. In A. Mahboob and N. K. Knight (eds) Appliable Linguistics (London and New York: Continuum), pp. 92–113. Lukin, A. (2013) ‘What do texts do? The context-construing work of news’. Text and Talk 33 (4–5): 523–51. Martin, J. R. (1992) English Text: System and Structure (Amsterdam and Philadelphia: John Benjamins). Martin, J. R. (2003) ‘Cohesion and texture’. In D. Schiffrin, D. Tannen and H. E. Hamilton (eds) The Handbook of Discourse Analysis (Oxford: Blackwell Publishers), pp. 35–53. Meares, R. (2012a) A Dissociation Model of Borderline Personality Disorder (New York and London: W.W. Norton and Co.). Meares, R. (ed.) (2012b) Borderline Personality Disorder and the Conversational Model: a Clinician’s Manual (New York and London: W.W. Norton and Co.). Otto, B. (1991) ‘Cohesive harmony in stories reconstructed by emergent readers’. Paper presented at Annual Meeting of the American Educational Research Association (72nd Chicago, Ill., 3–7 April), ERIC. Otto, B. (1992) ‘Evidence of cohesive harmony in reconstructed beginning reader texts and complex trade book texts by emergent readers’. Paper presented at Annual Meeting of the American Educational Research Association (73rd Chicago Ill., 20–24 April), ERIC. Pappas, C. C. (1985) ‘The cohesive harmony and cohesive density of children’s oral and written stories’. In J. D. Benson and W. S. Greaves (eds) Systemic Perspectives on Discourse: Selected Applied Papers from the 9th International Systemic Workshop (1982: Toronto) (Vol. 2 XVI) (Norwood, NJ: Ablex), pp. 169–86. Pappas, C. C. (1987) ‘Exploring the textual properties of “protoreading” ’. In R. Steele and T. Threadgold (eds) Language Topics: Essays in Honour of Michael Halliday (Amsterdam and Philadelphia: John Benjamins), pp. 137–62. Pappas, C. C. and Brown, E. (1987) ‘Learning to read by reading: learning how to extend the functional potential of language’. Research in the Teaching of English 21: 160–84. Parsons, G. (1991a) A Comparative Study of the Writing of Scientiﬁc Texts Focusing on Cohesion and Coherence (Nottingham, UK: Department of English Studies, University of Nottingham). Parsons, G. (1991b) ‘Cohesion coherence: scientiﬁc texts’. In E. Ventola (ed.) Func- tional and Systemic Linguistics: Approaches and Uses (Berlin and New York: Mouton de Gruyter), pp. 415–30. Parsons, G. (1996) ‘The development of the concept of cohesive harmony’. In M. Berry (ed.) Meaning and Form: Systemic Functional Interpretations (Vol. 57) (Norwood, NJ: Ablex Publishing), pp. 585–600. Rentel, V. M. (1988) ‘Cohesive harmony in children’s written narratives: a secondary analysis’. In J. L. Green and J. O. T. Harker (eds) Multiple Perspective Analyses on Classroom Discourse (Norwood: Ablex), pp. 281–307. Rentel, V. M., King, M. L., Pappas, C. and Pettegrew, B. (1983) A Longitudi- nal Study of Coherence in Children’s Written Narratives. Research Report: Dept of Education, National Institute of Education. Ohio: Ohio State University, Research Foundation. Retrieved from ERIC Document Reproduction Service No. ED 237 989. 330 Society in Language, Language in Society

Scott, C. (2009) ‘Reporting Armistice: a diachronic functional perspective’. Unpub- lished PhD thesis, Macquarie University, Sydney. Scott, C. (2010) ‘Peace and cohesive harmony: a diachronic investigation of structure and texture in “end of war” news reports in the Sydney Morning Herald’. Proceedings from ISFC 36: Challenges to Systemic Functional Linguistics: Theory and Practice, Beijing and Sydney. Spiegel, D. L. and Fitzgerald, J. (1990) ‘Textual cohesion and coherence in children’s writing revisited’. Research in the Teaching of English 24: 48–66. Yang, A. (1989) ‘Cohesive chains and writing quality’. Word 40 (1–2): 235–54. Part VI Literacy and Education 14 Reﬂection Literacy in the First Years of Schooling: Questions of Theory and Practice Geoff Williams

14.1 Introduction

In 1996 Hasan introduced a proposal for ‘reflection literacy education’ (Hasan, [1996] 2011). Its defining characteristic is that it will equip students eventually to participate in the production of knowledge rather than just its reproduction. It is a deceptively simple proposal, perhaps seeming at first glance to be similar to aspirational statements by educators. It is not, however, simple, because the proposal is informed by the range and depth of Hasan’s understanding of two key issues: how learning depends primarily on language; and how family social positioning can differentially affect students’ access both to knowledge and to the means of producing knowledge (Hasan, 2009). Reflection literacy education is not as well known as many other aspects of her work on, for example, semantics, cohesion, modelling of context, linguistic poetics and semantic variation. This is largely because there have been few, if any, attempts to implement the proposal. Consequently, the main component of this chapter is a report on developmental work on an early phase of reflection literacy pedagogy in school, a selection that seems justified by Hasan’s own view:

If the innovation is to surface at all, it will have to do so from below, through the actual practices of literacy pedagogy. (Hasan, [2007] 2011, p. 244)

Because of its deﬁning characteristic, i.e. enabling all students to learn to participate in the production of knowledge, there is no chance that it will seriously be adopted as ofﬁcial pedagogic discourse and implemented ‘from above’. Where, though, might one begin? Beginning with very early

333 334 Society in Language, Language in Society schooling seems to be a logical place, ﬁrst testing what might be built in those years as an orientation to later, more extended work. First, though, because Hasan’s proposal is not well known I begin by introducing her broad typology of approaches to literacy to contrast reﬂection literacy education with other, better-known approaches and to present her critiques. For consistency, the focus throughout is on early childhood education since many further issues arise, beyond what can be considered here, in other phases of education.

14.2 Hasan’s typology of literacy education

Hasan distinguishes three broad types of literacy education: ‘recognition’, ‘action’ and ‘reflection’ literacy education. Within each of the types there is of course a large range of subtypes, and each of these will lead to different educational outcomes. The fact that subtypes are grouped together does not at all imply that Hasan evaluates their outcomes similarly within the general type. For example, there are crucial differences between genre- based pedagogy and other subtypes of action literacy, as I will discuss shortly. ‘Recognition’, ‘action’ and ‘reflection’ literacy are arranged in a hierarchy of value in Hasan’s work: ‘action’ literacy is much more educationally desirable than ‘recognition’ literacy, but ‘reflection’ literacy has educational qualities that extend beyond ‘action’ literacy. Similarly, there is a hierarchy of knowledge and skill: ‘reflection’ literacy presupposes the knowledge and skills developed in both ‘action’ and ‘recognition’ literacy programmes but extends beyond them. This does not imply a necessary relationship between current pedagogies and the development of this knowledge and skill set, but rather that the outcomes to which each is oriented are presupposed by the other types: the pedagogies for achieving the outcomes are a separate matter.

[Reﬂection literacy] presupposes the expertise of both forms of literacy discussed above, but at the same time it aims to create in the pupil an understanding of reading and writing as bearers of deep social signiﬁcance ...(Hasan, [2003] 2011, p. 229)

This is a point crucial to an attempt to introduce reﬂection literacy education from below given a high general level of community anxiety about the so-called ‘basics of literacy’ and, to be blunt, the crassness of much critical discussion of literacy education. These discussions typically perpetuate very reductive contrasts such as ‘phonics’ versus ‘whole word’ or ‘whole language’ approaches.1 Reﬂection literacy education does not imply any lack of attention to children’s development of phonemic awareness, understanding of grapho-phonemic relations, ability to write a well-structured exposition, read poetry appreciatively and so on. Geoff Williams 335

Recognition literacy education makes learning about elements of language the main goal, and typically the end goal, of literacy development. In social-semiotic terms, recognition literacy is characterized by an intra-stratal orientation, typically in the earliest phase at the phonological/graphological stratum. Pedagogic activities are designed to develop skill in recognizing selected features within those strata. Consequently, understanding how these language features are implicated in meaning making is an incidental, secondary goal of teaching (cf. Halliday, 1975). In Halliday’s terms, there is no perspective on the functions of a feature under scrutiny ‘from above’, from meaning: instead, the perspective is from ‘around about’. To illustrate what Hasan means by recognition literacy, I cite an example from North America, the DIBELS scheme for early literacy development. In this scheme, children are assessed in kindergarten on their ‘nonsense word fluency’ by ‘reading’ configurations such as ‘zuz’, ‘jaj’ and ‘fec’. These examples are drawn from the first monitoring of students’ progress. By the seventeenth monitoring they are expected to have progressed to recognizing ‘muz’, ‘wod’ and ‘woc’, though there is no linguistic explanation as to why ‘muz’ is more difficult to recognize than ‘zuz’ (Good and Kaminski, 2014). Significantly, in recognition literacy military metaphors are frequently used to represent pedagogy. Children have to learn ‘word attack’ skills to ‘decode’ language quickly. Speed is of the essence so ‘drill’ is crucial. In the ensuing ‘battles’ there will inevitably be wounding so ‘first aid’ and ‘recov- ery’ are crucial (Williams, 1985). The main tragedy of all this is that under such a metaphor language is made to be an enemy rather than an intriguing social phenomenon for children to investigate. Beyond the first years of schooling, recognition literacy practices can be entrenched so that they appear to be ‘natural’ and ‘basic’. For example, pedagogy about ‘grammar’ in primary education, and in TEFL education too, is recognition literacy education. It is configured merely as learning to recognize tokens of word classes. The orientation is again intra-stratal, grammar dissociated from meaning, as is inevitably the case where learning about grammar comprises learning to classify items into ‘parts of speech’. What could seem more natural than to first teach young students to recognize classes of words since this appears basic to anything else they might learn about grammar? Equally, using systemic functional linguistic descriptions just to teach children to classify constituents such as Participants, Processes and Cir- cumstances is recognition literacy education. All it does is to replace one recognition category with another. This is very far from education that uses knowledge about language to reflect on meaning making in social contexts. So far my account of Hasan’s work has focused on Western education, but her proposal is also informed by her experience and detailed knowledge of literacy education in India and Pakistan. She argues that recognition literacy education, which is so frequently encountered in these and other 336 Society in Language, Language in Society countries with large, impoverished rural populations, is a betrayal of the hope for what education could and should contribute to personal, national and indeed international development. While recognition literacy education might appear to increase a population’s general literacy rate, it contributes nothing towards developing citizens who can use reading and writing to reflect on social practices and to change them through producing new knowledge. In this specific respect her concerns are very similar to Freire’s (1985) in Pedagogy of the Oppressed. Even in countries with much higher average per capita income, recognition literacy appeals to governments that plan to address literacy education in impoverished rural areas. For example, the Australian Government has just announced that it will allocate millions of dollars to a ‘new’ literacy programme using the ‘alphabetic’ literacy approach for rural indigenous Australian children, commenting:

By introducing or embedding the alphabetic teaching approach, in particular Direct Instruction or Explicit Instruction, into classrooms, teachers will strengthen student learning. (Australian Government, Department of Education, 2014a)

Recognition literacy education derives its pedagogy from the second half of the nineteenth century when universal, compulsory schooling was first introduced in recently industrialized countries, when professional expertise was generally low, and printed resources were scarce. Its approach is mechanical, and indeed that metaphor was often used, as in ‘the mechanics of reading and writing’. But its approach is linguistically chaotic, however much it prides in efficiency and systematicity in the administration and delivery of literacy education programmes. Recognition literacy wil- fully ignores all that has been learned about the significance of meaning for language and literacy development through the twentieth and twenty-first centuries. Perpetuating it through expenditure of large sums is equivalent to giving funding to resuscitate the automotive industry by using pre-Fordist production techniques. In contrast, Hasan’s second type, action literacy education, is deeply concerned with meaning. This is a very broad type of literacy education in Hasan’s conceptualization, encompassing subtypes whose differences merit deeper scrutiny than can be engaged with here. For example, ‘creative writing’ approaches to literacy pedagogy are grouped with ‘critical literacy’ and ‘genre-based pedagogy’, amongst others. So the obvious question is: why does Hasan group them together? Answering the question can, I think, usefully be approached through considering one of the subtypes, genre-based pedagogy, about which Hasan has written extensively and positively. The key issue for understanding action literacy education as a general type is a major Geoff Williams 337 reservation to which I will turn shortly (Hasan, [1996] 2011, pp. 188ff.). The reservation is the criterion for the type. Genre-based pedagogy began during the early 1980s, when Jim Martin, Joan Rothery, Frances Christie and subsequently many other scholars investigated the writing pedagogy widely used in Australian primary schools (e.g. Christie, 1989; Martin, 1985; Rothery, 1990). Their research demonstrated, inter alia, that expectations for writing development were typically deeply implicit in pedagogy. The result was that young children who did not have access to ways of meaning in writing through their home discourse practices that were implicitly expected in school could be radically disadvan- taged. Making these expectations explicit became a research and curriculum development priority. Martin, Rothery, Christie and their colleagues thought about these problems in terms of Bernstein’s concept of ‘privileged and privileging texts’. Bernstein, in developing a general description of the principles that produce specific pedagogic practices, what he called ‘a grammar of the pedagogic device’, had drawn attention to the fact that certain types of texts were privileged in school, and that learners who could produce these types in relevant contexts were themselves very likely to be privileged. Not all of the ‘texts’ in Bernstein’s purview were written texts – to him, a text is anything that attracts evaluation – so any intervention about writing development specifically could not be expected to address all sources of covert privileging bias, but nevertheless written texts were clearly a crucial category.2 The key issue is textual reproduction. The orientation of genre-based pedagogy in primary schools is to support children to learn to effectively reproduce the types of text that are privileged in school, particularly in evaluation. Since there is clear evidence that a large proportion of children cannot learn to do so through implicit instruction, i.e. they clearly do not ‘learn to write by writing’ and by ‘reading widely’, as was often suggested in the early 1980s when Martin and his colleagues began their work, then explicit instruction is required. In primary schools the foci of instruction have been stages of genres, a feature foregrounded in Martin’s definition of genres as ‘staged, goal-oriented social processes’ (e.g. Martin, 1985, 1992). There are many complex theoretical and descriptive issues for educational linguistics raised by this definition. However, here I focus on just the one key feature for this discussion, the pedagogical orientation of genre-based pedagogy. The purpose of explicit instruction under this approach is to enable students to successfully write instances of genres that are privileged in school evaluation. In Hasan’s terms, the setting is necessarily towards successful textual reproduction through instantiation. While that is an imperative for any pedagogy seriously concerned with social justice, the central feature of Hasan’s proposal is that the education 338 Society in Language, Language in Society system should aim for a larger outcome for all learners, which is the linguistic ability to participate eventually in the production of knowledge through critical and imaginative reflection on language in use. She writes: ‘Action literacy is certainly a considerable advance on recognition literacy, and educating pupils to understand how things are done in their culture by using language is a necessary part of education; nonetheless, we also need to go beyond it’ (Hasan, [2003] 2011, p. 229). Hasan has not been the only scholar to make this observation about the reproduction setting of genre-based pedagogy in early literacy development. For example, in the same volume in which Hasan first advanced her proposals Allan Luke also criticized genre-based pedagogy as reproductive (Luke, 1996). Nevertheless, there is a crucial difference between their positions. Whereas Luke, arguably the chief exponent of critical literacy education internationally, proposes this approach to avoid reproductive literacy (Luke, 2012), Hasan argues that the reproductive nature of literacy education is unavoidable in societies in which

the brutal fact about the human social condition is that to live in a society is to collude to maintain at least some of the ways of being, doing and saying that are prevalent in that society; no one has yet shown us that to live in a society can mean anything else, and the natural line of literacy development is largely a process of reproduction. (Hasan, [1996] 2011, p. 192)

In Hasan’s view, critical literacy is itself another form of action literacy education so the criticism of genre-based pedagogy is well founded but equally applicable to critical literacy. Its critiques do little to enable students to independently critique relationships between norms of knowledge and norms of discourse through a deep understanding of the function of language in knowledge reproduction and production. The central point here is the profoundly different view about the importance of learners’ knowledge about language in the production of knowledge because

the possibilities of knowledge and the potentiality of human language cannot be understood in isolation from each other. Discourse and knowledge go hand in hand. (Hasan, [1996] 2011, p. 195)

Luke, writing recently about the foundations of critical literacy, commented:

The term literacy refers to the reading and writing of text. The term critical literacy refers to use of the technologies of print and other media of communication to analyze, critique, and transform the norms, rule systems, and practices governing the social ﬁelds of everyday life. (Luke, 2012, p. 5, emphasis in original) Geoff Williams 339

What, though, of the significance of language itself in literacy education? Language is much more than just a technology, whether of print or speech, so if students really are to possess the means to transform the practices governing the social fields of everyday life, knowledge about language, both as a system and as actualized in specific texts, is essential. Otherwise, without knowledge of how language works in the social fields of everyday life, students’ understanding is constrained to specific instances. These might be generalizable ad hoc, but that is exactly the problem for understanding how language use is implicated in governing the social fields of everyday life. Language understood as a meta-semiotic, as well as a semiotic resource in literacy education, language turned back on itself in J. R. Firth’s famous statement, is surely a precondition for understanding how ‘discourse and knowledge go hand in hand’. Most particularly, it is not possible to understand how ‘the rule systems ...and practices governing the social fields of everyday life’ result in such deeply inequitable, unjust outcomes from schooling without being aware of the processes through which language varies in relation both to its contexts of use and to the social positioning of its speakers. Critique of specific textual practices is indeed very valuable for social change, but enabling literacy education to go beyond reproduction to empower students to participate in knowledge production requires that pedagogy engages systematically with relations between ways of knowing and ways of saying (Hasan, 1996). Reflection literacy cannot entirely escape reproduction for the reason Hasan states. What it might do, however, is to provide students with meta-semiotic resources for a literacy education that can lead to the kinds of social transformation desired by both. Turning then to reflection literacy education, it is so far just a proposal for what education might be: so far as I know, there has as yet been no attempt to introduce it in schooling.3 In trying to think about it in practice I have found it particularly helpful to use Hasan’s formulation, written into a discussion of Popper’s World 3:

Knowledge evolves through reﬂection and analysis of the elements of communal knowledge in relation to social life. and her further observation that:

For knowledge to evolve, what we need is someone – anyone – to perceive the problem(s), to develop the implications; this implies a search for explanations, raising how-questions and why-questions, using existing knowledge only as a point of departure, not as the end of intellectual journey. Vygotsky (1978) would add that these abilities develop within society, and to this extent knowledge depends on a society which allows such mental activities. (Hasan, [2003] 2011, p. 196) 340 Society in Language, Language in Society

14.3 A case example

The first year of schooling is one in which children already do a lot of work towards developing knowledge language. A useful, straightforward question for characterizing the type of literacy education is: what features of language are the objects of attention, through what pedagogic materials? That, in turn, entails a further interesting question: what about language is typically excluded from the first year of schooling but might realistically and productively become a focus of attention? This is the core issue for an attempt to introduce reflection literacy education ‘from below’ and from the beginning. It was Vygotsky who first saw the importance of this relationship between attention and social activity, as the series of experiments he conducted to identify the functions of language in directing attention showed so clearly (Vygotsky, 1981). Language, he concluded, is the primary resource through which children first learn to attend, and therefore the primary means through which they learn to sustain attention voluntarily. The meanings of the signs that first mediate their attention will be restricted to the specific social activities in which they first engage, but as mediation using the sign continues and as the range of social activities in which it is used expands, then the meaning of the sign will also expand. This is now a familiar idea in studies of children’s language development but its significance is yet to be well understood in studies of children’s metalanguage development or, more generally, literacy education. That is, if we turn language back on itself so that language mediates children’s attention to language, what social activities does the mediating (meta)language enable, and through what social activities might the initial meaning of the metalinguistic signs be expanded? Taking work on the lexicogrammatical stratum as an example, if the mediating metalanguage initially directs children’s attention to nouns and verbs, through what pedagogic (social) activities could the meanings of these class terms be developed? The certain answer is the pedagogic activities of recognition literacy. However, thinking about the inclusion of linguistic features that have not yet figured in the kindergarten curriculum but might conceivably do so is actually a complex task. It is simple enough to list possibilities, but the key problem is that children experience an intensive curriculum in kindergarten so the expected pace of learning is typically high, with a potential for detrimental effects for precisely those children about whom Hasan has been most concerned in her work on semantic variation (Hasan, 2009). Addi- tional to the official curriculum, there is also much informal learning to be achieved during the first year through the ‘hidden curriculum’, the problems of managing peer relationships, learning to relate to a large group of older children, to follow institutional routines, and so on. One would therefore not want to add more content if doing so would result in an increase in ‘the Geoff Williams 341 expected rate of acquisition’ (Bernstein, 1990). One approach is to find ways of recalibrating and extending on what is currently taught under action literacy. This is a comparatively simple matter when genre-based pedagogy is in play. The strategy that most strikes me as feasible is to extend and intensify discussion of the relationship between contexts of language use and ways of saying and meaning that ‘fit’ those contexts. In other words, the starting point for this pedagogy could be selected social contexts of language use rather than a particular genre. This seems feasible because it would be a natural extension from aspects of everyday talk that are such a prominent part of the first year of schooling. As children begin this experience there is necessarily a lot of talk about how to mean in salient social contexts. Typically, children suddenly encounter the need to learn to use new, and often subtly calibrated, registers. For example, they need to learn how to talk with peers in the playground, inviting, resisting, sharing, informing, resolving and so on; how to talk effectively to teachers and school administrators about personal needs, learning enthusiasms, and relationship problems; how to relate home language practices, politeness features for example, to different school practices; and so on. Much of this learning already occurs through talk about ways of talking and meaning – teachers, after all, often speak of ‘giving young children the language’ to manage these new needs. However, despite its personal importance from the perspective of developing knowledge about language its value is restricted unless it becomes a part of more general context–meaning explorations from a meta-perspective. It is restricted precisely because it is so contextually specific, but there seems no good reason why some of it could not be freed from contextual specificity to become an object of attention in metalinguistic investigation. Might the more general relationship between contexts of language use and ways of saying and meaning be reflected on, systematically and at a more abstract level, in this first year? If so, it could represent a first move towards developing a deep understanding of realization along with the more frequently used principle of instantiation, whether in the sense of an instance of a genre or an instance of racist discourse, as in critical literacy. Another aspect of current curriculum practice that could readily be extended is play. The types of play one usually sees in kindergarten classes range from free play with apparatus to highly structured play with attrac- tively designed materials for practising the recognition of small units of language, a tradition that extends back at least as far as Walter Crane in the nineteenth century. Somewhat more rarely, pedagogy also includes literacy play through discourse. For example, in one of the classrooms to be discussed shortly, the teacher established a pretend veterinary surgery, complete with white coats and instruments. The children played ‘vet’ for several weeks, speaking and writing extensively in the roles of staff and clients. 342 Society in Language, Language in Society

Imagine, however, that such play were to be extended to be play about the nature of discourse in context. That is, by distorting discourse, exaggerating it, rearranging it, with contexts of its actual use in mind, could young children begin an exploration of discourse in social contexts? Children engage in a closely related kind of activity in their play with jokes, riddles, playground chants and so on. What if, though, play were to be extended to enable children to reflect on how language is used in actual social contexts, and how differences in wording bring about differences in meaning and therefore the relative effectiveness of texts? There is a cue here again from Vygotsky, for whom play was crucial to mental development. Noting that ‘in play it is as though the child were trying to jump above the level of his normal behavior’ (Vygotsky, 1933), he argued that it is through play young children learn to separate signifiers from signifieds, as in his example of a ‘horse’ and a ‘stick’. It is through play that this decoupling and recoupling, so crucial to intellectual development, takes place. Might it be possible for children to engage in metalinguistic play with contexts, meanings and wordings in ways which would also enable a similar kind of decoupling and recoupling, though at such a different level of abstraction? It was with these questions that we began the exploratory work on reflection literacy education that I will now report. However, to foreshadow, as we worked during the final school term a further, previously unimagined possibility arose, one Hasan herself has stated is crucial in developing reflection literacy education but one, frankly speaking, we had not considered a possibility at the kindergarten level. I return to this point after presenting the initial phases of our work. The institutional context for our work was an inner-city Sydney primary school, a comparatively small school of 240 students, but nevertheless one enrolling children from a wide range of cultural and linguistic backgrounds, and family social positions. The children were in kindergarten and were typically five years old, though a few were in the second half of their fifth year, so 4;7+ months. My co-investigators in this work were the two kindergarten teachers and the school principal.4 Our work commenced in the third school term, after the children had been in school for approximately five months. We took as given the need for systematic instruction in learning to read print and to write. The instruction, continuing the school’s long-standing practices, was based on reading and writing across a range of text types: information books, stories, poetry, small recounts of personal experience, classroom organizational texts, selected ‘school readers’, and so on. It also included systematic instruction to develop phonemic awareness and knowledge of grapho-phonic relationships. That is to say, the preoccupations of recognition literacy were part of the programme but were embedded in work on meaningful text, and no less important because of that. Geoff Williams 343

The speciﬁc questions we addressed were:

• How might teachers develop children’s ability to reﬂect systematically on language use in various contexts as they learn to read and write? • Which language concepts might be the focus of study? • How might these be introduced to make them accessible and attractive to kindergarten children?

Our approach was through register and context. While we did not use the term ‘register’ in the classroom we did use ‘context’, at first as an informal aid for starting to talk about a variety of material situational settings (Hasan, 1996) such as playgrounds, shopping precincts and parks, and rep- resentative uses of language within those settings. Much later, context was to become functionally associated with varieties of wordings and meanings. Register was a significant term in use throughout the planning discussions with the teachers because it provided a way of talking succinctly about effects of the patterning of linguistic features in meaning making. Thus, our first answers to the questions above were: (through) context, register and realization. The choice to foreground realization together with instantiation was deliberate. Both are, of course, crucial concepts in SFL theory and both must inform SFL literacy curriculum design, but so far realization has figured much less prominently in SFL primary-school educational linguistics than instantiation. To privilege instantiation so that learners understand texts primarily, even exclusively, as instances of a type is to lose substantial critical and evaluative power for the interpretation and critique of the relationships between texts and contexts. Therefore, our aim was to begin to orient children to notice, in informal terms, ‘links’ between social contexts and meaning, then to develop this sense gradually by exploring effects of wording on meaning in selected social contexts. Appendix 1 presents an overview of the elements and phases of our pedagogy. Our first practical move was to ask the children to begin to notice, then record and talk about, examples of language in use. We began from the assumption that for some children noticing language in use would be a familiar activity in home discourse but for others it would be quite new, and this assumption was confirmed in the first phase of work. We asked the children to become ‘language detectives’ by making a record of examples of spoken and written language they knew about in four ‘contexts’: on the street, in a shop, at the park and at school. These individual records became a focus for class discussion, especially for comparisons of what each person had discovered. Figure 14.1 presents the worksheet written by a girl (age 6;2), and Table 14.1 a transcription of her writing. This first activity was subsequently extended through a variety of other recordings of language in social contexts over about four weeks. For example, 344 Society in Language, Language in Society

Figure 14.1 A child’s initial recording of some examples of language in use

Table 14.1 Transcription of written language in the child’s notes in Figure 14.1

On the street In a shop At the park At school

Words or There’s an ice That’s what Let’s go on Let’s go on the phrases you cream van we need the slide monkey bars might hear My turn! Writing you There’s a stop Vege Shop No Dogs No swearing might see sign. Open No Parking

the children went on community walks to a local park and through the neighbouring shopping precint, taking photographs and notes of street and other community signs. This was followed by extensive discussion of language use at school. Figure 14.2 presents an example, in this instance a collaborative recording on a smartboard of frequent classroom sayings. Most of the sayings are the teacher’s, but there are two from the children: ‘I need to go to the toilet’ and ‘Which reader are you on?’ The next main move in noticing language involved play about discourse through dramatization. The main purpose was to assist children to learn to ‘disembed’ language from its immediate context of use and, by disrupt- ing it in play, to notice the ‘link’ between the particular wording (instance) and context. Much has been written about the signiﬁcance for intellectual Geoff Williams 345

Figure 14.2 Collaborative recording of oral language in use at school in a kindergarten class development of disembedding meditational means from their contexts of initial acquisition but much less about children themselves learning to disembed language from contexts of use to make it an object of reflection, at least at the level of discourse. A nonsense book in the Mr Men series, Mr Topsy Turvy, was our informal aid in beginning this phase. His problem is that he speaks the ‘wrong way round’, saying such things when he enters a hotel as ‘Afternoon good, I’d room a like’. This type of playfulness is of course infectious to children of kindergarten age, so Mr Topsy Turvy’s ways of speaking became a prominent element of discourse in the playground and amongst some families for a short time. These phases of work recall Whorf’s comment about the significance of ‘innumerable small momenta’ in developing and sustaining relations between language and thinking, and of ways of digging oneself out of the linguistic ruts of habitual thought (Whorf, 1956). In themselves the moments of the activities seem inconsequential. However, the patterned consistency of the activities appears to be useful for making relations between linguistic meaning making and social contexts of use visible to young children. It thus also appears to be a significant first step for young 346 Society in Language, Language in Society children towards reflecting on the potential of language to be differently configured in social contexts, and perhaps even eventually to ‘transform’ the ‘social fields of everyday life’. We moved subsequently to focus more specifically on the wordings of commands. The teachers gave the children familiar classroom contexts in which commands were often given and received, such as getting a book from the book corner, organizing the lunch basket and so on. Through dramatization they experimented with ‘long’ and ‘short’ ways of giving commands. They noted that some ways of wording soon became dysfunctional for the required meanings. They also considered some of the interpersonal meanings activated in commands by different wordings, as a result noting that ‘short’ commands were not always preferable, as was their first impression, because they could seem rude. The choice depended on the social context and ‘what we need to mean when we’re ...’. Photographs of signs involving commands that had been taken during the community walks then became the focus of study. After noting these were typically very succinct, the children discussed the contextual features making this necessary and, once again, played with the language to make the commands dysfunctional. The children produced a large number of simple examples such as the following:

Teacher: Michaela, how could we say ‘No stopping’ in a silly way, a long way? Michaela: ‘No stopping because this park is bad!’

Figure 14.3 presents one example of a dysfunctional sign that the children developed, including the choice of image features. Of note is the fact that the wording plays with both experiential and interpersonal meanings. The final move in this phase of work was to talk about the wording of commands in recipes, and it was this work that eventually opened up possibilities for developing knowledge about language that we had not envisaged at the beginning of the experience. This was the possibility of beginning to introduce young children to grammatics.5 We had not initially envisaged introducing any grammatical concepts in kindergarten. Our dominant concerns had been to, on the one hand, interest children in researching relations between contexts of language use, meanings and wordings (i.e. contexts, registers and realization) and on the other, to avoid adding a lot of new content to the existing K (kindergarten) curriculum and therefore increase ‘the expected rate of acquisition’ in Bernstein’s terms (Bernstein, 1990). Before reporting our strategies and findings in this phase, I divert briefly to raise a more general issue about developing children’s knowledge of lexicogrammatical description. Currently, there is no research- based knowledge about how or when to begin the study of grammatics in early childhood education, nor about what sequence of concepts to Geoff Williams 347

Figure 14.3 The children’s example of dysfunctional wording in a command

arrange. By necessity, the few scholars who have conducted research with this age group have selected concepts to best suit their research environments, which have always included some prior instruction in grammatics (e.g. French, 2010, 2013; Schleppegrell, 2013; Williams, 1998, 2000, 2004, 2005). This gap notwithstanding, we do have clear evidence from this research that primary-school children can learn to make grammatical descriptions, to use this knowledge in practical literacy-learning contexts, and that this knowledge can have a strongly positive effect on their literacy learning. The relevant contemporary question is not whether children can learn functionally based descriptions of grammatics, but rather when this learning could commence and how it could best be sequenced. For reﬂection literacy education, learning of grammatics is central because knowledge of grammatics is central to understanding how meanings are realized in wordings, and hence to how these meanings achieve their effects 348 Society in Language, Language in Society in social life. However, de facto, much recent literacy pedagogy in early childhood has foregrounded instantiation over realization through focusing on such questions as ‘What type of text is this? What stages does it have?’ These questions are pedagogically valuable, as has been shown so clearly in genre-based pedagogy, but if literacy education really is to develop students’ abilities to participate in knowledge critique and production then work on textual instantiation must be complemented by work on realization. After all, it is through the patterning of wordings that meanings in social contexts are both construed and activated, the bidirectional elements of realization relations (Hasan, 2010). As Hasan has commented:

Literacy as a process of making sense should therefore take seriously the construing power of language, and this means paying attention to the meaning-construing power of the lexicogrammar. The lexicogrammar of language is not simply dry stuff without any relation to the living of life: rather than presenting it as a catalogue of mind-less categories such as ‘gerundives’ or ‘subject raising’, grammar should be seen as a theory of human experience. ([2007] 2011, pp. 242–3)

Returning to the world of the two kindergarten classes, as we worked on the first set of recipe commands we were intrigued by the strength of children’s interest in the wording so we decided to extend to more detailed questions of lexicogrammatical structure. At first we talked about the activity sequences in these procedural texts. We printed out large versions of each step of a recipe for pizza on single cardboard sheets, and the children sorted these into the appropriate sequence. Subsequently, we took each step, deliberately written in a single clause, and further divided these into experiential constituents, the Participants, Processes and Circumstances. The children then rebuilt the clauses from the stack of constituents and arranged subsets of clauses into an appropriate sequence. At this stage we did not introduce any grammatical terms, working only from meaning. From this experience we moved to comparisons of different ways of wording individual steps in a recipe. This was a means of spiralling back to the preceding work on the efficacy of wording of commands in different contexts, but now focused on effective wording for recipes. We used a strategy that had previously proved productive in our work on grammatics in the Sydney Children’s Development of Knowledge about Language projects (Williams, 2000). We distorted a text to make the wording obviously dysfunctional and then asked children to edit it. The first object text was Mr Very Confused’s Recipe for Fairy Bread, as shown in Figure 14.4. From these discussions, which were a lot of fun, we moved to more specific comparisons of individual commands, again thinking about which wording was clearer in this context. Table 14.2 presents a discussion example for three steps in making pizza.6 Geoff Williams 349

Figure 14.4 Screenshot of ‘Mr Very Confused’s Recipe for Fairy Bread’

Table 14.2 Example of differences in wordings of commands in a recipe (transcribed from a smartboard screen)

Which Instruction is Better?

The first thing you need to do is make Turn the oven on sure you have the oven on Cut the muffin in half Get a knife and cut the muffin down the middle Spread the tomato paste on the Then get some tomato paste and muffin spread it on the muffin

The following discussion excerpt shows how involved these kindergarten six-year-olds were in the editing task. I believe it also shows their perceptiveness about the functionality of the wordings.

Sarah: If we got ...um ...a bit of the top right it would say ‘and spread it on ...the mufﬁn’. But we don’t know what we’re ...‘it’ would kind ...it would kind of tell us what we’re putting on but it wouldn’t tell us what to use with it, if we got rid of the top row kind of. 350 Society in Language, Language in Society

Teacher: Oh, that’s ...fantastic thinking. You’re saying we could get rid of this top row here. And we don’t need to say ‘Then get some tomato paste and ...’ We could just say ‘Spread it on the muffin’. But you’re right Sam, if we just said ‘Spread it on the muffin’, have we said what the ‘it’ bit is? ...No. But if you look on this side do we tell them what they have to spread on the muffin? Several students: Yes ...Yeah ...Yes Teacher: We’ve said, ‘Spread the tomato paste’. So we’ve got a bit more information than just the end of this one but we don’t have so much information that it’s really a long instruction. Hey? ...Great thinking!

Thinking of Hasan’s contributions to the understanding of cohesion and coherence, we might ask, with tongue only slightly in cheek, if such a sense of cohesion in written text can begin in kindergarten will a sense of cohesive harmony be far behind (Hasan, 1984)? Finally, as has been foreshadowed, we took a tentative step into grammatics itself through commands. The work on activity and clause sequencing was the basis because the patterning of wording was so regular. Returning to Vygotsky’s argument about the significance of directing attention through (meta)language, we pointed out that a particular kind of word, the Event, was usually placed first in each step because doing so made the command easier to follow. Event is an element of the experiential structure of the verbal group in English, realized by the lexical verb (Halliday and Matthiessen, 2014, pp. 396–8). Why was Event selected, rather than the more familiar Process as we and others had used in previous research on the ontogenesis of grammatics? For this age group and phase of learning, we hypothesized that a term referring to a single word that was readily understandable from everyday use would be more accessible. Processes usually involve other elements such as Finites and auxiliaries, so developing attention would have been much more difficult at this first stage if it involved multiple words. We envisaged that Event could be the starting point for subsequently building a sense of Process, after which it could be used again to think about process types in a cyclical process of elaboration, perhaps extending again in time to be a focus in the discussion of experiential grammatical metaphor. In any case, selection of Event as the lexicogrammatical focus proved to be effective in this specific context. The children used the concept readily in the comparative editing. In a later exercise, which they completed independently, they were able to identify Events accurately. The analysis examples were: Geoff Williams 351

Figure 14.5 An example of a child’s playful text to display his ability to identify Events Transcription: Get 19 skewers then get chocolet and marshmeleas and mainggos and then wertermelen and then Eat it !!!!!!!!!!!!!!!!!!

1. Cut the fruit. 2. Eat the fruit skewers. 3. Wash the car. 4. Put the egg in the bowl. 5. Next, spread the butter on the bread. 6. Tomorrow, plant the seed in the ground.

Of the 19 students who completed this exercise, 13 analysed all six examples correctly, 5 misanalysed one example, typically the ﬁnal one, which involved both a Marked Theme and ‘plant’, a word that can be used either as a verb or a noun. One child analysed four examples correctly. Given our stated concern to make the children’s initial experience of systematic reﬂection on context, register, meaning and realization as positive and inclusive as possible, the range of these results was particularly important. Of even greater interest to us was the fact that many of the children spontaneously created further playful examples and then analysed these accurately. Figure 14.5 presents one of these texts, which is included to try to convey some sense of the playfulness with which the children engaged in this work.

14.4 Concluding comment

In overview, the evidence from this study clearly justifies systematic further research into kindergarten as a site for beginning reflection literacy education. Our evidence shows that children of this age can begin to enquire, playfully and reflectively, into the nature of language itself, including 352 Society in Language, Language in Society specific features of social context–meaning relations and lexicogrammar. A key point is that children appear to be able to begin to develop an understanding of realization, complementing the orientation to instantiation in genre-based pedagogy. To reiterate, coming gradually to understand realization ‘from above’ as activation and ‘from below’ as construal (Hasan, 2010) is, in my view, crucial to achieving the outcomes desired for reflection literacy education. Many further questions of course arise about sequencing of concepts, pacing of instruction under different demographic and linguistic conditions, and developing more pedagogic strategies, to name just a few. Much remains to be done to form a well-sequenced curriculum over the primary-school years, and to determine resources for teachers to be well equipped to use it. The source of such a curriculum is highly unlikely to be the producers of official pedagogic discourse, as the introductory quote from Hasan states. But that does not mean it cannot be done. Significant work towards it could be completed using digital and online resources at an international level, and these could be ‘infused’ into existing hospitable curricula, such as those using some of the theoretically well-informed genre- based initiatives. We have barely begun to touch the potential of digital technology to support children’s learning about language using new visualization resources. Even some reflection on language in use using the optics of realization and instantiation is, on this experience, likely to provide a very powerful resource for young children’s orientation to uses of literacy.

Appendix 1: Summary table of pedagogic activities7

Pedagogic Description Comment activity

Language Students were asked to record their This ﬁrst activity was ‘detectives’ observations of spoken and written designed to engage language in a variety of known students in ‘noticing’ contexts, e.g. at the park, and in a language in use, i.e. the shop. This was done in collaboration functionality of with parents at home. language. Community Students walked around their local This activity commenced walk community and were asked to notice discussion of relations the language used in different between contexts of contexts, e.g. street signs, café menus. use and language. Their observations were recorded It represents our ﬁrst once back at school. Students were attempt to use register as encouraged to look for patterns in the the orienting concept types of language used in different for the study of contexts. lexicogrammar. 353

Language at Students made lists of the This was another activity school different phrases they would hear designed to increase in different places in the school, children’s awareness of e.g. library, assembly, as well as how language varies who was speaking. The class according to its use in worked together to ﬁnd common different contexts, this patterns in the language used. time within the school. Commands Through drama activities, the students practised giving and responding to commands. By ‘playing’ with the length and experiential content of each command they were encouraged to notice what makes a command function more or less effectively. Students told each other ‘silly’ commands as a basis for comparison. Street signs Using images from the community walk the students focused on local street signs with simple commands, e.g. Stop. They talked about what makes a street sign effective, including aspects of visual semiosis (colour, size of letters, etc.). The students then designed their own street signs that were deliberately less effective. Recipes – fairy In this activity the students made We used the activity as a bread ‘fairy bread’ by following a recipe. basis for discussing However, the recipe had been relations between mixed up and they were required wording and meaning, to rearrange the sentences so that focusing on the the instructions made sense. importance of sequence. Recipes – mini After making mini pizzas the pizzas students worked together to choose the most effective instruction for each step. They were given two options for each instruction and were asked to talk about if, and why, one was more effective than the other. Once they had decided on a complete recipe there was a brief discussion about Events and they noticed an Event often came at the beginning of a sentence. 354 Society in Language, Language in Society

(Continued)

Pedagogic Description Comment activity

Recipes – fruit Students worked in small groups to skewers make an instruction using the word cards they were given. For some groups there were several different ways the words could be ordered. This was talked about as a whole class. The whole class then sequenced the instructions into a full recipe that explained how to make fruit skewers. After making the fruit skewers the students revised their recipe. Together the class identiﬁed the Event words in each instruction. The students then worked individually to complete a worksheet to identify the Event. They needed to apply their knowledge to other forms of commands, not exclusively to recipe instructions.

Notes

1. As a striking example in Australia, see the recent Review of the Australian Curriculum Final Report (Australian Government, Canberra, 2014b). 2. This contextual point is important because it underscores the fact that Hasan’s and her colleagues’ theoretical and empirical work on semantic variation (e.g. Hasan, 2009; Cloran, 1994; Williams, 1999) and genre-based pedagogy both have close links to Bernstein’s analysis of pedagogic discourse. Both systematically examine linguistic bases for covert privileging through texts. 3. Hasan has herself presented an engaging example of what a student educated through reﬂection literacy education might be able to do to critique relations between dominant social practices, language use and language variation. Her focus was the disputes, often violent, around the World Trade Organization’s (WTO) meeting in Seattle in 1999 and the process of resemanticization used to defend the WTO position. Resemanicization is a process of shifting the con- ventionally established meanings of terms, often with dizzying speed, to new meanings to better suit the interest of the dominant social group. Critics of the resemanticized term are made to appear irrational or extremist because they oppose what appear to be obvious initiatives for the common good (Hasan, [2003] 2011]. 4. I am grateful to the Principal, Sally Hogan, and the two kindergarten teachers, Emily Moreton and Phyllis Bird, for this collaboration, which involved many hours of conversation. However, responsibility for this report is entirely mine. Geoff Williams 355

5. This term is widely used in SFL research and pedagogy, following Halliday’s proposal ([1996] 2002, pp. 384–6) to distinguish terminologically between the study of lexicogrammar, grammatics, and the phenomenon of language itself, as in ‘the grammar of Mandarin’. 6. This table was presented on a smartboard but unfortunately no clear screenshot is available. 7. This table is based on a paper co-presented with Emily Moreton at the 40th Annual Congress of the International Systemic Functional Linguistics Association, Sun Yat- sen University, Guangzhou, China, 2014. I thank Emily Moreton for this further phase of our collaboration.

References

Australian Government, Department of Education (2014a) Programme Guidelines Flex- ible Literacy for Remote Primary Schools Overview (Canberra: Australian Government). Australian Government (2014b) Review of the Australian Curriculum Final Report (Canberra: Australian Government). Bernstein, B. (1990) Class, Codes and Control,Vol.IV:The Structuring of Pedagogic Discourse (London: Routledge). Christie, F. (1989) ‘Curriculum genre in early childhood education: a case study in writing development’. Unpublished PhD dissertation, University of Sydney. Cloran, C. (1994) Rhetorical Units and Decontextualisation: an Enquiry into Some Relations of Context, Meaning and Grammar (Monographs in Systemic Functional Linguistics) (Nottingham: Department of English Studies, University of Nottingham). Freire, P. (1985) Pedagogy of the Oppressed (Harmondsworth: Pelican). French, R. (2010) ‘Primary school children learning grammar: rethinking the possibilities’. In T. Locke (ed.) Beyond the Grammar Wars (New York and London: Routledge), pp. 206–29. French, R. (2013) ‘Teaching and learning functional grammar in junior primary classrooms’. Unpublished PhD dissertation, University of New England. Good, R. H. III and Kaminski, R. A. (eds) (2014) Dynamic Indicators of Basic Early Liter- acy Skills (6th edn). Dibels Nonsense Word Fluency Progress Monitoring. https://dibels. uoregon.edu/ (accessed 2 July 2014). Halliday, M. A. K. (1975) Learning how to Mean (London: Edward Arnold). Halliday, M. A. K. (1996 [2002]) ‘On grammar and grammatics’. In J. J. Webster (ed.) The Collected Works of M. A. K. Halliday, Vol. 1: On Grammar (London: Equinox), pp. 384–417. Halliday, M. A. K. and C. M. I. M Matthiessen (2014) Halliday’s Introduction to Functional Grammar (London: Routledge). Hasan, R. (1984) ‘Coherence and cohesive harmony’. In J. Flood (ed.) Understand- ing Reading Comprehension: Cognition, Language and the Structure of Prose (Newark: International Reading Association), pp. 181–219. Hasan, R. (1996) Ways of Saying: Ways of Meaning. Edited by C. Cloran, D. Butt and G. Williams (London: Cassell). Hasan, R. ([1996] 2011) ‘Literacy, everyday talk and society’. In J. J. Webster (ed.) The Collected Works of Ruqaiya Hasan,Vol.3:Language and Education: Learning and Teaching in Society (London: Equinox), pp. 169–206. Hasan, R. ([2003] 2011) ‘Globalization, literacy and ideology’. In J. J. Webster (ed.) The Collected Works of Ruqaiya Hasan,Vol.3:Language and Education: Learning and Teaching in Society (London: Equinox), pp. 207–31. 356 Society in Language, Language in Society

Hasan, R. ([2007] 2011) ‘Literacy pedagogy and social change: directions from Bernstein’s sociology’. In J. J. Webster (ed.) The Collected Works of Ruqaiya Hasan, Vol. 3: Language and Education: Learning and Teaching in Society (London: Equinox), pp. 232–50. Hasan, R. (2009) The Collected Works of Ruqaiya Hasan,Vol.2:Semantic Variation: Meaning in Society and in Sociolinguistics. Edited by J. J. Webster (London: Equinox). Hasan, R. (2010) ‘The meaning of “not” is not in “not” ’. In A. Mahboob and N. K. Knight (eds) Appliable Linguistics (London: Continuum), pp. 267–306. Luke, A. (1996) ‘Genres of power? Literacy education and the production of capital’. In R. Hasan and G. Williams (eds) Literacy in Society (London: Longman), pp. 308–38. Luke, A. (2012) ‘Critical literacy: foundational notes’. Theory into Practice 51 (1): 4–11. Martin, J. R. (1985) Factual Writing: Exploring and Challenging Social Reality (Geelong, Vic.: Deakin University Press). Martin, J. R. (1992) English Text: System and Structure (Amsterdam: John Benjamins). Rothery, J. (1990) ‘Story writing in primary school: assessing narrative type genres’. Unpublished PhD dissertation, University of Sydney. Schleppegrell, M. J. (2013) ‘The role of metalanguage in supporting academic language development’. Language Learning 63 (1): 153–70. Vygotsky, L. S. (1933) Play and its Role in the Mental Development of the Child. Psychol- ogy and Marxism Internet Archive (marxists.org) 2002. http://www.marxists.org/ archive/vygotsky/works/1933/play.htm (accessed 31 January 2013). Vygotsky, L. S. (1978) Mind in Society: the Development of Higher Psychological Processes (Cambridge: Harvard University Press). Vygotsky, L. S. (1981) ‘The genesis of higher mental functions’. In J. V. Wertsch (ed.) The Concept of Activity in Soviet Psychology (New York: M. E. Sharpe), pp. 144–88. Whorf, B. L. (1956) Language, Thought and Reality: Selected Writings of Benjamin Lee Whorf. Edited by J. B. Carrol (Cambridge, Mass.: The MIT Press). Williams, G. (1985) ‘Texts and training: some notes on the metaphor of reading as a military operation’. Signal Approaches to Children’s Books 48: 182–8. Williams, G. (1998) ‘Children entering literate worlds: perspectives from the study of textual practices’. In F. Christie and R. Misson (eds) Literacy and Schooling (London: Routledge), pp. 18–46. Williams, G. (1999) ‘The pedagogic device and the production of pedagogic discourse: a case example in early literacy education’. In F. Christie (ed.) Pedagogy and the Shaping of Consciousness: Linguistic and Social Processes (London: Cassell), pp. 88–122. Williams, G. (2000) ‘Children’s literature, children and uses of language description’. In L. Unsworth (ed.) Researching Language in Schools and Communities (London: Cassell), pp. 111–29. Williams, G. (2004) ‘Ontogenesis and grammatics: functions of metalanguage in pedagogical discourse’. In G. Williams and A. Lukin (eds) The Development of Lan- guage: Functional Perspectives on Species and Individuals (London and New York: Continuum), pp. 241–67. Williams, G. (2005) ‘Grammatics in schools’. In J. J. Webster, C. M. I. M. Matthiessen and R. Hasan (eds) Continuing Discourse on Language (Vol. 1) (London: Equinox), pp. 281–310. 15 Reﬂection Literacy and the Teaching of History Mariana Achugar and Mary Schleppegrell

15.1 Learning from Hasan’s work

How can we educate those who have a different language and experience from ours? How can we communicate ‘across unshared worlds’ and ‘different points of departure’ (Hasan, 2004)? How can we address issues of equity and social justice in the education of language minority learners? Ruqaiya Hasan’s work has provided us with some key ideas that have guided our search for answers to these questions. As applied linguists who work on education in multilingual contexts, we are confronted with variation at the linguistic, cultural and social levels that demands new ways of looking at language in education. Language is a form of activity of people in society, and ways of speaking cannot be separated from social life. This dialectical relation between language and society is conceptualized in systemic functional linguistics as register, the variation of language according to situation. Each situational context is realized in linguistic choices that construe representational, interpersonal and textual meanings, and through linguistic choices the meanings we make differ from one social situation to another. The classroom offers students opportunities for participation in registers they do not encounter outside of school, but Hasan (1996) reminds us that students will be differently positioned to participate in this register variation:

Register variation needs to be seen from two perspectives ...: ﬁrst, the use of language varies depending on variation in social processes; and secondly, not all speakers enjoy the possibility of engaging in all social processes. It follows that pupils will come to the school with differing discursive abilities, with differing experience of participation in registers – simply because of the differences in their social position. (Hasan, 1996, p. 396)

357 358 Society in Language, Language in Society

Even when students are engaged in the same social process in the classroom, their experiences of this process will vary because they have different ideological perspectives and social positioning. Hasan’s work has shown us how the analysis of language can help us understand the unequal distribution of knowledge in schools and how social class affects orientations to situations and ways of meaning. As Hasan states:

what might be taken to be the ‘same’ in terms of its appearance is not necessarily the ‘same’ for everyone in terms of its meaning: what gets in the way of arriving at the same meaning is the difference in the experiences of persons who are socially differentially positioned, because meaning is refracted through the experience of one’s social location. It follows that a given instance of systematic instruction may in fact not be the same for children coming from different social locations. (Hasan, 2004, p. 66)

This means that just being in the classroom does not necessarily lead to the learning outcomes for students that teachers intend and plan for. It is the interpretation of context, not its physical features, that produces shared or conﬂicting understanding or meanings. According to Hasan (2004), one of the most important reasons for unequal distribution of knowledge in schools is the educational systems’ reluctance to acknowledge this difference and open up a space to work with the different voices in the classroom and enable meanings of different kinds to be shared. Recognition of and explicit work with different ways of meaning are required for the goals of education to be achieved. Hasan’s research on the literacy practices of mothers and children in the early years shows how different ways of meaning emerge, and how children who are differently positioned in the social structure are socialized in different ways to recognize and realize (Bernstein, 2000) different kinds of meanings. She also demonstrates how formative modes that shape more interpersonal aspects of social participation are as relevant to the child’s emerging sense of literacy as the informational modes through which knowledge is shared. For example, in Hasan’s (2002) paper on mother–child interaction, she identiﬁes questions framed in the negative (e.g. Why don’t you like that?) as realizing assumptive meaning. This type of construction has an explicit meaning: seeking information (e.g. what is the reason for not liking something) but also an implicational meaning (e.g. you should like it). In asking this type of question the speaker demonstrates assumptions about the addressee’s orientation that reduce the personal distance between them. Hasan’s research establishes that children’s experiences of such implicational meanings vary. Not all children develop the same orientation to interpersonal meanings, as these are construed in different ways by different cultural and social groups. Mariana Achugar and Mary Schleppegrell 359

In educational contexts, the interpersonal has great pedagogic potential and relevance, as the enactment of social relations in the classroom also shapes the habits of mind and participation roles through which knowledge is produced. The relationships instantiated in the classroom set up expectations for the kinds of relationships children are expected to establish with each other and with adults in that context and position them in particular ways as actors within the institutional context of the school. This means that the regulative (interpersonal) aspects of teaching are as pedagogically important as the instructional (ideational). When teachers model ways of encountering and responding to new knowledge that make knowledge open to question and revision, and offer students roles in thinking about how meaning is being made or taking the role of the one who questions the text, students experience the context as one that recognizes their own agency and experience. In this way the interactional roles that the pedagogy enables and the opportunities for leadership in thinking and discussing that the students are encouraged to take up shape in important ways the learning that is accomplished. Designing a learning space where dialogue between different voices is validated enables explorations of language and knowledge that do not just passively inform students, but instead engage them in powerful ways of speaking. Hasan’s work reminds us that to produce social change, education needs to integrate processes of acculturation from what is already known with processes of expanding knowledge in order to recognize and engage learners’ agency in constructing something new. She has made particularly important contributions in this regard in the area of literacy education. For Hasan, literacy is using language to mean, linguistic semiosis. Literacy includes recognition (identifying what is in the text: decoding/noticing), action (using language to achieve certain goals: rules of production), and questioning (asking about who benefits from certain ways of saying/doing, what interests are served); all of these aspects of literacy are important if students are going to be able to produce knowledge, not only to reproduce others’ knowledge. Literacy is thus conceptualized as a social activity by which people participate in the world, reproducing it and transforming it. Hasan’s (1996) work on literacy in society provides us with an important means to respond to this dialectic of reproduction and change through her construct ‘reflection literacy’, contributing a critical perspective to literacy education by introducing a reflexive aspect to the concept of literacy. In the pedagogic context, literacy is conceptualized in various ways: (1) as coding (sound–shape correspondence); (2) as action with a focus on making meaning; and (3) as reflection whose goal is producing new knowledge. Hasan refers to conceptions of literacy as coding as ‘recognition literacy’, focused on structure, sentence level and rules, and suggests that this conception of literacy has the potential to focus only on the expression plane and forget 360 Society in Language, Language in Society meaning, producing a learner who may conform to norms without questioning the social motivation behind them. She asks ‘[w]hat more is needed to make reading and writing semiotic acts and what does it contribute to being educated?’ (Hasan, 1996, p. 392). She posits a second type of literacy, ‘action literacy’, which focuses on academic knowledge as specialized discourses. This literacy extends recognition literacy to doing something with literacy in the service of meaning making. This action-oriented, text-based approach to literacy has the potential to produce a learner who reproduces valued discursive practices. We have seen much growth in understanding of this conception of literacy in recent research on disciplinary literacy and academic language (Anstrom et al., 2010; Moje, 2008; Shanahan and Shanahan, 2008). But the question Hasan (1996) poses to this type of literacy is ‘whether in learning discursive ability through [action literacy] pedagogy, one is also learning the ability to analyse and to challenge the desirability of the prevalent ways of being, doing and saying’(Hasan, 1996, p. 405). If action literacy merely teaches the norms of what is accepted in the educational context, learners may become conformists who respect convention without reflection. Hasan suggests that

then we end up teaching literacy for schooling, creating the literate subject in the image of the existing standards of education. It could be argued then that it is an instrument for the perpetuation of the standards of the existing educational system, and one might go on to argue that if the ideal goal of teaching is to enable pupils to PRODUCE knowledge, not just replicate it, then genre-based literacy falls short of this ideal. (Hasan, 1996, p. 406, emphasis in original)

For knowledge to evolve it has to be subjected to analysis and reflection. To describe literacy that accomplishes that goal Hasan describes a third type of literacy, ‘reflection literacy’, which focuses on literacy as knowledge that is changing and whose goal is the production of knowledge requiring pedagogies that focus on reflection, enquiry and analysis. This type of literacy requires pedagogic practices that question a text:

to ask why the said is being said, what it implies, and on what grounds – call[ing] for a much deeper understanding of language as a resource for meaning. So teachers will need ...to show what alternative ways there are of saying the ‘same thing’. The point is that one can never say exactly the same thing using a different wording; so in fact, the teachers will need to make pupils aware of the sorts of difference in meaning that may arise from putting it one way as opposed to another. [ ...] there are many ways in which a text can be read. This does not mean that meaning is in the reader, as the post-modernist sentiment has it (Rosenau, 1991, pp. 25ff.); rather, it means that our perceptions of meaning are coloured by our Mariana Achugar and Mary Schleppegrell 361

point of view, and our point of view – in as much as it is capable of being studied and taken into account – is related to who we are, socially speaking. To say that a community has many voices is to say that there are experiences of saying and meaning which differ from one social group to the next; this includes the possibility that the way a locution is evaluated in one segment of the community might be critically different from that in another. So it becomes important to ask whose point of view does the writing represent? whose point of view is implied in which reading? It is from this kind of deeper understanding of what ‘the’ text means that we can move to explanation questions. (Hasan, 1996, p. 411, emphasis in original)

To be able to engage in this kind of explicit pedagogic practice, teachers need to conceptualize language as a meaning potential and be familiar with the ways in which their discipline constructs knowledge. This produces a student who has a disposition to question knowledge instead of accepting it on the sole basis of the authority of the author. If we recognize, as Hasan demonstrates, that experience and social position shape consciousness, education needs to focus on designing environments where various coding orientations are acknowledged and included, while expanding the meaning-making potential resources of everyone participating in knowledge construction. This can result in new ways of understanding and making meaning that will push forward our knowledge construction practices and produce new knowledge while tapping into everyone’s potential (see Williams, this volume). In pedagogical terms, Hasan’s reﬂection literacy demands researchers and teachers of literacy to be explicit about how meanings are made and investigate the history and social signiﬁcance of meaning making, challenging naturalized knowledge and discursive practices. Asking questions about what, how and why meanings are made in particular ways by particular social actors in particular socio-historical contexts makes visible the choices made from the system’s potential and whose interests they serve.

15.2 Implications of Hasan’s work for our practice

Hasan’s work challenges us both in the analyses we do and in the pedagogies we develop. We have attempted to consider the features of the particular contexts, to acknowledge and value diversity, to foreground the interpersonal, and to enable explicit attention to language and meaning that involves learners in recognizing and participating in new ways of making meaning that they themselves help shape. In a variety of contexts we have confronted the challenges of developing and promoting reflection literacies. In this chapter we present brief descriptions and examples that illustrate the ways in which we have tried to implement educational interventions 362 Society in Language, Language in Society that use the construct of ‘reflection literacy’. The examples come from work we have done over the past years and have been reported on in previous papers as cited below. Here we take the opportunity to offer new interpretations in connection to Hasan’s ideas. Our goals in the projects described in this chapter have been to offer teachers a view of language together with a metalanguage and analytic tools for talking about text in ways that make explicit the meanings at stake, as well as to design pedagogical activities that engage teachers and students in the exploration of the meanings in text and enable the sharing of alternative meanings. Our focus in this work has been on ‘teacher development’: helping teachers build the deeper understanding of action literacy and disciplinary literacy that they need as a foundation for building reflection literacy, and providing them with tools for asking questions in discussion about the texts they read and engage with. In our work we draw on systemic functional linguistics (SFL) for a theoretically grounded grammar and way of thinking about language that enables exploration of texts and registers through a connection between form and meaning. This is a crucial tool for promoting both action and reflection literacy, as it offers a metalanguage that enables a meaning-based characterization of texts and registers, where the relation between meaning and form is conceptualized as solidary, inseparable (form is meaning and meaning is realized in form). In addition, it enables an exploration of meaning from a paradigmatic orientation that recognizes valuer in terms of potential alternative choices, so that language can be conceived of as a meaning-making resource used to participate in and construct the world, and as a form of social activity that contributes to the production and reproduction of both content/information and ways of being (social relations). As the examples in this chapter illustrate, the SFL conceptualization of language and its metalanguage have enabled us to be explicit in the exploration of meaning in text in ways that set up opportunities for discussion about why an author has made a linguistic choice, and for consideration of its meanings and the alternative meanings that could have been made. This pushes the classroom discourse beyond decoding of local instances to more socio-historical approaches that explore the history of ways of meaning and linguistic choices, and that offers learners a means of bringing their own perspectives into dialogue about text. Learning in multilingual classrooms requires designing educational spaces where different voices are acknowledged and valued. Understanding variation in the ways situations are interpreted, in terms of what is considered relevant, and variation in the potential inferences readers and interlocutors take from instances of language use, helps us design educational interventions that address variation as potential instead of problem. Mariana Achugar and Mary Schleppegrell 363

15.3 Building from Hasan’s work: illustrations from our work

The diversity of language in education involves not only the diversity the students bring, but also the disciplinary diversity that calls for different ways of constructing knowledge in different fields. This means that a key goal of education is to enable students’ expansion of ways of saying and meaning so they get access to knowledge as well as the means to construct new knowledge. Academic language development is not about adopting privileged forms of language; it means mediating learners’ consciousness to enable the possibility that they will see the world in new ways at the same time they benefit from learning about how previous generations have seen and interpreted the world. Academic language development both enables participation in new domains and enables new ways of seeing the world and understanding reality. For example, access to disciplinary ways of seeing allows one to look at a tree and understand its meaning from different perspectives. In a literary sense the tree can be represented as an object of beauty through a poem, while from a scientific perspective the same tree could be represented in terms of the biological processes that it is part of (i.e. pho- tosynthesis). Education in today’s diverse classrooms cannot be successful if it assumes that there is one authoritative voice that can project privileged knowledge that is then taken up by all learners as given. Instead, the interpersonal relationships that need to be established in classroom learning vary according to the field of knowledge being developed and the ways information needs to be organized to establish particular logical relations. Teaching and learning activities need to integrate and take account of the experiences the learners bring and the contexts in which they live and work in order to offer them ways of taking the initiative in their own learning and of contributing their own knowledge and perspectives. In this chapter we focus on our work in the disciplinary context of learning history. History educators discuss the challenges of engaging students in thinking historically (e.g. Bain, 2000). They highlight the role of analysis skills in history pedagogy, urging that students learn the means to identify and evaluate diverse interpretations, recognize human motive in the texts they read, and acknowledge that certainty is elusive (Stearns, 1998; Wineburg, 2001). In learning history, students need to engage in causal reasoning and to deconstruct and recognize the causal links in the texts they read (e.g. Leinhardt, 2000; Voss et al., 1994). They need to learn that every interpretive perspective limits or focuses the questions that can be asked and answered, the evidence used and how it is evaluated, the view of linkage or causation that is offered, and the point of view that is taken. Every author students read, including textbook authors, presents a particular interpretation, and helping students recognize this and explore the perspectives presented in the texts they read is a key goal in teaching history. Understanding the challenges of engaging students in these history discourses has helped us 364 Society in Language, Language in Society focus the professional development we have offered teachers on developing action literacy in history on which reflection literacy can be built.

15.4 The California History Project

In the early 2000s a change in education policy in California brought us an opportunity to apply some of Hasan’s principles in classroom contexts. A voter initiative that greatly decreased opportunities for students learning English as an additional language to have bilingual education meant that many students who previously had not been in mainstream classroom contexts were now in those classrooms. Teachers were struggling to accommodate those students and were asking for new strategies. We were approached by leaders of the California History Project (CHP), a professional development initiative of the University of California that had been offering seminars for history teachers where history professors offered new ways of thinking about historical topics and provided them with additional teaching resources. The project leaders asked us about resources that would be history-speciﬁc to support teachers in their work with these students. We saw this as a social justice issue. Without a focus on language in the history classroom, students learning English were unable to fully participate. Their teachers, however, were not positioned to address the students’ language development, as many of the history teachers who participated in the project did not have backgrounds in language that gave them tools or strategies for talking about language and relating it to the learning of history. Aware of the work that had been done to analyse history discourse by SFL researchers in Australia, we began a collaboration with the project to introduce teachers to new ways of thinking about the language–content connection in history and enable them to engage all of the learners in their classrooms with the challenging content of secondary-school history. We observed secondary-school history teaching that helped us better understand the need for more effective ways of focusing on the interpretation built into history texts by their authors and how to make this interpretation a focus of discussion in the classroom. The teachers we were engaged with were very concerned about developing such understanding based on what they were learning about historiography and the ways historical events and actors were positioned by those writing history. But the tools they had for making interpretation a focus of attention in the classroom were not robust enough to enable a nuanced discussion about how interpretation is infused into texts about history. For example, in an observation of a teacher considered highly effective as a history instructor, the teacher asked students to look through the textbook for speciﬁc words that imply bias or a point of view (the teacher assigned students to different chapters). They were to look for ‘positive/good’ and Mariana Achugar and Mary Schleppegrell 365

‘negative/bad’ words and make a list in their notebooks. The examples provided by the teacher included ‘best’ and ‘cruel’, and the teacher presented this as ‘the textbook is making the decision for you, where it is supposed to be [presenting] just facts’. The students struggled to identify such words and characterize them as good or bad. While the teacher noted the importance of context in identifying words that imply bias, the activity did not support looking at meaning in context. In the end the heuristic the teacher offered the students was written on the board for them to copy:

Question: How do we deal with the bias in books and stuff? Answer: Be aware of it. 1. Omit or get rid of the bias words. 2. Substitute the bias word with another word.

We saw that the focus on bias was at the level of individual words, and did not lead to rich discussion about the meanings in the history textbook. This and other observations of teaching practice helped us continue to shape the professional development (PD) we offered to enable teachers to work with meaning in context and explore author interpretation in more robust ways. We made the goal in our PD with teachers to raise their awareness about how language works in the construction of meanings in history texts, and to offer them a metalanguage for making the meaning-making resources of language a focus of classroom discussion. We developed a focus on exploration of the register features of history texts, introducing the notions of ideational, interpersonal and textual meanings by using three questions to engage teachers in exploration of text at the level of the clause and sentence: ‘What is it about?’, ‘What is the perspective of the author?’, and ‘How is the text organized?’ We introduced SFL metalanguage and constructs to engage teachers in exploring text to answer these questions. For example, we introduced the notion of ‘processes of different types’ to focus teachers on the kinds of meanings presented in different clauses. We identified history passages that had dominant patterns of different clause types and developed activities that showed teachers how these resources were used by authors to present meanings with different purposes in the development of the text; for example, passages with ‘verbal/mental’ processes that construct some kind of historical debate; passages with ‘material’ processes that recount events; and passages with ‘relational’ processes that provide descriptions of settings in time and space to set up the context for events. We introduced the notions of participants in those processes and 366 Society in Language, Language in Society their roles as Actors, Sensers, or Sayers, and the Circumstances of time, place and manner that enhanced the meaning of the process. This focused on the ideational meanings in history discourse to help teachers and learners consider the ways authors present meanings of different kinds as they reconstrue the events of the past. Our initial introduction of these concepts identified for us further issues that were going to emerge as challenges. Many of the teachers we worked with had difficulty identifying the grammatical elements; for example, the verbs that construed the processes in a clause. We realized that the teachers would need knowledge about these linguistic features in order to do a systematic analysis of a text. Figure 15.1 is an example of the kind of scaffold we provided to respond to this issue. By identifying the processes for the teachers, they were supported in analysing the text, identifying the participants and circumstances, and thinking about the meanings being presented by the authors, as well as in learning more about how to identify verbs and processes. Analysis of who is acting here puts the Declaration itself in an agentive role of ‘capturing’ and ‘strongly criticizing’, and then moves King George III into the agentive position as perpetrator of a set of actions that eventually led the colonists to revolt. Discussion in the classroom helps students recognize these history discursive practices, such as positioning documents, acts, laws, etc. as agentive Actors in historical practices. We linked the discussion of process types with a focus on what the different process types enable the writer to do, and offered teachers and students ways of thinking about the choices the author was making in construing history through those process types. As in the Declaration example, they often recognized that in analysing and discussing material processes, the

The Declaration captures the spirit of Thomas Paine’s ideas by strongly criticizing King George III for depriving the colonists of trial by jury, imposing taxes without the colonists’ consent, dissolving colonial charters and legislatures, and committing two dozen other alleged crimes. (Stuckey and Salvucci, 2000, p. 193).

Actor Material process Receiver/Goal captures strongly criticizing depriving imposing dissolving committing

Figure 15.1 Support for deconstructing the Declaration of Independence passage Mariana Achugar and Mary Schleppegrell 367 focus was on who is presented as agentive in history and who is acted upon rather than presented as an Actor. In discussions of texts about slav- ery, for example, teachers recognized that African Americans were seldom represented as agentive (Schleppegrell and Achugar, 2003). In examining verbal and mental processes, teachers and students focused on whose ideas and words the history writer had chosen to present and whose words and ideas were not presented. This provided alternative ways of thinking about bias that helped teachers go beyond a good/bad dichotomization and think more deeply about the choices a historian makes in presenting history. In the PD events we developed, teachers worked together to deconstruct texts and explore their meanings and then to design lessons that engaged their students in the same kind of exploration and deconstruction that they had experienced in their work with us. The pedagogy involved teachers modelling followed by students working and talking together about the language and meaning in the texts. The ensuing conversations about text created opportunities in the PD context for teachers to focus closely on the affordances of the texts they were using in their classrooms, and this often led them to new insights about what students would be able to learn from their reading of those texts. Often they realized that the texts did not offer the learning opportunities they had expected them to offer, and were motivated to bring additional texts into the discussion or to engage students in conversation to expand the historical information provided beyond what was offered in the text. During the subsequent teaching of the units of study the teachers had developed in the PD, their feedback and our observations of their teaching showed that this approach was enabling teachers to engage students in new ways in the classroom. Where previously many of the students who were learning English were observed to mainly work alone, often distracted, we saw that the focus on language and meaning enabled teachers to engage in richer conversations with students about history. Students were gaining new insights and exploring meaning more deeply, and the conversations about text created opportunities for students to ask questions and not only answer them. Teachers were enthusiastic about these outcomes and reported that the work was engendering new kinds of discussions in their classrooms that were enabling them to go much deeper in their discussions about history. The more they engaged in this work, the more they were able to go beyond a focus on action literacy to reﬂection literacy that enabled questioning and discussion about the ideologies informing the author of the history texts. For example, after working with the project for a while, one of the teacher leaders began to inform her teaching with contrasting examples of the ways historians position particular events. To support discussion of the Alien and Sedition Acts, laws that positioned immigrants as dangers to the US in the 368 Society in Language, Language in Society early twentieth century, she highlighted for students this sentence from their textbook:

Claiming that the nation needed protecting from treasonous ideas and actions, the Federalist-controlled Congress passed the Alien and Sedition Acts in the summer of 1798. (Stuckey and Salvucci, 2000, p. 324)

In contrast with this, she also had them consider this sentence from a text written by the Pulitzer-prize-winning historian Alan Taylor:

To suppress domestic opposition, Congress passed a Sedition Act, virtually outlawing public criticism of the federal government. (Taylor, 1996, p. 264)

Comparing the two sentences, students were able to analyse the different ways the two perspectives present interpretations of the same historical events. The teacher used close analysis of the sentences to explore with students questions like ‘which part of the nation was being protected? From what or whom? What is the relationship between treasonous ideas and actions and domestic opposition?’ Students discussed these different ways of presenting the Acts by looking closely at the language; for example, at the difference between the ‘claiming’ and the ‘to suppress’ clauses, and by talking about the different ways these two sentences position them as readers. They explored the meaning of ‘treasonous’, ‘domestic opposition’ and ‘public criticism’ in this context to recognize the attitudes and judgements that the authors had infused. This helped them develop more critical perspectives on the history they were learning while at the same time they were learning how the disciplinary discourse works. The goal is to help students recognize the historical thinking that is represented in all of the language that they encounter in history, and to give teachers tools for unpacking text and generating discussion in the classroom about history. Students who are able to talk about interpretation, to recognize how they are positioned as readers, and to use language in powerful ways themselves are developing practices of reflection literacy that will serve them throughout their education and their lives by helping them understand that they can ‘talk back’ to the texts they read, ask about who is positioned in different ways in the texts, and consider their own stance. Key to the success of this work was the SFL notion that the language represents choices by an author, and that students can reflect on the meanings presented by an author and the significance of the choices that were made, thus seeing the historian as someone who makes choices they can question. (Other publications that report on this work include Achugar and Schleppegrell, 2005; Achugar et al., 2007; Schleppegrell, 2006, Mariana Achugar and Mary Schleppegrell 369

2011; Schleppegrell and Achugar, 2003; Schleppegrell et al., 2004, 2008; Schleppegrell and de Oliveira, 2006.)

15.5 The teaching and learning history in multilingual contexts project

In 2008, Mariana Achugar and Brian Carpenter built on the CHP work to design an intervention in collaboration with teachers to integrate a functional approach to disciplinary literacy into the high-school history curriculum in multilingual classrooms in the south-west of the United States (Achugar and Carpenter, 2012). In this project, they added to the focus on language as meaning-making resource the heuristic practices typical of historians to make visible the linguistic choices, contextual knowledge and metacognitive skills experts draw on to make meaning within the discipline. The work of Leinhardt (e.g. 2000) and Leinhardt and Young (1996) and other educational researchers who have explored history learning provided a larger set of principles to enhance the previous work.1 The basic principle guiding this work was that to learn from historical documents, students needed to think not only about what was said, but also about how it was said. The core of the intervention focused on designing lessons where there was close reading of primary source texts using linguistic tools to get to historically relevant questions to think about the content. This approach was a text-based reflection on the use of meaning-making resources that allowed teachers and students to explore the significance of the choices authors made within the potential the linguistic systems afford. This exploration allowed them to read texts and think about them in more critical ways by focusing on who benefits from the choices authors make. Engaging teachers as collaborators in the design of the intervention positioned them as intellectuals and constructors of knowledge, not only the recipients of others’ knowledge. This collaboration also impacted the researchers, who had to position themselves as learners and less expert than the teachers in relation to the classroom environment and knowledge about the students’ backgrounds. The process entailed a shift in roles throughout the project, including a first phase where the teachers took on the role of learners in PD sessions; a second phase where there was collaboration as peers with the researchers where knowledge was distributed across them and integrated in the design of focal lessons; a third phase where the lessons were implemented and teachers reflected on the work; and a fourth phase where teachers and researchers collaborated in sharing the results with larger audiences. In addition, students’ learning was documented and assessed by the researchers. The design of the project allowed for learning to happen at two levels: in the interactions between teachers–researchers, and in those between teachers–students (see Achugar, 2009; Achugar and Carpenter, 2012; Carpenter et al., 2015). 370 Society in Language, Language in Society

The intervention was conducted in five high-school American history classrooms taught by the same teacher. The students in these classes included students learning English as an additional language as well as mainstream speakers of English who had been placed in this course because of their lower performance in the state-wide history test. The teacher designed three focal disciplinary literacy lessons that were part of the regular curriculum in collaboration with the researchers and modified them during the year while implementation was being documented and assessed by researchers and the teacher based on the students’ responses. More detailed descriptions of the PD (Achugar, 2009), the focal lessons (Carpenter et al., 2014), the students’ learning (Achugar and Carpenter, 2012, 2014), and the teacher’s learning through conversations with the researchers can be found in previous publications (Carpenter et al., 2015). The examples selected below illustrate the teacher’s learning and the students’ learning in understanding and using language as a resource to learn history. These conversations around texts show how teachers and students take different roles and display different understandings of what meaning making in history entails. These examples were selected to show the changes through time in ways of talking and organizing the lesson as evidence of a qualitative difference in their understanding of how language constructs meanings in history and of what the instructional tasks were like (i.e. text analysis focus and demands placed on students). This example is an excerpt of a classroom discussion where there was a shared reading and analysis of the US Declaration of Independence. The teacher had already introduced a transitivity analysis to explore the representation of the events and social actors involved in a different text. Here they do the joint activity of going over the text to find out what events and actors are represented in the Declaration of Independence.

1. T – ‘What do you know of the Declaration of Independence?’ 2. S – ‘It made us who we are.’ (Teacher talks about the context of the Declaration) 3. T – ‘You’ll have to pay close attention to the language.’ 4. T – ‘I’ll give you some easy examples to start us off here’ 5. T – ‘We are going to identify just like the verb, the participant (the Subject) and the message,2 6. T – Puts up slide with paragraph on it ‘We hold these truths to be self-evident ...’ 7. T – ‘What’s our verb here?’ 8. S – ‘hold’ 9. T – ‘What’s the participant?’ 10. S – ‘the government’ 11. S – ‘the people of the US’ 12. S – ‘The people that wrote the declaration’ Mariana Achugar and Mary Schleppegrell 371

13. T – ‘It could just be those who wrote’ 14. T – ‘What does ‘self-evident’ mean?’ 15. T – ‘What other words could you put in here for ‘self-evident’?’ [ ...] there was some discussion of the meaning of ‘self-evident’ and then the teacher displayed another slide with the ‘are created’ example. 16. T – ‘What’s the verb “are created” ’? It’s kinda passive.’ 17. T – ‘Who are they created by?’ 18. S – ‘By their creator’ 19. T – ‘Does it say by whom?’ 20. T – ‘It’s kinda vague’

In this example from the first focal lesson, the teacher is trying to apply the approach through an unpacking of meaning by identifying processes and participants in the text. This identification allows the class to make inferences about who the social actors are when those social actors are not explicitly rendered, as in ‘we’. Such analysis also points to what is not in the text, forcing students to look for more information outside it. The students in this case provide not only information to answer the teacher’s questions with discrete pieces retrieved from the text as in line 8 (‘hold’), but also provide inferences in lines 10–12 about who ‘we’ refers to (e.g. ‘the government’, ‘the people of the US’) that makes their point of departure and thinking visible to the teacher. The teacher takes up their offers and provides an indirect evaluation ‘could just be those who wrote’ to tie back the inferences to the reference in the text. It is of interest here that the teacher wants the students to focus on the actual authors of the text, rather than ‘the government’ (line 10) or ‘the people of the US’ (line 11). Later on the conversation expands to discussions about who was involved in the process of producing the text and whose interests were served by this document. While we see this as a good beginning, we also recognize that at this point, the conversation about the text is closely focused on what is represented in the text. While there is some exploration of other choices the author could have made in the context of discussing the meaning of ‘self-evident’ (lines 14–15), in other instances the teacher did not take up potential opportunities for discussing alternative options an author might have taken up and how those would have affected the meaning being made. For example, the focus on how events are represented does not highlight the notion that language choices offer ways of eliding agency (see lines 16–20, where the teacher notes that the verb group ‘are created’ is ‘kinda passive’ and ‘kinda vague’, without exploring the motivation for that choice). In a second example, from the middle of the semester, the teacher focuses attention on interpersonal meanings by looking at finites in the text (tense and modality) to explore how certain choices construct orientations to the information and evoke reactions from the audience. The activity 372 Society in Language, Language in Society involved discussing sentences from different texts that had been recast by the teacher/researchers to modify the meaning in subtle ways. The students’ task was to identify the sentence from the historical document based on its linguistic features, and say why they thought that was the actual sentence. The class is discussing a sentence from The Jungle, by Upton Sinclair, in a lesson where the goal was to investigate the role of Muckrakers in influencing public policy. The text analysis explored interpersonal meanings to help students explain how the author’s choice of wording may have impacted the audience and helped produce a legislative change. The identification of linguistic features allowed the students to predict the context of situation and potential effects on the audience. On the screen the teacher had presented three texts (colour coded by teacher):

1. Let a man so much as scrape his finger pushing a truck in the pickle rooms, and he might have a sore that would put him out of the world; all the joints in his fingers might be eaten by the acid, one by one. 2. I saw a man that scraped his finger in the pickle rooms while pushing a truck; and the resulting sore nearly killed him, as the acid had a terrible effect on his hand. 3. When a man scrapes his finger pushing a truck in the pickle rooms, he has a sore that will most certainly kill him; as the joints in the finger will be eaten by acid, one by one.

This activity was completed as a whole group and the students had to justify their answers using evidence from the texts. The texts were highlighted with different colours so that students could visualize the changes. The teacher began by saying ‘We are going to look at these sentences and decide which one Upton Sinclair might have written.’

1. T – ‘What difference do you see between the sections?’ 2. S – ‘The second one is past tense and it actually happened. The ﬁrst one is hypothetical.’ 3. T – ‘Why do you say the second one is past tense?’ 4.S–‘sawaman’ 5. S – ‘number 3 has “has” not “might have” ’ 6. S – ‘the third one sounds like it always happens’ 7. T – ‘why does it sound like this?’ 8. S – ‘will kill him’ 9. T – ‘What does the word “let” do in the ﬁrst sentence?’ 10. S – ‘talks about potential’ 11. T – ‘what is a synonym for “let”?’ 12. S – ‘allow’ 13. T – ‘and who is doing the letting?’ Mariana Achugar and Mary Schleppegrell 373

14. S – ‘the plant’ 15. T – ‘why does he use “would put a man out of the world”?’ 16. S – ‘because it is more dramatic’ 17. T – ‘so which one is it?’ [that Sinclair wrote?] 18. T – ‘it’s the ﬁrst one.’ 19. T – ‘What do you think his motive is in this ﬁrst one?’ 20. S – ‘to make people aware’ T [talks about consumer’s rights and workers’ rights, and relates that this book ‘shocks Americans’]

Here we can see how the teacher has expanded his understanding of disciplinary literacy to include a more focused exploration of how linguistic choices have historical significance through manipulating texts and building in an explicit reflection about the effects of these choices. The students have also been socialized into using some grammatical metalanguage and have internalized the idea of meaning making as choices from a potential. In addition, the explicit discussion of the effects different choices produce in terms of reception foreground why there can be different readings of the same text. In this case, the teacher and students explored the interpretations of Sinclair’s text that had historical impact (i.e. food safety regulations were created), and contrasted them with other possible interpretations; for example, regarding workers’ safety, that were not made at the time (i.e. no labour safety regulations were produced). This project found positive results in students’ reading comprehension of texts and understanding of history as well as growth in their academic language use (Achugar and Carpenter, 2012, 2014). In addition, the teacher’s involvement in the project enabled him to position himself as an intellectual (Giroux, 1988) and engage in action research by constructing knowledge in the form of writing papers, sharing his experience with colleagues and taking the role of expert on his practice (through a prestigious fellowship for teachers). His growth in understanding of language was also evident in his practice as seen through the changes in the lessons he designed and his reflections on the meaning of language analysis in history (Carpenter et al., 2015).

15.5.1 Summary In this work with teachers, we focused especially on what teachers took up from the PD in which we engaged them in analysing texts, using functional grammar to explore the register features and disciplinary meanings in history. We recognize that the examples we have provided here do not yet rise to the level of the powerful reﬂection literacy that Hasan has inspired us to work toward; however, we have seen that the development of action literacy in the form of disciplinary discourses is a required foundation to being able to interrogate texts and explore their embedded ideologies and the ways 374 Society in Language, Language in Society they position readers. We need further research that builds on this work and extends it through longer-term studies that analyse the ways students take up and use reﬂection literacy and how it affects their learning success. There needs to be documentation of growth in linguistic knowledge and registerial variation across disciplines for students from various linguistic backgrounds. In particular, we need more research that shows how teachers can respond to the needs of students in ways that create classroom contexts where it is a normal practice to build on students’ contributions and make them a focus of discussion toward developing disciplinary literacy.

15.6 Conclusion

In broad terms, the goal of education is to pass on to our children what our societies consider important from our heritage and history, as well as to enable them to continue to develop that knowledge to respond to the problems of today’s world. In order to accomplish that goal, we need to develop in students and teachers the habits of mind or dispositions toward learning that enable them to construct, share and question knowledge. Knowledge and language are intrinsically connected and develop together over time; today’s knowledge and language are not the same as in the past. As learning is accomplished primarily through language, classroom activities need to position teachers and students to use language in ways that connect to what they bring to the context but also enable them to take up new knowledge and transform it to shape a new future. This requires dialogic engagement as a learning process. While the notion that teaching requires dialogic engagement is not new, Hasan’s focus on the social positioning of learners and the need for teachers to develop a solid ‘sense of the addressee’s social identity’ (Hasan, 2011, p. 41) adds an important perspective when we are working in multilingual and multicultural classrooms. Ways of thinking and being are naturalized for individuals as they engage with and take up the semiotic potential of the community in learning the mother tongue, where language plays a major role in the construal of a reality. In education, however, students can experience a semantic disjuncture as the naturalized frameworks operating in the classroom may not be readily recognized or may conflict with frameworks the child experiences in other contexts. This is not just because they are encountering ‘academic language’ or register differences; it comes out of the fact that classroom language may not typify experience in ways that relate to the students’ reality. For example, what a teacher expects students to take from a reading of a text or how a teacher anticipates students’ responses to that text may not comport with the actual readings or responses of students in their classroom (Hasan, 2004). By making explicit and visible the ways of interpreting that are relevant and meaningful in particular learning contexts, teachers enable recognition Mariana Achugar and Mary Schleppegrell 375 of new ways of understanding. While it may not be possible for a teacher to be familiar with the socialization that every learner has experienced, teachers can make it possible for all learners to bring their voices, questions and perspectives into the classroom dialogue. By enabling students’ sharing of the interpretations they bring to the learning context, teachers facilitate each student’s participation and contributions, and at the same time respect differences. In such dialogue, alternative ways of thinking can be acknowledged. Hasan and others have demonstrated how reasoning about linguistic choices can assist teachers in bringing to the surface important differences in the ways students who are differently positioned understand and engage with the texts they read. However, reflection literacy does not end there. For Hasan, the goal of education is ‘to produce such habits of mind in human beings which enable them to appreciate the nature of a problem, to examine the merits and demerits of proposed solutions, so as to act with discrimination not simply from the point of view of what’s in it for me but from a consideration of the quality of life in the community’ (Hasan, 2011, p. 22, emphasis in original). Hasan’s notion of reflection literacy – the idea that students need to learn to talk about language and its meaning potential as a means of interrogating the texts they read and develop the resources to challenge and construct knowledge, offers opportunities for developing in students new ways of approaching social problems and considering their solutions. Reflection literacy entails making explicit how knowledge is presented and constructed, and identifying cues that are present in language that help us recognize and make sense of what is happening. That requires discussion of the social motivation for and consequences of texts in their social contexts, so that learners develop sensitivity to and an understanding of the social practices of the larger communities in which they participate. And it also means recognizing learners as agents who can construct and not just acquire knowledge. Establishing dialogue around different ways of making meaning, and recognizing and questioning the value of different manners of making meaning, enables students in multilingual and multicultural contexts to interact and relate to each other and to the knowledge being developed in ways that support their participation and engagement in manners of their own choosing. This can support the participation of all students and nurture in them a sense that they can contribute to and shape new knowledge as well. These are very difficult goals to meet, but important ones to work toward in educational research that attempts to go beyond describing, and which aims to effect change in educational contexts. We recognize that our attempts to design interventions with the goal of developing reflection literacy need further development. In particular we have found that systematic engagement with teachers, and sensitivity to local needs and constraints, are important factors in producing interesting results. Further work needs also to continue developing ways to make 376 Society in Language, Language in Society

SFL concepts and metalanguage more accessible and usable for teachers with varied backgrounds. Another aspect for continued development is the exploration of ways to integrate students’ everyday knowledge into the process to bridge the differences between different types of knowledge that are typically presented as completely separate. As Hasan (2004) has shown, pedagogic potential and education into scholarly thinking occur from the earliest years of life and in seemingly mundane situations. We need to identify pedagogical practices that recognize this knowledge base and tap into it in order to scaffold students’ acculturation into disciplinary ways of making meaning that also enable them to reﬂect on and participate in those discourses.

Notes

1. This project was part of a collaboration with the Disciplinary Literacy History Group of the Institute for Learning of the University of Pittsburgh which offered PD to school districts nationwide. Members of that group included Gaea Leinhardt, Anita Ravi, Joan Mohr and Katherine Stainton. This project was funded by the Spencer Foundation. 2. This metalanguage was an adaptation by the teacher of the metalanguage presented by the researcher (see Achugar, 2009, for more on the PD). The teacher asked students to identify participants (deﬁned as the subject), verbs (processes) and message (deﬁned as the rest of the clause that represented new information).

References

Achugar, M. (2009) ‘Designing environments for teaching and learning history in multilingual contexts’ (Special Issue on Education and Immigration edited by L. P. Stevens). Critical Inquiry in Language Studies 6 (1–2): 39–62. Achugar, M. and Carpenter, B. (2012) ‘Developing disciplinary literacy in a multilingual history classroom’. Linguistics and Education 23 (3): 262–76. Achugar, M. and Carpenter, B. (2014) ‘Tracking movement toward academic language in multilingual classrooms’. Journal of English for Academic Purposes 14: 60–71. Achugar, M. and Schleppegrell, M. J. (2005) ‘Beyond connectors: the construction of cause in history textbooks’. Linguistics and Education 16 (3): 298–318. Achugar, M., Schleppegrell, M. J. and Oteíza, T. (2007) ‘Engaging teachers in language analysis: a functional linguistics approach to reﬂective literacy’. English Teaching Practice and Critique 6(2):8–24. Anstrom, K., DiCerbo, P., Butler, F., Katz, A., Millet, J. and Rivera, C. (2010) A Review of the Literature on Academic English: Implications for K-12 English Language Learners (Arlington, Va: The George Washington University Center for Equity and Excellence in Education). Bain, R. B. (2000) ‘Into the breach: using research and theory to shape history instruction’. In P. N. Stearns, P. Seixas and S. Wineburg (eds) Knowing, Teaching, and Learning History: National and International Perspectives (New York: New York University Press), pp. 331–53. Bernstein, B. (2000) Pedagogy, Symbolic Control and Identity: Theory, Research, Critique (rev. edn) (London: Taylor & Francis). Mariana Achugar and Mary Schleppegrell 377

Carpenter, B., Earhart, M. and Achugar, M. (2014) ‘Teaching with documents in the multilingual classroom’. The History Teacher 48 (1): 91–103. Carpenter, B., Achugar, M., Walter, D. and Earhart, M. (2015) ‘Developing teachers’ critical language awareness: a case study of guided participation’. Linguistics & Education. Giroux, H. (1988) Teachers as Intellectuals: Toward a Critical Pedagogy of Learning (Westport: Bergin and Garvey). Hasan, R. (1996) ‘Literacy, everyday talk and society’. In R. Hasan and G. Williams (eds) Literacy in Society (Harlow, Essex: Addison Wesley Longman), pp. 377–424. Hasan, R. (2002) ‘Ways of meaning, ways of learning: code as an explanatory concept’. British Journal of Sociology of Education 23 (4): 537–48. Hasan, R. (2004) ‘Reading picture reading: a study in ideology and inference’. In J. A. Foley (ed.) Language, Education and Discourse: Functional Approaches (London: Continuum), pp. 43–75. Hasan, R. (2011) ‘On the process of teaching: a perspective from functional grammar’. In J. J. Webster (ed.) The Collected Works of Ruqaiya Hasan, Vol. 3: Language and Education: Learning and Teaching in Society (London: Equinox), pp. 7–47. Leinhardt, G. (2000) ‘Lessons on teaching and learning in history from Paul’s pen’. In P. Seixas, P. Stearns and S. Wineburg (eds) Teaching, Learning, and Knowing History (New York: New York University Press), pp. 223–45. Leinhardt, G. and Young, K. M. (1996) ‘Two texts, three readers: distance and expertise in reading history’. Cognition and Instruction 14 (4): 441–86. Moje, E. B. (2008) ‘Foregrounding the disciplines in secondary literacy teaching and learning: a call for change’. Journal of Adolescent and Adult Literacy 52 (2): 96–107. Rosenau, P. M. (1991) Post-Modernism and the Social Sciences: Insights, Inroads and Intrusions (Princeton, NJ: Princeton University Press). Schleppegrell, M. J. (2006) ‘The linguistic features of advanced language use: the grammar of exposition’. In H. Byrnes (ed.) Advanced Language Learning: the Contribution of Halliday and Vygotsky (London: Continuum), pp. 134–46. Schleppegrell, M. J. (2011) ‘Supporting disciplinary learning through language analysis: developing historical literacy’. In F. Christie and K. Maton (eds) Disciplinarity: Functional Linguistic and Sociological Perspectives (London: Continuum), pp. 197–216. Schleppegrell, M. J. and Achugar, M. (2003) ‘Learning language and learning history: a functional linguistics approach’. TESOL Journal 12 (2): 21–7 (AAAL 2003). Schleppegrell, M. J., Achugar, M. and Oteíza, T. (2004) ‘The grammar of history: enhancing content-based instruction through a functional focus on language’. TESOL Quarterly 38 (1): 67–93. Schleppegrell, M. J. and de Oliveira, L. C. (2006) ‘An integrated language and content approach for history teachers’. Journal of English for Academic Purposes 5 (4): 254–68. Schleppegrell, M. J., Greer, S. and Taylor, S. (2008) ‘Literacy in history: language and meaning’. Australian Journal of Language and Literacy 31(2): 174–87. Shanahan, T. and Shanahan, C. (2008) ‘Teaching disciplinary literacy to adolescents: rethinking content-area literacy’. Harvard Educational Review 78 (1): 40–59. Stearns, P. N. (1998) Why Study History? Website of the American Historical Associa- tion: http://www.historians.org/pubs/free/WhyStudyHistory.htm Stuckey, S. and Salvucci, L. K. (2000) Call to Freedom (Austin, Tex.: Holt, Rinehart & Winston). Taylor, A. (1996) William Cooper’s Town: Power and Persuasion on the Frontier of the Early American Republic (New York: Vintage Books). 378 Society in Language, Language in Society

Voss, J. F., Carretero, M., Kennet, J. and Silﬁes, L. N. (1994) ‘The collapse of the Soviet Union: a case study in causal reasoning’. In M. Carretero and J. F. Voss (eds) Cognitive and Instructional Processes in History and the Social Sciences (Mahwah, NJ: Erlbaum), pp. 403–29. Wineburg, S. (2001) Historical Thinking and Other Unnatural Acts (Philadelphia: Temple University Press). Part VII In Her Own Words 16 In Her Own Words: an Interview with Ruqaiya Hasan David G. Butt and Jennifer Yameng Liang

R = Ruqaiya Hasan J = Jennifer Yameng Liang D = David Butt W = Wendy Bowcher

J: Working in the profession of linguistics, we are often asked questions like ‘How can linguistics make a difference to social institutions and the culture?’ And when applying for research grants, we are asked to write the social implications of the project. What are your views on the usefulness of teaching linguistics and how it beneﬁts society?

R: Well, Michael [Halliday] has pointed out, on many occasions, why linguistics would be a good discipline to be introduced and encouraged. My own view is that no matter what you are doing in life, any kind of planning or negotiation largely depends on how you use language. If you don’t understand how language works, you are quite likely to miss out on some of the elements of meaning that somebody is making or as is often the case with second language speakers, you might negotiate in a way which is perhaps not the best for the outcome. That is the practical aspect. The other aspect is (and I think it is just a continuation of that point, in a way) that you do need to understand people. I think that linguistics is important from that point of view, as well. I don’t think there is any other science or discipline that gives you as much information about the nature of language – certainly not as accurately as linguistics is capable of doing. Of course you have to choose your linguistics. Then there are professions that depend on understanding how language works – teaching is one example which everyone is aware of, but I think many professions depend on how well you understand about human interaction.

381 382 Society in Language, Language in Society

D: What did linguistics ‘look like’ to you as you entered Edinburgh Univer- sity, coming from Pakistan?

R: When I came to Edinburgh, I was so ignorant (yeah, yeah! ignorance is a matter of degree, you know! And it’s always there, about one thing or another). I was terrified and thought linguistics was largely grammar and of course I didn’t have a single term of grammar right. The ‘finite’ and ‘infinitive’ really got me in a flap! For me, coming to linguistics was, in my humility (laughs), really a test of linguistics. Because I had brought my problems – two of them – with me: I wanted to see if the linguistics that was being done there was going to be able to do anything for solving either of those problems. It was interesting, in a way, because the only linguist that I had read before I came to Edinburgh was Bloomfield. Then John Mountford, who became a good friend, bought me Firth’s The Tongues of Men (Firth, 1937), from a second-hand book store – for six pence, or something like that. It was a revelation to me, because for the first time, I felt that this linguistics actually could tell me something about meaning. I came to Edinburgh by my good luck and as a result of good consultation and advice from the British Council, so that I came to a linguistics that was unique in the sense of being more focused on linguistics making you understand about meaning rather than just doing something that was shuffling grammatical patterns.

D: Coming to Edinburgh from India and Pakistan and meeting a different kind of upper middle class must have given you a strong sense of what you wrote about as ‘semiotic distance’ and ‘semantic distance’.

R: Well really, that train of thought was more the consequence of having been teaching English as a second language; in Pakistan I was teaching classes with two sets of pupils. Some were actually brought up in an English medium school so their spoken English was very fluent, whereas there was another group that came up from the Urdu medium, so they kind of struggled without getting very far – because suddenly they were faced with doing everything in English, without any practice. It must have been really soul- destroying for them because, in the end, the issue of language goes much deeper than how to ‘say something’. What it was doing to this second group of pupils, those who were struggling and putting in a terrific amount of time in their work in order to get their meaning across, was to destroy their self-confidence. The other lot that were quite fluent in English just didn’t care much. So when you read the essays, you could either see a lot of thought behind the struggle for words from the Urdu medium pupils, or there would be essays where there might be hardly anything behind the fluency in English – all nicely strung together but little thinking. It was really an interesting situation, but also disturbing. David G. Butt and Jennifer Yameng Liang 383

And, two things I learned as soon as I began teaching English as a foreign/ second language to these students: one was that none of the grammars that I read and tried to base my lessons on was trustworthy, because I will tell my students, for example, a very simple thing: ‘Noun is a word that names something, etc.’. And then within a couple of minutes, I will find out that there is the matter of this ‘mountain spring’: what was mountain doing there, was it still a name, a noun? And how many nouns can you put one after the other and in what order? You try to get that kind of information from the early grammar books like, say, Nesfield (e.g. 1908/2007), [but] you wouldn’t get very far. So that was one thing I learnt fast: clearly I had to find someone who really and truly knew about English language and could tell me how to teach it. The other issue that became very clear to me was that being able to use language in a way that satisfies your purpose, allows you to achieve what you had set out to do, and becomes a central issue. The struggle and the failure one encounters have a serious effect on some pupils: that experience did something to the students’ self-confidence. It profoundly affects your students’ ability to raise their head with confidence. In the end, the pupils in the classroom whom you wanted to hear, because they had something to say, kept quiet because they just didn’t know if they would be able to say what they wanted to say. And there were the others that you didn’t really particularly want to hear: they were fluent enough to say nonsense, they hardly gave any thought to anything.

D: One of your papers, perhaps one of my favourites, ‘Ways of saying: ways of meaning’ (Hasan, 1984), sets out so many ways in which two systems can be tracked as different. The levels of implicitness are brought out with so much structure in that article.

R: Yeah, I still quite like parts of that paper. As a matter of fact, I wasn’t focused on class particularly, or ethnicities, or concepts like that. These issues came in because there were questions that naturally invoked reference to them – I was not talking about any dialect of Urdu: the main issue I had in mind in doing that paper was something like the ‘cut of language’ viewed in a rather different perspective than that which Michael has recently reminded us of (Halliday, 2014) – I didn’t know anything at that time – but I was beginning from what the characteristic nature of a language does, to ways of saying in a community: what can you say if the language is like this ...oh you might say, it was like exploring Whorf’s ‘fashions of speaking’. So, comparing the cohesive devices in English and Urdu, I said to myself the extent to which Urdu can be implicit is impossible for English. I like to put it that way rather than the extent to which English is explicit, because English needs to do certain things which add to textual continuity in a way that reduces the dependence on the knowledge of the text’s immediate context. 384 Society in Language, Language in Society

Now with hindsight, one can see how many things are involved in ways of speaking. Malinowski had really made a distinction between kinds of registers – though he didn’t use that term. He talked about two peaks of the pragmatic power of words. His example for one peak was ‘words as action’ such as issuing directives to get something done materially (funnily, one of his examples was the war cries that lead the attack!). The other peak of a language’s pragmatic power he called creative or magical. His example here was the language of magic, prayer and such like. This created something that is entirely an artefact of language; if the language was not used in that particular way, then that kind of action would not occur – you know, something like the distinction between ‘ancillary’ and ‘constitutive’. The ancillary assists the material activity but much of that activity can transpire without words and often does. The constitutive use of language is actually the creator of the activity, which has only this way of being performed. The roots of this distinction involve both the different sorts of activity as well as ways of saying; what kind of talk goes with what kind of social practice, what social practice with what kind of cultural institution. Both those things – the ways of doing and of saying, that is to say, the social practices, and registers – are involved. The ways of meaning in English contribute to explicitness to an extent that it becomes easy for you to take a message and as it were, create the whole context on the basis of the language of that message: you recon- stitute the context in interpreting the message; you don’t need to depend as much on the immediate context as you would do with Urdu. In Urdu, I could for example come in from outside and turn to someone in the house and ask literally ‘is at home?’ and that person would not ask ‘who are you talking of?’; they would know exactly who I am referring to. Now those one- word clauses – they are not there because Urdu-speaking people are reluctant to talk – on the contrary, I mean I am a living example, so you can see! No, no, it’s only because supporting this way of speaking is ‘how we normally be, how we habitually do, and how we mean’: that’s our ‘fashion of speaking’. So naturally, language is implicated in this cultural game; it has to happily tolerate this kind of ‘laconic brevity’, which is not viewed by the speakers as laconic or brief style; it is just the habitual way of saying. Thinking on this way of speaking led me to ask: what’s going on in grammar then; how could one explain such a style and what does it say about Urdu grammar? Where I am a bit critical now of that paper is the kind of ‘causality’ that I had implied. Today I wouldn’t do that because it’s like asking which came ﬁrst: the chicken or the egg? The fact is, it now seems to me, that the claim that we talk like that because the grammar is like that is true just as much as it’s also true that the grammar is like that because that’s how we talk. May be, not just the framework of language, but also the sorts of patterns viewed as rational in a language are as they are because of the way we put language to use. (Remember that paper much later, 1992, I think, called ‘Rationality in everyday talk’.) In the end, you’ve got to think of communal conventions, David G. Butt and Jennifer Yameng Liang 385 and in the same breath ask: How did the communal conventions manifest themselves? How did they get known? How does a certain form of behaviour become conventional?

D: When did that turnaround come from your original view of grammar? Was that after you had arrived in Edinburgh?

R: That diploma course that I did in Edinburgh, which was really wonderful, made me see what one could do with grammar that would bring you closer to meaning. The penny kind of dropped at the end of the ﬁrst six months or so. By then my views about grammar had changed enough so that the research for my diploma – I am referring to the analysis of Neces- sity’s Child – it was purely a grammatical analysis. There was a lot in Firth about meaning and, of course, how to study it, but it was not operationalized. So it was not very easy to see how you could begin working on the semantic motifs in a short story. But the lectures on grammar began to make sense in time. It was also around that time that that book Content Analysis was published (Krippendorff, 1980). And I could see that it was not really an analysis any more than much of literary critique would be; that was an important moment in my life to realize that a grammar-based analysis of meaning is qualitatively different from one that is based on the understanding of meaning simply because you speak the language. In terms of Bernstein, the difference is that between the ‘ordinary’ – based on understanding of meaning – and ‘exotic’ based on an understanding of how meaning is construed.

J: How do you see the centrality of meaning in linguistic endeavours, and how is meaning tied up with the concept of the ‘social’ in your work?

R: The primary thing is somehow to be able to get ‘inside’ the language enough to get the connections and the relations of signs, which is the secret of language. If you take the functionally oriented studies of child language learning, something is quite noticeable: the child is really focused on something that is meaningful to him, i.e. there is something significant to the child that he wishes to achieve. All that he is able to do by way of communicating with another – and another’s intervention is typically mandatory – is to use his body as the means of expression: the limbs moving, the eyes in contact, gestures, facial expression, force of vocal postures, and so on. In this early stage, how to manage a successful exchange of meanings? The beginning of learning how to mean in Halliday’s words (Halliday, 1975) (later replicated by both Clare Painter and Jane Torr) depends a good deal on those who have been able to relate the child’s body movement and certain types of elements in the material context to each other: parts of the context foregrounded by the child’s embodied behaviour patterns become 386 Society in Language, Language in Society central – in a sense, some elements of material context become the mediator of the child’s meaning. As we get into the mother tongue or perhaps even a little before that, we begin to, as it were, free the communication from this kind of dependence on material context: we somehow begin to learn the relations of this linguistic sign with that. Just to give an example that’s hopelessly simplified: the relations of thirst, drink, water, cup develop rather than having to foreground the sensation of being thirsty, its connection to that object ‘cup’ by the straining of the body towards it so as to mediate the material action of drinking that is associated with that object: the child is beginning to learn how to mean linguistically with his mother tongue rather than by jumping up and down, pointing to objects that relate to the satisfaction of his desires. Whereas previously events and objects in the context mediated the meanings to the caregiver, now it is the indicative vocal acts that do the mediation. In other words, his grammar is developing – Saussure’s ‘associative bonds’, that lead the analyst to the ‘morphology of meaning’, which is what we in SFL refer to as the paradigmatic relations of sign – elements of these patterns begin to be internalized: interactants begin to see languaging rather than physical moves as the mediator of meaning: in time the physical/material phenomena will count for less than this abstract means of meaning. The child is beginning to be a ‘semantic person’. It is really an infinitely more developed sense of the relations of linguistic signs that forms the object of enquiry for semantics: the content of semantics is the meaning that can be exchanged by exploiting the language-internal, i.e. semiotic devices – well I prefer Saussure’s term ‘the semiological device’ by which the relations of signs are able to create linguistic meaning. To my mind semantics is the study of linguistic meanings. I believe that without ignoring other systems of meaning, we need to be clear that linguistically made meanings represent a sort of meaning making that is specific to language alone; and for the understanding of semantics to develop in this sense, it is necessary to set aside a term that refers unambiguously only to that domain: thinking of the history of language study ‘semantics’ seems to be the best for that purpose. So I would say the semiological system of images/figures or of gestures has meaning: but that meaning does not belong to semantics: let’s find another name for those meanings. Let’s find out the structure of those semiological systems; let’s create their language of description. Mixing up everything is misleading. There is one paradox, that has apparently not received much attention: what Saussure called the ‘associative bonds’ – the in absentia relations of signs, which to him were the heart of langue, develop and become know- able only through parole: that is where they are experienced – I mean, in syntagms – instances of those relations are witnessed as the in praesentia relations of signs in the community’s parole (these are Firth’s ‘colligation’; underlying the possibilities of his ‘collocations’ are the associative bonds, I David G. Butt and Jennifer Yameng Liang 387 believe). So, ironically, as I have pointed out (Hasan, 2013, 2014), Saussure’s langue cannot be studied except by relying on the working of parole. And all parole, that Peter Doughty used to call ‘languaging’, happens in some context, in performing one social practice or another: the context becomes the measure of the relevance of discourse. Relevance is ‘social’ and also ‘linguistic’: if your talk appears irrelevant to the occasion of talk, well, people begin to wonder about your sanity – yes, a large part of rationality lies in the obser- vance of social conventions and that’s a sobering thought. Anyway, what this means is that the working of context is as relevant to understanding semantics as are the relations at the stratum of lexicogrammar. Almost from the beginning of thinking about semantics, I have found that viewing the realizational relations of these three strata as bidirectional solves some serious problems in linguistics. In fact, I see the three strata of context, meaning and wording as engaged in a dialectic of activation and construal. So, the process of meaning exchange – both the presentation of meaning and its interpretation – can be described in general terms by reference to that dialectic. For the discourse to be relevant, the perception of context must activate the speaker’s meanings; the semantic choices are conceptualized as activating the lexicogrammatical choices, which are expressed in patterns of sound primarily and of writing secondarily (in terms of the history of language evolution). I am using the word ‘activate’ approximately the same way as in neurocognitive linguistics (e.g. Lamb, 1999); by contrast, interpretation begins from access to the signals at the plane of expression: the patterns of this plane signal the boundaries of ‘wording units’; the syntagmatic bonds of these units – their in praesentia relations – construe the sign’s identity at the lexicogrammatical stratum, which in turn construes the sign’s semantic values, and the semantic values construe the relevant context. To understand a text is to know what the relevant context of that text is which is being heard or read by the receiver. I am aware that there have been some misgivings on this formulation; in any case it is highly condensed and makes the whole process appear stochastic, but really it is no more stochastic than the postulate of strata of language related by realization: clearly we adults don’t think walk and then the ed bit when we say walked! Nor is it the case that in conceptualizing the exchange of meaning in this way, I want to imply that you move word by word, sign by sign or even sentence by sentence: I know that as soon as there is a recognized meaning, the entire set of strata have already played their part: there is no such thing first ‘intending’ the meaning and then expressing it. What I mean to do is to establish a schematic construct, a theoretical fiction of the same kind as the stratal structure of language. As you begin to interact, the texture and structure of your developing meaning–wording begin to delineate the points of relevance between the context and your talk: as a listener, you can begin to see what the speaker is planning to achieve, and how that process is progressing. As a matter of fact my view of the relations 388 Society in Language, Language in Society of context, meaning and grammar and phonology is pretty traditional: I simply pursue the spirit of Firth’s approach as I do that of Saussure’s, ignoring on serious reflection simply those parts of the writing which appear contradictory to their professed goals. The full justification ought to be spelt out in both cases but, you know the saying, ‘life is short, but art is long’: well this applies to science as well; ‘life is short, but science is long’. This is why vertical disciplines make more mileage than our egocentric horizontal ones: they build on the already built, and if they are lucky, in this way they defeat the impermanence of individual lives. Take just the scope of the word ‘context’ alone: it is really amazing. How do speakers recognize those aspects of the occasion of talk that are likely to influence what you are likely to say? It is this ‘same’ context that plays one role in the earlier stage of a child’s life: it’s a resource that the child can use as the mediator of his meaning when juxtaposed with his bodily gestures in material objects: in cooperation with its meaning group the child actively creates sensation, event and object like a unique sign language made up of some material phenomena, some body movements in relation to those material phenomena. As for the adult, the minutiae that somehow correlate with the character of our meanings is a marvel! This discussion we are having just now relates to a discipline; but you wouldn’t describe it as ‘technical’ – its style is rather informal for that. We can see how the features in contextual configuration are working: there is a reflection of our relationship; but also note I couldn’t have used this way of speaking with just anyone that I happen to be close to – technical in style or not, it is still in a disciplinary field; it calls for a particular kind of expertise and an interest in an ‘esoteric reality’; the addressee has to be someone willing and able to participate in this kind of activity and with whom the social distance is fairly close, and the interaction is spoken. Well there you go!

D: Can I ask about the role of variation in your thinking? In the 1960s and 1970s and maybe even up to the end of the 1980s, a lot of world linguistics was geared to seeking universals and lots of linguists tended to emphasize Saussure’s notion of langue as though it was one form for all. But you have always emphasized that variation is the essence of the endeavour of linguistics. How has variation been an organizing principle for your research?

R: Well, I’d say that variation is central to doing linguistics, though sometimes thinking about the term, it has seemed to me that its rather careless use has to do with how one approaches the term ‘variation’. What does one mean by ‘variation’ speaking ordinarily, not technically? Everyone knows that the base of the word is the verb ‘vary’, and so, thinking analogically about how things work in English, you might say okay ‘vary’ is a ‘doing word’ and ‘variation’ is ‘a naming word’. We know that ‘to vary’ is to differ; so it follows that ‘variation’ means difference, and so it does – that’s why David G. Butt and Jennifer Yameng Liang 389 sociolinguists chose that word for technical use. But it has to be remembered that while in everyday use all ‘variation’ is difference, not all difference is ‘variation’ when it comes to the technical use of the word. There is hardly anything you study where you cannot locate differences and similarities. I first began to appreciate the significance of that term as I began to read people such as Labov (1972a, b). Trevor Hill, you know the inventor of the term ‘institutional linguistics’ (Hill, 1958), had always been in Edinburgh. But it was after reading Labov and listening to Michael on functional semantics that I began to see what it means to have institutional linguistics and what kind of difference is referred to as ‘variation’. All that sort of coincided with the time when Halliday et al. were writing their book, The Linguistic Sciences and Language Teaching (1964). And as a matter of fact, one way or another from that time on, the concept of variation has been in the forefront for me. It is a central concept in linguistics as I just said. Why? Because it is one point in the study of language where you can see how certain aspects of culture actively enter into the process of language, perchance making their way to the system. This happens when you look closely into what people do with language, who these people are in terms of Basil Bernstein’s ‘social positioning’, and what kinds of activity are ‘open’ to them. I guess working with Basil (e.g. 1971, 1990) was another factor that had really made me see more clearly the significance of variation from the social angle: it is in this area that linguistics has a tremendous contribution to make by sensitizing us to how language in use, and its evaluation, play an active part in the life of the speech community. In the 1973 paper on ‘Code, register and social dialect’, I had tried to show that varieties in a language and their effects are central to the study of languages. It doesn’t matter what kind of variation it is – dialect or register. Whatever it is, it is inherent ‘variation’ in language, and the secret of its persistence lies in the fact that society has been run by classifying human beings – language is a part of that game of classification and differentiation. Even the valuation of registers bears a deep relation to these issues. When I came to London, I was again surrounded by variation studies. All of Bernstein’s work was about variation; the project on register in Halliday’s department was about variation. And later, as I began to teach, the experience of teaching sociolinguistics particularly forced me to recognize how much more there is to variation than just A being different from B – it put a lot of pressure on me to understand the term as used in Labov, and the implications of what he was saying. By the time I came to that big sociolinguistics research about mothers and children based on Basil’s ideas about the role of language in the formation of consciousness, I was already thinking how great Labov’s work was as a model for reliable studies on sociolinguistic variation and how much more elaboration is needed to unpack the meaning of ‘variation’ – some of these thoughts you might have come across in the second volume of my collected works (2009), especially, the first two chapters. 390 Society in Language, Language in Society

So all this was in the background, before I began to pay closer attention to Saussure. It was exciting reading him, and it seemed to me, it’s okay for him to establish état de langue, but an isolated stage of language had to be a theoretical fiction; its value lies not in any kind of physical reality, but in what it allows the analyst to explain. After all, each such ‘synchronic stage’ is, as it were, sealed off from the others – those adjacent to it. Say that the temporal layers are like those we recognize in archaeology, then the stages of language are comparable to individual layers – does this remind you of the figure in the Course of General Linguistics (Saussure, 1966)? Well, fair enough, this draws a boundary of some kind between the various historical stages of language, but the fact is that language is never in just one stage. So there are problems in this: first, no état de langue will give you such a tidy picture of language as a homogeneous system; it will always show variation at any one moment, any ‘state’ whatever; some changes will have completed; some are in the middle of taking a sort of stable direction, and some are just beginning. Then again, no matter how much you draw the boundaries by reference to temporal layers, you have got to be thinking of how language is experienced by members of the speech community – what happens, who is talking to whom and why, and so on. Every use implies users – typically in the plural; that implies some relation. It could be age-based relation or sex based, or social status based. You see where I am going? Variation in language cannot be avoided except by fiat – at least, in the world in which we human beings live, it is always present in every single culture. What do you gain by pretending not? There is some kind of tension between ‘variation’ and our conception of ‘doing grammar’; grammar is about generalizations that apply to categories as a whole, or so we say. By definition, grammatical descriptive statements are what Saussure (2006) would call statements about ‘templates’. (Wonder if his templates are like Firth’s ‘typical’ – you know, Firth always said that the events worth describing are the ‘typical’ ones, not individuals.) What Labov’s work has brought to attention is that, just as with grammar so also with ‘variation’, an empirical (‘synoptic’, would you say?) study has to be about systematic behaviour in communities. Weinreich, Labov and Herzog (1968) convincingly argued that variation is orderly. It follows that if grammatics has to be about templates/types/categories as a whole, then these have to belong to one variety: no one has shown how to describe in one breath, so to speak, both the patterns in AAVE1 and in Standard English, except by superimposing one on the other. That is a challenge to any grammatics, universalist or not.

D: You seem to be saying that variation has to be looked at in terms of a total model, a theory of language. It can’t just be nibbled at as instances here and there – the whole model requires reconstruction. David G. Butt and Jennifer Yameng Liang 391

R: Well yes; reconstruction or at least first of all, some reflection – that stands to reason. It is a theoretical concept, variation; it has to be reconciled with the other theoretical concepts unless what we have is a conglomeration of theories about this, that and the other. I tend to think a theory has to be unified; it cannot sprawl in a serendipitous manner. This is why it seemed to me that variation is just absolutely central to anything that you do in linguistics. Take the question of what grammar did I learn first, which dialect did I speak? And did everyone around me speak the same way both grammatically and phonologically? The grammars that we write, how much do they represent of any one état de langue? They certainly don’t tell the whole story. Until very recently all your examples in grammars came from written language. So these were things that made me think if you do not pay attention to variation properly, you are doing ‘partisan descriptions’ without realizing what you’re doing. You can’t really have one theory of this bit, another of that bit – you can’t have a theory of variation and another of homogeneous grammar without recognizing that other valid grammars of the same ‘language’ exist; so you have to think more comprehensively about language. Well you may have heard people who just talk about this little text being different in some respects from that little text and they think they are talking of variation. But variation has to be understood in its fullness (David: fully across the strata!). Yes, certainly strata. I feel that the distinction between what Abercrombie called ‘accent’ (one aspect of Labov’s ‘social dialect’) and what was seen as social dialect that involved variation in wording, as well as the possibility of variation at the stratum of semantics – these specific kinds of variation at each stratum seem to in a way ratify the recognition of linguistic strata. Then again, nothing else in language brings the necessity of recognizing context of culture and context of situation as a part of the theory of language: neither register varieties nor variation can be properly studied without an understanding of the context of culture.

D: Yes, that’s an important point, isn’t it? You’re already in a culture when you get to the context and language.

R: Yes and going back to ‘instance’ that you just mentioned, it has seemed to me for some time – actually since Michael theorized the word ‘instance’ – that the popular usage of the term displays a confusion: the usage does not distinguish ‘instance’ from materially located occurrence, and it seems to me we need to spell out the difference between the two. An instance is an instance of some systemic path: it realizes a recognizable category deﬁned by linguistic description as a part of the potential of language. A speech event seen as a materially located speech event may be more than a recognized category or less. Take, for example, the stories I discussed in that unpublished chapter called ‘Measuring the length of a language piece’ – I wrote that in 392 Society in Language, Language in Society

1973, and it got used by many research students to establish message or clause boundary. Now the stories were extempore production by young children, and the messages/clauses in the story often had features that at least I have never seen discussed in terms of the systemic potential of an English clause type, like quite often the kids would pause, turning away from one ongoing structure to another leaving the first one dangling, or there would be semantically unmotivated repetitions and so on in most of the stories. Now whether you took the clause as your focus or the message or the story, you couldn’t say these things were elements of any such category’s potential. But they were certainly part of those occurrences, and they were displaying a feature of the context of situation that has no inherent relation to the potential of clause or message or sentence, though it is a feature that does tell you that the language piece was an extempore production. You describe categories, not language pieces: so the instance represents some valid configuration of the choices that are available to a unit: these make up the systemic path being realized by some instance under focus. And as a linguist, I’d say my interest is in describing the potential of some categories, rather than a language piece. One way you could put it is that sometimes a language piece might completely coincide with a recognizable instance; but whenever there are ‘bits and pieces’ that do not coincide with a recognizable instance of some category, then the choice is to ignore the overflow. For example, imagine that, in preparing a script for analysis, we find that the actual occurrence runs as ...and um um when you when if someone calls a baby it ... what you identify as a clause to be analysed is really the segment if someone calls a baby it; what you are implying is that this segment is the only part of the language piece here that can instantiate a systemic path at the rank of clause; the rest connotes the extempore nature of phonically presented dialogue, and it indicates the speaker’s effort to bring orderliness without having had time to think much about it or something of that kind. Well, you are going to find out the truth of all this when you have to impose order on this ‘naturally occurring interview’! You will screen out those sorts of things when you turn it into ‘respectable prose’; all the comings and goings from this structure to that one, and you will iron out many of my formulations. What do you suppose you’d be doing? You are actually turning it into a series of instances that are the realization of some established systemic paths. You are taking out things from the actual occurrences, because there is no system that can ever consciously produce occurrences. (Abercrombie once wrote an excellent paper on the difference between natural conversation and simu- lated conversation as found in novels or plays.) By and large, all speakers have to respect grammar because without the regularities of grammar the intelligibility of the dialogue might be jeopardized. The evaluation of variation by reference to who speaks a given variety has muddied the situation. We think PC (political correctness) is a recent phenomenon. Not true. David G. Butt and Jennifer Yameng Liang 393

But focus on the instance is a good reminder of the underlying system. In an ethnographic analysis where this kind of looking to the system might be absent, one is simply elaborating one’s understanding of the meanings as if there was a direct link to meaning from the sound – like some of those introductory linguistics books used to announce ‘language is a binary system of sound and meaning’. (Don’t misunderstand me please: I am not damn- ing the ethnographic analysis ...) Take, for example, Carmel Cloran’s article (1989) on the learning of gender. One part of it was largely ethnographic, but because she had a theory behind it, it comes across as a different kind of thing than simply picking up a text and pointing out, ‘oh look at this, this is an interesting part’ and ‘look at that, that sounds fascinating’ and just saying ‘this is a variation on that attitude’: this is neither a language analysis nor a proper study of social attitudes. You can do excellent ethnographic analysis – don’t forget Garﬁnkel (1967) and his group of scholars. But as a linguist, the analysis that is most signiﬁcant to me is the analysis that attends to language in a particular case – looking at it from the point of view of how it was possible to create that construction of meaning.

D: How did all this bring you to semantic variation? Is that what carried you into the studies in the 1980s of socialization of young children?

R: No, no! Actually, Basil did. Working with Basil, and of course Michael’s interest in recording the language of Nigel and his writing on learning mother tongue as learning how to mean. So there were two pushes. One was coming from Basil Bernstein’s work, which was all focused on language in a way, because his main thesis was about the central role of language in developing different forms of consciousness. The hypothesis was that classes are not transmitted genetically; they are transmitted linguistically. A person prepares him/herself to occupy a particular position in society by the patterns of interaction he/she experiences. And do you know, I feel proud that, together with everybody like Carmel [Cloran], David [Butt], Rhondda [Fahey] and a lot of others that helped me, that I was able to show that this hypothesis is tenable, that this was actually the dominant message of our research based on that data. I was able to show that yes, the mothers were transmitting the mores known to them from their own positioning to their children, not directly but by implication of what they said. They were generating in their kids forms of consciousness. Another phrase Basil used was creating ‘different orders of relevance’. He often pointed out that there are certain things that teachers say in the classroom that do not appear relevant to some children, so they ﬁlter them out. If you have observed little children still learning their language, you might have noticed an interesting situation. There is this little child sitting on the ﬂoor, playing with his toys. And here is you, the mother, talking to a friend and you are talking about linguistics. Now, the child will absolutely ignore all that. Just turn your talk 394 Society in Language, Language in Society to being about a drink of chocolate and immediately the child is in the discourse, or, talk about anything to do with the sorts of things that the child is able to be interested in. I do feel that in a very concrete kind of way, the issue of relevance should be given more attention. The formation of consciousness in a way was also coming from my own background as a teacher and being concerned about how does one change people’s ways of thinking. My interest in literature was because I wanted to go back to Pakistan and teach literature in such a way that it would develop sensibilities and humanistic attitudes; that would have counteracted the fun- damentalism of social convention and especially religion, though religion was nowhere as vulgarized as it is at present. So, in a way, I put my own dream together with Basil’s hypothesis, and it seemed to me that the formation of consciousness was what education was all about. In fact, I do believe that the failure of Marxism is truly the failure of the leaders of revolution to create the kind of consciousness that is needed to practise true commu- nism: they never seriously tried to understand, much less to explain, the meaning of living in a communist society. Nearer home, the problem for a researcher was the operationalization of the enquiry. Here, Labov’s writing was very helpful in the project design, and the combination of Bernstein and Vygotsky gave one sociological and psychological insights – Halliday’s SFL enabled the analysis. I am grateful to many people. I hope all these details are clearly presented in my volume 2 [The Collected Works of Ruqaiya Hasan] (Hasan, 2009).

D: It is interesting to me that the ﬁrst volume (Hasan, 2005) in your set of collected papers is on language, society and consciousness. Out of all of the papers that you might have started with, these are the most cerebral and in a technical sense. They are to do with the construction of the mind.

R: Well I dedicated that volume to the memory of Basil and the topic still speaks to my passion. During my ﬁrst visit to the US in the mid 1960s, what made me wonder was their denial of ‘classes’ in the country: they thought they had ethnic differences not classes! And again later, my early 1970s visit there again made me realize how much work needs to be done to clarify the relations of language and the human mind. Of course, one did read people that were topical then like Piaget (e.g. 1959) – interesting in so many ways but still leaving me with a lot of questions. Somehow, despite having seen Vygotsky’s Thought and Language (1962), I never stumbled on to his more important materials (e.g. Vygotsky, 1978). I really got to that work after settling down in Sydney. Now, it seems to me, that to study Basil without Vygotsky, or Vygotsky without Basil, is to read only half the story: the link between the social and the psychological in ways of saying and ways of meaning is provided by these two scholars – and it is Halliday whose linguistics has unlocked the secret of how to study language as a potential David G. Butt and Jennifer Yameng Liang 395 for meaning, so you can be consistent in your analysis of data while keeping in mind the point of the analysis – the meaning exchanged within a community. These were the kinds of theme that were then uppermost in my writings, particularly one of the papers included in that volume (Hasan, 2005) that I am very fond of, about three exotropic theories.

D: Yes, of course, the other one is Halliday’s work. I found it working on your project; you explained the Marxist understanding of the construction of society and mind through an example like the one mother who had the resources to replace any toy had choices in that she could turn any accident or mistake or breakage or carelessness that the child made into an exemplum, whereas for another mother, it was essentially catastrophic because things couldn’t be replaced so easily. One could see the material condition actually gave very different choices about meaning. It wasn’t just choices about material.

R: Oh, did I talk about Marx – or was it Bernstein? Anyway we do talk about time. We are interested in managing time. Think of animals: they have to invest so much of their time in getting their nourishment – there is hardly any time for anything else that might have perhaps evolved at least in the higher mammals. But time also matters for human beings; I’m thinking in terms of managing time practically with little children when you don’t have any help. Among human beings the management of time across social classes is of a different order so that it has other signiﬁcant manifestations. You don’t have time if you have to get the dinner ready, to get the house tidy, and to also watch the child – the primary thing is that the child should keep alive, keep out of mischief, and create no chaos; those are the primary concerns (laughs). Obviously, mothers in certain social conditions would not have the time or the habit of mind, which would incline them towards explaining rather than immediately putting a stop to whatever is going ‘wrong’ from their point of view, by direct directives, threats or whatever. And one has to remember also that explanations do occur in both groups of mothers: the difference often lies in what is considered worth investing time in explaining.

D: One of the things for people who are not familiar with these statistics that emerge in the project is that the only outliers in the patterns that you found were people who had had dramatic changes in their life circumstances.

R: Yes, yes, that was interesting! One of them was a middle-class mother whose change of circumstances was extraordinarily recent, just like ten months or less. And you can see how much difference that made – how time is to be invested became an issue. That was the outlier mother in the middle class. In the working class, the outlier mother had been a teacher 396 Society in Language, Language in Society and came from a slightly different stratum of life; she had married into a different sort of family, and that made a difference.

D: An interesting thing about the semantic variation study is that it also led you to network representations of semantic options so that in a sense the theory, with the strata and realizational relations, then connected semantics to context and lexicogrammar. What achieved this was the elaboration of semantic networks, so in that sense, your research project was ﬁlling out a more general issue in the study of semantics.

R: Yes, this is what I was saying a little while ago. There is so much that is interesting about meaning mainly because if you take people in everyday life, whatever they may be saying, whatever the reason – it is difficult to put it in words you know, but I’d say speakers have a kind of preview of their forthcoming meaning. You know how you sometimes go over the same ground; I think what you are doing is to bring your meanings closer to the meaning that you set out to express – it doesn’t matter how remote it is, but there is some shadow of an awareness of meanings that you want to get across. It has always seemed to me that the way we talked about meaning being ‘produced’ by grammar did not really do justice to that sense of ‘knowing where you are going semantically’, which is so obviously there in human speech behaviour. Michael Halliday has said something to the effect that interactants understand meaning because they meet the meanings halfway – they actually anticipate them. This to me is really a thought-provoking claim. The urge to articulate meaning has to come from the speaker and it has to arise in him, be activated by the kind of situation in which the speaker finds himself. Now it is interesting, though in some ways an absolutely obvious fact, that the only material entity that must always be present in the material situational setting of a discourse is the speaker: his material presence at the point of production is mandatory and sufficient for some varieties of text production; and it is a necessary and a sufficient condition for his semantic presence in the text. Put it this way: there is nothing in the text that the interactants do not find relevant in some way or other. Actually they provide the degree of relevance, because their desire/need to get something done somehow or other is what determines the relevancies in the contextual configuration. Sometimes I think I am perhaps not being fair to Firth: you know I am thinking just now of how he commented on the speaker carrying his culture within him wherever he goes. It is obvious of course, like Canterbury Tales, once it has actually been laid in front of you, but it’s brilliant. If you ask about where the experiential stuff comes from in the exposition of, say, blood circulation, rain cycle, or the Egyptian burial rites – you have to go to the speaker’s experience of knowledge in his community which he carries about within him as he carries his culture around. David G. Butt and Jennifer Yameng Liang 397

But Firth did deny the need for a stratum of semantics by dividing all the labour of meaning description across the strata, though you might say he didn’t have a very clear and fixed sense of the positioning of strata. All the same, by distributing the semantic description across the entire ‘prism’, he denied semantics the possibility of having its potential represented in a distinct stratum, and along with that also the possibility of establishing the explicit realizational relations – of both kinds, the activation and the construal. If grammar or phonology has developed, wouldn’t you say that it is to some extent because these relations between them have been made precise – the potential of each is made explicit. It is not really true, is it? That we first think of grammar and then the meaning comes. Or we first think of meaning and then the grammar comes: even when we have the preview of meaning as speaker or addressee, the sense of meaning without the sense of grammar must be just a phantom. And I still don’t know where in the inner organization of language he positioned his context of situation. And I can’t say that I like his extension of the scope of reference for context to what is say collocation, colligation, sandhi and whatnot. However, we are not aware and we could not – at least I can’t – say at what stage meaning is actively achieved. We create these strata because they help us explain how we see the relations, not how we articulate. What the semantic network did in the case of my project was that it explicitly laid out at least some of the realizational relations between grammar and semantics at the rank of message and clause in semantics and grammar, respectively. It is work that needs to be followed up – I am not saying that’s the only way to go, but whatever it is, it’s got to be at least as explicit as that is. You should not overburden grammar as if it is the alpha and omega of language.

D: That’s right. They – I mean the strata – look artiﬁcially bounded and so, you can have people like Chomsky saying this stratum is autonomous. No stratum is autonomous.

R: No, there is no autonomy at any point and you are not really getting close to any kind of truth in saying, for example, that if you go into grammar far enough, you get into meaning. What does it mean to say ‘if you go far enough’? This in itself is a fairly clear indication that the strata are a convenience, a theoretical ﬁction, which helps us explain things as we experience them rather than a physical force out there like the wind, or gravity or whatever.

D: I’ve always felt that Firth’s notion of ‘exponence’ was trying to get at the point that you are making, the way he used ‘exponence’. He did have that division where you had the grammar and then above that context of situation. At least at one stage in his writings. 398 Society in Language, Language in Society

R: In a sense, if you look at the child learning language, he is doing precisely that. He has signals; he has some signing devices; and he has context. What does he use all those resources for? Well, he makes meaning.

D: It’s the twenty-ﬁrst century now. Where do you see the important developments going in linguistics? We’ve now got corpus at a higher scale than we had in the past. Most people can carry corpora on their computers. We’ve got interactions with other disciplines that were rather nascent in previous eras. On the one hand, there are the opportunities that do seem to be emerging, and then there are the opportunities maybe that you feel ought to be emerging that people haven’t paid attention to. So the things we ought to be looking towards. Maybe some that are positive developments. Do you have any views about that?

R: Well, you could look at this question from various points of view. One which makes you feel the most pessimistic is the current developments in education, where short-termism has become so powerful that there is no sense of why the university is important; we don’t ask where will the disin- terested search for knowledge – not for material gain but just out of curiosity; where is the knowledge going to come from that might become available for one to apply. All of these issues are as if buried under that one driving force: ‘let me be more powerful than the next guy’. In the light of that, it seems somewhat like trying to build Utopia if you say that every university must have a department of linguistics, a department devoted to the study of language and languages in relation to each other, and in relation to the life of the community, with variation in the centre, because an understanding of this would be to many of us news – for example the fact that variants do not have to be graded along the axes of right and wrong. The explanation why we do this – simply to create differentiation amongst human beings – this is an eye opener because you know we do have this mythology that we are a just, compassionate society: but perhaps when we say these things, we kind of close our eyes to what we do in the name of God, of nation, of my country, my family, and so on. Humanity as a unifying concept comes last to us. These big questions that it is necessary for communities to understand to make any real progress as opposed to the gimmicky progress of ‘packag- ing’ – these big questions are necessary for creating a mentally/emotionally healthy society. Then, the analysis of language has typically had a bad press: it is difficult; it is not interesting; it is not scientific. The flexibility of language, the permeability of its categories is treated as the result of some kind of lack of rigour in analysis. At the bottom of all this is the fact that language is a highly complex system: we need to understand the sources of this complexity. We have often tried to fit the theory of language into the mould of the theory of some other discipline – let the linguistic theory be like that of the hard sciences, let it be like that of psychology, view it from the point David G. Butt and Jennifer Yameng Liang 399 of the user, and on and on. Now the pursuit of questions that go beyond the here and now seems like a dream – a distant dream. We are fighting over who pays for all this, while the majority of us are paying heavily to maintain a pretty dysfunctional sociopolitical system, that is ostensibly for the good of the country and the nation but neither seems to be doing too well.

D: What do you mean by ‘permeability’? You mean you can’t tell the difference or what ...

R: It’s not that you can’t tell the difference. It is that you describe a category no matter how carefully, but sometimes, the category changes its face – for example take the present tense in English; you say it construes ‘habituality’ as for example in he walks to the office. But then you also have such uses as he leaves for London tomorrow, clearly not a habitual use and there are other meanings of simple present tense changing its semantic face. Now this is what I would describe as: the category of simple present tense shows permeability; it has the potential of one default meaning construal which is its typical meaning, but also given certain specifiable conditions, it will construe other, different meanings. Those ‘conditioned meanings’ are not haphazard; they are, like all potential, systemic and regular – but in an environment that is specifiable. There is really quite a bit more to it. I wrote a chapter for Donna’s [Miller] book on hybridity earlier this year; my chapter discusses this concept in some detail. I first began thinking along these lines in that paper, the one that is the opening chapter of my volume 3 (Hasan, 2011). Perme- ability is a feature of well-defined categories – not all categories of grammar but some – and most probably those also at the level of context have been flagged. Perhaps one way of looking at it is to say that permeability presents an argument against a one-to-one realizational relation across the strata.

D: Let me just jump in there; it is interesting that, in this week, I’ve been asked in two professions, one in surgery and one in psychiatry about being parts of centres or projects to do with communication. Yet I ﬁnd trying to defend linguistics on my own campus, I’m not ﬁnding it easy.

R: Oh well I might say – that’s Bernstein’s horizontal knowledge structure for you: what you call linguistics is not what someone else does, and instead of the department being a place where we problematize issues of language, what we do is problematize our own needs; doing linguistics becomes a source of my fame or yours – the world is not large enough for two people to be innovative at once, have fame at once. Our governments see to it that we fight for ‘higher ranking’: we don’t get the benefits that the other one does – so it’s not easy, is it? Now you might remember that Michael Halliday has also talked about this situation: he believes other linguists cannot be interested in SFL; they ask questions that interest them, whereas SFL always 400 Society in Language, Language in Society attempts to consider questions used by those who are using language for some purpose. And this seems to be true. What can we do about the situation? Now this may sound absurd to you, but if I were to answer that in one sentence, I’d say it like this: let us be very careful how we use words in doing linguistics. Whenever we have to justify our linguistics, it will have to turn on what it can do, what it can tell us about language, about language use, about its relation to brain, to human progress, whatever – all this we can explain if we have used language precisely in talking linguistics. Forget, if you can, about others: figure out how to present your linguistics clearly to those who work in the areas where language is critical. The important thing really is to get through to those who want a theory for describing discourse from some specific perspective – as for example in the work by Di Slade, her team both here [Sydney] at UTS2 and at Poly U3 in Hong Kong, or by Geoff Williams, or that by yourself [David] and Wendy [Bowcher] in the domain of forensic linguistics, these all show that language is a part of our life. Conflict resolution, analysis of political speeches and so on, these all need linguistically informed, semantically oriented descriptions of language – because you know the world really moves not on love; it moves on language. So there are lots of people who need linguistics that is written out for them and that states in clear terms what linguistics is, what it does, and how it can help them achieve their goals.

D: I want to return to the twenty-ﬁrst century: there are noticeable new trends today that were not present in the formative years of SFL. For example, what would you say about corpus linguistics, or multimodality? Both of these trends are strong now. How do they relate to linguistics? Do they relate to linguistics?

R: Well, those are interesting questions, David, if not the ‘leading ones’ (laughs). Now I have the impression that my position on both these developments – especially on multimodality – is perhaps not exactly as I actually think of these developments. You know one advantage of having worked so closely with Michael Halliday is that you begin to see the strength of his position – it is not one of exclusion; it does not draw strong boundaries around language; rather, it includes anything and everything that bears any relation to language as a semiotic system. Though sometimes I have also felt frustrated. Put it this way: it is pretty difficult to create a theory of everything. But even if you succeed in doing that, there’s a serious problem: if a theory explains everything, then of course you have no theory for the umpteen things across which you perceive some differences. You have to follow the middle path – something like Gandhi you know. It was this kind of thinking that lay behind my distinction between endotropic and exotropic theories (Hasan, 2005, ch. 2). A theory might segregate its object of enquiry, tightly excluding it from other phenomena experienced by human beings: David G. Butt and Jennifer Yameng Liang 401 you explain everything about the object of enquiry by just looking into that object and never ask what this object is related to in the experience of the speakers. Such a theory is endotropic. By contrast, an exotropic theory iden- tifies its object of enquiry by relating it to those phenomena that impact on it, changing it, supporting it, whatever, while it (the object of enquiry) in turn bears some strong relation to those phenomena. Simple example: you might say ‘language is an instrument of communication’: now, it is clear that forms of communication lead to the development of language; languages die when no one communicates with them. At the same time, the quality of communication when language is used is qualitatively different from that by pictures or gestures; the way that language is used makes a big difference to the success or failure of acts of communication. Now, language is pervasive: we use it everywhere in life for doing a large number of things. This does not mean linguistics has to include the description of the various social institutions whose practices depend on communication. This would be in fact undoable. Instead, linguistics enquires into forms of communication, registers, what makes them vary systematically and so on. In most societies today, the various social institutions are under study and the description of those disciplines is a resource for the interpretation of the conventions of how they work in practice. But as a linguist, our focus is on how language works in the life of the community and why it is able to do these things. So the exotropic theory of language has that other virtue that Michael is very keen on: eat your cake and have it. Recognize the institutions in the culture, recognize every discipline in the world but as a linguist your focus has to be on language: what connection can language have with biology, sociology, geography, human history and so on. There is tremendous advantage here: you are not locking yourself into one small box, but you have identified what matters most to linguistics: the description of language from as many different points of view as appear significant in human experience. With this kind of orientation to the study of language, you are never going to say that other systems of meaning are irrelevant – well they are a kind of semiological system, aren’t they, so let’s see what’s in common, and what differs. Similarly, you wouldn’t say corpus is irrelevant: after all, corpus is a collection of language in use on many different kinds of occasions. This collection may be useful to you as a linguist and how much might depend on the method of its collection, the degree to which it is possible to say something reliable about its history, identify its context and so on. Corpus is one thing – and you could paraphrase it as ‘collection of samples of language in use’; corpus linguistics is another. It says there is a linguistics that is based on data: what does that mean? Based how? If I use the library of literature as data, that doesn’t make the data corpus linguistics? But corpus linguistics has methods and means that are available to an analyst only under certain circumstances. The many methodologies that have been invented to interrogate the data are a great advance in providing evidence for claims. Of course 402 Society in Language, Language in Society it would depend on what kind of questions it allows to be asked? On what research is the method based? Why some questions are more important than others? How relevant are they to writing systematically organized grammars? Is there advantage in doing grammar word by word? All these are open questions – I am afraid I am speaking from no first-hand knowledge of this mode of analysis. I can see its power – particularly if it could keep both variation principles and whole language in view – but I have no access to any corpus nor have I been able to use the methodology unique to corpus analysis except once thanks to Wu Canzhong and Annabelle [Lukin] with whom I worked on something to do with the grammar of space many years ago now, perhaps in 2009 or 2010. But it is obvious that the elements in its methodology identify some kind of ‘units’ of language. Some knowledge about language is coming from somewhere. It is a way of describing language. Is it the best? That remains to be seen. The point is as a linguist, you need to give reliable information about language. And you can do that only if you know why you analyse this unit this way rather than any other; what place does this little bit of a pattern have in the ecology of the whole language? How does this register differ from that? Do I know where my data are coming from? And a lot of such questions. So in some way it would seem that corpus linguistics duplicates work that has been being done by linguistics. Whether it has tried the available linguistics and found it failing is something I do not know.

D: I am thinking of a quote in Christopher Hookway’s book on C. S. Peirce. In the foreword to the book is a sentence attributed to Oliver Wendell Holmes: ‘Theory is the most practical thing there is.’ Whenever I listen to you talk, it is this message that seems to be coming across. Get the theory in a better form and you will solve and resolve issues more clearly.

R: Yes. In fact, I couldn’t agree more with myself (laughs). In my view, all analysis, any analysis, that we undertake, is really doing something about the theory. Either it is saying ‘this theory is such that it can explain this here pattern I am analysing’, or it is saying ‘I’m having to fudge here because this theory hasn’t provided the right resources for this kind of pattern’. I think it is always necessary to ask: what is the nature of this pattern and where is it positioned among the patterns identiﬁed at a general level in my theory of language?

J: Regarding the science of linguistics, how can one apply rigour and precision into doing linguistic analysis, including the testing and refutation of statements, and the attempt to make statements clear and less equivocal?

R: The method I’ve taken is this (I’m not saying everybody else will agree with me): I begin with the assumption that SFL theory is a good theory and that it will work. This may predispose me to ﬁnd that it is so and that it David G. Butt and Jennifer Yameng Liang 403 works, but at each point, I do try and remember to ask myself: ‘I’m saying this. Why? What is the justiﬁcation?’ In fact, my papers sometimes get so tor- tuous and long because I insist on asking, and explaining – I won’t give up, like a dog with a bone, as one of my supervisors told me. Of course, there is the point in the words of Wittgenstein: explanations must stop somewhere.

D: It seems that your papers have two parts particularly, if I may just talk as someone who is responding to hearing you speak on many topics. In one sense, like you did the issue of variation – what do we mean by variation? So there is the broader theoretical essentialism (the verb ‘vary’ is the root; where do we go with that and how do we deﬁne it) and there is the other aspect which you rail over often, which is we’ve got a tool like a network, so how do we apply the tool with maximum consistency. There are sorts of two levels of theoretical pressure that you bring to bear.

R: I say to myself usually, if I make a statement, ‘x is the case’, it should be possible for somebody else being given the same situation, to come back and say ‘yes’ or ‘no’ (David: Yes or no). But I don’t insist that the task of a theory is to show how it will fail. Why not the positive environment check: under these conditions this principle will work; so long as the environment is not going against any theoretical maxim I feel this way of formulating your theory is much better (Maybe it is a misreading of Popper); nor do I believe that a theory is the theory when it starts. Theories are evolving things because they are really a source of explanation, frames for explaining something that you are putting together. You would have to be superhuman to be able to take a virtual not material area of analysis and get everything right in the first go. Michael’s ‘the law of the failed first try’ is significant. It seems to me that precision is needed precisely because you want to specify neither more nor less than just what the theory is calling. You know, for me the first problem that was raised in SFL is a wonderfully rich and open question: how does language work in the social life of its speakers and why? If you don’t ask precise questions, if you don’t use the theoretical terms precisely, hiding behind some kind of metaphor, well: it might make for interesting reading or listening, but it is too difficult then to check on the validity of your statement. For example, I have problems with the usage into which we’ve got with regard to the word ‘realization’: realization is a relation that applies up and down the strata in linguistics. Then also realization is a relation that holds between system and structure: but this is a relation across ranks, whereas the former was across strata. These are two rather different kinds of relations; and what’s more, as I was saying just a little while ago, even within the stratal realization there is a difference in how realization works between the strata above Halliday’s ‘line of arbitrariness’. You can’t really maintain that for example, the final voiceless sibilant /s/ construes the grammatical category plural; if that was true then /fars/ (farce) should be plural. To cut the arguments clearly, the value assigned to /s/ comes from grammar. So, we cannot talk 404 Society in Language, Language in Society about the relations of grammar and phonology in exactly the same way as we do when we talk about the relations of grammar to semantics to context of situation: the structuring within the content plane seems to be governed by a different principle from that which governs the structuring of content and expression plane into one unified language. Whenever I’ve lectured to my classes, particularly the semantic ones, I’ve pointed out these things. This was not to say ‘this theory is no good; go get another’, but to alert them that there is this inconsistency. And inconsistencies in theoretical relations must not be ignored; they need to be debated. So think on it – see what you can make out of it. This was necessary, because I was saying all this at a rather interesting time – mid-1970s to late 1980s you know, when Popper was very much in vogue. Also the talk about the so-called ‘context- metafunction hook-up’; oh, it does not work and we need another theory that will solve that problem and so on. Well I just felt that was the wrong stance – for one thing, the extension of the ‘hook-up’ metaphor to ‘context- metafunction’ was a rather unfortunate extension: it had originally been used in relation to grammar: how the structures derived from the different metafunctions conflated to reappear as one syntagm; were there generalizable principles for the conflation of functions across the distinct systemic choices, and so on. In any case, I became very curious about throwing out theories. I asked a few scientists at MU4 – and we had some international figures such as Jim Piper, a well-known physicist. He was quite amused: of course not; we never throw out any theory just like that; we tinker with it until it begins to work. So much for Popper and falsifiability in scientific theories. Besides, let us think a little about changing theories: the way I look at it, I believe every time you propose that a particular term that had been interpreted as ABC should now be interpreted as XYZ; this is a proposal for changing theory – not entirely, but it is changing something in the theory. So all such statements require supporting arguments. That’s one reason for insisting on precision, particularly in linguistics. I suppose all theories deal with generalizations and, so with abstractions. So your terms often refer not to something we now called carrot or cat that will stay carrot or cat even if you call them radish and potoroo. Theory is one place where a rose by any other name cannot smell as sweet, because your only connection to the terms of the theory is by precise idea of their relations – their ‘tool power’ (David, as you said yourself). So changing the meaning of a term is changing the theory to a certain extent. To my mind there is something I might call the alpha principle: this is the goal of the theory. If you say something as SFL does in my reading: the goal of linguistics is to show how language works in the social life of the speakers. This is a comprehensive formulation of the goal, because that ‘how’ carries the implication of ‘why’; there cannot be a more extensive goal. Explaining this ‘why’ and ‘how’ will take you to all the double-barrelled discipline titles that have arisen in the last century, like David G. Butt and Jennifer Yameng Liang 405 sociolinguistics, historical linguistics, corpus linguistics, psycholinguistics or neurocognitive linguistics. Any ‘tinkering’, little or big, that one might wish to make in a theory by way of changing it must be viewed in the light of the alpha principle: one has to ask how is this going to impact on the goal of the theory? Do I wish to change the goal of this theory? I am not saying that the goal of a theory cannot be changed, but if this order of change is introduced, then it is highly likely that the concept concerning the object of enquiry will also change. That’s a different kind of problem. Usually, whenever I want to introduce a change with terminology, I say to myself, ‘How would this change work in the theory?’ If I feel satisfied that it will help in attaining that goal, then this seems to be a good way to go. For example, I have been talking about the dialectic of the three strata at the content plane; I feel it is a move that helps to explain why anyone says anything and how is it that what they say turns out to be relevant; at the same time, it also shows how the addressee gets to know ‘this is what the speaker wanted to say: this is the relevant context of this text’.

D: In terms of Jennifer’s question about the science of linguistics, you use a kind of predictability issue, don’t you? For example, when you talk about meanings at risk with register.

R: Yes. Fact is, you haven’t got registers until you have a classification. You haven’t got a classification until you have identified meanings at risk. Predic- tion in linguistics is present in a lot of things we say: realization statements are a kind of prediction. Every system is a kind of prediction about the community’s understanding of a term and so on.

D: I’m just trying to show that there are ways in which people are wrong if they think linguistics is not scientiﬁc; but networks, for example, can be falsiﬁed in certain ways even though network principles might be standard, and particular networks might not always work as you and I constantly discover. Then there is the fact that there is the regularity of language for which you’ve got the very good terminology, thinking of register not as a block of meanings but meanings at risk, so there is the probabilistic.

R: Ah no, the credit goes to Michael – ‘meanings at risk’, that’s his expression, not mine. (Don’t know the text it comes from but it was he who used it ﬁrst.) But I use it a lot precisely because it is so competent – I say exactly the right things. Well it’s not easy to know about thinking, especially when you work closely with people – but I would never hesitate to say that I owe much of my thinking and writing to the stimulus of Michael’s work.

D: You’ve gone in directions that Michael hasn’t taken to the same degree of elaboration. 406 Society in Language, Language in Society

R: For example, it’s Michael who has set that goal for the theory. But then I also feel very strongly about one thing (and not because it’s Michael). Someone initiates a train of thought – they present some hypotheses. But the initiator is not superhuman you know, [so] s/he can’t do everything. The theory is not for the initiator’s use only, is it? Theories may begin with some one person, but they are, in an important sense, a social gift – otherwise it’s a theory that’s a one-day wonder. I do agree with Popper to some extent: what people say carries implications – including descriptive, theoretical writing. For example, what I did, say, about the study of literature, semantic variation, semantics or context was implicit in the goal of the theory as articulated by Halliday. This is what I mean: that was the text and my work is the elaboration.

D: Do you think that some of the big problems that linguistics has are not the things that Saussure pointed out that we don’t have any natural units, but that the natural units have this ineffability problem, what Halliday talks about. We have to use language to describe language and that puts us at a disadvantage when we are dealing with sciences that appear to have more natural objects.

R: I think this is an interesting thought! You know the other day I was reading this book by Davies and Gregersen (2010) and I found it quite fascinating that there are still certain unresolved questions about matter itself – its identity does not seem to be as ‘concrete’ as one imagined! The material sciences, the physical sciences, which are about matter, do not know how to describe matter. But they talk to each other: they are discussing precisely these difficult questions about the nature of matter! So I think in a way, simply because we understand so little of physics, we don’t really know what it is that we don’t know, or they don’t know – perhaps they are no better off than we are except in one thing: they understand each other; we seem not to. I just don’t know that I find it easy to explain why negotiation, rational discourse, discussion of disagreements is so difficult amongst the social sciences both across the various disciplines and also within the same one. It seems like we cannot discuss our own constructs, our models and what they are modelling. I doubt though that it is because our field deals in the nature of language, though ‘language turned back upon itself’ could be a problem. And there seems to be a sort of insouciance how we use our own terminologies – I think, like Humpty Dumpty, each of us is insisting the theoretical term will do as it is told by that individual – I mean there may be some exaggeration in this but not far wrong you know. This may hinder discourse. With language, people begin by the assumption that they know everything about language because they use it. It is just like saying ‘because David G. Butt and Jennifer Yameng Liang 407

I digest my meals, I know how the digestive system works’. It’s interesting the way people both undervalue language and also seem to have a visceral relationship to it. In some way, they don’t want to realize at any point that they don’t understand something that is being said in their language, about their language but which at the same time, sounds quite bizarre. At least this much is true they do not very often understand anything much that linguists might be saying: it’s just jargon.

D: That sort of captures one of the conundrums that make it difﬁcult for us to compete in today’s university environment.

W: The issue of engaging with other kinds of theories about language or approaches to language, like pragmatics or psycholinguistic theories. You did touch on Labov, but you have dealt with pragmatic issues and psycholinguistics, particularly with child language development, which is a psycholinguistics-dominated ﬁeld. I’m just wondering how you see the future of that kind of engagement, or how you’ve seen that kind of engagement in the past where linguistics has been taken over by other ﬁelds, like philosophy and sociology, rather than linguistics.

R: Perhaps the problem is that language is so big, and in a way linguistics has to be bigger, because it is language about language. When you have something big like that, it helps to see them as connected modules and to have recognizable areas of discourse inside linguistics. But so long as we think that a theory is good enough to extend into other areas, it is a good idea to show on what basis that extension is taking place: in what way they are alike. The absence of this has been my main problem with ‘multimodality’: I can see it is ‘multi’ and it is ‘modality’, but to me, it is not obvious that any of these modalities has as yet built a framework for describing its specific semiological nature – and that’s probably true of any ‘modality’ you pick, except maybe kinetics. Now I would say that you do need this understanding in order to be doing a multisemiotic/multimodal analysis of a text: how can you analyse a text for the patterns in some modality if you don’t know what units of analysis there are? On the other hand, I do have a sense that most probably my understanding is outdated, because people say ‘oh yes, but have you read Bateman’s or O’Halloran’s work’. So perhaps there are frameworks which specify the nature of these many modalities with which we are playing. Then again language is so big: no one linguist can do everything. At least speaking from my experience, you focus on some problem and that ushers you into the area and the scholars who are doing significant work there. For example, I talked about Labov, because Labov’s work in variation was absolutely central: no variation study could ignore it; it was good and the 408 Society in Language, Language in Society actual analyses are fascinating. But I took issue with his conception of variation. I tried to bring the tools from my kind of linguistics and attempted to show why the conception of meaning that he puts forward just does not sit with the kind of variation that he works with. I took the debate to his field: you talk about variation, but variation cannot really be discussed without including meaning, and if you include meaning, then the idea that variation, like transformation, is always meaning preserving would not work. So either you have to show that there is some convincing reason for maintaining that variation does not affect meaning, or you have to re-examine your approach; you need to reconceptualize how variation works. It seems to me that it was not a bad idea to do that. But I don’t know that it has made a lot of difference to anybody, though once in a while there are documents that seem to be bringing meaning into variation – how they will reconcile it with a universal semantic space is an interesting question. On the other hand, I believe that reflection on these matters has actively contributed to my work, as volume 2 (Hasan, 2009, especially chs 5 and 7) shows. I came out of that research quite a bit better informed than I was before! It was bouncing off Labov’s work that I was able to redesign my idea of how variation presents itself in languages. No discussion has taken place between SFL and Labov’s variation theory. So what? It didn’t really worry me much, that outcome; I had a problem in hand and I analysed it to my satisfaction: it is in conformity with the theory of language I think of as the best at the moment. Similarly, I have got into the area of psycholinguistics, mainly due to my interest in the formation of consciousness. I turned away from the kind of psychology according to which the body is preprogrammed to do certain things at certain points, but I found Vygotsky’s thesis of sociogenetic development of human beings far more satisfactory: the issue is not about chronological maturity; it is about how social interaction can bring about this sociogenetic maturity. But that is not to say that I have made any difference to the relationship between SFL and the work being done on Vygotsky. You don’t see flocks of SFL linguists looking into Vygotsky to see how his ideas on semiotic mediation might be of interest to that linguistic theory, and you certainly see no Vygotskian scholars paying any attention to anything except what Vygotsky and his colleagues have said. But you see, this does not bother me: I looked into Vygotsky’s work because I wanted to understand if the concept of semiotic mediation was relevant to a functional theory: I have found my answer. Yes it is relevant but, though Vygotsky is a great thinker, his approach to language cannot replace a functional model like SFL. The world of knowledge has no boundaries – you enter where you find something that raises your hope that it might offer some thinking on the problems that engage you. And if you read to gain rather than to treat with disdain, then you win. So I don’t care if Searle a philosopher worked on language: what’s great is that I can learn from what he says, whether I agree with him or not. That’s a minor issue. David G. Butt and Jennifer Yameng Liang 409

D: Vygotsky is arguing against Piaget and those other models as well as afﬁrmative models.

R: Yes, I agree – but to continue with SFL and related disciplines, let me turn to another example – literature. Here again over some years I have done some work as has David and other folks, but I don’t know any place where they have even a course as an option on ‘Language in Literature’ or ‘Stylistics’ or whatever. And interpersonally speaking, it is perhaps the most difficult of all fields, perhaps, because its history as a discipline is very long; it is very difficult to introduce in that citadel such a course. There is no deny- ing that the scholarship over the centuries has been often excellent, but at the same time when you think of what we teach by way of teaching literature as a discipline, it’s not all that impressive. It seems a pity that somehow the question is never raised: what are they supposed to be learning: How to catalogue the respected opinions on specific pieces of work? To spin a style that comes close to the style of a writer whose work we might be responding to? Or should you be asking sometimes: What do these linguists mean by ‘aesthetic function’? Where does it come from into a text? How is it produced? Certainly these are questions that a course on literature should be concerned with – it should make sure that the claims are in some sense open to scrutiny; they are based on something that might be counted as objective analysis, going above the replication of approved taste. But I don’t think that’s going to happen in a hurry: anyway, I am not holding my breath. So that’s pretty disappointing. What shall we do? Give up exploring the role of language in literature? Well, no, I don’t think so. Speaking for myself, I had planned to teach literature along those lines, based on insights from linguistics, when I got back to Pakistan: it just didn’t happen. I am glad though that I have done all these researches. And anyway I have perhaps been lucky because I had confidence that I was ‘doing’ a linguistics that’s probably the best, not because it has answered all the questions but because it enables rather than restricts; it guides rather than forcing; and above all it encourages exploration because it is exotropic. And nearly all of the fields that I have entered have strengthened my belief in the superiority of SFL. And this is true even today. I have been lucky that I have never had to say ‘I am doing this kind of analysis but I don’t really believe it’s going to be of any use‘. I don’t know what to say about this century, in which ‘glibspeak’ is the best policy for short-term gain. If one has confidence in the validity of what one is doing, well then, I’d say one should continue and enjoy the work one is doing – there is still so much in linguistics to do, and an army is not enough! And if I had a doubt, a nig- gling feeling ‘this isn’t right – it needs another look’, I wouldn’t hesitate, I would take that look. There is nothing in a theory that is sacred: that cannot be questioned. And you know, I’d remember Firth’s advice: ‘Above 410 Society in Language, Language in Society all know what you are doing!’ That takes quite a bit of reflection. All this may not bring one fame – simply that it will keep one at peace with oneself.

Notes

1. African American Vernacular English. 2. University of Technology, Sydney. 3. The Hong Kong Polytechnic University. 4. Macquarie University.

References

Bernstein, B. (1971) Class, Codes and Control,Vol.I:Theoretical Studies towards a Sociology of Language (London: Routledge & Kegan Paul). Bernstein, B. (1990) Class, Codes and Control,Vol.IV:The Structuring of Pedagogic Discourse ( London: Routledge & Kegan Paul). Cloran, C. (1989) ‘Learning through language: the social construction of gender’. In R. Hasan and J. R. Martin (eds) Language Development: Learning Language, Learning Cul- ture. Meaning and Choice in Language: Studies for Michael Halliday (New Jersey: Ablex), pp. 111–51. Davies, P. and Gregersen, N. H. (2010) Information and the Nature of Reality: From Physics to Metaphysics (Cambridge: Cambridge University Press). Firth, J. R. (1937) The Tongues of Men (London: Watts). Garﬁnkel, H. (1967) Studies in Ethnomethodology (Englewood Cliffs: Prentice Hall). Halliday, M. A. K. (1975) Learning How to Mean: Explorations in the Development of Language (London: Edward Arnold). Halliday, M. A. K. (2014) ‘That “certain cut”: towards a characterology of Mandarin Chinese’. Functional Linguistics 1 (2). Doi: 10.1186/2196-419x-1-2. Halliday, M. A. K., McIntosh, M. and Strevens, P. (1964) The Linguistic Sciences and Language Teaching (London: Longman). Hasan, R. (1973) ‘Code, register and social dialects’. In B. Bernstein (ed.) Class, Codes and Control, Vol. II: Applied Studies towards the Sociology of Language (Routledge & Kegan Paul), pp. 224–54. Hasan, R. (1984) ‘Ways of saying: ways of meaning’. In R. Fawcett, M. A. K. Halliday, S. Lamb and A. Makkai (eds) The Semiotics of Culture and Language,Vol.1:Language as Social Semiotic (London: Frances Pinter), pp. 105–62. Hasan, R. (1992) ‘Rationality in everyday talk: from process to system’. In J. Svartvik (ed.) Directions in Corpus Linguistics, Proceedings of Nobel Symposium 82, Stockholm (Berlin: de Gruyter). Reprinted in Hasan (2009), pp. 313–56. Hasan, R. (2005) Collected Works of Ruqaiya Hasan, Vol. 1: Language Society and Consciousness. Edited by J. J. Webster (London: Equinox). Hasan, R. (2009) Collected Works of Ruqaiya Hasan,Vol.2:Semantic Variation: Meaning in Society and in Sociolinguistics. Edited by J. J. Webster (London: Equinox). Hasan, R. (2011) Collected Works of Ruqaiya Hasan, Vol. 3: Language and Education: Learning and Teaching in Society. Edited by J. J. Webster (London: Equinox). Hasan, R. (2013) ‘Choice taken in the context of realization’. In L. Fontaine, T. Bartlett and G. O’Grady (eds) Systemic Functional Linguistics: Exploring Choice (Cambridge: Cambridge University Press), pp. 269–99. David G. Butt and Jennifer Yameng Liang 411

Hasan, R. (2014) ‘Towards a paradigmatic description of context: systems, metafunctions and semantics’. Functional Linguistics 1: 9. Doi: 10.1186/s40554-014-0009-y. Hill, T. (1958) ‘Institutional linguistics’. Orbis 7 (2): 441–55. Krippendorff, K. H. (1980) Content Analysis: an Introduction to its Methodology (Commtext Series) (London and Thousand Oaks, Calif.: Sage Publications). Labov, W. (1972a) Sociolinguistic Patterns (Oxford: Blackwell). Labov, W. (1972b) Language in the Inner City: Studies in the Black English Vernacular (Philadelphia: University of Pennsylvania Press). Lamb, S. M. (1999) Pathways of the Brain: the Neurocognitive Basis of Language (Amsterdam: John Benjamins). Nesﬁeld, J. C. (1908/2007) Outline of English Grammar (Delaware: Adamant Media Corporation). Piaget, J. (1959) Language and the Thought of the Child (London: Routledge & Kegan Paul). Saussure, F. de (1966) Course in General Linguistics (Translated and introduced by W. Baskin) (New York: McGraw-Hill). Saussure, F. de (2006) Writings in General Linguistics (Translated by C. Sanders and M. Piers) (Oxford: Oxford University Press). Vygotsky, L. S. (1962) Thought and Language (Translated by E. Hanfmann and G. Vakar) (Cambridge, Mass.: The MIT Press). Vygotsky, L. S. (1978) Mind in Society: the Development of Higher Psychological Processes. Edited by M. Cole, V. John-Steiner and E. Souberman (Cambridge, Mass.: Harvard University Press). Weinreich, U., Labov, W. and Herzog, M. (1968) ‘Empirical foundations for a theory of language change’. In W. P. Lehmann and Y. Malkeil (eds) Directions for Historical Linguistics: a Symposium (Austin: University of Texas Press), pp. 95–188. Part VIII In Honour of Ruqaiya Hasan 17 In Honour of Ruqaiya Hasan: Notes from the Contributors

Mariana Achugar Carnegie Mellon University, USA

The work of Ruqaiya Hasan has been influential in my understanding of social positioning, ideology and socialization practices in language development. Her explicit engagement with social theory from a linguistic background has informed my work on language development and cultural reproduction. Her work has provided a theoretical and analytic framework to explore issues of social justice and language socialization by connecting the unequal distribution of knowledge with language variation and coding orientations. Hasan’s studies of home literacy practices helped me understand how situations that may be considered the same are not necessarily the same for everyone in terms of their meaning. The idea that meaning is refracted through one’s social location has been key in understanding why it does not make sense to only teach the ‘genres of power’. In my work designing educational interventions to contribute to students’ development of disciplinary literacy, I have been particularly inspired by her proposal for reflection literacy. This approach to education foregrounds the importance of not only expanding students’ meaning-making resources, but also of questioning and reflecting about how those meaning-making choices connect to particular social positionings and power differences in society. Meaning-making practices are ideologically driven and as such our exploration of these practices and the design of educational experiences to socialize newcomers into them need to be critically analysed, not simply reproduced. This questioning and challenge of established meaning-making practices opens up the possibility of transforming and constructing knowledge, allowing students to have a more active role in their educational experiences.

Tom Bartlett Cardiff University, Wales, UK

I ﬁrst met Ruqaiya in 2001, in Ottawa. I was still working on my PhD and had decided to attend not only the International Congress, but also

415 416 Society in Language, Language in Society the Pre-Congress Institute, at which Ruqaiya was presenting a workshop on semantic networks. The workshop was also attended by Inger Lassen and Alison Moore, both now key ﬁgures in SFL. As usual, the fortnight was a heady mixture of work and pleasure (though I remember my systems networking rather more clearly than the details of the social events). At that time, attending only my second international event, I thought I was more interested in grammar than in context, but in the years that followed I became increasingly aware of the importance of Ruqaiya’s work and continue to learn from her extensive writings, which never fail to leave me feeling enriched through their subtlety of argument and their relentless attention to detail. I still feel a novice in reading Ruqaiya, however, as the depth and signiﬁcance of her writings become clearer the more you read and the elegant architecture of her collected works takes shape. Tout se tient is as true for Ruqaiya’s work as for any system, and with each new piece the valeur of each of her other works becomes clearer. Above all it is Ruqaiya’s call for SFL to play a leading part in an integrated sociolinguistics that has inspired me to study language as it is situated in social life without losing sight of the integral unity between language as system and language as practice; not only asking why it is used thus, here and now, but how it is shaped through and for those uses. My chapter in this book is intended as a contribution to the continuing discussion of these questions that Ruqaiya has led over six decades. Ruqaiya has been very supportive of the SFL community in Cardiff and having both her and Michael Halliday at our inaugural Summer School in 2010 got the event off to a more successful start than I and the other organizers, Lise Fontaine and Gerard O’Grady, could ever have anticipated, with participants coming from far and wide, across several continents. And in their most recent visit to the UK, Ruqaiya and Michael kindly took a detour to talk to our postgraduate students and to take part in a question and answer session that drew an appreciative and somewhat star-struck audience (and that applies to the staff and visiting academics who attended as well). I remain star-struck and feel incredibly privileged to have been asked to contribute to this volume in honour of one of the most important and original linguists of the last half-century.

Margaret Berry University of Nottingham, UK (retired)

Ruqaiya is a friend of many years, and one with whom it is a pleasure to discuss linguistic matters. I have always admired her work and before I retired I made great use of it in my teaching, particularly her work on generic structure and on cohesion. Her articles on the nursery tale and on cohesive harmony were special favourites with my students and led to much interesting project work. I now regard her as a leader in developing SFL work In Honour of Ruqaiya Hasan: Notes from the Contributors 417 on context and on semantics, both of these areas being central to SFL aims of relating linguistic analysis to sociological, ideological and educational issues. It is a great pleasure to be able to contribute to this volume in her honour.

Wendy L. Bowcher Sun Yat-sen University, China

I first met Ruqaiya Hasan when I was an undergraduate student at Macquarie University. I was enrolled in her course ‘Language and the Child’. Later, after finishing my BA Dip Ed, I went on to do a BA honours degree, and I had the great fortune of having Professor Hasan as my thesis supervisor. She was a tough, no-nonsense supervisor and I believe she brought out the best in me. It was also around this time that I met Carmel Cloran who was studying at Macquarie University and who later worked closely with Ruqaiya, and David Butt, too, was teaching at Macquarie University during the time I was a student there. After I graduated from Macquarie University, I took up a position as a secondary-school teacher, and so I was not involved in academia for quite a few years. However, soon I completed an MA at Sydney University around the time that Michael Halliday was retiring, and it was interesting to see the book An Introduction to Functional Grammar and to listen to Michael’s lectures, because I had only seen notes from drafts of his book when I was a student of Ruqaiya’s during my BA honours year. My main reason for doing the MA was so that I could work as a language teacher overseas. But doing the MA sparked my interest again in academic research. Nevertheless, it was not until after I had moved to Japan in 1989 that I became more interested in academic research, and it was during that time that I crossed paths with Ruqaiya again. In 1993 I initiated the founding of the Japan Association of Systemic Functional Linguistics (JASFL), and in 1994, we invited Ruqaiya to deliver the plenary for our first major international annual conference. When undertaking doctoral research at Liverpool University (UK), I wanted to catch up on what Ruqaiya had written and so I reviewed the recently published collection of her work Ways of Saying: Ways of Mean- ing edited by Carmel Cloran, David Butt and Geoff Williams (1996) for the Australian Journal of Linguistics. This wonderful collection led me on a trail of many of her other published works, some which I remembered her writing when I was her student back in 1982 and having them typed by ‘Judy the typist’, who also typed my honours thesis – back in the days before computers (at least how we know them today)! I now had a renewed interest in her work which challenged and inspired my own thinking. When I was back in Australia for a couple of years from 2007, I had the pleasure of attending some small group meetings on ‘context’ at Ruqaiya’s home. Those meetings were attended by David Butt, Annabelle Lukin, 418 Society in Language, Language in Society

Rebekah Wegener, Carmel Cloran, and occasionally Marilyn Cross, with occasional visits by one or two others. These meetings ignited my interest in working on developing the concept of context of situation. Since living in China, I have had many opportunities to engage in lively discussions with Ruqaiya and with Michael Halliday, who have both visited China regularly. I have seen how generous they continue to be with young scholars and researchers – nurturing and advising them in worthwhile and interesting research pursuits. I consider Ruqaiya my academic mentor, and it has been a privilege to have known her and to have had her encouragement, support and critical eye on my research especially in the latter part of my career. Throughout the years she has demonstrated through her own work and through her engaging discussions with me, academic rigour and integrity. And I am very happy to have had the opportunity to co-edit this collection of essays in her honour.

David Butt Macquarie University, Australia

It was my privilege, and good fortune, to be supervised by Ruqaiya Hasan. Furthermore, I was a junior colleague to Ruqaiya at Macquarie, in a department founded and developed by the also remarkable Professor Arthur Delbridge. Together they created, with other staff like Colin Yallop and Theo van Leeuwen, an environment of great linguistic intensity and of liberal thought. My goal in taking up a scholarship for the study of linguistics was to understand and illustrate how symbolic behaviour guided (even directed) human experience, namely: it really counts, the way we talk about a ‘thing’! To achieve this goal, I knew literature would be the best ‘laboratory’ for me, and that an exacting, detailed structuralism would be required – a tool that might enable me to capture those evanescent ﬂutterings of verbal arrangements I had gleaned through textual analysis, but only ‘through a glass darkly’. Well, I certainly landed in the right place! Ruqaiya Hasan is a brilliant, energetic, even combative supervisor. That is, she will argue point by point, removing shadows, lacunae, or glib transitions from expository writing. And, in relation to the problem I wished to research, she was deeply engaged with literatures and languages, across cultures and eras. As most readers will know, Ruqaiya was also profoundly engaged with the writings of Whorf and with the implications of his seminal ideas. Then there were the conversations about Vygotsky and Luria, and the grounding of meaning in the materiality of sociocultural arrangements: the context of situation. Ruqaiya worked harder than anyone to make this concept ‘carry’ a consequential load in linguistic description. In Honour of Ruqaiya Hasan: Notes from the Contributors 419

Throughout this time, I had to set about and learn more about linguistics in its established subdisciplines, from semantics and text linguistics with Ruqaiya to phonetics and phonology with John Clark and Colin Yallop. Vis- itors and speakers to the department included Arthur Capell, Ken Pike, Charles Ferguson, Lesley Milroy, John Laver, Anna Wierzbicka, Adrienne Lehrer, Jay Lemke, Christian Matthiessen and Michael Halliday. Halliday’s organizing principle of ‘choice’, along with the networks he developed, provided the tool for capturing pathways through the semiotic potential that I needed to describe; but An Introduction to Functional Grammar provided the immediate means for displaying the motivated arrangements around which a poet or novelist might extend the meaning potential of a community without necessarily invoking any ‘poetic licence’, without any overt ﬂouting of linguistic conventions. Ruqaiya had worked through the complexities of analysing verbal art: it had been her way into linguistics when she started at Edinburgh Uni- versity. Her theory dovetailed with the insights that emerged from the Prague School, in particular Mukaˇrovský’s writings on the aesthetic function. In detailed analyses, Hasan produced tight empirical arguments for ‘consistency of foregrounding’, typically unconscious, in verbal art. These supported her claim there was a form of ‘double articulation’ in verbal art, best understood as the realization of a higher-order ‘thesis’ for which the whole text functioned as a metaphor. The remarkable result for me was that, through my time of sharing discussions with Ruqaiya, I came to produce the work I actually set out to do: she had found a way of carrying our discourse to where my aspiration was realized. Our contested meanings had meant I found a way to articulate my thesis. I could not have hoped for more.

Carmel Cloran University of Wollongong, Australia (retired)

As a result of cross-cultural/cross-linguistic teaching experiences in various educational settings, I became interested in the multifaceted role of language in the process of the transmission of knowledge and of culture. This interest led me to undertake studies in linguistics at Macquarie University, Sydney, in the latter part of the 1970s. In my second year there, I enrolled in an undergraduate subject called ‘Language and the Child’ taught by Ruqaiya Hasan. This subject was my introduction to (a) Ruqaiya’s breadth and depth of scholarship, and her insistence on the same academic rigour from her students that she herself displayed; and (b) the systemic functional model of language. In my fourth year I asked Ruqaiya to supervise my dissertation (I still look back in wonder at my temerity!) At this time Ruqaiya was preparing her application for the funding of her major research into the relation between language and social class in the 420 Society in Language, Language in Society

Australian setting. On the completion of my undergraduate degree Ruqaiya invited me to work with her on this project. Since the project clearly fell within the area which had motivated my original interest in linguistics, I happily accepted. Thus began a steep learning curve (certainly for me though possibly for both of us) into the ontogenesis, development, and socio-semantic variation in that variety of language use required for success in school. Halliday’s systemic functional model of language and Bernstein’s sociological theory of education informed this research. Ruqaiya’s familiarity with both theories enabled her to build the framework for their extension and practical application. Witnessing her rigorous operationalization of the realizational relationship between the strata of language was nothing short of an eye-opener for me, giving as it did the interpretive framework and evidence-based rationale for claims about what was going on in the conversations that comprised the project’s database. Coding the data using the semantic networks she developed, and discussing with her the problems encountered, leading to modiﬁcations of the network, occupied many lunch hours in the early years of the project. I owe a great debt of gratitude to Ruqaiya for her openness and generosity in inviting me to study, develop and publish on aspects of the data that interested me, and for her encouragement in these endeavours. I appreciate not only her mentoring, showing me how true scholarship is enacted in practice, but also, and most particularly, her lasting friendship.

Andy Fung Hong Kong Polytechnic University, China

Since my undergraduate study, I have always been aware that Ruqaiya Hasan is one of the leading figures in SFL. However, the impression at that time was still a bit abstract and brief. Perhaps, the most remarkable one is that her works appeared from time to time in the reference lists of journal papers! Several years later, I started my PhD study under the supervision of Christian Matthiessen and collaborated with the healthcare communication research team led by Diana Slade in the Hong Kong Polytechnic University. At that time, we were fortunate to have Geoff Williams as our academic advi- sor, offering profound insights in various research aspects. I am extremely grateful to Geoff Williams who introduced Hasan’s semantic networks to me, which eventually came to serve as the main analytical tool in my current study. However, what Geoff did was not merely to introduce her works, but to offer opportunities to learn about Ruqaiya herself. I still remember the first time I met Ruqaiya – at the 40th International Systemic Functional Congress held at Sun Yat-sen University. Being a young researcher, it was my first time to present my initial finding in an international conference. Geoff was there and reminded me to grab the chance to chat with her. And finally the chance came. Having introduced myself to her In Honour of Ruqaiya Hasan: Notes from the Contributors 421 briefly, Ruqaiya had already recognized who I was, because Geoff had talked about my study to her when they were in Australia! Ruqaiya was so nice and supportive, offering valuable comments concerning semantic networks after my conference presentation. Her kindness extended beyond this short exchange to a longer academic dialogue when she, together with M. A. K. Halliday, came to the City University of Hong Kong for an academic visit. Never had I imagined I could have an opportunity to have academic dialogue with a renowned scholar like her for the whole afternoon! My discussions with her have definitely answered a lot of questions that I have had when reading her works, and most importantly, her current thoughts on semantic networks and context networks. The contribution I have made in this volume would not have been possible without her valuable and constructive comments as well as her encouragement.

M. A. K. Halliday Emeritus Professor, University of Sydney, Australia

I have tried in these few paragraphs to select from Ruqaiya’s impressive body of scholarly enterprise the motifs that have stood out for me most prominently, looking back over the years that we have spent together. She is an altogether remarkable person, and she has cast her intellectual net so widely that no specific theme within her writings can be taken as capturing the ‘essence’ of her work. What is clear is that she has contributed in very many ways to the striving towards a general ‘functional’ theory of language – that is, one where explanation is given and understood in functional terms; specifically to the development of a ‘systemic’ functional theory – that is, one where the organizing concept is that of choice, the elaboration of the human body and brain’s potential for meaning. I have referred to some of her relevant publications in my chapter in the present volume. Since we share this point of focus, on a systemic functional theory, it will provide a thread that runs through these few observations, which can be seen as a kind of time-reversed narrative on aspects of Ruqaiya’s intellectual life. The driving force behind her work has always been a quest for understanding, in which she tried to make explicit what questions she was asking and, wherever possible, what she would accept as a valid and relevant answer. Her point of departure is that, whatever language does in our lives, or whatever we get our language to do for us, it does by making meaning; and that is what determines how, as linguists, we go about recording and studying it, as well as and how we define and limit the scope of our studies.

1. The meaning of a text depends on its relation to its context; so when we want to understand what the text is doing, we need to locate it in its context of situation and of culture. For Ruqaiya, this means that ‘context’ belongs firmly within the domain of a theory of language. The situation 422 Society in Language, Language in Society may be analysed in terms of dimensions of social practice (her ‘action, relation, contact’, or ARC) that are embodied in the contextual parameters of the discourse; thus the field, tenor and mode of the discourse, its ‘address’ in the lattice of functional varieties or ‘registers’, is validated from ‘outside’ the language in terms that relate it to the social processes of the culture. At the same time, the stratum of context may be modelled paradigmatically, by the same framework of systems, and system networks, as had been developed, and tested in the modelling of the lexicogrammar. Ruqaiya has played a major part in bringing Firth’s ‘exterior relations of language’ within the compass of a system-structure theory. 2. Context as a distinct stratum was central to Ruqaiya’s thinking about language also from another important consideration: it is what remains constant under semantic variation. Variation depends on a higher-stratum invariant, and in semantic variation different semantic patterns act out the same contextual configuration (Ruqaiya’s ‘CC’). Ruqaiya saw that where different semantic features and structures were selected in a regular fashion by different groups in a given sociocultural environment, this gave value to Bernstein’s ‘codes’, more fully characterized as ‘socio-linguistic coding orientation’. This had been largely misunderstood and even caricatured by those who found it ideologically unacceptable (or simply did not understand it); Ruqaiya’s studies in semantic variation provided Bernstein’s theory with an explanatory foundation in linguistics. 3. This insight was validated in spectacular fashion in Ruqaiya’s groundbreaking research on interaction between mothers and 3½- to 4-year-old children in the home. With the collaboration of her research associate Carmel Cloran, she analysed a substantial body of natural conversation in lexicogrammatical and semantic terms, and this showed that there was a statistically significant pattern in the semantic variation: the same situation occasioned consistently different meaning styles in the mother’s discourse (reflected also in the discourse of the child), both according to the sex of the child and according to the social class status of the household. The procedure that they followed – gathering of data (100 hours of natural conversation), theory-based description (linguistic analysis of a sample of 20,000 clauses, out of 60,000 transcribed), statistical processing (cluster analysis in terms of principal components), and explanation and interpretation (variation in data accounted for by independent social factors) – provides a model of serious scientific research into language, given additional value by its obvious relevance to real-life problems – in this case, its significance for learning theory and the study of cultural transmission, and thus for early childhood education. 4. The success of this research depended on Ruqaiya’s ability to undertake a semantic analysis that was distinct from, but at the same time explicitly In Honour of Ruqaiya Hasan: Notes from the Contributors 423 related to, the categories of the lexicogrammar; this meant identifying the domains in which significant semantic choices can be shown to operate – as one example, the options available to mothers in answering their children’s questions. It also depended on her ability to conceive, design and carry out a complex programme of research, formulating aims and methods that were compatible with the theoretical resources available to her, but also identifying and negotiating with potential collaborators among various social groups across the culture. Bernstein’s researches had been based on interviews, with questions such as ‘what would you say if ...?’; but faced with this kind of experimental probe into their meaning behaviour, people’s answers, though offered in honesty, are likely to be very different from what they will actually say when they encounter the situation in real life. Research into how people mean, naturally in everyday situations, is extremely difficult to organize and manage; the teamwork of Hasan and Cloran seems to have got it all just right. 5. This in turn depended on Ruqaiya’s ability to recognize and to elucidate the structure and texture of everyday natural language. She had worked intensively on the notion of text; and on the nature of any particular instance of a text, both its structure, in terms of a generalized structure potential (her ‘GSP’) of a register or family of registers, and its texture, in terms of the powerful concept of cohesive harmony. Cohesive harmony is the cumulative effect of all the various processes that create patterns of meaning which go coursing through the text, picking up, overlapping and dying out, giving a clear sense of ‘textuality’ (which we usually become aware of only when it is not there). Ruqaiya showed that this was not just a feature of highly valued, ‘literary’ texts, but rather was inherent in the nature of human discourse. 6. It was in making use of, and extending, our joint work on the lexicogrammatical resources of cohesion that Ruqaiya had set herself the task of exploring its essential foundations in semantics (this was the context in which she introduced the term ‘cohesive harmony’). This involved investigating the nature of the stratal relationship between the lexicogrammar and the semantics; and also the intra-stratal relationship between grammar and lexis. Her research into the concept of lexis as ‘most delicate grammar’ showed clearly that a lexical item could be modelled as a complex of features which are themselves organized in networks of choices; and that these in turn could be derived by steps in delicacy from very general system networks in the grammar, especially the ideational networks embodied in transitivity. This provided a further insight into the semantic activation (in her terms) of lexicogrammatical forms. It also explained the powerfully cohesive effect of lexical ‘chaining’ in discourse, where the underlying paradigmatic order is manifested in an observable syntagmatic process. 424 Society in Language, Language in Society

7. This sort of semantic patterning of formal, lexicogrammatical resources is recognizable as one of the mainstays of verbal art, the context in which, as a student and then a teacher of English literature, Ruqaiya had ﬁrst challenged herself to explain the creative powers of language. She had always rejected the notion that there was a special ‘language of literature’; and she always disliked what she called ‘verbal pyrotechnics’, where the writer puts the text on display as a literary artefact. What impressed her was rather the potential of everyday language, as something which could be seen at work in texts of every variety: in a highly valued masterpiece of poetry or prose, in a nursery tale or childhood rhyme, in a conversation with children of all ages in school, or in informal family interaction in the home – all of which she has investigated at some period in her working life. At the same time, however, Ruqaiya has always been aware of a special quality in the language of verbal art, recognizing in it a distinct dimension of meaning, beyond the symbolic articulation, which is where the listener, or reader, is enabled to construe an underlying, more general theme. The congruent meanings may be those of everyday discourse; but the ‘literary’ text is one in which these are further distilled to give an extra, metaphoric momentum to the language’s construction of reality. I have always been dismayed by the amount of knowledge that is lost as fashions of scholarship swerve from one vantage point to another, oscillat- ing around the centres of symbolic power. One strand running throughout Ruqaiya’s thinking has been her awareness of the ideological conditions within which scholarship in the human sciences is conducted – and, by the same token, of how these conditions are themselves brought about and strengthened by unformulated assumptions about the nature of reality (and of language). There is such a wealth of insight in her writings, including her close critical readings of the work of others, such as Saussure, Firth, Malinowski, Whorf, Vygotsky, Labov and Bourdieu, that I think they will go on being read, with increasing appreciation and acknowledgement, for a good time to come. Our thanks go to Jonathan Webster, and to her publishers at Equinox, for ensuring that her key contributions are brought together so they can be taken in as a whole, as they should be; and to the two editors of the present volume, Wendy Bowcher and Jennifer Liang, for giving space to us, her colleagues, to signify our regard.

Kristin M. Khoo Macquarie University, Australia

The work of Ruqaiya Hasan has greatly influenced my understanding of a social-semiotic theory and development as a linguistic researcher. As a first- year undergraduate at Macquarie University, I was introduced to the work of Hasan in the course, ‘Language, Culture and Communication’. And since In Honour of Ruqaiya Hasan: Notes from the Contributors 425 this time, I have had the opportunity, not only to read (and reread!) her work, but to listen to her speak on many occasions. My introduction to Ruqaiya’s work and continued efforts to read and wres- tle with the concepts and methodology have also been much in part, due to my teachers – David Butt, a student of Ruqaiya’s himself – and more recently, Annabelle Lukin. It is my teachers who have encouraged my continued learning and questioning and they, in turn, have been shaped and challenged by Ruqaiya Hasan. It is also through the continued discussion of Ruqaiya’s ideas with colleagues and peers, and in exploring these ideas with students myself, that I have the opportunity to engage with both elements of theory and the transition to research practice. As the research focus for my PhD involves the use of cohesive harmony in therapeutic conversation, a method developed by Hasan, her work on cohesive harmony and texture has been of particular significance. A copy of Language, Context and Text rarely leaves my desk. Ruqaiya has also been generous in the use of an unpublished manuscript, for which I am very grateful. I have found cohesive harmony analysis not only to be a valuable source of evidence in the pursuit of a specific research area, but it has also led me to further conceptualize linguistic theory, driven me to further research questions, and opened up opportunities for future collaborations. It is a privilege to contribute to this volume in her honour.

Marvin Lam Hong Kong Polytechnic University, China

My chapter in this volume outlines Professor Hasan’s vast influence on my work in modelling and measuring personal distance, the inspiration of which was Professor Hasan’s discussion on biography of interaction and personal distance in Halliday and Hasan (1985). My work on context of situation and language arts also follows Professor Hasan’s work, particularly her work on verbal art. Apart from having an ideational impact, I have been very much privileged to learn from her interpersonally. I first met Professor Hasan when I was working on my PhD project under the supervision of Professor Jonathan Webster at the City University of Hong Kong. With Professor Webster’s guidance, I was very lucky to know many world-renowned scholars in person instead of as names printed on articles and books, and among them was Professor Hasan. She visited Hong Kong to work on projects with the Halliday Centre for Intelligent Applications of Language Studies at CityU. By then my work had already been much influenced by her, and my first international conference paper, presented at the 33rd International Systemic Functional Congress held at the Catholic University of São Paulo, already referred closely to her work. I took the opportunity to show her my paper 426 Society in Language, Language in Society and asked for comments. After a few days, she returned the paper to me with detailed comments on every page of the paper and two pages of additional comments (I still keep the paper to remind myself of the occasion), and we had a very serious discussion. She was concerned not only with the content of the paper, but also my academic writing and thinking skills. She then helped organize my three-month visit to Sydney to work with Profes- sor David Butt at Macquarie University. Sydney was an academic paradise for me. I got to know friends and colleagues who are very important to me, and Professor Hasan and I met weekly to work on my PhD project. By the time I returned to Hong Kong, I felt like a different person. The most important lesson I have learned from Professor Hasan is to be as critical of oneself as of others, which is what I have observed as her practice over the years in different contexts. This is hugely influential on me for my every intellectual pursuit, and I would like to take this opportunity to express my sincere thanks to her for that.

Jennifer Yameng Liang University of Science and Technology Beijing, China

The first time I had the chance to meet Professor Hasan was when I was an undergraduate student at Sun Yat-sen University. One day, we were told that Professor Halliday and Professor Hasan would come to speak at a symposium. Even though we had to look up the word ‘symposium’ in the dictionary to know what it was like, we were still thrilled at the thought of meeting the two world-renowned linguists in person. The auditorium that day was packed with faculty and students. Although I could only understand a tiny bit of their talk, I was mesmerized by what linguistics offers to the understanding of the world and what Professor Halliday and Professor Hasan offer to the development of linguistics. Later, I decided to continue my study at Sun Yat-sen University, which turned out to be one of the best decisions I have ever made, for over the years, I got to learn from Professor Hasan more and more in person, not just from her lectures, but also face to face. The biggest bonus we had as PhD students at Sun Yat-sen is that every time they visited, Professor Halliday and Professor Hasan would kindly arrange a one-on-one discussion with us on our doctoral research. With each discussion Professor Hasan would always enlighten me one way or another. In one discussion, she drew upon the philosophical distinction between ‘sensible’ and ‘intelligible’ as an analogy for the different nature of non-verbal and verbal semiotic modes. This idea is rather enlightening in providing a rational argument for the discussion of the front and back side of tourist site entry tickets which were examined in my PhD thesis. Mean- while, on several occasions, she has also expressed her deep concerns about the possible limitations of linguistic frameworks applied to the study of other In Honour of Ruqaiya Hasan: Notes from the Contributors 427 non-verbal semiotic modes. Being aware of that and under the supervision of Professor Wendy Bowcher, I started to investigate more meaning-driven questions, such as the generic structural potential of multimodal texts, how meanings of multimodal texts are motivated by a particular context of situation, if the context system network can be modified to fit with the multimodal texts, etc. Looking back on those discussions, I appreciate so much the precious guidance from Professor Hasan. In Chinese, a person with a doctoral degree is called (boshi), which literally means a scholar with extensive knowledge. Nowadays, with some of the doctors only focusing on their particular area of research, they tend to be nicknamed (zhuanshi), which means a scholar with specialized knowledge. I am not raising here a discussion of which one of the two we should be, but I strongly believe that Professor Hasan is definitely one of the few scholars who deserve both the name of and of .

Annabelle Lukin Macquarie University, Australia

My introduction to Ruqaiya Hasan’s work came through my PhD supervisor, Professor Geoff Williams, who had been Ruqaiya’s PhD student. I was capti- vated by Geoff’s research from the first conference presentation I heard, on how teaching about grammar as a meaning-making resource could revolu- tionize the primary-school classroom. Later, as a Master’s student at Sydney University, I came to know Geoff’s PhD research, which Ruqaiya had supervised in the late 1980s and early 1990s, on the language used by caregivers and children during bedtime story reading and how patterns of interaction varied depending on the family’s socio-economic location. Geoff bravely gave his Master’s students long, and sometimes difficult, articles by Ruqaiya on the relations of text, language, context and society. Some complained. I felt a kind of intellectual stimulation I had been search- ing for through much of my twenties. It worked to my advantage that I was not easily put off by reading ideas I did not fully understand. To read linguistic theory that made meaning in society the central ‘problematic’, and was both abstract in its conceptions but grounded in the ‘typical actual’, was totally absorbing. Now that I have read Saussure, Malinowski, Firth, Vygotsky and Bernstein, I understand why I found Ruqaiya’s research both brilliantly insightful, and deeply practical. Although I began by reading her discourse on context, and on semantic variation, I became enamoured of her linguistic analysis of verbal art. From my mother I had understood the power of literature to show the many realities of human experience. Ruqaiya’s work finally gave me critical tools worthy of the literary texts themselves. I combined her theoretical work on verbal art, with my own experience and interest in pedagogic contexts, and wrote my PhD under Geoff’s supervision on the teaching and assessment 428 Society in Language, Language in Society of literary text (in particular, the poetry of one of Australia’s finest contemporary poets, Bruce Dawe) in NSW high schools. In this process I brought together her work on the conception of context in text, her method for the description of texture, her semantic network descriptions, her description of the generic structure potential for reasoning in natural discourse, and of course her model of the two orders of semiosis in verbal art. I also then read the various other systemic functional studies of literature, including by Halliday, Butt and O’Toole. When I began to work on media and political discourse, I naturally added to the mix Hasan’s work on ideology, which, despite being set out across a small handful of book chapters, is the most coherent linguistic account of the relations of language and ideology that our discipline has offered. I have also increasingly drawn on her writings on lexis as delicate grammar – on what she has come to call ‘microgrammatics’. I have also found Rhetorical Unit analysis, developed by Carmel Cloran in PhD research also supervised by Ruqaiya, enormously clarifying in my studies of news discourse. I hope this volume will bring more readers to understand how Hasan’s contributions to the Hallidayan tradition not only extend it, but explain more fully its enormous theoretical and descriptive power.

Donna R. Miller University of Bologna, Italy

I have ‘known’ Ruqaiya Hasan for 25 years, and known of her for at least another 10, having first come upon her name in the late 1970s, through her and Halliday’s 1976 ‘best seller’, even in Italian academia, Cohesion in English. At the time I was teaching EFL in Rome and aiming to enhance the all in all unsatisfactory communicative teaching materials I was adopting with explicit teaching of language awareness. This text led me firstly into further reading of Halliday’s work; Halliday led me to Bernstein, who led me to Hasan again, and the attraction intensified and continued. My academic spe- cialization had been literature and ‘lit crit’, but I felt like an unenlightened novice when I first discovered Hasan’s work on the language of literature, admittedly not until the 1989 OUP edition of Linguistics, Language, and Ver- bal Art: a revelation indeed, and one I enthusiastically embraced. In the spring of that same year Hasan and Halliday came to Bologna, where I was by then teaching. I religiously attended all they offered, including Hasan’s three-day seminar on ‘Ideology, Social Class and Semantic Change’ and lecture entitled ‘You know it as learning to be reasonable; I know it as learning to reason: The language of rationality in mother–child talk’ – an early testi- mony to her ongoing research into semantic variation. My admiration was clinched; I was at her feet. In the autumn of 1996, Hasan was again in Italy. Happily she agreed to give a conference ‘On Reasonable Reasons: Some Semantic Orientations in In Honour of Ruqaiya Hasan: Notes from the Contributors 429

Mother–Child Dialogues’ at the University of Trieste, where I was doing my ﬁnal semester before transferring back to Bologna. At about this time email came in, and we exploited the medium to stay in touch. As our exchanges grew, so did my delight in Ruqaiya’s keen wit and feisty spirit, but also in what I came to experience as her candid human warmth. I wanted and was lucky to get Hasan to come once more to Bologna as keynote speaker at the international symposium I organized in 2004: ‘Towards a Linguistic Approach to Verbal Art: Theory and Practice’. She spoke on the theme of ‘Private pleasure, public discourse: Multiple modalities in the reading of literature’. Her excellent paper is included in the 2007 Equinox volume edited by Miller and Turci. In 2010, I was honoured by Hasan’s invitation to collaborate with her in a seminar on ‘Semiotic Stylistics and Verbal Art’ for the LinC Summer School and Workshop at Cardiff University which focused on just how the deepest level of meaning in verbal art is realized in the patterns of the literature text’s worded meanings. The seminar went extremely well and from then on our contact has been intense, most recently over her important contribution to the volume: D. R. Miller and P. Bayley (eds) Hybridity in Systemic Functional Linguistics: Grammar, Text and Discursive Context (London: Equinox). I am mindful of how privileged I have been to know Ruqaiya Hasan and how impoverished my academic and personal life would have been without the experience. For this enduring bond – but also for having had this wonderful opportunity to celebrate her, I am gratiﬁed, and very grateful.

Alison Rotha Moore University of Wollongong, Australia

My connection with Ruqaiya Hasan began when I had the great privilege of taking some of her undergraduate subjects at Macquarie University in the 1980s. Ruqaiya’s research on semantic variation and children’s socialization was in full swing. I recall the energy and excitement she brought to our sociolinguistics lecture one afternoon after making a breakthrough on how to code familial social class as an ‘independent variable’ in her quantitative study. The decision to use the level of autonomy in the household breadwinner’s occupation had been informed – if I recall correctly – by discussions with R. W. Connell, Macquarie’s founding professor of sociology. Using level of autonomy made sense because it addressed the issue of how social location might inﬂuence a person’s semantic orientation, and Ruqaiya was able to weave this into the set topic for class that day (a topic which itself has faded from memory). The issue of coding for social class remains a complex one, but Ruqaiya’s solution from the 1980s was already much more nuanced than the ‘occupational prestige scales’ that are the industry standard in other disciplines. 430 Society in Language, Language in Society

In that sociolinguistics class I had a glimpse of many of Ruqaiya’s particular strengths as a researcher and scholar. What I wish to draw attention to here is Ruqaiya’s power to connect her work to big issues outside linguistics, without colonizing those other disciplines, on the one hand, and without being in thrall to them, on the other. Being in Ruqaiya’s classes was like being let into a big secret – that linguistics was extraordinarily relevant, not just intellectually stimulating. I must confess I did not really ‘get’ the full scope and power of Ruqaiya’s approach at the time, and I went on to do honours in developmental phonology with another inspirational teacher, Barbara Dodd, who taught me other ways of doing research involving children’s consciousness, ingenuity and life opportunities. After that, while studying and working in public health, I began to see how useful it would be to bring in tools and concepts like those I had learnt from Ruqaiya. And so I enrolled in a PhD in linguistics. I have now studied both highly personalized contexts of care such as HIV medicine, psychotherapy and palliative care. I have studied contexts in which the patient is not present as an interactant, such as surgical teamwork, and contexts with mass audiences, such as the illegal but fairly common practice of direct-to-consumer advertising of prescription drugs. I have also begun to study the textual practices that are normatively involved in keeping non-human animals out of the category of social subjects in public discourses, and some of the ways that activists attempt to deautomatize such practices. In each of these projects, Ruqaiya’s work has been invalu- able, particularly her work on modelling context, her semantic networks, and her analysis of cohesive harmony. There is still much to read, reread and interrogate, but it is an honour and a pleasure to present some of my own engagement with her ideas in this volume.

Mary Schleppegrell University of Michigan, USA

Ruqaiya Hasan’s work has been an inspiration to me throughout my academic career. She has offered ways of thinking about the role of everyday talk and its profound implications in shaping students’ consciousness and identities that were important in my conceptualization of the language of schooling and the relationship between dialect and register. Her foregrounding of the interpersonal in education has consistently highlighted the need to consider who the students are and their prior socialization contexts as we move toward pedagogies that draw on what students bring to learning. She has demonstrated how language shapes the development of knowledge both in the individual and in society, urging teachers and researchers to continually ask why things are as they are and explore other ways they could be. In my work in the contexts of history pedagogy and early literacy development, we have attempted to engage both teachers and students through exploration of In Honour of Ruqaiya Hasan: Notes from the Contributors 431 interpersonal meaning, helping them recognize the interpretation and perspectives of authors in the language choices that shape the way knowledge is presented. In particular, Hasan’s view of reﬂection literacy has underpinned the work I have done to develop tools and processes based in SFL theory that engage students in a literacy pedagogy that values their perspectives. Such pedagogy is manifested in talk about text that enables students to share their perspectives and make varied interpretations visible in the classroom and a focus of discussion.

Rebekah Wegener Aachen University, Germany

It was Hasan’s work with context and in particular her analysis of mother– child interaction that ﬁrst captured my attention as a linguist. I was fas- cinated by her contextual description of the discussions between mothers and their children and the potential to link these to the work I was examining in using knowledge of discourse to improve outcomes for children at risk of abuse. Hasan brought the social context into contact with the highly detailed analysis of the linguistic world. Several years later I was able to discuss this work with Hasan at length at the context discussion groups that she organized. More than anything I appreciate her willingness to engage fully with the work of new researchers and take the time to talk with students about their work. I know that I am not the only one to have beneﬁted from her guidance in arguing about theoretical points. The energy and depth of her engagement with the ideas and arguments of others show the strength of her intellectual integrity.

Geoff Williams Emeritus Professor, University of British Columbia, Canada

Just at the time Ruqaiya Hasan and Michael Halliday arrived in Australia in 1976 I was completing an MA (honours) thesis about primary-school children’s peer conversations. I was very puzzled about how to analyse these conversations. Having transcribed the audio recordings, I knew something very interesting was going on in the language, but the only analytic approach I knew about was content analysis. I had no better way to look at the data and was very frustrated. Fortunately for me, Hasan heard me talking about this research a short while after the MA was completed and invited me to lunch. I expected she wanted to talk more about the findings. Though she was polite about them, she was not very impressed with the analysis. Before long she turned the conversation to the possibility of my doing a PhD with her. I was both grateful and intrigued by this idea. Having no idea what writing a doctoral thesis with her would involve, I enrolled a year later. This 432 Society in Language, Language in Society was just after the first edition of An Introduction to Functional Grammar was published, and in my first course with her, that was the textbook. Simply put, there was my intellectual life before that lunch with Hasan and my intellectual life after the lunch with Hasan. It was she who first suggested I read de Saussure, Whorf and, of course, most of all Halliday, especially Learning How to Mean and Language as Social Semiotic alongside ‘the grammar’. That was in the first week of the PhD candidature (or so it seemed). Vygotsky and Luria followed soon after. Then, after some months of my thrashing around with the topic I had first proposed, she suggested I read Bernstein and consider writing my PhD on semantic variation. At about the same time I read her paper ‘Ways of Saying: Ways of Meaning’, and soon after, her book in the Deakin University Press series, Linguistics, Language and Verbal Art, which introduced her thinking about symbolic articulation and theme. That is the experiential account. It has to be supplemented by the interpersonal. Most of all, there has been her deep and genuine interest in my development as a scholar, which continues. There was no praise, but there was a strong, obviously genuine sense of collaboration toward a shared vision of the PhD thesis. Alongside this there was her personal interest in all members of my family, and an introduction to members of her family, especially her sister Zakia in Karachi. By that time I knew Ruqaiya well, but being with her close family in Karachi gave me a much deeper understanding of the cultural resources on which she draws so productively. So now it is the turn of the textual metafunction. There is much to say, so I highlight just the one dominant feature, coherence. Looking back on those early PhD days, I realize now that what I encountered, and what was so attractive, was a coherent way of understanding language in social life. I know of no adequate way to thank her for these continuing intellectual gifts. A Bibliography of Work by Ruqaiya Hasan

This section sets out a comprehensive and up-to-date list of Ruqaiya Hasan’s publications throughout the course of her career.

Books

1. ‘Cohesion in written and spoken English’. Programme in English Language Teach- ing: Paper No. 7. Longman, Greens & Co. 1968. 2. Cohesion in English (with M. A. K. Halliday). Longman, 1976. 3. Language, Context and Text (with M. A. K. Halliday). Oxford University Press (ﬁrst published by Deakin University Press). 1985. 4. Linguistics Language and Verbal Art. Oxford University Press (ﬁrst published by Deakin University Press). 1985. 5. Discourse on Discourse (ed.). Applied Linguistics Association of Australia, Special publication no. 7, 1985. 6. ‘A semantic network for the analysis of messages in everyday talk between mothers and children’. (Mimeo.) 7. Language Development: Learning Language Learning Culture (co-edited with J. R. Martin). Norwood, NJ: Ablex. 1989. 8. On Subject and Theme: a Discourse Functional Perspective (co-edited with Peter H. Fries). Amsterdam: John Benjamins. 1995. 9. Literacy in Society (Applied Linguistics and Language Studies Series) (co-edited with Geoff Williams). London: Longman. 1996. 10. Functional Descriptions: Theory in Practice (co-edited with Carmel B. Cloran and David G. Butt). Amsterdam: John Benjamins. 1996. 11. Ways of Saying: Ways of Meaning: Selected Papers of Ruqaiya Hasan, edited by Carmel Cloran, David Butt and Geoff Williams. London: Cassell. 1996. 12. Language, Society and Consciousness: the Collected Works of Ruqaiya Hasan,Vol.1 (edited by Jonathan J. Webster). London: Equinox. 2005 (pp. 3–370). 13. Continuing Discourse on Language: Functional Perspectives, Vol. 1 (co-edited with Christian Matthiessen and Jonathan J. Webster). London: Equinox. 2005. 14. Continuing Discourse on Language: Functional Perspectives, Vol. 2 (co-edited with Christian Matthiessen and Jonathan J. Webster). London: Equinox. 2007. 15. Semantic Variation: Meaning in Society and in Sociolingistics: the Collected Works of Ruqaiya Hasan, Vol. 2, edited by Jonathan J Webster. London: Equinox. 2009 (5–478; with CD). 16. Language and Education: Learning and Teaching in Society: the Collected Works of Ruqaiya Hasan, Vol. 3. Edited by Jonathan J Webster. London: Equinox. 2011 (5–408). 17. Selected Works of Ruqaiya Hasan on Applied Linguistics (in Series of World Famous Applied Linguists). Beijing: Foreign Languages Teaching and Research Press. 2011 (3–461).

433 434 A Bibliography of Work by Ruqaiya Hasan

Articles

1966 ‘On the subject matter of English’ (co-authored with S. Lushington). Pro- gramme in English Language Teaching: Paper No. 2. Longman, Greens & Co. 1967 ‘Linguistics and the study of literary texts’. Etudes de Linguistique Appliqué, 5. Paris, Didier. 1970 ‘The verb “Be” in Urdu’. In The Verb ‘Be’ and Its Synonyms, edited by J. W. M. Verhaar. Foundations of Language Supplementary Series 7. 1971 ‘Rime and reason in literature’. In Literary Style: a Symposium, edited by S. Chatman. New York: Oxford University Press. 1971 ‘Syntax and semantics’. In Biological and Social Factors in Psycholinguistics, edited by J. Morton. London: Logos. 1973 ‘Code, register, and social dialect’. In Class, Codes and Control, Vol. II. (ed.) Basil Bernstein. London: Routledge & Kegan Paul (pp. 253–92). 1975 ‘The place of stylistics in the study of verbal art’. In Style and Text: Studies Presented to Nils Erik Enkvist, edited by H. Ringbom. Stockholm: Skriptor. 1976 ‘Socialization and cross-cultural education’. In Language and Education in the Third World, edited by J. Berry. International Journal of the Sociology of Language 8, Special Issue. The Hague: Mouton. 1978 ‘Some sociological considerations in second language teaching’. In Language Learning in Australian Society: Proceedings of the 1976 Congress of the Applied Linguistics Association of Australia, edited by D. E. Ingram and T. J. Quinn. Melbourne: Australia International Press (pp. 228–46). 1978 ‘Text in the systemic functional model’. In Current Trends in Text Linguistics, edited by W. U. Dressler. Berlin: Walter de Gruyter. 1979 ‘On the notion of text’. In Text vs Sentence: Basic Questions of Textlinguistics, Vol. 2, edited by J. S. Petoﬁ. Amsterdam: Helmut Buske (pp. 369–90). 1981 ‘What’s going on?: a dynamic view of context in language’. In The Seventh LACUS Forum 1980, edited by J. E. Copeland and P. W. Davis. Columbia: Hornbeam Press (pp. 106–21). 1984 ‘The nursery tale as a genre’. In Nottingham Linguistic Circular, 13. Nottingham University Press (pp. 71–102). 1984 ‘The structure of the nursery tale’. In Linguistica Testuale: Proceedings of the 15th International Congress of the Italian Linguistic Society, edited by L. Coveri. Rome: Bulzoni. 1984 ‘Coherence and cohesive harmony’. In Understanding Reading Comprehension, edited by J. Flood. Delaware: International Reading Association (pp. 181–219). 1984 ‘What kind of resource is language’? Australian Review of Applied Linguistics 7 (1): 57–85. 1984 ‘Ways of saying: ways of meaning’. In Semiotics of Language and Culture,Vol.1, edited by R. P. Fawcett, M. A. K. Halliday, S. M. Lamb and A. Makkai. London: Pinter. 1985 ‘Lending and borrowing: from grammar to lexis’. In The Cultivated Australian: Festschrift in Honour of Arthur Delbridge, edited by J. E. Clark. Amsterdam: Helmut Buske. 1985 ‘Meaning, text and context: ﬁfty years after Malinowski’. In Systemic Perspec- tives on Discourse, Vol. 1, edited by J. D. Benson and W. S. Greaves. New Jersey: Ablex. 1986 ‘The ontogenesis of ideology: an interpretation of mother child talk’. In Semiotics, Ideology Language, edited by T. Threadgold, E. Grosz, G. Kress and M. A. K. Halliday. Sydney University, Sydney Studies in Society and Culture. A Bibliography of Work by Ruqaiya Hasan 435

1986 ‘The implication of semantic distance for language in education’. In Studies in Bilingualism edited by A. Abbi. New Delhi: Bahri Publications. 1987 ‘Reading picture reading: invisible instruction at home and in school’. In The Proceedings of the 13th Annual Conference of the Australian Reading Association. July 1987, Sydney (pp. 33–82). 1987 ‘The grammarian’s dream: lexis as most delicate grammar’. In New Develop- ments in Systemic Linguistics, Vol. 1, edited by M. A. K. Halliday and R. P. Fawcett. London: Frances Pinter (pp. 184–211). 1987 ‘The analysis of one poem’. In Functions of Style, edited by D. Birch and M. O’Toole. London: Frances Pinter. 1987 ‘Directions from structuralism’. In The Linguistics of Writing, edited by D. Attridge, A. Durant, N. Fabb and C. McCabe. Manchester: Manchester University Press (pp. 101–21). 1988 ‘Language in the processes of socialisation: home and school’. In Language and Socialisation: Home and School: Proceedings from the Working Conference on Language in Education, edited by L. Gerot, J. Oldenburg and T. Van Leeuwen. North Ryde: Macquarie University (pp. 17–21). 1989 ‘Semantic variation and sociolinguistics’. Australian Journal of Linguistics 9(2): 221–75. 1990 ‘A sociolinguistic interpretation of everyday talk between mothers and children’. In Learning, Keeping and Using Language: Selected Papers from the 8th World Congress of Applied Linguistics, Vol. 1 edited by M. A. K. Halliday, J. Gibbons and H. Nicholas. Amsterdam: John Benjamins (co-authored with C. Cloran) (pp. 67–100). 1991 ‘Questions as a mode of learning in everyday talk’. In Language Education: Interaction and Development: Proceedings of the International Conference, Vietnam 30 March–1 April 1991, edited by T. Le and M. McCausland. Launceston: University of Tasmania (pp. 70–119). 1992 ‘Rationality in everyday talk: from process to system’. In Corpus Linguistics: Proceedings of the Nobel Symposium 82, Stockholm 4–8 August 1991, edited by J. Svartvik. Berlin: Mouton de Gruyter (pp. 257–307). 1992 ‘Meaning in sociolinguistic theory’. In Sociolinguistics Today: International Per- spectives, edited by K. Bolton and H. Kwok. London: Routledge (pp. 80–119). 1992 ‘Speech genre, semiotic mediation, and the development of higher mental functions’. Language Sciences, 14 (4): 489–528. 1993 ‘Contexts for meaning’. In Georgetown Round Table on Language, Commu- nication, and Social Meaning 1992, edited by J. E. Alatis. Washington, DC: Georgetown University Press, pp. 79–103. 1994 ‘Situation and the deﬁnition of genre’. In What’s Going On Here: Complemen- tary Studies of Professional Talk, edited by A. D. Grimshaw. Norwood, NJ: Ablex (pp. 27–172). 1995 ‘Learning to function with the other tongue: a systemic functional perspective on second language teaching’. In Perspectives on Pedagogical Grammars, edited by T. Odlin. New York: Cambridge University Press (pp. 179–226). 1995 ‘On social conditions for semiotic mediation: the genesis of mind in society’. In Knowledge and Pedagogy: the Sociology of Basil Bernstein, edited by A. R. Sadovnik. Norwood, NJ: Ablex (pp. 171–96). 1995 ‘The conception of context in text’. In Discourse in Society: Systemic Func- tional Perspectives, edited by P. H. Fries and M. Gregory. Norwood, NJ: Ablex (pp. 183–283). 1995 ‘Reﬂections on subject and theme: an introduction (with P. H. Fries)’. In On Subject and Theme: a Discourse Functional Perspective (ed. with P. H. Fries). Amsterdam: John Benjamins. 436 A Bibliography of Work by Ruqaiya Hasan

1995 ‘English process, English tense: foreign learner, foreign teacher’. SPELT Newsletter Vol. X, No. 4, December: 2–23. 1996 ‘On teaching literature across cultural distances’. In The Language–Culture Connection (Anthology Series 37: RELC, pp. 440–96), edited by J. E. James. Singapore: SEAMEO Regional Language Centre. 1996 ‘Literacy, everyday talk and society’. In Literacy in Society, edited by R. Hasan and G. Williams. London: Longman (pp. 377–424). 1996 ‘Semantic networks: a tool for the analysis of meaning’. In Ways of Saying: Ways of Meaning: Selected Papers of Ruqaiya Hasan, edited by C. Cloran, D. Butt and G. Williams. London: Cassell (pp. 104–31). 1998 ‘Educating the language teacher: a social-semiotic approach’. In Teaching Language and Culture: Building Hong Kong on Education, edited by B. Asker. Hong Kong: Addison Wesley Longman China Ltd. 1999 ‘The disempowerment game: Bourdieu and language in literacy’. Linguistics and Education 10 (1): 25–88. 1999 ‘Society, language and the mind: the meta-dialogism of Basil Bernstein’s theory’. In Pedagogy and the Shaping of Consciousness: Linguistic and Social Processes, edited by F. Christie. London: Cassell (pp. 10–30). 1999 ‘Speaking with reference to context’. In Text and Context in Functional Linguistics, edited by M. Ghadessy. Amsterdam: John Benjamins (pp. 219–328). 2000 ‘Bourdieu on linguistics and language: a response to my commentators’. Linguistics and Education 10 (4): 441–58. 2000 ‘The uses of talk’. In Discourse and Social Life, edited by S. Sarangi and M. Coulthard. London: Longman (pp. 28–47). 2000 ‘Basil Bernstein 1924–2000: Obituary’. Functions of Language 7 (2): 279–91. 2001 ‘Understanding talk: directions from Bernstein’s sociology’. International Jour- nal of Social Research Methodology: Theory & Practice 4(1):5–9. 2001 ‘The ontogenesis of de-contextualized language: some achievements of clas- siﬁcation and framing’. In Towards a Sociology of Pedagogy: the Contribution of Basil Bernstein to Research, edited by A. Morais, I. Neves, B. Davies and H. Daniels. New York: Peter Lang. 2001 ‘Wherefore context?: the place of context in the system and process of language’. In Grammar and Discourse Analysis, edited by R. Shaozeng. Macau: The Publication Centre, University of Macau (pp. 1–33). 2002 ‘Semiotic mediation and mental development in pluralistic societies: some implications for tomorrows schooling’. In Learning for Life in the 21st Century: Socio-Cultural Perspectives on the Future of Education, edited by G. Wells and G. Claxton. Oxford: Blackwell (pp. 112–26). 2002 ‘Ways of meaning, ways of learning: code as an explanatory concept’. British Journal of the Sociology of Education 23 (4): 537–48. 2003 ‘Globalization, literacy and ideology’. World Englishes 22 (4): 433–48. 2004 ‘Reading picture reading: a study in ideology and inference’. In Language, Education and Discourse: Functional Approaches, edited by J. Foley. London: Continuum (pp. 43–75). 2004 ‘Analysing discursive variation’. In Systemic Functional Linguistics and Critical Discourse Analysis, edited by L. Young and C. Harrison. London: Continuum (pp. 15–52). 2004 ‘The concept of semiotic mediation: perspectives from Bernstein’s sociology’. In Researching with Bernstein, Working with Bernstein, edited by J. Muller and A. Morais. London: Routledge (pp. 80–92). A Bibliography of Work by Ruqaiya Hasan 437

2004 ‘The world in words: semiotic mediation, tenor and ideology’. In Language Development: Functional Perspectives in Evolution and Ontogenesis, edited by G. Williams and A. Lukin. London: Continuum (pp. 158–81). 2005 ‘Introduction: a working model of language’. In Continuing Discourse on Lan- guage, Vol. 1, edited by R. Hasan, C. Matthiessen and J. J. Webster. London: Equinox (pp. 37–54). 2005 ‘Language and society in a systemic functional perspective’. In Continuing Dis- course on Language, Vol. 1, edited by R. Hasan, C. Matthiessen and J. J Webster. London: Equinox (pp. 55–80). 2005 ‘Language, society and consciousness: transdisciplinary orientation and the tradition of specialisation’. In Language, Society and Consciousness: the Col- lected Works of Ruqaiya Hasan, Vol. 1, edited by J. J. Webster. London: Equinox (pp. 3–17). 2005 ‘Semiotic mediation and three exotropic theories: Vygotsky, Halliday and Bernstein’. In Language, Society and Consciousness: the Collected Works of Ruqaiya Hasan, Vol. 1, edited by J. J. Webster. London: Equinox (pp. 130–56). 2006 ‘Literacy pedagogy and social change: directions from Bernstein’s sociology’. In Knowledge, Power and Educational Reform: Applying the Sociology of Basil Bernstein, edited by R. Moore, M. Arnot, J. Beck and H. Daniels. London: Falmer (pp. 211–25). 2006 ‘Retrospective on SFL and literacy’ (M. A. K. Halliday Part I and Hasan Part II). In Language and Literacy: Functional Approaches, edited by R. Whittaker, M. O’Donnell and A. McCabe. London: Continuum (pp. 15–44). 2007 ‘Semantic networks: the description of linguistic meaning in SFL’ (with Carmel Cloran, Geoff Williams and Annabelle Lukin). In Continuing Discourse on Language, Vol. 2, edited by R. Hasan, C. Matthiessen and J. J. Webster. London: Equinox (pp. 697–738). 2007 ‘Private pleasure, public discourse: reﬂections on engaging with literature’. In Language and Verbal Art Revisited: Linguistic Approaches to the Study of Literature, edited by D. R. Miller and M. Turci. London: Equinox (pp. 13–40). 2008 ‘Interview on questions in applied linguistics’ (with M. A. K. Halliday). Mitchell, K. (ed.) Proceedings of the BAAL Annual Conference, 2007 (Pit Corder Colloquium, 7 September 2007), edited by M. Edwards. London: Scitsiuguil Press (ISBN 978-0-9559533-0-9). 2008 ‘Modes of learning, modes of teaching: semiotic mediation and knowledge’. SPELT Quarterly 23 (3): 2–19. 2009 ‘Wanted: a theory for integrated sociolinguistics’. In Semantic Variation: Mean- ing in Society and in Sociolinguistics: the Collected Works of Ruqaiya Hasan,Vol.2, edited by J. J. Webster. London: Equinox (pp. 5–40). 2009 ‘On semantic variation’. In Semantic Variation: Meaning in Society and in Sociolinguistics: the Collected Works of Ruqaiya Hasan,Vol.2,editedby J. J. Webster. London: Equinox (pp. 41–72). 2009 ‘A sociolinguistic interpretation of everyday talk between mothers and children’ (with C. Cloran). In Semantic Variation: Meaning in Society and in Sociolinguistics: the Collected Works of Ruqaiya Hasan, Vol. 2, edited by J. J. Webster. London: Equinox (pp. 75–118). 2009 ‘The place of context in a systemic functional model’. In Continuum Com- panion to Systemic Functional Linguistics, edited by M. A. K. Halliday and J. J. Webster. London: Continuum (pp. 166–89). 438 A Bibliography of Work by Ruqaiya Hasan

2010 ‘The meaning of “not” is not in “not” ’. In Appliable Linguistics, edited by A. Mahboob and N. Knight. London: Continuum (pp. 267–305). 2011 ‘On the process of teaching: a perspective from functional grammar’. In Lan- guage and Education: Learning and Teaching in Society: the Collected Works of Ruqaiya Hasan, Vol. 3, edited by J. J. Webster. London: Equinox (pp. 7–47). 2011 ‘Forms of discourse and forms of knowledge: reading Bernstein’ (with D. G. Butt). In Language and Education: Learning and Teaching in Society: the Col- lected Works of Ruqaiya Hasan, Vol. 3, edited by J. J. Webster. London: Equinox (pp. 99–165). 2011 ‘A timeless journey: on the past and future of present knowledge’. Selected Works of Ruqaiya Hasan on Applied Linguistics. Series World Famous Applied Linguists. Beijing: Foreign Language Teaching and Research Press (pp. xiv– xliii). 2012 ‘A view of semiotics in a social semiotic perspective’. Linguistics and the Human Sciences 5 (3): 251–79. 2013 ‘Choice, system, realisation: describing language as meaning potential’. In L. Fontaine, T. Bartlett and G. O’Grady (eds) Systemic Functional Linguistics: Exploring Choice. Cambridge: Cambridge University Press (pp. 269–99). 2014 ‘Linguistic sign and the science of linguistics: the foundations of appliability’. In F. Yan and J. J. Webster (eds) Developing Systemic Functional Linguistics: Theory and Application. London: Equinox (pp. 106–37). 2014 ‘Towards a paradigmatic description of context: systems, metafunctions, and semantics’. Functional Linguistics 1: 9 doi:10.1186/s40554-014-0009-y.

Books in press or in advanced preparation

The Collected Works of Ruqaiya Hasan, edited by Jonathan J. Webster. Vols 1–7. London: Equinox.

Vol. 4: Context in the System and Process of Language (in press) Vol. 5: Describing Language: Form and Function Vol. 6: Unity in Discourse: Texture and Structure Vol. 7: Verbal Art: a Social Semiotic Perspective Author Index

Abousnnouga, G. 271–2 Bowcher, W. 14, 22, 42, 51, 117, 136, Achugar, M. 367–70, 373, 376–7 149, 157, 163, 179–80, 182, 186, Acosta-Cabronero, R. 242, 247 192, 194–5, 198, 204, 235, 241, Aggarwal, G. 90, 111 245–6, 261, 264, 269, 271–3 Anstrom, K. 360, 376 Bredmar, M. 105, 113 Armstrong, E. 306, 309–12, 316–17, Brown, D. 215, 220 322, 325–7 Brown, E. 306–8, 310, 313, 329 Arnold, R. 242, 247 Brown, P. 207, 225 Aronin, L. 246 Brown, R. 207, 225 Butler, C. 228, 246 Butler, F. 360, 376 Bain, R. B. 363, 376 Butow, P. 89–90, 109, 111 Balboni, T. 89, 114 Butt, D. 14, 22, 25–6, 31, 42, 48, 51–2, Baldry, A. 256, 272 63, 78, 88, 100, 107, 109–10, 113, Bamberg, B. 312, 327 115, 117, 122, 136, 149, 160, 163, Barthes, R. 265, 272 165, 179–80, 182, 186, 198, 210, Bartlett, T. 109, 111, 168, 171, 175–8, 211, 225, 240, 246, 308–10, 316–17, 180–2, 189, 203–4, 241, 246 320, 322–4, 327 Bateman, J. 254–6, 258, 260, 262, 271–2, 407 Caffarel, A. 107, 111, 120–1, 135–6 Bateson, G. 40, 51, 145 Carpenter, B. 369, 370, 373, 376–7 Beelders, T. 260, 272 Carretero, M. 363, 378 Berazhny, I. 258, 260, 262, 272 Cheng, W. 134, 137 Bergh, L. 260, 272 Cheong, Y. Y. vii, 262–4, 272 Bernstein, B. 3, 10–13, 88, 106, 110, Cheung, H. N. S. 131, 137 145, 152–4, 163, 168, 171, 179, 182, Christie, F. 106, 111, 337, 355 214, 224, 244, 275–7, 283–4, 299, Chu, P. Y. 116, 137 337, 341, 346, 354–5, 358, 376, 385, Chung, W. Y. J. 134, 137 389, 393–5, 399, 410 Ciccone, N. 325, 327 Berry, M. 93, 110, 175, 178, 181–2, 184, Clark, K. 90, 111 189–90, 196, 198, 200, 204, 240, Clayton, J. M. 89, 90, 109, 111 246 Cléirigh, C. 41–2, 52, 54, 88, 107, Biven, L. 48, 54 113 Björklund, M. 308, 327 Cloran, C. 12–13, 31, 52, 83–6, 88, 97, Blackmur, R. P. 27, 51 105, 108–9, 111–12, 115–17, 119, Block, S. 89, 114 126–7, 136–8, 156, 163, 175–7, Blommaert, J. 167, 170–1, 180, 182–3, 187, 204, 216, 225, 240, 246, 182 276, 286, 293, 298–9, 308, 309, Bloom, H. 40, 51 314–15, 317, 320, 323–4, 326–7, Bloomﬁeld, L. 382 354–5, 393, 410, 435, 437 Bogardus, E. S. 213, 225 Cook, G. 233, 246 Bohle, R. 263, 272 Coupland, N. 207, 225 Bortoluzzi, M. 262–4, 272 Cox, B. E. 306–7, 310, 312, 318, 320, Bottenberg, D. 306, 309–10, 324, 327 325, 327

439 440 Author Index

Cross, M. 163 Giroux, H. 373, 377 Currow, D. 89–90, 111 Glock, H.-J. 37, 52 Godecke, E. 325, 327 Damasio, A. 37, 52 Goffman, E. 207, 225 Davies, M. 42, 52 Good, R. H. III 335, 355 Davies, P. 406, 410 Goodwin, C. 166, 172, 183 Demantik, M. 308, 310, 327 Greatbatch, D. 93, 112 Demjen, Z. 89, 114 Greaves, W. S. 22, 52, 111, 137 de Oliveira, L. C. 369, 377 Greer, S. 369, 377 Devine, R. J. 90, 111 Gregersen, N. H. 406, 410 Devlin, P. 106, 111 Grossman, J. 106, 113 DiCerbo, P. 360, 376 Drew, P. 90, 111 Haliburn, J. 31, 52, 100, 110, 309–10, Duranti, A. 166, 172, 183 316–18, 320, 327 Hall, P. 109, 111, 116, 137 Earhart, M. 370, 377 Halliday, M. A. K. 11, 22, 25, 35, 40, 41, Edwards, A. D. 208, 225 42, 48, 52, 59, 61, 64, 67, 78, 84, 93, Edwards, C. L. 28, 52 107, 109–11, 115–22, 130–1, 133–5, Eggins, S. 115, 137 137–8, 143, 145, 147, 151–3, 162–3, Elliott, L. 90, 111 169–72, 175, 177, 181, 183, 187, El-Menoufy, A. M. E. S. 42, 52 190, 192, 198–9, 204–5, 207–14, Engeström, Y. 227, 234, 246 216–17, 219, 224–6, 231, 235–40, Exley, C. 105, 111 244–7, 271–2, 301–4, 312, 314–16, Eyal, R. 100, 110 320–1, 328, 335, 350, 355, 381, 383, 385, 389, 394–6, 399–400, 403, 406, Fahey, R. 88, 113, 115, 139, 393 410 Fang, Z. 306–7, 312–13, 318, 325, 327 Hancock, K. M. 89, 111 Fanshel, D. 93, 113 Harris, S. 93, 111 Fawcett, R. 240, 244, 247 Hasan, R. 3–4, 11, 13–14, 22–33, 35–8, Fiedler, K. 207, 226 41–2, 49–51, 52–3, 59–68, 74–9, Fine, J. 324, 327 83–90, 92–101, 104–12, 115–23, Fink, R. J. 309–10, 318, 320, 326, 328 125–8, 130–1, 133, 135–8, 143–58, Firth, J. R. 13, 22, 43, 52, 145–6, 163, 160–5, 166–83, 185–92, 194–5, 286, 339, 382, 385–6, 388, 390, 198–9, 201, 203, 205–19, 222, 396–7, 409–10 224–6, 227–8, 231–4, 236–7, 239–41, Fitzgerald, J. 306–7, 310, 312, 318, 330 244–5, 247, 251–5, 262–4, 268–9, Fontaine, L. 109, 111 271–3, 275–6, 286, 293, 298–9, Fowler, R. 59, 62, 64, 66, 78 300–2, 304–7, 309–17, 319–26, Freebody, P. 106, 111 333–40, 342–3, 348, 350, 352, Freeman, D. 28, 52 354–64, 373–7, 383, 387, 394–5, Freire, P. 336, 355 399, 400, 408, 410–11 French, R. 347, 355 Hedberg, N. 306, 309–11, 318, 320, Friedman, L. B. 313, 323, 327 324, 326–8 Fries, P. H. 240, 247, 308, 325–7 Henderson-Brooks, C. 31, 48, 52, 100, 110, 309–10, 316–18, 320, 327 Garcia Marrugo, A. I. 308, 315, 327 Heritage, J. 90, 93, 111–12 Garﬁnkel, H. 393, 410 Herke, M. 88, 107, 113, 116, 139–40 Gibbon, D. 245, 248 Herzog, M. 390, 411 Giles, H. 207, 225 Hiippala, T. 256, 258, 267, 273 Gilman, A. 207, 225 Hill, T. 389, 411 Author Index 441

Hillier, H. 184–5, 192, 196, 198, 203–5 Li, E. S. E. 120, 139 Hjelmslev, L. 105, 113 Liang, J. Y. 157, 245, 261, 264, 269, Hodge, J. R. 242, 247 271–3 Hoey, M. 326, 328 Lindholm, C. 93, 113 Hofvendahlm, J. 93, 113 Linell, P. 93, 105, 113–14 Holmqvist, K. 260, 273 Luke, A. 106, 111, 338, 356 Holsanova, J. 260, 273 Luke, K. K. 130, 139 Hymes, D. 171, 183, 208, 226 Lukin, A. 30, 31, 53, 84–6, 88, 97, 107–9, 112–13, 115–16, 136, 138–9, Jackson, V. 89, 114 243, 248, 308, 314–15, 318, 321–3, Jakobson, R. 32, 35, 50, 53, 59–60, 63, 329 66–8, 77, 79, 100, 113 Janet, P. 28, 53 Machin, D. 271–2 Maciejewski, P. 89, 114 Kaminski, R. A. 335, 355 Mack, J. 89, 114 Katz, A. 360, 376 Malcolm, K. 31, 53 Kennet, J. 363, 378 Maley, Y. 88, 113, 115, 139 Kermode, F. 26, 53 Malinowski, B. 13, 143, 145–6, 165, Khoo, K. M. 31, 33, 53, 149, 252, 309, 384 316–18, 320–2, 328 Mann, W. 42, 53, 115, 139 King, M. L. 306–7, 310, 329 Martin, J. R. 14, 23, 42, 53, 67, 72, 79, Kipps, C. M. 242, 247 93, 100, 104, 106, 110–11, 113, 115, Kok, B. 325, 327 120, 136, 139, 144–5, 161, 163, 165, Koller, V. 89, 114 190, 203, 205, 237, 240, 248, 254, Korner, A. 49, 53, 100, 110 261, 273, 308, 312, 320, 323, 326, Kress, G. 244, 247, 256–7, 259–60, 329, 337, 356 267–8, 271, 273 Matthews, S. 130–1, 139 Krippendorff, K. 385, 411 Matthiessen, C. M. I. M. 22, 31, 35, 42, Kruglanski, A. W. 207, 226 52–4, 88, 93, 107, 109–11, 113, Kwok, H. 131, 138 115–16, 120, 122, 131, 134–6, 138–40, 163, 165, 168–9, 173, 181, Labov, W. 88, 93, 113–14, 145, 389–91, 183, 207, 209–14, 217, 219, 224–6, 394, 407–8, 411 235–7, 239–40, 247–8, 301, 304, Lacey, J. 90, 111 321, 328, 350, 355 Lam, M. 163, 165, 168, 173, 183, McAllister, I. 213, 226 209, 211, 214–15, 219–21, 223, McDonald, E. 120, 131, 133, 137, 139 226 McIntosh, A. 9, 25, 53, 143, 147, 163, Lamb, S. M. 387, 411 271–2, 389, 410 Lanham, R. A. 43, 53 Meares, R. 31, 48, 52, 54, 100, 110, Larsson, U. S. 105, 114 309–10, 316–18, 320, 325–7, 329 Lawson, R. 213, 226 Mehler, A. 245, 248 Lee, P. J. 90, 111 Miettinen, R. 227, 234, 246 Leinhardt, G. 363, 369, 376–7 Miller, D. R. 77, 79, 160, 399 Lemke, J. 121, 138, 240, 247, 253–4, Millet, J. 360, 376 256, 258–60, 273, 326, 328 Mitchell, S. 89, 114 Lemme, M. 306, 309, 310, 324, 327 Moje, E. B. 360, 377 Lepschy, G. 236, 247 Moore, A. 31, 48, 52, 88, 90, 100, Leung, C. S. 131–3, 139 106–8, 109–11, 113, 116, 139, Levin, S. R. 29, 53 243–5, 248, 309–10, 316–18, 320, Levinson, S. 94, 113, 123, 138, 207 326–7 442 Author Index

Moore, R. 213, 226 Scott, C. E. 308, 310, 315, 319, 321–2, Mukaˇrovský, J. 61, 65, 66, 78, 80 326, 330 Seale, C. 105, 114 Nesbitt, C. 88, 107, 113 Semin, G. R. 207, 226 Nesﬁeld, J. C. 383, 411 Semino, E. 89, 114 Nestor, P. J. 242, 247 Shanahan, C. 360, 377 Neumann, S. 109, 113 Shanahan, T. 306–7, 310, 320, 327, Noel, M. 90, 111 360, 377 Shukman, A. 50, 54 O’Donnell, M. 107, 114 Silﬁes, L. N. 363, 378 O’Grady, G. 42, 54, 109, 111, 182 Silverstein, M. 180, 183 O’Halloran, K. 256, 259, 273, 407 Simpson, E. 27, 54 Ohmann, R. 29, 54 Simpson, J. 90, 111 O’Toole, L. M. 50, 54, 271, 273 Simpson, P. 59, 60, 80 Otto, B. 306–8, 310, 320, 329 Slade, D. 115, 137, 400 Oteíza, T. 368–9, 376–7 Smith, B. 42, 51 Spencer, J. 144, 165 Pabjan, B. 213, 226 Spiegel, D. L. 306, 307, 310, 312, 318, Panksepp, A. 48, 54 330 Pappas, C. C. 306–8, 310, 312–13, 316, Starkhammar, H. 105, 114 329 Stearns, P. N. 363, 377 Parsons, G. 307, 310, 320, 326, 329 Steele, T. 50, 54 Pettegrew, B. 306–7, 310, 329 Steiner, E. 109, 114, 120, 140 Phillips, J. 4, 5, 21, 23 Stöckl, H. 257, 266, 268, 271, 273 Piaget, J. 394, 409, 411 Strevens, P. 61, 78, 143, 147, 163, Porges, S. W. 48, 54 271–2, 389, 410 Prakasam, V. 120, 140 Stuart-Smith, V. 177, 183, 308, 314, Prigerson, H. 89, 114 326–7 Punamäki, R.-L. 227, 234, 246 Stuckey, S. 366, 368, 377 Sulzby, E. 306–7, 310, 313, 320, 323, Rahm, H. 260, 273 327 Ray, A. 89, 114 Swales, J. 254, 273 Rentel, V. M. 306–7, 310, 312, 329 Rivera, C. 360, 376 Tam, H. S. 125, 130–1, 140 Robbins, D. 162, 165 Tang, T. C. 130, 140 Romary, L. 245, 248 Tattersall, M. H. N. 89–90, 109, 111 Rose, D. 120, 140 Taylor, A. 368, 377 Rosenau, P. M. 360, 377 Taylor, S. 369, 377 Rothery, J. 337, 356 Taylor Torsello, C. 77, 80 Royce, T. 256, 273 Teich, E. 109, 114, 120, 140 Rubini, M. 207, 226 Tench, P. 42, 54 Ryan, L. 29, 54 Teruya, K. 109, 113, 120, 140, 163, 165, 168, 173, 183, 209, 214, 226 Salvucci, L. K. 366, 368 Thai, M. D. 120, 140 Sandén, I. 105, 114 Thibault, P. 256, 258, 272, 274 Saramago, J. 166–7, 183 Thompson, G. 184, 204 Saussure, F. de 36, 49, 54, 145–6, 158, Thompson, S. A. 42, 53 165, 386–8, 390, 406, 411 Threadgold, T. 106, 111 Schleppegrell, M. J. 347, 356, 367–9, Thyne, M. A. 213, 226 376–7 Trice, E. 89, 114 Author Index 443

Tsui, A. 93, 114 Weinreich, U. 390, 411 Tuckwell, K. R. 42, 54, 89, 100, Westby, C. E. 309, 311, 320, 328 109–11 White, P. R. R. 72, 79, 100, 113 Turner, G. J. 115, 117, 119, 140 Whorf, B. 21, 63, 145, 165, 345, 356, 383 Ure, J. 210, 226 Wilkinson, F. 213, 226 Williams, G. 84–6, 88, 93, 97, 106, van Dijk, T. 171–2, 176, 183 108–9, 112, 114, 115, 128, 135, 138, van Leeuwen, T. 179, 183, 244, 247, 140, 153, 165, 335, 347–8, 354, 356, 257–61, 267–8, 270–1, 273–4 361, 400 Ventola, E. 123, 140, 271, 274 Wineburg, S. 363, 378 Vološinov, V. 54, 161, 165 Wittgenstein, L. 25, 36, 144, 165, 403 Vonwiller, J. 42, 54 Wolf, G. 233, 248 Voss, J. F. 363, 378 Wong, M. C. 116, 120, 123, 134, 137, Vygotsky, L. 145, 165, 339–40, 342, 140 350, 356, 394, 408–9, 411 Wright, A. 89, 114 Wu, C. 88, 107, 109, 113, 116, 139–40 Wake, B. J. 116, 140 Wu, K. Y. 130, 131, 132, 140 Waller, R. 255, 274 Walter, D. 369–70, 373, 377 Walter, T. 105, 114 Yang, A. 306, 308, 325, 330 Walton, C. 106, 111 Young, K. M. 369, 377 Watts, R. J. 174, 183 Young, L. 177, 183, 308, 314, 324, Webster, J. 30, 42, 52–4, 302, 328 326–7 Wegener, R. 88, 113, 116–17, 122, 136, Yip, V. 130–1, 139 139, 157, 244, 248 Weiner, J. 88, 114 Zhang, B. 89, 114 Subject Index

AAVE 390, 410 n. 1 central entity (CE) 176, 276 abstraction 63, 117, 157, 169–70, 212, see also message 233, 252, 269, 342, 404 chain academic activity 200, 202 chain formation 304, 320–1, 323 see also social activity chain interaction 31, 149, 152, 300, academic language 196, 360, 363, 304–6, 310, 314–16, 318–19, 373–4 322–3, 325–6 accent 153, 391 cohesive chain 216, 304–5, 314, 322 act of meaning 25, 48 focal chain 316 action see material action; sphere of identity chain (IC) 32, 149, 300, 304 action; verbal action lexical chain 31, 33, 304 see also ARC realization chain 240, 246 n. 4 action literacy 334, 336, 338, 341, significant chain 326 n. 15 359–60, 362, 364, 367, 373 Action RU 282–5, 292, 295–7 similarity chain 31, 33, 149, 300, 304 activation–construal dialectic 117, 172 chaining 177, 304, 320, 323, 423 adjacency pair 305 see also lexical rendering; tokens affordance 179, 246 n. 6, 256, 259, channel (contextual, language), 267–8, 271–2, 367 CHANNEL (as system) 151, 169, agency 89, 107, 109 n. 6, 359, 371 172–4, 179, 204 n. 8 alignment 41–3, 170, 181, 310 choice 11, 13–14, 27, 29, 31, 38, 40–3, AMPLIFICATION (logical meaning) 119, 49–50, 66, 68, 77–8, 83, 92–3, 96–7, 122, 127, 325 102, 109 n. 9, 116–18, 123, 126, ancillary 150, 156, 174, 182 n. 10, 128, 131–2, 149, 160, 179–80, 215–16, 218, 220–3, 286, 384 182 n. 13, 184–5, 187, 189–91, 195, see also constitutive 197, 208, 210–12, 215, 217, 228, aphasic discourse 309–10, 322, 325 n. 4 233, 254, 258–9, 261, 264–5, 293, apposition 49 301–2, 343, 346, 357, 361–2, 366–9, appraisal analysis 72, 326 n. 21 371–3, 375, 387, 392, 395, 404 ARC (action, relation and contact) 14, classification (as in weak or strong) 147, 157, 422 155–7 associative bond 386 classification (in cohesion) 304–5 assumptive (also non-assumptive) 87, CLASSIFICATION (network) 88, 110 95, 97, 101–2, 107, 110 nn. 18, 19, n. 16, 119, 122, 127 358 cline of de/contextualization 156 cline of instantiation see instantiation biography of interaction 209–13, 215, clinical communication 85, 100 219–22, 224 clinical contexts of research 134, 311 biology(ical) 37, 48, 145–6, 157, 214, code 10–11, 18, 37, 85, 101, 106, 108, 237–40, 363, 401 154–5, 157–8, 161, 167, 179, 236, see also eco-social environment 241–4, 259, 293, 318, 326 n. 14, 422, 429 California History Project (CHP) 364 code-switching 179 Cantonese semantics 135 elaborated vs restricted 10

444 Subject Index 445 coding orientation 10, 13, 153–6, 361, context of creation 29, 49, 73, 415, 422 78 n. 7, 269 coherence 14, 33, 252, 256, 300–2, context dependent, context 304–6, 310–20, 322, 324–6, 350, 432 independent 156 cohesion 11, 14, 28, 33–4, 67, 171, 173, context of situation 13–14, 63, 88, 180, 230, 301–5, 310, 312, 318, 143–4, 147, 151–3, 227, 232, 234, 320–1, 323–6, 333, 350, 416, 423 237, 241, 251, 259, 269, 271, cohesive harmony 11, 31, 33, 67, 149, 300–1, 315, 372, 391–2, 397, 404, 216, 252, 300–2, 304–6, 309–12, 418, 421 350, 423, 425, 430 context potential 175, 186, 203 n. 4 Cohesive Harmony Analysis (CHA) context theory 227, 234 300, 304, 306, 311, 313, 317–18, contextual configuration (CC) 14, 323, 425 89, 117, 136 n. 5, 148–9, 151, 162, cohesive harmony coding sheet 311 171, 181, 185, 208, 252–3, 262, cohesive harmony index (CHI) 306, 271, 314–15, 325 n. 2, 388, 396, 309 422 cohesive link con/textual integration 150–1 co-reference/co-classification/ contextualization 149, 203 co-extension 304–5, 323 contextual parameters 14, 63, 143, COLLABORATIVE CONTEXT see under 147–50, 169–71, 190, 209, 210, context 232, 422; see also field, tenor, compositional meaning 257, 259 mode communicative gestalt 271 contextual/registerial consistency see computer-mediated communication consistency (CMC) 120, 228, 241 contextual significance 90 conjunctive 305, 326 n. 20 DEPENDENT CONTEXT 151 consciousness 27, 85, 86, 161, 189, DOMINANT CONTEXT 151 361, 363, 389, 394, 408, 430 first- and second-order context consistency (of foregrounding) 29, 39, 168–70, 172, 175 42, 51, 61, 65, 67, 78 n. 5, 419 instructional context 275–7, 282, consistency (of register, context) 285–90, 293 151–2, 188, 314–16 open context 118 constituency 254, 298 n. 3 PRE-TEXT relevant contextual features constitutive 150, 174, 182 n. 10, 191, 175, 187–91, 195, 197–8, 201, 203 215–24, 286, 384 specific context vs unique context see also ancillary 118 construe (construal) 5, 18, 20, 24–6, 70, VIA-TEXT relevant contextual features 73, 87, 94, 102, 104, 108, 117–18, 175, 187–91, 195, 197–8, 201, 203 121, 144, 147–51, 160, 168, 172–6, contextually open (networks) 86, 106, 178–81, 187–90, 217–19, 221, 227, 108, 109 n. 1, 115–18 234, 236–7, 254, 256–7, 259, 263, CONTINUATION network 107, 119, 122, 266, 268, 271, 275–6, 279–81, 283, 127 285–9, 292–4, 298, 315, 326 n. 18, continuative 305 348, 352, 357–8, 366, 374, 385, 387, Conversational Model (of 397, 399, 403, 424 Psychotherapy) 48, 49, 317, content-based activity 199–200 325 n. 8 see also people-based activity critical discourse analysis (CDA) 167, context 171 actual/virtual context 154, 156 cultural activity theory 227, 234 COLLABORATIVE CONTEXT 151 cultural intervention 145 446 Subject Index de-automatization 67, 78 n. 4 Event Orientation (EO) 176 decontextualized (language) 154, exophora 304, 316 156–7, 176, 245, 275–6, 293, 298 exophoric deixis 174, 180 default dependency 149 exotropic (theory) 395, 400–1 delicacy 85–6, 96, 116, 147, 151, see also endotropic 210–11, 215, 217, 221, 235, 301, experiential 14, 48, 67, 71–2, 74, 322, 323, 325 n. 8, 423 78 n. 8, 90, 96, 102, 109 n. 8, degree of delicacy 151, 210–11, 215, 110 n. 14, 119, 126, 176–7, 199, 221 235, 286, 300, 302, 303–4, 322–3, demand 326 n. 14, 346, 348, 350, 353, 396, (goods and services) imperative 20, 432 93, 126–7, 134, 285 see also ideational; metafunction (information) interrogative 20, 93, exponence 86, 397 107, 126–7, 134, 294–5 expressive resource 145, 158 DEMAND INFORMATION (network) 86, see also proxemics modality 88, 90, 93–5, 101, 106–7, 110 n. 18 extended reference 320 dementia 242 DEPENDENT CONTEXT see under context fashions of meaning 84, 86 discourse (spoken and written) 232, fashions of speaking 34, 383 321 field (of discourse) 14, 63, 93, 143, display text 262 147–50, 160–1, 167–71, 174, 182 double articulation 26, 49, 51, 62–3, n. 10, 185, 190–1, 198, 200, 209–12, 77, 419 215, 218–19, 231, 244 double entendre 39 FIELD (system network) 201–2, 210–12, double-seeing 40 215, 219, 220–2, 241 dyad 12–13, 22 n. 5, 123, 128, 153, see also context of situation 209, 294, 296, 298 n. 4 finite 98, 101, 176, 303, 350 foregrounding 61–2, 65–9, 77, 78 n. 4, ecosocial (eco-social) environment 4, 430 13, 237, 269, 271 see also consistency (of foregrounding) ellipsis 97–8, 195, 207, 302–3, 305, formative motif 89, 108 320, 326 n. 13 framing 257 see also cohesion; substitution see also information value; salience end-of-life (EOL) discussion 85, 89–90, framing (in relation to coding 92–3, 96–8, 100–6, 109 n. 7 orientation) 155–7 endotropic (theory) 400–1 framing (questions) 93, 96 see also exotropic function Epithet 5, 304, 322 interface/organizing function 238–9 état de langue 390–1 event (language, communicative event) structuring/layering/pre-selecting 90, 155, 167, 171, 177, 185–6, 240 188–91, 193, 196–202, 203 n. 4, 211, 228, 241, 271, 298, 391 GeM (genre and multimodality) model event (in relation to ‘message’) 276, 254–6, 258 293, 296 gender 175, 231, 393 event (as stage in a genre, e.g. Initiating generalized structure potential 14 Event, Sequent Event, Final Event) see also GSP 253, 313 generalization 28, 62, 64, 176–7, 213, event (as in verbal group) 350, 351, 276, 279–81, 285, 288, 290–3, 353, 354 296–8, 314, 390, 404 Subject Index 447 genre-based pedagogy 336–8, 341, 348, interpersonal (meaning) 20, 67, 86, 352, 354 n. 2 93–4, 99, 109 n. 8, 119, 123, 126, genre-specific semantic potential 148, 136 n. 7, 176–7, 199, 213, 219, 235, 212, 252 244, 301, 303, 326 n. 21, 346, see also GSP 357–9, 365, 371–2 goal orientation 161, 210, 286 interrogative 4, 8, 17, 20, 47, 73, 93–4, graduation 100, 110 96, 110 n. 12, 130–3 grammatical cohesive device 302, 305 intersemiotic complementarity 256 grammatical intricacy 321 intersubjective agreement 198 grammatical metaphor 321, 326 n. 14, intonation 4, 8, 22 n. 4, 42, 110 n. 12, 350 133, 182 n. 12, 302, 324 grammatical parallelism (GP) 60, 66–9, intra-registerial/inter-registerial variation 72, 76–7 83–5, 100, 104 grammatics 107, 346–8, 350, 355 n. 5, intra-stratal 240, 246 n. 4, 335, 423 390 invisible semiotic mediation 156 graphology 63 GSP (generic structure potential) 14, knowledge 154–6, 213, 275, 277, 286, 148, 212, 251–6, 258, 261–4, 268–9, 312, 333–6, 338–9, 341, 346–7, 270–2, 275–7, 283–4, 287–9, 313, 358–60, 361, 363, 369, 373–6, 383, 423, 428 396, 398–9, 408 knowledge relations 193, 196–7, 200, HAP (higher autonomy profession) 13, 202 22 n. 5, 153 hierarchical relation 195–6 langue 146–7, 170, 386–8, 390–1 hyponomy 323 see also parole hypotactic, hypotaxis 17–18, 42–3, 45, language as action 19 47, 49, 303 language event see event language-in-use 167–8 ideational (meaning) 20, 31, 177, 199, LANGUAGE MEDIUM 172 212, 215, 235, 264, 301, 303, 359, latent patterns 35, 40–1, 43, 51 365–6 layer (of multimodal text) 255, 271 implicit/implicitness 85, 89, 92, 100–1, lexical cohesion 302, 304, 320–1, 104–5, 207, 337, 383 325 n. 1 incipience 85, 89, 100, 104–5 lexical rendering 304, 320 indexical signs 49 see also chaining indexical orders 180 lexicogrammar 11, 13, 27, 40–1, 49, see also order of indexicality 63–6, 83, 90, 115–17, 120, 135, individuation 85, 87, 89, 104–5, 135 n. 3, 150, 162, 173, 176, 190, 163 n. 6 207, 219, 235–7, 239, 298, 348, 352, information value 257 355 n. 5, 387, 396, 422–3 see also salience; framing lexis 12, 18–19, 31, 33–4, 38, 40, instantiation (also cline of) 27, 77, 235–6, 319, 321, 423, 428 116–17, 152–3, 235, 237, 251–2, line of arbitrariness 49, 162 n. 2, 403 254, 264, 270, 281–2, 337, 341, 343, linguistic category model 207 348, 352 linguistic pointillism 42 institutional(ized) 161, 209, 275, 293 literary (criticism, language, studies) 3, see also individuation 9, 10, 25, 39, 41, 50, 64, 66, 68, 115, institutional linguistics 389 149, 160, 308, 363, 385 integrated context/text see con/textual literature 28–9, 38–9, 50, 59–64, 67, 77, integration under context 160, 308, 394, 406, 409 448 Subject Index logogenetic, logogenesis 32–3, 42, 51, mood 305, 314, 316, 319 declarative 4, 96, 99, 110 n. 12 LAP (lower autonomy profession) 13, imperative 72 22 n. 5, 153 morphology of meaning 386 motivation 26, 61, 65, 78 n. 4 see also consistency (of foregrounding) macro-element 277, 283, 286–7 Mukaˇrovský–Jakobson Theory 66–8 Malinowski, B. 13, 143, 145–6, 384, multimodal(ity) 157, 227, 228, 230, 424, 427 234–5, 241–3, 245 n. 2, 251–2, marked/unmarked 254–64, 266–72, 309, 400, 407, 427 attitudinally marked/unmarked 130, multimodal communication 227–8, 133, 135 n. 10 241–2 material action 144, 150, 174, 191, multimodal context 243 200, 210, 215, 217–18, 221, 386 multimodal document 254–6 material situational setting (MSS) 14, see also GeM model 148, 161, 163 n. 4, 175, 181 n. 6, multiple coding system 63 185, 188–9, 232, 295, 314, 343, 396 multiple contextual framework 160 mathetic/pragmatic 5, 15, 19–20 multiplying meaning 25, 256 meaning potential see under potential multisemiotic 259, 407 MEDIUM 172, 179–80, 204 n. 8 multi-stratal 116, 135 n. 2 message (semantic unit) 83–90, 93, 98, multi-unit design 93 102, 104–8, 109 n. 8, 118, 122, 126, mutual prehension 149 276, 298 n. 2 see also assumptive; prefaced; progressive; punctuative narrative 8–9, 17, 19, 50, 263, 309–10, metafunction 109 n. 8, 116–19, 313, 421 122–3, 126, 135, 136 n. 12, 143, 177, natural vs irrealis 210 179, 212–13, 235, 240, 301–4, 404, negative-final particle contraction 132 432 neurocognitive linguistics 387, 405 see also ideational; interpersonal; neutral polarity 131–2 textual Nigel 4–9, 15–21, 22 nn. 1, 4, 393 metafunctional configuration 179 nominal group 33, 45, 71, 190, 322 metalanguage 340, 362, 365, 373, nominalization 326 n. 14 376 n. 2 non-verbal communication 242 metaphor 10, 28, 37–8, 63, 65, 68, 71, see also multimodal 78 n. 8, 143, 166, 256, 321, norm 61, 67, 161, 171, 184, 338, 360 326 n. 14, 335–6, 350, 403–4, 419, normal coherence 311 424 meta-semiotic 339 obligatory vs optional (elements) 99, modality 4, 126, 157–8, 207, 219, 148, 251–4, 256, 264, 267, 270 241–5, 303, 371, 407 see also GSP; ordering mode (of discourse) 14, 63, 143, 147–9, ontogenesis 3, 144, 154, 156–7, 350, 155, 160–1, 166, 168–72, 174, 420 176–80, 185, 190–2, 198, 206, ordered typology of systems 213–14 209–11, 216, 218, 227–8, 232, ordering 253, 261 234–6, 241–2, 244, 255–8, 261, 266, see also GSP; obligatory vs optional 268, 300, 358, 402, 422, 426–7 order of indexicality 180 monologic 7, 9, 160, 211, 241, 300, see also indexical order 325 n. 7 orders of relevance 153–4, 393 mono-stratal 236 overall tone (of a text) 89 Subject Index 449 palliative care (context, corpus, potentially relevant vs actually discourse) 89–90, 93, 96–100, 104, relevant features of context 188–9 106, 109 syntagmatic potential 35 paradigmatic 15, 48, 83–4, 89, see also GSP 117–18, 149, 235, 304, 314, 362, pragmatic environment 271 386, 422–3 prefaced (feature of message) 86, 87, parameters of context 14, 117, 147–50, 88, 94, 98–9, 101, 102, 107, 161, 209–10, 215–16 110 n. 19, 118 see also context of situation; field; see also message tenor; mode pre-text context 175 paratactic, parataxis see also first-order context; PRE-TEXT paratactic insistence 47 relevant contextual features; parole 144, 146–7, 149, 153, 170, 386–7 VIA-text context see also langue PRE-TEXT relevant contextual features see participant role 11, 206–8, 212, 214 under context pedagogic practice (official vs local) primary knower, PRIMARY KNOWER 193, 275, 290, 293–4, 296, 337, 360–1 196–8, 204 n. 10 people-based activity 199 PRIMARY TEXT 151 see also content-based activity see also SUBTEXT; DEPENDENT CONTEXT; personal distance 206–14, 219, 221–2, COLLABORATIVE CONTEXT 224, 358, 425 probe 95–9, 102, 104, 107, 110 n. 13, see also social distance 126, 423 pervasive parallelism 59–60, 68–9, professional development (PD) 90, 73–4, 76–7 364–5, 367, 369–70, 373, 376 n. 1 phonetics 20, 116, 135 n. 3, 235–9, 419 progressive (as in message) 98, 119–20, phonic/graphic opposition 172–6, 125–7 178–9 see also message (semantic unit) phonology 13, 27, 41, 49, 63–4, 66, prosodic 43, 49, 69 116, 135 n. 3, 235–9, 388, 397, 404, protolanguage 5, 20, 135 n. 2, 145, 236 419, 430 proxemic modality 158 phylogenetic function 237 see also expressive resource plane (expression/content/contextual psychotherapy 49, 317–18, 430 plane) 11, 135 n. 2, 227, 234–9, see also Conversational Model 245, 245 n. 2, 359, 387, 404–5 psychotherapy discourse 309, 317, 318, play (theatrical) 26, 29–40, 50 320, 324, 325 n. 7 playing (e.g. children) 153, 156, 159, punctuative (as in message) 98, 110 160, 341–2, 344, 351, 393 n. 14, 119–20, 122–3, 125–6 poet, poetry 24, 28, 40, 45, 48–50, see also message (semantic unit) 60–1, 66–9, 71, 77, 78 n. 6, 308, 324, 334 pyrotechnics (lexicogrammatical, verbal) polarity 4, 98–9, 101, 130–3, 173, 303 27–8, 424 post hoc analysis 188 potential 392, 394 question, types of questions 4, 7, 9, 14, instance vs potential 117–18, 235 32, 47, 73, 85, 87, 88, 90–4, 96–9, meaning/semiotic potential 7, 9, 12, 101–4, 107, 110 n. 18, 116, 120, 15, 29, 40–1, 51, 63, 148, 203 n. 4, 126–7, 129–34, 136 n. 10, 208, 224, 236, 242, 244–5, 264, 301, 361, 232, 281, 285, 287–8, 290–1, 293, 369, 373, 374–5 305, 339, 348, 358, 362, 365, 368, potential of context/context potential 371 186, 203 n. 4, 314 see also framing (questions) 450 Subject Index quotidian 150, 198, 220 salience 257, 259–60, 263 see also specialized see also information value; framing scale 11, 28, 39, 41–2, 66, 108, 116, 118, 159, 170, 172, 174–6, 178–81, rank (scale) 41, 66, 108, 116, 118, 200, 244, 312, 320, 325 n. 4, 398, 136 n. 12, 176, 235, 240, 303 429 rationality 3–4, 384, 387, 428 schooling 154, 157, 333–6, 339–41, reading path 257, 259–62 360, 430 realization 11, 13, 27, 48, 84, 86, 97–9, sctx 175–82 105–8, 110 n. 13, 116–17, 120, 125–6, 130–3, 153, 155, 170, 173, see also second order 180, 195, 203, 211–12, 215, 217–19, secondary addressee 193–4, 200 221–2, 235, 240–1, 246 n. 4, 251–2, secondary knower, SECONDARY KNOWER 258, 266, 268–9, 271, 312, 314–15, 193, 196, 197, 204 n. 10 317, 326 n. 14, 341, 343, 346, 348, secondary school history 364 351–2, 387, 392, 396–7, 399, 403, second order (context, semiosis) 62, 64, 405, 419–20 150, 168–70, 172, 175, 178, 180, 181 referential 32, 67, 216 n. 8, 217, 220 reflection-based activity 199, 218–19 selection expression 117, 127 see also relation-based activity semantic boundary 301, 314 relational (semiotic system) 36 semantic disjuncture 374 RELATIONAL ENACTMENT system 86, 93, semantic distance 382 122, 126–7, 136 n. 7 semantic drift 100, 102, 108, 227, 234 relation-based activity 199, 218–23 semantic flow 257–8 see also reflection-based activity semantic innovation 28 relevant context 14, 148, 160, 167–8, semantic intensity 43 175, 181 n. 6, 185, 187–8, 337, 387, semantic network 83, 85–6, 88–90, 405 92–3, 106–9, 110 n. 18, 115–23, see also second-order context 126–8, 133–6, 396–7, 416, 420–1, register 3, 14, 27, 59–61, 63, 83–6, 89, 428, 430 98–100, 105–9, 122, 136 n. 5, 148–9, semantic person 386 151–3, 160–1, 188, 206, 251–3, semantic potential 148, 212, 244, 252 270–1, 300–1, 306–8, 312–15, semantic progression 28, 41 319–26, 341, 343, 346, 351–2, 357, semantics 3, 7, 9, 11, 13, 15, 27, 44, 49, 362, 365, 373–4, 384, 389, 391, 63–5, 78 n. 4, 83–6, 88–90, 98, 100, 401–2, 405, 422–3, 430 102, 104–9, 115–18, 120–1, 126, register variation 83, 357 135 n. 3, 150, 172–5, 178–80, 199, representational 158, 244, 257, 313, 210, 235–41, 333, 386–7, 389, 391, 357 396–7, 406, 417, 419, 423 resource integration principle 256 semantic shape 275 rhetorical mode 169, 185, 270 semantic stratum 83, 118, 120, 173, rhetorical structure 42, 255, 258, 176, 244, 297–8 326 n. 23 semantic system 12, 84, 86, 109 n. 8, rhetorical unit 118, 136 n. 5, 176, 240 256, 276–8, 298 n. 3, 315, 326 n. 23, semantic unit 48, 118, 122–3, 190, 271, 428 301 rhetorical unity 256 semantic variation 12–13, 83–4, 88, ROLE ALLOCATION (interpersonal 102, 104, 109, 115, 333, 340, meaning) 109 n. 10, 119, 123, 354 n. 2, 393, 396, 406, 420, 422, 136 427–9, 432 Subject Index 451 semiosis symbolic articulation 27, 34–5, 39–40, early (proxemic) semiosis 145, 147 42, 51, 59, 64–5, 67, 69, 77, 424, 432 linguistic semiosis 359 system network 14, 83–4, 93, 118–19, multimodal semiosis 157 185–6, 189–90, 192, 198, 200, semiotic action see ARC 202–3, 210–12, 215, 219, 221–2, semiotic distance 24, 29, 49, 214, 382 254, 301, 315, 422–3, 427 semiotic metaphor 256 semiotic risk 134 semiotics 158, 236, 259 teacher development 362 signification 146 tenor (of discourse) 14, 63, 143, social action 144, 147, 160, 169, 198, 147–50, 154, 160–2, 168–71, 174, 227, 233–4 190–5, 197, 198, 208–9, 213, 218, social activity 151, 157, 198–200, 202, 232, 244 206, 215–17, 219, 221–2, 224, 231, text-centric focus (view) 234 340, 359, 362 textual (meaning) 31, 40, 65, 67, 90, see also academic activity 93, 109 n. 8, 119, 126, 173, 176–8, social dialect 108, 389, 391 180, 235, 301–3, 357, 365 social distance 62,151, 160, 187, 189, texture 3, 11, 14, 33–4, 42–3, 49, 144, 192–5, 213–14, 232, 388 148–9, 152, 252, 300–2, 310, see also personal distance 312–16, 318, 324, 387, 423, 425, 428 social semiotic stylistics (SSS) 59, 63, 77 Theme (Given and New) 49, 324 sociolinguistics 108, 153, 167–8, 170–2, Theme constellation 51 389, 405, 416, 429–30 theoretical dynamism 240 socio-material world (context) 237, 239 therapeutic conversation 317–18, 425 socio-semiotic approach 208, 212, 214 token 31, 304 socio-semiotic distance 214 central token (CT) 304, 306 socio-semiotic process 211 peripheral token (PT) 304, 306, 321 solidary relations 105, 107 relevant token (RT) 304, 306 specialized 198, 210, 220 total token (TT) 306 see also quotidian tone (intonation) 8, 22 n. 3 speech accommodation theory 207 tourist brochures 256, 269 speech fellowship 145 tourist site entry ticket 261, 264–7, sphere of action 150, 161, 210, 215, 269–71 220 transitivity 11–12, 20, 70, 74, 203 n. 5, stage 261–2, 266, 269, 271 300, 303, 326 n. 20, 370, 423 stratification 116, 135 n. 2, 212, 235–40 trinocular perspective 121, 172 stylistics 28–9, 41, 59, 62, 77, 409, 429 tripartite entity 227, 234 Subject 176–7, 195, 370, 376 n. 2 see also context of situation subject matter 77, 177, 198, 206, 286 tri-stratal 135 n. 3 subliminal patterning 35 typology (systemic functional) 120, see also latent patterns 122, 135, 210–11 substitution 302, 303, 305, 320 typology of literacy education see also cohesion; ellipsis action literacy education 336, 338, SUBTEXT (see also PRIMARY TEXT) 360 151–2, 163 n. 5 recognition literacy education 335–6, syllable 22 n. 3, 41, 133 359 syntactic imagery 67 reflection literacy education 333–4, SYSconcordance 31 339–40, 342, 347, 351–2, 354 n. 3, systemic paradigm 240 359–62, 364, 367–9, 374–5 452 Subject Index unimodality 234 verbal art manquée 68 see also multimodality verbalization 27, 64 unity of structure /unity of texture 148, via-text context 175 252, 300–1, 313 see also VIA-TEXT relevant contextual Urdu 25, 37, 207, 382–4 features; pre-text context utterance (pragmatic/mathetic) 5, 20 VIA-TEXT relevant contextual features see under context valeur 25, 37–8, 146, 416 virtual addressee 194 valuation 34–5, 209, 407 voice 71, 109 n. 6, 160, 171, 187, 189, variation 388–9, 403 191, 215, 303, 358–9, 361–3, 375 verbal action (in field) 150, 215–19, tone of voice 325 n. 8 221, 224 voiceless 403 see also ancillary; constitutive verbal art 3, 9–10, 24–30, 34, 40–2, 49–51, 59, 60–4, 67–9, 77–9, 160, wording unit 387 181 n. 3, 322, 419, 424–5, 427–9, 432 Zone (in composition of an image) 257