The Routledge Handbook of Language and Dialogue 9781138808584, 9781138808591, 9781317612582, 9781317612575, 9781315750583

Table of contents :
Cover
Half Title
Series Page
Title Page
Copyright Page
Table of Contents
List of Illustrations
Figures
Tables
List of Contributors
Preface
Turning Points in Linguistics: From Language to Language Use and Dialogue
Part 1 The State of the Art
To Sum Up
References
Chapter 1 Pragmatics: From Language as a System of Signs to Language Use
1.1 Introduction
1.2 Twenthieth Century: Structuralism Including Generativism
1.3 Peirce and Formal Semantics
1.4 Pragmatics
1.5 Non-structuralist Linguistics
1.6 The Relationship of Pragmatics to Context
1.7 Post-structuralist Linguistics
1.8 Outlook
References
Chapter 2 Conversation Analysis
2.1 Introduction
2.2 Origins and History
2.3 Methodology
2.4 Key Notions
2.5 Current and Future Trends: Embodied Conducts in Interaction
2.6 Conclusion
Transcription Conventions
References
Chapter 3 Corpus Linguistics
3.1 Introduction
3.2 (Spoken) Corpora, Corpus Annotation and Corpus Tools
3.3 Corpus Linguistics and Language in Use
3.4 Interactivity in Spoken and Written Discourse
3.5 Corpus Linguistics and Language Variation across Speech and Writing
3.6 Conclusions
Note
References
Chapter 4 Discourse Analysis
4.1 Background
4.2 Towards a ‘Unified Theory’ of Discourse Analysis
4.3 Extending Linguistic Discourse Analysis
4.4 Design
4.5 Conversation (Extended Sense)
4.6 Face-to-Face Verbal Conversations
4.7 Discourse Analysis and Video Games
4.8 Remaining Issues
References
Chapter 5 From Pragmatics to Dialogue
5.1 Introduction
5.2 The Dialogue Approach
5.3 Current Pragmatic Theories
5.4 Why Do We Need the Dialogue Approach?
5.5 Salience Effect in the Dialogic Sequence
5.6 Conclusion
References
Chapter 6 Psycholinguistic Approaches
6.1 Introduction
6.2 Traditional Psycholinguistics: Language-as-Product vs. Language-as-Action
6.3 Toward a New Psycholinguistics: Bridging the Traditions
6.4 Conclusion
Acknowledgement
References
Chapter 7 Intersubjectivity in Dialogue
7.1 Introduction
7.2 Intersubjectivity and Alterity
7.3 Alterity, Polyphony, Heteroglossia
7.4 Intersubjectivity and Trust
7.5 Forms of Intersubjectivity: A Developmental Theory
7.6 The “Problem of Other Minds”
7.7 Intersubjectivity as a Property of Interpersonal Interaction
7.8 Intersubjectivity, Self and Internal Dialogue
7.9 Interactivities Are More Basic Than Intersubjectivities
7.10 Intersubjectivity and Subjectivity
7.11 Intersubjectivity and Objectivity
7.12 Conclusion
Acknowledgements
Notes
References
Chapter 8 Dialogue and Literature
8.1 Introduction
8.2 Dialogue in Literature
8.3 Literature as Dialogue
8.4 Literary Dialogicality and Communicational Ethics
References
Chapter 9 Computational Approaches to Dialogue
9.1 Introduction
9.2 Human Computer Dialogue
9.3 Dialogue Manager Organizing Principles
9.4 Issues for Dialogue Systems
9.5 Conclusions
Acknowledgements
References
Chpater 10 From Speech Act Theory to Dialog: Dialog Grammar
10.1 Introduction
10.2 The Traditional Speech Act Theory: Language as a System of Constitutive Rules
10.3 Dialog Grammar
10.4 From Intentions to Interactions: Habermas’ Theory of Communicative Action
10.5 Tying Up Loose Ends: Weigand’s Theory of Dialog
References
Chapter 11 The Mixed Game Model
11.1 The Issue
11.2 What Constitutes a Theory
11.3 How to Cope with Complexity
11.4 Language as Dialogic Action
11.5 A Dialogic Speech Act Taxonomy
11.6 Minimal versus Complex Games
11.7 Conclusion
References
Part 2 Theoretical Key Issues and Open Debates
Reference
Chapter 12 Shifting Concepts of Language: Meeting the Challenge of Modelling Interactive Syntax
12.1 The Point of Departure: Grammar versus Language Processing
12.2 Shifting Perspectives: Problems in Modelling Natural-Language Semantics
12.3 Dynamic Syntax
12.4 Reflections
Notes
References
Chapter 13 The Concept of Language in an Utterance Grammar
13.1 The Issue: Shifts in the Concept of Language
13.2 Grammar and Performance
13.3 Fundamentals of an Utterance Grammar
13.4 The Structure of Meaning
13.5 Dialogic Means of Expression
13.6 Correlation of Meaning and Expression: The Core of Grammar
13.7 Speaking and Perceiving
13.8 Speaking and Thinking
13.9 Conclusion
References
Chapter 14 The Issue of Theorising: Object-of-Study and Methodology
14.1 Introduction
14.2 Dichotomy of Language
14.3 The Question of the Norm
14.4 Linguistic Methodology: General Questions
14.5 Linguistic Methodology: Language-Oriented Research
14.6 Communication-Oriented Research
14.7 Approaching Communication with Different Tools
14.8 Methodology-Driven Research Compared to Phenomenon-Driven Research
14.9 Conclusion
References
Chapter 15 Theory and Practice
15.1 Introduction
15.2 Theories of Dialogue
15.3 Applied Linguistics: Dialogue in the Practice of Teaching and Learning
15.4 Examples That Question the IRF
15.5 Conclusion: Applying the Holistic Theory to the Practice
Notes
References
Chapter 16 The Sociobiology of Language
16.1 Introduction
16.2 Mirror Neurons
16.3 The Anatomo-Functional Hypothesis
16.4 Neuroscience Studies on Neural Mirroring and Speech
16.5 Correlation and Causation: TBS and the Simulation Model
16.6 Conversation, Imitation, Empathy, and Embodied Concepts
16.7 The Need for Implicit Control and the Study of Stimulus-Driven, Synchronized Brains
16.8 Future Directions and Concluding Remarks
References
Part 3 Components of Dialogic Interaction: Human Nature, Institutions and Cultures
References
Chapter 17 Emotion, Reason and the Human Brain
17.1 Introduction
17.2 Historical Research (Seventeenth and Eighteenth Century)
17.3 Descartes’ Error – Descartes View of the Brain
17.4 Damasio’s Approach (1994) – Emotions and Reason in the Brain
17.5 Language Analysis and Emotions
17.6 The Holistic Model MGM By Edda Weigand
17.7 Conclusions
Notes
References
Chapter 18 Self-interest and Social Concerns
18.1 Introduction
18.2 Mixed Game Theory
18.3 Persuasion
18.4 Argumentation
18.5 Audience
References
Chapter 19 Language as the Originative House of Dialogic Ethics
19.1 Introduction
19.2 Ethics and Linguistic Inquiry
19.3 Ethics in Dialogue
19.4 Linguistic Contextual Weight as Call of Responsive Action
19.5 Discerning What Matters: Linguistic Weight and Height
Note
References
Chapter 20 Dialogue in Institutions
20.1 Introduction
20.2 Institutional Action Games: Formal and Informal Communication
20.3 The Functions of Institutional Communication
20.4 Institutional Knowledge
20.5 A Theoretical Framework for Institutional Action Games
20.6 The Influence of Culture
20.7 Conclusion
References
Chapter 21 Dialogue and the Law
21.1 Introduction
21.2 Language and Law
21.3 Complexity
21.4 Position Games
21.5 Consensus Games
21.6 Structured Thinking
References
Chapter 22 How Culture Affects Language and Dialogue
22.1 The Impact of Culture
22.2 Concepts of Culture, Language and Dialogue
22.3 The Mixed Game Model By Edda Weigand
22.4 Culture Within the MGM
22.5 Conclusions
Note
References
Outlook
Towards the Unity of Knowledge
References
Index

Citation preview

i

The Routledge Handbook of Language and Dialogue

The Routledge Handbook of Language and Dialogue is the first comprehensive overview of the emerging and rapidly growing sub-discipline in linguistics, Language and Dialogue. Edited by one of the top scholars in the field, Edda Weigand, and comprising contributions written by a variety of likewise influentials, the Handbook aims to describe the history of modern linguistics as reasoned progress leading from de Saussure and the simplicity of artificial terms to the complexity of human action and behaviour, which is based on the integration of human abilities such as speaking, thinking, perceiving, and having emotions. The book is divided into three sections: the first focuses on the history of modern linguistics and related disciplines; the second part focuses on the core issues and open debates in the field of Language and Dialogue and introduces the arguments pro and contra certain positions; and the third section focuses on the three components that fundamentally affect language use: human nature, institutions, and culture. This handbook is the ideal resource for those interested in the relationship between Language and Dialogue, and will be of use to students and researchers in Linguistics and related fields such as Discourse Analysis, Cognitive Linguistics, and Communication. Edda Weigand is Professor Emeritus of Linguistics at the University of Münster, Germany. She is Honorary President and Founding Vice-President of the International Association for Dialogue Analysis (IADA, Bologna) and has been elected Assistant Secretary-General to the Committee of UNESCO’s Fédération Internationale des Langues et Littératures Modernes. She is chief editor of the journal Language and Dialogue and the Dialogue Studies series (Benjamins). On the basis of a notion of ‘language as dialogue’ she developed a dialogic speech act theory and a holistic theory of human action and behaviour in performance in her book Dialogue: The Mixed Game (2010).

Routledge Handbooks in ­Linguistics

Routledge Handbooks in Linguistics provide overviews of a whole subject area or sub-discipline in linguistics, and survey the state of the discipline including emerging and cutting edge areas.  Edited by leading scholars, these volumes include contributions from key academics from around the world and are essential reading for both advanced undergraduate and postgraduate students. The Routledge Handbook of Systemic Functional Linguistics Edited by Tom Bartlett, Gerard O’Grady The Routledge Handbook of Handbook of Heritage Language Education From Innovation to Program Building Edited by Olga E. Kagan, Maria M. Carreira and Claire Hitchins Chik The Routledge Handbook of Language and Humor Edited by Salvatore Attardo The Routledge Handbook of Language and Dialogue Edited by Edda Weigand The Routledge Handbook of Language and Politics Edited by Ruth Wodak and Bernhard Forchtner The Routledge Handbook of Language and Media Edited by Daniel Perrin and Colleen Cotter The Routledge Handbook of Ecolinguistics Edited by Alwin F. Fill and Hermine Penz Further titles in this series can be found online at www.routledge.com/series/RHIL

The Routledge Handbook of Language and Dialogue

Edited by Edda Weigand

First published 2017 by Routledge 711 Third Avenue, New York, NY 10017 and by Routledge 2 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN Routledge is an imprint of the Taylor & Francis Group, an informa business © 2017 Taylor & Francis The right of Edda Weigand to be identified as the author of the editorial material, and of the authors for their individual chapters, has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging in Publication Data Names: Weigand, Edda, editor. Title: The Routledge Handbook of language and dialogue / edited by Edda Weigand. Description: New York, NY: Routledge, [2017] | Series: Routledge Handbooks in Linguistics | Includes bibliographical references and index. Identifiers: LCCN 2016054610 | ISBN 9781138808584 (hardback) | ISBN 9781138808591 (pbk.) | ISBN 9781317612582 (epub) | ISBN 9781317612575 (mobipocket/kindle) Subjects: LCSH: Dialogue analysis—Handbooks, manuals, etc. | Language and languages—Handbooks, manuals, etc. | Linguistic analysis (Linguistics)—Handbooks, manuals, etc. Classification: LCC P95.455 .R69 2017 | DDC 401/.41—dc23 LC record available at https://lccn.loc.gov/2016054610 ISBN: 978-1-138-80858-4 (hbk) ISBN: 978-1-315-75058-3 (ebk) Typeset in Times New Roman by Deanta Global Publishing Services, Chennai, India

Contents

List of Illustrations List of Contributors Preface: Turning Points in Linguistics: From Language to Language Use and Dialogue

viii x xv

Part I

The State of the Art

1

1

Pragmatics: From Language as a System of Signs to Language Use Jens Allwood

9

2

Conversation Analysis Lorenza Mondada

26

3

Corpus Linguistics Marina Bondi

46

4

Discourse Analysis James Paul Gee

62

5

From Pragmatics to Dialogue Istvan Kecskes

78

6

Psycholinguistic Approaches: Meaning and Understanding Susan E. Brennan and Joy E. Hanna

93

7

Intersubjectivity in Dialogue Per Linell

109

8

Dialogue and Literature Roger D. Sell

127

9

Computational Approaches to Dialogue David Traum

143

v

Contents

10 From Speech Act Theory to Dialog: Dialog Grammar Sebastian Feller

162

11 The Mixed Game Model: A Holistic Theory Edda Weigand

174

Part II

Theoretical Key Issues and Open Debates

195

12 Shifting Concepts of Language: Meeting the Challenge of Modelling Interactive Syntax Ruth Kempson

197

13 The Concept of Language in an Utterance Grammar Edda Weigand

214

14 The Issue of Theorising: Object-of-Study and Methodology Arto Mustajoki

234

15 Theory and Practice Dale A. Koike

251

16 The Sociobiology of Language: What Mirror Neurons Can Tell Us Marco Iacoboni

264

Part III

Components of Dialogic Interaction: Human Nature, Institutions and Cultures

275

17 Emotion, Reason and the Human Brain Marion Grein

277

18 Self-interest and Social Concerns Jennifer L. Adams

294

19 Language as the Originative House of Dialogic Ethics Ronald C. Arnett

307

20 Dialogue in Institutions Sebastian Feller

318

vi

Contents

21 Dialogue and the Law Fritjof Haft

330

22 How Culture Affects Language and Dialogue Marion Grein

347

Outlook: Towards the Unity of Knowledge 367 Index369

vii

Illustrations

Figures

  4.1  A Screenshot from Thomas Was Alone  10.1 Action Structure   10.2  Minimal Speech Act Sequence   10.3  INSIST Structure. Based on Hundsnurscher, 2007: 69   10.4  Ideal Bargaining Dialog   11.1  Components of a Post-Cartesian Theory   11.2  Specialization of Subsystems in a Complex Hierarchy   11.3  The Action Principle   11.4  The Classical Speech Act Formula   11.5  The Extended Speech Act Formula   11.6  The Dialogic Principle Proper   11.7  Restructuring the Three-Part Analysis   11.8  The Basic Two-Part Sequencing Principle   11.9  The Principle of Coherence 11.10  The Balance of Self-interest and Respect in Different Cultures 11.11  Dialogic Typology of Basic Action Games 11.12  Derived Types of Directives 11.13  Derived Representative Types of the Simple Claim to Truth 11.14  Derived Representative Types of the Modal Claim to Truth 11.15  Differentiating the Proposition 11.16  The ‘Commissive’ Sequence 11.17  Exploratives versus Directives 11.18  Embedded Clarifying Sequence 11.19  Argumentation in a Representative Game 11.20  The Action Game of Planning   12.1  Processing Example 23: Input and Output   12.2  Building Unfixed Nodes : *Adjunction   12.3  Modelling an Utterance Split   12.4  Building and Resolving an Unfixed Node: Long-Distance Dependency   13.1  Correlation of Meaning and Expression in an Utterance Grammar   13.2  Meaning Structure of the Speech Act   13.3  Correlation of Meaning and Means viii

72 164 166 167 168 177 179 181 181 181 182 182 182 183 184 187 188 189 189 189 190 191 191 192 192 204 206 208 209 215 219 219

Illustrations

 13.4 Comparative Grammar 220   13.5  The Universe of Meaning 221   14.1 Source of Information in Linguistic Research (The Two States of Language, Langue and Parole)239   14.2  Main Constituents Influencing the Course of Communication 241   14.3  Source of Information in Communication-Oriented Research 244   20.1  Components of the Institutional Action Game 327   22.1  Monologic vs. Dialogic View 354   22.2  Communicative Function and Variation of Utterances 355   22.3  Communicative Function Request and Variation of Utterances 355   22.4 Communicative Function and Variation of Utterances (Direct, Indirect, Idiomatic) 356  22.5 Action Principle 356   22.6  Revised Action Principle 356  22.7 Dialogic Principle 356   22.8  Holistic Concept of the Action Principle 361   22.9  Example Dialogic Principle (Request) 362 Tables

14.1  Basic Elements of Language as a Research Object 22.1  Example Analysis Action Principle 22.2  Subtypes of representatives

237 357 358

ix

Contributors

Jennifer L. Adams is Associate Professor in the Department of Communication and Theatre at DePauw University, USA. Her research interests include rhetoric and language, with a focus on historic public discourse. Jens Allwood is Professor of Linguistics and Director of the interdisciplinary center SCCIIL at the University of Gothenburg, Sweden. His research interest is mainly in the intersection of the six concepts which make up SCCIIL (Semantics, Cognition, Communication, Information, Interaction and Language). He is the author of more than 250 peer reviewed articles, book chapters and books and is editor-in-chief of the Journal of Intercultural Communication, besides serving on the editorial board of several scientific journals. Ronald C. Arnett (PhD, Ohio University, 1978) is Professor and Chair of the Department of Communication and Rhetorical Studies and the Patricia Doherty Yoder and Ronald Wolfe Endowed Chair in Communication Ethics at Duquesne University, USA. He is the author or coauthor of ten books and the recipient of five book awards. His most recent works are Levinas’s Rhetorical Demand: The Unending Obligation of Communication Ethics (in press, Southern Illinois University Press) and Communication Ethics in Dark Times: Hannah Arendt’s Rhetoric of Warning and Hope (2013, Southern Illinois University Press). He is in his third editorship of the Journal of Communication and Religion. Marina Bondi is Professor of English Linguistics at the University of Modena and Reggio Emilia, and Director of the CLAVIER centre (Corpus and Language Variation in English Research). She has published extensively in the field of genre analysis, English for Academic Purposes (EAP) and corpus linguistics, with a focus on argumentative dialogue and language variation across genres, disciplines and cultures. She is on the Advisory Board of the series Studies in Corpus Linguistics (Benjamins) and on the editorial boards of Journal of English for Academic Purposes, Asian ESP Journal, Profile, Revista de Lenguas para Fines Especificos (Journal of Languages for Specific Purposes), Language and Dialogue, and ESP Across Cultures. Susan E. Brennan, PhD, is Professor in the Cognitive Science Program of the Departments of Psychology, Computer Science, and Linguistics at Stony Brook University, USA. As a cognitive scientist she studies the psychology of language use—in particular, interactive spoken dialogue—and has published widely in this field. Her recent interests include the neural circuits that support interactive communication and the human use of computer technology. x

Contributors

Sebastian Feller has conducted research in the area of dialogue analysis. Following a lecturer position at the English Studies Department at Majan University in Muscat, Oman, he was appointed Research Scientist at the Institute of High Performance Computing in Singapore. There he worked in a number of different research projects including the Social Situation Awareness Project, in which he contributed to the conceptualisation and implementation of a tutor robot for language learning. Sebastian has published and edited books and journal articles on topics spanning from dialogue over learning to technology. He is currently working as an English teacher in Germany and is launching Growth, his own education consultancy (www.learn-with-growth.com). Sebastian obtained his PhD in theoretical linguistics at the University of Muenster, Germany, in 2010. James Paul Gee is the Mary Lou Fulton Presidential Professor of Literacy Studies and a Regents’ Professor at Arizona State University. He is a member of the National Academy of Education and serves on a number of editorial boards for journals devoted to language, learning, literacy, or digital media. He has published widely on discourse analysis, learning and literacy, and digital media and education. Marion Grein is Associate Professor and Director of the Master’s Programme in German as a Foreign Language at the Johannes Gutenberg-University in Mainz, Germany. The focus of her work is research into language teaching (‘neurodidactics’), cross-cultural studies, and dialogue studies as well as the use, deployment and effectiveness of multimedia and e-learning. She is currently researching learning processes in older learners and the influence of emotions on learning. She is a member of the Advisory Board of Dialogue Studies, a member of the Editorial Board of Language and Dialogue and advisor of the Hueber Verlag, Germany, which specialises in language teaching books. Fritjof Haft is Professor Emeritus at the University of Tübingen, Germany. He held the chair for criminal law, philosophy of law, and jurimetrics from 1982 to 2005. There he founded the Tübingen Negotiation Seminar and was involved in establishing the Alternative Conflict Resolution (ADR) in Germany. In addition to his activities as professor, he worked as criminal defense counsel. In 1999 he founded the Normfall GmbH in Munich. As its CEO, he developed software with the focus on legal activity. In 2015, he won a Europe-wide tender of the state of Bavaria with his software Normfall Manager, which is now being used by several thousand users in courts, authorities and firms all over Germany, Austria and Switzerland. In the course of his work, he published several hundred scientific publications. Joy E. Hanna is Assistant Professor of Psychology at Daemen College, USA. She earned her PhD in Brain and Cognitive Sciences at the University of Rochester in 2001. Hanna’s research focuses on the interaction of linguistic and non-linguistic information on human language processing. Marco Iacoboni, MD, PhD, is Professor in the Department of Psychiatry and Biobehavioral Sciences in the Semel Institute for Neuroscience and Human Behavior at the David Geffen School of Medicine at UCLA. He is also a member of the UCLA Brain Research Institute and Director of the Neuromodulation Lab at the Ahmanson-Lovelace Brain Mapping Center (http://www.bmap.ucla.edu). His research focuses on the use of brain imaging and non-invasive neuromodulation to study neural plasticity and repair in humans, and neural xi

Contributors

systems and mechanisms supporting social cognition. He has authored more than a hundred peer-reviewed original research articles, more than 50 book chapters, and is the author of Mirroring People, a general reader book on mirror neurons research, including his own. Istvan Kecskes is Distinguished Professor of the State University of New York System, USA. His research interest is in pragmatics, second language acquisition and bilingualism. He is the President of the American Pragmatics Association (AMPRA) and the CASLAR (Chinese as a Second Language Research) Association. Dr. Kecskes is the founding editor of the journal Intercultural Pragmatics and the Mouton Series in Pragmatics, as well as the bilingual (Chinese–English) journal CASLAR (Chinese as a Second Language Research) published by Mouton. He sits on the editorial board of various journals, among them the Journal of Pragmatics, Pragmatics & Society and the International Journal of Multilingualism. Dr. Kecskes is also on the editorial board of book series, for instance, Pragmatics, Philosophy and Psychology, and Language and Dialogue. Ruth Kempson, FBA, is Emeritus Professor of Linguistics of King’s College London and Honorary Research Associate with the Cognitive Science research unit of Queen Mary University of London and with the School of Oriental and African Studies. She is the lead developer of the Dynamic Syntax framework, and has a long-term research interest in the interface of syntax and pragmatics. She is on the editorial board of Glossa. Dale A. Koike is Professor of Hispanic and Portuguese Linguistics at the University of Texas at Austin, USA. Her research interests include Discourse Analysis and Pragmatics, in both native Spanish and Portuguese speaker contexts, and as applied to sociolinguistic issues and Second Language Acquisition. She is the editor or co-editor of several volumes on Pragmatics and Dialogue, and co-authored a textbook on Spanish Applied Linguistics. She has published in venues such as the Journal of Pragmatics, Modern Language Journal, Foreign Language Annals, and Journal of Spanish Language Teaching, and is a new series editor for Routledge. Per Linell is Professor Emeritus in Communication, Language and Culture from the Universities of Linköping and Göteborg, Sweden. He took his PhD in linguistics (Uppsala, 1974), and in the years 1981–2006 he was engaged in research and research training in the interdisciplinary unit of Communication at Linköping. His main areas of expertise are dialogical theory, and studies of institutional, professional and professional–client discourse. He has authored about 120 peer-reviewed papers, and several books, including The Written Language Bias in Linguistics (Routledge, 2005), Rethinking Language, Mind and World Dialogically (IAP, 2009) and Memory Practices and Learning: Interactional, Institutional, and Sociocultural Perspectives (co-edited with Åsa Mäkitalo and Roger Säljö; IAP, in press). Lorenza Mondada is Professor for Linguistics at the University of Basel, Switzerland, and Finland Distinguished Professor at the University of Helsinki. Her research deals with social interaction in ordinary, professional and institutional settings, within an ethnomethodological and conversation analytic perspective. She has extensively published in Journal of Pragmatics, Discourse Studies, Language in Society, ROLSI and co-edited several collective books (for CUP, De Gruyter, Benjamins, Routledge) on multimodality, mobility, multiactivity, and knowledge in interaction. xii

Contributors

Arto Mustajoki is Professor of Russian and Dean of the Faculty of Arts at the University of Helsinki, Finland. His research interests include contemporary Russian, linguistic theory, corpus linguistics and functional syntax. Recently he has also carried out research on miscommunication, the Russian mind and societal impact of research. He is a member of the Editorial Council of the journals Questio Rossica (Ekaterinburg) and Zhanry retchi (Saratov). Roger D. Sell is Emeritus H.W. Donner Research Professor of Literary Communication at Åbo Akademi University, Finland. He has written widely on several different periods of English and American literature, and has developed a re-humanising account of literature as one among other forms of communication. He is Honorary President of UNESCO’s Fédération Internationale des Langues et Littératures Modernes (FILLM) and Founding Series Editor of FILLM Studies in Languages and Literatures (John Benjamins). He is an Officer of the British Empire, a Knight First Class of the White Rose of Finland, and a Member of the Humanities Section of the Finnish Society of Sciences and Letters. David Traum is the Director of Natural Language Research at the Institute for Creative Technologies (ICT) and a research faculty member of the Department of Computer Science at the University of Southern California (USC), USA. He leads the Natural Language Dialogue Group at ICT (http://nld.ict.usc.edu/group/). Traum’s research focuses on Dialogue Communication between Human and Artificial Agents. He has authored more than 200 refereed technical articles, is a founding editor of the Journal Dialogue and Discourse, and is a past President of SIGDIAL, the international special interest group in discourse and dialogue. Traum earned his PhD in Computer Science at the University of Rochester in 1994.

xiii

This page intentionally left blank

Preface

Turning Points in Linguistics: From Language to Language Use and Dialogue The Routledge Handbook of Language and Dialogue aims to give an overview of the whole history of modern linguistics, starting with de Saussure and arriving at the present state of the art. The goal is not only to describe but also to explain what happened. Explaining presupposes that the historical sequence of schools in modern linguistics not only represents an arbitrary sequence of events but can at least to some extent be justified by internal reasons. The reasons are considered to be turning points of reflection, and evidence that certain schools of thought have come to an end, thus compelling scientists to change direction if they wanted to move forward in their search for progress. To meet such a goal represents a huge challenge, considering the multiplicity of approaches developed during the twentieth and the beginning of the twenty-first centuries. Plurality of models, however, cannot mean giving up reasoning. On the contrary, a plurality of models means we have to reflect, to compare and to ask, for instance, why some models can be called orthodox or outdated from our current point of view, or why others are not well-founded or inconsistent. Such a critical stance constitutes the thread that runs through the whole handbook. I think that progress in science has indeed been made, even if not immediately and clearly visible and often called into doubt. Having experienced the sequencing of different schools in my academic life, from structuralism via generative grammar to pragmatics and dialogue analysis, I personally could observe how underlying questions changed and provoked a paradigm change in theory and analysis. These turning points are the pillars that bear the structure of this handbook. Obviously these pillars depend on our view of the object ‘language and dialogue’. The whole history of modern linguistics can be seen as a history of attempts to grasp this object which first meant language in abstraction, then proceeded to a concept of language as natural language use and eventually addressed use as dialogic action. These changes in the object are, of course, connected with changes in methodology and the basic question of the correlation between object and methodology. What should the starting point of science be: methodology or the object? Defining the object by abstraction results in an artificial methodology. When we are confronted with a natural object of use, that means that we are confronted with multiple purposes of use and an extreme variability of empirical means. This calls for a change in methodology from reductionism to holism, from constructing patterns to uncovering the ‘architecture of complexity’. The basic impulse that kept me alive in the scientific struggle between different schools was always the desire and pleasure of finding what constitutes language. In this search for xv

Preface

an adequate answer to this crucial question of linguistics, I met Franz Hundsnurscher, who was my teacher for a long time and a life-long friend. I am very grateful to him for the neverending discussions we had on fundamental questions even if, at some point, I left his way of thinking in patterns and started my own way in the adventure of the complex. It goes without saying that I am also grateful to many other colleagues and friends. I will not attempt to list their names as it would of necessity be an open list. I would like to thank the community of scientists who are truly interested in their subject matter, many of them members of the ‘dialogue family’ but also many others from other disciplines who helped me find my way in the stormy sea of competitive approaches. In the end the fundamental question of what constitutes language requires us to go beyond narrow disciplinary boundaries and join genuine like-minded friends, all searching for the unity of knowledge. Our object is not only language but language embedded in human dialogic action and behaviour. Finally, I want to extend a particular thank you to Kathrene Binag and her colleagues from the Routledge Publishing Company, for their precision, efficiency and kindness in carrying through the formal editorial and printing process. Edda Weigand October 2016

xvi

Part I

The State of the Art

Part I of the Handbook delineates the sequence of schools which more or less followed each other in the twentieth and the beginning of the twenty-first centuries. The order of the chapters is determined by consideration of the scientific interests that have brought about changes in the objects-of-study. All these schools can be seen as attempts to grasp the object ‘language’. To some extent they run into each other and cannot always be strictly kept apart. These attempts did achieve partial results and paved the way, step by step, towards greater insight into the complexity of human life. The Handbook represents a collective volume which gives the floor to different voices. These voices are on the one hand representative of their field but, of course, they express individual views. Differences of opinion remain which reflect science as a field of argumentation and discussion. Nonetheless, the chapters clearly mark paradigm changes in the history of modern linguistics. We start with an overview by Jens Allwood of the first period from structuralism to pragmatics, from language as a system of signs to the concept of language-in-use. Having studied and analysed language as an abstract system over decades, the interest of the researchers in artificially constructing language diminished and turned instead to language as it is used, not an object on its own but an object dependent on the user. In use there are no signs which have meaning on their own. On the contrary, there are communicative means whose meanings differ dependent on the individual user and the specific environment they are used in. The change to the new object ‘language use’ constituted the pragmatic turning point. Changing the object requires a change in methodology, a challenge pragmatics is still struggling with. This first period, based on a system of signs, mainly comprises the structuralistic and generative schools. Both focused on the ‘sentence’, which was analysed by structuralistic approaches as a static system, and by generativists as a dynamic system of production. Language use, however, does not stop with the sentence nor is it based on rule-governed ‘sentences’ at all. Language use consists of ‘utterances’. Looking beyond the sentence we were struck by the complexity that opened up. Beyond the sentence there was the world. We did not even know what ‘language use’ meant and considered it an illuminating insight to

1

Edda Weigand

discover that ‘language use’ meant ‘natural language use’ and ‘natural language use’ meant ‘authentic language use’. However, how was the new concept of language as ‘authentic language use’ to be analysed? In principle, two possibilities opened up, first, searching ‘in the underground’, and second, scrutinizing empirical details at surface level. Searching in the underground meant focusing on the meaning of ‘sentences-in-use’, which was heavily influenced by language philosophy and Austin’s book How to Do Things with Words. Austin and later Searle drew our attention to the crucial insight that we not only speak with words but act with words in utterances. Language-use means language-action. This decisive insight, however, meant clarifying the relationship between communicative means of use and their purposes in action. After a first wave of approaches based on philosophical action theory, interest diminished. Searle’s speech act theory was considered to be too abstract and not directly applicable to authentic language use. However, instead of making efforts to modify and change it in order that it should become apposite to dialogic performance, researchers turned their interest upside down; they went from analysing meaning to analysing the means at the empirical surface level. Pragmatics splits up into various types of approaches all trying to come to grips with the empirical details of ever-changing authentic language use. ‘Data’ and ‘discourse’ became the core terms of these attempts. Any empirical element was picked out and investigated meticulously as so-called ‘data’, although half a century ago Chomsky (1959) provided a thorough critique of this type of approach in his review of Skinner. Empirical elements can count as ‘data’ only if they can be meaningfully related to the purpose of language use. But even genuine ‘data’ do not provide the key to the complex whole of language use. Another strong argument against devoting all our energies to any empirical detail whatsoever can be derived from the way we as competent users of language deal with them in performance by simply passing over them. The concept of language changed due to the focus on ‘data’ of authentic language use. Language was equated with spoken language and ‘text’ as a sequence of sentences became ‘discourse’ as a sequence of utterances. At the beginning of this process, ‘discourse’ still meant ‘text’ or the sequence of sentences, for instance, with Zellig S. Harris (1952). As soon as the verbal level of text no longer sufficed for the claims of performance analyses, the term ‘discourse’ developed into a concept which even included the cognitive world of the users. ‘Discourse analysis’ became a “pan-discipline” (Van Dijk, 1985: 10) which claimed to encompass everything connected with language use and embraced various diverging schools under the premise ‘anything goes’. The criterion of consistency “of a homogeneous ‘discipline’ with a unified theory” was declared to be “pointless” (Tannen, 1989: 7f.). The catchword of the ‘plurality of models’ was welcomed by many scholars as it easily allowed unlimited access to a view of language which was declared to be mainstream linguistics. In effect the ‘plurality of models’ only means eclecticism and arbitrarity as Frawley (1987: 363ff.) very early critically pointed out. The integration of perceptual means, such as gestures or body movements, was crucial to the emergence of a new type of approach, the multimodal analysis. Lorenza Mondada describes this development, starting from conversational analysis which was primarily introduced in its exclusively formal version by Sacks, Schegloff and Jefferson (1978). She shows that the change to the multimodal version is based on the insight that the empirical level not only consists of verbal means but includes multimodal means of speaking and perceiving in interaction. This insight was an important step forward to a view of language use based on the integration of different human abilities. 2



The State of the Art

Among the schools focusing on the empirical component of language use there is also the school of corpus linguistics which—as Marina Bondi describes—provided a highly valuable tool for describing the multiple ways of use of verbal expressions. Corpus linguistics does not aim to demonstrate what lies below the verbal surface but describes it and justifies verbal evidence by its frequency. This step reveals important insights into the system of language as a network of language-specific phrases or collocations. Consequently, lexical semantics can no longer be based on structuralistic terms such as polysemy and disambiguation, or on an artificial system of single words but needs to take account of the way words are used in phrases. Corpora of written and oral language provide a vast amount of material of authentic language use in different languages far beyond the reach of any artificial language system based on the competence of the ideal speaker. Discourse analysis is the topic of James P. Gee’s chapter. This is only a brief outline of the broad field of different ‘discourse schools’ and mainly introduces Gee’s extended version of a theory of discourse which no longer puts the emphasis exclusively on the empirical surface but includes categories of action as well as various types of dialogic means. Besides discourse-based models there is another methodology within pragmatics which developed into a school of its own, the neo-Gricean approach, which believes in the addition of orthodox assumptions of truth-conditional semantics and individual inferences (see Chapter 5). It is not surprising that it got stuck in a jungle of terms and levels (e.g., Levinson, 2000). Attempts to settle the crucial problem in pragmatics of how to address language use in performance are manifold. Among them, there is a technique which simply uses one term for all the different cases, the term ‘pragmeme’ (Mey, 2001). This reminds us of the techniques of structuralism. However, whereas structuralist terms, for instance, the phoneme or morpheme, were coined as the result of clear operations defined by structuralist methodology, the term ‘pragmeme’ is applied as a circular classification to very different pragmatic issues. There are no internal criteria which define the term, and therefore it remains without any explanatory power. Such an attempt is doomed to fail from the very outset. The change in the object, from language competence to language use in performance, requires a change in methodology. The question is where to begin, with the object or with methodology. Whereas modern linguistics started with the dogma of language as a sign system, that is, with an abstract model of language, pragmatics and its natural object of language use resisted being dominated by pre-given terms. We should be careful not to distort the natural object by our methodology, as Martinet (1975) told us. A complex natural object requires us to start from the complex whole and to recognize that there is no language as such but human beings who use language in dialogic interaction. Even if pragmatics continues to focus on one utterance and to invent new terms, the insight that language use does not stop with the utterance can no longer be ignored. We approach the next turning point from pragmatics to dialogue, which Istvan Kecskes describes in his chapter. Kecskes can be considered as a forerunner among the pragmatists who emphasized that language use does not mean making utterances but acting and reacting in dialogue. He indicated the direction towards dialogue and towards a concept of language as action. Scholars of psycholinguistics opened up the view from the speaker who means something to the hearer who tries to understand or to interpret the speaker’s utterance. Meaning and understanding are at centre stage in the chapter by Susan E. Brennan and Joy E. Hanna. The concepts of meaning and understanding still refer to the speaker’s utterances and cut off the natural process of interaction, which is a process of coming to an understanding, 3

Edda Weigand

in which the hearer not only listens but changes to an interlocutor who reacts by means of their own action, verbal, perceptual by a gesture or in the mind. Per Linell points to the role of shared understanding and intersubjectivity and argues that understanding will mostly be only partially achieved. The direction towards dialogic interaction became more and more manifest. The view of ‘language as dialogue’ spread gradually and even extended to literary texts (Chapter 8). Whereas literary scholars often view literature as monologic texts, Roger D. Sell very early on considered literary texts as communication between the author and their audience. In his groundbreaking Communicational Criticism: Studies in Literature as Dialogue (2011) he assigned literature a specific dialogic quality of addressivity and thus strengthened the view that language in general is used dialogically. We need not develop different theories but can describe any use of language in a unified theory of ‘language as dialogue’ which, of course, manifests itself differently in different genres. The view of ‘language as dialogue’ also spread to computer linguistics as is described in Chapter 9. David Traum gives an overview of how computer linguists tried to come to grips with the considerable expansion of linguistic theories from an artificial system of signs to dialogic action and multimodality. It goes without saying that the problems are enormous. Encouraging progress has been made even if many issues, including a coherent computer linguistic model of the whole, have still to be settled. Taking the direction from pragmatics to dialogue finally led to the next turning point, which was achieved by recognizing that language use not only means speaking but acting. The challenge was to change Searle’s monologic speech act theory (1969) into an action theory capable of dealing with dialogic sequences. Action in general is defined by the correlation of purposes and means, be it verbal action or perceptual or cognitive action—just as, for instance, if we want to fell a tree, we have to look for an appropriate tool. The first question to be raised refers to the purpose of dialogic language use. The hearer changes to an interlocutor and the purpose of their dialogue is not only the hearer’s understanding but their mutual attempt to come to an understanding. The premises of this view were laid out by Weigand (1991): Language is primarily used for communication. Communication is always dialogic.

Thus the concept of ‘language as dialogue’ was born. We do not speak in order to express our thoughts. Of course, we can do that but expressing our thoughts is directed to our fellow beings and needs to be negotiated if their thoughts are different. Also, we do not speak in order to gain knowledge. Gaining knowledge might be included in the purpose of certain types of dialogue in which we ask for and provide information. Moreover, it is mostly not a self-sufficient purpose. We need knowledge in order to collaborate with our fellow beings in tackling the issues of life. Language use as dialogic action was first addressed by Franz Hundsnurscher in his influential article on ‘Konversationsanalyse versus Dialoggrammatik’ (1980). He argued against the empiricistic conversation analytic approach of turn-taking by Sack, Schegloff and Jefferson (1978) and introduced his counter-model of ‘dialogue grammar’. The arbitrary sequence of ‘not fixed’ turn taking at the level of verbal means is changed to conventional patterns of speech act sequences determined at the level of action function. Hundsnurscher did not change Searle’s concept of the speech act but added specific reactive speech acts dependent on the initial speech act. The focus on dialogic action inspired Weigand to 4



The State of the Art

develop a dialogic speech act theory (1989 [2003]) which changed Searle’s concept of the ­illocutionary speech act and replaced it by introducing two functionally different types of action, the initiative and the reactive action. On this basis, Weigand developed a comprehensive dialogic speech act typology starting from the general purpose of dialogue and reaching out to any particular speech act of the minimal two-part sequence of action and reaction. In this way finally the concept of dialogue has been functionally defined as sequence of action and reaction which constitutes the dialogic turning point. Both approaches, Hundsnurscher (1980) and Weigand (1989 [2003]), represent pattern approaches of communicative competence and remained within the Cartesian rule-governed methodology (Chomsky, 1966). In his chapter Sebastian Feller deals with the step “from speech act theory to dialogue: Dialogue grammar” and describes approaches which extend Searle’s monologic speech act theory towards an interactive action theory of communicative competence, among them also Habermas’s (1981) sociological interactive theory. Even if the insight that language use means dialogic action considerably improved our understanding of human action and behaviour, the pragmatic challenge of addressing authentic natural language use had not yet been fully met. Natural language use meant language use in performance, which included individual use. Traditionally, the goal of a theory is considered to restrict empirical variability to an underlying structure of rules. How should individuality be included in a theory? Conversational analysis and discourse analysis have offered case studies but failed to provide a justified theory of performance. This challenge meant going beyond Cartesian science and tackling the adventure of complexity. A theory of performance thus requires yet another turning point from reductionism to holism. ‘Complexity’ and ‘holism’ became catchwords in post-Cartesian science. Fortunately there have been a few studies which helped us gain insight into the complex whole, among them Herbert A. Simon (1962) and Edward O. Wilson (1999). They demonstrate that a holistic endeavour will, of necessity, be a genuinely interdisciplinary affair. There are a few premises which need to be met in order to understand the ‘architecture of complexity’, among them: • Complexity does not mean addition of parts but integration of components. • Components need to be derived from the complex whole. • A holistic theory has to start from the complex whole and derive the components by specialization. • Specialization requires that the components are nearly decomposable.

Applied to natural language use, the first premise refers to the complex whole which needs to be an autonomous object. In the end any object in the humanities and social sciences leads us to human beings and their abilities in coming to grips with the conditions of the social and cultural environment. They orient themselves according to rules as far as they go and are prepared to adapt to particular cases on the basis of principles of probability. It is their extraordinary ability of competence-in-performance which allows them to bridge the gap between competence and performance in their minds (e.g., Weigand, 2010: 4f.). Weigand’s Mixed Game Model (Chapter 11) represents a first comprehensive holistic model of a theory of competence-in-performance. The unit is the dialogic action game which is not a chess game but something like children’s games, played partly according to rules, to some degree competitive but always creative and as ‘a mixed game’ prepared to adapt to ever-changing action conditions by any means available. The theory of competence-in-performance makes the step from Cartesian linguistics to post-Cartesian thought, 5

Edda Weigand

from reductionism to holism, from a methodology of rules to a methodological basis of principles of probability. Any approach which claims to be a theory should be verified. Fortunately progress in neuroscience can verify how human abilities work, at least in their essential features. We finally can go beyond the restricted Cartesian view and accept our object as a complex whole which is not based on separation but on integration. The insight in the co-evolution of genes, mind and culture marks an important turning point in the human sciences (Lumsden and Wilson, 2005). It is not only the biology of the human brain but sociobiology which determines human behaviour and action. There has been a vast discussion about the human species which can be cut short by reference to sociobiology: human beings are by their very nature dialogic individuals. They are social beings but nonetheless have their own individual interests. It is this double nature which determines the negotiation of mutual goals in different action games dependent on the environment.

To Sum Up In the past hundred years of modern linguistics, schools of thought have changed and various types of approach have been developed in search of what actually constitutes language. Roughly speaking, it is a search which started with the sign and ended up with dialogue. Moving ahead and changing position is determined by the scientific ‘pleasure of finding things out’ and gaining increased insight into the complex (Feynman, 2001). History is more than a list of data. History means finding out the relevant events and how they are connected to form a meaningful picture. In this sense, describing history will inevitably contain an individual judgement. The sequence of schools described in Part I is characterized by a series of turning points which mark decisive changes in the direction of research. There is the important pragmatic turning point from language as an abstract system in structural and generative grammar to natural language use in pragmatics, which is also a turning point from approaches on competence to those on performance. Pragmatics has been struggling to bridge this gap over decades. There was no key to opening up the uncircumscribed field of authentic language use. The key was provided by the next turning point from pragmatics to dialogue. Dialogue not only represents an object but also includes a methodology with which to structure and analyse the object. Dialogue means action and reaction and requires a change from approaches which stick to the surface level of empirical means to approaches which are determined at the level of action function. The object language has been extended to ever-more complex units. In the end it is human beings who carry out actions which are dependent on the social and cultural environment. The focus on the co-evolution of genes, mind and culture means another fundamental turning point in methodology, from Cartesian patterns of rules to postCartesian thought based on principles of probability.

References Chomsky N (1959) Review: B. F. Skinner: Verbal behavior. New York 1957. Language 35: 2658. Chomsky N (1966) Cartesian Linguistics: A Chapter in the History of Rationalist Thought. New York: Harper & Row. Feynman RP (2001) The Pleasure of Finding Things out. London: Penguin Books. Frawley W (1987) Review article: Van Dijk (ed.). Handbook of Discourse Analysis, I–IV. Language 63: 361–397. 6



The State of the Art

Habermas J (1981) Theorie des kommunikativen Handelns, 2 vols. Frankfurt, Main: Suhrkamp. Harris ZS (1952) Discourse analysis. Language 28: 1–30. Hundsnurscher F (1980) Konversationsanalyse versus Dialoggrammatik. In: Rupp H and Roloff H-G (eds) Akten des VI. Internationalen Germanisten-Kongresses. Basel 1980, Part 2. Bern: Lang, 89–95. Levinson SC (2000) Presumptive Meanings: The Theory of Generalized Conversational Implicature. Cambridge: The MIT Press. Lumsden CJ and Wilson EO (2005) Genes, Mind, and Culture: The Coevolutionary Process. New Jersey: World Scientific. Martinet F (1975) Functional linguistics: La linguistique fonctionnelle. In: Martinet F (ed.) Studies in Functional Syntax: Études de syntaxe fonctionnelle. München: Fink. Mey JL (2001) Pragmatics: An Introduction, 2nd ed. Oxford: Blackwell. Sacks H, Schegloff EA and Jefferson G (1978) A simplest systematics for the organization of turn taking for conversation. In: Schenkein J (ed.) Studies in the Organization of Conversational Interaction. New York: Academic Press, 7–55. Searle JR (1969) Speech Acts: An Essay in the Philosophy of Language. Cambridge: Cambridge University Press. Sell RD (2011) Communicational Criticism: Studies in Literature as Dialogue. Amsterdam/ Philadelphia: Benjamins. Simon HA (1962) The architecture of complexity: Hierarchic systems. Proceedings of the American Philosophical Society 106: 467–482. Tannen D (1989) Talking Voices: Repetition, Dialogue, and Imagery in Conversational Discourse. Cambridge: Cambridge University Press. Van Dijk TA (1985) Introduction: Levels and dimensions of discourse analysis. In: Van Dijk TA (ed.) Handbook of Discourse Analysis, vol. II. London: Academic Press, 1–11. Weigand E (1989 [2003]) Sprache als Dialog: Sprechakttaxonomie und kommunikative Grammatik, 2nd, rev. edn. Tübingen: Niemeyer. Weigand E (1991) The Dialogic Principle revisited: Speech acts and mental states. In: Stati S, Weigand E and Hundsnurscher F (eds) Dialoganalyse III: Referate der 3. Arbeitstagung, Bologna 1990, vol. 1. Tübingen: Niemeyer, 75–104. (Reprinted in: Weigand E (2009) Language as Dialogue (Feller S ed.). Amsterdam/Philadelphia: Benjamins, 21–44. Reprinted in: Capone A and Mey JL (eds) (2016) Interdisciplinary Studies in Pragmatics, Culture and Society, 209–232. Cham: Springer.) Weigand E (2010) Dialogue: The Mixed Game. Amsterdam/Philadelphia: Benjamins. Wilson EO (1999) Consilience: The Unity of Knowledge. New York: Vintage Books.

7

This page intentionally left blank

1 Pragmatics From Language as a System of Signs to Language Use Jens Allwood

1.1 Introduction In classical Greece and Rome, rhetoric and dialectics, the most popular studies of language, were concerned with persuasive speech making and argumentation. Besides this, and less popular, there was also a study of language concerned with the preservation of the Homeric epics, the Odyssey and the Iliad. In the Alexandria grammar (compare Robins, 1997), this study developed into a full fledged morphologically based grammar, containing more or less the parts of speech and subcategories we still use in most of modern language descriptions. Two characteristics of this classical description were that, first, it was normative—the correct forms of the Homeric epics were to be preserved against the changes taking place in Alexandrian Koiné Greek, and second, it was concerned with written language. Over time, up until today, this type of grammar—a normative description of primarily the morphological characteristics of the written representations of a language, especially when complemented with a lexicon, has been seen as a way of capturing the essential features of a language. It was, and still is, the basis of both linguistic theorizing and language teaching in many countries of the world. Grammar (from the Greek techné grammatiké, the art of letters, i.e. writing) was seen as the core of a language and as an indispensible aid to write, but also, somewhat surprisingly, to speak a language. If the use of language was studied at all, it was studied in the literary works of great writers and to some extent in rhetoric and dialectics. Moving to the twentieth century, there have been three main trends in linguistic research on language: 1. Historical linguistics—this trend has continued the traditions from the nineteenth century, when it was the dominant interest of linguists. 2. Structuralism—the main trend of the twentieth century. 3. Communication-, usage- and function-oriented studies of language—a minority trend.

Historical linguistics has had a constant but diminishing presence during the century, but has not really been the primary focus of interest for linguists, especially not in the Saussure 9

Jens Allwood

influenced tradition of “synchronic” linguistics (see below). Outside of linguistics, communication has been more widely studied, above all in the United States within the discipline of communication studies. In what follows, I will concentrate on the two last trends, starting with structuralism.

1.2  Twenthieth Century: Structuralism Including Generativism Perhaps the most influential view of language during the twentieth century comes from the Swiss linguist Ferdinand de Saussure, who, inspired by the social emergentism of the French sociologist Emile Durkheim (Durkheim, 1895) and by Hegelian ideas of how identity can be given by a system of oppositions, conceived of language as a static network of syntagmatic and paradigmatic oppositions (de Saussure, 1959 [1916]). The syntagmatic axis is made up of written (and in principle also spoken) language strings of words that create a text, while the paradigmatic axis is made up of the set of (similarity based) associations, linked to every word in a syntagm, for example, the set of words that could be substituted in a particular syntagmatic position. Inspired by Durkheim, Saussure furthermore proposed to distinguish between the biologically given “faculty of language” (faculté de langage), giving rise to an underlying social system of language, which he called “langue” and its realization in individual acts of speaking, which he called “parole”. He also proposed that it is “langue” which is the proper object of study of linguistics. Even though “langue” is a fairly abstract system, it is not Platonic in nature but social in the sense of Durkheim, that is, not reducible to properties or activities of individuals but sui generis social. Saussure then added to the trichotomy between “langage”, “langue” and “parole” a distinction between a synchronic (at a given point in time) and a diachronic (or historical) study of language and suggested that synchronically we could regard language as a social static system (langue). This enables us to think of language diachronically as a series of changes between synchronic systems. Synchronic language descriptions in this way become a basis for diachronic descriptions and are therefore presupposed by diachronic descriptions. A consequence of the Saussurean conception of language as a social static system (langue) is that the traditional grammar and non-communication oriented view of language was reinforced. Language could be seen as a static system of signs organized syntagmatically and paradigmatically. In fact, Saussure’s conception of language as a synchronic description of the irreducible social system of “langue”, in opposition to the variation in “parole”, also politically helped to strengthen the normative abstraction and idealization of a cultivated national language devoid of variation in terms of dialects, age, gender or social activities and, in this way, supported or at least was compatible with the cultivation and construction of nationally homogeneous states. Many of the features of the view of language promoted by the Saussurean concepts of “langue”, “synchrony” and “system of syntagmatic and paradigmatic oppositions” were supported and strengthened by the structuralist movement in linguistics which roughly speaking can be divided into North American (US) structuralism and European structuralism. American structuralism has as its leading proponents Leonard Bloomfield (Bloomfield, 1933), Bernard Bloch and George L. Trager (Bloch and Trager, 1942) and Zellig Harris (Harris, 1951) and is oriented to “expression” side of language rather than meaning and empirical data rather than theory, focusing on field work and distribution analysis in the areas of phonology and morphology with very little time for meaning and semantics. American structuralism was simultaneous with and harmonious with the Behaviorist school in psychology, which like American linguistic structuralism was very empiricist, in its perspective 10



From a system of signs to language use

on how scientific studies should be pursued, represented, for example, by B. F. Skinner in his widely read book Verbal Behavior (Skinner, 1957). Even though the American structuralist movement was dominant in the United States, there were also other voices less in harmony with the expression-oriented empiricism of the American structuralists. Two such voices were Edward Sapir (1921) and Benjamin Lee Whorf (1942), who both, in the tradition of Wilhelm von Humboldt and Franz Boas, were more open to the content side of language and its connection with culture and cognition. European structuralism, following Ferdinand de Saussure (1959 [1916]), had as its leading proponents Louis Hjelmslev in Denmark (1928/1929, 1943), Roman Jakobson (1960, 1971) and Wilhelm Mathesius (1929) and the Prague School. In contrast to the empiricism of the American structuralism, European structuralism is theoretical, more abstract and also more open to an analysis of the content side of language. Especially the work of Hjelmslev (1943) is relevant in this connection, since he reinforced and in many ways clarified the abstract view of language as a sign system based on oppositions that had been suggested by Saussure. Drawing on the Danish linguist Jespersen, Hjelmslev added to the Saussurean structuralist approach a clearer distinction between the “expression” and “content” of a linguistic sign, by connecting it to an Aristotle-inspired distinction between the “form” and the “substance” of a sign. Although mainly working on phonology, the Russian linguist Roman Jakobson, in his proposal of an account of the functions of language (referential, poetic, emotive, conative, phatic and meta-lingual), also showed an openness to the content side of language. As in America, the European continental structuralists can be seen in contrast to other contemporaneous schools of linguistics. One important such school was the functional type of linguistics, which was developing in Great Britain with Bronislaw Malinowski (1922), Alan Gardiner (1932) and John Rupert Firth (1957). As in the United States with Sapir and Whorf, here the content side of language was connected to cognition and culture, albeit in a different, more functional and less epistemological and ontological way than in the United States. According to both Saussure and Hjelmslev, every linguistic sign is the result of a twoplace relation, where there is an “arbitrary connection” between a signifier (expression) and something signified (its content). In contradistinction to Saussure, who thought the substance was basically psychological, Hjelmslev did not think linguistics should take a stand on the nature of the “substance”, neither of the expression nor of the content of the linguistic sign. Both expression and content had “form” and “substance” but linguistics should only be concerned with the form of both the content and expression of the signs, leaving the investigation of the substance to other sciences. In this way, Hjelmslev hoped, among other things, to bring about a rapprochement between linguistics and mathematics (his father was a professor of mathematics) and to create in linguistics a study of the “algebra of language” dealing only with the forms of language. Thus, through the distinction between “form” and “substance”, the view of the linguistic sign and of language promoted by Hjelmslev takes another step towards abstraction and decontextualization—in its essence, language is form, not substance. Linguistics should be concerned with language as an abstract system of disembodied and decontextualized signs. Outside of linguistics, another source of the non-communication-oriented view of language comes from the descriptions of language and grammar developed in logic and mathematics during the twentieth century. Usually, in such descriptions the conception of grammar as a normative written language description is taken for granted. A consequence of this is that a language here often is conceptualized as a set of well-formed discrete strings of (discrete written) signs. The conceptualization presupposes and becomes plausible 11

Jens Allwood

against a background of reflecting on language as a set of grammatical sentences made up of word-like signs. This perspective is often very clear, if we open a text book in mathematical linguistics or formal semantics, for example, Grishman (1986: 12), writes “Formally, a language is a set of sentences, where each sentence is a string of one or more symbols [by symbol Grishman means words] from the vocabulary of the language”. The quote reflects a view of language that really only allows for a study of syntax, morphology and possibly for graphemics, since content and function are not taken into consideration. Language is seen as a set of sentences, consisting of sequences of discrete symbols. In most of these formal accounts, the use of language as a medium of communicative interaction, in which non-discrete phenomena like prosody and gesture play a crucial role, is not seriously considered or even noticed. Rather, with inspiration from writing, language is seen as an abstract set of rules for manipulating well-formed strings of discrete signs. At the side of the formal tradition, partly reinforcing it, there was a majority of linguists who were becoming structuralists in the tradition of Saussure, Bloomfield, Hjelmslev, and Jakobson (Jakobson, 1971). These linguists were, on the one hand, developing the structuralist theory of language, the clearest example of this being Louis Hjelmslev (1928/1929, 1943) and the glossematic approach to linguistic theory being developed in Copenhagen. On the other hand, others, mainly in the United States, were developing the empirical methods for studying language and here the clearest examples are probably Bloch and Trager (1942) as well as Zellig Harris (1951), where the latter exemplifies how an empiricist approach can also be dealt with formally and mathematically.

1.2.1  Generative Syntax In the 1950s, the structuralist tradition took a new turn through the work of Noam Chomsky (1957) who took as his point of departure the syntactic description of a language, proposing to describe this with the kind of rewrite rules proposed by Emile Post (1944) and earlier by Henrik Abel (1881) in mathematics. Like Hjelmslev, Chomsky, thus, suggested a mathematization of linguistics, and in an elegant way demonstrated how his approach could be used to give an elegant, precise description of several syntactic constructions in English (passive, negation, interrogatives) (compare Chomsky, 1957). In 1959, Chomsky broadened his approach to language by publishing a very influential critical review of B. F. Skinner’s Verbal Behavior (Skinner, 1957). In this review, Chomsky (1959) criticized the empiricist approach of Skinner for being largely vacuous and being unable to account for a child’s acquisition of complex grammatical rules. In the books Aspects of the Theory of Syntax and Cartesian Linguistics which appeared in 1965 and 1966, Chomsky elaborated his proposals from 1957 and his critique of behaviorism from 1959 by proposing a framework for linguistic theory, which became known under the names of “generative grammar” or “transformational grammar”. In this framework, Chomsky suggested that the formal rules of phrase structure produced so called “deep structures”, which, via transformational rules, produced the actual “surface structure” of sentences in a language. He suggested that these formal rules formed the essential core of a language, its grammar, which in turn could be said to capture the linguistic “competence” of an “ideal speaker–listener”. This competence could then be contrasted with actual linguistic “performance”, which for various reasons like fatigue, illness, or memory limitations might contain “incompetent behavior” i.e. behavior not directly reflecting the ideal “linguistic competence”. 12



From a system of signs to language use

In line with the anti-empiricist, rationalist tradition of Descartes, Chomsky also suggested (1966) that the inadequacies of the empiricist-behaviorist approach to language acquisition exemplified in Skinner’s Verbal Behavior could only be remedied by engaging in a more philosophically rationalist approach making use of “innate ideas” and relying on an inborn language instinct or, as Chomsky called it, “language acquisition device” (LAD). Since all human children can learn all human languages in spite of the extensive typological differences between human languages, this language acquisition device contains the potential for a universal grammatical core of all human languages, often referred to as “universal grammar”. The Chomskyan views have, in general, supported an abstract decontextualized view of language. Chomsky’s notion of “competence” cannot be identified with an underlying, social reality like Saussure’s “langue” nor can it be identified with actual genetic, neurological or psychological models of linguistic competence, since it concerns an abstract model of the grammatical principles of an “ideal speaker–hearer”. It is perhaps not so strange that some of Chomsky’s followers and perhaps Chomsky himself at times have espoused a Platonic approach to language; see Chomsky’s discussion of Plato’s problem, Chomsky (1986), or Fodor’s discussion of the nature of the “language of thought” in Fodor (1975). Even if American structuralism was criticized by Noam Chomsky for being too empiricist and non-theoretical, Chomsky was more tolerant of European structuralism. In fact, the methodological goals of Chomsky and Hjelmslev were rather similar, aiming at a mathematical account of language structure. This means that, in a wide sense, generativism can be seen as a part of the structuralist movement in linguistics. The decontextualized view of language and of the linguistic sign was further reinforced by another feature of the analysis. In Saussure and Hjelmslev as well as in Chomsky, in both the structuralist and generativist movements, the nature of the linguistic sign was assumed to be a two-place relation between a signifier (expression) and something signified (content), with no consideration of the user or context and their respective influence on either expression (signifier) or content (signified). However, this view was not generally shared but, in fact, before the rise of structuralism and generativism it was already challenged by Charles Sanders Peirce (1839–1914).

1.3  Peirce and Formal Semantics The Saussurean/Hjelmslevian structuralist view of the nature of the sign stands in contrast to the view suggested by Charles Sanders Peirce in a series of papers from 1868 to 1914 (see Peirce, 1940). In the Peircean analysis, all types of signs are three-place, not two-place. The two-place relation between signifier and signified is always mediated by a relation to the sign user, who can be a sign producer or sign perceiver. The sign user creates the sign by creating a mental representation (“interpretant”), connecting the signifier (“representamen”) with the signified (“object”). This analysis of the sign, thus, makes it crucially dependent on the sign user and in this way prepares the ground for the current widespread interest in communication and pragmatics (see below). Peirce further distinguished three types of signs (symbol, icon, index) depending on the type of representing link (interpretant) the sign user uses to connect the representamen (the signifier) with what is signified or represented (the object). On the Peircean analysis, language is a semiotic system, primarily involving linguistic symbols, which are constituted by users. This view of language is, thus, less disembodied 13

Jens Allwood

and less decontextualized than the views of Saussure, Hjelmslev and Chomsky that we have discussed above, and for this reason, also more open to considering the role of language in communication. Inspired by Peirce, Charles Morris (1938) proposed that any sign system could be studied from the three points of view suggested by the Peircean three-place relation analysis of signs: 1. “Syntax”: If we restrict our study to relations between the signifiers or representamens, we are studying the “syntax” of the system. 2. “Semantics”: If we include the relation between signifiers and what is signified—the object (between representations and what is represented)—we are studying the “semantics” of the system (the sign system as a two-place system). 3. “Pragmatics”: If we include not only the relation between signifiers and signified, but also the relation to the users, that is, the relation between the representations, what is represented and the interpretants, we are studying the “pragmatics” of the system (the sign system as a three-place system).

The proposal made by Charles Morris only gradually became noticed and used. One of the first persons to use it was Rudolf Carnap, who in the 1930s (see Carnap, 1942), through Gödel’s incompleteness theorem for logic and the work on semantics by Alfred Tarski, had become convinced that formal systems could not be studied merely syntactically, but had to also involve a semantics. Later in the 1950s, Carnap further became convinced that, in fact, including a semantics was also insufficient, so that a pragmatics had to be included as well (Carnap, 1959).

1.4 Pragmatics In general, the necessity to include the use of language in communication in the theoretical analysis of language became accepted within the philosophical study of language in the 1950s and 1960s, where we find the appearance of a body of work on language which is clearly pragmatic in nature. Three very influential examples of this are Ludwig Wittgenstein’s (1953) Philosophical Investigations and John Langshaw Austin’s (1962) How to Do Things with Words as well as Paul Grice’s 1967 William James Lectures, partly published in Grice (1975). It is interesting that, in general, the tradition of formal semantics, which unlike mathematical linguistics was often influenced by the same sources as Carnap, first became convinced that semantics was indispensable in the description of a language, and then in the late 1950s and 1960s also became convinced that aspects of context, that is, pragmatics, had to be included in the formal description of language, especially for the description of deixis, compare Bar-Hillel (1954) and Montague (1968, 1970). Perhaps the most important contribution here was made by Richard Montague, who in his papers giving formal syntactic and semantic descriptions of formal and natural languages (Montague, 1974), followed Bar-Hillel’s suggestions and included pragmatics in the form of spatio-temporal coordinates (indices), used to assign referents and truth values, especially to sentences containing deictic and temporal expressions. Note here that the introduction of context, in the form of spatio-temporal coordinates, to give a more complete account of truth conditions and inference does not mean that a study of the role of language in interaction or communication is introduced. Context is merely 14



From a system of signs to language use

conceptualized as a set of parameters or indices (see Lewis, 1970 and 1983a, for a well known example of this), that are used to determine reference and truth and no consideration is taken of multimodality or the dynamics of dialog. In linguistics the “pragmatic turn” was gradually introduced in the 1970s and 1980s, following what might be called the “syntactic turn” championed by Chomsky and the generativist school and the “semantic turn” championed by Montague (1974) as well as the so-called “generative semantics”, advocated by scholars such as McCawley (1976), Lakoff (1969) and Jackendoff (1976). Several different perspectives on pragmatics were represented. Compare the following definitions given of pragmatics:

1.4.1  Allwood (1976) Allwood, in 1976, characterizes pragmatics in the following way (Allwood, 1976: 173): Pragmatics could be viewed as the study of the factors which determine what information is actually communicated by a sender and apprehended by a receiver; the different communicative intentions, purposes, motives and reasons; the different ways in which a receiver can react to information; the relationship between sender intention, receiver reaction and other psychological and sociological phenomena, such as systems of emotions and attitudes (including cognitive ones) and phenomena like social structure, role relations, power and solidarity. Further, it should study the relationship between conventional content and the intended content of a sender or the apprehended content of a receiver. It should determine to what extent this relationship is dependent on various contextual factors such as shared background assumptions. It should study how such dependence on pragmatic factors affects the phonology, morphology, syntax, content and purpose of the utterances expressed and apprehended. Finally, it should determine in greater detail the strategies, conventional and nonconventional norms that govern communicative interaction. Allwood then further elaborated the consequences of this characterization/definition in Allwood (1978) and in 1981, in addition, argues that actually there is no good pre-theoretical, non-stipulative grounds for the distinction between semantics and pragmatics, since all ways of trying to delimit semantics turn out to presuppose so called pragmatic factors.

1.4.2  Levinson (1983) Levinson discusses several possible definitions of pragmatics like the following: (i) “Pragmatics is the study of those relations between language and context that are grammaticalized or encoded in the structure of a language” (9). (ii) “Pragmatics is the study of all those aspects of meaning not captured in a semantic theory” (12). (iii) “Pragmatics is the study of the relations between language and context that are basic to an account of language understanding” (21). (iv) “Pragmatics is the study of the ability of language users to pair sentences with the contexts in which they would be appropriate” (24). 15

Jens Allwood

(v) “Pragmatics is the study of deixis (at least in part), implicature, presupposition, speech acts, and aspects of discourse structure” (24). Basically, Levinson’s definitions (i), (iii) and (iv) are compatible with and foreshadowed by Allwood’s characterization of pragmatics from 1976. Levinson himself, however, has a weak preference for definition (ii) but does not press the point. The weak point of this definition is that it is dependent on giving semantics a definition that is independent of the definition of pragmatics. As is discussed in Allwood (1981), this is not possible if we define semantics as being concerned with truth conditions or conventional meaning, the two primary candidates, leading to the consequence that semantics and pragmatics cannot really be theoretically distinguished. As Levinson points out, we can then retreat to definition (v), which is not really a definition but a stipulative list of phenomena that we should regard as pragmatic.

1.4.3  Leech (1983) Geoffrey Leech (1983), the same year as Levinson’s book Pragmatics, defines pragmatics the following way: (i) “Pragmatics is the study of meaning in relation to speech situations” (6, 15) and (ii) “Pragmatics—the principles of language use” (4). We can see that both of these definitions are less specific than the definitions or characterizations provided by Allwood and Levinson. The first definition runs into the same problems as Levinson’s definition (ii). It is dependent on there being enough substantial aspects of meaning that can be studied without a relation to speech situations. The second definition includes all aspects of language use and therefore means that pragmatics cannot be an aspect of language in the same sense as phonology, morphology, syntax and semantics.

1.4.4  Mey (1993) Ten years later, Jakob Mey defines pragmatics the following way: “Pragmatics is the study of the conditions of human language uses as these are determined by the context of society” (42). In addition, Mey wants to distinguish two types of context: (i) Societal context (determined by societies, institutions) (ii) Social context created in interaction. Given this definition, one wonders if not all properties of a given language—phonological, morphological, syntactic, semantic—are to some extent determined by the society in which the language occurs, so that pragmatics becomes the study of the social conditions of any aspect of language. It also seems that the definition relates pragmatics more to contexts of type one: macro-social factors, than to contexts of type two: micro-social factors, which would have the consequence that pragmatics would be more or less co-extensional with socio-linguistics. Later Mey develops his societal view of pragmatics further into an interdisciplinary study (see Mey, 2016). 16



From a system of signs to language use

1.4.5  Verschueren et al. (1995) Jef Verschueren comes fairly close to Mey’s view, when he says that pragmatics should be seen as a functional perspective on language: Pragmatics as a general functional perspective on (any aspect of) language, i.e., as an approach to language which takes into account the full complexity of its cognitive, social, and cultural (i.e. ‘meaningful’) functioning in the lives of human beings. (19) Neither Mey nor Verschuren see pragmatics as a component to be added to linguistic theory, following syntax and semantics. Rather they see it as a perspective through which any aspect of language can be studied; in Mey’s case the societal conditions and in Verschueren’s the functions the aspect has.

1.4.6  Neo-Gricean Approaches Sometimes scholars like Sperber and Wilson (1986), Levinson (2000) and Horn (2004) are referred to as “Neo-Gricean”. What these scholars have in common is that they take the conversational maxims used by Paul Grice (see Grice, 1975) to characterize rational communication as a starting point for a more general pragmatic approach to language. Since the approaches all rely on some notion of “literal meaning”, while disagreeing on what is required to reach actual meaning in communication, they all run into the problems related to trying to clarify what “literal meaning” is, seen already above in relation to approaches of Levinson (1983) and Leech (1983). To summarize, we can see that all the definitions and characterizations agree that pragmatics is essentially concerned with the way language, especially meaning, interacts with context. But there is no agreement about what aspects of meaning and context this concerns, or about how to draw the line between semantics and pragmatics or, indeed, about whether there is a substantial line to be so drawn. Maybe one can venture to say that the “pragmatic turn” has meant that pragmatically sensitized linguists have abandoned the structuralist credo that language essentially involves a two-place relation between expression and content (form and meaning) in favor of the idea that the analysis of language always involves a three-place relation between expression, content and context and that this relation should be studied by focusing on language use rather than on language seen as a static system.

1.5  Non-structuralist Linguistics Structuralism, especially including generativism, and later pragmatism have been the dominant schools of twentieth century linguistics, but the general development has been more complicated and different schools have been competing with each other. While in the first half of the century the majority of linguists were engaged in historical or structuralist linguistics, there was also a minority of linguists who saw language as connected with context and to some extent with communication. This minority was most strongly represented in Britain where Bronislaw Malinowski (1922), who was an anthropologist, Alan Gardiner (1932) and John Rupert Firth (1957) (later continued by Michael Halliday, 1973), all ­published work in the 1920s and 1930s stressing the role of context. 17

Jens Allwood

In the early twentieth century, an attempt at a use-oriented linguistics was made by Alan Gardiner (1932) in Great Britain and later by J. R. Firth and Michael Halliday. In Germany, the psychologist Karl Bühler presented a communication-oriented view of language (Bühler, 1934). One of the ideas of J. R. Firth was that national languages were to some extent an artificial creation and that a language instead should be viewed as a collection of registers (or genres), capturing the linguistic practices of particular activities and organizations. In such a view, context plays an important role in determining the conditions for the register to appear and stabilize. Besides these use-oriented studies of language, the Alexandria grammatical tradition has been a major and persistent influence on studies of language from antiquity until today. This influence has had two features that have prevented linguistics from becoming an empirical science of linguistic communication: a normative focus, and a focus on written language. Both features have been preserved through the centuries and appear again in twentieth century linguistic structuralism, as promoted by linguists like Saussure, Bloomfield, Hjelmslev, Jakobson and Chomsky, even though they have often paid lip service to empirical descriptivism. Chomsky’s interest in the “competence” of the ideal speaker–listener, in contrast with the actual behavior (performance) of the speaker–listener, is, for example, completely normative, in its classification of linguistic expressions into “grammatical and “ungrammatical”. It also relies a lot on written language since our intuitions of “grammaticality” are clearer for written language than for multimodal face-to-face interaction. The same is true (although in slightly milder form) for linguists who believe that their object of study is something akin to the Saussurean notion of “langue” which they believe should be distinguished from actual use “parole” in which the norms of “langue” (often identified with the norms of a national language, see Harris, 1980) are not always upheld. The written language bias is more implicit but can also be seen in how the choice of examples and empirical data, in both structuralist and generativist linguistics, usually does not focus on the use of authentic language in communication, but instead uses constructed examples highlighting various systemic features of language, which are usually most clearly apparent in written language (see also Linell, 2005). Added to the focus on rules and focus on written language, there is also an additional feature of abstraction leading to disembodiment and decontextualization. This is the result of identifying the object of study of linguistics with either the Saussurean doctrine of “langue” as separated from “parole”, the Hjemslevian conception of form as separated from substance or the Chomskyan idea of an idealized competence as separated from actual performance. In all these approaches, actual linguistic data from real interaction can be seen as of secondary interest, to be superseded by a more abstract notion of language. Linguistics was dominated by this abstract conception of language in the 1960s and 1970s. However, beginning in the 1970s, and with increasing strength in the 1980s and 1990s, there was a shift of focus, so that linguistics gradually began to discover the two other aspects of Morris’s trichotomy: semantics and pragmatics. Besides the English Firthean tradition, movements like cognitive semantics, the linguistic pragmatic tradition and the sociological school of conversation analysis (CA) start to have an influence. One important reason for the gradual shift away from a focus on language as an abstract system is the gradual discovery of the philosophical texts on language from the 1950s and 1960s. A second reason is the new availability of first, analog recording technology, and then of digital recording technology. Both of these types of recording make it possible to seriously record, transcribe and store data in the form of corpora and linguistic data bases which then 18



From a system of signs to language use

can be used to study language as an instrument of face-to-face communication. The richness of the new types of data makes it possible to study both collective patterns of language use and more individualistic idiosyncratic ways of using language, for which existing models of language still, generally, have no clear role. A third reason is the critical theoretical discussion within linguistics, which increasingly, like the earlier discussion in philosophy and logic, calls into question the possibility of pursuing an adequate description and explanation of language from a purely syntactic point of view. Language, after all, exists to share meaning, so a syntactic description necessitates a semantic description that in turn leads to pragmatic description. Thus, following on the acceptance of semantics, a next step is to show that a syntactic and semantic description and explanation of language is not possible without taking context and language users into account, which means that neither syntax nor semantics can be pursued without pragmatics. For other more radical criticism of the idea of describing a language, compare Harris (1980, 1981) and Baker and Hacker (1984). In this way, it seems as if a majority of linguists gradually are moving from a view of language as an abstract and decontextualized system to an appreciation of language as an instrument for communication. However, what this means has been interpreted in several different ways.

1.6  The Relationship of Pragmatics to Context Even if there has been a gradual realization of the role of context for language, one problem is that the conceptions of context have not been the same. The appreciation of the role of context for language has taken several different routes. One way to gain an appreciation of the routes that have been taken is to use the definition of a sign provided by Peirce, in combination with the characterization of the aspects of a sign system suggested by Morris (syntax, semantics and pragmatics) as a starting point to capture the different conceptions of context. The basic idea is to go from the user dependence of the constitution of the sign claimed by Peirce to a more general notion of a context dependence of the sign. In this way, we can basically distinguish three different approaches to context, while also being aware that some of the approaches combine several of these three types of context. Using the terminology suggested by Peirce, we can distinguish the following types of context: 1. 2. 3. 4.

The context of the representamen—syntax The context of the object—semantics The context of the interpretant—pragmatics Combining all of the three above types and perhaps adding other aspects.

1.6.1  The Context of the Representamen (the Signifiers) This is the original sense of context, namely, that of the accompanying text. Investigating context in this sense, we consider how a linguistic element (phoneme, morpheme, word, sentence, text) is influenced by the linguistic elements surrounding it. This is the context of signifiers given by their syntagmatic surroundings. For example, this is the notion of context, usually involved, when in translation we need a context to give the proper translation. For instance, if we want to translate the English verb “know” to German, we need to have information if the context is that of “knowing the answer” or that of “knowing a language” or yet something else. The translation will depend on what the context is. 19

Jens Allwood

This notion of context, over time, became broader and broader and was eventually used in the sense of “extra-linguistic context”, so some authors felt the need to reintroduce the original notion by creating a distinction between “context” and “co-text”, where “context” was given a vague, wider and basically undefined sense while “co-text” reintroduced the original text-based notion (compare Catford, 1969). It is this text-based notion of context that first becomes popular in linguistics, where it appears, for example, in the movement of “text linguistics” which was popular in the 1970s and 1980s (Harweg, 1968; Petöfi, 1971; Van Dijk, 1972; Enkvist and Kohonen, 1976; Halliday and Hasan, 1985). In a wide sense, the representamen-based notion of context is also present in “multimodal communication”, that is, communication employing more than one sensory modality, where we can regard body language and prosody as context for speech or pictures as context for text. In line with the Peircean starting point, we are now considering all accompanying representamens and since gestures and prosody qualify as such representamens, we can now abandon and extend the representamen-based notion of context merely as “accompanying text”.

1.6.2  The Context of the Object (the Signified) The context of the object is the semantic notion of context. This is the notion of context that formal language philosophers discovered was necessary to get around the purely syntactic approaches to language. Here context is used to assign reference and truth conditions to the expressions of a language. In practice, the language considered was usually a formal language that was gradually being made more sophisticated by incorporating new features from natural language. Clear examples of such incorporated features are deixis, tense, pronouns and modal expressions (in a different sense of modality than sensory modality). For example, in order to give an interpretation of the reference and truth conditions of the sentence “I was hungry”, we need to know who said it, and when it was said. For this, contextual information is needed. Without this contextual information, we will not know what proposition was expressed and no inferences can be drawn from the sentence. One may of course be skeptical of this approach, since it indeed is artificial and does not do justice to the real complexity of natural language. The question is whether the austere ideal of conceptual clarity, which was espoused by Carnap and others working in formal semantics, really is compatible with a consideration of the real complexity of natural language. At present, at least, the alternatives seem to be either an imprecise theory doing justice to the complexity of linguistic communication, or a more precise theory, which does not do justice to this complexity. In Bar-Hillel (1954), it was suggested that the distinction between “pure” and “descriptive” should also apply to pragmatics and so pure or “formal pragmatics” was conceived. Carnap and Morris later both concurred with this proposal (Schillp, 1963; Morris, 1938). We can see that the semantic conception of context is fairly different from the conception of context as accompanying representamens (signifiers). Here we are instead focusing on what is signified conceptualized as objects, indices, possible worlds or sets of believed propositions.

1.6.3  The Context of the Interpretant (the User) The third type of context brings us to the “interpretant” which is the interpretation given by the user of a particular signifier (representamen) in terms of the object it signifies (the 20



From a system of signs to language use

signified). We thus arrive at the context of the user and by extension also at the context of usage. This is the pragmatic notion of context, that is, the notion of context that invites us to study language in use, to study the functions of language, no matter whether they are given by accompanying signifiers or by the need to identify truth conditions and reference (propositions). The notion thus also opens up for a consideration of other contextual factors than accompanying signifiers, truth conditions and reference determining possible worlds, indices or propositional beliefs. Now contextual factors can include, for example, culture and social activity, since these factors can influence how a user interprets and in other ways uses language. This is the notion we see employed in the Malinovski–Firthean tradition, in Halliday (1973), Levinson (1979), Allwood (1976), sociolinguistics, anthropological linguistics, intercultural communication, communication studies and computer-based dialog systems. It is also a notion of context that allows inclusion of the syntactic and semantic notions of context which then become dimensions in a more inclusive notion of context. What the gradual acceptance of the pragmatic, more inclusive notion of context means is that, in a sense, the interests that were most popular in ancient Greece, rhetoric and dialectics, are being reinstated and that dialog and communication increasingly are seen as central phenomena for an understanding of the nature of language (see Hundsnurscher, 1980; Weigand, 1989/2003 and 2009). A second type of classification makes a distinction between (i) the context of communicative action, (ii) the context of other actions simultaneous with communication, (iii) the social activity that both types of action are part of and (iv) the cognitive content (e.g. beliefs, values) that is activated from memory by the actions and activities mentioned in (i), (ii) and (iii); compare also Allwood (1995).

1.7  Post-structuralist Linguistics By the 1960s, computers and recording technology had made possible the collecting and establishing of corpora of different types of language. This trend has continued, and over the past fifty years an increasing number of corpora of different registers and genres have appeared, see, for example, the corpora collected by Randolph Quirk, Jan Svartvik and Sidney Greenbaum (compare Svartvik and Quirk, 1980). Most of these corpora have been oriented to written language like Sinclair’s COBUILD (compare Moon, 2009), but some have attempted to also capture the spoken language, like the British National Corpus (2015) (see Leech and Smith, 2000) and to some extent Sinclair’s COBUILD, or only spoken language like the Gothenburg Spoken Language Corpus (GSLC) (Allwood et al., 2000). The appearance of these corpora has meant that language now could be studied with a basis in what Saussure called “parole” and Chomsky called “performance”, that is, in actual concrete linguistic usage data, rather than in idealized accounts of language as an abstract system of signs. Above all, the appearance of corpora has made possible the observation of individual language use in communication and descriptive statistics on how different linguistics expressions are actually used. During the last decade, linguistic corpora have increasingly become video based, making possible the establishment of multimodal corpora, allowing for studies of the simultaneous multimodal interaction of gestures (in the sense of informative body movements), prosody and vocal verbal elements. 21

Jens Allwood

So, even if the development of linguistics has been greatly influenced by the philosophical discussions of the nature of language and communication, it has also been influenced by communication-oriented research in sociology, anthropology and communication science. Particularly important has been the influence from the school of sociology called conversation analysis (CA) (see, for example, Sacks, 1992 and Sacks, Schegloff and Jefferson, 1974), which has lead to a much greater interest in interaction and communication in linguistics. However, the focus only on transcriptions of authentic data by CA, without taking the context into account, has met with criticism from many scholars, see, for example, Allwood (1976), Hundsnurscher (1980) or Weigand (1989/2003). Strangely enough, the influence in linguistics from the field of communication studies, which have been very popular especially in the United States, although highly relevant, has, so far, been much smaller. Perhaps because of the status of formalized accounts in linguistics, a somewhat stronger influence has come from computer simulation of multimodal communication and the construction of artificial dialog systems. All of these influences are slowly forcing a reconceptualization of the nature of language from being essentially based on a normative version of written language captured in concepts like “langue”, “form” or “competence” to a conception of language as being an empirical phenomenon, essentially multimodal, involving, besides vocal verbal (or written) elements, also prosody and gestures, with the primary function of language being a medium of both communicative interaction and thinking in different social activities, sensitive to the needs of different types of users.

1.8 Outlook Based on the discussion above, it might now be possible to point to the following potential developments in contemporary linguistics for the three subfields we distinguished in the introduction: 1. Historical linguistics—Historical linguistics has long been an important part of linguistics and will, in all likelihood, continue to be so. It is successively incorporating and providing a historical perspective on new aspects of language, like cognitive semantics and pragmatics as well as on more communication- and dialog-oriented studies. It is also making an increasingly extensive use of the historical linguistic corpora that are becoming available. 2. Structuralism (including generativism)—Due to the availability of corpora of actual linguistic data and the theoretical needs of various practical concerns involving language use, even though structuralism has been the dominant approach during the twentieth century, it will probably become less dominant. However, structuralist linguistics will still be pursued for traditional linguistic research but also probably in connection with new attempts to provide formal “mathematical” models of language. 3. Communication, usage and function oriented studies of language—The “pragmatic turn” and the growing popularity of communication- and cognition-oriented studies, inside and outside of linguistics, often pursued in an interdisciplinary way, combined with cognitive science and studies of corpora, experimental methods and computer simulation, has meant that this type of linguistics is slowly becoming the most common type. The field will probably over time see a theoretical discussion between more cognitive approaches, like cognitive linguistics, cognitive semantics and so forth, and more communication-, interaction- and dialog-oriented approaches, like interactional linguistics or dialog-oriented linguistics—leading (it is to be hoped) to new insights, and resulting in new proposals for how to bridge this gap between interaction and cognition related approaches. 22



From a system of signs to language use

In general, because of the new availability of ways to capture linguistic interaction, linguistics is becoming more of a full-fledged empirical science where theories and descriptive accounts can be tested against empirical data in much the same manner as they can in other ways of studying human activity. This means that there can be room both for intensive qualitative case studies (both experimental and with more ecological validity) of what communicative means (content and expression) are involved in particular instances of communication in particular social activities and for more extensive quantitative, corpus-based descriptions and statistical analyses of aspects of collective language use. It also means that there can be a place for inductive as well as abductive and deductive building of a theory of communication, cognition and language, which has the power to provide descriptions and explanations of human employment of language (languaging) in cognition and communication in different social activities and cultures.

References Abel NH (1881) NH Abel Œuvres complètes. Ed. by Sylow L and Lie S, 2 vol, Christiania. Allwood J (1976) Linguistic communication as action and cooperation. Gothenburg Monographs in Linguistics 2, Gothenburg: University of Gothenburg, Department of Linguistics. Allwood J (1978) A bird’s eye view of pragmatics. In: Gregersen F (ed.) Papers from the Fourth Scandinavian Conference of Linguistics. Odense: Odense University Press, 145–159. Allwood J (1981) On the distinctions between semantics and pragmatics. In: Klein W and Levelt W (eds) Crossing the Boundaries in Linguistics. Dordrecht: Reidel, 177–189. Allwood J (1995) On the contextual determination of communicative functions. In: Moen I, Simonsen HG and Lødrup H (eds) Papers from the XVth Scandinavian Conference of Linguistics. Oslo: University of Oslo, Department of Linguistics, 25–35. Allwood J, Björnberg M, Grönqvist L, Ahlsén E and Ottesjö, C (2000) The spoken language corpus at the Department of Linguistics, Göteborg University. FQS—Forum Qualitative Social Research 1(3): 22. Aristotle (1908–1952) The Works of Aristotle Translated into English Under the Editorship of W. D. Ross, 12 vols. Oxford: Clarendon Press. Austin, JL (1962) How to Do Things with Words. Oxford: Oxford University Press. Baker GP and Hacker PMS (1984) Language, Sense and Nonsense. Oxford: Basil Blackwell. Bar-Hillel Y (1954) Indexical expressions. Mind 63: 359–379. Bloch B and Trager GL (1942) Outline of Linguistic Analysis. Baltimore: Linguistic Society of America. Bloomfield L (1933) Language. Chicago: University of Chicago Press. British National Corpus (2015) Available at: www.natcorp.ox.ac.uk/ Bühler K (1934) Sprachtheorie. Jena: Fischer Carnap R (1942) Introduction to Semantics. Cambridge, Mass: Harvard University Press. Carnap R (1959) Introduction to Semantics. Cambridge, Mass: Harvard University Press. Catford JC (1969) A Linguistic Theory of Translation, An Essay in Applied Linguistics. London: Oxford University Press. Chomsky N (1957) Syntactic Structures. The Hague/Paris: Mouton. Chomsky N (1959) Review: B. F. Skinner: Verbal Behavior. New York 1957. Language 35: 26–58. Chomsky N (1965) Aspects of the Theory of Syntax. Cambridge, Mass: MIT Press. Chomsky N (1966) Cartesian Linguistics. New York: Harper & Row. Chomsky N (1986) Knowledge of Language: Its Nature, Origin, and Use. New York: Praeger Scientific. Durkheim E (1895) Les règles de la méthode sociologique. Paris: Payot. 23

Jens Allwood

Enkvist NE and Kohonen V (eds) (1976) Reports on Text Linguistics: Approaches to Word Order. Åbo: Åbo Akademi. Firth JR (1957) Papers in Linguistics 1934–1951. London: Oxford University Press. Fodor JA (1975) The Language of Thought. Cambridge, Mass: Harvard University Press. Gardiner A (1932) The Theory of Speech and Language. Oxford: Clarendon Press. Grice HP (1975) Logic and conversation. In: Cole P and Morgan JL (eds) Syntax and Semantics, Vol 3: Speech acts. New York: Seminar Press, 41–58. Grishman R (1986) Computational linguistics: An introduction. Cambridge, UK: Cambridge University Press. Halliday M (1973). Explorations in the Functions of Language. London: Edward Arnold. Halliday MAK and Hasan R (1985) Language, Context, and Text: Aspects of Language in a SocialSemiotic Perspective. Geelong: Deakin University. Harris R (1980) The Language-Makers. London: Duckworth. Harris R (1981) The Language Myth. London: Duckworth. Harris Z (1951) Methods in Structural Linguistics. Chicago: University of Chicago Press. Harweg R (1968) Pronomina und Textkonstitution. München: Fink. Hjelmslev L (1928/1929) Principes de grammaire génerale. Copenhagen: Bianco Lunos Bogtrykkeri. Hjelmslev L (1943) Omkring Sprogteorins Grundleggelse. Copenhagen: Akademisk Forlag. Horn LR (2004) Implicature. In: Horn LR. and Ward GL (eds). The Handbook of Pragmatics. Oxford: Blackwell, 3–28. Hundsnurscher F (1980) Konversationsanalyse versus Dialoggrammatik. In: Rupp H and Roloff H-G (eds) Akten des VI. Internationalen Germanisten-Kongresses, Basel 1980. Part 2. Bern: Lang, 89–95. Jackendoff R (1976) Toward an Explanatory Semantic Representation. Linguistic Inquiry 7(1): 89–150. Jakobson R (1960) Closing Statements: Linguistics and Poetics. Style in language. New York: T. A. Sebeok. Jakobson R (1971) Selected Writings, Volume II. Word and Language. The Hague, Paris: Mouton. Lakoff G (1969) On Generative Semantics. Bloomington: Indiana University Linguistics Club. Leech GN (1983) Principles of Pragmatics. London: Longman. Leech G and Smith N (2000) The British National Corpus (Version 2) with Improved Word-class Tagging. UCREL, Lancaster University. Levinson SC (1979) Activity types and language. Linguistics 17: 365–399. Levinson SC (1983) Pragmatics. Cambridge: Cambridge University Press. Levinson SC (2000) Presumptive meanings: The theory of generalized conversational implicature, Cambridge Mass: MIT Press. Lewis D (1970) General Semantics. Synthese 27: 18–67. Lewis D (1983a). Postscript to ‘General Semantics’. In: Lewis D (ed.) Philosophical Papers, Volume I. Oxford: Oxford University Press. Lewis D (1983b). Philosophical Papers, Volume I. Oxford: Oxford University Press. Linell P (2005) The Written Language Bias in Linguistics: Its Nature, Origin and Transformations. London: Routledge. McCawley JD (ed.) (1976) Syntax and Semantics 7: Notes from the Linguistic Underground. New York: Academic Press. Malinowski B (1922) Argonauts of the Western Pacific: An Account of Native Enterprise and Adventure in the Archipelagoes of Melanesian New Guinea. London: Routledge and Kegan Paul. Mathesius W (1929) Functional Linguistics. Prague: Proceedings of the First Congress of Slavists. Mey JL (1993) Pragmatics: An Introduction. Oxford: Blackwell. Mey JL (2016) Pragmatics seen through the prism of society. In: Capone A and Mey JL (eds) Interdisciplinary studies in pragmatics, culture and society. Berlin: Springer Montague R (1974) Formal philosophy: Selected Papers of Richard Montague. New Haven, CT: Yale University Press. 24



From a system of signs to language use

Montague R (1968) Pragmatics. In Thomason RH (ed.) (1974). Formal Philosophy. New Haven, CT: Yale University Press. Montague R (1970) English as a formal language. In: Visentini B et al. (eds) Linguaggi nella Societa et nella Technica. Milan: Edizioni di Communita, 188–221. Moon R (2009) Words, Grammar, Text: Revisiting the Work of John Sinclair. Amsterdam: Benjamins. Morris C (1938) Foundations of the theory of signs. In: Morris, C (1971). Writings on the General Theory of Signs. The Hague: Mouton. Peirce CS (1940) The Philosophy of Peirce: Selected Writings. Buchler J (ed.) London: Routledge. Petöfi JS (1971). Transformationsgrammatiken und eine ko-textuelle Texttheorie. Grundfragen und Konzeptionen. Frankfurt: Athenäum. Post E (1944) Recursively enumerable sets of positive integers and their decision problems. Bulletin of the American Mathematical Society 50: 284–316. Robins RH (1997) A Short History of Linguistics. London: Longman. Sacks H (1992) Lectures on Conversation, Volume I and II. Oxford: Blackwell. Sacks H, Schegloff EA and Jefferson G (1974) A simplest systematics for the organization of turntaking for conversation. Language 50: 696–735 Sapir E (1921) An Introduction to the Study of Linguistics. New York: Harcourt Brace. Saussure F de (1959 [1916]) Course in General Linguistics. Bally C and Sechehaye A (eds) (1983) Translated by Roy Harris. La Salle, Illinois: Open Court. Schillp PA (ed.) (1963) The Philosophy of Rudolf Carnap. La Salle, Illinois: Open Court. Skinner BF (1957) Verbal Behavior. Acton, MA: Copley. Sperber D and Wilson D (1986) Relevance: Communication and Cognition. Oxford: Blackwell. Svartvik J and Quirk R (eds) (1980) A Corpus of English Conversation. Lund: CWK Gleerup. Van Dijk TA (1972) Some Aspects of Text Grammars. A Study in Theoretical Poetics and Linguistics. The Hague: Mouton. Verschueren J, Östman JO and Blommaert J (1995) Handbook of Pragmatics. Manual. Amsterdam: Benjamins. Weigand E (1989) Sprache als Dialog. Tübingen: Niemeyer; 2nd rev. ed. 2003. Weigand E (2009) Language as Dialogue. Amsterdam/Philadelphia: Benjamins. Whorf BL (1942) Language mind and reality. Madras, India: The Theosophical Society. The Theosophical 63(1): 281–291. Wittgenstein L (1953) Philosophical Investigations. Oxford: Blackwell.

25

2 Conversation Analysis Lorenza Mondada

2.1 Introduction Conversation analysis has emerged in the last decades as a key perspective to study social interaction within a variety of disciplines, including sociology where it originated and linguistics where it develops nowadays in a dynamic and innovative way. The chapter offers a sketch of the history of this research paradigm and an overview of its conception of social action, interaction, and language (Section 2.2). Then it synthetizes the specific methodological procedures built on this conception in Section 2.3, concerning audio/video recordings of naturalistic settings, their transcription and their specific forms of analysis. Analysis is undertaken thanks to some key notions presented in Section 2.4, dealing with the organization of turn-taking, sequences, and the overall organization of social interaction. Finally, the chapter ends with some considerations about current challenges raised by the multimodal analysis of video data (Section 2.5). Through the chapter, abundant references are given to the literature in conversation analysis (for which a useful starting point is Sidnell and Stivers, 2012).

2.2  Origins and History Conversation analysis (CA) is a paradigm that emerged in California during the 1960s within sociology—and which has now spread in the United States, Europe, and Japan within various disciplines, such as linguistics, anthropology, and communication. At its origins, CA developed in close relationship with ethnomethodology, a research tradition inaugurated by Harold Garfinkel (1967). Garfinkel reacted against sociology’s abstract models of macro structural entities and Parsons’s model of action as governed by internalized external norms, by proposing instead to re-focus on social action and practical reasoning as constitutive of the way in which action is situatedly produced and rendered intelligible (accountable) in normative terms, as well as, more generally, constitutive of the endogeneous achievement of society, everyday life and institutions alike. Among the founders of CA, Harvey Sacks worked with Garfinkel and was also a student of Erving Goffman, who introduced the very idea of an “interaction order” (Goffman, 1963, 1983). Thanks to the impetus of Harvey Sacks, as well as Emanuel Schegloff and Gail Jefferson, CA focused on social interaction as the primordial locus of social order and engaged in the exploration of its systematic organization. 26



Conversation Analysis

Garfinkel’s ethnomethodology is a mode of inquiry aimed at studying the practical methods of commonsense reasoning that members of a society use in order to produce and to interpret social actions in an orderly and intelligible way. The very name of ethnomethodology suggests this program: it refers to the study (-logy) of the methods—understood as an organized, rational, and systematic (i.e., methodic) set of practices, procedures, or techniques—through which the members of a society (ethno-) achieve their reasoning, talking, debating, decision-making, understanding, and so on in situated and lived social contexts. These methodic practices enable members to achieve social order, in the sense that they locally produce, reproduce, maintain, and transform the intelligibility of facts, events, and situations as they are interpreted and understood in and through situated action. In this sense, ethnomethodology is dedicated to the investigation of the foundations of social order through the study of reasoning practices and, more generally, of the methodic organization of social action. As he was a law student, Harvey Sacks met Harold Garfinkel and consequently moved to study sociology and work with him at UCLA and at the Center for the Scientific Study of Suicide. During this time, Sacks progressively developed his own take on social action, shifting his attention from practical reasoning to a more direct interest for the details of how a given utterance is formatted and how it is interpreted by the next speaker—within an emergent analysis of turns at talk, of what happens next as revealing what has been understood of the previous, and a specification of member’s methods as practices mobilized for and within the methodic organization of social conversation. This exploration emerged from the data Sacks collected both at the Suicide Prevention Centre and in group therapy sessions. One example from the former setting can be useful to situate the questions Sacks was beginning to ask and that would have a tremendous impact on the further evolution of the field. This fragment was used as a material for the very first lecture Sacks gave in the fall 1964: (1) (Sacks, 1992: I: 3) A: This is Mr Smith, may I help you B: I can’t hear you A: This is Mr Smith B: Smith The initial turn at the opening of a telephone call to a help line creates a slot for a next action, in second position, responding to the first. Actions come in pairs, and the second is both projected and expected. Moreover, the second turn could be expectedly “this is Mr Brown”. By saying “this is Mr Smith” at the beginning of his turn, A is soliciting B’s name without explicitly asking for it (“what is your name?” would be a very different action). By saying “I can’t hear you”, B is occupying the slot that has been offered to him, but with another action than the one expected. By initiating repair, B is not claiming that he has hearing problems, but is using the slot for an action that gives the turn back to A—it does the job of “skipping a move” (1992: 7), as does the subsequent repetition of the name. Thus, B mobilizes practices that constitute “ways of providing that the place where the return name fits is never opened” (1992: 7). These practices reveal how the sequential organization of conversation works. This inaugural study constitutes the beginning of what will become CA, studying talk as an object of analysis in its own right, treating talk as action, and showing how detailed arrangements of resources produce the intelligibility of that action and project more actions to come (Schegloff, 1992a: xvii–xviii). 27

Lorenza Mondada

On the basis of such observations, CA progressively grew up, first focusing on the fundamental building blocks of ordinary conversation, offering a model for turn-taking (Sacks, Schegloff and Jefferson, 1974), for sequence organization, adjacency pairs and overall organization (Schegloff and Sacks, 1973), and for repair (Schegloff, Jefferson, and Sacks, 1977). Then, these key aspects have been explored in a range of contexts, showing that ordinary conversation works as the fundamental form of turn and sequence organization, which gets specialized and restricted within other forms of social interaction, namely in institutional settings (Drew and Heritage, 1992; Boden and Zimmerman, 1991). So, whereas the first studies established how conversation as the primordial form of social interaction works, studies of the second generation explored a variety of setting, showing how their institutionality is both characterized and achieved as such by specific sequential formats—like the interviewer asking questions and the interviewee giving answers. In this way, a diversity of institutions have been described, such as doctor–patient interactions, news interviews, court hearings, and so forth. Within a few decades, CA has become a recognized distinct approach to social interaction, further expanding the contexts and the phenomena studied. Beside the study of institutional talk-in-interaction, investigations of professional interactions have been achieved within workplace studies (Heath and Luff, 2000), insisting on the complexity of tasks, environments, material and technological artifacts characterizing work activities. This interest has fueled the analysis of not only talk but more generally embodied action in interaction (Goodwin, 1981; Heath, 1986; Streeck, Goodwin, and LeBaron, 2011) and a development of video methodologies and multimodal analysis. Parallel to the exploration of embodied resources, linguistic resources have been also been more systematically studied within interactional linguistics (Ochs, Schegloff, and Thompson, 1996; Ford, Fox, and Thompson, 2002; Hakulinen and Selting, 2005). Within this field—which has shown how CA, initially originating in sociology, has been successfully implemented in linguistics—grammar (mainly syntax and lexis), and prosody (Couper-Kuhlen and Selting, 1996; Local and Walker, 2005), have been looked at under a new light: not only do they represent resources that are significantly mobilized within sequential organization, but in turn possibilities and constraints characterizing social interaction have been shown to play a configuring role in the emergence of grammar. Thus, conversation analysis has been crucial in demonstrating the systematic organization of social interaction as a constitutive dimension of society, culture and language. First, interaction is the “primordial scene of social life” (Schegloff, 1996a: 4) in the sense that it is “a form of social organization through which the work of the constitutive institutions of social life gets done” (1996a: 4). Second, it is the context in which language is learned by the child and in which s/he also gets socialized; it is the arena in which social norms are acquired and culture is enacted. Third, interaction is the natural habitat of language: social interaction is the ecology within which grammar is shaped in service of the actions it expresses and implements.

2.3 Methodology In a nutshell, CA aims at describing the organization of social activities in their ordinary settings, considering interaction as collectively organized by the co-participants, in a locally situated way, achieved emergently through its temporal and sequential unfolding, by mobilizing a large range of vocal, verbal, visual and embodied resources, which are publicly displayed and monitored in situ. The analysis of these features insists on their indexicality and contingency. This has consequences for the way in which data are collected. The 28



Conversation Analysis

importance of audio and video recordings emerges from these analytical demands: they involve the study of “naturally occurring activities” as they ordinarily take place in social settings, and, consequently, the necessity of recording (Section 2.3.1) and transcribing (Section 2.3.2) situated activities for a detailed analysis (Section 2.3.3) of their relevant endogenous order.

2.3.1  Recordings of Naturally Occurring Social Interactions Over the last decades, CA scholars have worked on an increasingly diverse range of data which has evolved alongside technological developments. Audio and visual recording technologies such as the phonograph and photograph, developed in the nineteenth century, were first used by anthropologists, folklorists and dialectologists in the early twentieth century. In the 1950s, Fries based his The Structure of English (1952) on a corpus of telephone conversations. An interdisciplinary project which began in 1955 titled the Natural History of an Interview, involved the analysis of human behavior based on “spontaneous conversational materials” in “a variety of contexts” (McQuown, 1971: 9–11), and, in particular, a video recording of a psychiatric interview between Bateson and a patient, Doris. However, CA was the first paradigm to systematically ground its empirical analyses in recordings of naturally occurring interactions. Early seminal CA analyses used audio recordings of telephone calls. Sacks’s doctoral dissertation (1966) was based on calls for help that were answered and occasionally recorded by the Los Angeles Suicide Prevention Center. Schegloff’s dissertation on telephone openings (1967) was based on calls to the police collected by Ohio State University’s Disaster Research Center, in connection with the study of organizations under disaster conditions (Schegloff, 1968: 1093, fn. 1). Conversation analysts initially focused on phone conversations simply because they represented a form of naturalistic data easily available for close analysis. Furthermore, because the organization of phone conversations is based on what both parties can hear (and not on embodied resources, to which they have no access) audio-only recordings were sufficient for analytical purposes (Schegloff, 2002: 288). But early CA did not draw exclusively on telephone calls: Sacks, Schegloff, and Jefferson (1974) used a variety of data in their collaborative study of turn-taking, as well as in their individual work: Sacks’s lectures make extensive use of a set of recordings of group therapy; Schegloff’s early work on repair is based on, as he puts it, “a collection of audio- and videotapes of naturally occurring ordinary conversations. The parties to these conversations are diverse—students, housewives, janitors, etc.—as are the ‘contexts’—telephone, co-present, at home, at work, etc.” (1979a: 283; see also Jefferson, 1974: 199, listing the materials used). Moreover, film recordings were also used quite early in the development of CA. In the 1970s, Charles and Candy Goodwin were making film recordings of everyday dinner conversations and other social encounters. Some of these recordings were transcribed by Gail Jefferson and subsequently used from 1973 onward by Jefferson, Sacks and Schegloff in research seminars and publications. In 1975, Schegloff presented a paper co-authored with Sacks at the Annual Meeting of the American Anthropological Association (published much later on, as Sacks and Schegloff, 2002) on “home position”. This is an early attempt to describe bodily action systematically that drew upon film recordings from a variety of settings. In 1977, Charles Goodwin defended his dissertation (published in 1981), in which he presented analyses based on about 50 hours of filmed conversations in various settings (1981, fn. 33). 29

Lorenza Mondada

Further developments of CA were increasingly based on institutional settings: in the 1970s, Atkinson and Drew used tape recordings of court hearings (Atkinson and Drew, 1979), thereby initiating a program of comparative studies of speech exchange systems (compare Drew and Heritage, 1992: 19; Sacks, Schegloff, and Jefferson, 1974: 729). Other institutional settings studied from the 1980s onward include medical interactions in primary care (compare Heath, 1986; Heritage and Maynard, 2006), calls for emergency service (compare Whalen and Zimmerman, 1987) and news interviews (Clayman and Heritage 2002a, Greatbatch, 1988). Moreover, video has been increasingly used in workplace studies for documenting the details of professional interactions in complex environments (Heath and Luff, 2000; Heath, Hindmarsh, and Luff, 2010).

2.3.2 Transcription Another distinctive feature of conversation analysis is the thorough attention given to the transcription of recorded data. Transcribing is both a practical and analytical activity that allows the analyst to pursue the details of talk and embodied conducts and to discover and substantiate the claim that there is order at all points (Sacks, 1984: 22). In this sense, transcribing conforms to the CA conception of social action: it starts from the assumption that “no order of detail in interaction can be dismissed a priori as disorderly, accidental, or irrelevant” (Heritage, 1984a: 241); it engages in a textual and visual representation of talk and embodied conduct that carefully considers the orientation of the participants towards the issue of “why that now?” and “what’s next?” (Schegloff and Sacks, 1973), which characterize the fundamental principles of temporality and sequentiality on which the organization of social interaction is based. In this sense, it is particularly careful in representing the emergent, incremental, situated, and contingent unfolding of action in time. The conventions used by conversation analysis for transcribing talk developed by Jefferson have been based on these considerations and attention for the temporal dimensions of talk, such as silences and overlap (Jefferson, 1973, 1983; Sacks, Schegloff, and Jefferson, 1974). They also have been based on the concern to transcribe what the analyst hears (and participants orient to)—and not to describe it: Jefferson (1985) gives a powerful example of this distinction, based on the transcription of a dirty joke. Here are two possible versions of it: (2a) Jefferson (1985: 28–29, ex. 7) Ken: And he came home and decided he was gonna play with his orchids from then on in. Roger: With his what? Louise: heh heh heh heh Ken: With his orchids. [He has an orchid. Roger: [Oh he h hehheh Louise: ((through bubbling laughter)) Playing with his organ yeah I thought the same thing! Ken: No he’s got a great big[glass house [I can see him playin with his organ hehh hhhh (2b) Jefferson (1985: 28–29, ex. 8) Ken: An’e came home’n decided’e wz gonna play with

30



Conversation Analysis

his o:rchids. from then on i:n. Roger: With iz what? Louise: mh hih hih[huh Ken: [With his orchids.= Ken: =Ee[z got an orch[idRoger: [Oh:. [hehh[h a h ‘he:h]’heh Louise: [heh huh ‘hh PLAYN(h)W(h)IZ O(h)R’N ya:h I [thought the [same Roger [uh:: [‘hunhh’hh’hh Ken: [Cz eez gotta great big[gla:ss house]= Roger: [I c’n s(h)ee]= Ken: =[( ) Roger =[im pl(h)ay with iz o(h)r(h)g’(h)n ’uh When Ken tells that the guy “was gonna play with his orchids” this generates a repair initiated by Roger (“with his what?”), to which Louise responds with a slight laughter— which is transcribed with the annotation of variously aspired syllables—and Ken with a repair (“with his orchids”). Now Roger laughs too, in overlap with Ken’s continuation of the turn and before he has completed his explanation. Louise, in the next turn, formulates again the gist of the joke: in the early transcription, which uses in fact a description of the laughter, Louise’s turn is in the clear and is orthographically rendered in a unique way, making the relation between “orchid” and “organ” totally explicit. In a later transcription, Jefferson shows the payoff of transcribing it: “O(h)R’N” is a very different form, in which laughter “invades the talk” (Jefferson, 1985: 29). Here both “uncontainable” laughter and “difficulty in hearing the punchline” (1985: 29) are constitutive elements of the dirty joke—that get lost in a normalized orthographic transcription and in a description of the laughter. These challenges have been even more important as scholars have begun to transcribe not only verbal but also embodied conducts. Multimodality (including language, gesture, gaze, head movements, body postures, body movements, objects manipulations and the like) complexifies the issue of transcription because it concerns not only the relatively linear course of verbal action, but many courses of both simultaneous and successive embodied conducts. If temporality is fundamental in the transcription of talk, it is even more crucial for the transcription of body movements. This has prompted researchers engaging in a detailed representation of multimodality to develop specific conventions for its notation (Goodwin, 1981; Mondada, 2007) (see below Section 2.5).

2.3.3  Forms of Analysis Within CA, data analysis typically follows two types of procedures. The first: single-case analysis, focuses on an episode of interaction and aims at understanding it in detail, as it unfolds sequentially and temporally, and as it is organized thanks to a variety of resources, formats, sequences and methodical practices. The second: analysis of collections, focuses on a precise phenomenon, defined in terms of a set of resources or formats being used in a specific sequential environment for achieving a specific action. Often, single-case analysis is conducted as an exploratory procedure, especially on complex settings and multimodal sets of resources. The analysis of collections is fundamental for the systematic study of

31

Lorenza Mondada

many instances of more local phenomena, which can be found in one or several datasets (see Schegloff, 1988, 1996b). Collections based on audio data (or on the verbal dimension of recordings) are still the majority: they can be built starting from a form (such as a particle—see for example Heritage, 1984b—a syntactic construction, a format for accomplishing an action, etc.), from a sequential format (see, for example, Schegloff’s early work on identification in conversation openings, 1979b) or from an action (see Schegloff 1996b on “confirming allusions”). Analyses of collections based on video data and taking into account the complexity of multimodal details and embodied conduct are scarcer (but for systematic studies see Goodwin’s pioneering work on gaze at turn beginnings, 1981; Heath and Luff’s study of the closing of bids in auctions, 2007; and Mondada’s study of pointing at turn beginnings, 2007). Naturalistic recordings, detailed transcriptions and fine-grained analysis constitute the basis of the CA methodology. They have allowed both the discovery and the demonstration of substantial phenomena characterizing the organization of social interaction at different levels of analysis, to which we now turn.

2.4  Key Notions Sequentiality is the fundamental principle organizing social interaction. It is defined by Schegloff as “any kind of organization which concerns the relative positioning of utterances or actions” (2007: 2). It concerns the production of turns at talk, as well as of larger stretches of talk, like adjacency pairs and other sequences. Participants orient to the sequential position at which a resource, an action or a sequence is produced (through the underlying question setting up their relevance: “Why that now?”, Schegloff and Sacks, 1973). Sequentiality is closely related to time: actions and turns are characterized by the fact that they are produced in real time by speakers, by adjusting in real time to the responses of their co-participants. Time matters for the complex deployment of sequentiality, which is both oriented progressively—constantly projecting more to come and building normative expectations (generating the underlying question “What’s next?”, Schegloff and Sacks, 1973)—and retrospectively, constantly displaying how previous actions are being treated, interpreted and responded to, consequently making it possible for the first speaker to repair them if they have been misunderstood (Schegloff, 1979a, 1992b). Sequentiality shapes various levels of complexity in the organization of social interaction: turns (Section 2.4.1), sequences (Section 2.4.2), and the overall structure of interactions (Section 2.4.3).

2.4.1  The Organization of Turns-at-Talk The organization of turns-at-talk constitutes the basis of the practices through which participants configure their contributions, smoothly organize turn-taking, normatively orient towards one party speaking at a time, and thus coordinate their action and create or constraint opportunities to participate. Turn-taking has been described in Sacks, Schegloff and Jefferson’s seminal paper (1974) as being structured by two dimensions, the turn-constructional component and the turnallocation component. The first reveals how turn formatting is interactively achieved; the second explains how participants manage the selection of the next speaker. Here is an example taken from the turn-taking paper: 32



Conversation Analysis

(3) Sacks, Schegloff and Jefferson (1974: 707 – exc. 4 [Crandall: 2-15-68 :93]) A: Well if you knew my argument why did you bother to a:[sk. B: [Because I ’ d like to defend my argument. A asks a question to B, and thereby selects him as next speaker. The slightly overlapping response displays that B has identified the possible completion point or transition-relevance place even before A’s turn is brought to completion. This anticipation is made possible by the projectability of the trajectory of the current turn—a possibility that co-participants use to select and start speaking in a finely tuned manner, reducing both overlaps (which is limited here to the last syllable) and pauses. Turn construction has been of fundamental interest for the way in which speakers arrange the resources they mobilize in an incremental and emergent way while building their turn and paying attention to the responses of their recipients. Turn organization orients to the temporal emergence of a unit (turn-constructional unit, TCU) as it reaches a possible completion. This point also represents a transition-relevance place (TRP) at which another speaker can select. Thus, participants constantly monitor the ongoing production of units in order to identify the relevant sequential position at which the current turn will be complete and other turns will be relevant, expected and legitimate—minimizing both gaps and overlaps. In this sense, the production of a turn is a reflexive collective accomplishment, even when the other participants do not utter a word (Goodwin, 1979). The recognition of completion points is achieved by the participants through their online endogenous monitoring, analyzing and parsing of the turn, in which they pay attention to all of the resources used. If linguistic resources have been widely studied in this respect, multimodal resources have been shown to play a crucial role too (see Goodwin, 1979 about gaze; Mondada, 2007 about pointing; Schegloff, 1984 about gesture; Streeck and Hartge, 1992 about [o] face, etc.). Turn allocation becomes relevant as soon as a TRP has been recognized. At that sequential position, participants have a series of opportunities (the speaker might otherselect another participant, another speaker might self-select, the current speaker might continue to speak; Sacks, Schegloff, and Jefferson, 1974). Turn-allocational issues have been of special interest for contrasting what happens in everyday conversation with more specialized speech exchange systems, characteristic of formal occasions (e.g. ceremonies) or of institutional settings (e.g. debates, courtroom interrogations, interviews). While in the former all possible methods for turn allocation are observable, in the latter turn-taking is guided by prearranged formats that constrain and specify the order of speakership, as well as the length, the form and the content of turns (Drew and Heritage, 1992).

2.4.2  The Organization of Sequences The organization of sequences builds upon the alternation of turns at talk. Two subsequent turns might constitute an adjacency pair (Schegloff and Sacks, 1973), if the first action (first pair part, 1PP) projects a second action (second pair part, 2PP) in such a way that the latter is normatively expected (is made conditionally relevant) and, if it is not produced, it is oriented to as observably absent. Adjacency pair is the basic form a sequence can take— prototypical examples are pairs such as greeting/greeting, invitation/acceptance, request/ requested action, or question/answer, like in the following example: 33

Lorenza Mondada

(4) Lerner (2003: 191) exc 12 [Auto discussion]) 1 Curt: Wul how wz the races las’night. 2 (0.8) ((Mike nodes head twice)) 3 Curt: Who w’n [th’feature.] 4 Mike:    [A l w o n,] Curt asks a question (1), which is formulated as a request for an eyewitness account (Lerner, 2003), which implicitly selects the co-participants possessing expertise in racing activities. The qualified participant, Mike, actually responds silently and in an embodied way (2). However, his response is not the expected type of answer, a report—and is thus treated as an absence of (an adequate) second pair part by Curt, who pursues and specifies the initial question (3)—getting an answer (4). So, sequence organization is based on normative expectations not only relative to a next action, but also to the type of action and to the type of legitimate speaker expected to achieve it. Furthermore, sequences are characterized by the fact that the 1PP projects a type of action expected for the 2PP, and can be realized with different polarities: an invitation can be responded to positively by accepting it, or negatively by rejecting it. Interestingly, different formats and arrangements of forms characterize two types of responses: preferred responses tend to be shorter and simpler, and dispreferred responses are longer and structurally more complex (Schegloff, 2007). Pomerantz (1984) has shown how adjacency pairs constitute a basic environment for the systematic positioning and choice of some turn and action format: for example, upgraded assessments are expected after first assessments (such as “gorgeous” responding to “nice”) occasioning a gradual organization of lexical resources, sensitive to their sequential positioning: (5a) Pomerantz (1984: 61) [exc. 18 MC:1] ((“he” refers to a neighborhood dog)) B: Isn’t he cute A: O::h he::s a::DORable (5b) Pomerantz (1984: 76–77) [exc. 22 NB:IV:11.-1] L: D’they have a good cook there? (1.7) L: Nothing special? A: No. – Every- everybody takes their turns In fragment 5a the second speaker responds with a positive assessment, which constitutes an upgrade both lexically and prosodically. Quite the opposite, in fragment 5b, the first interrogative assessment is followed by a silence, projecting disagreement. This is oriented to by the first speaker, who reorients her question, now tilted towards the negative. This permits the second speaker to respond negatively—in this position aligning with the previous turn (although disagreeing with the original formulation). The basic format of the adjacency pair can expand within more complex structures. These expansions have been described according to their position, before, within or after the basic paired actions (pre-expansion, insert expansion, post-expansion; Schegloff, 2007). Beyond sequences that are describable on the basis of expansions of adjacency pairs, there are even more complex activities, such as telling a story, arguing, explaining, and so forth. (Drew, 1995; Heineman and Traverso, 2009; Jefferson, 1988). 34



Conversation Analysis

The three expanded formats of adjacency pairs present rather different features. A preexpansion might check whether a subsequent action can be successfully performed; it consists of a first pair (often a question/answer) prefacing the projected action (the invitation). Pre-expansions are, even more than other types of expansions, related to the “why that now?” question, and lead the participants to consider that some actions are done not for their own sake only, “but for the relevance to and bearing on some action/utterance projected to occur” (Schegloff, 2007: 28). Inserted expansions are distributed between “post-firsts”, which consists of repairs and are oriented backwards to the already produced talk, and “preseconds”, which are oriented forwards. They might be devoted to getting information that is necessary in order to complete the second pair part. Post-expansions can take the form either of minimal post-expansions (like “oh”, “okay”, and acknowledgement tokens; Heritage, 1984b; Jefferson, 1994), or of longer expansions, related to the negotiation of an agreement, the elaboration of a topic or the completion of a complex activity (such as complaining, explaining, narrating, etc.). Sequence organization is oriented towards the progressivity of talk and of the activity. However, the ongoing course of the interaction can be suspended to address possible problems, affecting, suspending, delaying or even interrupting that progressivity. Repair is the practice that allows participants to address the “trouble” problem: how to deal with trouble in speaking, hearing and/or understanding the talk so that the interaction does not freeze in place when trouble arises, that intersubjectivity is maintained or restored, and that the turn and sequence and activity can progress to possible completion. (Schegloff, 2007: xiv) Participants turn to the preference for self-repair (Schegloff, Jefferson, and Sacks, 1977) but the practice can take different forms and formats. First of all, analysts have distinguished between who initiates the repair (self vs. other) and who carries out the repair work proper (self vs. other). Second, repairs can be initiated and achieved at different sequential positions—within the same TCU, the same turn, the transition space, the next turn, in third position, and so forth. Third, repairs (using them as resources) are sensitive to the specific grammatical constraints of diverse languages (see Fox, Hayashi, and Jasperson, 1996). Fourth, repair might be in service of other actions, such as redesigning and fine-tuning ongoing actions, profiling disagreement or trouble, adjusting to epistemic constraints or to conditions that have emerged as inapposite (Drew, 1997; Lerner et al., 2012).

2.4.3  Overall Organization Finally, sequential organization is also visible in the overall organization of social interaction, for instance as it is achieved in the openings and closings of an encounter. Openings are a crucial sequential moment, in which the coordinated participants’ entry into a joint activity is collectively achieved. Openings in telephone calls have been systematically described by Schegloff (1968) through a series of sequences (summons/ answer, identification, greetings, how are you?). Openings of face-to-face interactions have also been widely studied in a large range of languages and situations (for example, medical encounters, Heath, 1981; Heritage and Maynard, 2006), putting to the fore on the one hand the sequential role played by these first exchanges for launching the

35

Lorenza Mondada

turn-taking system, and on the other hand the routinized feature of the lexical and syntactical forms recurrently used at this place. Multimodal approaches to openings based on video recordings are scarcer (but see De Stefani, 2011; Mondada and Schmitt, 2010) and involve a range of multimodal resources achieving the progressive approach and conjunction of the participants. Closings delimit the boundaries of social interaction and are also structured through adjacency pairs (pre-closings followed by proper closings, Schegloff and Sacks, 1973), achieving a smooth conclusion of the topical exchange and the transition to final greetings. In pre-closings, pairs of discourse particles might display that both participants stop to talk topically (such as “okay”/ “okay”; Schegloff and Sacks, 1973). Closings are sequential environments where specific resources might be used (such as tag-­positioned address terms, Jefferson, 1973; idiomatic expressions, Drew and Holt, 1998; specific prosodic patterns, Goldberg, 2004). Closings also mobilize a range of embodied resources, such as the contingent mobilization of drinks in café conversations (Laurier, 2008), body reorientations (De Stefani, 2006), or the dissolution of the interactional space (Mondada, 2009).

2.5  Current and Future Trends: Embodied Conducts in Interaction The current challenges of CA are many, demonstrating the vitality of the field and its expansion towards new topics and issues. One key issue emerging in the field is comparative analysis, raising interesting questions about how to compare situated actions—rather than forms, as linguistic typology has done. This issue has been pursued both across time, by some rare historically oriented studies (Clayman and Heritage, 2002b), as well as by studies dealing with the evolution of impairments in interactions with patients (Wilkinson et al., 2007) and with learning trajectories (Pekarek Doehler, 2010), as well as across space (Haakana et al., 2009; Sidnell, 2009). In both fields, comparison supposes a clear identification of the action targeted, of the sequential environment studied, and of the resources mobilized by speakers. It raises challenges for doing that in a way that both respects the situated organization of these actions in specific contexts and tracks them across contexts. Another key issue that has vividly emerged in the last decade concerns the study of multimodality. On the one side, the role of gesture and gaze as resources for conversation has been recognized for a long time in a variety of disciplines (Kendon, 1990, 2004) but also in conversation analysis (Goodwin, 1981; Heath, 1986; Schegloff, 1984). On the other side, more recent studies have shown that not only gesture and gaze, but the entire body, are concerned by multimodality (which covers all linguistic and embodied resources used for the organization of social interaction, such as language, gesture, gaze, head movements and facial expressions, body postures, and body movements) (Mondada, 2013). Moreover, this implies also an attention for the space in which the participants’ bodies are coordinated (conceptualized as F-formation by Kendon, 1990; contextual configuration by Goodwin, 2000; interactional space by Mondada, 2009) and for more dynamic movements involving mobility (Haddington, Mondada, and Nevile, 2013) as well as more complex praxeological environments like in work spaces (Heath and Luff, 2000) and within multi-activity (Haddington et al., 2014). The advances in this field do not merely concern the addition of more layers of complexity to be analyzed, they rather raise fundamental questions about the detailed organization of multiple temporalities of embodied action, about the way in which sequentiality is configured by multiple simultaneities and successivities characterizing the 36



Conversation Analysis

multimodal formatting of turns and actions, and about the way in which participants demonstrably orient to them. In order to give an example of the kind of issues that multimodality raises, I discuss below two fragments of a guided visit in which participants ask questions. Questions are a type of action that has been substantially studied in conversation analysis, as an adjacency pair which points at issues such as sequence organization, turn format, epistemic asymmetries, and such like (Enfield et al., 2010; Heritage, 2002, 2012; Raymond, 2003; Steensig and Drew, 2008). A multimodal perspective on video recorded instances of questions reveals other aspects concerning their situated and embodied dimension, contributing to the ‘ecology’ of questions. In the next fragment, we join a group of citizens visiting a construction site within a participatory project in urban planning in which they have been involved. Ligour is the architect of the project and guides the group across a park, commenting on the kind of greenery that will develop: (6) (CAB 21-140613 - 1.02.50 là)

1 LIG et qui à terme, (0.3)# donneront une strate de chênes  and which at the end, (0.3) will produce a series of oak trees fig     #fig.2.1 2 *eu::h# .hh *euh:: qui s’installera. eh::m .hh ehm:: which will grow up. *turns fwd-*walks------> fig     #fig.2.2

Figure 2.1 

Figure 2.2 

3 (0.6) 4 LOU → °°là vous avez *une#* petite allée* là?°°* °°there you have a small path there? lig        ->*turns*pivots-----*stops*



fig     #fig.2.3

37

Lorenza Mondada

Figure 2.3 

5 LIG  voilà. y a là y a *une allée, (0.2) qui passe à travers  right. there is over there there is a path, (0.2) which crosses    *LH points----> 6  toute ça: (0.3) euh avec des bancs à l’intérieur,* qui  all this: (0.3) ehm with some benches inside, which    --->*RH--> 7 rejoint le::: le::: #le:: (0.3) le chemin qu’on a vu meets the::: the::: the:: (0.3) the footpath we saw fig    #fig.2.4

Figure 2.4 

8 tout à l’heu*re. a short while ago.    -->*,,,,,,,-->

38



Conversation Analysis

9 (0.2) 10 LOU °d’accord°. * °alright°.      -->* 11 *(2.6) # lig *walks forward-->> fig   #fig.2.5

Figure 2.5 

Ligour finishes his explanation at line 2, which was addressed to a participant behind (Figure 2.1). Verbally, turn completion corresponds with the syntactic completion of the double relative phrase; bodily, turn completion is projected by the fact that the guide turns his head and body forward (Figure 2.2) as he hesitates (2) and then begins to walk away even before he stops talking. So, turn completion is multimodally achieved and exhibited by the speaker (Broth and Mondada, 2013; Mondada, 2015). While he is walking away, a person asks a question (4). This question is prefaced by the deictic adverb “là”/“there” and takes the form of a request for confirmation. The guide immediately turns back (Figure 2.3) and pivots towards the person asking the question, stopping his walk. His answer (5) confirms what Louvier says and points (Figure 2.4) at the footpath she is commenting on. At completion of the answer, Ligour walks forward again. In sum, this short example shows the situatedness and spatial positioning of the question, which not only refers to a place visible and accessible for the participants, but makes relevant a change in the trajectory of the guide’s walk. The answer is given by him stopping and addressing the object mentioned by the questioner. As soon as the sequence is achieved, the guide walks again . The sequence is made recognizable and accountable by the alternating between walking and stopping—and walking again.

39

Lorenza Mondada

This is by far not the only occurrence of the phenomenon, which occurs systematically in the dataset. In the next fragment, the same group is observing the foundations of an aerial walkway above the garden. The architect explains that within a few years the greenery will make the foundations invisible (1-2). (7) (CAB 21-140613 – 1.05.00) 1 LIG et à terme dans deux trois ans, je pense que  and at the end in two three years, I think that 2  la (.) ça se verra plus trop hein  the (.) this will not be too visible anymore right 3 la: la jon*ction >entre les deux 4 (0.2) 5 AUB °°ah bon°° * °°yes well°°   -->*walks---> 6 LIS °+et +là: *eu::::h (.) qu’est ce que c’est* là?° and there: eh:::m (.) what is that there?  +...+points-----> lig   ->*turns back---------------------*stops--> 7 (0.3) • (0.6) lig       •points--> 8 LIG ici? here? 9 (0.4) 10 LIS °oui° °yes° 11 (0.4) 12 LIG  ça:, c’est la:::: la partie: de f- l’autre  this:, this is the:::: the part: of the f- the other 13  partie de fondation y a d- y a •des double:s  part of the foundations there’s there’re double       -->•gesticulates--> 14 comment dire y a des doubles fondations how to say there’re double foundations 15 l’une sur +l’au•tre. one on top of the other.    --->• lis -->+ 16 LIS a::h oui:, d’accord. oh:: yes:, alright. 17 LIG voilà. •y a des semelles très très grandes  right. there are very very big soles    •gesticulates---> 16  (0.3) et %d’autres semelles à l’intérieur,  (0.3) and other soles inside hum   %points-> 19 et encore celui• d’une• autre se•melle. and moreover the one of other soles.    --->•,,,,,,,,,,,,,,,,• 40



Conversation Analysis

20 +(0.6)+ lis +lowers head+ 21 HUM   °et le faux- on voit pas d’ici, le faux  °and the fals- we don’t see from here the false 22 tunnel% en béton qui a été créé juste là-bas? concrete tunnel that has been created just over there?  --->%points w 2H----> 23 (1.0) 24 LIG on va• %all%er vo•ir. we are going to see    •points fwd-• hum     -->%,,,% 25 (0.5) * (0.3) lig     -->*walks fwd---->> 26 LIG c’est >la montée d’la passerelle.< it’s >the access to the footbridge.< 27 (0.3)

As Ligour completes his explanation, he turns forward (2) and begins to walk (5). At that point, another visitor, Lisandre, asks a question (6), prefaced by “et là” and achieved by pointing towards the architectural detail it concerns. As in the previous extract, as soon as the deictic form is uttered, Ligour turns back and stops before the end of the question. Looking back he points at the same spot, initiating a repair of the spatial reference (8), and, as the repair is achieved, engaging in the explanation of the visible detail (12–14). Lisandre displays her understanding (16), but Ligour continues (17–18). Even before Ligour comes to the completion of his explanation, where he retracts his gesture (19), another participant, Humaire, begins to point very early on (17). As the former completes his turn, the latter asks another question (21). Interestingly, this new question, although accompanied by an early pointing towards the greenery, is not “là”-prefaced. Quite the opposite, Humaire refers to a detail which is not visible “here” (18) and which is located “là-bas”/“over there” (22). This question is responded to by Ligour by postponing the answer and by inviting the participants to move on towards the spot referred to. In this way, Ligour responds to these two types of questions in different ways: he stops to answer questions that are referring to a proximal accessible—visible and talkable—detail, but he continues to walk for questions that refer to something not yet visible. His ways of responding is sensitive to the ecology of the question, which gives it its relevance, accountability, and legitimacy. Likewise, the person asking the question is careful in exploiting its situatedness and spatial anchoring in the choice of a specific turn format, prefaced by “là”, further embodied by a pointing gesture. In this sense, the study of multimodality and indexicality enrich the grammar and sequence organization of questions—as of other actions and resources for achieving them.

2.6 Conclusion This chapter has offered a presentation of CA, showing how it has emerged historically, and sketching its original contribution to the study of social interaction. Its originality concerns its conceptual foundations, focused on action as a situated, embodied, temporal, interactional phenomenon; its methodology, motivated by the goal to capture social action as it 41

Lorenza Mondada

naturally occurs in its context; and its analytic mentality, based on the detailed study of the way in which actions, turns, and sequences are formatted. This approach has been consequential for shaping the way in which action is treated in sociology, cultural practices in anthropology, and language in linguistics.

Transcription Conventions For Jefferson’s conventions for transcribing talk, see Jefferson (2004). For multimodal conventions for transcribing embodied conducts, see Mondada (in press, 2014) and https://franz.unibas.ch/fileadmin/franz/user_upload/redaktion/Mondada_conv_ multimodality.pdf.

References Atkinson JM and Drew P (1979) Order in Court. London: Macmillan. Boden D and Zimmerman DH (eds) (1991) Talk and Social Structure. Cambridge: Polity Press. Broth M and Mondada L (2013) Walking away: The embodied achievement of activity closings in mobile interactions. Journal of Pragmatics 47: 41–58. Clayman S and Heritage J (2002a) The News Interview: Journalists and Public Figures on the Air. Cambridge: Cambridge University Press. Clayman S and Heritage J (2002b) Questioning presidents: Journalistic deference and adversarialness in the press conferences of US Presidents Eisenhower and Reagan. Journal of Communication 52: 749–775. Couper-Kuhlen E and Selting M (eds) (1996) Prosody in Conversation: Interactional Studies. Cambridge: Cambridge University Press. De Stefani E (2006) Le chiusure conversazionali nell’interazione al banco di un supermercato. In Y Bürki and E De Stefani (eds) Trascrivere la lingua: Dalla filologia all’analisi conversazionale. Bern: Lang. De Stefani E (2011) “Ah petta ecco io prendo questi che mi piacciono”. Agire come coppia al supermercato. Un approccio conversazionale e multimodale allo studio dei processi decisionali. Roma: Aracne. Drew P (1995) Interaction sequences and anticipatory interactive planning. In E Goody (ed) Social Intelligence and Interaction. Cambridge: Cambridge University Press, 111–139. Drew P (1997) ‘Open’ class repair initiators in response to sequential sources of troubles in conversation. Journal of Pragmatics 28: 69–101. Drew P and Heritage J (1992) Talk at work: Interaction in institutional settings. In P Drew and J Heritage (eds) Talk at Work. Cambridge: Cambridge University Press, 3–65. Drew P and Holt E (1998) Figures of speech: Idiomatic expressions and the management of topic transition in conversation. Language in Society 27: 495–523. Enfield N, Stivers T and Levinson S (eds) (2010) Question–response sequences in conversation across ten languages. Special issue: Journal of Pragmatics 42: 2615–2860. Ford C, Fox B and Thompson SA (eds) (2002) The Language of Turn and Sequence. New York: Oxford University Press. Fox B, Hayashi M and Jasperson R (1996) Resources and repair: A cross-linguistic study of syntax and repair. In E Ochs, EA Schegloff and SA Thompson (eds) Interaction and Grammar. Cambridge: Cambridge University Press. Fries CC (1952) The Structure of English. New York: Harcourt Brace. Garfinkel H (1967) Studies in Ethnomethodology. Englewood Cliffs, NJ: Prentice-Hall. Goffman E (1963) Behavior in Public Places. New York: Free Press. Goffman E (1983) The interaction order. American Sociological Review 48: 1–17.

42



Conversation Analysis

Goldberg JA (2004) The amplitude shift mechanism in conversational closing sequences. In GH Lerner (ed) Conversation Analysis: Studies from the First Generation. Amsterdam: Benjamins, 257–297. Goodwin C (1979) The interactive construction of a sentence in natural conversation. In G Psathas (ed) Everyday Language: Studies in Ethnomethodology. New York: Irvington, 97–121. Goodwin C (1981) Conversational Organization: Interaction Between Speakers and Hearers. New York: Academic Press. Goodwin C (2000) Action and embodiment within situated human interaction. Journal of Pragmatics 32: 1489–1522. Greatbatch DL (1988) A turn-taking system for British news interviews. Language in Society 17/3: 401–430. Haakana M, Laakso M and Lindström J (eds) (2009) Talk in Interaction: Comparative Dimensions. Helsinki: Finnish Literature Society. Haddington P, Keisanen T, Mondada L and Nevile M (eds) (2014) Multiactivity in Social Interaction: Beyond Multitasking. Amsterdam: Benjamins. Haddington P, Mondada L and Nevile M (eds) (2013) Interaction and Mobility: Language and the Body in Motion. Berlin: De Gruyter. Hakulinen A and Selting M (eds) (2005) Syntax and Lexis in Conversation. Amsterdam: Benjamins. Heath C (1981) The opening sequence in doctor-patient interaction. In P Atkinson and C Heath (eds) Medical Work: Realities and Routines. Farnborough: Gower, 71–90. Heath C (1986) Body Movement and Speech in Medical Interaction. Cambridge: Cambridge University Press. Heath C and Luff P (2000) Technology in Action. Cambridge: Cambridge University Press. Heath C and Luff P (2007) Gesture and institutional interaction: Figuring bids in auctions of fine art and antiques. Gesture 7/2: 215–240. Heath C, Hindmarsh J and Luff P (2010) Video in Qualitative Research. London: Sage. Heinemann T and Traverso V (eds) (2009) Complaining in interaction. Special issue: Journal of Pragmatics 41: 2381–2578. Heritage J (1984a) Garfinkel and Ethnomethodology. New York: Polity Press. Heritage J (1984b) A change-of-state token and aspects of its placement. In JM Atkinson and J Heritage (eds) Structure of Social Action. Cambridge: Cambridge University Press, 299–345. Heritage J (2002) Designing questions and setting agendas in the news interview. In J Mandelbaum, P Glenn, C LeBaron (eds) Studies in Language and Social Interaction. Mahwah: Erlbaum, 57–90. Heritage J (2012) Epistemics in action: Action formation and territories of knowledge. Research on Language and Social Interaction 45: 1–29. Heritage J and Maynard D (eds) (2006) Communication in Medical Care: Interaction between Primary Care Physicians and Patients. Cambridge: Cambridge University Press. Jefferson G (1973) A case of precision timing in ordinary conversation: Overlapped tag-positioned address terms in closing sequences. Semiotica 9: 47–96. Jefferson G (1974) Error correction as an interactional resource. Language in Society 2: 181–199. Jefferson G (1983) Notes on some orderlinesses of overlap oßnset. Tilburg Papers in Language and Literature 28. Jefferson G (1985) An exercise in the transcription and analysis of laughter. In TA van Dijk (ed) Handbook of Discourse Analysis Volume 3. New York: Academic Press, 25–34. Jefferson G (1988) On the sequential organization of troubles: Talk in ordinary conversation. Social Problems 35(4): 418–441. Jefferson G (1994) Notes on a systematic deployment of the acknowledgement tokens “Yeah” and “Mmhmm”. Papers in Linguistics 17: 197–216. Jefferson G (2004) Glossary of transcript symbols with an introduction. In Gene H Lerner (ed) Conversation Analysis: Studies from the First Generation. Amsterdam: Benjamins, 13–31. Kendon A (1990) Conducting Interaction: Patterns of Behavior in Focused Encounters. Cambridge: Cambridge University Press. 43

Lorenza Mondada

Kendon A (2004) Gesture: Visible Action as Utterance. Cambridge: Cambridge University Press. Laurier E (2008) Drinking up endings: Conversational resources of the cafe. Language and Communication 28/2: 165–181. Lerner GH (2003) Selecting next speaker: The context-sensitive operation of a context-free organization. Language in Society 32: 177–201. Lerner GH, Bolden G, Hepburn A and Mandelbaum J (2012) Reference recalibration repairs: Adjusting formulations for the task at hand. Research on Language and Social Interaction 45: 191–212. Local J and Walker G (2005) Methodological imperatives for investigating the phonetic organisation and phonological structures of spontaneous speech. Phonetica 62: 120–130. McQuown NA (ed) (1971) The Natural History of an Interview. Chicago University of Chicago Library Microfilm collection of manuscripts on cultural anthropology, no 95 series xv. Mondada L (2007) Multimodal resources for turn-taking: Pointing and the emergence of possible next speakers. Discourse Studies 9/2: 195–226. Mondada L (2009) Emergent focused interactions in public places: A systematic analysis of the multimodal achievement of a common interactional space. Journal of Pragmatics 41: 1977–1997. Mondada L (2013) Embodied and spatial resources for turn-taking in institutional multi-party interactions: The example of participatory democracy debates. Journal of Pragmatics 46: 39–68. Mondada L (2014) Bodies in action: Multimodal analysis of walking and talking. Language and Dialogue 4/3: 357–403. Mondada L (2015) Multimodal completions. In A Deppermann and S Günthner (eds) Temporality in Interaction. Amsterdam: Benjamins, 267–307. Mondada, L (in press). Multiple temporalities of language and body in interaction. Challenges for transcribing multimodality. Research on Language and Social Interaction. Mondada L and Schmitt R (eds) (2010) Situationseröffnungen: Zur multimodalen Herstellung fokussierter Interaktion. Tübingen: Narr. Ochs E, Schegloff EA and Thompson SA (eds) (1996) Grammar and Interaction. Cambridge: Cambridge University Press. Pekarek Doehler S (2010) Conceptual changes and methodological challenges. On language and learning from a conversation analytic perspective on SLA. In P Seedhouse, S Walsh and C Jenks (eds) Conceptualising Learning in Applied Linguistics. New York: Palgrave and Macmillan, 105–127. Pomerantz A (1984) Agreeing and disagreeing with assessments: Some features of preferred/dispreferred turn shapes. In JM Atkinson and J Heritage (eds) Structures of Social Action: Studies in Conversation Analysis. Cambridge: Cambridge University Press, 57–101. Raymond G (2003) Grammar and social organization: Yes/no interrogatives and the structure of responding. American Sociological Review 68: 939–967. Sacks H (1966) The search for help: no one to turn to. Unpublished PhD dissertation, University of California Berkeley. Sacks H (1984) Notes on methodology. In JM Atkinson and J Heritage (eds) Structures of Social Action. Cambridge: Cambridge University Press, 21–27. Sacks H (1992) Lectures on Conversation. 2 vol. Oxford: Blackwell. Sacks H and Schegloff EA (2002 [1971]) Home position. Gesture 2: 133–146. Sacks H, Schegloff EA and Jefferson G (1974) A simplest systematics for the organization of turntaking for conversation. Language 50: 696–735. Schegloff EA (1967) The First Five Seconds: The Order of Conversational Openings. Unpublished PhD dissertation, University of California Berkeley. Schegloff EA (1968) Sequencing in conversational openings. American Anthropologist 70: 1075–1095. Schegloff EA (1979a) The relevance of repair for syntax-for-conversation. In T Givon (ed) Syntax and Semantics Vol. 12. New York: Academic Press, 261–288. Schegloff EA (1979b) Identification and recognition in telephone openings. In G Psathas (ed) Everyday Language. New York: Erlbaum, 23–78. Schegloff EA (1984) On some gestures’ relation to talk. In JM Atkinson and J Heritage (eds) Structures of Social Action. Cambridge: Cambridge University Press, 266–296. 44



Conversation Analysis

Schegloff EA (1988) Discourse as an interactional achievement II: An exercise in conversation analysis. In D Tannen (ed) Linguistics in Context: Connecting Observation and Understanding. Norwood, NJ: Ablex, 135–159. Schegloff EA (1992a) Introduction. In H Sacks (ed) Lectures on Conversation. Oxford: Blackwell. Schegloff EA (1992b) Repair after next turn: The last structurally provided for place for the defence of intersubjectivity in conversation. American Journal of Sociology 95: 1295–1345. Schegloff EA (1996a) Issues of relevance for discourse analysis: Contingency in action interaction and co-participant context. In EH Hovy and DR Scott (eds) Computational and Conversational Discourse: Burning Issues—An Interdisciplinary Account. New York: Springer, 3–35. Schegloff EA (1996b) Confirming allusions: Toward an empirical account of action. American Journal of Sociology 102/1: 161–216. Schegloff EA (2002) Beginnings in the telephone. In JE Katz and M Aakhus (eds) Perpetual Contact: Mobile Communication Private Talk Public Performance. Cambridge: Cambridge University Press, 284–300. Schegloff EA (2007) Sequence Organization in Interaction: A Primer in Conversation Analysis Vol. 1. Cambridge: Cambridge University Press. Schegloff EA and Sacks H (1973) Opening up closings. Semiotica 8/3: 289–327. Schegloff EA, Jefferson G and Sacks H (1977) The preference for self-correction in the organization of repair in conversation. Language 53: 361–382. Sidnell J (ed) (2009) Conversation Analysis: Comparative Perspectives. Cambridge: Cambridge University Press. Sidnell J and Stivers T (eds) (2012) The Handbook of Conversation Analysis. London: Blackwell-Wiley. Steensig J and Drew P (eds) (2008) Questioning. Special issue: Discourse Studies 10: 5–133. Streeck J and Hartge U (1992) Previews: Gestures at the transition place. In P Auer and A di Luzio (eds) The Contextualization of Language. Amsterdam: Benjamins, 135–157. Streeck J, Goodwin C and LeBaron C (eds) (2011) Embodied Interaction Language and Body in the Material World. Cambridge: Cambridge University Press. Whalen M and Zimmerman DH (1987) Sequential and institutional contexts in calls for help. Social Psychology Quarterly 50: 172–185. Wilkinson R, Gower M, Beeke S and Maxim J (2007) Adapting to conversation as a language-impaired speaker: Changes in aphasic turn construction over time. Communication and Medicine 4: 79–97.

45

3 Corpus Linguistics Marina Bondi

3.1 Introduction Corpus linguistics (CL) is the study of language on the basis of corpora, that is, principled collections “of naturally occurring language text, chosen to characterize a state or a variety of a language” (Sinclair, 1991: 171). A corpus is a ‘body’ of language (from Latin corpus, pl. corpora) stored as computer files; it consists of real language data collected for language analysis. The principles used for data collection constitute the basis of the representativeness of the corpus. A corpus of general language typically includes data that are representative of different modes (written, spoken), different types of speakers (considering gender, age, ethnicity, social class, etc.) and different contexts and communicative events (conversation, classroom interaction, government meetings, etc.). Specialized corpora refer to one context or one communicative event only (e.g. business discourse, business negotiations), depending on the potential research interest of the compilers. CL has provided the most impressive results in the field of studies on lexis and grammar. Pragmatics and the study of discourse have initially been slower in picking up the advantages of corpus tools, but they seem to have become one of the most fertile grounds in the past decade or so. Indeed, the greatest resonance of CL can be attributed to figures like John Sinclair and Geoffrey Leech, whose work was also closely related to the study of discourse, semantics and pragmatics. CL has established itself at a time when linguistics was clearly looking beyond the limits of the sentence and aiming to capture the essence of language in use: its techniques are particularly helpful in exploring the interface of lexis and grammar and discovering new patterns of words in combination, but their potential for the study of interaction is also undeniable. Studies on dialogue, and especially studies on language as dialogue (see Weigand, 2009, 2010), privilege holistic views, with human beings at the centre acting and reacting in cultural surroundings: in Weigand’s “mixed game”, for example, natural language is only one of the elements in dialogic communication, also integrating perception (including emotional perception) and cognition. This obviously privileges qualitative analysis and thick descriptions of language use, as language is seen as always performed dialogically for communicative purposes and discourse is seen as communicative (inter)action, whether in speech or in writing. 46



Corpus Linguistics

CL has extended the possibilities of studies on discourse, communication and interaction by adding to their theoretical-methodological perspectives in many ways. Corpus work has first of all provided scholars with large amounts of language data, as well as with computational tools for the analysis of frequencies and recurrent patterns, so that qualitative analysis can be integrated with quantitative data. Different types of corpora have made available different types of data for comparison (across speech and writing or across different communicative situations), thus highlighting the need to go beyond the simple opposition between speech and writing, to explore the many different contexts, communicative situations and dialogic action games in which language realizes its interactional potential. Comparison has allowed a focus on what is typical and what is distinctive in different contexts of language use. A major focus on looking at words in their co-text (i.e. the text that surrounds them) has also favoured the study of specific structures and patterns that are employed in interaction to regulate the organization of language above the level of the sentence. If advances in CL are often related to the development of computers and appropriate software, the origins of corpus work can be traced back to a special attention to language in use. The reasons for a slower uptake of corpus linguistic techniques in the field of spoken discourse analysis seem to lie above all in the complexity of representing spoken discourse and its interactional features in particular, including pragmatic meanings and contextual elements. The great potential of CL for variational analysis, however, should not pass unnoticed, especially in the context of studies on diachrony, sociolinguistics and register. The most prominent contribution of corpus approaches to language description lies probably in comparative analysis, highlighting similarities and differences in frequency and distribution of language. The rest of the chapter explores the following issues: Section 3.2 provides some background on CL, with special emphasis on spoken corpora and tools for corpus analysis; Section 3.3 focuses on the complexity of language in use as seen from the perspective of CL; Section 3.4 focuses on the interactional features of language use in speech and writing, while Section 3.5 highlights the importance of CL for studies on language variation. The Conclusions draw attention to the impact of corpus linguistics in teaching language in a dialogic perspective.

3.2  (Spoken) Corpora, Corpus Annotation and Corpus Tools CL has grown rapidly since the 1960s. The development of the discipline has been facilitated by the increasing storage and processing capacity of computers. The first corpora were collections of both written and spoken data. The earliest examples, such as the Brown Corpus of American English and the Lancaster-Oslo/Bergen (LOB) Corpus of British English, comprising a million words of different genres in writing, were soon complemented by the London-Lund Corpus (LLC) of spoken English (see e.g. Svartvik, 1990).1 The LLC derives from the Survey of English Usage (SEU), launched in 1959 at University College London by Randolph Quirk (and later directed by Sidney Greenbaum) and the Survey of Spoken English (SSE), a sister project started by Jan Svartvik at Lund University in 1975. The potential of corpus linguistics for the study of language variation was soon apparent. Internal analysis of the data collected in each corpus could easily lead to comparative analysis: the Brown and the LOB corpus were comparable in design and comparative analysis offered insights into the differences between American and British English, whereas the LOB and the LLC offered a picture of spoken versus written English. Studying variation across speech and writing was soon found to be extremely fruitful in highlighting convergences and divergences in the use of the language in the two media and problems of transcribing oral data were soon 47

Marina Bondi

prominent. The London-Lund corpus was annotated for prosodic features such as tone units, onsets, location of nuclei, direction of nuclear tones, pitch levels, pauses and stress (normal and heavy), as well as speaker identity, simultaneous talk, contextual comment (ʻlaughs’, ʻcoughs’, ʻtelephone rings’, etc.) and incomprehensible words. The beginning of the 1990s saw the massive growth of general reference corpora, for example, the 100-million-word British National Corpus, the Bank of English (450 million words) and the Brigham Young University Corpus of Contemporary American English (COCA) (450 million words). Availability, of course, has influenced the ease with which materials were collected: if issues of copyright often determined the forms of written discourse to be included, collecting large quantities of spoken discourse required a major recording and transcribing effort. Because of the greater availability of written materials, general corpora are often more easily representative of variation in written than spoken discourse and may provide diverging forms of speech data: the British National Corpus (BNC), for example, includes a lot of informal conversation, while the COCA privileges more formal media interactions. Finally, large general corpora can be seen to be lexically rich but contextually much poorer than specialized corpora (especially spoken corpora, which need to be transcribed and annotated by the researcher). Even if interactivity is by no means restricted to spoken language, it is manifest that the emergence of a number of exclusively spoken corpora was meant to give greater representation to spoken discourse and to provide materials that could be more relevant to scholars of dialogue and interaction. Spoken corpora—corpora of naturally occurring speech compiled for the use of linguists and discourse analysts—are often contrasted with speech corpora, as commonly used in speech technology and containing various forms of elicited data (e.g. Wichmann, 2007). Examples of spoken corpora are the COLT (Bergen Corpus of London Teenagers, later included in the BNC) (Haslerud and Stenström, 1995), the CANCODE (Cambridge and Nottingham Corpus of Discourse in English) (McCarthy, 1998), the LCIE (Limerick Corpus of Irish English) (Farr et al., 2004), the Santa Barbara Corpus of Spoken American English (Du Bois et al., 2000) and the Hong Kong Corpus of Spoken English (HKCSE) (Cheng and Warren, 1999, 2000; Cheng, Greaves and Warren, 2005, 2008). A special role has been played by academic contexts, as for example in the MICASE (Michigan Corpus of Academic Spoken English) (Simpson et al., 2002) and its British counterpart, the British Academic Spoken English Corpus (BASE) (Thompson and Nesi, 2001) or the Irish Li-Bel (Limerick Belfast Corpus of Academic Spoken English). More recently the interest has extended to non-native speakers, considering both interlanguage contexts, as in the LINDSEI corpus (Louvain International Database of Spoken English Interlanguage) (Granger et al., 2002) and English as a Lingua Franca, with the VOICE (Vienna and Oxford International Corpus of English) (Seidlhofer, 2010) and the ELFA corpus (English as a lingua franca in Academia) (Mauranen, 2003). The web has added to the availability of corpora in different ways. On the one hand, many corpora have become available to different kinds of users, including learners and casual browsers (Anderson and Corbett, 2009). On the other hand, the huge repository of textual data available on the web has enhanced the possibility of using the web itself as corpus (e.g. Baroni et al., 2009), building ever larger corpora with materials from different web sources. This has increased the ease with which corpora can be built, and favoured the study of new forms of communication specifically developed on the web (e.g. blogs, micro-blogs, fora, chatrooms, etc.), with their interactional peculiarities. Different research interests may suggest using not only different corpora, but also different corpus analysis software. Most packages—whether standalone like MonoconcPro, 48



Corpus Linguistics

Wordsmith tools (Scott, 2017), ConcApp (Greaves, 1996) or Antconc, or web-based, like Sketchengine, the Cobuild concordance sampler or View—will offer wordlists with frequency (and distribution) of individual items (or strings of words), as well as the classic KWIC (Key Word In Context) concordances, “the most basic way of processing corpus information” (Hunston, 2002: 38). Wordlists reveal the range and frequency of occurrence of language items. Concordances—retrieving all the occurrences of a word or expression in the corpus and showing the node word at the centre of the concordance line in their lexico-grammatical, semantic and pragmatic environment—are particularly useful in the study of collocation, the tendency of two words to occur near or next to each other in discourse (Firth, 1957). Specific types of software have been developed for specific research purposes: Concgram (Greaves, 2009), for example, has been created to find all the cooccurrences of words in a text or corpus irrespective of constituency variation and positional variation, so as to reveal the full phraseological profile of a text or corpus (Cheng, Greaves and Warren, 2006). The compilation of spoken corpora is more expensive and time-consuming than that of written corpora. Ethical issues and permissions further complicate the question. Recording quality has recently improved impressively, but the most complex task remains the definition and transcription of relevant features. The design of a transcription system is an obvious element of collecting spoken data, starting from the decision whether to adopt an orthographic, prosodic or phonetic transcription and to what nonverbal data to include: contextual information, paralinguistic features, pauses, overlaps, and such like. Basic transcription focuses on the production of a written record of the audio or video recordings. Coding, or annotation, involves the process of recognizing, analyzing, and taking note of phenomena that are perceived as relevant in transcribed speech. Annotation has become standard practice in CL in general, to allow use of plain text corpora with different platforms. One of the best-known standards is TEI (Text Encoding Initiative), using a form of SGML (Standard Generalized Markup Language) with standardized codes within diamond brackets, such as

representing the beginning of a new paragraph. This also allows encoding features of spoken interaction typical of any transcription of spoken data (laughter, pauses, overlapping speech, outside noises). Automatic taggers have been developed to tag every single word in a corpus with part-of-speech information (POS-tagging) or semantic field, to allow for disambiguation of homographs. Software may combine different types of annotation functions: Wmatrix (Rayson, 2009), for example, combines POS-tagging and semantic tagging and allows for the identification of the dominant conceptual domains of a corpus or text. Pragmatic annotation is still more experimental, as automatic assignment still lacks precision. Programs exist for manual speech-act annotation, allowing the study of co-occurrences of speech acts and language forms. Manual coding of large corpora is obviously particularly time-consuming, but speech act annotation has moved from a mostly theoretical issue to an area of great advances in semi-automated systems (see Weisser, 2015—author of freeware DART, Dialogue Annotation and Research Tool—for a recent discussion of the issues involved). If all corpora need some sort of mark-up system identifying “metadata” (e.g. file name, author(s), title, date of publication, genre, intended audience, etc.), corpora to be used for the study of dialogic language use will tend to require different layers of (often ad hoc) annotation. Corpus builders may annotate the corpus with information about the identity of their speakers (sex, age, roles and relationships, location, etc.) or other features of the event (date, genre, register), or features of multimodality in interaction, such as visual or auditory 49

Marina Bondi

elements: rhythm, stress, intonation, but also gesture, gaze, head movements, facial expressions and posture. The number of corpora providing prosodic annotation has grown over the years, but the design and the extent of the annotation are often related to the aims of the research that initiated them. The Hong Kong Corpus of Spoken English (Cheng et al., 2005, 2008) is an interesting example of a prosodically transcribed corpus showing specific attention to discourse features. The corpus contains 108 hours of spoken discourse (academic interactions, business interactions, conversations and public discourse) and is annotated for different intonation choices for each tone unit (prominence, tone, key, termination), following Brazil’s communicative model of intonation in English (Brazil, 1997). Prosody plays a major role in determining the function of many expressions and prosodic annotation can advance our understanding of core units of meaning. In some corpus projects based on spoken discourse—as for example the C-ORAL corpora of romance languages (Cresti and Moneglia, 2005)—prosody is regarded as a central element of language description, defining the boundaries of the utterance (taken as the primary unit of analysis) and signalling the most important functional linguistic levels of speech, such as illocution, attitude and information structure (see Raso and Mello, 2014 for an extensive exploration of these issues in spoken corpora of very different languages.) Although the issue of linking transcripts to the original recordings has been long debated (Roach and Arnfield, 1995), the procedures for aligning audio and transcription have only recently become an important element of systems for processing spoken corpora. The MICASE, for example, only includes the audio files, the Santa Barbara Corpus also includes alignment files. The alignment of audio streams to transcript of spoken data has obvious advantages for functional analysis and can contribute substantially to the description of spoken discourse, but even text-to-sound alignment leaves out many other features. Following groundbreaking work on multimodality (Baldry and Thibault, 2006), aligning transcript and audiovisual recordings has become an important frontier. Adolphs and Carter (2013), for example, illustrate the important role of gesture in communication using the Nottingham MultiModal Corpus (NMMC), which includes transcripts and video data of academic lectures and PhD supervision sessions. The architecture and the interface of the Digital Replay System devised for the project provides a tool for the analysis of head and hand movements, revealing for example how head nods can act as ‘back channels’ intended to offer relevant feedback to speakers. The meaning of individual gesture types, however, is not easily discernible outside a sequence of gestures, and the addition of hand movements further complicates the issue of coding. In spoken discourse analysis, it is not only difficult to separate problems of assembling spoken corpora from questions of annotating them, but also from the perspective of processing and archiving spoken language to make it accessible for the larger public. The issue of corpus usability and sharing has developed increasing importance in spoken discourse studies, bringing together scholars whose work falls primarily in linguistic approaches to speech and scholars who specialize in developing infrastructures for spoken corpora (see Ruhi et al., 2014 for an overview of best practices). Appropriate systems have been specifically devised to support the creation and analysis of spoken language corpora: EXMARaLDA, for instance, has been created to facilitate pragmatic analysis of spoken language corpora and provide tools for synchronizing the recording and the navigation in the transcript, including the complexities of multi-part interaction, different linguistic levels and different modalities of interactional behavior (Schmidt and Wörner, 2009: 566–567). 50



Corpus Linguistics

3.3  Corpus Linguistics and Language in Use The main original thrust of corpus linguistics can be seen as shifting the emphasis from what is possible in the language to what is probable, and developing models based on principles of probability (rather than possibility) that can help tackle the complexity of language use. The main results in the field of CL were initially linked to lexicography (see for example the COBUILD project and the dictionaries produced in that context), but the interest rapidly extended to the study of grammar and to the production of reference grammars (see Biber et al., 1999) as well as to contrastive studies (Johansson, 2007). Corpus-based research has been particularly influential in integrating the study of grammar and lexis. Growing attention has been paid to observing and describing phraseological units combining grammatical and lexical patterns. In a Sinclairian perspective, the “idiom principle” (Sinclair, 1991) views language as essentially made up of strings of co-selected words, that is, “semi-preconstructed phrases that constitute single choices, even though they might appear to be analysable into segments” (Sinclair, 1991: 110), thus paving the way for the study of phraseology as the ideal starting point for the exploration of systematic relations between text and form (Sinclair, 2004). The key role played by phraseology in major approaches to corpus linguistics is also linked to the problem of integrating semantic and pragmatic meanings. The vast majority of Sinclair’s work from the 1990s onward, for example, shows very well how attention to frequency and patterns highlights the existence of systematic relations in texts on a functional or semantic basis. Moving beyond the Firthian notions of collocation (a frequent co-occurrence of words) and colligation (the co-occurrence of grammatical choices) (Firth, 1957), Sinclair postulates the notion of semantic preference (the tendency of a word to occur in the context of words sharing the same semantic features) (Sinclair, 2004: 32–33) and semantic prosody (the further possibility of words to correlate with attitudinal, pragmatic and functional choices) (Sinclair, 2004: 33–34). Attention to these patterns shows that “the word is not the best starting point for a description of meaning, because meaning arises from words in particular combinations” (Sinclair, 2004: 148). The emphasis is on an ‘empty lexicon’, where “a word becomes associated with a meaning through its repeated occurrence in similar contexts” (Sinclair, 2004: 161) and a ‘holistic orientation’ is advocated in language studies, looking at messages as “not initially separated from the particular expression used in the interactive circumstances of the utterance” (Sinclair, 2004: 163). Sinclair’s work is not only noticeable for developing the theoretical implications of both a corpus approach to language and the study of dialogic language use: it is also explicitly focused on dialogicity, or the “internalization of dialogue”. He develops the hypothesis that “much of the complexity of sentence grammar can be explained as the internalization of features of spoken interaction” (Sinclair, 2004: 102), concluding that it is “unsafe to separate grammar below the main clause from grammar above it” (Sinclair, 2004: 114), and ultimately from a grammar of discourse. The different contributions offered by his work in discourse analysis and corpus linguistics can be seen to result in the integration of linear and hierarchical (strategic) perspectives on language. Making sense of text is seen as including: “the ability to apply a hierarchical model to the linear string”, “the ability to prospect, i.e. to look ahead for features that might help the interpretation” and “the ability to hold provisional interpretations in mind and to abandon them if they are superseded by more plausible ones” (Sinclair and Mauranen, 2006: xiii). The Linear Unit grammar proposed by Sinclair and Mauranen (2006) is an attempt to describe both written and spoken language by organizing the text into units for further 51

Marina Bondi

analysis. The basic distinction is between organizational units, action-oriented units organizing discourse, and message-oriented units, units of incrementation of the topic, reflecting the complex interactional work on the part of the participants, in a perspective that integrates consideration of speech and writing. A growing body of literature has focused on the integration of corpus and discourse approaches (Partington and Morley, 2004; Baker, 2006; Ädel and Reppen, 2008; Jucker et al., 2009). This means relating textual practices to language choice, so that statements about language use can be supported with reference to data and, on the other hand, corpus data are not only described but also interpreted in terms of textual structure and social action (Bondi, 2008). As I have claimed elsewhere, adapting the corpus-based versus corpusdriven distinction (Tognini Bonelli, 1993), “corpus work can be seen as both ‘catalysing’ the analysis (in a corpus-driven direction) and supporting the interpretation of data in terms of discourse (in a corpus-based direction): the two perspectives co-construct the research process’ (Bondi, 2014a: 44). Spoken corpus studies often refer to conversation analysis as an interpretative paradigm: Walsh et al. (2011), for example, show how CL and conversation analysis (CA) can be used together to provide powerful insights into how interactants establish understanding in spoken interaction, thus overcoming potential limitations of both, where the risk is for CL to ignore context and for CA to be unable to generalize. The process is seen as one of “interdependence” (Walsh et al., 2011). CL has been increasingly applied to specific contexts and domains of language use, and corpus approaches to language analysis have greatly contributed to register studies, discourse analysis and genre analysis. Specialized corpora have been used to describe the practices of particular discourse communities, ranging from courtroom and forensic contexts (Cotterill, 2010) to business discourse (Koester, 2006; Handford, 2010), political discourse (Ädel, 2010), advertising and the media (O’Keeffe, 2006), healthcare discourse (Adolphs et al., 2004) and classroom and educational contexts (O’Keeffe and Farr, 2003; Walsh, 2006; Walsh et al., 2011; Farr, 2003, 2008). The area of academic discourse has probably witnessed the greatest upsurge of corpus-assisted discourse and genre analysis, focusing primarily on the research article, but also including a wide range of genres such as reviewing genres (Hyland and Diani, 2009), abstracts (Bondi and Lorés Sanz, 2014), student writing (Nesi and Gardner, 2012) and academic lectures (Csomay, 2002; Thompson, 2006). Dialogic aspects of academic communication have also been paid particular attention, as will be seen in the next section.

3.4  Interactivity in Spoken and Written Discourse The integration of different perspectives is essential to dialogic approaches to language, where the attempt is to capture the ever-changing complexity of linguistic performance. CL is often used as a tool in combination with other interpretative tools drawn from the study of discourse, pragmatics or interaction. Corpus pragmatics, though recently called “a relative newcomer on the pragmatics and corpus linguistics scene” (Rühlemann and Aijmer, 2015: 1), shows evidence of growing in both synchronic (Adolphs, 2008; RomeroTrillo, 2008) and diachronic studies (Taavitsainen et al., 2014). The lack of form/function correlation in pragmatic phenomena has influenced the focus of research combining the tools of corpus linguistics and pragmatics: this has either focused on expressions with a reflexive function (typically identifying a speech act or a pragmatic function, such as sorry or I mean) or required specific annotation and close analysis of contextual variables that could reveal discourse patterns as well as lexico-grammatical ones. 52



Corpus Linguistics

Looking at language as dialogue draws attention to the language resources that bear traces of interactivity in both spoken and written language, the means by which interaction is created and sustained. If action (and reaction) is still the central theoretical concept of dialogicity, language studies often look at the resources that realize this dialogic dimension in text, bearing traces of the interpersonal and interactional nature of discourse. The ease with which corpus analysis techniques provide information on lexico-grammatical features has often determined the specific focus of many analysts: if a study of hesitations requires manual annotation of what counts as hesitation, studies of specific expressions or grammatical features can be based on rough data or POS-tagged data. Studies of interactional elements of language have produced extensive results in fields that can be tracked down to specific expressions. Pragmatic markers are a case in point. Corpus studies of speech have often focused on single pragmatic or discourse markers or specific functional categories with easily identifiable markers. Following Schiffrin’s (1987) seminal work on “discourse markers”, corpus work has given great attention to “discourse markers” (Jucker and Ziv, 1998; Pons Bordería, 2008), “pragmatic markers” (Aijmer, 2002) and “discourse particles” (Fischer, 2006), often using different terms to identify the vast area of expressions of discourse relations or speaker’s stance. The class itself is more often defined in terms of a core and periphery (Pons Bordería, 2008) than of necessary and sufficient conditions. The multifunctional nature of discourse markers is further complicated by the range of units they can refer to: the speech situation, the utterance act, the illocutionary act, the propositional act, knowledge, other mental processes (e.g. evaluation, planning etc.), cooperation (interaction-management) (Rehbein et al., 2007: 4). Some of the earliest corpus-based studies of pragmatic markers were based on the LLC or the COLT (Stenström, 1990; Andersen, 2001; Aijmer, 1996, 2002), but the development of the field has greatly profited from being able to compare different contexts of use in a variational perspective (Aijmer, 2011a), ranging from register studies (Biber et al., 1999) and cross-cultural and cross-linguistic analyses (Aijmer and Simon-Vandenbergen, 2006) to comparing native-speaker use with English as a lingua franca (Baumgarten and House, 2010) or learner language (Müller, 2005; Fung and Carter, 2007; Gilquin, 2008; Aijmer, 2011b). A growing body of studies have also been based on the analysis of diachronic corpora (see e.g. Lenker and Meurman-Solin, 2007 for a range of approaches). Similarly, work on written and spoken academic discourse has often paid particular attention to interactional perspectives and polyphony: the original influence of Bakhtin’s (1981) notion of dialogism in writing has drawn attention to the heteroglossia and internal stratification of any language, the dynamic process of taking and using another person’s words, and the different degrees of distance that one may assume from one’s own discourse, for example through the resources of appraisal (Martin and White, 2005). Dialogic approaches combine the Bakhtinian notion of dialogism with attention to linguistic and communicative perspectives, as well as attention to the role played by argumentation in discourse, as shown for example by Livnat (2012) in her work on written scientific discourse or Rowley-Jolivet and Carter-Thomas (2005) on spoken discourse Corpus-informed studies of spoken academic discourse have often looked at markers of textual interaction. Lectures, for example, have been studied extensively in their resources for textual structuring (Thompson, 2003; Lorés Sanz, 2006; Nesi and Basturkmen, 2006), audience orientation (Ädel, 2012; Bamford, 2005a, 2009), markers of identity and evaluative language use (Anderson and Bamford, 2004; Bamford, 2005b; Deroey and Taverniers, 2012; Diani, 2008, 2013; Fortanet, 2004a, 2004b; Okamura, 2013) as well as general interactive features (Crawford, 2007, 2008). 53

Marina Bondi

Studies of academic writing have also paid increasing attention to interactive aspects. Dialogicity has been strategic in the definition of the notion of ‘writer’s voice’ (Fløttum et al., 2006; Bondi, 2007, 2009, 2012, 2014b; Hyland and Sancho Guinda, 2012) and of ‘metadiscourse’, discourse about the evolving discourse (Bamford and Bondi, 2005; Hyland, 2005b; Ädel and Mauranen, 2010). Following Tardy (2012), authorial voice can be seen to have individual aspects (uniqueness, authenticity and authoritativeness), social aspects (repertoires of self-representation and authorial presence), and dialogic features (interaction between the individual and social dimension, including writer–reader interaction). The multiplicity of textual voices in academic writing encompasses the virtual voices of readers, thus highlighting the enhanced dialogism of argumentative discourse: writers often address potential objections by showing temporary agreement with claims, which are then refuted, or revised in subsequent claims. In Hyland’s terms, authorial voice draws attention to expressions of both stance—the writer’s expression of personal attitudes and assessments—and engagement—the ways writers textually establish the presence of their readers in discourse (Hyland, 2005a), involving the most overt interactional features of metadiscourse (reader’s mentions, directives, questions, etc.) but also other choices accommodating the audience’s “probable knowledge, interests, rhetorical expectations and processing abilities” (Hyland, 2005b: 49). Hyland’s work, combining corpus and discourse tools, has been very influential in stimulating studies on discourse identities (Hyland, 2012) and the inherent dialogicity of writing (e.g. Gil-Salmon and Soler-Monreal, 2014), as well as on variation across genres and disciplines (Hyland and Bondi, 2006). As I have shown elsewhere (Bondi, 2010, 2012), the study of authorial voice and metadiscursive practices requires a dialogic approach to academic writing and attention to how signs of reader–writer interaction vary across texts and genres. Introductions, for example, often frame discourse through semantic sequences (Hunston, 2008) of ‘attributions to personal or locational sources’ (e.g. I, we, you, in Section 3.2), ‘references to discourse or research procedures’ (e.g. examine, consider, discuss) and nominal elements pointing to a ‘cognitive construct’ (advantages, characteristics, etc.). Textbooks, however, favour personal forms (we discuss) and combinations highlighting topic-setting (look at notions) and explanation (provide examples, explain concepts), whereas research articles favour non-personal forms (Section 3.1 discusses) and combinations highlighting purpose (present model) and research structure (test hypotheses, review literature, provide results). Similarly, a comparative analysis of research articles and popular articles in history (Bondi, 2014a) shows that deictic pronouns I and you establish different interactional patterns in popular and academic history. Popular texts make extensive use of you to engage readers in the text, draw them into the time frame of history and guide them towards the selected perspective, while mainly using first person pronouns to construct the identity of ‘us today’. Academic journals, on the other hand, use forms of self-mention to enhance authorial visibility and highlight the role of the interpreting scholar, treating the reader as a partner in the development of arguments from premises to conclusions.

3.5  Corpus Linguistics and Language Variation across Speech and Writing The study of language variation contributes to a better understanding of language as dialogue by looking at how different speech events or genres realize specific communicative action, thus bringing together the focus on the complexity of language use and the interactional view of language. Dialogic perspectives are interested in how human beings use language in (both written and spoken) interaction, in ways that cut through the written–spoken 54



Corpus Linguistics

distinction. The difference between speech and writing is often a major focus of attention, although the idea is that there is a cline or spectrum of variation between what is prototypically written and prototypically spoken, involving a range of different communicative event types. In a dialogic perspective, the distinction between speech and writing will obviously not be clear-cut, as the specific type of “dialogic action game” (Weigand, 2009, 2010) becomes prominent, with the type of interaction and the conditions of production influencing, for example, the degree of planning and the features of language that may be related to impromptu production. Most notable in the context of corpus-based studies of written and spoken registers is the work carried out by Biber and his associates. Following pioneering work (Biber, 1988) on a multidimensional analysis of the linguistic features that co-occur in texts with markedly high frequency, the “Flagstaff school” has produced extensive studies of spoken and written academic discourse (Biber et al., 2002; Biber, 2006; Biber et al., 2007; Biber and Barbieri, 2007; Cortes and Csomay, 2015), mostly using the T2K-SWAL corpus (TESOL2000 Spoken and Written Academic Language). The corpus includes a range of spoken and written registers that students encounter in US universities: institutional writing, syllabi, course packs, textbooks, service encounters, study groups, office hours, labs, class management, classroom teaching. Attention has also been paid to including different levels of interactiveness when collecting class sessions. The corpus thus allows different types of comparisons and a systematic mapping of the general distribution of lexical and grammatical features across the different registers. Multidimensional analysis helps identify the linguistic features that characterize each register in terms of a few dimensions of variation, such as oral versus literate discourse, procedural versus content-focused discourse, reconstructed account of events, teacher-centred stance (Biber, 2006: 185). The analysis shows, for example, that spoken and written instructional registers (textbooks and classroom teaching) have very few shared lexico-grammatical characteristics. E-discourse (electronic discourse) and CMC (computer mediated communication) provide an interesting testing ground of the impossibility to keep a clear-cut distinction between speech and writing, as well as of the need to consider the extremely wide range of possible combinations in the development of different forms of human interaction and hybridization of spoken and written discourse. The notion of literacy has long been extended to the new media, in explorations of the way electronically-mediated communication has affected our language practices (Baron, 2008, 2015). Early studies on web discourse focused on the impact of the medium on discourse—its multimodality, hypertextuality, granularity, or the hybridization of spoken and written discourse in its “silent orality” (Soffer, 2010). It is now the time to look more closely into how the extended participatory framework of the Web influences both language choice and communicative practices by increasing the range of interactive patterns and discursive identities construed in each text. The web poses a number of interesting questions of discourse pragmatics (see Herring et al., 2013). Multilingualism and language change are in action all the time and virtual discourse communities are rapidly created, extended and maintained through shared knowledge and the forms of communal bonding (Yus, 2011: 110) that characterize social networks. The Web has also generated a wide range of new specific communicative genres and to new forms of dialogue in particular (chats, fora, blogs, etc.). Blogs and fora, for example, have proved to be interesting arenas for persuasion and exchange of opinions. From the point of view of analysts, they have provided excellent material for the study of writer– reader interaction such as expressions of writer’s stance, reader’s engagement and identity construction (see Part II of Gil-Salmon and Soler-Monreal, 2014). The ‘evaluatively 55

Marina Bondi

charged’ nature of blogs’ posts—and comments especially—makes them ideal territory for an analysis of ‘subjectivity markers’ and evaluative language use (Bondi and Seidenari, 2012). The self-disclosure elements typically characterizing personal blogs have generally survived genre migration over to news, corporate or research blogs. Research blogs have introduced new collaborative practices in academic discourse: “Linguistically there is much of the informality and spontaneity of spoken language. The unforeseen practice of involving audiences in open commentary means that unknown, heterogeneous, and varied audiences may participate in co-constructing research debates” (Mauranen, 2013: 30–31).

3.6 Conclusions The chapter has analyzed the contribution of CL to the study of dialogue and language as dialogue, with a focus on English linguistics and academic discourse studies. The impact of corpus linguistics is seen primarily as that of an approach that has extended the possibilities of the study of dialogic language use and added to its theoretical-methodological perspectives, while highlighting the presence of markers of interactivity in both written and spoken discourse. The preliminary overview of the developments of CL has shown a growing interest in capturing the complexity of spoken discourse and pragmatic perspectives, as reflected in the corpora made available to other researchers and the analytical tools developed. These have moved towards the inclusion of different aspects of human interaction and more flexible tools for research. The main contribution of CL can be seen in its attention to language in use and to probabilistic studies of language, as well as in new approaches to how meaning is created in text and how text ‘internalizes’ interaction. Studies of language in use have also pointed towards another important element of dialogic approaches: the need to pay attention to the different ‘action games’ that constitute our interactional abilities, beyond simplistic distinction between monologue and dialogue or spoken versus written language. In the field of academic discourse studies, the impact of CL has also been extremely relevant in teaching and learning practices, but a discussion of this issue is well beyond the limits of the present chapter. It will suffice to say that the close connection between language research and language teaching in the field (see Campoy and Luzón, 2007 on spoken corpora) has often resulted in practices that combine activities oriented to ‘noticing’ with ‘task-based’ communicative practice (see Bondi, 2014a). Emphasis on communication highlights the need to provide students with meaningful and authentic tasks and looks at text as a hierarchical sequence of rhetorical structures, whereas corpus work can contribute to discovery learning and to developing learners’ autonomy through observation and hypothesis formulation by looking at text as a linear sequence of meaning units. The integration of the two perspectives may contribute to a better understanding of textual competence and of different learning styles, variously integrating top-down and bottom-up processes. Whether this integration may fit with the idea of ‘competence-in-performance’ (Weigand, 2009) remains to be discussed.

Note 1 Given the original input provided by English studies to CL, my examples will primarily (but not exclusively) refer to English language corpora. The field has also grown so extensively that no attempt can be made at mapping it exhaustively. The area of academic discourse has also been privileged for exemplification, for the impact that corpus studies have had on its development. 56



Corpus Linguistics

References Ädel A (2010) How to use corpus linguistics in the study of political discourse. In O’Keeffe A and McCarthy M (eds) The Routledge Handbook of Corpus Linguistics. London: Routledge, 591–604. Ädel A (2012) What I want you to remember is… : Audience orientation in monologic academic discourse. English Text Construction 5(1): 101–127. Ädel A and Mauranen A (eds) (2010) Metadiscourse. Special issue: Nordic Journal of English Studies 9(2). Ädel A and Reppen R (eds) (2008) Exploring Discourse through Corpora. Amsterdam: Benjamins. Adolphs S (2008) Corpus and Context: Investigating Pragmatic Functions in Spoken Discourse. Amsterdam: Benjamins. Adolphs S and Carter R (2013) Spoken Corpus Linguistics: From Monomodal to Multimodal. London: Routledge. Adolphs S, Brown B, Carter R, Crawford P and Sahota O (2004) Applying corpus linguistics in a health-care context. Journal of Applied Linguistics 1(1): 9–28. Aijmer K (1996) Conversational Routines in English: Convention and Creativity. New York: Longman. Aijmer K (2002) English Discourse Particles: Evidence from a Corpus. Amsterdam: Benjamins. Aijmer K (2011a) Understanding Pragmatic Markers: A Variational Pragmatic Approach. Edinburgh: Edinburgh University Press. Aijmer K (2011b) Well I’m not sure I think …: the use of well by non-native speakers. International Journal of Corpus Linguistics 16(2): 231–254. Aijmer K and Simon-Vandenbergen A-M (eds) (2006) Pragmatic Markers in Contrast. Oxford: Elsevier. Andersen G (2001) Pragmatic Markers and Sociolinguistic Variation: A Relevance-Theoretic Approach to the Language of Adolescents. Amsterdam: John Benjamins. Anderson L and Bamford J (eds) (2004). Evaluation in Oral and Written Academic Discourse. Rome: Officina Edizioni. Anderson W and Corbett J (2009) Exploring English with Online Corpora. Basingstoke: Palgrave. Baker P (2006) Using Corpora in Discourse Analysis. London: Continuum. Bakhtin M (1981) The Dialogic Imagination: Four Essays (Michael Holquist ed). Austin/London: University of Texas Press. Baldry A and Thibault P (2006) Multimodal corpus linguistics. In Thompson G and Hunston S (eds) System and Corpus: Exploring Connections. London: Equinox, 164–183. Bamford J (2005a) Interactivity in academic lectures: The role of questions and answers. In Bamford J and Bondi M (eds) Dialogue within Discourse Communities: Metadiscursive Perspectives on Academic Genres. Tübingen: Niemeyer, 123–145. Bamford J (2005b) Subjective or objective evaluation? Prediction in academic lectures. In Tognini Bonelli E and Del Lungo Camiciotti G (eds) Strategies in Academic Discourse. Amsterdam: Benjamins, 17–29. Bamford J (2009) Patterns of description in lectures in science and technology. In Radighieri S and Tucker P (eds) Point of View: Description and Evaluation across Discourses. Rome: Officina Edizioni. 195–210. Bamford J and Bondi M (eds) (2005) Dialogue within Discourse Communities: Metadiscursive Perspectives on Academic Genres. Tübingen: Niemeyer. Baron N (2008) Always On: Language in an Online and Mobile World. Oxford: Oxford University Press. Baron N (2015) Words Onscreen: The Fate of Reading in a Digital World. Oxford: Oxford University Press. Baroni M, Bernardini S, Ferraresi A and Zanchetta E (2009) The Wacky Wide Web: A collection of very large linguistically-processed Web-crawled corpora. Language Resources and Evaluation 43(3): 209–226. Baumgarten N and House J (2010) I think and I don’t know in English as a lingua franca and native English discourse. Journal of Pragmatics 42: 1184–1200. 57

Marina Bondi

Biber D (1988) Variation Across Speech and Writing. Cambridge: Cambridge University Press. Biber D (2006) University Language: A Corpus-Based Study of Spoken and Written Registers. Amsterdam: Benjamins. Biber D and Barbieri F (2007) Lexical bundles in university spoken and written registers. English for Specific Purposes 26: 263–286. Biber D, Connor U and Upton TA (2007) Discourse on the Move: Using Corpus Analysis to Describe Discourse Structure. Amsterdam: Benjamins. Biber D, Conrad S, Reppen R, Byrd P and Helt M (2002) Speaking and writing in the university: A multidimensional comparison. TESOL Quarterly 36(1): 9–48. Biber D, Johansson S, Leech G, Conrad S and Finegan E (1999). Longman Grammar of Spoken and Written English. Longman: Harlow. Bondi M (2007) Authority and expert voices in the discourse of history. In Fløttum K (ed) Language and Discipline Perspectives on Academic Discourse. Newcastle: Cambridge Scholars, 66–88. Bondi M (2008) Emphatics in academic discourse: Integrating corpus and discourse tools in the study of cross-disciplinary variation. In Ädel A and Reppen R (eds) Exploring Discourse through Corpora. Amsterdam: Benjamins, 31–55. Bondi M (2009) Poliphony in academic discourse: A cross-cultural perspective. In Suomela-Salmi E and Dervin F (eds) Cross-cultural and Cross-linguistic Perspectives on Academic Discourse. Amsterdam: Benjamins, 83–108. Bondi M (2010) Metadiscursive practices in introductions: Phraseology and semantic sequences across genres. Nordic Journal of English Studies 9(2): 99–123. Bondi M (2012) Authorial voice in textbooks: Between exposition and argument. In Hyland K and Sancho Guinda C (eds), Stance and Voice in Written Academic Genres. London: Palgrave Macmillan, 101–115. Bondi M (2014a) Integrating corpus and genre approaches: Phraseology and voice across EAP genres. In Gotti M and Giannoni DS (eds) Corpus Analysis for Descriptive and Pedagogic Purposes: English Specialised Discourse. Bern: Peter Lang, 43–62. Bondi M (2014b) Changing voices: Authorial voice in abstracts. In Bondi M and Lorés Sanz R (eds) Abstracts in Academic Discourse. Bern: Peter Lang, 243–269. Bondi M and Lorés Sans R (eds) (2014) Abstracts in Academic Discourse: Variation and Change. Bern: Peter Lang. Bondi M and Seidenari C (2012) and now i’m finally of the mind to say i hope the whole ship goes down …: Markers of subjectivity and evaluative phraseology in blogs. In Mukherjee J and Huber M (eds) Corpus Linguistics and Variation in English. Amsterdam: Rodopi, 17–27. Brazil D (1997) The Communicative Role of Intonation in English. Cambridge: Cambridge University Press. Campoy MC and Luzón MJ (eds) (2007) Spoken Corpora in Applied Linguistics. Bern: Peter Lang. Cheng W and Warren M (1999) Facilitating a description of intercultural conversations: The Hong Kong corpus of conversational English. ICAME Journal 23: 5–20. Cheng W and Warren M (2000) The Hong Kong Corpus of Spoken English: Language learning through language description. In Burnard L and McEnery T (eds) Rethinking Language Pedagogy from a Corpus Perspective. Frankfurt: Peter Lang, 133–144. Cheng W, Greaves C and Warren M (2005) The creation of a prosodically transcribed intercultural corpus: The Hong Kong Corpus of Spoken English (prosodic). ICAME Journal 29: 47–68. Cheng W, Greaves C and Warren M (2006) From n-gram to skipgram to concgram. International Journal of Corpus Linguistics 11(4): 411–433. Cheng W, Greaves C and Warren M (2008) A Corpus-Driven Study of Discourse Intonation: The Hong Kong Corpus of Spoken English (Prosodic). Amsterdam/Philadelphia: John Benjamins. Cortes V and Csomay E (eds) (2015) Corpus-Based Research in Applied Linguistics: Studies in Honor of Doug Biber. Amsterdam: Benjamins. Cotterill J (2010) How to use corpus linguistics in forensic linguistics. In O’Keeffe A and McCarthy M (eds) The Routledge Handbook of Corpus Linguistics. London: Routledge, 578–590. 58



Corpus Linguistics

Crawford B (2007) The Language of Business Studies Lectures: A Corpus-Assisted Analysis. Amsterdam: Benjamins. Crawford B (2008) Interaction in academic lectures vs. written text materials: The case of questions. Journal of Pragmatics 40(7): 1216–1231. Cresti E and Moneglia M (eds) (2005) C-ORAL ROM. Integrated Reference Corpora for Spoken Romance Languages. Amsterdam: Benjamins. Csomay E (2002) Variation in academic lectures: Interactivity and level of instruction. In Reppen R, Fitzmaurice S and Biber D (eds) Using Corpora to Explore Linguistic Variation. Amsterdam: Benjamins, 203–224. Deroey K and Taverniers M (2012) Just remember this: Lexico-grammatical relevance markers in lectures. English for Specific Purposes 31(4): 221–233. Diani G (2008) Emphasizers in spoken and written academic discourse: The case of really. International Journal of Corpus Linguistics 13(3): 296–321. Diani G (2013) The construction of academic identity in English and Italian university lectures. In Salvi R and Bowker J (eds) Space, Time and the Construction of Identity: Discursive Indexicality in Cultural, Institutional and Professional Fields. Bern: Peter Lang, 47–64. DuBois J, Chafe W, Meyer C and Thompson S (2001) Santa Barbara Corpus of Spoken American English. Philadelphia, PA: Linguistic Data Consortium. Farr F (2003) Engaged listenership in spoken academic discourse: The case of student-tutor meetings. Journal of English for Academic Purposes 2(1): 67–85. Farr F (2008) Evaluating the use of corpus-based instruction in a language teacher education context: Perspectives from the users. Language Awareness 17(1): 25–43. Farr F, Murphy B and O’Keeffe A (2004) The Limerick Corpus of Irish English: Design, description and application. Teanga 21: 5–29. Firth JR (1957 [1935]) A synopsis of linguistic theory, 1930–1955. In Firth JR et al. Studies in Linguistic Analysis. Special volume of the Philological Society. Oxford: Blackwell, 1–32. Fischer K (2006) Approaches to Discourse Particles. Oxford: Elsevier. Fløttum K, Dahl T and Kinn T (2006) Academic Voices. Amsterdam: John Benjamins. Fortanet-Gómez I (2004a) The use of we in university lectures: Reference and function. English for Specific Purposes 23(1): 45–66. Fortanet-Gómez I (2004b) I think: Opinion, uncertainty or politeness in academic spoken English? RAEL: Revista Electronica De Linguica Aplicada 3: 63–84. Fung L and Carter R (2007) Discourse markers and spoken English: Native and learner usage in pedagogic settings. Applied Linguistics 28(3): 410–439. Gilquin G (2008) Hesitation markers among EFL learners: Pragmatic deficiency or difference? In Romero-Trillo J (ed) Corpus and Pragmatics: A Mutualistic Entente. Berlin: De Guyter, 119–149. Gil-Salmon L and Soler-Monreal C (2014) Dialogicity in Written Specialised Genres. Amsterdam: Benjamins. Granger S, Hung J and Petch-Tyson S (eds) (2002) Computer Learner Corpora, Second Language Acquisition and Foreign Language Teaching. Amsterdam: Benjamins. Greaves C (1996) ConcApp. Hong Kong: Edict Virtual Language Centre. Greaves C (2009) ConcGram 1.0: A Phraseological Search Engine. Amsterdam: John Benjamins. Handford M (2010) The Language of Business Meetings. Cambridge: Cambridge University Press. Haslerud V and Stenström A-B (1995) The Bergen Corpus of London Teenage Corpus (COLT). In Leech G, Meyrs G and Thomas J (eds) Spoken English on Computer. London: Longman, 235–242. Herring S, Stein D and Virtanen T (eds) (2013) Pragmatics of Computer-Mediated Communication. Berlin: De Gruyter. Hunston S (2002) Corpora in Applied Linguistics. Cambridge: Cambridge University Press. Hunston S (2008) Starting with the small words: Patterns, lexis and semantic sequences. International Journal of Corpus Linguistics 13(3): 271–295. Hyland K (2005a) Stance and engagement: A model of interaction in academic discourse. Discourse Studies 7: 173–192. 59

Marina Bondi

Hyland K (2005b) Metadiscourse. London: Continuum. Hyland K (2012) Disciplinary Identities: Individuality and Community in Academic Discourse. Cambridge: Cambridge University Press. Hyland K and Bonoli M (eds) (2006) Academic Discourse across Disciplines. Bern: Peter Lang. Hyland K and Diani G (eds) (2009) Academic Evaluation: Review Genres in University Settings. Basingstoke: Palgrave MacMillan. Hyland K and Sancho Guinda C (eds) (2012) Stance and Voice in Academic Writing. London: Palgrave-MacMillan. Johansson S (2007) Seeing through Multilingual Corpora: On the Use of Corpora in Contrastive Studies. Amsterdam: John Benjamins. Jucker A and Ziv Y (eds) 1998 Discourse Markers: Description and Theory. Amsterdam: Benjamins. Jucker A, Schreier D and Hundt M (eds) (2009) Corpora: Pragmatics and Discourse. Amsterdam: Rodopi. Koester A (2006) Investigating Workplace Discourse. London: Routledge. Lenker U and Meurman-Solin A (2007) Connectives in the History of English. Amsterdam: Benjamins. Livnat Z (2012) Dialogue, Science and Academic Writing. Amsterdam: Benjamins. Lorés Sanz R (2006) The referential function of metadiscourse: Thing(s) and idea(s) in academic lectures. In Hornero AM, Luzón MJ and Murillo S (eds) Corpus Linguistics: Applications for the Study of English. Bern: Peter Lang, 315–334. Martin J and White P (2005) The Language of Evaluation: Appraisal in English. London and New York: Palgrave Macmillan. Mauranen A (2003) The corpus of English as lingua franca in academic settings. TESOL Quarterly 37: 513–527. Mauranen A (2013) Hybridism, edutainment, and doubt: Science blogging finding its feet. Nordic Journal of English Studies 13(1): 7–36. McCarthy M (1998) Spoken Language and Applied Linguistics. Cambridge: Cambridge University Press. Müller S (2005) Discourse Markers in Native and Non-Native English Discourse. Amsterdam: Benjamins. Nesi H and Basturkmen H (2006) Lexical bundles and discourse signalling in academic lectures. International Journal of Corpus Linguistics 11(3): 283–304. Nesi H and Gardner S (2012) Genres across the Disciplines: Student Writing in Higher Education. Cambridge: Cambridge University Press. Okamura A (2013) Common features and variations in the use of personal pronouns in two types of monologic academic speech. In Bamford J, Cavalieri S and Diani G (eds) Variation and Change in Spoken and Written Discourse: Perspectives from Corpus Linguistics. Amsterdam: Benjamins, 33–44. O’Keeffe A (2006) Investigating Media Discourse. London: Routledge. O’Keeffe A and Farr F (2003) Using language corpora in language teacher education: Pedagogic, linguistic and cultural insights. TESOL Quarterly 37(3): 389–418. Partington A and Morley J (eds) (2004) Corpora and Discourse. Bern: Peter Lang. Pons Bordería J (2008) Do discourse markers exist? On the treatment of discourse markers in Relevance Theory. Journal of Pragmatics 40: 1411–1434. Raso T and Mello H (eds) (2014) Spoken Corpora and Linguistic Studies. Amsterdam: Benjamins. Rayson P (2009) Wmatrix: A web-based corpus processing environment. Computing Department Lancaster University. http://ucrel.lancs.ac.uk/wmatrix/. Rehbein J, Hohenstein C and Pietsch L (eds) (2007) Connectivity in Grammar and Discourse. Amsterdam: Benjamins. Roach P and Arnfield S (1995) Linking prosodic transcription to the time dimension. In Leech G, Myers G and Thomas J (eds) Spoken English on Computer. Harlow: Longman, 149–160. Romero-Trillo J (ed) (2008) Corpus and Pragmatics: A Mutualistic Entente. Berlin: De Gruyter. Rowley-Jolivet E and Carter-Thomas S (2005) The rhetoric of conference presentation introductions: 60



Corpus Linguistics

Context, argument and interaction. International Journal of Applied Linguistics 15(1): 45–70. Ruhi S, Haugh M, Schmidt T and Wörner K (eds) (2014) Best Practices in Spoken Corpora in Linguistic Research. Newcastle: Cambridge Scholars. Rühlemann C and Aijmer K (2015) Introduction. Corpus pragmatics: Laying the foundations. In Aijmer K and Rühlemann C (eds) Corpus Pragmatics: A Handbook. Cambridge: Cambridge University Press, 1–26. Schiffrin D (1987) Discourse Markers. Cambridge: Cambridge University Press. Schmidt T and Wörner K (2009) EXMARaLDA—creating, analysing and sharing spoken language corpora for pragmatic research. Pragmatics 19(4): 565–582. Scott M (2017) Wordsmith Tools. Version 7. Liverpool: Lexical Analysis Software. Seidlhofer B (2010) Giving VOICE to English as a lingua franca. In Facchinetti R, Crystal D and Seidlhofer B (eds) From International to Local English—and Back Again. Frankfurt: Peter Lang, 147–163. Simpson R, Briggs S, Ovens J and Swales J (2002) The Michigan Corpus of Academic Spoken English. The Regents of the University of Michigan. Sinclair J (1991) Corpus Concordance Collocation. Oxford: Oxford University Press. Sinclair J (2004) Trust the Text: Language, Corpus and Discourse. London: Routledge. Sinclair J and Mauranen A (2006) Linear Unit Grammar: Integrating Speech and Writing. Amsterdam: Benjamins. Soffer O (2010) Silent Orality: Toward a conceptualization of the digital oral features in CMC and SMS texts. Communication Theory 20(4): 387–404. Stenström A-B (1990) Lexical items peculiar to spoken discourse. In Svartvik J (ed) The London-Lund Corpus of Spoken English: Description and Research. Lund: Lund University Press, 137–175. Svartvik J (ed) (1990) The London-Lund Corpus of Spoken English: Description and Research. Lund: Lund University Press. Taavitsainen I, Jucker A and Tuominen J (eds) (2014) Diachronic Corpus Pragmatics. Amsterdam: Benjamins. Tardy C (2012) Current conceptions of voice. In Hyland K and Sancho Guinda C (eds) Stance and Voice in Written Academic Genres. London: Palgrave, 34–48. Thompson S (2003) Text-structuring metadiscourse, intonation and the signalling of organisation in academic lectures. Journal of English for Academic Purposes 2(1): 5–20. Thompson P (2006) A corpus perspective on the lexis of lectures, with a focus on economics lectures. In Hyland K and Bondi M (eds) Academic Discourse across Disciplines. Bern: Peter Lang, 253–270. Thompson P and Nesi H (2001) The British Academic Spoken English (BASE) Corpus Project. Language Teaching Research 5(3): 263–264. Tognini Bonelli E (ed) (1993) Text and Technology: In Honour of John Sinclair. Amsterdam: Benjamins. Yus F (2011) Cyberpragmatics: Internet-Mediated Communication in Context. Amsterdam: Benjamins. Walsh S (2006) Investigating Classroom Discourse. London: Routledge. Walsh S, Norton T and O’Keeffe A (2011) Analysing university spoken interaction: A CL/CA approach. International Journal of Corpus Linguistics 16(3): 325–344. Weigand E (2009) Language as Dialogue (Sebastian Feller ed). Amsterdam: Benjamins. Weigand E (2010) Dialogue: The Mixed Game. Amsterdam: Benjamins. Weisser M (2015) Pragmatic annotation. In Aijmer K and Rühlemann C (eds) Corpus Pragmatics: A Handbook. Cambridge: Cambridge University Press, 84–116. Wichmann A (2007) Corpora and spoken discourse. In Facchinetti R (ed) Corpus Linguistics 25 Years On. Amsterdam: Rodopi, 73–86.

61

4 Discourse Analysis James Paul Gee

4.1 Background There is some ambiguity about what the term ‘discourse’ in ‘discourse analysis’ means. For some, it means the study of language in use and for others it means the study of the level of language above the sentence (above syntax). In both these cases, discourse analysis is partially tied to syntax as a theory about how language structure (form) is put to use in communication (content). For still others discourse analysis means largely thematic (contentbased) studies of language, texts, or communication as they work in society and history, with little tie to syntax. There are forms of discourse analysis in many different disciplines, but in linguistics proper the term ‘discourse analysis’ overlaps with pragmatics (the study of how speech acts, implicature, and inferencing work in communication). Historically, in linguistics, discourse analysis has one of its origins in the close study of language styles and interpretation in sacred texts like the Bible and overlapped with a field called hermeneutics. Perhaps, because of this, much early work in discourse analysis focused on the interpretation of written texts. However, with the advent of structural linguistics and later generative linguistics in the United States, movements which focused on oral language and language structure, discourse analysis became more and more the study of how the grammar of oral language was put to use for communication. Functional theories of grammar—which flourished more outside than inside the United States—argued for a close relationship among syntax, meaning making, and interpretation across talk and texts because such theories saw syntactic structures as historically shaped by, and well designed for, meaning making and communication. Generative theories of grammar argued for a good deal of autonomy between syntax (form/structure) and meaning (content) and argued that there was not a close fit between them, but rather a complex relationship. Even restricting ourselves just to linguistics, there are still many different approaches to discourse analysis (for surveys, see: Gee and Handford, 2013; Jaworski and Coupland, 2014; Tannen et al., 2015; van Dijk, 1997a, 1997b, 2011). Some approaches to discourse analysis are called ‘critical’ as in ‘critical discourse analysis’ (Fairclough, 2003, 2010;

62



Discourse Analysis

Rogers, 2011; van Dijk, 2008; Wodak and Meyer, 2009). Critical approaches are interested in how talk and texts function in terms of politics, equity, and power in society. Non-critical approaches are interested in detailed accounts of grammar-in-use in a wider variety of settings (and sometimes eschew power and politics as themes, for fear of becoming advocates rather than ‘dispassionate’ scholars). I myself (Gee, 2014a, 2014b) have argued that all instances of language-in-use signal ‘political’ meanings (as well as many other types of meanings)—the term ‘politics’ here is used in the sense of the distribution of public goods—and, thus, that all discourse analysis does or ought to have a critical element. This chapter is not an attempt to summarize or integrate different approaches to discourse analysis. Rather, the chapter will seek to place discourse and discourse analysis in a wider context than language (talk and texts) alone (van Dijk, 1985); it will seek to include the world, media, and multimodality (a growing, but still new, interest for many in discourse analysis) inside the field of discourse analysis (Gee, 2015a; Jones et al., 2015; Kress and van Leeuwen, 2001; Mondada, Chapter 2 this volume). There has, of course, long been the realization that language in use, whether in writing or speech, often communicates in tandem with gestures, images, and other symbolic forms (as in mathematics). But as digital and social media, as well as multimodality, have proliferated in our world, the issue of how discourse analysis of language relates to analyses of meaning making using digital media and multiple modes (with or without language) has become pressing. In an attempt to relate talk, text, world, media, and multimodality I will focus on an extended notion of ‘conversation’. While several approaches to discourse analysis—most noticeably work in conversational analysis (CA) (Sidnell and Stivers, 2014)—have argued that conversation is the primordial form of human communication, my extended notion of conversation will add some novel claims to the mix, though primarily by arguing that there are wider bodies of literature relevant to discourse analysis than have been traditionally thought to be relevant.

4.2  Towards a ‘Unified Theory’ of Discourse Analysis It is clear that there are many ways of analyzing themes and images in language use, in media, and in multimodal texts. However, here we will focus on the following question: Is there a relationship between discourse analysis of a linguistic sort (that is, as traditionally applied to the study of grammar-in-use) and an extended notion of discourse analysis that would apply beyond language to the world, media, and multimodality (Gee, 2015a)? Grammar is a system (in the sense of rules or conventions). By a system I simply mean that grammar is made up of parts and wholes composed of those parts and that the wholes inherit many of their properties from the properties of the parts and how they are combined (though wholes can have, in some systems, emergent properties as well). So, for example, in English, ‘red’ can combine with ‘flower’ to make up ‘red flower’, a whole that means something that is both red and a flower (and a whole that can combine as a bigger part to make bigger wholes, as in ‘the red flower wilted’). Or, ‘big’ can combine with ‘mouse’ to make up ‘big mouse’, a whole that means, not something that is both big and a mouse, but something that is big  for a mouse (big relative to other mice though not relative to other sorts of animals like dogs and cats). Grammars and many other sorts of systems offer us categories, categories like nouns and verbs, subjects and objects, or count nouns and mass nouns. Categories can get as specific 63

James Paul Gee

as nouns naming forms of government such as democracy, oligarchy, plutocracy, and so forth, or nouns naming economic systems such as capitalism, socialism, communism, and so forth. For reasons well known since Saussure, the members of categories are defined in part by the other terms in the category and, thus, adding or subtracting terms changes the meanings of all the terms in the category. A set of terms in a category—like forms of government or pet animals—defines a sort of grid in terms of which we see and ‘cut up’ the world and, thus, adding or subtracting terms in the category changes the grid and, therefore too, how we see and talk about the world. Discourse analysis, as I will use the term here, is the study of how language works when a system (in the case of language the system is grammar) is put to use in actual situations of use. Note that the relationship of system (grammar) to situation creates a dynamic interaction between social groups (tradition, culture, society) and individuality. In a sport like basketball, dribbling the ball is defined as a legal move that must meet certain parameters in how it is done. This is a convention (‘rule’). How a player actually fully implements dribbling and integrates it into the other elements of the game is a product of individual style (influenced by social trends, of course). In fact, in the history of basketball, how people implement the system can change the system when people agree to change the system to accommodate innovations or problems stemming from performance. Why even make a distinction between system (conventions) and situations (performance)? Because system sets up the possibility for individuality and performance to become shared enough to work for common understandings and joint endeavors, while still allowing individuality and performance as sources of change. Of course, common understanding is always potentially problematic and potentially ‘endangered’ by innovations—in need of monitoring and work—but grammar is one anchor in the process, just as the rules of basketball are in basketball. The same could be said for the morphology (structure) of the human body and the actual ways each of us, thanks to genetics and life-style choices, implement it in practice. By the way, for Chomsky, grammar is more akin to the morphology of the body—set by natural history—than to a set of social conventions set by social agreements, thanks to the fact that Chomsky sees grammar as formed from our biology as humans and not from sets of social conventions.

4.3  Extending Linguistic Discourse Analysis We will look at human interactions with the world, media and multimodality as working within different sorts of systems than grammars for natural languages, but systems (types of ‘grammars’) nonetheless. Within this perspective on discourse analysis, discourse analysis involves two main tasks to which we now turn (Gee 2014a, 2014b).

4.3.1  First Task: The Situating Meaning Task First, there is the situating meaning task (Barsalou, 1999a, 1999b; Bergen, 2012; Clark, 1989, 1997; Gee, 2004, 2014a, 2014b, 2015; Glenberg and Gallese, 2012). While words and phrases have system (general, type) meaning (e.g. coffee, tea, soda, etc. are beverages), any word or phrase in actual use must be given a specific, situated, more nuanced (token) meaning in the context in which it is actually being used. System meanings are the types of things that are in dictionaries (or in mental lexicons). So in ‘The coffee spilled, go get a mop’, ‘coffee’ means a fluid; in ‘The coffee spilled, go get a broom’, it means grains or beans (depending on the situation); in ‘The coffee spilled, stack it again’, it means cans; and 64



Discourse Analysis

in ‘The coffee spilled, scoop it up’, it might mean ice-cream. In specific situations, ‘coffee’ takes on what I will call situational meanings (in contrast to general system meanings). Speakers are both in situations and help to create and construe them in certain ways. This dynamic and reflective process gives rise to the creativity of situational meanings. The process of assigning situational meanings (i.e., aligning meaning and situation) can be routine (in the sense that we have done it a lot before and rather routinized it) or it can be more novel and proactive—done on the spot for the first time. For example, if I say ‘Big Coffee is as bad as Big Oil’, you can use your knowledge of how ‘Big Oil’ has been used in talk, media, and texts in the past to ‘make up’ a novel sort of situational meaning for ‘Big Coffee’ (as, for example, institutionalized corporate self-interest of coffee companies gone amok). Another example: What for you is the situational meaning of the phrase ‘things you would save first in a fire’? To fill this phrase with situated meaning, very likely now for the first time, you have to draw on past experiences, imagined situations, and your present situation in life to make the phrase mean something, say, like ‘children, pets, heirlooms, important documents …’ (or whatever your list would be). We humans situate meaning by first, using our past experiences as a guide (Bergen, 2012); second, considering how the speaker (or writer) is using language (and things like gesture) and what this signifies for what the speaker or writer might mean and intend (Gee, 2004; Gumperz, 1982; Levinson, 2000); and third, considering what count as relevant aspects of the context of situation in which the communication is going on to guide our guesses about what the speaker means and intends (Sperber and Wilson, 1986). The experiences we have had in life (and stored as long-term memories), how we use them to make judgments about what words and phrases mean in situ, including how we interpret other people’s language and what count as relevant aspects of context are all deeply affected by the social and cultural groups to which we belong and by their values and normative ways of interacting and operating in the world (Gee, 2004, 2014a, 2014b; Holland et al., 1998). Thus, how we situate meaning is not a private mental affair but also a social and cultural one. Thus, in a given situation someone might interpret ‘Relationships take work’ to be a bad thing because ‘work’ is situated to mean arduous effort for extrinsic rewards like money, while another person might interpret it to be a good thing because ‘work’ is situated to mean effort that gives meaning and purpose to life.

4.3.2  The Second Task: The Integration Task The situating meaning task is ‘spatial’ in the sense that language is tied to situations in memory or in the world so that words and phrases relate to each other not just as terms in a system but as links to experience and the world. The second major task of discourse analysis is temporal. Language as a system becomes temporal when it is put to use. It flows through time one element after another. Hearers and readers need to link various elements together and assign them meanings in terms of their connections. Hearers and readers have to interpret elements as belonging more or less closely together or not at all belonging together. I will call this the ‘integration task’ (Gee, 2014a, 2014b; Harris, 1952; Longacre, 1996; van Dijk and Kintch, 1983). In the integration task the border between syntax and discourse becomes porous. Consider the example below (Gee, 2014a): LAST YEAR / Maria Gomez / 65

James Paul Gee

who is our curriculum coordinator here / had a call from Susie // at Clarkson // And called me / and said (pause) / “We have a person from Clarkson [College]/ who’s in the History Department / she’s interested in doing some RESEARCH / into black / history // in Worthell // And she would like to get involved with the school / And here’s her number / Give her a call” // In this transcript capitals mean extra emphasis, a single slash means an intonation contour that is heard as non-final and a double slash is an intonation contour that is heard as final. All the material on a single line is a tone unit (or what is something called an information unit). I have put a space between material that constitutes something like a sentence in writing, but there is no good match here between intonation and sentence structure—for example, ‘at Clarkson’ and ‘in Worthell’ are after thoughts, that is, material that could have been part of the preceding material syntactically to make up a sentence, but has been separated out in terms of speech. Hearers have to give meaning not just to words and phrases but to smaller and larger units or elements as they flow through time and as hearers build up an interpretation of larger and larger stretches of connected language (this interpretation is sometimes called a ‘discourse model’, a type of mental model). Order matters. So consider the different meanings you might assign (based on past experiences and shared knowledge) to these different ways of saying things (and there are many, many more options): LAST YEAR / Maria Gomez / who is our curriculum coordinator here / had a call from Susie // at Clarkson // LAST YEAR/ Susie a professor from Clarkson/ called Maria Gomez / who is our curriculum coordinator here // LAST YEAR our curriculum coordinator here / 66



Discourse Analysis

Maria Gomez / had a call from Clarkson // from Susie / an historian there //

4.4 Design I am going to take a design stance on language and say that speakers and writers actively design their language-in-use to help the listener or reader engage in the situating meaning task and the integration task. Listeners and readers, in turn, consider these design features and use them as guides (clues, cues) in context, together with their past experiences, to give meaning to talk and texts. This is essentially the stance the New London Group took to grammar and meaning making (New London Group, 1996). However, it is crucial that in conversations people take turns at being the designer. They take turns alternating as producer and consumer of designed language. Turn taking is important because it allows speakers when they become listeners to check on the meanings they have activated by their previous turn at talk. They can check whether they realized the purposes that informed their previous turn at talk or not. Since people give situational meanings based on their own previous (socially and culturally informed and mediated) experiences, and these experiences can be quite different between different people and different groups, something must ensure that people and groups stay on track in meaning activation well enough for communication to work (and, of course, it does not, by any means, always work, Gee, 1992, 2015b, 2015c). We have to check up on each other and turn taking allows this. So I am claiming that humans take a ‘design stance’ towards communication. Listeners or readers assume that the speaker or writer has designed his or her talk or text to achieve certain goals or purposes. This implies that the speaker or writer not only takes their turn but acts. The basis of communication is actions that are achieved through grammar put to use (and informing is one such action) (Weigand, 2010). Listeners or readers actively look for the design features of talk or text to make judgments (informed guesses) about what the speaker or writer intends (Gumperz, 1982). Since listeners have no access to the internal workings of other people heads, they need to check on whether their judgments about the speaker’s or writer’s intentions seem to be on track when it is their turn to talk. They use their response to the speaker (when it is their turn at talk), in part, to probe (or even try to change) the previous speaker’s intentions as a foundation on which they can continue mutual meaning making. This means that speakers continually try to adjust and adapt themselves to each other in the ongoing dialogue, as in a dance (Weigand, 2010). (My approach here is related to the massive amount of work on updating situation models, discourse models, or mental models, see, for example: Givon, 1992; Kelleher et al., 2005; Kintsch and van Dijk, 1978; Zwaan and Radvansk, 1998.) All this raises a deep problem about writing—the reader cannot probe the writer’s intentions by designing a probe that will illicit an illuminating and useful response from the author about whether interpretation is on track or not. This is a problem that was seen clearly already by Plato, one of the first great writers in Western history (Gee, 2015a, 2015c). So for the moment we have to leave writing aside as a problem (and, remember, writing is a late and uncommon invention in human history) and concentrate on conversation, turn taking, and interaction (remember that ‘turn taking’ means not just taking turns at talk but taking turns at acting, since language in use is always not just speaking but acting). Meaning 67

James Paul Gee

is mutually made—that is, it is inherently turn-based or interactive—and humans are in danger of going far off course if they do not take turns (and even then if they do not share enough culture, knowledge, and experience things can go badly awry).

4.5  Conversation (Extended Sense) When we humans interact with books, paintings, and movies, they do not talk back to us if we interrogate them (Sell, Chapter 8 this volume). But when we interact with the world, with others in face-to-face communication, or with video games, these things talk back to us, they do respond to our actions. In a sense we can have a conversation with the world, with interactive media like video games, and with other people. While approaches to discourse analysis like conversational analysis (Sidnell and Stivers, 2014) have argued that conversation is the primordial form of human communication, I want to advance the same claim here but with an extended sense of the term ‘conversation’. Note that I am giving an extended sense to the word ‘conversation’—redefining it—so that, for me, saying that the world ‘talks back’ (i.e. responds meaningfully) is not just a metaphor. Work on embodied cognition (Barsalou, 1999a, 1999b; Bergen, 2012; Gee 2004; Glenberg and Gallese, 2012; Mondada, Chapter 2 this volume), as well as work on reflective practitioners (Schon, 1983), has argued that we humans have evolved to have reflective conversations with the world, a process that we can call ‘the cycle of reflective action’ or, to make a point here, ‘the probe cycle’ (Weigand, 2010 sees what I am calling the ‘probe cycle’ in terms of the process of adaptation and, indeed, it is how we adapt to each other and to the world if we want to communicate and survive). In the probe cycle, we 1. Form a goal and need to take an action to accomplish it. 2. Then we use our minds to imagine or simulate what might happen if we act in certain ways (these simulations are based on previous experiences we have had and using elements of these experiences to imagine even novel things). 3. Then we choose an action and act on the world in an attempt to move closer to our goal. This action is a probe of the world, a probe meant to test whether the world responds in the way our simulation made us hope and expect. 4. Then we pay attention to what the world’s response is to our action probe (this is the world’s turn ‘at talk’). 5. Then we reflect on whether the world’s response (the result we got) is good, bad, or indifferent for accomplishing our goal. This is called the ‘appreciative capacity’ because we must value or appreciate (evaluate) the result of our action is some way in order to know what to do next. 6. Then, if we have not yet accomplished our goal (or given up), we either rethink our goal or act again based on the world’s response and our appreciative judgment about it.

This conversation with the world—where we pay attention to the world’s response to our actions—is primordial for humans (Tomasello, 2014) because, on Darwinian terms, people who could not do it are gone. Not paying attention to the world’s responses to our actions in consequential cases is dangerous (something our present-day politics seem not to understand) and bad for survival. The key to the probe cycle is the appreciative capacity (Gee, 2007; Schön, 1983). If we do not know how to assess the world’s response to our actions we do not know what to do next to accomplish our goals. But where do we learn how to assess the world’s actions? We 68



Discourse Analysis

learn this—we develop an appreciative capacity in different domains—through our socialization into families, social groups, cultures, institutions, and interest-driven groups of all sorts. Our appreciative capacity is formed by internalizing social and cultural interactions in a Vygotskian way. It is part of what makes the mind social. Note that the probe cycle—a primordial form of reflective, purpose-driven conversation—is core to interactive digital media like video games as well (Gee, 2007). A video game is just a (virtual) world designed to facilitate solving problems of a certain sort. Players form goals, plan in their minds, act, appreciate the game world’s response (response being not just words or symbols but actions and images as well), and then, with due consideration, change their goals, give up, or act again and repeat the cycle. There are two differences between conversations with the world and conversations with game worlds. One is that the real world is ‘real’ and the game world is ‘virtual’. This is not as severe a contrast as one might think, since when humans are given new tools (like telescopes or electron microscopes) they see ‘new worlds’ or see the ‘real world’ in new ways or even see altogether new things. To a scientist armed with an electron microscope, what we humans see without it is not really ‘real’ but sort of virtual (Gee, 2015c). The second difference is that the game world is well designed by game designers to facilitate the probe cycle, that is, to help players have good conversations (which include practical and effective actions) that eventually, with persistence, lead to success. But this, too, is not as big a difference as one might think, since parents, teachers, mentors, and social groups design (constrain, guide, facilitate) experiences (constrain and define conditions for successful action) in the world for ‘newbies’ (so they do not get overwhelmed or hurt) (Gee, 2015c).

4.6  Face-to-Face Verbal Conversations So what about conversations in the literal sense of turn-taking talk among people? Here, too, we humans apply the probe cycle. Each turn of talk is an action probe based on planning (some of it unconscious and some not) to test whether our intentions (goals) are on track to be accomplished or not. Here, too, when we listen to our interlocutor’s response we have to appreciate it (assess it) in a certain way to know what to say (and do) next. In this sense, we humans are each worlds to be probed for each other in a conversation. When we say that each turn of talk is an action probe this is really what recipient design (Gee, 2014a; Sacks and Schegloff, 1974) is all about. We design our turn at talk to fit with who we think our respondent is or who we want our respondent to be and to get from the respondent the response we want. Just so, we shape our actions in the world and in games in terms of what we think the world or game is like (or what we hope we can make it act like) and in terms of getting the sort of response we want. In turn, we pay attention to the response we get to our probe (recipient design), appreciate (assess) it, and, then, probe (speak) again based on that assessment. In the process we dynamically adjust and adapt to a changing world or person in the service of individual goals and actions, but often mutual and collaborative ones as well.

4.6.1  An Example from Language I will give a brief example from language here and then move to a slightly more extended example from video games. Consider the turns at talk below (Gee 2014a): 69

James Paul Gee

Janet: 9. 10. 11. 12. 13. 14.

And also, once you finally get into the war situation / Because you have Fort Derby / Everybody would get a ride in to come to Bright City / To the amusement park // So it was this influx of two, three cars worth of guys / That were now available to meet the girls that suddenly were there //

Sara: 15. 16. 17. 18. 19.

Well actually street, street cars had just come in in this / And as I recall um from a student who wrote a paper on this / Bright City and Park was built by the street car company / In order to have it a sort of target destination for people to go to / And to symbiotically make money off of this //

Janet: 20. 21. 22. 23.

Because once you got there / You could take a boat ride / And go up and down a lake / And there were lots of other ways to get the money from people //

Four people are talking informally after a formal meeting of an educational project involving middle-school teachers and college professors working to design oral history curricula for middle-school students in a post-industrial struggling town. Janet is one of three teachers in the group of four and Sara is a university professor. The teachers are all third-generation natives of the town and Sara is a typical mobile professional who was neither born in the town nor dedicated to staying there if advancing her career demands that she move. The teachers have been discussing shared family history centering around an amusement park where their parents’ generation met up for dates. Janet is designing her turn at talk for recipients who are—like her—natives, she is speaking out of a native identity (or Discourse). This leaves Sara at a loss when it is her turn to talk, since she is not a native and does not share their family backgrounds in the town. Sara is forced to design a turn that speaks to Janet and the others differently and seeks a different sort of response while speaking topically. Sara designs her turn to speak out of her identity as an historian referring to a scholarly paper and to a ‘conspiracy’ theory that the links between the park and the people coming there were based on money, not shared communal space. Sara’s use of ‘actually’ surely signals to Janet that her previous turn of talk has not worked smoothly for Sara (who seems to be contesting her claims as information and not speaking to them as communal bonding). This signal to Janet means Janet in her next turn of talk must rethink her goals (of bonding). She can, among other options: ignore Sara’s more academic-identity talk and keep to her insider bonding talk; or she can contest Sara in some (possibly conflictual) way, or she can agree with Sara and switch to a conversation about facts and realities; or she can both accommodate Sara and still speak to (design for) her other (native) recipients. Janet deftly does the latter. She continues with activities that made for dating and bonding in the park (and whose memories, shared by their parents) and links them back to Sara’s 70



Discourse Analysis

turn at talk with her ‘And there were lots of other ways to get the money from people’. Note that Janet’s ‘get’ resonates back with the ‘conspiracy’ feel of Sara’s turn. Janet has avoided conflict, she has incorporated Sara into the talk, though it is hard to say how much sarcasm is added to her turn (it seems to me, on hearing it, that there is a little but not a lot). This sort of interaction does seem video-game-like to me. Janet’s original turn, as a probe, worked for the other natives but not for Sara (the goal was bonding). So when Sara talks, Janet has to re-think and probe again, with due deference to not losing the ground she has already gained with her fellow natives. There is an art to this, the art of design.

4.7  Discourse Analysis and Video Games The question, Can there be a linguistic discourse analysis of video games?, taken literally, asks whether games have a system (a syntax), a semantics (system meanings for parts and wholes), and require people to engage in the situating meaning task and the integration task. If they do, they are in that sense ‘like language’ and open to discourse analysis (Gee, 2015a). Note that the larger combination of a grammar, a semantics (‘literal meaning’) and situational meaning need not—and certainly is not—just the sum of its parts, but a larger complex system (in the technical sense of ‘complexity’ here) with emergent properties. But that is not now relevant to my question here, which is: Do games have grammars in anything like the way in which natural languages have grammars? To demand that a system have a ‘syntax’ is to demand that it have basic units that combine in predictable ways into larger units. It is to demand, as well, that the meanings of the larger units be computable in some fashion from the meanings of the smaller units. To demand that a system gave a ‘semantics’ (basic categorical or ‘type’, not ‘token’, meanings) is to demand that its basic units and their combinations have basic meanings or meaning ranges fixed by conventions. While these conventional meanings can vary across different contexts (and new contexts can extend or change their meaning ranges), there must be a conventional core that sets some limits on contextual variation and shapes how contextual variation in meaning operates. The world has a syntax and semantics for us humans thanks to how human vision works (Marr, 1991). The eye sees the world in vaguely bounded 2D (upside down) images. The eye and brain then process these images in order to construct 3D images with bounded edges and clear shapes. These edges, angles, and bounded surfaces and the way they are combined into spaces and objects (and actions across the flow of time) constitute the syntax of the world for humans. We then assign names and conceptual labels to the spaces and objects and actions, based on context, cultural knowledge, and social conventions. This is the semantics of the world for humans. Scientists have special tools that allow them to see the world in a different way than do ‘everyday people’. With their telescopes and microscopes, they see different units (like atoms, cells, and stars) that combine in different ways (into molecules, organs, and galaxies). For them, the world has a different syntax and semantics. Games are made out of a flow of visual images. So they share the syntax and semantics of the human visual world. But, like scientists, gamers have special tools that allow them to see the game world in a different way. Gamers have controllers and avatars through which they can manipulate the game world to accomplish goals and solve problems. Thus, they see the game world not just in terms of spaces, objects, and actions, but in terms of what these things in the game world are good for in terms of accomplishing their goals for winning the game and solving its problems (Gee, 2007, 2014b). 71

James Paul Gee

Gamers see the game world in terms of what we can call ‘game mechanics’ (Gee, 2007, 2015a). Game mechanics are what you can do with things in a game. So gamers see the game world in terms of verbs (actions): crates are good for breaking, ledges are good for jumping, shadows are good for hiding, and so forth. Additionally, things in game worlds can combine in various ways to enable certain actions. For example, a ledge, gap, rope, and wall can in some games combine to enable a deft set of moves to get across the game world (as in Tomb Raider games, for example, games in which the famous character Lara Croft dexterously explores caves, ruins, and other mysterious sites). Though games are built on the syntax and semantics of human vision, their distinctive syntax is composed of the objects and spaces relevant to action in the game. The semantics of a game is a conceptual labeling of these spaces and things not just in terms of their realworld identity (e.g., a crate) but in terms of what they are functionally good for in the game (e.g., breakable to get a power-up). Given that games have a syntax and semantics, we now can ask: Do the ways video games package things, the ways they sequence things, what things mean in actual contexts (situated meaning), and how situated meanings relate to context and culture work, in any significant way, like they do in dialogue in language? To get at these questions in the small space I have here, I will discuss but one game, Thomas Was Alone (TWA).

Thomas Was Alone TWA is a game that uses very simple 2D images. It is about as minimal as a game can get, but for that very reason exposes the basic structure and function of game worlds quite well. Below is an image from TWA. Note that while this image is static, the placement of the characters in it was determined by the player. The image is a result of action. Further, the very next screen will be created by the player’s actions, as well, based on his or her assessment of the problems to be solved. Each sequence created by the player will reflect, too, the player’s interactions with the story and with the emotions of the characters (even though they are shapes!):

Figure 4.1  A Screenshot from Thomas Was Alone. 72



Discourse Analysis

If you play the game TWA, you immediately see that the small colored rectangles in the game are like words that can be combined together (like a phrase). For example, players can stack the rectangles on top of each other in certain orders (e.g., red on orange on yellow on blue). Order matters in a minimal way. Each shape has a characteristic movement of its own and each can be move independently, but when they are combined, the bottom one determines the movement. The bottom one functions like a predicate in language. In the game, the player must get all the shapes to the end of each level, using their different actions and combinations. So we have a clear, albeit simple, syntax and semantics here. For example, ‘red on orange on yellow on blue’ means ‘stack can move over water’ (because that is the blue rectangle’s basic action). The order of the stack matters in terms of what subsequent actions are possible, since the shapes can jump off only from the top down. So how do situated meanings work in TWA? Do the shapes and their actions take on specific and extended meanings in actual contexts of use (play)? TWA has a story. The story is narrated by a narrator whose narration is heard and whose words are printed on the screens. Within the story each shape has a name and something of a backstory as an artificially intelligent agent inside a computer whose programming has gone awry. The shapes are trying to escape the system. In an interesting twist, each shape has certain unique abilities and limitations (determined by the game’s game mechanics/ semantics) that fit with the character’s personality trait and role in the story. For example, Thomas, the red rectangle, has an up-beat attitude and can do an average jump. John, the yellow rectangle, is arrogant and eager to show off and can jump quite high. Claire, the blue square, who starts off feeling bad about herself but comes to see herself as a super-hero, cannot jump well or move fast, but she can float and move in water and thereby save others by giving them rides across water. The game’s story allows us to assign meanings far beyond ‘short jumping yellow rectangle’. They allow us to assign emotional and narrative meanings to the rectangles as they act alone and together. The story—and our cultural knowledge about escape stories and about computers—give us cultural models or frames within which we can give richer interpretations to what is happening. In TWA the game’s oral narration is also printed on the screen. Since TWA prints the words of the narration on the screen, this, in a way, subtracts the words from the oral narration and means that the oral narration mainly functions to carry the intonation of the narrator’s voice, the musical and affective part of speech. This affect is created in part because we can read much more quickly than we can hear, so the player has often read all the (short) material on the screen before the narrator has finished saying it. The player has the ‘meaning’ but still must pay attention to the intonation contours. The narration in TWA is in a British accent that is amazingly good at indicating the emotions of the characters (rectangles though they be), emotions like fear, self-loathing, loneliness, liking and love, caring, arrogance, humility, and trust. We attribute these emotions and attitudes as deeper meanings for each character by considering the contexts they are in. Consider the image from TWA above. Given the words on the screen, the positions of the characters, and the situation we are in within the game at this point, as well as our earlier play in the game, we can attribute to John (the tall yellow rectangle) a situated meaning or inference like: John wants to help, though not necessarily for altruistic reasons, but because he likes to show off and look good to others. Let’s turn now to the ways meanings and inferences are built up in the context of the order, sequence, and flow of screens in the game. Let’s assume, for the sake of argument, that a screen prior to the image above had John, the yellow triangle, down on the same 73

James Paul Gee

level as Thomas (the red rectangle) and Chris (the orange square). Assume further that the player has then jumped John—a very good jumper—up to where we see him in the image, up above Thomas and Chris. From this vantage point, a player can clearly see that he or she could move John to the right, down the little alley, and then again further to the right and away from Thomas and Chris. But the player can also readily see that moving John to the right will not get Thomas and Chris up the ledge so they can move on in the game as well. They cannot alone or together jump high enough to get up the ledge. John must come back down and allow Chris to jump on top of Thomas and then to jump from Thomas to John and, finally, to jump up to the ledge. Thomas can then jump on John and then up to the ledge. And only then can John jump back up by himself. This creates a sequence in the player’s mind, a sequence that he or she can then create. In the image we see that John is higher than Thomas and Chris and that he can easily go on without them. We see, too, that if the game (and its story) is to continue, he must go back down, place himself again on the same level as Thomas and Chris, help them, and then move on last (not first) himself. All these meanings derived from sequence reinforce the sorts of situated meanings we have drawn from the story and contexts of play. John thinks more highly of himself than he does of the others. Forced to go back and help, he has to rationalize this as not a weakness, but as a strength. This strength is not only that his help is essential to mitigate Thomas’s and Chris’s weaknesses. It is also that John will look good in the act and others will see how special he is. It is clear from even this simple analysis that how game designers and players (through their choices and actions) create situations or contexts as meaningful (for action taking) in certain ways is a crucial way in which games take on meanings beyond their game mechanics (semantics). Note that in the case of language, context is a co-creation of the world and of how humans construe things in certain ways. As in the case of language, gamers have to know what is relevant (and how it is relevant) in the contexts of their play, which often involve games that are much more graphically complex and rich than TWA. Games and players co-create contextual relevance and meaning (including conditions action taking), as do speakers and writers.

4.8  Remaining Issues There is not space enough here to go into the many remaining issues that must be dealt with in a unified theory of discourse analysis of the sort I am gesturing to here, a theory that places language, the world, media, and multimodality under the umbrella of discourse analysis. I have hardly done justice to any of the domains here and most certainly not to all of them. And, indeed, there are other related approaches, such as Weigand’s Mixed Game Model (2010), that look at grammar, conversation, and action in different ways. The problem of literacy remains: What to do when authors design for readers who cannot take a turn and allow the author to check whether things are working and allow themselves to design a response that could change the course of the conversation? Clearly we need theories about how text is turned into conversations (dialogues) in people’s heads and through talk and other texts in the world. In one sense, readers must voice both sides of a conversation for themselves and in the act re-write the text or, otherwise, leave the text inert (Gee 2015a, 2015c). So, too, for paintings and film, the other content-driven media. I have here only had space to concentrate on probes, recipient design, situating meaning, and integrating—with only sideways glances to social and cultural groups (and, thus, to 74



Discourse Analysis

things like frames, schemes, cultural models, figured worlds, and so forth, see: Gee 2004, 2014a, 2014b; Holland et al., 1998). Elsewhere (Gee 2014a, 2014b) I have argued that people use language—often in conversations—to actively and collaboratively design (attempt to make, construct, call into being in the world) seven things. Below I list these seven things and give examples relevant to the conversation between Janet and Sara above: 1. Significance (What is to be taken as important or not; for example, meet ups among friends at Bright City or ways of making money from such meet ups?); 2. Activities (What are we doing or what practice are we participating in here and now; for example, bonding or disputing facts?); 3. Identities (What are we being or what socially significant identities or ‘kinds of people’ are we enacting or recognizing; for example, natives or professional historians?); 4. Relationships (What sorts of relationships exist among people or things; for example, are we here and now colleagues, fellow professionals, natives, or in conflict?); 5. Politics (How are social goods to be or meant to be distributed; for example, who is an insider or an outsider, who has access ‘privileged insight’ into reality or not?); 6. Connections (How are things connected or linked, or not, to each other; for example, what are the connections between an amusement park, meet ups, and profit, and how strong or weak are these relationships?); 7. Sign Systems and Knowledge (Whose language or language styles and whose claims to knowledge count or don’t and how; for example, what counts more: teacher talk or professor talk? What knowledge matters most: claims based on personal oral histories or claims in scholarly papers?).

I would argue that our conversations (adaptive meaningful interactions) with the world, with other people, and with video games (and other media) can be viewed through these seven building tasks. It is a useful exercise to ask how these tasks work out for the language example here (the natives and the professor, which I have gestured to above in the list of building tasks) and the video game example (Thomas Was Alone). Clearly in both examples things are actively made to matter and come to be in certain ways via design and interaction.

References Barsalou LW (1999a) Language comprehension: Archival memory or preparation for situated action. Discourse Processes 28(1): 61–80. Barsalou LW (1999b) Perceptual symbol systems. Behavioral and Brain Sciences 22(1): 577–660. Bergen BK (2012) Louder than Words: The New Science of How the Mind Makes Meaning. New York: Basic Books. Clark A (1989) Microcognition: Philosophy, Cognitive Science, and Parallel Distributed Processing. Cambridge, MA: MIT Press. Clark A (1997) Being There: Putting Brain, Body, and World Together Again. Cambridge, MA: MIT Press. Fairclough N (2003) Analyzing Discourse: Textual Analysis for Social Research. London: Routledge. Fairclough N (2010) Critical Discourse Analysis: The Critical Study of Language. Second Edition. London: Longman. Gee JP (1992) The Social Mind: Language, Ideology, and Social Practice. New York: Bergin & Garvey. Reprinted 2014: Champaign-Urbana: Common Ground. Gee JP (2004) Situated Language and Learning: A Critique of Traditional Schooling. London: Routledge. 75

James Paul Gee

Gee JP (2007) What Video Games Have to Teach Us About Learning and Literacy. Second Edition. New York: Palgrave/Macmillan. Gee JP (2014a) An Introduction to Discourse Analysis: Theory and Method. Fourth Edition. London: Routledge. Gee JP (2014b) How to Do Discourse Analysis: A Toolkit. Second Edition. London: Routledge. Gee, JP (2015a) Unified Discourse Analysis: Language, Reality, Virtual Worlds, and Video Games. London: Routledge. Gee JP (2015b) Social Linguistics and Literacies: Ideology in Discourses. Fifth Edition. London: Routledge. Gee JP (2015c) Literacy and Education. New York: Routledge. Gee JP and Handford M (eds) (2013) The Routledge Handbook of Discourse Analysis. New York: Routledge. Givon T (1992) The grammar of referential coherence as mental processing instructions. Linguistics 30(1): 5–55. Glenberg AM and Gallese V (2012) Action-based language: A theory of language acquisition, comprehension, and production. Cortex 48(7): 905–922. Gumperz JJ (1982) Discourse Strategies. Cambridge: Cambridge University Press. Harris ZS (1952) Discourse analysis. Language 28(1): 1–3. Holland D, Lachicotte W, Skinner D and Cain, C (1998) Identity and Agency in Cultural Worlds. Cambridge, MA: Harvard University Press. Jaworski A and Coupland N (eds) (2014) The Discourse Reader. Third Edition. London: Routledge. Jones RH, Chik A and Hafner CA (eds) (2015) Discourse and Digital Practices: Doing Discourse Analysis in the Digital Age. London: Routledge. Kelleher J, Costello F and van Genabith J (2005) Dynamically structuring, updating and interrelating representations of visual and linguistic discourse context. Artificial Intelligence 167(1–2): 62–102. Kintsch, W and van Dijk TA (1978) Toward a model of text comprehension and production. Psychological Review 85(5): 363–394. Kress G and van Leeuwen T (2001) Multi-Modal Discourse: The Modes and Media of Contemporary Communication. London: Arnold. Levinson, SC (2000) Presumptive Meanings: The Theory of Generalized Conversational Implicature. Cambridge, MA: MIT Press. Longacre RE (1996) The Grammar of Discourse. New York: Plenum Press. Marr D (1991) From the Retina to the Neocortex: Selected Papers by David Marr. Boston, MA: Birkhauser. New London Group (1996) A pedagogy of multiliteracies: Designing social futures. Harvard Educational Review 66(1): 60–92. Rogers R (2011) An Introduction to Critical Discourse Analysis in Education. Second Edition. Mahwah, NJ: Lawrence Erlbaum. Sacks H and Schegloff E (1974) A simplest systematics for the organization of turn-taking for conversation. Language 50(4): 696–735. Schön, DA (1983) The Reflective Practitioner: How Professionals Think in Action. New York: Basic Books. Sidnell J and Stivers T (eds) (2014) The Handbook of Conversational Analysis. Oxford, England: Blackwell. Sperber D and Wilson D (1986) Relevance: Communication and Cognition. Cambridge: Cambridge University Press. Tannen D, Hamilton HE and Schiffrin D (eds) (2015) The Handbook of Discourse Analysis. Second Edition. Oxford, England: Blackwell. Tomasello M (2014) A Natural History of Human Thinking. Cambridge, MA: Harvard University Press. Van Dijk TA (ed) (1997a). Discourse as Social Interaction. London: Sage. Van Dijk TA (1997b) Discourse as Structure and Process. London: Sage. 76



Discourse Analysis

Van Dijk TA (2008) Discourse and Power. New York: Palgrave/Macmillan. Van Dijk TA (ed) (2011) Discourse Studies: A Multidisciplinary Introduction. London: Sage. Van Dijk T and Kintch W (1983) Strategies of Discourse Comprehension. London: Academic Press. Weigand E (2010) Dialogue: The Mixed Game. Amsterdam/Philadelphia: Benjamins. Wodak R and Meyer M (2009) Methods of Critical Discourse Analysis. Third Edition. London: Sage. Zwaan RA and Radvansky GA (1998) Situation models in language comprehension and memory. Psychological Bulletin 123(2): 162–185.

77

5 From Pragmatics to Dialogue Istvan Kecskes

5.1 Introduction This chapter focuses on how the limits of pragmatics—as long as it is restricted to one utterance—are overcome by including the hearer not only as interpreter who tries to understand the speaker’s utterance but as an interlocutor who tries to come to an understanding with the speaker. The goal of the chapter is not to describe and analyze the dialogue approach, rather the goal is to explain what inner developments in the pragmatics paradigm have made it necessary to move in a dialogic direction, specifically emphasizing the importance of evaluating speaker meaning from the perspective of the speaker rather than from the perspective of the hearer and the double role of the interlocutor (speaker–hearer). Pragmatics is an utterance-based inquiry. However, recently there have been many attempts to extend the boundaries of the discipline to include dialogic sequences or discourse segments to make better understanding of utterance meanings, especially speaker meaning. This endeavor has become especially articulate with dialogic studies, intercultural pragmatics studies and computational pragmatics becoming strong and independent inquiries. Research in intercultural pragmatics (e.g. House, 2002; Kecskes, 2007, 2013) demonstrated that in intercultural communication participants are creative on discourse level rather than on utterance level. This is mainly due to limited language proficiency that may result in, among others, not-very-well formulated utterances. Consequently, intercultural interactions may require not only a bottom-up, sequential utterance by utterance analysis but also a top-down, holistic dialogue segment analysis if we want to make sure that we understand the message of interlocutors. Another field of inquiry that includes a dialogue approach is computational pragmatics that focuses on the application of pragmatics to dialogue modelling, especially the development of spoken dialogue systems intended to interact with human beings in task-oriented scenarios such as providing travel information and shows how and discusses why computational pragmatics differs from ‘linguistic’ pragmatics, and how pragmatics contributes to the computational analysis of dialogues (e.g. Morante et al., 2007; Bunt, 2011). Recently several papers have talked about ‘narrow pragmatics’ and ‘wide pragmatics’ discussing the relationship of pragmatics and discourse analysis (e.g. Puig, 2003; Taboada 78



From Pragmatics to Dialogue

and Mann, 2006; De Saussure, 2007). The issue of beyond utterances analysis has also been on the agenda of dialogue studies (e.g. Weigand, 2001, 2010a, 2010b; Carbaugh, 2013; Cooren, 2010). The importance of these studies is that they consider language as action that is always something that is shared. Whenever someone appears to act, others also proceed into action. They also emphasize that communicative function and communicative agenda are two separate things.

5.2  The Dialogue Approach Because ‘utterance’ is hard to define, and its meaning is both in the building elements of an utterance (lexical units) and in the subsequent utterances produced in response, pragmatics, which is also defined as a theory of meaning in context, has been looking for meaning ‘ingredients’ both inside the utterance and outside the utterance. As a result, there has developed an approach that can be called pragma-dialogue, which calls attention to the dialogic nature of communication by emphasizing that interlocutors are actors who act and react (e.g. Weigand, 2010a; Kecskes, 2012). So the speaker–hearer not only interprets but also reacts to the other interlocutor’s utterance. The basic dialogic principle is that human beings are dialogic individuals (social individuals) who communicate in dialogic interaction not only by producing and understanding utterances but also by acting and reacting (e.g. Weigand, 2010a, 2010b). The dialogic principle defines dialogue as a sequence of actions and reactions. A sequence of speech acts can be considered a dialogue which is a highly structured activity involving (at least) two agents. The structure of real dialogues has been extensively studied by ethnomethodologists (Garfinkel, 1972; Psathas, 1979; Schenkein, 1978; Turner, 1974), who have advocated the necessity of non-quantitative, ethnic methods in the analysis of social interactions. Their work on naturally occurring conversation provides a significant amount of data on the way in which different types of dialogue actually evolve. A dialogue is a sequence of utterances, a reciprocal conversation between two or more entities. Buber believed that genuine dialogue can be obtained between two people “no matter whether spoken or silent” (Buber, 1955). Buber further claimed that, “each of the participants really [must have] in mind the other or others in their present and particular being and turn to them with the intention of establishing a living mutual relation between himself and them” (Buber, 1955: 22). More simply, genuine dialogue can be understood as a momentary experience in which participants are consciously aware of the other/s. Preserving the original Buberian idea of dialogue being a reciprocal conversation in which interlocutors are consciously aware of the partners, current researchers of dialogue studies (e.g. Weigand, 2010a, 2010b; Cooren, 2010; Tracy and Craig, 2010) go one step forward and emphasize that dialogue is constituted by the interactive purpose of coming to an understanding, which is based on the sequence of actions and reactions as the following conversation demonstrates. (1) Action-directive Info-request Signal-non-understanding Info-request Agreement

Bill:  Can I get a cup of coffee? Sara: Milk? Bill: Hm? Sara:  Do you want your coffee black? Bill:  Oh yes, thanks.

79

Istvan Kecskes

Weigand (2010b) argued that action and reaction are not two actions of the same type which are arbitrarily connected and only formally distinguished by their position in the sequence. Action and reaction are functionally different types of actions: action is initiative while reaction is responsive. According to Weigand, orthodox speech act theory that is exclusively based on illocution should be adjusted. She continues arguing that Searle’s theory of conversation is based on single illocutionary acts that are put together by what Searle (1995: 26) calls “collective intentionality” ignores the interdependence between action and reaction. Although Searle’s collective action aims at a communal goal it is not a dialogic act yet. Weigand claims that individual actions and reactions are organized into a dialogic sequence by their interdependence. Searle’s speech act theory focuses on the communicative functions of speech acts (assertives, directives, commissives, expressives, declaratives) and says less about the communicative agenda of interlocutors. To interpret an utterance properly, the interlocutor has to arrive at an understanding not only of its communicative function but also of the communicative agenda of his or her dialogue partner, that is, what the partner really wants to achieve in the dialogue. The underlying interests of the dialogue partners become evident through dialogic sequences. This is in line with what the socio-cognitive approach (Kecskes, 2010, 2013) claims: occasionally we should go beyond single utterance to understand the communicative agenda of interlocutors. Intercultural communication provides several examples that support this point. In the following excerpt the Korean speaker is trying to explain the sandy wind they experience in Seoul in the spring. (2) Korean student is talking with an American student 10 NS: In Korea? = 11 NNS: =Yeah::. In (0.2) in ↑spring? 12 NS: Hm mm 13 NNS: There is a (0.5) um (0.7) how- how can I ‘spl- ah:: how can I say::? (0.5) Send wind? 14 (1.7) ((NS displaying mental effort)) 15 NS: Uh:::m 16 NNS: Sen::d (0.2) the wind? 0:30 17 NS: ((now nodding)) Yeah::, it would blow around [the pollen? 18 NNS: [Yes. 19 NNS: Yeah yeah [yeah (0.2) yeah. 20 NS: [(from) the trees? 21 NNS: Ye[ah 22 NS: [Yeah:: 23 NNS: Ah, the:: (0.7) the wind came from China? 24 NS: (0.7) 25 NNS: Do you know that wind? 26 (0.2) 27 NS: ↑No::. 28 NNS: The the:: (0.7) the- many sand? ((“sand” sounds like “send”)) 29 NS: (0.5) ((flashes little smile)) 30 NNS: Sen:d in a: (0.5) desert? 31 NS: Oh, (.) sa:nd. 32 NNS: Yeah, [sand ((shift in pronunciation to match NS)) 33 NS: [Oh:::, [god, yeah 80



From Pragmatics to Dialogue

34 NNS: [in deser::t. We’ve had- (0.2)wi:nd (0.2) go to- uh lan- (0.5) come- (0.2) over 35 to Korea: also. (0.5) So in spri:ng, it’s very difficult to (0.5) little bit difficult to (0.7) □ (?)hhh um:: In this dialogue the Korean speaker has a communicative agenda to explain to her American partner what Koreans call morae baram [sandy wind]. However, her pronunciation is not correct and the American partner does not understand what kind of wind the Korean speaker refers to. Not until the Korean connected ‘sand’ with ‘desert’ in line 30 did the American understand what her Korean partner really meant. The action-reaction sequence shows that they could work out the misunderstanding after all. The segment as a whole makes perfect sense and directs our attention to how hard the Korean speaker has worked on making her partner understand what exactly she wanted to say. Airenti et al. (1993) and Bara (2011) argued that a distinction must be made between a global and a local structure of dialogues. Local structure is about what binds utterances together and what makes one utterance coherent (or incoherent) when following another. Turn taking thoroughly studied by conversation analysts (e.g. Sacks et al., 1974), is part of the local management system. The relationships among speech acts within a single turn also pertain to the local structure. The global structure determines the flow of conversation. It involves, in particular, the scheduling of dialogue phases, for instance, the opening and closing sections. In conversation analysis the case of telephone calls is often studied, because in those exchanges the general structure of the conversation is especially strictly determined (see, e.g., Schegloff, 1979). Dialogues share a global structure with all kinds of interpersonal activities. The global structure of these dialogues derives from mutual knowledge of an action plan, executed in the course of the activity. According to Bara (2011) cognitive pragmatics views the global structure of dialogues as deriving from sharing the knowledge of an action plan. Consequently, the global structure of a dialogue does not derive from linguistic rules, but from behavior games.

5.3  Current Pragmatic Theories Kecskes (2008: 404) argued that in order to give an adequate explanation of communicative processes, we need a dialectical model of pragmatics that combines the perspective of both the speaker and hearer. This change is warranted because current pragmatic theories, both those that have grown out of Grice’s theory, such as the various neo-Gricean approaches and the approach proposed by Relevance Theory, appear to be hearer-centered. They derive from the Gricean modular view that divides the interpretation process into two stages: what is said and what is implicated. Although the Gricean theory, with its cooperative principle and maxims, was supposed to embrace conversation as a whole, basically its further development has remained hearer-centered, with less emphasis on interest in the speaker’s position—a rather paradoxical turn, as Grice himself always emphasized speaker’s meaning. Even so, the Gricean divide of truth-conditional semantics and pragmatics has led to a somewhat impoverished speaker meaning, without much regard for the pragmatic features embedded in the speaker meaning. The division between what is said and what is implicated was made for the sake of utterance interpretation and for the sake of distinguishing the semantic meaning from the pragmatic meaning of an utterance. However, a theory that is concerned about the speaker’s [or speakers’] meaning should focus not only on the truth values of the speaker’s utterance, but also on its pragmatic elements and on the speaker’s commitment and egocentrism (in the 81

Istvan Kecskes

cognitive sense of the term that will be explained later) deriving from prior experience and dominated by salience. Kecskes (2013) argued that the dominance of the hearer’s perspective in current pragmatic theories can be explained by three facts: first, the focus on what is said, and the truth values of the speaker’s utterance without due regard for its pragmatics features in the neo-Gricean approaches; second, emergence of the powerful Relevance Theory; and third, the partial misinterpretation of the Gricean approach. Although the neo-Griceans’ main concern is the speaker’s meaning, they still view communication as designed with a focus on the recipient and the recipient’s recognition of speaker’s intention. In this approach the speaker designs his or her utterance for the hearer and the hearer’s task is to recognize the speaker’s intention. But what is recovered is not always what was intended because of the interlocutors’ differences in their private cognitive contexts and prior experience. So an adequate account of interaction should consider interlocutors not only as common-ground seekers, but as individuals with their own agendas, with their own prior experience, with their specific mechanisms of saliency (based on prior experiences), and their individual language production systems. This issue is especially important in intercultural communication where interlocutors representing different L1s and cultures cannot be certain how much they have in common with each other. They may have to create common ground almost from scratch. Consequently, the balance between the individual and societal factors in intercultural interaction is rather lopsided towards the individual factors. However, individuals in these cases also work under the constraints of societal factors but those happen to be different for each participant depending on their L1 and cultural background. This may lead to too much emphasis on individual factors like personality, experience, egocentrism, and so forth in intercultural interaction. So we may have an intercultural exchange that is affected by societal factors that are different on each side about how much self-promotion you should do in a job interview (these aren’t individual factors), but it may be further affected by individual differences, such as condescension by one party or defensiveness by one party because of those societal differences (and these are entirely individual factors). The second reason for hearer-centeredness is the Relevance Theory. Unlike the neoGriceans, who attempt to give an account of the speaker’s meaning, relevance theorists focus on developing a cognitive psychological model of utterance interpretation, which does not address the question of how and why the speaker, given what she or he wants to communicate, utters what she or he utters. Saul (2002) said that the main difference between the neo-Gricean theory and Relevance Theory lies in ‘whose meaning’ they model. While the neo-Griceans follow the original perspective and consider utterance meaning, including implicature, to be the speaker’s intended meaning, relevance theorists discuss intentional communication from the perspective of the addressee’s reconstruction of the speaker’s assumptions. There is a third reason why current pragmatics is basically hearer-centered rather than interlocutor-centered. Grice’s implicature is often misinterpreted by pragmaticians. The Gricean implicature is an aspect of speaker meaning: what is said and what is implicated by the speaker. Although Horn (2004) and Bach (2001) called attention to the difference between implicating and inferring, not everybody has listened. Horn said: “Speakers implicate, hearers infer” (Horn, 2004: 6). Bach (2001: 18) argued that “People sometimes confuse infer with imply … . When we say that a speaker or sentence implies something, we mean that information is conveyed or suggested without being stated outright. ... Inference, on the other hand, is the activity performed by a reader or interpreter in drawing conclusions that are not explicit in what is said.” 82



From Pragmatics to Dialogue

Grice may (unwillingly) also have contributed to this misunderstanding by bringing the ‘audience’ into the explanation of implicature. According to Grice, a speaker’s intending to convey that P by saying that Q is not enough for the speaker to implicate that P. The audience must also need to believe that the speaker believes that P to preserve the assumption of the speaker’s cooperativeness. With this claim Grice attempted to give some degree of intersubjectivity to the notion of conversational implicature. However, speakers have authority over what they utter-implicate, but they can’t fully control what they conversationally implicate (Saul 2002). The following short conversation (example 3) demonstrates this very well. (3) Roy: - Is there something wrong, Susie? Susie: - I am fine, Roy. Roy: - I would have believed you if you hadn’t said “Roy”. Susie: - OK, OK, just stop …… Susie’s utterance “I am fine, Roy” gives way to different possible interpretations. The expression “I am fine” is a situation-bound utterance, a formulaic phrase. However, Susie added the name of her interlocutor to the formula, which basically broke the formula. It was not a formula any more. Roy had a choice to interpret it as a formula and say nothing. However, he understood the speaker’s intention clearly. She wanted to indicate that her words should not be taken at face value. She was, in fact, not fine. Roy got that, and responded accordingly. So the speaker can direct the hearer to the right understanding of his utterance with something additional to the utterance such as discourse markers, gestures, intonation, sentence formulation, and such like. From the speaker’s perspective what really matters is what the audience is required by the speaker to believe, not what the audience does believe. Davis argued that Grice was wrong to include audience-oriented criteria in his characterization of conversational implicature. The speaker’s intentions do not depend on what anyone else presumes. “To mean or imply something is to have certain intentions” (Davis, 1998: 122). Saul (2002) found it important to distinguish between ‘utterer-implicatures’ and ‘audience-implicatures’. Utterer-implicatures are claims that the speaker attempts to conversationally implicate (intended by the speaker, but not necessarily recognized by the addressee). Audience-implicatures are claims that the audience takes to be conversationally implicated (recognized by the addressee but not necessarily intended by the speaker). In example (3) the utterer-­implicature was recognized by the hearer so the utterer-implicature coincided with the audience-implicature. However, this is not always the case, as example (4) demonstrates: (4) Allan and Sherri (of the sitcom Two and a Half Men) are sitting in a restaurant. Allan’s right eye is covered with a bandage so he does not see Sherri very well. Allan: - You know, Sherri, I would really like to see more of you. Sherri: - Maybe, we should wait and see how the night goes. Allan: - Oh, no. I mean I have only got one good eye. Can we change places? Sherri: - Sure. The conversation demonstrates that Sherri completely misunderstood Allan’s utterance “… I would really like to see more of you.” This may be due to the fact that she relied exclusively 83

Istvan Kecskes

on linguistic salience and ignored perceptual salience in processing the utterance. The utterer-implicature is the literal meaning of the sentence, that is, Allan expressed his desire to see more of Sherri right then. He wanted to change places hoping that from that angle he would see more of Sherri with his one eye. The hearer-implicature is significantly different because Sherri relied on the figurative meaning of “see” [meet someone romantically] and thought that Allan wanted to be more involved with her romantically. The discrepancy between utterer-implicature and hearer-implicature has caused the misunderstanding between the interlocutors. There have been several attempts in pragmatic theorizing to revise/correct the problems of the modular view and recognize pragmatic features of the speaker’s meaning (e.g. Sperber and Wilson, 1986; Carston, 2002; Moeschler, 2004: explicature/implicature; BurtonRoberts, 2006: what-is-A-said/what-is-B-said; Bach 2001: what is said/ impliciture/implicature). However, they have not gone far enough because they still were interested primarily in utterance interpretation, without paying due attention to private or individual knowledge, prior experience, and the emergent, rather than the a priori only intentions of the speaker. The main problem with the hearer-centered views is that they want to recover speaker meaning from a hearer perspective. As a result, what is actually ‘recovered’ is hearer meaning, in the sense of how the hearer interprets what the speaker said. The proposition the speaker produces will not be exactly the same as that which will be recovered by the hearer, because, as we said above, interlocutors are individuals with different cognitive predispositions, different commitments, different prior experiences, and different histories of use of the same words and expressions.

5.4  Why Do We Need the Dialogue Approach? As discussed above current pragmatic approaches struggle with two issues. First, it would be important to explain speaker intention and speaker meaning from the perspective of the speaker. Second, we need to treat the hearer as an interlocutor, that is, a hearer-speaker, who is not only an interpreter but also a producer of utterances. The dialogue approach could help to solve both of these problems with its principle according to which human beings are dialogic individuals (social individuals) who communicate in dialogic interaction not only by producing and understanding utterances but also by acting and reacting. Bach (2007: 5) said that (actual situational) context does not literally determine, in the sense of constituting, what the speaker means. What the speaker really means is a matter of his communicative intention although what he could reasonably mean depends on what information is mutually salient for him/her and the hearer. Bach further argued that taking mutually salient information into account goes beyond semantics, for what a speaker means need not be the same as what the uttered sentence means. This claim can be best demonstrated in intercultural communication where ‘mutually salient information’ is quite problematic. Interlocutors in intercultural interactions have different first languages and use a second (or Lx) language to communicate with each other, sharing very little common ground (Kecskes, 2007, 2013). Salience is based on familiarity, frequency, and common prior experience (Giora, 1997, 2003). Mutually salient information is something lingua franca speakers usually lack because they speak several different L1s and represent different cultures. For them, mutually salient information should be directly connected with the actual situational context, tied to some universal knowledge and/or encoded in the common linguistic code (lingua franca) so that it can be ‘extracted’ by the hearer without any particular inference that relies on non-existing common prior experience. It appears that inferencing 84



From Pragmatics to Dialogue

for the lingua franca interlocutor in intercultural interaction usually means something close to decoding. This can be demonstrated through the following dialogue. (5) The interaction below takes place between the clerk and a Korean student in the Office of Human Resources. Lee: - Could you sign this document for me, please? Clerk: - Come again ... Lee: - Why should I come again? I am here now. The Korean student misunderstands the expression “come again”. He processes it literally in spite of the fact that the actual situational context is completely unsupportive to this interpretation. It is essential therefore that pragmatics for intercultural interlocutors cannot be something “… they communicate over and above the semantic content of the sentence” as King and Stanley (2005: 117) assumed. For lingua franca speakers the semantic content is usually the conveyed content. If this is not clear from their utterance they try to reinforce it with repetition, paraphrasing or other means. This can be seen by the analyst only if she goes beyond utterance and analyzes the dialogic sequence. So for nonnative speakers (especially with lower language proficiency) participating in intercultural interactions, “what is said” often coincides with “what is communicated”. Most of the time what the utterance says is meant literally. This fact gives a strong support to Bach’s claim: it is a mistake to suppose that ‘pragmatic content is what the speaker communicates over and above the semantic content of the sentence’ (King and Stanley 2005: 117). Pragmatics doesn’t just fill the gap between semantic and conveyed content. It operates even when there is no gap. So it is misleading to speak of the border or, the so-called ‘interface’ between semantics and pragmatics. This mistakenly suggests that pragmatics somehow takes over when semantics leaves off. It is one thing for a sentence to have the content that it has and another thing for a speech act of uttering the sentence to have the content it has. Even when the content of the speech act is the same as that of the sentence, that is a pragmatic fact, something that the speaker has to intend and the hearer has to figure out. (Bach 2007: 5) In intercultural communication the content of the utterance is very often the same as that of the sentence. That is a pragmatic act, as Bach says, something that the speaker intends to say and the hearer has to figure out. In L1 communication on which current pragmatic theories are built, there is much more of a gap between what is said and what is meant than in intercultural communication in which it is of utmost importance that the speaker should mean close to what she says otherwise the hearer may have difficulty in making sense of the speaker’s intention because of limited core common ground, shared knowledge and mutual norms.

5.5  Salience Effect in the Dialogic Sequence As discussed above there is always an intuitive possibility of a distinction between what a speaker says and what she or he actually implicates. The likelihood of the distinction is especially high when a communicative process is affected by salience. The socio-cognitive 85

Istvan Kecskes

approach proposed by Kecskes (2010, 2013) argues that both cooperation (recipient design) and egocentrism (salience) are present in the communication process all the time to a varying degree, and the interplay of conscious recipient design (result of cooperation) and subconscious salience (driving force behind egocentrism) is what shapes speaker production. As a semiotic notion, salience refers to the relative importance or prominence of signs. It is the most probable out of all possible. The relative salience of a particular sign when considered in the context of others helps an individual to quickly rank large amounts of information by importance and thus give attention to that which is most important. We tend to overestimate the causal role of information (salience) we have available to us both perceptually and linguistically. The socio-cognitive approach (Kecskes, 2010, 2013) claims that interlocutors should be considered social beings searching for meaning with individual minds embedded in a sociocultural collectivity. Grice was right when he tied cooperation to speaker–hearer’s rationality. However, egocentrism must be added to speaker–hearer’s rationality. We human beings are just as egocentric (as individuals) as cooperative (as social beings). ‘Egocentrism’ in the socio-cognitive approach refers to attention-bias that is the result of prior experience of individuals. It means that interlocutors activate and bring up the most salient information to the needed attentional level in the construction (by the speaker) and comprehension (by the hearer) of the communication. So there is nothing negative about egocentrism if the term is used in this sense. It should not be confused with ‘egotistic’ that refers to a person who is self-centered and focuses only on his or her agenda. In pragmatics, based on Giora’s graded salience hypothesis (Giora, 1997, 2003), when we speak about salient information we mean given information that the speaker assumes to be in central place in the hearer’s consciousness when the speaker produces the utterance. Salience has two important features that affect the dialogic sequence and dialogic actions of interlocutors. First, as discussed above, salience affects not only comprehension (as the graded salience hypothesis of Giora says) but also production (Kecskes, 2010, 2013). Second, salience is based on prior experience that is usually different for each individual. This may result in misunderstanding that needs to be repaired in the dialogic sequence which gives less chance to the speaker to manipulate meaning according to his/her needs and unfold his/her agenda because repair is generally dictated by the hearer’s needs that the speaker tries to meet. The following short dialogue demonstrates what happens when both speakers’ production is dominated by salience rather than recipient design (excerpt is from the TV comedy Two and a Half Men). (6) Charlie: - Morning. Alan: - What’s so good about it? Charlie: - I ain’t say ‘good’. Neither Charlie nor Alan did much recipient design in this interaction. Charlie said “morning” (which may be considered an elliptical form of “good morning”) automatically when noticing Alan. Although the utterance was produced subconsciously it was relevant in that actual situational context. However, Alan did not pay close attention to what exactly Charlie had said, rather relied on the situational frame (morning meeting). He was in a bad mood and may have been affected by some negative prior experience the previous evening. He assumed that Charlie had uttered the usual formula “good morning”, and his response was formulated accordingly. 86



From Pragmatics to Dialogue

The analysis of dialogic sequences show very clearly what the difference is between an approach where we handle speaker and hearer separately and where we handle dialogue participants as interlocutors who are speaker–hearers. If we talk about speaker–hearers we need to accept that salience (just like cooperation in the Gricean sense) will affect whatever the speaker–hearer does, let it be production or interpretation. But Kecskes (2008: 401) argued that there is a difference between scenarios when the interlocutor acts as speaker and when s/he acts as hearer. When a lexical unit (labeled for private context) is used by a speaker to produce an utterance, private contexts (prior experience of the speaker) attached to this lexical expression are activated top-down in a hierarchical order by salience. This hierarchical order works differently for the speaker and the hearer. For the speaker, there is primarily an inter-label hierarchy, while for the hearer the intra-label hierarchy comes first. The interlabel hierarchy operates in the first phase of production, when a speaker looks for words to express her/his intention. As a first step, s/he has to select words or expressions from a group of possibilities in order to express his/her communicative intention. These words or expressions constitute a hierarchy from the best fit to those less suited to the idea s/he is trying to express. To explain how this works we will analyze an excerpt from a movie. (7) Excerpt from the movie: Angel Eyes. Situation: A policewoman in uniform is driving the car, and the man sitting beside her is staring at her. PW: - What? M: - I was trying to picture you without your clothes on. PW: - Excuse me? M: - Oh no, I did not mean like that. I am trying to picture you without your uniform. PW: - Okaay? M: - I mean, on your day off, you know, in regular clothes. Excerpt (7) demonstrates a deductively built-up dialogic sequence. Salience effects usually result in deductive sequences where the speaker has something on his/her mind, and this intention is formulated abruptly, rather carelessly without specific planning, as seems to be the case in example (7). This dialogue appears to support the claim of cognitive psychologists according to which the initial planning of utterances ignores common ground (egocentric approach), and messages are adapted to addressees only when adjustments are required (Barr and Keysar, 2005; Keysar, 2007; Barr, 2004). It looks like salience effect generally appears in the form of a deductive process that may contain repairs and adjustments. Why were the man’s attempts unsuccessful in the conversation in (7)? Because, on the one hand we have the speaker with an intention to tell the woman why he was looking at her the way he did, and on the other hand there is subconscious salience that affected how he formulated his intention. So the utterance was not designed by the speaker to fit his words into the actual situational context, taking into account his partner. As a result, the word selection was wrong. Why was word selection wrong? Because it was not directed by recipient design but was prompted by salience. As argued above, salience (motivated by prior experience), which operates subconsciously and automatically, may affect word selection and utterance formation. “I was trying to picture you without your clothes on.” Is this what the speaker wanted to say and mean? Yes, this is exactly what he wanted to mean but not necessarily what he wanted to say. I agree with Bach (2005) that a speaker can always mean something distinct from the semantic content of the sentence he is uttering. But even if he speaks 87

Istvan Kecskes

literally, and means precisely what his words mean, even that fact depends on his communicative intention (Bach 2005: 27). Wording, expressing intention in words is a tricky thing. Conceptualization is one thing, wording is another and meaning is a third one. There is no one-to-one relationship between any of the three. Now we should look at how salience works for the interlocutor when s/he is in the role of hearer. The hearer has to cope with a different type of hierarchy than the speaker from her/his perspective. Thus, an intra-label hierarchy is in force when the hearer processes (a) lexical unit(s) in an utterance (or even an entire utterance). The label (word) uttered by the speaker hierarchically triggers the history of that particular label as used by the hearer (but not by the speaker). This may also be a reason for misunderstanding in the communicative process, both in intracultural and intercultural interactions. Compare the following (source: American sitcom) interchange in (8): (8) Bob: Are you OK? Mary: I am fine. Bob: I know you are fine, but are you OK? Bob had several options to ask about Mary’s well-being: “Are you OK?”, “Are you fine?”, “Is everything all right?”, and so forth. His selection of “Are you OK?” caused a slight misunderstanding between the two because they interpreted “OK” differently. As we have seen, salience is both an individual and a societal phenomenon. Prior and actual situational experience is privatized/subjectivized and prioritized in the mind of interlocutors. Their different prior experiences, their different evaluations of the actual situational context, their dynamically changing intentions and individual degrees of salience result in a subjectivized process of production and comprehension. As a result, there may be no single point in the recovery process at which a speaker’s utterance fully matches a hearer’s interpretation. This is because both a speaker’s production and a hearer’s interpretation are ‘contaminated’ by individualized pragmatic elements. Pragmatic enrichment processes work differently for speaker–hearers depending on which role they are in. Consequently, the match between the two sides keeps varying in the communicative process. This is why we think that: “we almost always fail. […] Yet we almost always nearly succeed” (Rapaport, 2003: 402). And this is why a pragmatic theory should be interlocutor-centered that takes the individual as a speaker–hearer. Speaker’s production and hearer’s interpretation should be analyzed in their own rights. To get a full picture of what is going on in communication we need to look at a dialogic sequence where the role of salience is minimal and recipient design and intention recognition comes to the fore. (9) This is an excerpt from the film Coogan’s Bluff. Situation: A man and a young woman are sitting in a restaurant after a meal. The woman stands up and with a short move reaches for her purse. W: - I have to be going. M (seeing that she reaches for her purse): — What are you doing? W: - Dutch. M: - You are a girl, aren’t you? W: - There have been rumors to that effect. M: - Sit back and act like one. W: Oh, is that the way girls act in Arizona? 88



From Pragmatics to Dialogue

When the girl wants to pay the man expresses his objection with asking “what are you doing?”. The girl perfectly understands what the man is referring to, so she tells him “Dutch”, which means she wishes to pay for her share of the bill. The man indicates his dislike in a very indirect but still expressive way: “You are a girl, aren’t you?” The girl’s response shows that she follows where the man is getting to. Then the man hints at what he expects the girl to do: “Sit back and act like one”. This means that he does not want her to pay the bill. This inductively developed dialogic sequence is a good example for elaborated recipient design where nothing is said directly, still there is no misunderstanding because the speaker adequately alerts the hearer to what he means. According to the recipient design view in order to succeed speakers must correctly express intended illocutionary acts by using appropriate words, and make their attempt in an adequate context. In this process speakers relate propositional contents to the world (actual situational context; audience) with the intention of establishing a correspondence between words and things from a certain direction of fit. This is what happened in example (9). When looking at the dialogic sequence from the speaker’s perspective it appears that recipient design usually requires an inductive process that is carefully planned while salience effect generally appears in the form of a deductive process that may contain repairs and adjustments. The speaker has something on his/her mind, and this intention is formulated subconsciously, abruptly, rather carelessly without specific planning. From the hearer’s perspective intention recognition can also be driven by salience that may frequently lead to misunderstandings because of the private and cultural specific nature of salience. The interplay of recipient design and salience demonstrates very well how cooperation (social) and egocentrism (individual) are intertwined in the communicative process. This gives strong support to the dialogic principle which claims that human beings are dialogic individuals (social individuals) who communicate in dialogic interaction not only by producing and understanding utterances but also by acting and reacting.

5.6 Conclusion The aim of the chapter was to direct attention to how a dialogue approach can lead to better understanding of actions of interlocutors in communication. It was claimed that utterances in pragmatics research should be analyzed from the perspective of the interlocutor that comprises both roles: speaker and hearer. However, if we want to focus both on the speaker’s and hearer’s perspective, and want to figure out why the speaker said what s/he said the way s/he said it, and why the hearer opted for a given interpretation, we need to use a holistic, top-down approach combined with the bottom-up, utterance by utterance analysis. This means that we need to analyze the dialogic sequence or discourse segment rather than just the utterance in order to find cues that help us identify the real intention of the speaker and better understand the options of the hearer in interpreting the speaker’s utterance. Following Saul (2002), it was suggested that we make a difference between utterer-implicature and audience-implicature. In order for us to understand meaning in interaction we need to focus on the double role of interlocutors. While fitting words into actual situational contexts speakers are driven not only by the intent (conscious) that the hearer recognize what is meant as intended by the speaker (cooperation), but also by speaker individual salience that affects production subconsciously (egocentrism). Therefore it was proposed that speaker–hearer rationality should include not only cooperation but egocentrism as well (see also in Kecskes, 2013). 89

Istvan Kecskes

Through some examples it was pointed out that recipient design usually requires an inductive process in the dialogue that is carefully planned while salience effect generally appears in the form of a deductive process that may contain repairs and adjustments. The speaker has something on his/her mind, and this intention is formulated abruptly, rather carelessly without specific planning. The salience effect can lead to unwanted ways of expressing intention and occasionally misunderstanding, which require the speaker to make repairs and adjustment in the communicative process and work together with the hearer to come to an understanding. On the other hand, recipient design gives the speaker the opportunity to manipulate speaker meaning according to his/her needs: leave meaning conversationally open for interpretation or signal his/her intention with cues and markers. Here the goal of the speaker is to promote his/her own agenda by trying to manipulate the interpretation process. A dialogic, ‘wide pragmatics’ approach may give us the chance to better understand these complexities of the communicative process.

References Airenti, Gabriella, Bara, Bruno G. and Colombetti, Marco. 1993. Conversation and behavior games in the pragmatics of dialogue. Cognitive Science 17(2): 197–256. Bach, Kent. 2001. You don’t say? Synthese 128: 15–44. Bach, Kent. 2005. Context ex Machina. In Semantics vs. Pragmatics, Szabó Zoltán Gendler (ed.), 15–44. Oxford: Oxford University Press. Bach, Kent. 2007. Regressions in pragmatics (and semantics). In Pragmatics (Advances in Linguistics), Burton-Roberts Noël (ed.), 24–44. Houndmills: Palgrave-Macmillan. Bara, Bruno. 2011. Cognitive pragmatics: The mental processes of communication. Intercultural Pragmatics 8–3 (2011): 443–485. Barr, Dale J. 2004. Establishing conventional communication systems: Is common knowledge necessary? Cognitive Science 28(6): 937–962. Barr, Dale J. and Keysar, Boaz. 2005. Making sense of how we make sense: The paradox of egocentrism in language use. In Figurative Language Comprehension: Social and cultural influences, Colston Herbert L. and Albert N. Katz (eds.), 21–43. Mahwah, NJ: Lawrence Erlbaum. Buber, M. 1955. Dialogue. In Between Man and Man, R. G. Smith (trans.), pp. 1–39. Boston, MA: Beacon Press. Bunt, H. C. 2011. The Semantics of Dialogue Acts. Proceedings 9th International Conference on Computational Semantics (IWCS 2011). J. Bos and S. Pulman (eds.), 1–14. Oxford, UK: Oxford University. Burton-Roberts, Noël. 2006. Cancellation and intention. Newcastle Working Papers in Linguistics 12: 1–12. Carbaugh, Donal. 2013. On Dialogue Studies. Journal of Dialogue Studies 1(1) (2013): 9–28. Carston, Robyn. 2002. Thoughts and Utterances: The pragmatics of explicit communication. Oxford: Blackwell. Cooren, F. 2010. Action and agency in dialogue: Passion, incarnation, and ventriloquism. Amsterdam: John Benjamins. Davis, Wayne. 1998. Conversational Implicature: Intention, convention and principle in the failure of Gricean theory. Cambridge: Cambridge University Press. de Saussure, Louis. 2007. Pragmatic issues in discourse analysis. Critical approaches to discourse analysis. Across Disciplines 1(1): 179–195. Garfinkel, H. 1972. Remarks on ethnomethodology. In Directions in Sociolinguistics, J. J. Gumperz and D. H. Hymes (eds.), 301–309. New York: Holt, Rinehart & Winston.

90



From Pragmatics to Dialogue

Giora, Rachel. 1997. Understanding figurative and literal language: The graded salience hypothesis. Cognitive Linguistics 8(3): 183–206. Giora, Rachel. 2003. On Our Mind: Salience context and figurative language. New York: Oxford University Press. Horn, Laurence R. 2004. Implicature. In The Handbook of Pragmatics, Horn Laurence R. and Gregory Ward (eds.), 3–28. Oxford: Blackwell. House, Juliane. 2002. Developing pragmatic competence in English as a lingua franca. In Lingua Franca Communication, K. Knapp and C. Meierkord (eds.), 245–267. Frankfurt am Main: Peter Lang. Kecskes, Istvan. 2007. Formulaic language in English Lingua Franca. In Explorations in Pragmatics: Linguistic, Cognitive and Intercultural Aspects, I. Kecskes and L. Horn (eds.), 191–219. Berlin: Mouton de Gruyter. Kecskes, Istvan. 2008. Dueling context: A dynamic model of meaning. Journal of Pragmatics 40(3): 385–406. Kecskes, Istvan. 2010. The paradox of communication: A socio-cognitive approach. Pragmatics and Society 1(1): 50–73. Kecskes, I. 2012. Is there anyone out there who really is interested in the speaker? Language and Dialogue 2(2): 285–299. Kecskes, Istvan. 2013. Intercultural Pragmatics. Oxford, UK: Oxford University Press. Keysar, Boaz. 2007. Communication and miscommunication: The role of egocentric processes. Intercultural Pragmatics 4(1): 71–84. King, Jeffrey C. and Stanley, Jason. 2005. Semantics, pragmatics, and the role of semantic content. In Semantics versus Pragmatics, Szabó Zoltán Gendler (ed.), 111–164. Oxford: Oxford University Press. Moeschler, J. 2004. Intercultural pragmatics: A cognitive approach. Intercultural Pragmatics 1(1): 49–70. Morante, R., Keizer, S. and Bunt, H. C. 2007. A dialogue act based model for context updating. Proceedings of the Eleventh International Conference on the Semantics and Pragmatics of Dialogue (DECALOG 2007). Trento, Italy: 9–16. Psathas, O. (ed.). 1979. Everyday Language. Studies in Ethnomethodology. New York: Irvington. Puig, Margarida Bassols. 2003. Pragmatics and discourse analysis. Noves SL. Revista de Sociolingüística. Available at http://www.gencat.cat/llengua/noves [Winter 2003]. Rapaport, William J. 2003. What did you mean by that? Misunderstanding, negotiation, and syntactic semantics. Minds and Machines 13(3): 397–427. Sacks, H., Schegloff, E. A. and Jefferson, G. 1974. A simplest systematics for the organization of turntaking for conversation. Language 50: 696–735. Saul, Jennifer. 2002. What is said and psychological reality: Grice’s project and relevance theorists’ criticisms. Linguistics and Philosophy 25: 347–372. Schegloff, Emanuel A. 1979. Identification and Recognition in Telephone Conversation Openings. In Everyday Language: Studies in Ethnomethodology, G. Psathas (ed.), 23–78. New York: Irvington Publishers, Inc. Schenkein, J. (ed.). 1978. Studies in the Organization of Conversational Interaction. New York: Academic. Searle, John. 1995. The Construction of Social Reality. London: Allen Lane, Penguin. Sperber, Dan and Wilson, Deirdre. 1986/1995. Relevance: Communication and cognition (2nd ed.). Oxford: Blackwell. Taboada, M. and Mann, W. C. 2006. Rhetorical structure theory: Looking back and moving ahead. Discourse Studies 8(3): 423–459. Tracy, Karen and Robert T. Craig. 2010. Studying interaction in order to cultivate communicative practices: Action-Implicative discourse analysis. In New Adventures in Language and Interaction, Jürgen Streeck (ed.), 145–166. Amsterdam: John Benjamins.

91

Istvan Kecskes

Turner, R. (ed.). 1974. Ethnomethodology: Selected readings. Harmondsworth, England: Penguin. Weigand, Edda. 2001. Negotiation and Power in Dialogic Interaction, Edda Weigand and Marcelo Dascal (eds.). Amsterdam: Benjamins (Current Issues in Linguistic Theory 214). Weigand, Edda. 2010a. Language as dialogue. Intercultural Pragmatics 7(3): 505–515. Weigand, Edda. 2010b. Dialogue: The mixed game. Amsterdam: Benjamins.

92

6 Psycholinguistic Approaches Meaning and Understanding Susan E. Brennan and Joy E. Hanna

6.1 Introduction Consider what two friends passing in a hallway on a Monday morning can accomplish with just a few words: S: how was it? J: great! Previously, J had told S that she was going to a party on the weekend; when S asked how it went, J had no trouble understanding S’s intended meaning, or providing an answer that the party went well. This type of speaking and understanding is both completely unremarkable in that it is accomplished easily and without conscious effort, and yet remarkable given the amount of linguistic and non-linguistic information that must be combined in order to communicate successfully. In the interaction above, “it” and “great” could have referred to a burger that tasted fantastic, a sabbatical that was more productive than expected, or even an ironic response about a date that was actually cringe worthy; the word senses, prosody, identity of the speakers, their current setting, and previous interactions (along with the ability to activate relevant information in memory) all combined to achieve shared meaning and understanding. In addition, common ground between the friends, perceptual salience, hand gestures, body orientation, or eye gaze could also have played a part in both the linguistic choices that S made, and whether J was able to respond to S’s question without clarification. There have been two rather distinct traditions in mainstream psycholinguistics. The dominant tradition has considered meaning to be in the province of speakers, and understanding in the province of addressees. This conceptualization has led to the study of these topics as separate from each other, with a focus on individuals’ linguistic representations and processes. As Clark (1992) described this language-as-product tradition, the business of psycholinguistic inquiry was to start with linguistic structures and determine their effects (separately) on language production and comprehension. Rationales for this approach include an interest in studying pure ‘competence’ or linguistic knowledge (Chomsky, 1957), as well as the need to achieve sufficient experimental control. Focusing solely on linguistic 93

Susan E. Brennan and Joy E. Hanna

factors in individuals acting alone reflects the assumption that language is for thinking, rather than for communicating. But regardless of what language is presumed to be ‘for’, the products and processes uncovered in such language studies do not necessarily scale up to, or provide a complete picture of, the products and processes that unfold in interactive dialogue contexts. The other long-standing tradition in psycholinguistics (albeit a minority one) has considered meaning and understanding to be part of the same inquiry. This conceptualization cuts across cognitive and social psychology, linguistics, and ethnography to focus on language as it is used for communication. This language-as-action tradition (Clark, 1992) has examined psycholinguistic phenomena in communicative contexts. In conversation, arguably the fundamental arena for language processing (Clark, 1992), the current speaker is not only planning and articulating speech, but also monitoring her addressee for evidence of understanding and uptake, while the current addressee is simultaneously listening and planning what to say next. They closely time their speaking turns and nonverbal responses with each other’s behavior; on top of all that, they may be coordinating their actions in a collaborative task (Brennan, 1990; Clark, 1992; Garrod and Pickering, 2004). This focus leads to considerably messier data that the product tradition dismissed as language ‘performance’, distinct from the concerns of traditional sentence processing research. Another reason for avoiding data from spontaneous dialogue has been the difficulty of achieving sufficient precision in methods and measures to test particular hypotheses. Conversational exchanges were considered by many to be too noisy to use as a data source for the scientific study of language and cognitive processes, and so the dominant product tradition kept the focus on idealized sentences, leaving spontaneous language to the action tradition. In this chapter, we discuss the study of meaning and understanding as approached by these two traditions in psycholinguistics. Then we discuss technological advances in the last two decades that have enabled psycholinguistics to develop unobtrusive methods and methodological innovations for studying language processing online or moment by moment, while it is used in communicative contexts. These advances have enabled the two disparate psycholinguistic traditions to begin to converge toward a more holistic approach to uncovering the products and processes of meaning and understanding that, we believe, characterizes the current state and future of modern psycholinguistics.

6.2 Traditional Psycholinguistics: Language-as-Product vs. Language-as-Action 6.2.1  Language as Product The dominant approach to psycholinguistics research after the cognitive revolution of the mid-twentieth century took the autonomous individual’s language knowledge as the focus of investigation. This tradition emerged in the 1960s and combined Chomsky’s (1957) work in linguistics with the emerging information processing approach to cognition (Miller, 1962); it emphasized individual perceptual and cognitive processes used to activate and build linguistic (especially syntactic) representations, or products, during language production and comprehension. The actual use of language was seen as irrelevant, subject to behavioral and other non-linguistic cognitive factors (e.g., disfluencies and memory failures) that were not considered to be at the core of linguistic competence. As the cognitive revolution unfolded, the study of language became the focus not only of linguists, but of psychologists, philosophers, computer scientists, and engineers. Many 94



Psycholinguistic Approaches

cognitive scientists viewed communication (explicitly or implicitly) as a process by which speakers encode thoughts or intended meanings into words (presumed to be little packages of meaning) that are then conveyed via some medium to addressees, who in turn decode words back into meanings. Critics of this simple chain of events dubbed it, variously, the ‘message model’ (Akmajian et al., 1987), the ‘conduit metaphor’ (Reddy, 1979), the ‘encoding/decoding paradigm’ (Krauss and Chiu, 1997), or ‘autonomous processing’ (Schober and Clark, 1989). The assumption behind this simple view is that communication should succeed as long as the interlocutors speak the same language (i.e., know the same linguistic encoding and decoding rules). Empirical studies in this tradition largely ignored aspects of language that were assigned to “performance,” and concentrated on the rules and representations accessed and created during linguistic processing. Many theoretical debates centered on how information from different linguistic sub-systems was processed and the extent to which it was integrated across sub-systems that might or might not be modular. Popular experimental methodologies were those capable of collecting reaction times moment-by-moment during ‘online’ language processing (beyond more traditionally linguistic ‘offline’ measures such as grammaticality judgments provided at the end of a sentence), in order to provide a finer-grained picture of the time-course of processing. For example, the waxing and waning of lexical representations activated while listening to short spoken discourses could be detected by measuring reaction times to words in a simultaneous lexical decision task (see, e.g., Swinney et al., 1978), and the effects of variations in syntactic structure could be examined by monitoring the duration and patterns of eye fixations (delays and regressions at difficult-to-process text) made by readers. While these studies considered effects of context on linguistic processing, context was largely defined as limited to surrounding words and sentences. The nature of contextual information, especially that from task goals, nonverbal signals, or common ground from interlocutors’ previous interaction, was certainly not the emphasis of the research enterprise. Methodological constraints have played a strong role in psycholinguistics. It was assumed that only carefully controlled, idealized stimuli (typically, written) could achieve the millisecond precision required in comprehension experiments. With this control, lexical items could be matched for generic frequency and neighborhood density or number of phonological or semantic associates; alternative syntactic structures or grammatical and ungrammatical versions of the same syntactic structure could be presented in planned increments on a computer screen, or ambiguous and unambiguous referring expressions could be used within the same sentential context. Such control is harder to achieve in production experiments. Uncovering the processes underlying a particular surface structure depends on eliciting, from multiple speakers, utterances comparable in form. Some (e.g., Allbritton et al., 1996; Price et al., 1991) had subjects read aloud, but this fails to evoke the same processes of planning and articulating as speaking spontaneously; speakers have an abundance of expressive choices (as Bock, 1996, put it, “exuberant responsing”). The challenge has been to implicitly guide speakers to produce analyzable data without destroying the underlying cognitive phenomena; the solution is that speakers are given something to do with their utterances (rather than simply something to read). For example, speakers may be asked to respond to text, image, or schematic prompts (Bock, 1996; Ferreira and Swets, 2005; Griffin, 2001; Kraljic and Brennan, 2005), produce sentences or phrases from memory (Ferreira and Dell, 2000; Schafer et al., 2000), or speak after being primed by task instructions or previous turns that prime the targeted forms (e.g., Hanna and Brennan, 2007; Haywood et al., 2005). Such approaches represent initial steps toward bridging the product and action traditions. 95

Susan E. Brennan and Joy E. Hanna

6.2.2  Language as Action In parallel with the language-as-product tradition, another approach to psycholinguistics focused on the spontaneous use of language for communication, dubbed the ‘languageas-action’ tradition (Clark, 1992). This approach stressed the pragmatic nature of meaning in the world, finding inspiration from the philosophy of language use (e.g., Austin, 1962; Grice, 1957; Searle, 1969) as well as from ethnomethodological accounts of naturally occurring conversations (e.g., Goodwin, 1981; Jefferson, 1974; Schegloff and Sachs, 1973). The language-as-action tradition examined how people use language to perform actions in conversation, including many of the performance factors that the language-asproduct tradition ignored. These actions are related to, but not identical with, speech acts as proposed within philosophical speech act theory, and more recently dialogic speech act theory (see Chapters 10 and 11 in this volume). Psycholinguistic research within this tradition, pioneered by Bavelas, Clark, Krauss, and colleagues (e.g., Bavelas and Chovil, 2000; Bavelas and Coates, 1992; Clark and Brennan, 1991; Fussell and Krauss, 1992; Krauss, 1987; Schober and Clark, 1989; Clark and Wilkes-Gibbs, 1986), used collaborative tasks, typically in settings with observable referents and well-defined behavioral goals. In contrast to the product tradition’s emphasis on basic linguistic products and processes, the action tradition has treated the processing involved in language use as more than just the sum of individual linguistic parts. Conversational participants jointly and incrementally coordinate the creation of successful meaning and understanding (see Clark, 1992, 1996; Clark and Wilkes-Gibbs, 1986; Krauss, 1987; Krauss and Chiu, 1997). Consider this example from an experiment in which D is trying to get C (who is seated behind a barrier) to identify a card with an abstract geometric figure (or tangram) on it (from Stellmann and Brennan, 1993 [tangram in Fig. 1]). D: alright ... the- the fifth one kinda looks like New Jersey? C: New Jersey ... D: yeah, it- um it ... does it- looks like a man with either a big nose ... C: yeah, pointing towards his nose- pointing towards the left- right. D: yes! C: okay that’s number 5, right D: yeah Here, D is trying to orient C to an object they have never discussed before, complicated by the fact that the object in question is not conventionally associated with any label. So D presents to C what is essentially a proposal, marked as tentative by a hedge (“kinda”), and uses final-rising intonation to invite C’s input. C considers this proposal while echoing “New Jersey” but with intonation that appears to display a lack of understanding and uptake. D then tries again with “a man” (perhaps one with a big nose). C takes up this counter proposal, providing additional evidence (if a bit disfluently) about its appropriateness, which D ratifies with “yes!”, and on they go. The next time they refer to this object during a later round, one of them calls it “the man with the big nose”, to which the other readily agrees. D and C have constructed a joint meaning for this expression that reflects a shared perspective on its referent; they mark this by re-using the same referring expression when they rerefer to it (a process called ‘entrainment’; Garrod and Anderson, 1987; Brennan and Clark, 1996). Interlocutors seek and provide evidence about what they and their partner understand (in a process known as ‘grounding’; Clark and Brennan, 1991; Clark and Wilkes-Gibbs, 96



Psycholinguistic Approaches

1986), and in this way, they build up common ground that enables them to become more efficient with repeated referring. If others were to join the conversation at this point, they would not understand the referring expression as accurately or as easily as D and C even if they were equally familiar with the object outside of the conversational context (as demonstrated by Schober and Clark, 1989). Odd geometric objects are one thing; one might wonder whether referring to common objects works differently. After all, native English speakers have conventionalized associations for basic level terms like shoe, dog, car, and fish. However, even when common labels can be associated with objects, speakers do not always use them. In one study (Brennan and Clark, 1996, Experiment 1), speakers initially referred to common objects by their basic level terms (shoe, dog, car, or fish) only 77 percent of the time, even though these short, common words could have sufficed for identification. And these speakers produced over-informative expressions as often as 80 percent of the time when referring to a quite typical shoe, dog, car, or fish that had previously (but no longer) needed to be distinguished from other shoes, dogs, cars, or fish. Could speakers have just not bothered to re-evaluate how much information their addressees needed? Perhaps these speakers were just doing what was easiest—re-using expressions that worked the last time around (in fact, this account was proposed by Garrod and Anderson, 1987). However, in a follow-up study (Brennan and Clark, 1996, Experiment 3), speakers were over-informative more often when maintaining precedents they had set up with a particular addressee (with whom they had discussed the objects previously) than when beginning afresh with new addressees. With a new addressee, they often proposed terms similar to those they had used with the old addressee (which is not surprising, as these would have been highly available in memory), but in elaborated forms that marked their awareness of the need to ground this information anew (a finding replicated in Horton and Gerrig, 2005, 2002; Gorman et al., 2013). Thus, referring expressions can be viewed as products that emerge not only from solitary processes of comprehension and production, but also from interaction between interlocutors. Convenient as it may be to think of words as little packages of meaning by which autonomous speakers share their thoughts with addressees, things are not so simple. In order for referring to succeed in contexts like those in the examples above, interlocutors must have access to information that is mutually known—that is, in their common ground—that they can reasonably depend on being accessible to both of them in the moment. The languageas-action tradition has explored the use of common ground during spontaneous communication in controlled laboratory settings, where pairs of subjects do referential communication tasks that provide readily available visible evidence about what they mean and understand (e.g., picture cards to match, a map to traverse, objects to assemble, etc.). In such referential communication tasks (popularized by Krauss and colleagues: Glucksberg, Krauss, and Weisberg, 1966; Krauss and Weinheimer, 1964), pairs of subjects do a task where one has information that the other does not. For example, a director instructs a matcher to construct, draw, or arrange objects so that their displays match, repeating the task multiple times. Partners are typically separated by a barrier and allowed to communicate spontaneously (as in our previous example of the two subjects discussing the tangram that looked like New Jersey). The measures obtained in these tasks include numbers of words, turns, and errors (as proxies for effort and efficiency), codable elements such as referring expressions, disfluencies, hedges, and backchannels, and the time it takes to complete the task. Although these measures are informative, both qualitatively (supporting rich descriptions of language use) and quantitatively (enabling hypothesis-testing), typically the evidence they provide is 97

Susan E. Brennan and Joy E. Hanna

aggregated at a relatively coarse grain. Even if language-as-product researchers had wanted to explore the online cognitive and linguistic effects of being engaged in conversation, that would have proved difficult because the stimuli and responses in a dialogue are inherently idiosyncratic and contingent upon one another. Despite their advantages, referential communication tasks by themselves do not make it easy to assess precisely how information is integrated online, in the service of meaning and understanding. Unsurprisingly, then, many psycholinguists have studied linguistic processing abstracted away from dialogue (e.g., comprehension in static contexts, or linguistic phenomena distributed within text or speech corpora), but not as it occurs during interpersonal coordination between speakers and addressees, nor intra-personally (e.g., how hearing and interpreting an utterance is affected by simultaneously planning what to say next). To return to our examples, it wouldn’t have been feasible to interrupt S’s and J’s brief and informal exchange or D’s and C’s referential communication task every few milliseconds to ask what they were thinking, or to require them to metacognitively assess their own (likely inaccessible) language processes. To return to our opening example, how do we know whether S even considered using the phrase the party instead of the deictic it, or whether J was briefly puzzled by it and considered other possible meanings before landing on the correct one, or what non-linguistic information may have affected their processing? The challenge is how to study conversational interaction in the laboratory in a way that preserves its spontaneity and contingent structure, as well as the fine-grained incremental cognitive processes of production and comprehension. Both the action and product traditions were in need of better online measures in order to probe cognitive processing, whether in speakers or addresses or both.

6.3  Toward a New Psycholinguistics: Bridging the Traditions In the past 20 years, due in part to new technological and methodological developments, psycholinguistics has begun to combine the strengths of the language-as-product and language-as-action traditions to examine meaning and understanding on a moment-bymoment basis. This has allowed for the testing of less reductionist predictions made by more holistic theories of language processing. From the product tradition, some researchers first took advantage of these developments to reproduce effects that had also been found in monologues and non-communicative contexts. Others explicitly explored the effects of discourse context (that is, the broad communicative context in which goal-oriented spoken or written language is being used) on language understanding. From the action tradition, some researchers were eager to deploy new technologies to measure the cognitive processing that takes place during interactive discourse, with a level of precision that allowed synchronizing streams of verbal and nonverbal data. Such streams include spontaneous speech, gestures, postural sway, eye gaze, and mouse cursor movement during screen-based tasks. A few such investigations have synchronized one or more kinds of data streams generated by pairs of spontaneously interacting partners as they coordinate language and action in task-oriented dialogue. As language-as-product and language-as-action approaches have begun to converge, computational linguistics research in speech recognition, parsing, and message generation (much of which was associated with the product tradition) has embraced the use of natural text and speech corpora as data, or have become integrated with human–computer interaction research in dialogue. This has supported the development of spoken dialogue systems and animated conversational agents that are now in use by the public. In this section, we briefly discuss and illustrate these changes. 98



Psycholinguistic Approaches

6.3.1  Gaze Tracking and the Visual World Paradigm The development of accurate, relatively inexpensive head-mounted eye-tracking systems (see Henderson and Ferreira, 2004; Trueswell and Tanenhaus, 2005) made it possible to combine strengths of the product and action traditions. Subjects could wear a lightweight apparatus similar to a bike helmet (or be monitored by a desk-mounted camera) that recorded the direction and duration of gaze during referential communication or other tasks. This technology enabled its pioneers to study incremental language processing in what has come to be known as the “visual world” paradigm (e.g., Tanenhaus et al., 1995; Velichkovsky, 1995). In the typical visual world experiment, a subject follows spoken instructions to manipulate physical objects or objects displayed on a screen, while their eye-movements are monitored at a resolution between 30 and 2000 samples per second, depending on the equipment. Listeners tend to look at objects when they hear them mentioned (Cooper, 1974; Tanenhaus et al., 1995), especially when they need to manipulate them, and fixations are incrementally time-locked to the unfolding utterance so that they provide a fine-grained real-time measure of comprehension (and production; see Griffin and Bock, 2000; Meyer et al., 1998). It takes about 150 milliseconds (ms) on average for a syllable to unfold in running speech, and about 150 ms to plan and execute a saccade (with listening and looking proceeding in parallel, as listeners do not need to wait until they’ve heard the whole of a word before looking toward its referent). The looks toward relevant target objects are graphed over time, along with changes in the probability of looking at other similar or competing objects in the task. Measures such as fixation probability, first and last looks before a selection, fixation duration, patterns of gaze over sets of objects, and repeated looks to the target are recorded with respect to key points in time during the utterance, providing a measure of how long it takes a listener to commit to an interpretation relative to the spoken onset of the referring expression (or other element of interest). One such key point that figures prominently in many visual world studies is the point of disambiguation (POD). Many referring expressions are temporarily ambiguous among potential referents; as the speaker produces an instruction to select or move an object, the POD is when the hearer has enough linguistic information to disambiguate the expression. In one of the first studies to use the visual world paradigm, subjects viewed displays with four objects, including colored squares and rectangles, some of which were starred (Eberhard et al., 1995). On a critical trial, subjects were told to “touch the starred yellow square.” This instruction was paired with one of three different displays: one containing only a single starred object, so the target could be identified at starred; a second with multiple starred objects, but only one that was yellow, so the target could be identified at yellow; and a third containing a starred yellow square (the target) and a starred yellow rectangle, so the target could be identified only at the final word square. Listeners identified the target sooner when the display provided an earlier POD, demonstrating critical support for the incremental and pragmatic nature of language comprehension. Fixations also reflect anticipatory processing, such as when listeners generate expectations about what will be referred to, as they hear a referring expressions unfold over time, or when they anticipate that a contrast will be made with something in the discourse context (e.g., Altmann and Kamide, 2003; Brown-Schmidt et al., 2008; Dahan et al., 2002; Kim et al., 2015). This methodology has provided a powerful tool for examining the influence of higherlevel sources of information on reference resolution. The finding that information from a visual scene can eliminate or dramatically reduce syntactic ambiguity (Tanenhaus et al., 1995) 99

Susan E. Brennan and Joy E. Hanna

was a crucial turning point in understanding the role of context. These and subsequent experiments using the visual world paradigm (e.g., Chambers et al., 2004; Spivey et al., 2002) have demonstrated that information from the discourse context, such as the number and features of the potential referents, immediately constrains interpretation—there is no early, context-free processing stage (or module) where syntactic information is privileged. Subsequent research extended these findings, demonstrating that contextual information is combined with a variety of sources of information including verb bias, thematic fit, prosody, affordances between objects, the pragmatics of particular actions, perceptual salience, and so forth, to guide language processing decisions over time (e.g., Altmann and Kamide, 1999; Dahan and Tanenhaus, 2004; Hanna and Tanenhaus, 2004; Sedivy et al., 1999; Tanenhaus and Brown-Schmidt, 2008; Watson et al., 2008; Wilson and Garnsey, 2009). The visual world paradigm, therefore, met the methodological criterion of those who wanted an unobtrusive online measure of processing and was adopted relatively quickly to study product tradition questions, such as the influence of visual context on the activation of lexical representations or syntactic parsing commitments. Findings showed that language understanding takes into account not only unfolding linguistic input, but also the domain against which that input is being interpreted, and that visual and other perceived (non-linguistic) information changes the domain of interpretation. This was an important finding because it both met the scientific standards of those in the product camp and confirmed what those in the action camp had been arguing all along: that the process of understanding a sentence or an utterance is more complex than simply mapping words onto entities in the mind or in the world. In essence, researchers could read the minds of subjects as they coordinated meaning and understanding during conversational interaction.

6.3.2  From Flotsam and Jetsam to Signals for Coordinating Meaning Another advance that has been transforming psycholinguistic approaches to meaning and understanding is an increased interest in how pairs of people coordinate actions and processing during language use (a form of ‘joint action’; Clark, 1992, 1996; Galantucci and Sebanz, 2009; see also Kempson, Chapter 12 in this volume). Recall that the product and action traditions conceptualize meaning and understanding in distinctly different ways. The product tradition restricts domains of meaning to lexical items and their combinations, including abstract logical and grammatical constraints, conventionalized meanings, knowledge schemas, and propositional content; understanding consists of decoding. In contrast, the action tradition integrates individual linguistic and semantic knowledge with a critical focus on meaning as achieved jointly by speakers and addressees. Potential cues for coordinating meaning, such as gesture, gaze, prosody, speech repairs, disfluencies, pauses, interruptions, and side sequences, were first described in detail by conversation analysts (e.g., Goodwin, 1981; Jefferson, 1974; Schegloff and Sacks, 1973); these were the very elements that were discarded as irrelevant artifacts of language use by Chomskyan linguists and psycholinguists in the product tradition. The intuitions by conversation analysists that these phenomena were actually orderly and interpretable were mined by experimentalists in the action tradition, who attempted to apply (more or less controlled) experimental methods to establish the social and cognitive functions of such signals. In other words, the language-as-product tradition’s noise is, potentially, part of the language-as-action tradition’s signal. This distinction is captured well by Clark’s concepts of ‘Track 1’ and ‘Track 2’ signals (Clark, 1994, 1996). Track 1 signals (the principal concern of the language-as-product tradition) constitute the “official business” of an utterance, or 100



Psycholinguistic Approaches

messages encoded and decoded from syntactic and semantic knowledge that is presumed to be shared by the members of a language community. In our previous example, a perspective amounting to the man with a big nose pointing towards the right was co-created by D and C to refer to the fifth tangram, although this expression is an idealization that was actually uttered by neither. The language-as-action tradition, on the other hand, predicts that Track 2 signals provide critical metalinguistic information about a speaker’s stance toward her utterance, and are used to coordinate meaning and construct common ground. D, in our example, uttered two proposals with hedges (“looks like New Jersey?” and “looks like a man with either a big nose ...”) that she marked with cues that seemed to communicate that she was not yet committed to these proposals, suggesting that she was seeking (or open to) input from C. She interrupted herself several times (“the- the fifth one ...” and “it- um it ... does it looks like ...”) and produced pauses and the filler “um”, which appeared to provide evidence that she was formulating or reformulating what she was about to say next, and explicitly or implicitly seeking confirmation from C. Over the last decade or so, psycholinguists have been able to wrestle such phenomena under experimental control, to examine systematically and with temporal precision the causes and effects of spontaneously produced Track 2 signals within dialogue contexts. Here are just a few examples. Disfluencies. Track 2 signals can be both information-bearing (Smith and Clark, 1993) and interpretable as such (Brennan and Williams, 1995; Brennan and Schober, 2001; Fox Tree, 1995; Swerts and Krahmer, 2005) about speakers’ commitment toward or confidence in what they are saying, shaping how the message in Track 1 is understood, even during the earliest moments of interpretation. For example, in a visual world study (Arnold et al., 2004), disfluencies before a temporarily ambiguous referring expression (e.g., camel, when the display also included the cohort competitor candle) caused listeners to gaze more at the previously unmentioned versus the previously mentioned competitor, indicating that they interpreted the disfluency to mean the speaker was experiencing difficulty or increasing effort (to plan and articulate new rather than given information). Moreover, this rapid interpretive process is not reflexive, but is tailored to attributions about the speaker that guide inferences about the source of disfluency; when listeners were told that the speaker had an object-naming deficit (which would make their speech more disfluent), the effect went away (Arnold et al., 2007). Pitch accents and information status. Words repeated by a speaker within a discourse are typically articulated more clearly with the first use and attenuated with subsequent uses; new or unpredictable information receives a pitch accent, while given or predictable information is de-stressed (Firbas, 1992; Fowler and Housum, 1987; Lieberman, 1963; McAllister et al., 1994). Visual world evidence from ambiguous contexts established that pitch accents marking the information status of a referring expressions as given or new have an immediate effect on understanding, with more looks during accented words to new items and more looks during de-stressed words to given items (Dahan et al., 2002). This result shows that the “meaning” of a word includes its information status in the discourse. Backchannels. Some studies have used tasks and measures from the product tradition to examine the effects of Track 2 signals upon comprehension of speech recorded from spontaneous, naturalistic interactions. For instance, Tolins and Fox Tree (2016) showed that listeners to pre-recorded natural dialogues were able to use the backchannels in those dialogues as predictive cues; specific backchannels (“oh” and “really”) slowed monitoring for an upcoming word, appearing to focus listeners back onto material just mentioned, while generic backchannels (“umhm”, “uh huh”, and “yeah”) and pauses did not. In addition, specific backchannels sped word verification when the upcoming speech continued with 101

Susan E. Brennan and Joy E. Hanna

the same topic, and slowed verification when the upcoming speech elaborated with new information. Although this study involved subjects as overhearers to spontaneous dialogues, it is certainly possible that backchannels could have compatible effects on addressees within such dialogues, if a suitable method could be found to measure this online. Such hypotheses remain to be tested, as the methods of psycholinguistics continue to evolve. The studies used as examples in this section on coordinating signals have tried to bridge the product and action traditions by using pre-recorded speech (sometimes from natural dialogue) to test listeners’ understanding of linguistic and dialogue phenomena. That is of course a compromise. For the rest of this chapter, we focus on interaction between live partners.

6.3.3  Coordination between Speakers and Addressees, Online For many cognitive scientists who study dialogue, the ideal is to be able to examine incremental language processing—both comprehension and production—as it unfolds simultaneously within and between the minds of interacting speakers and addressees. Some consider this goal to be largely aspirational, given the experimental control needed in order to achieve interpretable and generalizable results. But some investigations have achieved sufficient control to support conclusions about online processes of coordinating meaning and understanding, both inter- and intra-personally. In this section, we discuss two approaches to such investigations. The first uses partially scripted experimental confederates interacting with naive subjects, in order to standardize certain lexical or syntactic aspects of critical utterances while allowing most of the interaction to be spontaneous. The second examines interactions where both interlocutors are naive subjects, but synchronizes their spoken utterances with streams of other behaviors. Interactions with partially scripted confederates. To test for early pragmatic effects on referential processing, a visual world study (Hanna and Tanenhaus, 2004) had confederates (the “cook”) instruct subjects (“helpers”) to hand them food and pantry items while the director followed a recipe; crucially, the cook read the steps in the recipe aloud, as a cover for having the instructions scripted, and was trained to add disfluencies during the noncritical instructions to make the utterances sound relatively spontaneous. Helpers knew they would need to hand the cook items from their own side of the table, which the cook could not reach, but that occasionally, when the cook was busy, they would need to get something from the cook’s area located in the middle of the table. Results included that when an ambiguous reference was included in the cook’s instructions (e.g., “Now hand me the cake mix” when there was a chocolate cake mix in the helper’s area but a vanilla cake mix in the cook’s area), helpers immediately looked at the object in their own area if the cook’s hands were empty, but looked at both cake mixes (and in fact, sometimes asked which one was meant) if the cook’s hands were full. In other words, the pragmatics of the task and the reaching ability of the cook were taken into account early in processing by the helpers to construct a domain of possible referents to consider during the creation of understanding. (See related work on the integration of human abilities in Weigand, Chapter 11.) Conversational partners are also capable of keeping close track of the process of grounding meaning; this includes tracking the partner’s mental state and common ground during interruptions and conversational repairs. In another visual world study of referential communication in which subjects answered questions posed by a confederate-experimenter (Brown-Schmidt, 2009), subjects’ gaze showed that they assumed referring expressions to be in common ground when the confederate had responded with “OK” to their previous 102



Psycholinguistic Approaches

answer, but not when the grounding process had been interrupted by the confederate saying “sorry, I didn’t get that.” In another study, after one object within a pair of competitors had already been mentioned, questions that the confederate produced with falling intonation (typical of questions about new information) were interpreted as referring to the object that had not yet been mentioned, whereas questions that used high/rising intonation throughout (as in “what’s? above the bear? that’s wearing a flower?”, similar to “what was that again?”) were interpreted as referring to the one already-mentioned—essentially, as initiating a repair (Brown-Schmidt and Fraundorf, 2015). Other studies that have tracked gaze while subjects interact with a partially scripted confederate have found that common ground established with the confederate-partner influences interpretation early in processing (Brown-Schmidt, 2009; Hanna and Tanenhaus, 2004; Metzing and Brennan, 2003; Matthews, Lieven, and Tomasello, 2010; Nadig and Sedivy, 2002; although see also Barr and Keysar, 2002; Keysar, Lin, and Barr, 2003). When can a confederate fulfill the role of conversational partner without changing the nature of the dialogue itself? The ways in which experimenters deploy confederates depend to a large degree on what they assume about the nature of dialogue. Kuhlen and Brennan (2013) considered the risks and benefits of using confederates as interactive partners, concluding that the rationale for using a confederate to standardize behavior in one of the partners is strongest when hypotheses concern low-frequency linguistic forms and behaviors. Although the use of confederates is risky, they concluded that it is less so when the task requires the confederate to act mainly in a speaking role (because behavior in that role is understood well enough to script convincingly) but more risky in the addressee role (because nonverbal behavior is more implicit, less well understood by experimenters, and therefore usually left unscripted, which is especially problematic when confederates know more than is warranted by a collaborative task) (Kuhlen and Brennan, 2013, p. 54). Interactions in pairs of naive subjects. To the extent that a visuo-spatial task affords measuring streams of behavior that can be synchronized with speech, it can enable documenting with reasonable confidence how interlocutors are coordinating and what they are intending and understanding (as well as misunderstanding). An early precursor to visual worlds gaze studies tracked mouse movements as an incremental measure of the time course by which naive pairs came to jointly understand locations on maps (Brennan, 1990, 2005). Remotely located directors instructed matchers to move a cursor to a target location; directors could either see or not see the matchers’ cursor moving toward the target. When visible, the moving cursor provided directors with a continuous stream of evidence about what matchers understood, speeding mutual understanding of the target location (especially near the end of each trial). Strikingly, even though the matcher always viewed the same kind of information (their cursor, as they moved it over the map), they provided many fewer verbal backchannels when they knew their partner could see their cursor than when their partner could not. This means that the mouse movements were not only instrumental in the task, but also communicative, replacing spoken backchannels and counting as the matcher’s turn in the conversation. The director’s and matcher’s mouse cursors served as representations of individual knowledge that converged non-linearly and nonmonotonically as they grounded meaning. Other action sequences can be both instrumental and communicative; in a study in which a director instructed a partner to build Lego block models, a detailed analysis showed that when the partner’s manipulations of the blocks were visible to the director, grounding was speeded, with the partner pointing at, exhibiting, orienting, poising, and placing the blocks in a way that was precisely timed with utterances (Clark and Krych, 2004). 103

Susan E. Brennan and Joy E. Hanna

Gaze is not only instrumental for doing visuo-spatial tasks and as a measure interpretable by experimenters, but it is also a source of evidence for communication between interacting partners. We documented the time course and flexibility with which a speaker’s gaze could be used by an addressee to disambiguate referents before the point of disambiguation (Hanna and Brennan, 2007). Naive pairs matched colored shapes with markings on them, such that there was a controlled POD (e.g., “the yellow square with a Q”, similar to Eberhard et al., 1995). Directors and matchers sat face-to-face, separated by a barrier so that they saw each other’s faces but not each other’s displays; their duplicate sets of cards were either in a spatially-congruent (mirrored) arrangement on either side of the barrier, or else arranged non-congruently. Directors’ gaze and accompanying head movements cued matchers early as to which temporarily ambiguous shape the director meant to refer to. Even when displays were in reverse orders, matchers still identified targets well before the POD, showing that they rapidly re-mapped and integrated directors’ gaze cues with the unfolding linguistic information in directors’ utterances. A few studies have used gaze as a measure and as a communication channel, but have synchronized these streams from both naive partners, simultaneously. A shared gaze system (Brennan et al., 2008; Neider et al., 2010) had pairs of remotely-located partners both wear a head-mounted eye tracker, the outputs of which were displayed in real time as moving gaze cursors superimposed over the other partner’s screen. Partners could speak freely in some conditions or had no speech channel in others, with or without the ability to monitor each other’s gaze cursor. Pairs were able to coordinate doing a visuo-spatial task using gaze alone. This involved, in one study, dividing the labor to find a target during a collaborative search task (Brennan et al., 2008) and in another, reaching consensus (simultaneous joint visual attention) on another task that involved detecting the location of a shooter in a city scene (adding speech in this task provided a useful alerting function; Neider et al., 2010). The use of locative and deictic expressions was shaped by the modes of evidence available to the partners. A third study (for which data analysis is still underway) tracked pairs of naive partner’s gaze, speech, and mouse cursors simultaneously (Churilov, 2014) as they discussed locations on a map. In that study, visual evidence (whether from mouse or gaze cursors) was more efficient than speech alone, replicating previous studies. Direct comparison of mouse and gaze dynamics showed gaze to be much noisier, but that it led the mouse movements as directors’ and matchers’ understanding of the target location converged. In addition, matchers changed the way they used language, the mouse, and their eyes, depending on the mode of evidence available to directors (replicating Brennan, 1990, 2005). Finally, recurring patterns of behavior and their temporal dynamics at different amounts of lag time have been examined using cross-recurrent gaze analysis, a technique based on nonlinear dynamical systems theory and adapted to the study of language and communication by Richardson, Dale, and colleagues. As two interacting partners coordinate meaning and understanding, their eye gaze over the same elements in a display is coupled, typically within an interval of about two seconds (with the speaker’s gaze leading the addressee’s); in fact, the more closely coupled their gaze, the more successful their communication (e.g., Richardson and Dale, 2005; Richardson et al., 2007; Shockley et al., 2009).

6.4 Conclusion Psycholinguistics is truly at an exciting point. New and unobtrusive measures, innovative tasks and experimental designs, and online computational methodologies are making it increasingly possible to test insights and extend theories about the nature of language. 104



Psycholinguistic Approaches

Testable theories are being harvested from diverse disciplines—including not only psychology and linguistics, but also ethnography and sociolinguistics. Moreover, promising approaches not covered here, such as studies of mediated communication via digital technologies or with increasingly realistic avatars, as well as cognitive and social neuroscience approaches that use brain imaging or electrophysiological measures, are being used to examine the neural systems that are active when people are engaged in interactive tasks (see also Iacoboni, Chapter 16 in this volume). All of these developments present unprecedented opportunities for testing and applying theories of meaning and understanding. We predict that these advances will continue to show that the fundamental nature of meaning and understanding in dialogue is both incremental and collaborative.

Acknowledgement This material is based on work done while Susan E. Brennan was serving at the National Science Foundation. Any opinion, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.

References Akmajian A, Demers RA and Harnish RM (1987) Linguistics: An introduction to language and communication (2nd ed). Cambridge, MA: MIT Press. Allbritton DW, McKoon G and Ratcliff R (1996) Reliability of prosodic cues for resolving syntactic ambiguity. JEP:LMC 22(3): 714–735. Altmann GTM and Kamide Y (1999) Incremental interpretation at verbs: Restricting the domain of subsequent reference. Cognition 73: 247–264. Altmann GTM and Kamide Y (2003) The time-course of prediction in incremental sentence processing: Evidence from anticipatory eye movements. Journal of Memory and Language 49: 133–156. Arnold JE, Hudson-Kam CL and Tanenhaus MK (2007). If you say thee uh—you’re describing something hard: The on-line attribution of disfluency during reference comprehension. JEP:LMC 33: 914–930. Arnold JE, Tanenhaus MK, Altmann R and Fagnano M (2004) The old and thee, uh, new. Psychological Science 15: 578–581. Austin JL (1962) How to Do Things with Words. Oxford: Oxford University Press. Barr DJ and Keysar B (2002). Anchoring comprehension in linguistic precedents. Journal of Memory and Language 46: 391–418. Bavelas JB and Coates L (1992) How do we account for the mindfulness of face-to-face dialogue? Communication Monographs 59: 301–305. Bavelas JB and Chovil N (2000) Visible acts of meaning: An integrated message model of language in face-to-face dialogue. Journal of Language and Social Psychology 19: 163–194. Bock JK (1996) Language production: Methods and methodologies. Psychonomic Bulletin and Review 3: 395–421. Brennan SE (1990) Seeking and providing evidence for mutual understanding. Unpublished doctoral dissertation, Stanford University. Brennan SE (2005) How conversation is shaped by visual and spoken evidence. In Trueswell JC and Tanenhaus MK (eds) Approaches to Studying World-Situated Language Use: Bridging the language-as-product and language-as-action traditions. Cambridge, MA: MIT Press, 95–129. Brennan SE and Clark HH (1996) Conceptual pacts and lexical choice in conversation. JEP:LMC 6: 1482–1493. Brennan SE, Chen X, Dickinson C, Neider M and Zelinsky G (2008) Coordinating cognition: The costs and benefits of shared gaze during collaborative search. Cognition 106: 1465–1477. 105

Susan E. Brennan and Joy E. Hanna

Brennan SE and Schober MF (2001) How listeners compensate for disfluencies in spontaneous speech. Journal of Memory and Language 44: 274–296. Brennan SE and Williams M (1995) The feeling of another’s knowing: Prosody and filled pauses as cues to listeners about the metacognitive states of speakers. Journal of Memory and Language 34: 383–398. Brown-Schmidt S (2009) Partner-specific interpretation of maintained referential precedents during interactive dialog. Journal of Memory and Language 61: 171–190. Brown-Schmidt S and Fraundorf S (2015) Interpretation of informational questions modulated by joint knowledge and intonational contours. Journal of Memory and Language 84: 49–74. Brown-Schmidt S, Gunlogson C and Tanenhaus MK (2008) Addressees distinguish shared from private information when interpreting questions during interactive conversation. Cognition 107: 1122–1134. Chambers CG, Tanenhaus MK and Magnuson JS (2004). Actions and affordances in syntactic ambiguity resolution. JEP:LMC 30: 687–696. Chomsky N (1957) Syntactic Structures. Gravenhange, Holland: Mouton. Churilov E (2014) Speaking, pointing, and gazing: Different communication strategies for different communicative constraints. Unpublished master’s thesis, Stony Brook University. Clark HH (1992) Arenas of Language Use. Chicago: University of Chicago Press. Clark HH (1994) Managing problems in speaking. Speech Communication 15: 243–250. Clark HH (1996) Using Language. New York: Cambridge University Press. Clark HH and Brennan SE (1991) Grounding in communication. In Resnick LB, Levine J and Teasley SD (eds) Perspectives on Socially Shared Cognition. Washington, DC: APA, 127–149. Clark HH and Krych MA (2004) Speaking while monitoring addressees for understanding. Journal of Memory and Language 50: 62–81. Clark HH and Wilkes-Gibbs D (1986) Referring as a collaborative process. Cognition 22: 1–39. Cooper RM (1974) The control of eye fixation by the meaning of spoken language. Cognitive Psychology 6: 84–107. Dahan D, Tanenhaus MK and Chambers CG (2002) Accent and reference resolution in spoken language comprehension. Journal of Memory and Language 47: 292–314. Dahan D and Tanenhaus MK (2004) Continuous mapping from sound to meaning in spoken-language comprehension: Immediate effects of verb-based thematic constraints. JEP:LMC 30: 498–513. Eberhard KM, Spivey-Knowlton MJ, Sedivy JC and Tanenhaus MK (1995) Eye-movements as a window into spoken language comprehension in natural contexts. Journal of Psycholinguistic Research 24: 409–436. Ferreira VS and Dell GS (2000) The effect of ambiguity and lexical availability on syntactic and lexical production. Cognitive Psychology 40: 296–340. Ferreira F and Swets B (2005) The production and comprehension of resumptive pronouns in relative clause “island” contexts. In Cutler A (ed.) Twenty-first Century Psycholinguistics: Four Cornerstones. Mahway, NJ: Erlbaum, 263–278. Firbas, J (1992) Functional Sentence Perspective in Written and Spoken Communication. New York: Cambridge University Press. Fowler CA and Housum J (1987) Talkers signaling “new” and “old” words in speech and listeners’ perception and use of the distinction. Journal of Memory and Language 26: 489–504. Fox Tree JE (1995) The effects of false starts and repetitions on the processing of subsequent words in spontaneous speech. Journal of Memory and Language 34: 709–738. Fussell SR and Krauss RM (1992) Coordination of knowledge in communication: Effects of speakers’ assumptions about what others know. Journal of Personality and Social Psychology 62: 378–391. Galantucci B and Sebanz N (2009) Joint action: Current perspectives. TopiCS 1: 255–259. Garrod S and Anderson A (1987) Saying what you mean in dialog: A study in conceptual and semantic co-ordination. Cognition 27: 181–218. Garrod S and Pickering MJ (2004) Why is conversation so easy? Trends in Cognitive Sciences 8: 8–11. 106



Psycholinguistic Approaches

Glucksberg S, Krauss R and Weisberg R (1966) Referential communication in nursery school children: Method and some preliminary findings. Journal of Experimental Child Psychology 3: 333–342. Goodwin C (1981) Conversational Organization: Interaction Between Speakers and Hearers. New York: Academic Press. Gorman KS, Gegg-Harrison W, Marsh CR and Tanenhaus MK (2013) What’s learned together stays together: Speakers’ choice of referring expression reflects shared experience. JEP:LMC 39(3): 843–853. Grice HP (1957) Meaning. Philosophical Review 66: 377–388. Griffin ZM (2001) Gaze durations during speech reflect word selection and phonological encoding. Cognition 82: B1–B14. Griffin ZM and Bock K (2000) What the eyes say about speaking. Psychological Science 11: 274–279. Hanna JE and Brennan SE (2007) Speakers’ eye gaze disambiguates referring expressions early during face-to-face conversation. Journal of Memory and Language 57: 596–615. Hanna JE and Tanenhaus MK (2004) Pragmatic effects on reference resolution in a collaborative task: Evidence from eye movements. Cognitive Science 28: 105–115. Haywood SL, Pickering MJ and Branigan HP (2005) Do speakers avoid ambiguity during dialogue? Psychological Science 16: 362–366. Henderson JM and Ferreira F (2004) The Interface of Language, Vision, and Action: Eye Movements and the Visual World. New York: Psychology Press. Horton WS and Gerrig RJ (2002) Speaker’s experiences and audience design: Knowing when and knowing how to adjust utterances to addressees. Journal of Memory and Language 47: 589–606. Horton WS and Gerrig RJ (2005) Conversational common ground and memory processes in language production. Discourse Processes 40: 1–35. Jefferson G (1974). Error correction as an interactional resource. Language in Society 2: 181–199. Keysar B, Lin S and Barr DJ (2003) Limits on theory of mind use in adults. Cognition 89: 25–41. Kim CS, Gunlogson C, Tanenhaus MK and Runner JT (2015) Context-driven expectations about focus alternatives. Cognition 139: 28–49. Kraljic T and Brennan SE (2005) Prosodic disambiguation of syntactic structure: For the speaker or for the addressee? Cognitive Psychology 50: 194–231. Krauss RM (1987) The role of the listener: Addressee influences on message formulation. Journal of Language and Social Psychology 6: 81–97. Krauss RM and Chiu C-Y (1997) Language and social behavior. In Gilbert D, Fiske S and Lindsey G (eds), Handbook of Social Psychology Vol. 2. Boston: McGraw-Hill, 41–88. Krauss R and Weinheimer S (1964) Changes in reference phrases as a function of frequency of usage in social interaction. Psychonomic Science 1: 113–114. Kuhlen AK and Brennan SE (2013) Language in dialogue: When confederates might be hazardous to your data. Psychonomic Bulletin and Review 20: 54–72. Lieberman P (1963) Some effects of context on the intelligibility of hearing and deaf children’s speech. Language and Speech 24: 255–264. Matthews D, Lieven E and Tomasello M (2010) What’s in a manner of speaking? Children’s sensitivity to partner-specific referential precedents. Developmental Psychology 46: 749–760. McAllister J, Potts A, Mason K and Marchant G (1994) Word duration in monologue and dialogue speech. Language and Speech 37: 393–405. Metzing C and Brennan SE (2003) When conceptual pacts are broken: Partner-specific effects on the comprehension of referring expressions. Journal of Memory and Language 49: 201–213. Meyer AS, Sleiderink AM and Levelt WJM (1998) Viewing and naming objects: Eye movements during noun phrase production. Cognition 66: B25–B33. Miller G (1962) Some psychological studies of grammar. American Psychologist 17: 748–762. Nadig AS and Sedivy JS (2002) Evidence of perspective-taking constraints in children’s online reference resolution. Psychological Science 13: 329–336. Neider MB, Chen X, Dickinson CA, Brennan SE and Zelinsky GJ (2010) Coordinating spatial referencing using shared gaze. Psychonomic Bulletin and Review 17: 718–724. 107

Susan E. Brennan and Joy E. Hanna

Price PJ, Ostendorf M, Shattuck-Hufnagel S and Fong C (1991) The use of prosody in syntactic disambiguation. Journal of the Acoustical Society of America 90(6): 2956–2970. Reddy MJ (1979) The conduit metaphor—A case of frame conflict in our language about language. In Ortony A (ed.), Metaphor and Thought. Cambridge: Cambridge University Press, 284–297. Richardson DC and Dale R (2005) Looking to understand: The coupling between speakers’ and listeners’ eye movements and its relationship to discourse comprehension. Cognitive Science 29: 1045–1060. Richardson DC, Dale R and Kirkham NZ (2007) The art of conversation is coordination: Common ground and the coupling of eye movements during dialogue. Psychological Science 18: 407–413. Schafer AJ, Speer SR, Warren P and White SD (2000) Intonational disambiguation in sentence production and comprehension. Journal of Psycholinguistic Research 29: 169–182. Schegloff EA and Sacks H (1973) Opening up closing. Semiotica 8: 289–327. Schober MF and Clark HH (1989) Understanding by addressees and overhearers. Cognitive Psychology 21: 211–232. Searle JR (1969) Speech Acts: An Essay in the Philosophy of Language. Cambridge: Cambridge University Press. Sedivy JC, Tanenhaus MK, Chambers CG and Carlson GN (1999) Achieving incremental semantic interpretation through contextual representation. Cognition 71: 109–147. Shockley K, Richardson D and Dale R (2009) Conversation and coordinative structures. TopiCS 1: 305–319. Smith VL and Clark HH (1993) On the course of answering questions. Journal of Memory and Language 32: 25–38. Spivey MJ, Tanenhaus MK, Eberhard KM and Sedivy JC (2002) Eye movements and spoken language comprehension: Effects of visual context on syntactic ambiguity resolution. Cognitive Psychology 45: 447–481. Stellmann P and Brennan SE (1993) Flexible perspective-setting in conversation. In Abstracts of the Psychonomic Society, 34th Annual Meeting, Washington, DC: 20. Swerts M and Krahmer E (2005) Audiovisual prosody and feeling of knowing. Journal of Memory and Language, 53: 81–94. Swinney DA, Onifer W, Prather P and Hirshkowitz M (1978) Semantic facilitation across sensory modalities in the processing of individual words and sentences. Memory and Cognition 7: 159–165. Tanenhaus MK and Brown-Schmidt S (2008) Language processing in the natural world. Philosophical Transactions of the Royal Society B 363: 1105–1122. Tanenhaus MK, Spivey-Knowlton MJ, Eberhard KM and Sedivy JC (1995) Integration of visual and linguistic information during spoken language comprehension. Science 268: 1632–1634. Tolins J and Fox Tree JE (2016) Overhearers use addressee backchannels in dialog comprehension. Cognitive Science 40(6): 1412–1434. Trueswell JC and Tanenhaus MK (eds) (2005) Approaches to World-Situated Language Use: Bridging the Language-as-Product and Language-as-Action Traditions. Cambridge, MA: MIT Press. Velichkovsky BM (1995) Communicating attention: Gaze position transfer in cooperative problem solving. Pragmatics and Cognition 3: 199–222. Watson DG, Tanenhaus MK and Gunlogson, CA (2008) Interpreting pitch accents in online comprehension: H∗ vs. L+H∗. Cognitive Science 32: 1232–1244. Wilson MP and Garnsey SM (2009) Making simple sentences hard: Verb bias effects in simple direct object sentences. Journal of Memory and Language 60: 368–392.

108

7 Intersubjectivity in Dialogue Per Linell

Intersubjectivity has […] to be taken for granted in order to be achieved. (Rommetveit, 1974: 56)

7.1 Introduction This chapter discusses the concept of intersubjectivity and its relation to dialogue. I will argue that intersubjectivity is hardly a monolithic notion, but rather a family of related concepts. I shall elaborate on various forms and aspects of intersubjectivity in both wider philosophical contexts and more local empirical contexts.

7.2  Intersubjectivity and Alterity Intersubjectivity can be defined in many ways. For example, Gillespie and Cornish (2009) provide at least six definitions. In this context, we can start by defining intersubjectivity as the relations in terms of similarities and differences within a constellation (e.g. in a communication situation) or a group/community of people, with regard to their knowledge, opinions, interests, feelings and preferences. Intersubjectivity is often thought of in terms of agreement, commonality and sharedness, but such general assumptions will turn out to be inadequate, except for quite special circumstances. As we will discover soon, intersubjectivities are usually partial, and penetrated by discrepancies and sometimes important differences between the participants involved. Nonetheless, preliminary accounts of intersubjectivity often home in on shared understandings among participants in the situation or community involved. Many scholars have taken intersubjectivity as a necessary assumption or a definitional point-of-departure for successful communication. For example, Clark (1996) has talked about “common ground” and Schegloff (1991) about “shared social cognition” as the prerequisites and goals of talkin-interaction. One type of common ground would be general assumptions about the world and proper interpersonal conduct, what one may call “common sense”. In addition, interlocutors must surely enter the conversation with something in common with regard to the topics raised, and their interaction will normally serve to increase the range of sharedness. 109

Per Linell

Accordingly, Clark talks about the successive establishment of “common ground” as ‘grounding’. Yet, the assumptions of commonality and sharedness must be duly qualified. Several scholars have pointed out that there is almost never any completely shared understandings, simply because people have divergent biographies and sometimes disparate interests. Garfinkel (1967, especially chapter 1) insists that it is meaningful only to talk about “sufficient” understandings for current “practical purposes”. There are no fixated interpretations of situated messages beforehand. Instead, participants need to add some pragmatic specifications ad hoc (Garfinkel, 1967: 21f). Taylor (1992), in his book Mutual Misunderstanding, developed a detailed argumentation for an attitude of profound scepticism towards any assumption of common perspectives or shared understanding among people communicating with each other, and this holds across situation types and different media such as face-to-face interaction and written communication. A language is not a code of fixed signs with static links between expression and content. Rommetveit (1974) too, while arguing that intersubjectivity is a basic notion in dialogue theories and that it “must be taken for granted in order to be achieved”, underscores that intersubjectivities are confined to occasional, only temporarily sustained and partially shared social worlds (Rommetveit, 1974: 29). Bakhtin (1986) proposes that dialogue builds upon “alterity”, that is, the “other” arrives with, or invokes, partly different (“alien”) perspectives, knowledge, experiences, interests and feelings other than those that the speaker entertains beforehand. The speaker can of course (partly) incorporate some of these perspectives, and the parties can get closer to one another’s understandings. However, the reverse can also happen. Parties to a dialogue normally do not share assumptions and understandings in anything like a complete manner; rather, in successful communication, there must be a substantial, or occasionally nearly “perfect” (Rommetveit, 1974: 29), complementarity between perspectives; people share framings of the situation they are involved in, but they bring different (“complementary”) knowledge and understandings to bear on the exchange, so to speak trading on each other’s truths. The addressee’s perspective is partly exhibited in her verbal or nonverbal responses; these represent the “visible cognitive life of the hearer” (Goodwin, 2017: 00). According to Goodwin, we should think of intersubjectivity as a “contingent achievement of relevant intersubjectivity” (2017: 00). Goodwin argues that intersubjectivity is (a) culturally and locally co-determined (rather than given), (b) contingent on dynamic conditions (that may change from moment to moment), (c) achieved (rather than postulated or assumed), and (d) relevant for current purposes only. To this we may add Rommetveit’s point that it is usually only partially shared among participants.

7.3  Alterity, Polyphony, Heteroglossia Intersubjectivity is not automatic, inevitable or complete (except perhaps in some forms of primary intersubjectivity; see Section 7.5). The other, whoever (s)he is, is not quite like oneself. Hence, as said, dialogical theories are very much about alterity, the role of the other as being different from self. Relatedly, communication is not always about striving for mutual understanding or consensus; parties may have different interests, knowledge, intentions, or limited willingness or practical opportunities to disclose all their ideas, thoughts, intentions, feelings, and so forth. Sharedness, commonality, reciprocity and mutuality are therefore partial. Instead of complete understandings, we aim for understandings, and intersubjectivities, that are sufficient for current practical purposes (Garfinkel, 1967). We can therefore say that intersubjectivities in the social world involve both partially shared premises and some mutually strange ideas. This is the nature of being with one 110



Intersubjectivity in Dialogue

another (what Heidegger called Mitsein). Others are always different from oneself in some respects, and they always come with at least partly different perspectives. As a speaker, self can often anticipate reactions of the other; objections, evaluations and points-of-view of the other, and thus; often he become aware of such differences in and through the interaction. This is a cornerstone of classical Bakhtinian dialogism. Alterity provides a dialogical counterpoint to dominant monological theories of commonality and sharedness, in which the essence of communication resides in making knowledge and attitudes entirely common (Latin communicare literally means ‘making things common’). In addition to alterity, Bakhtin also assumes that there can occur more profound differences, which are hard or even impossible to overcome: “outsideness” (Russian vnenakhodimost; Bakhtin, 1986; Emerson, 1997: 207ff). Alterity is related to at least two other foundational ideas in Bakhtin’s dialogism: heteroglossia of linguistic communities, and polyphony of discourse. Specific dialogues, discourses and texts contain several “voices”. The concept of voice is here intended to invoke several aspects, including, in particular, personal signature and perspective on topic (Bertau, 2007). However, a single utterance or text produced by a single person is also often “multi-voiced”, since the author (speaker, writer) cannot help being influenced by others (in accepting or countering their contributions), and (s)he actively includes explicit or implicit ‘quotes’ (in a broad sense) from others. Accordingly, a person’s thinking can also be regarded as an ‘internal dialogue’ which accommodates others’ ideas or prior utterances (there may be explicit or implicit instances of “reported speech”; Voloshinov, 1973). Thus, the dialogical self houses many partly different voices defining what Hermans and Hermans-Konopka (2010) call different “I-positions”. Such a conception may be compatible with a ‘postmodern’ split self, but dialogism might prefer a notion of a contextually sensitive self with some coherence over time. If “polyphony” (or “multivoicedness”) primarily applies to specific individuals and their discourses, there is a corresponding mixture of perspectives and tensions between viewpoints at the collective level, in or between groups, professions, communities and societies. This phenomenon is usually referred to as heteroglossia. We should realise that there is a positive value in alterity and heteroglossia, in the lack of complete intersubjectivity. Without differences, there would often be little point in communicating (e.g. Linell and Luckmann, 1991). Asymmetries of knowledge are a driving force in social interaction (e.g. Heritage, 2012), and we can learn from others and outsiders.

7.4  Intersubjectivity and Trust The points of alterity and polyphony might seem to be difficult to reconcile with Rommetveit’s above-mentioned dictum: “Intersubjectivity has to be taken for granted in order to be achieved”, which seems to build upon striving towards sharedness. Intersubjectivity is depicted as both premise and outcome of dialogue and interaction. But how can something that has not yet been achieved be, at the same time, taken for granted at the very outset? What is at stake here is that participants need some common premises in order to enter into communication (including sufficient trust in each other, see below), and that this is necessary for intersubjectivity at more concrete levels to be developed in and through the interaction. A reasonable interpretation of Rommetveit’s proposal would involve the notion of trust which penetrates all human relations. Trust and its opposite, distrust, come to life—and are made crucially relevant—when we engage in actual interaction. Trust basically means that, 111

Per Linell

despite the fact that we can never be certain about most circumstances in the world (including those having to do with the other’s intentions, knowledge and competence), we have to act on the premises that the world will turn out to remain normal and that we can rely on the likelihood that others to not do us any harm (of course, only as long as we do not discover evidence to the contrary). That is, we must trust that there is some common ground to begin with, if we want to achieve more reflected types of intersubjectivity, for example, through negotiations or through bringing something into language. As Linell and Marková (2014a, 2014b, and references there) argue, this shows that there is a close relationship of interdependency between intersubjectivity and trust: we need to take a lot for granted, take it as shared (i.e. as, by and large, intersubjectively valid), even though most of it we cannot know for sure. Trust and distrust are therefore phenomena closely linked to dialogicality; they are at the same time elusive and ubiquitous in human existence. In the subsequent section, it will be pointed out that there are several levels of intersubjectivity, which differ in the degree of awareness and other dimensions. One distinction, with intermediate degrees, concerns taken-for-granted (unreflected) intersubjectivity versus reflected intersubjectivity.

7.5  Forms of Intersubjectivity: A Developmental Theory Intersubjectivity takes many forms, which can partly be correlated with different stages of development. Trevarthen (1979, 1998) distinguishes between at least three forms: primary, secondary, and tertiary intersubjectivity. At first, we find primary (immediate, direct, unreflected, authentic, “genuine”) intersubjectivity between an infant and a carer in an ‘I–thou’ relationship. Here, it would not make sense to talk about shared (cognitive) understanding of things; rather, we have prelogical and pre-propositional interactions (Duranti, 2010: 8). This supports the contention that interactivities are more basic than intersubjectivities (see Section 7.9). Primary intersubjectivity involves little of individual subjectivity on the part of the infant. But the infant–carer interactivities build on several dialogical prerequisites: (a) the infant has a biologically induced disposition for interaction and mirroring, (b) the infant arrives in a world that is already inhabited by and meaningful for the adult partners, and (c) these carers usually treat infants and children as if they understand more than they actually do.

The first levels of intersubjectivity are further theorised by Zlatev (2013). With the advent of shared attention to external objects (a third entity intervening in the interaction: the ‘object’), there will be (what Trevarthen calls) secondary intersubjectivity (‘I–thou–it’). Still later, there is tertiary intersubjectivity, when participants start to address peripheral others (so-called third parties). The primary participants in the communicative encounters might then work with a “split” attention, with a “sideward glance” (to speak with Bakhtin) to third parties and remote others; not only ‘I, thou/you, and it’ but also ‘we/they’, i.e., “bystanders” and “generalised others”. Once we have interactional language at our disposal, we can have subjectivities enhanced by thinking and supported by written language (which is of course indirectly intersubjectively based). Intersubjectivities will now involve many more forms of partially shared knowledge of the world, and in addition to addressivity and responsivity, also other kinds of contextuality and feelings for genre differences. In the end, 112



Intersubjectivity in Dialogue

the tertiary stage includes virtually all available cultural forms, civilisations and literacies of late modernity. The third (tertiary) level—largely also Bråten’s (2007) notion—is complex with many dimensions; indeed, it might be better to talk about several “higher” levels (rather than just one “third” level), provided that “higher” need (must) not be taken to mean “better than the first or second levels”. Intersubjectivity in communicative projects (episodes) or cognitive events (Steffensen et al., 2010: 215) involves internal dialogue (see Section 7.8). Vygotsky (1978) describes internal dialogue as a further and later development of external interaction. He also talks about forms of intersubjectivity in different stages and at different levels of (ontogenetic, socio-historical, micro-interactional) development/genesis. In this context, one might mention our abilities to talk about imaginary objects and happenings, to joke, to cheat and lie, and to use language for various literary purposes. These activities obviously presuppose forms of intersubjectivity, and often literacy. Despite the fact that forms of intersubjectivity emerge at different developmental stages, they can sometimes live side by side in adult life. Kinds of primary intersubjectivity can be accessed also later in life. Adults can in exceptional cases establish dyadic high-intensity embodied contact. Hodges (2011) refers to exceptional instances of conversational contact involving ‘flow’ or ‘lostness’ (“lost in the moment”): “a fabulous conversation over a great meal with friends illustrates such unselfconscious disclosure. Even the burden of time seems absent. The experience is even like “a taste of heaven”, says Hodges (2011: 152). Such situations can also be exceptionally close and mutually attuning. Another case might be certain intimate erotic situations, with lovers being blind to the outside world. Or two dancers or several people in musicking totally engrossed in their activities. Primary intersubjectivity is usually considered to be more “authentic”, yet it is culturally penetrated too. Intersubjectivities can be disturbed in more fundamental, and abnormal ways. We can look at aberrations like ADHD and autism as disturbances of dialogicality. Another aberration is extreme egoism and Machiavellianism, which are forms of very monologising attitudes.

7.6  The “Problem of Other Minds” Intersubjectivities presuppose that the human mind is social (Valsiner and van der Veer, 2000), that is, that human beings are a social species, interacting with others for substantial parts of their time awake and learning to understand the world largely by observing and talking with other persons. An opposite to this assumption can be found in the theory of mind named itself “Theory of Mind” (ToM) (sic!; see e.g. Leslie, 2000), which is based on the assumption that individuals have great problems finding out what others think, feel, intend, want, mean, and so forth. It does not start out from assumptions of intersubjectivity in the human mind (Zlatev et al., 2008a); instead, for ToM theorists it is a (great) problem how individual minds can understand other minds (“the problem of other minds”). The extreme interpretation of such a position is solipsism. In other words, ToM is an exceptionally clear example of an individualist-monologist theory of a phenomenon (mutual social perception and understanding, primarily through interaction) that is in actual fact eminently “dialogical” in nature. The alternative dialogist theory, or an “embodied ‘Interaction Theory’ of social cognition” (Gallagher and Hutto, 2008: 17), would see the understanding of others’ minds as more or less “direct” and an outgrowth of interactional, embodied practices—practices that are from the innate start “emotional, sensory-motor, perceptual and non-conceptual” (Gallagher and Hutto, 2008: 20). Gallagher (2011) notes that embodiment, in its full sense, is linked to interaction and intersubjectivity. 113

Per Linell

ToM has been criticised in fundamental ways by many dialogically-minded scholars; see in particular Gallagher and Hutto (2008) and other articles in Zlatev et al. (2008a), Leudar and Costall (2009), De Jaeger et al. (2010), and references there. Some points can be summarised as follows. ToM constitutes an attempt on the part of cognitivists to tackle social interaction and others’ cognitions. Cognitivism is a theoretical approach within cognitive science which, briefly put, assumes the following: (a) cognition consists in information processing; representation and modelling are therefore central concepts; (b) cognition takes place in the brain (it is ‘intracranial’); the mind is ultimately assumed to be entirely reducible to brain functioning; (c) cognition is an individual (rather than social) activity; there are no collective minds that could process information.

ToM appears in two primary forms (Gallagher and Hutto, 2008: 18). The main variant assumes that understanding the other presupposes a theory of the other: “theory theory of mind” (TToM). Since we cannot directly perceive or understand what the other is up to, one’s self (or rather: mind/brain) has to make indirectly based inferences about what the other may have in mind. If, according to TToM, a person has a ToM, this is localised in a particular module of the brain. TToM has therefore strong similarities with Chomskyan modular theories of language (Language Acquisition Device, Universal Grammar). The other interpretation of ToM is the “simulation theory”. This does not involve a “theory” of the other; rather, one somehow simulates, or empathises with, the other’s situation, doings and thoughts. When I discuss ToM below, I deal primarily with TToM. ToM treats the mind of the other as something completely concealed, inside the skull, something about which one must develop “theories”—but normal social interaction supplies rich information about the other’s thoughts, reactions, intentions, emotions, doings, and this applies a fortiori to interaction directed to and co-constituted with infants and children. In most social interactions, the mind is thus partly on display; in other words, we live in a common life-world (e.g. Zlatev et al., 2008b: 3) and can “directly” observe a lot of “mental states” (Sharrock and Coulter, 2009: 66). Meanings, trust/distrust and moral stances are “displayed, and made visible and ascribable on the basis of actors’ actions and discourse” (Jayyusi, 1991: 243), thus, participation is more basic than intention. Others’ intentions leak out, that is, intentional behaviour can be observed by second or third persons. Of course, people (except young children) can lie, betray others, have secrets and fake trust and intentions, but this is not the default case, especially not with children. In addition, as already indicated, we talk about thoughts, feelings, intentions, and so forth (see Taylor, 2013, on these metadiscursive practices), often in the presence of observable evidence. Schegloff (1991) comments on “the socially shared world” and the role of interaction: “our understanding of the world and of one another is posed as a problem and resolved as an achievement, in an inescapably social and interactional context” (1991: 168; compare Trognon and Batt, 2010: 18). ToM has dealt mainly with children’s developing abilities to find out about other minds. Much research has focused on experiments inducing “false beliefs” in young children. It is beyond the scope of this chapter to discuss these experiments (see contributions to Leudar and Costall, 2009). However, that the mind is concealed, and especially intentionally 114



Intersubjectivity in Dialogue

concealed, from others in the real social world mainly applies to specific situations, usually involving adults rather than small children: (a) somebody is engrossed in solo thinking or silent reading or listening, situations in which little of the content of one’s thoughts is disclosed (or can be disclosed, for practical or cultural reasons); (b) somebody is involved in wilfully trying to conceal something from the other, cheating or lying; hence, wilful or intended deception is the problem, not normal social interaction.

It seems that ToM takes, in effect, cases like (a) and (b) to be basic in human existence, rather than (as we would argue) exceptional. Thus, contrary to ToM suppositions, the issue is more about concealing (and gradually becoming accustomed to concealing) what we know, believe, feel, and so forth; it is not primarily, as ToM argues, about unobservable mental states. This is of course not to deny that there are aspects of thinking and feeling that are largely inaccessible to others. Some are in fact inaccessible to ourselves as well, pace Freud and others. Intentions are derived and ascribed on the basis of direct observation of others, and on talk about actions. Monologism, for example, ToM, puts things upside down, deriving meaning from intentions.

7.7  Intersubjectivity as a Property of Interpersonal Interaction Interactivities can be regarded as the situated occasions in which persons work for and demonstrate their intersubjectivity. Rommetveit (1974) talked about “the architecture of intersubjectivity” with reference particularly to interactions, especially in and through talk. Any study of actual talk-in-interaction will demonstrate that there are plenty of reflections of self–other interdependencies (Linell, 2017). Here I will discuss such aspects of manifest sociodialogues under two headings: Global Units, and Local Reflections.

7.7.1  Self–Other Interdependencies I: Global Units When people enter a potential communicative exchange, they have to make assumptions about what situation type and communicative active type it represents, and these assumptions must be reasonably congruent. Situations when people meet face-to-face, what we might call ‘social occasions’, can be categorised. Goffman (1971: 13) gives examples like “weddings, family meals, chaired meetings, forced marches, service encounters, queues, crowds, and couples”. People don’t approach situations as if they are entirely novel. Instead, they have expectations of “what might happen here”. Goffman (1974: 8) explains: I assume that when individuals attend to any current situation, they face the question: “What is it that’s going on here?” Whether asked explicitly, as in times of confusion and doubt, or tacitly, during occasions of usual certitude, the question is put and the answer to it is presumed by the way the individuals then proceed to get on with the affairs at hand. Goffman goes on to introduce concepts like ‘frames’ and ‘situation definitions’: [M]uch use will be made of Bateson’s use of the term “frame”. I assume that definitions of a situation are built up in accordance with principles of organization which govern 115

Per Linell

events—at least social ones—and our subjective involvement in them; frame is the word I use to refer to such of these basic elements as I am able to identify. (Goffman, 1974: 10–11, italics added) Scheff (2006: 76) comments that “[t]he definition of the situation is the actors’ largest subjective response; frames are part of this subjective structure”. In this context we might prefer the attribute “intersubjective”. Other terms for these structures are activity type (Levinson, 1992) or communicative activity type (Linell, 2010). Levinson (1992: 69) mentions such examples as “teaching, a job interview, a jural interrogation, a football game, a task in a workshop, a dinner party, and so on”. John Gumperz, pioneer in interactional sociolinguistics, talks in the same vein: My basic assumption is that all understanding is framed understanding, that it ultimately rests on contingent inferences made with respect to presuppositions concerning the nature of the situation, what is to be accomplished and how to be accomplished. The term activity can be seen as a cover term to suggest what these presuppositions are. (Gumperz, 1992: 43; italics added)

Knowledge of the conversational activity entails expectations about possible goals or outcomes for the interaction, about what information is salient and how it is likely to be signalled, about relevant aspects of interpersonal relations, and about what will count as normal behavior. (Gumperz, 1982: 101) Accordingly, the gestalt properties of situation types provide a good deal of intersubjectivity to spoken interaction. They are the glue that binds situations together. Scholars have provided a variety of terms for these phenomena: for example, speech genres (Bakhtin, 1986), communicative genres (Luckmann, 1986), joint activities (Clark, 1996), activity types (Gumperz, 1982; Allwood, 2000; Sarangi, 2000) and dialogue games (Weigand, 2010).

7.7.2  Self–Other Interdependencies II: Local Reflections Bakhtin (1981: 293) suggests that “the word in language is half someone else’s”. This thesis refers both to speakers’ dependences on their prior communicative biographies, and to their relations within specific situated interactions. Irrespective of whether quotes from others’ contributions are due to unreflected routine or are carefully intended, they show that participants are dependent on one another. When we borrow words from our interlocutors, we make them co-authors of our own utterances. A single example from Du Bois (2009: 11), from a therapy session, will have to suffice in this context (see also updated version in Du Bois 2014): (1) (Deadly Diseases SBC015)1 1. Lenore: so your mother´s happy now. 2. (0.2) 3. Joanne: .hh my mother´s never happy. 4. [my mother wouldn´t be happy if] everything was g116



Intersubjectivity in Dialogue

5. Lenore: [excuse me hhh ] 6. Joanne: hrmm everything was great,and everything is 7. great. In line 1, Joanne and the therapist Lenore use the same words (‘mother’, ‘happy’; possessive pronouns: ‘your/my’) inserted in a predicative construction (with an inflected form of ‘be’) three times in lines 1, 3 and 4. In lines 6 and 7 (and in the interrupted line 4) Joanne repeats her own words ‘everything’ and ‘great’. Du Bois calls such lexical (and other) similarities across contributions ‘affinity’ between adjacent utterances. Such inter-turn links may be used to emphasise agreement and consensus, or difference and competition, depending on co-texts and prosodies, among other things. For example, in line 1 Joanne completely opposes Lenore’s suggestion (line 1), yet recycles the same words. In terms of M. Goodwin’s (1990: 177ff.) related notion of ‘format tying’, the partial repetition of others’ utterances is used mainly for competitive purposes (Goodwin is particularly concerned with ritual insults among young Black girls). Thus, interactions, whether entirely friendly or slightly competitive, turn out to be coactions and joint projects (Clark, 1996). A question like “Who came up with that idea?”, intended to refer to a particular topic, can often not be given an individual-based answer. A sociodialogue is not a series of self-sufficient actions by mutually independent, autonomous individuals; rather, it is a sequence of several persons’ contributions to a more or less coherent joint project. The mutual accommodations between interlocutors do not only concern the choice of words. They can involve the choice of language (when several languages are at the disposal of participants), and the attenuation (or sometimes accentuation) of dialectal differences, as was demonstrated in speech accommodation theory (Giles and Smith, 1979). Under friendly conditions, these may be expressions of mutual alignments, and even consensus and agreement. At a more detailed syntactic level, we find a lot of structural dependencies that help both speaker and listener in keeping track of an emergent utterance. In uttering some initial words the speaker projects more to come. The choice of a particular construction will anticipate what sorts of things may be upcoming (Auer, 2005). Syntactic and pragmatic projections may lead to units being produced collaboratively by two (or even more) participants. Let us look at an example of such a ‘collaborative completion’ (from Lerner, 2002: 227): (2) 1. Rich: if they come en´ pick it up 2. it´ll co[st yah 3. Mike: [they charge yuh Here, Rich, starting on an ‘if–then’ construction in line 1, comments on what will happen if a customer commissions a firm to transport something away. Lerner gives no additional context, but even the example stripped of this will allow us as readers (and the participants in situ) to expect an apodosis segment to follow the protasis part issued in line 1. (An explicit ‘then’ item is omitted from the beginning of line 2.) But Rich begins on an apodosis in line 2, but breaks off as his interlocutor contributes an alternative completion (line 3). In this example, the projection is both syntactic (the ‘if–then’ construction is bipartite, consisting of an initial protasis (condition) followed by an apodosis (consequent)) and pragmasemantic (“it’ll cost you” and “they charge you” meaning nearly the same). 117

Per Linell

In other cases, the completion could be partly competitive. In the following example from Szczepek (2000: 8): “Ken and Joe are telling a third person about a doctor whose diagnosing practices they both strongly disagree with. It is Ken who is directly affected by them as he suffers from a yet undiagnosed disease. Both participants have mutual knowledge of the incident discussed.” (3) (Transcription slightly modified in relation to Szczepek, 2000: 8) 1. Ken: he just di- did a blood test.= 2. =and said yeah well your b- your blood´s all shot;= 3. =and you have the liver of a ninety year old, .hh 4. uhm – 5. [and I w- AND I AND I THINK— 6. Joe: [DO YOU DRINK? 7. AND HE DOESN´T [DRINK. 8. Ken: [AND I THINK— 9. and I think uh::-10. you you picked up some uhm [(.) 11. Joe: [°°virus.°° 12. Ken: virus, Ken reports what the doctor said and begins listing the results from his blood test using list intonation (lines 2 and 3), but seems to have problems with retrieving the right words for the third list item (lines 4 and 5 and 8 and 9), which makes it possible for Joe to intervene (lines 6 and 7). However, Joe’s contribution does not seem to be the one Ken wants; unlike in most instances of other-completion there are no signs of concurring. Instead, Ken continues his word search from line 5 in lines 8 and 9, and then starts on an alternative of his own (line 10; note that “you” here refers to Ken, as allegedly talked about by the doctor). In this second phase (lines 8 to 12), Joe also makes a completion (line 11), which is taken up as cooperative and is affirmed by Ken (line 12). Thus, Joe’s ‘deviant’ (competitive) list item (line 6) becomes an isolated piece (the irrelevance of which is stated by Joe himself in line 7). The episode as a whole is clearly cooperative, and this holds for Szczepek’s data in general.2 There are many dependencies in interactional languaging that may generate syntactic projections, not only bipartite constructions (as in ‘if–then’; Günthner, 2011) and lists. Another kind of example is the peculiar Swedish construction initiated by two initial finite modal verbs, as in the following authentic example3: (4) Two persons A and B are looking at a rather rough coast-line by the sea: 1. A: kan man bada här? can you swim here? 2. B: kan kan man väl men vill vill man väl inte  you can (lit.: can can you) probably but you probably don’t want to (lit.: want want you not) The Swedish construction DIFA (double initial finite auxiliaries; Linell and Mertzlufft, 2014) is a specific responsive construction in the sense that it can only occur in a response to a prior utterance that contains a particular ingredient (‘source’). In this case, the source is a modal verb (here: kan ‘can’), typically occurring in an interrogative utterance (see line 1), 118



Intersubjectivity in Dialogue

and DIFA (line 2) starts by duplicated copies of this modal verb. The first copy appears in the front (so-called ‘fundament’ position of assertive main clauses in Swedish grammar), and the second copy follows suit, occupying the second-constituent position, that of the finite verb. The pragmatic function of DIFA is to admit that while the condition for the specific action asked about (“can you do it?”) is fulfilled, this condition (here: of possibility) is irrelevant because there is another precondition (in (4): the willingness to do it) that is negated and more crucial. Often, another modal verb expressing the other precondition is often (but not necessarily) duplicated in a second men-(‘but’)-clause following the first clause. DIFA is a member of a family of ‘reactive’ constructions that reacts to or against a claim or suggestion issued by a prior speaker (or, in fact, sometimes the same speaker) (see e.g. Linell and Mertzlufft, 2014). We note that DIFA is also bipartite, with a first clause uniquely and directly identifying the construction (Linell, 2013: ‘early identifiability’) with its duplicated initial auxiliaries, and then a second ‘but’-clause (see line 2 in (4)). The construction is solidly conventionalised in conversational Swedish, so that it can occur in a truncated form with the second clause omitted; such an instance would immediately be understood as meaning that there is a superordinated condition (here: e.g. unwillingness), although this is not actually said in the truncated version. With regard to Du Bois’s (2009, 2014) dialogical notion of affinity between adjacent utterances (compare discussion of example 1), we are here, in (4), confronted with both affinities across turns (both A and B use the modal kan) and within-turn affinity (two parallel double-auxiliary segments in line 2, the latter mirroring the first one). The affinities between utterances and projections followed by fulfillments (or sometimes changes leading to redirections) are local phenomena. Indeed, Schegloff (1992) seems to regard intersubjectivity primarily as manifest local phenomena. In discussing repair phenomena, he presents third-positioned turn (the third position after the source) as “the last structurally provided defense of intersubjectivity in conversation” (Schegloff, 1992: 1295: title of the paper) An interesting aspect of the affinities and projections within and across adjacent utterances is that they suggest that the predicaments of speaker and listener in sociodialogue are not so fundamentally different as conventional theories of speech production and understanding have usually assumed, that is, with the speaker starting from an intended message and going through various stages of production to an output, a linguistically structured acoustic signal (see Levelt, 1989), and the listener starting (only?) with that acoustic signal and then progressing in the opposite direction, finally arriving at a probable intended message. Instead, Pickering and Garrod (2004, 2003) have argued that speaker’s and listener’s predicaments are partly parallel, with the speaker starting to build the utterance, and the listener following the speaker’s course, anticipating and then hearing projected aspects (Linell, 2015). It would then be more motivating to talk about about them simply as parties to communication or participants in interaction.

7.8  Intersubjectivity, Self and Internal Dialogue As already noted (Section 7.3), societies and communities are heteroglossic, replete with tensions and partial agreements as well as disagreements. This has its counterpart within individual minds, which are populated by ideas, opinions and loyalties drawn from different parts of society. We learn to cope, or not cope so well, with differences between perspectives, and we know that we have to express ourselves and behave differently in different situations and 119

Per Linell

social contexts. Thus, selves are partially split, although there are also continuities in individual minds, for example, those due to each individual’s biographical trajectory. Different ideas struggle for influence in thinking and communication. Shakespeare portrayed such internal indecisiveness in the main characters of several of his dramas (such as Richard III and Hamlet). Tensions and polyphony within the single self have been particularly stressed within dialogical self theory (DST; Hermans, 2002; Hermans and Hermans-Konopka, 2010), which is distinctly Bakhtinian in orientation, and of special interest for clinical psychology, psychotherapy, and so forth, and for analyses of new media (computer games, etc.). Another field of application is international relations in a world of globalisation. The basic idea is that individuals too have internalised some of the heterogeneities and tensions of the world, and often get engaged in ‘internal dialogues’ (Section 7.3). The self is therefore not a monolithic one, but at least partially split between different ‘I-positions’ (different perspectives, ideas, opinions, feelings, etc.).

7.9  Interactivities Are More Basic Than Intersubjectivities In the history of the human sciences, the status of knowledge and objective propositions have usually been considered more basic than interactions, in which people have been regarded “merely” as “participating”, and to realise or exchange pieces of belief and knowledge. Cognition and thinking have been seen as more central than communication and information transfer. This conceptualisation underlies monological (i.e. non-dialogical) theories, such as the information processing model of cognition, the transfer theory of communication, and the code theory of language. Phenomenology of the twentieth century has of course on many points been different from this Cartesian meta-theory, but most of its proponents, including the ‘founding father’ Husserl, have continued to support the thinking-andsubjectivity-first approach. Duranti (2010), who in several ways embraces interactionism, folows Husserl and others on this point. By contrast, Linell (2014), in his commentary on Duranti, argues that interactivities (participation in interaction) are more basic than intersubjectivities (partially shared knowledge). Individuals are not there (in any strict sense) before interaction; they are constituted, in important respects, in and through interaction and the self–other relations that emerge there. Hence it is in interactions with others that our personalities and individualities are chiselled out. An individual self can embody ‘dialogical’ feelings, such as shame, guilt, pride, complacence, complaisance, conscience, consciousness, compassion, empathy, sympathy, morality (ideas about right and wrong), and so forth. These are all phenomena that would be impossible without direct or indirect relations to others, situationally but above all cross-situationally. The distinction between individual freedom (at the expense of others’ subordination) and solidarity (actions in others’ best interest) is also based on dialogicality and interactivity. Searle (2009) discusses these issues, as well as other related notions like the phenomenon of human civilization, and monologisation with the consent of others (e.g. in some democratic organisations).

7.10  Intersubjectivity and Subjectivity In psychology and related disciplines we find a fundamental distinction between objectivity and subjectivity. Objectivity is about ontology, the ‘objective’ reality of things and

120



Intersubjectivity in Dialogue

processes ‘out there’ in the physical world (and perhaps, by extension, parts of the human world), as they are ‘in themselves’, independently of human observation and description. Subjectivity concerns, in this theory, the individual (ap)perceptions and understandings of things and processes, as they ‘appear’ to us. Subjectivity is therefore related to epistemology, and it is usually seen as accidental, uncertain and non-reliable. In order to ascertain the objective nature of bodies and environments, we need ‘objectively’ reliable instruments and scientific inference models. Many scholars have argued that intersubjectivity necessarily presupposes subjectivity as a more basic notion, a position which is associated with classical phenomenology (e.g. Husserl). Yet, intersubjective understandings were central to phenomenology, and Husserl, especially in Cartesian Meditations (1973), struggled to end up with a tenable theory about them. Duranti (2010) relates his reasonings and tries to develop an interactionist account. But it is fair to say that classical phenomenology remained monological in orientation. A dialogical meta-theory that looks at interactivity as more basic than intersubjectivity (Section 7.9) will also regard intersubjectivity as underlying individual subjectivity. Individuality and subjective ideas and understandings can only emerge when the infants have accumulated sufficient experiences of participating in social interactions. It is in and through interactivities that both individualities and social representations (and society) can come into being. Thus, the dialogical theory will see subjectivities as dependent on intersubjectivity, rather than the other way around. It is true that infants explore their environment by handling, tasting and smelling objects, manipulating them and observing the consequences visually, haptically and auditively. Thus, this apperception is firmly grounded in action–perception cycles (Noë, 2004). Infants and children learn about their environment by exploring its affordances (Gibson, 1979). But this exploration is often monitored and guided by parents and other elders who tend to continuously comment on their offsprings’ discoveries and name the objects they deal with. In this way a systematic knowledge about the world opens up for the child, and their perceptual, cognitive, emotional and linguistic experiences become intersubjectively scaffolded and consolidated. In the dialogical interplay with others and with the ecosocial environment the knower, for example, the infant with its evolving sense-making, and the known become differentiated as two sides of the same activities (Dewey and Bentley, 1949). The subject’s perception is mediated by the senses, language and instruments the individual has at his or her disposal. This is also an insight that underlies Bohr’s theory of the acquisition of knowledge of more advanced aspects of the physical world (Barad, 1996). Even though subjectivity is partly the outcome of intersubjectivity, we cannot deny that once the child has acquired knowledge and language, he or she will gradually become capable of developing sophisticated individual ideas. Yet individuality is a consequence of social interaction too, not in the least when the child gets access to literate resources and artefacts (writing, computers, etc.). Subjective deliberations are part of many human activities. But intersubjectivity is not necessarily the sum of a bunch of random deviations from systematic observation and theorising. Rather, disciplined intersubjective discussions can be a corrective to subjective derailments. Thus, most sciences have set up rules of behaviour and inferencing, and consultations among practitioners can lead to increased stability and control. For example, this is a rationale behind the common use of data sessions wthin conversation analysis.

121

Per Linell

7.11  Intersubjectivity and Objectivity Accordingly, we have defined the relation of subjectivity within an intersubjectively established world. But what about the role of objectivity in a meta-theory based on intersubjectivity? Intersubjective apprehensions are filtered through the perceptions and sense-makings of human beings. On this point dialogism follows the lead of phenomenology: we are concerned with the understandings of the world as they appear to sense-making people (‘subjects’), not primarily with the material structure of the outside world as such. So what is then the status of objectivity in such an intersubjective theory? Let us first note that there are at least two senses involved in terms like ‘object’, ‘objective’ and ‘objectivity: 1. a reality independent of human observers (or analysts); something which is ‘objectively’ there; 2. delimited, boundaried bodies, things, entities, ‘objects’.

There is often some slippage between these. In addition, ‘object’ can be used in the sense of ‘goal of action, topic of discourse’, etc. Phenomenology deals with how human beings (and (some) animals?) perceive the world; humans construct the world perceptually and cognitively in terms of objects and bodies, that is, demarcated things that move or can be moved around, sometimes be assigned agency, and so forth; we are not perceiving clouds of molecules, or substances in terms of their chemical constitution. Evolutionary biology makes similar assumptions about (some) animals. Even aspects of nature are apperceived by active perceivers, with the help of their structure-­ generating senses, perhaps reinforced by instruments of various kinds (see Marková, 2014, on Bohr’s (1999) physics, based on quantum relations rather than on relations among macroscopic objects, as in classical mechanics). If there is a tendency to constructing ‘things’ (reification, entification) in sensory perception already, and this seems to be reinforced by language. Our shared use of names and content words adds to the perceived objectivity of things talked about as well as of the linguistic devices themselves. Naming things serves to categorise and objectify things, and this is a form of sense-making. Naming is a profoundly human habit (Moscovici, 2000). Naming and talking about ‘reality’ using (proper and common) names is a partially shared, that is, intersubjective, activity, which soon becomes routinised (Majlesi, 2014). Sociology deals with social realities which are partly human (social) constructions. Sociologists have a tradition of conceiving of social phenomena, such as societies, organisations, power relations, institutions like family, profession, and such like, as factual, existing as ‘social facts’ (Durkheim, etc.) independently of being observed, described and analysed by humans (observers, analysts). In this manner social phenomena (sometimes called ‘social objects’) seem to become nearly as objective as the objects of natural sciences. Yet, social phenomena are created, shaped and continually (re)negotiated in a more constitutive manner by human beings, in and through their sense-making.4 Most phenomenologists and dialogists, and even modern physicists (such as Bohr) would probably hold that objectivity should be understood as interactively (dialogically) accomplished under intersubjective conditions. Intersubjectivity, imbued as it is with linguistic aspects, should be seen as a central dimension of being in the world. The kind of ultimate objectivity often searched for in the philosophy of natural science is, in the opinion of many renowned natural scientists, an unattainable goal. 122



Intersubjectivity in Dialogue

7.12 Conclusion Intersubjectivity is a multi-faceted phenomenon, and a multi-ambiguous term. It is not a specific, easily localisable phenomenon, nor a precise descriptive notion. In a truly dialogical approach to human sense-making, interactivities are arguably more basic than intersubjectivities (Linell, 2009, 2014). As regards intersubjectivity, partial rather than complete types are the canonical cases. Both interactivity and intersubjectivity can take many forms.

Acknowledgements Work on this chapter was supported by a grant to LinCS, Göteborg University, from the Swedish Research Council (grant no. 349-2006-146).

Notes 1 SBC (Santa Barbara Corpus of Spoken American English) is available for free download at http:// www.linguistics.ucsb.edu/research/santa-barbara-corpus (Du Bois, 2014: 361, n.1). Boldface in this example and example 4 is used by the analyst simply to draw attention to specific parts of the utterances. It does not refer to a manifest feature in the pronunciation. By contrast, underlying (of a vowel sign, as in “kan” in example 4) signifies focal stress on the syllable in question. 2 The same is true in Bockgård’s (2003) comprehensive study of Swedish completions. I am indebted to Gustav Bockgård for calling my attention to Szczepek’s study. 3 It seems that the construction with two finite verbs at the front is more or less specific for Swedish, although German and several other European languages have a similar construction with largely the same pragmatic functions. This construction, however, starts with infinitive + finite form, i.e. the German counterpart of the Swedish example (4) would start with können kann man … See Linell and Mertzlufft (2014) for discussion of the Swedish and German constructions. 4 ‘Structure’, ‘activity’, ‘power’, ‘rank’, ‘relation’, ‘process’, ‘institution’, ‘family’, etc. are all nouns, but they are hardly ‘things’. Social phenomena need not be ‘things’ (they may actually be activities, relations, linguistic or other symbolic artefacts, etc.), but in combination with sensory and linguistic processes mentioned above, plus the scientific activities of categorising phenomena, they are often seen precisely as (abstract) stable entities.

References Allwood J (2000) An activity-based approach to pragmatics. In Bunt H and Black B (eds), Abduction, Belief and Context in Dialogue: Studies in Computational Pragmatics. Amsterdam: John Benjamins, 47–80. Auer P (2005) Projection in interaction and projection in grammar. Text 25: 7–36. Bakhtin M M (1981) The Dialogic Imagination: Four Essays. Translated by Emerson C and Holquist M, edited by Holquist M. Austin: Texas University Press. Bakhtin M M (1986) Speech Genres and Other Late Essays. A selection of essays from the Russian original “Estetika slovesnogo tvorchestva” [1979]. Translated by McGee V, edited by Emerson E and Holquist M. Austin: University of Texas Press. Barad K (1996) Meeting the universe halfway: Realism and social constructivism without contradiction. In Nelson L H and Nelson J (eds), Feminism, Science and the Philosophy of Science. Dordrecht: Kluwer, 161–194. Bertau M-C (2007) On the notion of voice: An exploration from a psycholinguistic perspective with developmental implications. International Journal for Dialogical Science 2: 133–161. Bockgård G (2003) Syntax som social resurs: En studie av samkonstruktionssekvensers form och function i svenska samtal. [Syntax as a social resource: A study of the form and function of collaborative construction sequences in Swedish conversations]. Uppsala: Department of Nordic Languages. 123

Per Linell

Bohr N (1999) Collected Works. Complementarity Beyond Physics (1928–1962), vol. 10. Edited by Favrholdt D. Amsterdam: Elsevier. Bråten S (2007) Dialogens speil i barnets og språkets utvikling. [The Mirroring of Dialogue in the Development of the Child and Language]. Oslo: Abstrakt Forlag. Clark H H (1996) Using Language. Cambridge: Cambridge University Press. De Jaeger H, Di Paulo E and Gallagher S (2010) Can social interaction constitute social cognition? Trends in Cognitive Sciences 14: 441–447. Dewey J and Bentley A (1949) Knowing and the Known. Boston: Beacon Press. Du Bois J (2009) Towards a dialogic syntax. Draft retrieved in December 2009 from [email protected]. Du Bois J (2014) Towards a dialogic syntax. Cognitive Linguistics 25: 359–410. Duranti A (2010) Husserl, intersubjectivity and anthropology. Anthropological Theory 10: 1–20. Emerson C (1997) The First Hundred Years of Mikhail Bakhtin. Princeton, NJ: Princeton University Press. Gallagher, S (2011) Interpretations of embodied cognition. In Tschacher T and Bergomi C (eds) The Implications of Embodiment: Cognition and Communication. Exeter: Imprint Academic, 59–71. Gallagher S and Hutto D (2008) Understanding others through primary interaction and narrative practice. In Zlatev J, Racine T, Sinha C and Itkonen E (eds) The Shared Mind: Perspectives on Intersubjectivity. Amsterdam: John Benjamins, 17–38. Garfinkel H (1967) Studies in Ethnomethodology (Prentice-Hall). New ed. 1984 [1992]. Cambridge: Polity Press. Gibson J (1979) The Ecological Approach to Visual Perception. Boston: Houghton-Mifflin. Giles H and Smith, P M (1979) Accommodation theory: Optimal levels of convergence. In Giles H and St. Clair R M (eds) Language and Social Psychology. Baltimore: University Park Press, 45–65. Gillespie A and Cornish F (2009) Intersubjectivity: Towards a dialogical analysis. Journal for the Theory of Social Behaviour 40: 19–46. Goffman E (1971) Relations in Public. Harmondsworth: Penguin. Goffman E (1974) Frame Analysis. Cambridge, MA: Harvard University Press. Goodwin C (2017/forthc) Co-Operative Action. New York: Cambridge University Press. Goodwin M H (1990) He-Said-She-Said: Talk as Social Organization among Black Children. Bloomington and Indianapolis: Indiana University Press. Gumperz J (1982) Discourse Strategies. Cambridge: Cambridge University Press. Gumperz J (1992) Contextualization revisited. In Auer P and di Luzio A (eds) The Contextualization of Language. Amsterdam: John Benjamins, 39–53. Günthner S (2011) “N be that-constructions in everyday German conversation: A reanalysis of ‘die Sache ist/das Ding ist’(‘the thing is’)-clauses”. In Laury R and Suzuki R (eds) Subordination in Conversation: A cross-linguistic perspective. Amsterdam: John Benjamins, 11–36. Heritage J (2012) Epistemics in action: Action formation and territories of knowledge. Research on Language and Social Action 45: 1–29. Hermans H (2002) The dialogical self as a society of mind: Introduction. Theory and Psychology 12: 147–160. Hermans H and Hermans-Konopka A (2010) Dialogical Self Theory: Positioning and CounterPositioning in a Globalizing Society. Cambridge: Cambridge University Press. Hodges B (2011) Ecological pragmatics: Values, dialogical arrays, complexity, and caring. In Cowley S (ed) Distributed Language. Amsterdam: John Benjamins, 135–159. Husserl E (1973) Cartesian Meditations: An Introduction to Phenomenology. Translated by Cairns D. Dordrecht: Kluwer. Jayyusi L (1991) Values and moral judgement: Communicative praxis as moral order. In Button G (ed) Ethnomethodology and the Human Sciences. Cambridge: Cambridge University Press, 227–251. Lerner G (2002) Turn-sharing: The choral co-production of talk-in-interaction. In Ford C, Fox B and Thompson S A (eds) The Language of Turn and Sequence. Oxford: Oxford University Press, 225–256. 124



Intersubjectivity in Dialogue

Leslie A (2000) Theory of mind. In Kazdin A E (ed), Encyclopedia of Psychology, vol. 8. Washington, DC: American Psychological Association; New York: Oxford University Press, 60–61. Leudar I and Costall A (2009) (eds) Against Theory of Mind. Basingstoke: Palgrave Macmillan. Levelt W (1989) Speaking: From Intention to Articulation. Cambridge, MA: The MIT Press. Levinson S (1979) Activity types and language. Linguistics 17: 365–399. Levinson S (1992) Activity types and language. In Drew P and Heritage J (eds) Talk at Work. Cambridge: Cambridge University Press, 66–100. (Originally published in 1979 in Linguistics 17: 365–399.) Linell P (2009) Rethinking Language, Mind and World Dialogically: Interactional and contextual theories of human sense-making. Charlotte, NC: Information Age Publishing. Linell P (2010) Communicative activity types as organisations in discourses and discourses in organisations. In Tanskanen S-K, Helasvuo M-L, Johansson M, Karhukorpi J and Raitaniemi M (eds), Discourses in Interaction. Amsterdam: John Benjamins, 33–59. Linell P (2013) The dynamics of incrementation in utterance-building: Processes and resources. In Szczepek Reed B and Raymond G (eds) Units of Talk—Units of Actions. Amsterdam: John Benjamins, 57–89. Linell P (2014) Interactivities, intersubjectivities and language: On dialogism and phenomenology. Language and Dialogue 4: 165–193. Linell P (2015) Mishearings are occasioned by contextual assumptions and situational affordances. Language and Communication 40: 24–37. Linell P (2017/in press) Self-other interdependences in dialogue. In Norén C and Gille J (eds) Parler les langues romanes. Vol. 1: Discourse and Interaction. Pieterlen, Switzerland: Peter Lang, 7–31. Linell P and Luckmann T (1991) Asymmetries in dialogue: Some conceptual preliminaries. In MarkováI and Foppa K (eds) Asymmetries in Dialogue. Hemel Hempstead: Harvester Wheatsheaf, 1–20. Linell P and Marková I (2014a) (eds) Dialogical Approaches to Trust in Communication. Charlotte, NC: Information Age Publishing. Linell P and Marková I (2014b) Trust and distrust in interaction: Some theoretical and methodological points. In Linell P and Marková I (eds) Dialogical Approaches to Trust in Communication. Charlotte, NC: Information Age Publishing, 213–235. Linell P and Mertzlufft C (2014) Evidence for a dialogical grammar: Reactive constructions in Swedish and German. In Günthner S, Imo W and Bücker J (eds) Grammar and Dialogism: Sequential, syntactic and prosodic patterns between emergence and sedimentation. Berlin: de Gruyter, 79–108. Luckmann T (1986) Grundformen der gesellschaftlichen Vermittlung des Wissens: Kommunikative Gattungen. Kölner Zeitschrift für Soziologie und Sozialpsychologie, Sonderheft 27: 191–211. Majlesi A R (2014) The intersubjective objectivity of learnables: Theoretical underpinnings of praxeological and dialogical research on opportunities for learning in teacher-student interactivities. In Majlesi, A R (ed) Learnables in Action: The embodied achievement of opportunities for teaching and learning in Swedish as a second language classrooms. (Studies in Language and Culture, no. 24). Linköping: Department of Culture and Communication. Marková I (2014) Complementarity as epistemology of life. In Wagoner B, Chaudhary N and Hviid P (eds) Culture Psychology and Its Future: Complementarity in a New Key. Charlotte, NC: Information Age Publishing, 33–50. Moscovici S (2000) What is in a name? In Chaib M and Orfali B (eds) Social Representations and Communicative Processes. Jönköping: Jönköping University Press, 12–28. Noë A (2004) Action in Perception. Cambridge, MA: The MIT Press. Pickering M and Garrod S (2004) Toward a mechanistic psychology of dialogue. Behavioral and Brain Sciences 27: 169–226. Pickering M and Garrod S (2013) How tightly are production and comprehension interwoven? Frontiers in Psychology 4: Article 238 (April 2013). Rommetveit R (1974) On Message Structure. London: Wiley. Sarangi S (2000) Activity types, discourse types and interactional hybridity: The case of genetic counselling. In Sarangi S and Coulthard M (eds) Discourse and Social Life. Harlow: Longman, 1–27. 125

Per Linell

Scheff T (2006) Goffman Unbound! A new paradigm for social science. Boulder, CO: Paradigm. Schegloff E A (1991) Conversation Analysis and socially shared cognition. In Resnick L, Levine J and Teasley S (eds) Perspectives on Socially Shared Cognition. Washington, DC: American Psychological Association, 150–171. Schegloff E A (1992) Repair after next turn: The last structurally provided defense of intersubjectivity in conversation. American Journal of Sociology 97: 1295–1345. Searle J (2009). Making the Social World: The Structure of Human Civilization. Oxford: Oxford University Press. Sharrock W and Coulter J (2009) ‘Theory of Mind’: A critical commentary continued. In Leudar I and Costall A (eds) Against Theory of Mind. Basingstoke: Palgrave Macmillan, 56–88. Steffensen S V, Thibault P and Cowley S (2010) Living in the social meshwork: The case of health interaction. In Cowley S, Major J, Steffensen S V and Dinis A (eds) Signifying Bodies: Biosemiosis, Interaction and Health. Braga: The Faculty of Philosophy, Portuguese Catholic University, 207–243. Szczepek B (2000) Functional Aspects of Collaborative Productions in English Conversation. InLiSt 21. University of Konstanz, Germany. Taylor T (1992) Mutual Misunderstanding: Scepticism and the theorizing of language and interpretation. London: Routledge. Taylor T (2013) Calibrating the child for language: Meredith Williams on a Wittgensteinian approach to language socialization. Language Sciences 40: 308–320. Trevarthen C (1979) Communication and cooperation in early infancy: A description of primary intersubjectivity. In Bullowa M (ed) Before Speech: The Beginning of Interpersonal Communication. New York: Cambridge University Press, 321–347. Trevarthen C (1998) The concept and foundations of infant intersubjectivity. In Bråten S (ed) Intersubjective Communication and Emotion in Early Ontogeny. Cambridge: Cambridge University Press, 15–46. Trognon A and Batt M (2010) Interlocutory logic: A unified framework for studying conversational interaction. In Streeck J (ed) New Adventures in Language and Interaction. Amsterdam: John Benjamins, 9–46. Valsiner J and van der Veer R (2000) The Social Mind: Construction of the Idea. Cambridge: Cambridge University Press. Voloshinov V (1973) [1929] Marxism and the Philosophy of Language. Translation by Matejka L and Titunik I R of Russian original (Marksizm i filosofia jazyka). Cambridge, MA: Harvard University Press. Vygotsky, L (1978) Mind in Society: The Development of Higher Psychological Processes. Cambridge, MA: Harvard University Press. Weigand E (2010) Dialogue: The Mixed Game. Amsterdam: John Benjamins. Zlatev J (2013) The mimesis hierarchy of semiotic development: Five stages of intersubjectivity in children. Public Journal of Semiotics, 4: 47–70. Zlatev J, Racine T, Sinha C and Itkonen E (2008a) (eds) The Shared Mind: Perspectives on Intersubjectivity. Amsterdam: John Benjamins. Zlatev J, Racine T, Sinha C and Itkonen E (2008b) Intersubjectivity: What makes us human? In Zlatev J, Racine T, Sinha C and Itkonen E (eds) The Shared Mind: Perspectives on Intersubjectivity. Amsterdam: John Benjamins, 1–14.

126

8 Dialogue and Literature Roger D. Sell

8.1 Introduction For the purposes of the present chapter, literature can be defined nominalistically as a sociocultural phenomenon. Seen this way, literature consists of the texts which are regarded as literature in any particular time and place, or of the texts which, if the term ‘literature’ is not current within the particular there and then, are regarded as important, valuable, admirable, illuminating, enjoyable. To spell it out, in different times and places the category literature is filled with different kinds of texts, as different kinds of features come to be valorized. As for the relationship between literature and dialogue, this can be approached in two main ways. One way is (a) to understand dialogue in the traditional prototypical sense of an interchange between two people talking face to face with frequent turn-switching, and (b) to consider the extent to which literary texts, especially novels, plays and poems with a story element, provide representations of people engaging in such dialogues. The other way of relating dialogue to literature is (a) to see, not just face-to-face interchanges, but all language use as fundamentally dialogical in nature, and therefore (b) to describe literature, too, as a mode of dialogue, between the people who create it and the people who respond to it—readers, audiences, performers, commentators. The chapter’s first section deals with the first-mentioned approach: with dialogue as represented in literature. Sometimes the linguists, literary scholars and critics who have adopted this frame of reference have been mainly concerned with the quality of literary mimesis. About any particular literary work the question they have asked is: Does its representation of human dialogues seem realistically life-like? And in cases where the answer has been negative, some investigators have followed up with a supplementary question: Was the mimetic inaccuracy a price that had to be paid for gains on some other level? Sometimes, however, the dialogues in a literary work have been subjected to a markedly different kind of questioning: Are they something to be admired or deplored? Perhaps in terms of style, perhaps in terms of ethics, do they, or do they not, deserve to be copied in everyday life? In certain cases the conclusion has been drawn that, instead of dialogues in literature imitating real life, dialogues in real life would be greatly improved if they were to imitate the ones in literature. 127

Roger D. Sell

The chapter’s second section explains the other approach: the view of literature as dialogue. A main concern here must clearly be to establish the exact nature of the dialogical communication that goes on between the writers of literary texts and those who respond to them. For instance, were the ideas once shared by both linguists and literary scholars about Literature (sic) as a form of Art (sic) that is timelessly impersonal and non-communicational simply wrong? Or can an account of literature as dialogue actually re-formulate that older sense of an a-historical ‘specialness’ in literary texts? But then again, what about the tendency of postmodern literary theoreticians and critics to describe literary activity as not just historical, but as historically determined, as if the human beings engaging in it had no real chance of communicating with each other as autonomous individuals? Many of the scholars and critics now approaching literature as dialogue would answer this last question by saying that human beings, though indeed powerfully influenced by cultural, social, political, ideological and linguistic factors, can nevertheless muster a modicum of empathy, imagination and will-power, and that this is what enables them to enter into meaningful communication with people unlike themselves. Seen within this perspective, dialogicality brings us straight back to ethical considerations. In the approach to literature as dialogue, ethics is in fact so central that literary discussion is positively re-humanized, as illustrated by the third and final section of the present chapter. This is where the study of dialogue in literature and the study of literature as dialogue can finally interweave with each other, one significant paradox being that, at precisely the points where the dialogues of characters within a literary work’s ‘story’ are, ethically speaking, least exemplary, the work’s writer may well be treating its addressees with a genuine respect for their mental autonomy which invites them to be just as respectful in return. The chapter’s closing suggestion is that scholarship, criticism and teaching which highlight such mutually respectful relationships as they develop within literary dialogicalities may ultimately help to improve the conditions for egalitarian habits of dialogue within the post-postmodern world at large.

8.2 Dialogue in Literature One example of the interest in the realism or otherwise of dialogue in literature is Norman Page’s book Speech in the English Novel (1974). Page develops his line of enquiry through more or less explicit comparisons with the findings of professional linguists, and this same basic procedure is still followed in recent studies of J.K. Rowling’s Harry Potter and the Deathly Hallows and Friedrich Schiller’s Wilhelm Tell by, respectively, Gill Philip et al. (2013) and Christel Björkstrand (2013). In cases where such comparisons confirm the verisimilitude of a literary text’s represented dialogues, discussion usually goes little further than to give the writer an approving slap on the back. In a case where there appears to be disagreement between a literary writer’s representation of dialogue and the observations of professional linguists, a sensible next step is to re-investigate, if possible, real-life instances of the type of dialogue under examination. This way, it may be possible to determine whether professional linguists have hitherto been less than fully alert, or the particular literary writer demonstrably unrealistic. Ordinary readers have their own sense of such matters. One of their most typical complaints is that some of the people in novels by Sir Walter Scott sound like characters in the high-style parts of Shakespeare. Scholars and critics, however, before damning any such apparent inaccuracies of representation too roundly, do well to check whether compensatory artistic considerations may apply. Scott, it could perhaps be argued, was not so much deaf to the tone and dynamics of natural dialogue as concerned to endow his historical fictions with 128



Dialogue and Literature

a measure of stylistic pondus. Shakespeare himself, certainly, was not primarily interested in rendering English dialogues as they actually happen. Not only did he set the high-style parts of his plays in blank verse, but his blank verse actually makes the contents and interpersonal interplays of his characters’ conversations seem far more important and memorable than those of most ordinary dialogues, as was perhaps his intention. Similarly, some of the characters conducting serious dialogues in the texts of Plato, though not expressing themselves in verse, nevertheless make speeches that can seem unnaturally long, coherent and elegant, possibly because Plato wanted to ratchet up a high general standard of argumentation, or possibly for a reason suggested by Friedrich Schlegel and to be mentioned later on here. Either way, Plato has had numerous successors in this mode—the Anglophone ones including Dryden (in Of Dramatick Poesie, An Essay, 1668), Bishop George Berkeley (in Alciphron, or The Minute Philosopher, 1734) and a number of American philosophers examined in a book edited by Till Kinzel and Jarmila Mildorf (2014)—and we can perhaps assume that most readers have found the unnaturalnesses justified. Another area in which a writer may legitimately strain the limits of mimetic likelihood has to do with characterization. Dickens in particular is well known for making some of his characters very idiosyncratic, not only in physical appearance, dress and general bearing, but also in the way they talk. Dialogic exchanges between his caricatures can be highly entertaining, and the comedy depends, not on total verisimilitude, but precisely on the element of grotesque exaggeration, which, as compared with a more sober kind of delineation, may sometimes get at more basic truths about the way human beings interact with each other. Nor, except in degree, is Dickens so very unusual in this. As shown by Roger Fowler (1977: 103–113; 1981; 1986: 150–167), in novels by other writers, too, there can be a marked link between characterization and manner of speech in dialogue, and the same is true of much drama and narrative poetry as well. Such links are part of what Fowler discusses under the heading ‘mind-style’—his term for a person’s world-view as communicated in linguistic behaviour. A slightly different, but often overlapping, kind of acceptable inaccuracy has been pointed out by Raymond Chapman (1989) in the regional novels of Thomas Hardy. When Hardy includes dialogues between characters in a dialect variety of English, he is making a very specific appeal to readers in their role as, so to speak, listeners or overhearers. His concern is not basically to give a detailed representation of dialogue in non-standard English. In fact, like most other novelists, he reports dialect speech by means of orthographic conventions that are pretty rough and ready. His real interest is in drawing readers into the sociocultural and interpersonal dynamics of the relationships described. In our experience both of life and of novels, dialect speech tends to be associated with differences of class and background, with particular individuals’ separateness within, or solidarity with, some particular grouping, with emotional pressure breaking down polite restraints, and with ease of intimacy. As Chapman says, a novelist who purports to convey dialogues in dialect “takes the risk of partially obscuring the text in order to make the conversational situation more realistic and effective.” Or as we might also put this, there is realism and realism; what is unrealistic at the level of language can be realistic at the level of feelings, attitudes, ethics. “The text may be less important than the hidden message” (Chapman 1989: 168). The unspoken human relationships which can lie hidden under the dialogues in literature are further explored by Alan Palmer. Palmer’s main suggestion is that readers understand novels in terms of the mental processes of the characters involved in the story. From all a novel’s scattered references to a particular character, readers develop a sense of that fictional person’s continuing consciousness as a kind of embedded narrative within the novel 129

Roger D. Sell

as a whole. And the combination of several different embedded character narratives feeds into readers’ sense of the overall plot. In showing how this happens, Palmer identifies what he describes as fictional characters’ social, public and dialogic mind—the “mind beyond the skin” (Palmer 2004: 157–169). His point is that, both for characters in novels and for the readers of novels in their own daily lives, many mental processes are actually ‘intermental’—they involve ideas and modes of thought that are shared and, indeed, collaborative. This, Palmer continues, is a main reason why readers can understand what makes a novel’s characters tick, even, or especially, when the dialogues between them leave the most important things unsaid. A major aspect of literature’s pleasurable challenge is in the opportunities afforded by its represented dialogues for such reading between the lines. And to return to the implications of class differences, our understanding here has been greatly augmented by M.M. Bakhtin’s The Dialogic Imagination ([1975] 1981). What Bakhtin offers is nothing less than a sociolinguistic poetics. He sees poetry and drama largely as representing the higher, ruling classes, and as therefore linguistically stable, monological, conservative. The genre of the novel, on the other hand, he sees as far more exciting and even subversive, because it can bring into play several different social “voices” from within society’s ever-ongoing “heteroglossia” (“different-tongued-ness”). By setting these different social voices “in dialogue” with each other through the interchanges between different fictional characters, novelists such as Dostoevsky, Rabelais and Dickens have, according to Bakhtin, resisted the monologic discourse of society’s power-holders and achieved effects that are sometimes liberatingly carnivalesque in their social topsy-turvydom. Even if many commentators have objected to Bakhtin’s low opinion of poetry and drama, his observations on socio-politically dialogic oppositions in novels have been widely influential, prompting some critics, for instance, to explore the sociological dimension of constellations of caricatures in Dickens (Fowler 1983; Sell 2001a). The political and, more particularly, the ethical thrust of conversing characters in Dickens has been further examined in George Goodin’s Dickens’s Dialogue: Margins of Conversation (2013). Drawing on Victorian conversation manuals, and also on more recent philosophical, sociological and linguistic insights into the nature of conversations, Goodin explores the rhetoric employed by three of Dickens’s major character types, whose interchanges he sees as often less cooperative than implicitly violent, positively deceptive or otherwise counterproductive. Bullies such as Eugene Wrayburn (in Our Mutual Friend) use interruption, interrogation, inattention, silence and several other devices in order to compete for conversational power. Con artists such as Sam Weller (in The Pickwick Papers) seek intimacy or reduced social distance, not only by habitually whispering or shaking hands, but also through what they actually say, whether in flattery or self-deprecation. And muddlers like Cousin Feenix (in Dombey and Son) often consciously avoid the perils of clarity by introducing various forms of incoherence, not least by inserting parentheses within parentheses. Goodin’s type of conversational analysis could be applied to character dialogues in the texts of other writers as well. And for reasons which will emerge below, it is important to note that both Goodin and any other commentator applying this approach, whether to Dickens or any other writer, are bound to come up with findings which prompt negative ethical assessments of many of the human relationships portrayed in literature. Goodin makes no secret of the fact that the motivations behind the conversational behaviour of Eugene Wrayburn, Sam Weller and Cousin Feenix are far from entirely benevolent, and countless other literary characters are no better. That said, an immediate qualification is necessary. Whereas all the types of dialogue analysis noted so far have assumed that dialogues in literature are to be praised or blamed 130



Dialogue and Literature

for their fidelity or otherwise to dialogues as they actually take place in real life, there have always been commentators who have found the dialogues in certain literary texts positively exemplary, sometimes because of their very difference from real-life practice. Granted, that difference is not always clear-cut. Faced with the brilliantly witty repartees in the English Restoration comedies of Wycherley, Congreve and Farquhar, how can we know what was the chicken and what the egg? Did the higher classes of society actually speak like that? Or did they only pick it up after repeated exposure to it in the theatre? Or was it six of one and half a dozen of the other?—did current dialogue praxis and stage representations of dialogue reciprocally encourage each other’s further development? In other cases, however, literature’s perceived instructiveness as a model for conversation is beyond doubt. One quality which rhetoricians and language teachers have often recommended for imitation is an idiomatic naturalness of style, as when Renaissance pedagogues encouraged schoolboys to act out and subsequently emulate the conversational Latin on display in the comedies of Plautus and Terence. Advice has also come from the writers of etiquette handbooks or other trendsetters, persuading some upwardly mobile individuals in late-sixteenth-century England, for example, to bolster their claims to gentility by speaking in the preciously artificial manner of interlocutors in John Lyly’s prose romance Euphues: The Anatomy of Wyt (1578)—a pretentiousness which Shakespeare and other dramatists jovially satirized.

8.3 Literature as Dialogue Interesting work is still being done on dialogue in literature. Judging by current levels of scholarly interest, however, most future research is likely to concentrate on the other relationship which can obtain between the two concepts: on the fact that literature can itself be seen as a form of dialogue, between the people who create it and the people who in one way or another, and in one role or another, respond to it. Among both professional scholars and lay people, this, too, has been much discussed over the centuries, but not until recently in exactly these terms. There has been much commentary on how writers have regarded their subject-matter, what impression they have wanted to make, and what thoughts and feelings they have wanted to arouse. And there have been plenty of testimonies to the way people have actually responded to particular texts by particular writers. But the implications of these two kinds of discussion for each other, the possible connections between a writer’s manner of writing and a respondent’s manner of response, have often been left unstated, in part because the relationship between writing a text and responding to it has not been explicitly recognized as dialogical. True, there have also been many cases in which literature’s dialogicality has been acknowledged in all but name. In ancient works of rhetoric, for instance, detailed advice was offered on how to achieve specific kinds of impact on listeners or readers by specific choices of language and subject matter. Even the effect of sublimity, which Longinus described as leaving readers of Homer or Sappho or the book of Genesis quite dumb with astonishment, and as simply unachievable by mere fidelity to rhetoric’s conventional rules, could be linked to certain specific features which a writer could choose to include—Longinus’s list included: great thoughts, strong emotions, certain figures of thought and speech, noble diction, and dignified word arrangement. And when the notion of the Sublime became freshly influential in the Europe of the eighteenth and nineteenth centuries, commentators such as Edmund Burke, and many pre-Romantic and Romantic poets, had an equally clear sense of writing and reading as means by which writers and readers interact with each other. An important part of the general understanding was that a sublime impression on readers was 131

Roger D. Sell

especially likely to be made by writers who included natural imagery evoking strong feelings of fear and terror—tempestuous storms, raging seas, craggy heights, dark forests and so on. Here the particularity of what went on in a reader’s mind was very clearly perceived as correlating with the particularity of what the writer was offering for consideration. Twentieth-century linguists, on the other hand, were in no hurry to think of literature as dialogical, given their frequent phonocentric bias and given, too, their no less frequent assumption that not even spoken monologues are interactive. For them, the spoken word was primary—the true stuff of language!—and was also, therefore, their only proper focus of attention. The written word was secondary, a mere reflection of the spoken, and only to be examined as an inadequate clue to the language of periods from which audible evidence did not survive. As for spoken monologues, here linguists were blinkered by the popular understanding of dialogue as an interchange involving frequent turn-shifting, a handicap which was pointed out by Emanuel A Schegloff. Whereas Schegloff’s own conversationanalytic approach recognized that, even in what appear to be monologues, speech exchange systems are involved “in which more than one participant is present and relevant to the talk, even when only one does any talking”, many other linguists still saw “the lecture, or sermon, or story told in an elicitation interview, campfire setting, or around the table, as the product of a single speaker and a single mind” (Schegloff 1981: 71–72). Since literary texts are written, then, and since many of them involve long, uninterrupted stretches of language emanating only from their writer, linguists were indeed unlikely to pick up on literature’s interactive dimension. Many nineteenth- and twentieth-century literary scholars were labouring under what was in effect an analogous handicap, though in fairness we must also note that it reflected their attempt to capture aspects of literature which until recently have perhaps seemed rather hard to discuss in terms of dialogical communication. For one thing, although certain texts intended for stage production have conveyed, often in allegorical form, some sort of didactic message directly from the playwright to the audience, most other play texts have given no such clear impression of an authorial intention, because the relationships between the characters on stage have been more fully dramatized as involving stresses and strains by no means easy to resolve or judge. What a theatre audience sees and hears in such a case does not seem to be the playwright’s way of ‘making a point’. Rather, it comes across as happening more or less independently in a world of its own, whose bearing on the world inhabited by the play’s writer and audience can be very open to debate. For another thing, there are also countless literary texts in which we immediately recognize that part of the writer’s aim, sometimes a very large part, has been to create something beautiful. Many Petrarchan sonnets, for instance, and many lyrical poems, especially those likely to be set to music, do not necessarily allow us close glimpses into their writers’ own life-stories. Often their expressivity is only apparent, as part of an aesthetically motivated, impersonalized fiction whose chief raison d’être is hedonistic. For nineteenth-century commentators, from these kinds of observation it was only a short step to seeing the whole of literature through the prism of aesthetics as developed by Immanuel Kant ([1790] 1951) and his followers, an approach which, though valuably sensitive to literature’s indisputable kinds of ‘otherness’, overgeneralized them within a theory of Art with a capital ‘A’. Here Art was seen as offering an aesthetic heterocosm that in and of itself had nothing directly to do with either the world of nature as studied by science or, even more to the point, the world of human interaction as explored by history-writing and philosophical ethics. Instead, Art was taken to be a pleasurable wellspring of lofty Beauty (sic) and 132



Dialogue and Literature

ideal perfection, and literary Art, or Literature (sic), even though it drew on the referential medium of language, was not perceived as straightforwardly holding up a mirror to life or stating an author’s own raw feelings and opinions, but as imaginative through and through, and frankly superior to everyday reality. So much so, that literary writers sometimes came to be venerated for performing a deeply spiritual function which official religion had allegedly mismanaged, a form of deference which could only increase authors’ distance from ordinary mortals, for whom there was of course no obvious feedback channel. Then as now, an author was usually somebody quite outside a reader’s own sphere of acquaintance, and in many cases was already dead and buried. For nineteenth-century thinkers, these circumstances merely confirmed that Literature did not in the first instance involve interpersonal agency, whether the author’s own or anybody else’s. In its special timeless realm, a literary work did not ‘mean’ or ‘do’ anything, but simply ‘was’. This idea, in addition to inspiring a wealth of Symbolist, Aesthete and Modernist creative writing, underpinned not only literary criticism as produced by the early- and mid-twentieth-century Russian Formalists and American New Critics, but also late-twentieth-century speech act theory of literature and, in many cases, literary pragmatics as well, approaches whose interdisciplinarity between literary scholarship and linguistics provided academic settings within which linguists’ phonocentric bias and reluctance to see literature as truly dialogical were inevitably reinforced (Sell 2000: 28–75). Within philosophy of language, one of the most widely influential challenges to phonocentricity was Derrida’s Of Grammatology ([1967] 1974), which spoke of linguistic semiosis as both prior to, and carried by, speech and writing alike. And within linguistics proper, during the last decades of the twentieth century there were increasingly successful attempts to map both spoken monologue and written texts onto interaction. Lauri Carlson, for instance, taking a hint from Hintikka’s game-theoretical semantics, showed the analytical benefits of dialogical extrapolations; these, by spelling out the “implicit dialogue steps”, could “make the connections between the sentences of the text explicit” (Carlson 1983: 146). Today, following pioneers such as Franz Hundsnurscher (2005) and Edda Weigand (2009), linguists are increasingly coming to work precisely within the framework of dialogue analysis. And as explained in other chapters in the present volume, one of their cardinal insights has to do with the interactional dialogicality of any kind of language use whatever. During those same last decades of the twentieth century, powerful objections were also raised against Kantian aesthetics and literary formalism. One of the earliest and most perceptive challenges had been that of the American pragmatist philosopher John Dewey, but Dewey’s Art and Experience (1934) had proved unable to halt the spread of American New Criticism, and has only much more recently begun to win the recognition it deserves (Sell 2000: 241–242). Far more influential to date have been further aspects of Bakhtin’s sociolinguistic poetics, and in particular his emphasis on the addressivity of all language use, both spoken and written, both non-literary and literary, and on addressivity’s sociocultural specifics. One of his examples is a passage from Dickens’s novel Little Dorrit describing the meteoric ascent of Mr Merdle, the celebrated financier who, much later in the story, will be exposed as a bankrupt fraud: That illustrious man and great national ornament, Mr Merdle, continued his shining course. It began to be widely understood that one who had done society the admirable service of making so much money out of it, could not be suffered to remain a commoner. A baronetcy was spoken of with confidence; a peerage was frequently mentioned. (As quoted in Bakhtin [1975] 1981: 306) 133

Roger D. Sell

The italics here are Bakhtin’s, drawing attention to the ironical solidarity of all the other phrasing in this passage with the hypocritically ceremonial and establishment view of Merdle. “Making so much money out of it” brings in a different, sardonically undercutting, outsider’s voice, which readers are invited to experience as a wavelength enabling them, for the moment, to enter into deliciously close communion with Dickens himself. When Bakhtinian readings began to proliferate, some of the most suggestive came from the novelist and critic David Lodge (1990). But no less effectively anti-Kantian and antiformalist were various kinds of literary theory and criticism grouping themselves under the label ‘postmodern’: Marxist, post-Marxist, feminist, queer, postcolonial, new historicist, cultural materialist, ecological, and poststructuralist. Commentators of all these different breeds were mostly seeking, like Dewey and Bakhtin before them, but in their own different ways, to re-situate literary activity within the world of history. Through the abundance and sharpness of their detailed findings, no less than through their reorienting theories and methods, they brought about a major scholarly paradigm shift. From now on, literature was perceived as decidedly more ‘for real’. The only problem was that postmodern approaches sometimes involved an element of historicist determinism, implying that it is not human individuals who write and respond to literary texts but entire societies, cultures, or structures of belief (Sell 2000: 88–118). In poststructuralist theory, language, too, could be viewed deterministically, as when Derridean deconstructionists argued that, thanks to the arbitrariness of the signified-­signifier relationship, the process of semiosis continues so interminably, and so totally beyond the control of individual language users, that no such thing as a literary author’s meaning can ever be pinned down. In fact, to Barthes, Foucault and Kristeva the role of authors—or ‘writers’, as they preferred to call them—seemed in need of a radical toning down, and so did the relationship between writers and other writers. In their view, writers were mere workers, whose job was to craft products from the determining material of language, and who could not be thought of as drawing on each other for inspiration, or as influencing each other, because they lacked sufficiently self-conscious powers of choice. Like everyone else, writers were perceived as adrift on a humanly ungovernable sea of intertextuality. So if the agency of both writers and their audiences had once been undermined by the Kantian and formalist notion of an aesthetic heterocosm, there were times when these later approaches undermined it just as clearly, but from within frameworks of ideas that were foundationally opposed to that kind of aestheticism. It was against this background that Roger D. Sell started to develop a humanizing literary pragmatics. His approach, unlike formalist accounts of literary pragmatics, was fully historical, but without being, like some postmodern analyses, deterministic, and it soon became a fully-fledged theory of literature as one among other forms of dialogical communication (Sell 2000, 2001b, 2011a). A central concept here is of the human being as a ‘social individual’, deeply influenced by society, but also endowed with a capacity for imagination, empathy and choice. Writers, like other social individuals, are bound to take account of language, intertextualities, genres of interaction, knowledge, attitudes and values as these already exist within society, but they do so in the hope that other people will be attracted to their own project, meeting them half-way, so to speak, in a communicational process which is essentially coadaptational, and which may thereby even tend to change prevailing social norms. Readers, similarly, capitalizing on their ability to contextualize another human being’s language use in the interests of understanding, a pragmatic ability seriously underestimated by deconstructionist critics, can do their best to grasp what a writer is inviting them to consider, and may either modify their own thoughts and feelings as a result of it, or respond in ways of their 134



Dialogue and Literature

own. At which point the literary-communicational approach, though viewing communication in the etymological sense of the term as ‘community-making’, nevertheless makes a strong distinction between a community and a consensus. The common ground resulting from communication can consist, not only of shared understandings and agreements, but also of uncertainties and even disagreements. The communities brought into being through literary, or any other kind of dialogicality can be at once large and heterogeneous. The many other scholars now working along these lines have collaborated in a series of anthologies (Sell [1991] 2015, 2002, 2007, 2012a, 2014b; Sell et al. 2013; Sell and Johnson 2009; Sell et al. 2016; Sell and Verdonk 1994). And in some of this work an underlying perception is that even those discoursal features which, for nineteenth- and early-twentiethcentury commentators, seemed to give literature a special aesthetic status have, in point of fact, a communicational dimension. For a start, even though a work of art can offer a pleasurable beauty that is psychologically very real for us, so real that we experience it as something positively ‘there’ in the work, and even though this impression will certainly not occur unless there are features in the work which give rise to it, a great deal also depends on our own prior conditioning—on what we ourselves bring to our appreciation. In ways explained by Richard Shusterman (1992), one of Dewey’s foremost successors in pragmatist aesthetics, there are actually social contracts as to what shall count as agreeable in any particular time and place, so that much of a reader’s enjoyment of a literary text is the result of prior expectations being met, exceeded, or modified. In a skilful piece of writing, co-adaptations between the reassuringly familiar and the surprisingly new give rise to the pleasure we as respondents take in metrical, stylistic, formal, narratological, thematic, and genre constructions of any kind at all. Nor would this happen unless we were communicating members of an audience within which matters of taste and value are under constant, though often tacit, dialogical negotiation. At any given time, the difference between the communally well established and the communally unexpected is something we have become accustomed to recognize at first sight. And when the one enters into fruitful co-adaptation with the other, our delight can itself be communally shared. Fictionality, similarly, which in the nineteenth and twentieth centuries was so often thought of as peculiar to the aesthetic ‘specialness’ of ‘Literature’, is actually an element in many everyday uses of language whose communicational function nobody would question. The point to grasp is that communication is not always transitive; it does not always communicate something; it is not limited to conveying hard-and-fast facts, opinions and feelings. So fiction can be a means by which a communicator explores general or moral truths which go beyond the detail of particular empirical cases, or probes feelings and opinions which have yet to be stabilized into constant attitudes. In most literature which readers have felt worthy of the name, the invitation extended by any fictional elements to a truly dialogical comparing of notes is very powerful. Conversely, high literary status is not usually accorded to fictional works that are allegorical and didactic unless they expose their dogma to dialogical challenge from its own inversion, so becoming tensely dramatic. Then there is the question of writing’s lack of a feedback channel, this, too, having once been thought to make literature distinctively uncommunicative and aesthetic. Here again, special note needs to be taken of Bakhtin’s insights into addressivity. Even when literary authors have written under the auspices of a formalist aestheticism, even when they have written non-didactic drama, they have written with other people in mind, their texts implying addressee personae which we, as the real readers of a novel or a poem, or as the real readers or spectators of a play, can try on for size, not least in terms of the emotional 135

Roger D. Sell

responses they seem to assume in us (Müller-Wood 2014), and just as we negotiate the addressee personae we are constantly offered in everyday conversation. By the same token, writers also imply a persona for themselves: an addresser persona, whose function is to create an impression of their own knowledge, assumptions, attitudes and values, an impression which is neither truer nor more false than most of the other self-images we all employ in life at every moment. Literary writing, through its use of these basic communicational devices, can not only connect people who are otherwise separated by space, but can extend its dialogicality across time as well.

8.4  Literary Dialogicality and Communicational Ethics The human, interactive aspect of the dialogicality between literary writers and those who respond to them is sharply captured by scholars attuned to communicational ethics, who are not necessarily alone in this concern. For any linguist, dialogue analyst or rhetorician setting out to examine the way human beings enter into dialogical co-adaptation both with each other and with the forms and opportunities for interaction available within the surrounding culture, ethics could in principle be of central importance (Sell 2011b). The main distinction to be drawn is between modes of communication in which people fully acknowledge and respect each other’s human autonomy, and modes of communication in which people’s way of treating each other is more dictatorial, manipulative, coercive, instrumentalist. This second, less ethically acceptable kind of intercourse tends to be less than fully dialogical. More often than not, it conforms with what was the most prevalent model of communication in semiotics, linguistics, rhetorical studies and literary studies as pursued during the twentieth century: an A sends a message to a B which is to be interpreted within a single context, also set by the A. Communication structured in this transitive mode so as to communicate something is not only very frequent but often entirely blameless. To send such a message can be far and away the most suitable course of action under prevailing circumstances, as when, for instance, the A is providing the B with some helpful information. Yet the flow is always unidirectional here, and while the A can be a very active party, the B is almost inevitably more passive, and in worst case scenarios can actually be passi­ vized, marginalized, victimized. Conversely, the modes of communication in which people respect each other’s human autonomy tend to be much more fully dialogical, and to have a form which, thanks to Hundsnurscher and Weigand’s insights into interpersonal processes of “adaptation and negotiation” (Weigand 2009: 373), is now increasingly likely to receive due scholarly attention: the form of an A and a B comparing notes about something (which can include themselves) as viewed from within their two different life-worlds. These are the kinds of communication by which the overlap between one person’s life-world and another’s is most likely to get larger, as the two of them come to share more and more knowledge and understanding. Even here, of course, communicants may have ulterior motives. To empathize with somebody else’s innermost dreams and aspirations can be a cynical preliminary to selling them something they do not really need. Perhaps more typically, however, these bi-directional and bi-contextual modes of communication are communicational in the word’s etymological sense: they tend to make or consolidate a community, often a non-consensual one. In current literary-communicational research, communication of this high ethical order is often described as ‘genuine communication’, a term whose deliberately evaluative loading does not detract from its descriptive accuracy, since the ethical appropriateness it claims to identify in an utterance or text is analytically demonstrable. The concept is squarely 136



Dialogue and Literature

grounded in the Kantian take on ethics which (even though Kantian aesthetics now seems dated) is still so central for our thinking about justice (Kant [1785] 1998). Other connections are with Levinas’s account of how we recognize the human other (Wehrs and Haney 2009), and with Habermas’s (1984 & 1987, 1998a, 1998b) exploration of the ethical protocols which apply for any kind of communicative interaction at all. One form of literary-communicational scholarship can be described as ‘mediating criticism’. A mediating critic concentrates on the respondent’s side of literary dialogicality, seeking to ensure that readers, performers and interpreters do not strain the writer-respondent relationship by disregarding the human autonomy of writers. The critic acts as a kind of go-between, basically helping respondents to appreciate and relate to the full richness of a writer’s otherness. This is not intended to deprive respondents of their right to assess a text by their own standards of judgment. The mediating critic merely encourages them to refrain from the solipsistic arrogance which would recreate the text’s author in their own image. What the critic gently reminds them is that they will be entering into much more genuine communication, and will have a correspondingly more rewarding experience if, while forming their own opinion, they also make an effort to understand the writer on his or her own terms—the net result will be so much the more interestingly dialogical. After all, not even mediating criticism itself can claim to issue from an Archimedean point outside of history. Like constructive mediators within any other sphere of human interaction, mediating critics are bound to acknowledge their own positionality, partly because this inspires trust. On which footing, they can then offer the close readings and various kinds of biographical, historical, literary-historical and cultural-historical commentary which can help to ensure that, as audiences negotiate literary texts, their independence of response is tempered by a fair-minded empathy with writers’ otherness. An example of mediating criticism is Guillaume Lejeune’s (2014) discussion of the genre of the Romantic fragment as used by Friedrich Schlegel. Lejeune’s main point is that Schlegel’s fragments, though far from succeeding in their aims, deserve that those aims be fully appreciated for their sheer nobility of spirit. Schlegel was using fragments in an effort to make intuitions which were otherwise incommunicable a topic for genuine dialogue within the larger community. In the background here was the dialectics of Plato, since Schlegel realized, just like the present-day literary-critic-cum-dialogue-analyst David Fishelov (2014), that some of the ostensible dialogues reported by Plato are communicationally ungenuine, at least within the world created in the texts—where Socrates can very much rule the roost. But Schlegel, again anticipating, as we shall see, present-day students of literary dialogicality, also thought that Plato, in registering this communicational dysfunctionality as something to compare notes about, was trying to set up a more truly dialogical relationship between himself and his readers in the real world. Certainly Schlegel’s own aspiration was to bring dialectics up to date, so that social intercourse would no longer revolve around the truth as defined by a monopolar discourse of authority, but would become more democratic in its orientation, allowing a more collaborative development of thought and sensibility. Hence, precisely, his valorization of fragments, texts which did not purport to convey a definitive and comprehensive truth, and whose very fragmentariness cried out for complementation from their readers. For Schlegel, there was actually a sense in which the readers of any text, and especially the readers of a fragment, themselves became authors, by whom the tradition of communicable wisdom was carried on into the future. Viewing his own fragments within this same perspective, he wanted to believe they would generate real sociabilities and reciprocities between people of different temperaments and backgrounds. Admittedly, they did not live up to this high hope. In the 137

Roger D. Sell

end Schlegel himself acknowledged that they were fragmentary to the point of incomprehensibility, their indirection becoming, not a round-about way of reaching mental territory otherwise inaccessible, but simply a way of getting lost. Yet even if, in writing them, he was hoping they would offer far more than any text can ever in fact deliver, the fragments were already valuably highlighting the issue of communicational genuineness. Great writers of the past had been communicationally genuine, needless to say. But had anybody ever actually praised them for it in the way that Keats, for instance, ultimately under Schlegel’s influence, would go on to praise Shakespeare for “negative capability”—for the undogmatic, non-dictatorial capability of being “in Mysteries, uncertainties, doubts, without any irritable reaching after fact and reason” (Keats 1954: 53)? Still more commendably, Schlegel practised what he preached. Even at their most difficult, his texts were endorsing, not his own individuality, but the no less precious individuality of each and every reader he was inviting to complete them. From Lejeune’s mediation, then, Schlegel emerges as a magnanimous writer whom students of literary dialogicality may even want to hail as a tutelary spirit. Without Lejeune’s kind of assistance, readers today would probably find the fragments utterly baffling. With such helpful mediation, readers can try to reciprocate Schlegel’s generosity. Or as Levinas would put this, they can recognize Schlegel just as Schlegel has already recognized them. Another type of literary-communicational commentary focuses mainly on the writer’s side of literature’s dialogicality. Such work, sometimes referred to as ‘communicational criticism’, directly raises the question of whether or not particular writers have respected the autonomy of the human beings for whom they are writing. A communicational critic tends to spotlight texts which, like Schlegel’s fragments, allow respondents a certain freedom of manoeuvre—the type of generosity which is the hallmark of genuine communication in any time or place or sphere of interaction whatever. In fact, one main suggestion here is that communicational genuineness may in practice be a literary sine qua non, even though its workings are often not explicitly acknowledged, and even though different cultural epochs will always have their own additional literary expectations as well. The perception is that great authors do not force themselves upon their addressees, but rather compare notes with them in a spirit of negative capability that is essentially egalitarian (Sell 2012b). Which is not to say that their magnanimity merely capitulates to the values and tastes of audiences. Significant writing is always more co-adaptational than that, and magnanimity does not rule out some very direct indications of a writer’s own character and thought-world. As long as such directness is truthful and stimulates free discussion, it, too, is a mark of respect for addressees (Finch 2013). Nor is a more indirect manner of expression necessarily disrespectful either, since indirectness, at least when not so extreme as in Schlegel’s fragments, can encourage addressees to make fruitful contributions of their own to the construction and assessment of ideas, experiences or stories (Lindgren 2013; Chen 2014). When communicational critics discuss what happens between characters “in the story” of a literary work, they often concentrate on those less edifying features in dialogue and interaction which, as noted earlier, have been highlighted by Goodin (2013) in exchanges between some of Dickens’s characters. The paradox is the one tentatively registered by Schlegel in Plato’s treatment of Socrates: that whereas what goes on within the world represented by a text certainly can involve some such communicational dysfunction, literary texts themselves are socially ameliorative, thanks to a genuinely communicational relationship which arises between their writers and their addressees as they compare notes about the text-internal people and events. By praising literary authors here for their fair dealing

138



Dialogue and Literature

vis à vis respondents, communicational critics are by implication holding them up as models for anyone attempting worthwhile interaction in the world at large. No less than some of the commentators from earlier ages, they are seeing literature as exemplary. An instance of communicational criticism is Bénédicte Ledent’s (2014) comparison between two novels, Caryl Phillips’s Higher Ground (1989) and Marie NDiaye’s Trois femmes puissantes (2009) (translated into English as Three Strong Women), both of which address their readers in a manner which can seem oddly counter-dialogical. Not to put too fine a point on it, at first glance both Phillips and NDiaye seem to suffer from a curious inability to communicate successfully with readers. Although the books are both described on their covers as novels, they both consist of three novellas which are only loosely interconnected, and to this confusion about genre must be added a discrepancy between their titles and their actual stories (if that is the right word). The titles are suggestive of superiority and strength, yet Higher Ground seems to be about three characters who have hit rock bottom, and Trois femmes puissantes about three women who are to all intents and purposes utterly powerless. Further bafflement, perhaps even verging on positive distaste, is likely to be caused by both books’ open-endedness, by some of their more unusual linguistic features, and by their explicit descriptions of very intimate aspects of extreme human suffering. And at least for many a Western-based reader, the hermeneutic challenge will be all the greater in that several of the life-stories apparently narrated (or half-narrated) take place in Africa or within the African diaspora. All in all, says Ledent, the textuality of both novels seems hardly less than autistic. Her conclusion, however, is that such uninviting appearances are deceptive. She demonstrates that the textual, epistemological, and to some extent cultural disorientation that can affect readers here is not necessarily so frustrating after all. Our perplexity as readers can actually pave the way for a consideration of our common humanity’s full ethical complexity. The two novelists’ powerful but less than obvious invitation to addressees does bear fruit, as soon as readers are prepared to ‘listen’. Phillips and NDiaye themselves are not only energetic but negatively capable, encouraging a high level of freespirited activity from their partners in literary dialogue as well. In the last analysis, their writing is rewardingly robust, honest and empowering. So both mediating criticism and communicational criticism see literature’s dialogicality as one among other types of human interaction, and are explicitly ethical in orientation. Indeed, these two types of criticism obviously complement each other. Lejeune’s mediation between Schlegel and present-day readers is very much an account of his fragments as a form of communication. Ledent’s communicational analysis of Phillips and NDiaye would not work at all, unless she were simultaneously mediating the difficulties which can arise for readers of their novels. In David Fishelov’s Dialogues with/and Great Books (2010), the complementarity between the two types of critical emphasis is even more pronounced. As a result he is able to rehabilitate the pre-poststructuralist study of literary sources and influences, and in this way to shed new light on the process by which certain books come to be canonized as great books. Instead of seeing canonical books as singled out for aesthetic properties of the kind valorized by Kant and the literary formalists, and instead of seeing canonization as resulting from sociocultural, political and ideological power struggles of the kind detected by postmodern critics, or from the blind intertextual forces described by the poststructuralists, Fishelov highlights the workings, on both sides of the literary dialogue, of a sheer human warmth. He also stresses that authors are in dialogue, not only with other people in general, but with each other. And when an author’s work includes a dialogical echo of some other

139

Roger D. Sell

author, this can even be “an act of love” (Fishelov 2010: 25)—something to which Harold Bloom’s (1973) theory of authors’ anxiety lest they be influenced by other authors is quite oblivious. For Fishelov, the communicational genuineness of both address and response is especially striking in the way successive treatments of one and the same motif are frank but undomineering, each reflecting their own new world of values, but also accepting each other’s otherness. So although in one sense the story of Samson, for example, is always the same, between the Old Testament, Milton’s Samson Agonistes (mid-seventeenth century), Zeev Jabonitsky’s novel Samson (1927) and Cecil B. DeMille’s film Samson and Delilah (1949) there are differences which enter into open dialogue. The more a writer’s magnanimity enters into communicationally genuine dialogue with the magnanimity of other writers who are inspired to respond, the more firmly that writer’s works become embedded in the canon. A literary canon is a dialogical product. Especially, but by no means only, when in tandem like this, communicational and mediating criticism have much to tell us of literature’s dialogicality. But as partly noted already, they may also be able to make for better communication all round—for higher ethical standards of both address and response in the world at large (Siebers 2013). The hope here is that, to however small an extent, research into literary dialogicality will improve the chances for a post-postmodern community where the politics of recognition—the politics which brought so many benefits to previously underprivileged groupings during the late-twentiethcentury phase of postmodernity (Gutman 1994)—would be extended even further, but without perpetuating the postmodern tendency to define various groupings in a deterministic manner, as if the sociocultural differences between them prevented mutual understanding and two-way communication (Sell 2014a). In a post-postmodern era of rampant globalization, scholars who focus on ethical exemplariness in the give and take of literary dialogicality may help promote a non-consensual community that is at once indefinitely large and indefinitely heterogeneous, potentially global, but without being hegemonic, the kind of community within which the dialogicality of genuine communication fosters hybridities and rainbow coalitions (Rimon 2014). This rosy vision does not privilege literary activities, however, as if they were the only site of ethical exemplariness. The plain fact is that genuine communication is human history’s greatest untold story, with countless cases of it having occurred everywhere and always, but with the paradigms of relevant scholarly disciplines having usually been much better geared to describing conflict and communicational dysfunction. What remains to be seen is whether students of genuineness in the composition of, and responses to, literary texts can encourage other kinds of scholars to examine its workings in other kinds of circumstance.

References Bakhtin, MM [1975] (1981) The Dialogic Imagination: Four Essays. Austin, Texas: University of Texas Press. Björkstrand, C (2013) Politeness and social utopia in Friedrich Schiller’s Wilhelm Tell. Language and Dialogue 3: 34–55. Bloom, H (1973) The Anxiety of Influence: A Theory of Poetry. New York: Oxford University Press. Carlson, L (1983) Dialogue Games: An Approach to Discourse Analysis. Dordrecht: Reidel. Chapman, R (1989) The reader as listener: Dialect and relationships in The Mayor of Casterbridge. In Hickey, Leo (ed.) The Pragmatics of Style. London and New York: Routledge, 159–178. Chen, Y (2014) Silence and dialogue: The hermetic poetry of Wáng Wéi and Paul Celan. In Sell, RD (ed.) Literature as Dialogue: Invitations Offered and Negotiated. Amsterdam and Philadelphia: John Benjamins, 41–66. 140



Dialogue and Literature

Derrida, J [1967] (1974) Of Grammatology. Spivak, GC (trans.). Baltimore: Johns Hopkins University Press. Dewey, J (1934) Art and Experience. London: George, Allen and Unwin. Finch, J (2013) Genuine and distorted communication in autobiographical writing: EM Forster’s ‘West Hackhurst’ and its contexts. In Sell, RD, Borch, A and Lindgren, I (eds) The Ethics of Literary Communication: Genuineness, Directness, Indirectness. Amsterdam: John Benjamins, 61–80. Fishelov, D (2010) Dialogues with/and Great Books: The Dynamics of Canon Formation. Brighton: Sussex Academic Press. Fishelov, D (2014) Dialogue and dialogicity: Swift’s A Modest Proposal and Plato’s Crito. In Sell, RD (ed.) Literature as Dialogue: Invitations Offered and Negotiated. Amsterdam: John Benjamins, 23–40. Fowler, R (1977). Linguistics and the Novel. London: Methuen. Fowler, R (1981) Literature as Social Discourse: The Practice of Linguistic Criticism. London: Batsford Academic and Educational. Fowler, R (1983) Polyphony and problematic in Hard Times. In Giddings, Robert (ed.) The Changing World of Charles Dickens. London: Barnes and Noble, 91–108. Fowler, R (1986) Linguistic Criticism. Oxford: Oxford University Press. Goodin, G (2013) Dickens’s Dialogue: Margins of Conversation. Brooklyn: AMS Press. Gutman, A (ed.) (1994) Multiculturalism: Examining the Politics of Recognition. Princeton: Princeton University Press. Habermas, J (1984 & 1987) The Theory of Communicative Action. Vols. 1 & 2. Boston: Beacon Press. Habermas, J (1998a) Justification and Application: Remarks on Discourse Ethics. Cambridge, MA: MIT Press. Habermas, J (1998b) On the Pragmatics of Communication. Cambridge, MA: MIT Press. Hundsnurscher, F (2005) Studien zur Dialoggrammatik. Stuttgart: Verlag Hans-Dieter Heinz. Kant, I [1785] (1998) Groundwork of the Metaphysics of Morals. Cambridge: Cambridge University Press. Kant, I [1790] (1951) The Critique of Aesthetic Judgement [Part I: Critique of Judgement]. New York: Hafner. Keats, J (1954). Letters of John Keats. Page, Frederick (ed.). London: Oxford University Press. Kinzel, T and Mildorf, J (eds) (2014) Imaginary Dialogues in American Literature and Philosophy: Beyond the Mainstream. Heidelberg: Universitätsverlag Winter. Ledent, B (2014) The Dialogic Potential of ‘Literary Autism’: Caryl Phillips’s Higher Ground (1989) and Marie NDiaye’s Trois femmes puissantes (2009). In Sell, RD (ed.) Literature as Dialogue: Invitations Offered and Negotiated. Amsterdam: John Benjamins, 99–114. Lejeune, G (2014) Early romantic hopes of dialogue: Friedrich Schlegel’s Fragments. In Sell, RD (ed.) Literature as Dialogue: Invitations Offered and Negotiated. Amsterdam: John Benjamins, 251–270. Lindgren, I (2013) Kipling, his narrator, and the public sphere. In Sell, RD, Borch, A and Lindgren, I (eds), The Ethics of Literary Communication: Genuineness, Directness, Indirectness. Amsterdam: John Benjamins, 99–113. Lodge, D (1990) After Bakhtin: Essays on Fiction and Criticism. London: Routledge. Müller-Wood, A (2014) The role of the emotions in literary communication: Joyce’s A Portrait of the Artist as a Young Man. In Sell, RD (ed.) Literature as Dialogue: Invitations Offered and Negotiated. Amsterdam: John Benjamins, 137–159. Page, N (1974) Speech in the English Novel. London: Longman. Palmer, A (2004) Fictional Minds. Lincoln: University of Nebraska Press. Philip, G, Bongelli, R, Canestrari, C, Riccioni, I and Zuczkowski, A (2013) Negotiating narrative: Dialogic dynamics of known, unknown and believed in Harry Potter and the Deathly Hallows. Language and Dialogue 3(1): 7–33. Rimon, H (2014) Dialogues of cultures and national identity: Teuven Asher Braudes’ The Two Poles. In Sell, RD (ed.) Literature as Dialogue: Invitations Offered and Negotiated. Amsterdam: John Benjamins, 237–250. 141

Roger D. Sell

Schegloff, EA (1981) Discourse as an interactional achievement: Some uses of ‘uh huh’ and other things that come between sentences. In Tannen, D (ed.) Analyzing Discourse: Text and Talk. Washington, DC: Georgetown University Press, 71–93. Sell, RD (ed.) [1991] (2015) Literary Pragmatics. London: Routledge. Sell, RD (2000) Literature as Communication: The Foundations of Mediating Criticism. Amsterdam: John Benjamins. Sell, RD (2001a) Decorum versus Indecorum in Dombey and Son. In Sell, RD (ed.) Mediating Criticism: Literary Education Humanized. Amsterdam: John Benjamins, 165–193. Sell, RD (2001b) Mediating Criticism: Literary Education Humanized. Amsterdam and Philadelphia: John Benjamins. Sell, RD (ed.) (2002) Children’s Literature as Communication: The ChiLPA Project. Amsterdam: John Benjamins. Sell, RD (ed.) (2007) Special Issue: Literature as communication. NJES: Nordic Journal of English Studies 6. Sell, RD (2011a) Communicational Criticism: Studies in Literature as Dialogue. Amsterdam: John Benjamins. Sell, RD (2011b) Dialogicality and ethics: Four cases of literary address. Language and Dialogue 1: 79–104. Sell, RD (ed.) (2012a) Literary Community-Making: The Dialogicality of English Texts from the Seventeenth Century to the Present. Amsterdam: John Benjamins. Sell, RD (2012b) Dialogue versus Silencing: Coleridge’s The Rime of the Ancient Mariner. In Sell, RD (ed.) Literary Community-Making: The Dialogicality of English Texts from the Seventeenth Century to the Present. Amsterdam: John Benjamins, 91–129. Sell, RD (2014a) A communicational criticism for post-postmodern times. In Fludernik, M and Jacob, D (eds) Linguistics and Literary Studies: Interfaces, Encounters Transfers. Berlin: De Gruyter, 127–146. Sell, RD (ed.) (2014b) Literature as Dialogue: Invitations Offered and Negotiated. Amsterdam: John Benjamins. Sell, RD, Borch, A and Lindgren, I (eds) (2013) The Ethics of Literary Communication: Genuineness, Directness, Indirectness. Amsterdam: John Benjamins. Sell, RD and Johnson, AW (eds) (2009) Religion and Writing in England, 1558–1689: Studies in Community-Making and Cultural Memory. Farnham: Ashgate. Sell, RD, Johnson, AW and Wilcox, H (eds) (2016) Community-Making in Early Stuart Theatres: Stage and Audience. London: Routledge. Sell, RD and Verdonk, P (eds) (1994) Literature and the New Interdisciplinarity: Poetics, Linguistics, History. Amsterdam: Rodopi. Shusterman, R (1992) Pragmatist Aesthetics: Living Beauty, Rethinking Art. Oxford: Blackwell. Siebers, J (2013) The utopian horizon of communication: Ernst Bloch’s Traces and JohannPeter Hebel’s The Treasure Chest. In Sell, RD, Borch, A and Lindgren, I (eds) The Ethics of Communication: Genuineness, Directness, Indirectness. Amsterdam: John Benjamins, 189–212. Wehrs, DR and Haney, DP (eds) (2009) Levinas and Nineteenth-Century Literature: Ethics and Otherness from Romanticism through Realism. Newark: University of Delaware Press. Weigand, E (2009) Language as Dialogue: From Rules to Principles of Probability. Amsterdam: John Benjamins.

142

9 Computational Approaches to Dialogue David Traum

9.1 Introduction Computational approaches to dialogue generally fall into two categories of computational task: dialogue modeling and dialogue management. Dialogue modeling refers to developing and updating an explicit representational model of the dialogue, with a purpose of inferring other information about the participants, the content or the process of the dialogue. Dialogue modeling takes the perspective of an observer of the dialogue, but not necessarily a participant. Dialogue management refers to the process of deciding what to do next in dialogue, and thus takes the perspective of a participant in dialogue, not just an observer. Generally a dialogue system will have both dialogue modeling and dialogue management components. There are some computational systems that include only dialogue modeling, and either analyze a dialogue retrospectively, or assist one or more dialogue participants, without the system making any decisions about dialogue participation. Some examples of roles for computational systems that do dialogue modeling only are: Speech to speech translation: systems such as Verbmobil (Kay et al. 1994) do not act as an autonomous participant in dialogue, but serve as an intermediary between two people who speak different languages. The system translates the utterances of one to the language of the other, and in order to accomplish this task, maintains a model of the ongoing dialogue (Alexandersson et al. 1998), so that this dialogue model can assist in reducing language understanding perplexity, as well as use of references to prior linguistic context, such as inter-sentential anaphora or ellipsis. Meeting summarization: people have tried to model multi-party dialogues such as meeting transcripts or chat logs (Tur et al. 2010; Elsner and Charniak 2008) in order to detect things such as action items (Murray and Renals 2008) and decisions in meetings (Fernández et al. 2008). Medical corpus analysis: Dialogue modeling is also done by systems that analyze corpora of dialogue interaction in order to detect medical or psychological conditions, such as mild cognitive impairment (Lehr et al. 2012), schizophrenia (Howes et al. 2012), depression and post-traumatic stress disorder (DeVault et al. 2013). 143

David Traum

In the rest of this chapter we will focus on dialogue systems that include a dialogue management process, and can engage in dialogue. These systems may require profoundly different dialogue models from the above examples, since they have to be adequate not just for understanding (some aspect of) dialogue, but also for providing information sufficient to decide what to say. Another big difference is that, because the system takes a role in the dialogue, it can shape the contributions of other participants as well, explicitly signaling when it has problems of understanding, or taking initiative to direct the kinds of contributions of others. Thus human–computer dialogue can often be much easier for a system to model than human—human dialogue. In the next section, we introduce human–computer dialogue, look at several orientations towards systems that can engage in dialogue, and several roles for computer dialogue systems. In Section 9.3 we review several organizing principles for dialogue systems, including what sort of information state they represent, and how they decide on next moves to make in dialogue. We follow this in Section 9.4 with a discussion of some current challenge issues for dialogue systems, and conclude in Section 9.5.

9.2  Human Computer Dialogue Currently, computer systems are not able to engage in fully human dialogue, communicating using the full range of human expressions and capable of comprehending and intending the full range of meanings and purposes that such communication is put to. Indeed, some people doubt that computers could ever engage in ‘real dialogue’. On the other hand, there are certainly many similarities between human dialogue and some forms of human–machine interaction, such as purposeful use and understanding of natural language expressions and interactive patterns. Examining the types of formal and computational models that are used to model and manage computer dialogue systems can be instructive as to both how much can be captured with simplifications as well as what is missing. A general strategy for most of these systems is to focus on only one or a small set of specific genres, roles and activities for dialogue interaction. The goals of many of these systems are thus much more modest than trying to handle all types of dialogue interaction, but rather capturing some of them well enough to be useful or interesting. There are two broad types of systems that participate in dialogue, differentiated on how their designers see their function. Larsson calls these two types simulations and interfaces (Larsson 2005). Interfaces function fairly similarly to the translation systems described above, the difference being that they are not translating between two humans using two human languages but rather between one human speaking a natural language, and a back end system that communicates according to a computer language or application program interface (API). In this case, the back end system will determine the content that should be expressed to the human, and the dialogue system will take care of the translation to human language, but also, importantly, including dialogue conventions of length of turns, and will also make some autonomous decisions about how best to convey the information to the user, and engage in various feedback (Allwood et al. 1992), repair (Schegloff et al. 1977), and grounding (Clark and Wilkes-Gibbs 1986) processes to establish reliable communication. Interfaces, or more generally, the whole system, including the back end, are generally designed to perform some sort of service for a human user, such as providing information (Raux et al. 2005; Hajdinjak and Mihelic 2004; Nakano et al. 2011; Sadek et al. 1994) or completing a simple transaction (Walker et al. 1998, 2001), but allowing the user to speak in natural language rather than use a different kind of interface. These systems are seen as tools of a user, and fully subordinate to the user’s goals (within the confines of the capabilities of the system). The goals of these 144



Computational Approaches to Dialogue

systems are to accomplish the user’s goal as efficiently and effectively as possible. It is seen as perfectly acceptable if these systems diverge from actual human–human dialogue patterns, as long as they perform these tasks effectively and efficiently, and satisfy their users. On the other hand, the simulation type of systems attempt to model human dialogue capability. They do not necessarily submit completely to the user’s goals and desires (especially when it would be unnatural or unproductive to do so), and are judged not just on task efficiency, but on how human-like their behavior is. The ultimate expression of this kind of evaluation metric is the Turing Test (Turing 1950). This test posits that computers have reached human-level intelligence if an interlocutor, communicating through a text-based interface, cannot tell whether their partner is a human or computer with any more accuracy than they could tell whether it is a man or a woman. More modest evaluation schemes for this kind of system do not try to assess whether a computer dialogue participant is indistinguishable from a human, but rather whether their contributions are coherent and motivated (Traum et al. 2004) or whether it has certain qualitative abilities that are present in human communication (Bos et al. 1999). Simulation type systems include: Those primarily aimed at passing the Turing Test (or the slightly more modest Loebner competition (Loebner Prize 2015)). Systems to support role-play training of conversational and other human interaction skills, such as mission rehearsal (Hill et al. 2003), negotiation (Traum et al. 2005a), tactical questioning (Traum et al. 2007), leadership training (Hays et al. 2012), patients for medical (Stevens et al. 2006) and psychological (Kenny et al. 2007) patient interaction, and language learning (Johnson et al. 2004). Systems to educate a user, acting as tutors (Graesser et al. 2005; Litman and Silliman 2004; Zinn et al. 2002) or experts (Swartout et al. 2013; Traum et al. 2015). Systems designed to act as confederates in a psychology experiment designed to measure behavior in a specific kind of interactive context (Gratch et al. 2013).

One can also generalize from systems primarily intended to simulate human dialogue behavior to intelligent agents that intend to model human-like cognitive, intentional, and communicating entities, without specifically aiming at human emulation. These systems, like the interface systems, may prioritize effectiveness over human-likeness, but, more like the simulation systems, may put other goals ahead of user-intention satisfaction. For any given system, it can often be difficult to tell whether the goals of the designers are for it to be a tool/interface, or a simulation/agent, but these differences are often implicit in the design and evaluation methods for such systems. There are a large number of types of applications for computer systems that can act as dialogue participants. Some of these line up well with the previous dichotomy of interface versus agent, while others may cut across either design. Some examples of applications for dialogue systems include: Information provider: the system can answer user queries about some domain, e.g. information in a database (Meng et al. 1996; Raux et al. 2005). Instruction giver: the system can provide step-by-step instructions for a user to follow (Cassell et al. 2002). Service provider: the system can carry out a sequence of instructions to provide a service for the user, such as checking email (Walker et al. 1998) or booking an air flight (Walker et al. 2001). 145

David Traum

Advisor: the system can provide advice on courses of action that a user should take (Carberry et al. 1999). Collaborative partner: the system can work with one or more human participants to plan and/ or complete a task. These situations can be described as either symmetric if the humans and computers can perform the same actions (Rich et al. 2001) or asymmetric if there are complementary actions that can be performed by human and system, for example human sensing the world, and system providing expert knowledge (Smith and Hipp 1994). Tutor: the system doesn’t instruct the user how to do things, but rather helps the user learn some material and/or how to think about and solve a class of problems. Competitor: the system acts as an opponent in a game, or competitor for resources in a negotiation.

These different points of view of dialogue systems’ main function: interface or agent, as well as the different types of role for the system, as exemplified above, do much to characterize the types of interactions that the systems can engage in, as well as evaluation metrics for determining whether they are successful.

9.3  Dialogue Manager Organizing Principles In this section, we consider the decision-making process for a dialogue manager, and how it decides what kinds of moves to make in a dialogue. Key for this process is also the dialogue modeling component used to inform these decisions. We consider several approaches, including those based on modeling the structure of a dialogue, and characterizing legal or desirable end-states, as well as those based more on principles to apply, seeing the structure of dialogue as a collective achievement, rather than a goal in its own right.

9.3.1  Structure-Based Approaches Structure-based approaches to dialogue management explicitly model some notion of dialogue structure and use that model to decide on specific moves that will create valid or desirable structures. Perhaps the simplest example is a ‘ritual’ or ‘script’, specifying the dialogue contributions as an ordered list of actions by actors playing different roles. Dialogue systems can use scripts to decide what to do next based on observation of what has just happened and the script specifying the next action in the sequence. This kind of formulation may be adequate for very simple kinds of interactions, like ‘knock knock jokes’. However, these systems are very limited, because they have no way to respond intelligently if their interlocutors go ‘off-script’. Other structural accounts specify constraints on future actions but provide some flexibility as to the specific contributions. We can further classify structure-based approaches into two groups—those that focus on local structure and local coherence between utterances, and those that focus on the structure of a whole dialogue and whether it includes all of the required elements in a proper order, essentially forming a grammar of a dialogue. We examine each of these in turn. 9.3.1.1  Local Structure

Most local coherence accounts are based on adjacency pairs (Schegloff and Sacks 1973), in which dialogue units are composed of two parts, one for each speaker, and in which 146



Computational Approaches to Dialogue

expectations for the second part are set up by recognition of the first part. Examples are Question–Answer, or Offer–Acceptance. Often there is a choice of type of second part, including preferred or dispreferred responses, such as rejecting an offer or (the legitimacy of) a question. Other names for this kind of local structure unit include ‘exchanges’ (Sinclair and Coulthard 1975) and ‘initiative-response (IR) units’ (Dahlbäck and Jönsson 1998). User-initiative systems often use this kind of structure, in which the system responds to a user initiative following a policy to generate an appropriate response. Information-retrieval based approaches (Chu-Carroll and Carpenter 1999; Leuski et al. 2006) use information from the initiative to select from a set of possible responses, using relationships between the words in the initiative and the possible responses. IR units can also be used by system-initiative systems, in which a series of questions or other prompts are given to a user to respond to. These systems simplify the natural language understanding (NLU) task, since the system is concerned with understanding only those answers that make sense given the context of the system initiative, rather than the full set of possibilities. There are also some extensions to the dyadic IR structure. First, some systems allow mixed initiative in which either system or user can start an IR unit. Second, recursive structures can be allowed, in which one IR unit can embed another, for example a question is initially responded to by a clarification question and the answer to the original question occurring only after receiving the clarification response. Turns can also consist of more than one move, so that an “RI turn” can be introduced that both responds to one initiative and also starts a new one for the original speaker to respond to. Eliza, one of the first computer dialogue systems (Weizenbaum 1966), used pattern-matching to construct responses from input initiatives, using some syntactic replacements from the initiative in some of the responses. Many system turns were RI turns that generated follow-up questions to part of the user’s utterance. If there is no matching pattern, the system can also go back to earlier utterances. Eliza, and similar chatbot systems, such as ALICE (Wallace 2009) allow conversational interaction that follows many of the rules of local sequential dialogue structure, but, despite minimal use of context to refer back to previous topics, are generally not able to carry on more globally coherent interactions. Some have also looked at local exchange structures that can include more than two moves. Sinclair and Coulthard (1975) introduce an IRF (initiative response feedback) unit, including also a feedback move, which is particularly common for instructional dialogue, in which this move can contain an evaluation of the student’s response, but more generally can also include feedback on contact, perception, understanding, and attitudinal reaction (Allwood et al. 1992). The key distinction between serial IR structures and an IRF structure is that the feedback move in an IRF structure does not require a response or feedback (while the initial response does require further feedback). A similar structure was proposed by Carletta et al. (1996), who describe dialogue games as a local structure involving a variable amount of individual dialogue moves to establish the intention behind an initiating act. These local structure-based accounts can capture some of the reactive aspects of dialogue, but don’t have much to say about why particular local structures were chosen or how to transition from one to another. 9.3.1.2  Global and Grammar-Based Structural Approaches

In addition to local dialogue structure, an important set of concerns involve the highlevel structure of dialogue: how it begins and ends, why it is undertaken, how do topics 147

David Traum

flow logically from one to another to accomplish these purposes? Many system designers approach this problem using dialogue grammars or equivalent finite-state networks that define legal dialogues (e.g., Sutton et al. 1996). These formalisms allow system designers to specify the set of legal and preferred moves at any point in the dialogue, and insure that all necessary actions occur and that actions occur in a prescribed order. These kinds of systems are good for system-initiative dialogue systems, where a fairly simple conversation structure is allowed. They are less well suited when there is a large number of options as to what can be said or the order in which parts can fit into the whole, because these options generate much larger sets of possible dialogue states, and also, the states provide less context to guide the NLU component or the decision-making component of the dialogue manager. The structural approaches mentioned above are able to represent some aspects of dialogue context that can be used for further interpretation of utterances, dialogue modeling, and dialogue management decisions. One problem is that while these approaches can capture prescriptive notions of local and global dialogue coherence, they do not express or allow a system to reason about why these guidelines exist, the consequences of not following the guidelines, or how severe different kinds of divergence might be (e.g. uncooperative versus incoherent dialogue). Moreover, it can be very difficult to efficiently capture regularities in behavioral patterns that lead to similar, but not identical structures.

9.3.2  Principle-Based Approaches Principle-based approaches to dialogue modeling focus not so much on the sequential relationship of dialogue contributions, but on the impact that contributions have on the interactive dialogue context and how aspects of the context make different kinds of contributions conditionally relevant. Some of these approaches allow the system to reason about both local and global coherence of a dialogue, but also often give more indications about the nature and relative cost of violations. In this section, we review a few popular methods of framing dialogue context for dialogue management decisions. 9.3.2.1 Frame

Tree and finite-state models are suitable for modeling the logical progression of a dialogue when each piece of information provided leads to a logical next initiative; however, these models are less felicitous when what is important is just the set of information received, rather than the order in which it has been received. A dialogue system that looks up information from a database may require a set of parameters in order to formulate an appropriate query. For example, in trying to locate an appropriate air flight, a system might need to know the user’s departure and arrival cities and date of travel. However, it does not matter which order these are presented in, or whether they all are provided as responses to individual queries or offered unsolicited in a single sentence, or any combination of methods. A frame or form (Goddeau et al. 1996) model of context keeps track of the relevant set of ‘slots’ for a query, and whether they have already been ‘filled’ by user responses or are still pending in dialogue. This structure can be used by the NLU component to recognize potential fillers in user utterances, and can also be used by the dialogue manager to constrain the set of relevant next utterances. To contrast with the finite-state approach, if the finite-state model dictates first asking about the departure and then the destination, it might still ask the latter question even after a response of ‘from Los Angeles to Miami’. This could be avoided by explicitly keeping track of separate states for all possible sets 148



Computational Approaches to Dialogue

of information provided at each state, but that quickly gets unwieldy. On the other hand, the frame-approach simply updates two rather than one slot after the response, and both queries are removed from the set of pending slots that generate queries. 9.3.2.2  Logic and Plan-Based Approaches

Another kind of principle used to drive dialogue management decisions is logic and practical reasoning, to either deduce new propositional attitudes (Hintikka 1971) or actions for the system to do. The circuit-fixit shop (Smith and Hipp 1994; Smith et al. 1995) can instruct a user in how to assemble circuits from a circuit-board kit. It can tutor the user about how to perform actions or names of pieces, as well as give instructions to complete a circuit. The main method for engaging in dialogue is the missing axiom theory, in which the system attempts to prove that the circuit is working—a successful proof leads to dialogue termination, but inability to complete the proof because of ‘missing’ axioms leads to dialogue contributions to establish support for the proof. These contributions can be questions about an unknown state of the circuit, or directions to do physical actions on circuit components. The Artimis system (Bretier and Sadek 1996) is based on a theory of rational balance, including basic attitudes of belief, desire, and intention, and rationality principles, in which communication is seen as a special case of rational interaction to align the attitudes (e.g. adopting intentions to perform actions that will satisfy desires). Plan-based approaches are similar to logic-based approaches; however, the focus is more on means-end reasoning than general inference. The roots of plan-based approaches are in Speech Act theory (Austin 1962; Searle 1969), formalized using artificial intelligence planning systems (Allen et al. 1990), which are able to reason back from goal states to the initial state, using planning operators. Planning operators establish effects and can be applied if their pre-conditions hold. If not, other actions can be selected to be performed first, whose effects establish the needed pre-conditions. Speech acts are defined as actions, using planning operators (Cohen 1978; Cohen and Perrault 1979), with pre-conditions and effects related to beliefs and desires of participants. Using this framework, systems can form a complex plan, involving sequences of speech acts to achieve dialogue goals. Unlike the structural approaches and even the frame-based approaches, plan-based approaches can be more dynamic and responsive to changes in conditions, so that a new plan can be formulated whenever the system notices other changes, such as unsolicited information being provided or a change in the main goal. Moreover, plan-based accounts of speech acts can also be used in reverse to understand user intentions behind utterances (Allen and Perrault 1980; Allen 1983), and facilitate cooperative behavior that addresses the user’s presumed deeper goal rather than surface obligations. There have also been several attempts to model joint attitudes between teams of agents working on the same task, or conversing, or both. Cohen and Levesque (1990, 1991) develop an account of joint action based on primitives of mutual belief and intention, that leads to individual intentions to keep coordinated. Grosz and Sidner (1990) develop a similar model of ‘Shared Plans’, that has been the foundation of a series of dialogue systems using the ‘Collagen’ architecture (Rich et al. 2001). 9.3.2.3  Information State

The Information State Approach (Traum and Larsson 2003) is an attempt to abstract away from the specific structures and principles used for guiding dialogue management decisions, 149

David Traum

and cast them in a way where they can be more easily combined and compared. For example, many dialogue systems need to deal with the coherence between question and answer, and why a system should answer a user question. Structural approaches focus on the sequential connection between question and answer, while principle-based approaches focus on how the occurrence of the question creates a motivation for the answer; however, as described above, there can be many ways of representing the structure or the motivations. The information state approach consists of five abstract components to model developing dialogue context. Informational components are a functional specification of the aspects of context (e.g. structures, principles) that are part of the dialogue model. Formal representations are data structures, including accessibility relations for each of the informational components. For example, a history of previous utterances might be represented as a simple list, while a set of topics might be represented as a stack (Grosz and Sidner 1986). Informational components and formal representations together model the statics of dialogue context at any moment in the dialogue.

The dynamics are modeled by three additional components: Dialogue moves are abstractions of the kinds of utterance information that is seen as relevant for updates. This could be the same sort of speech acts or dialogue acts as used by many approaches (e.g., Cooper and Larsson 1999; Lewin 1998), but also more nuanced information, including the contributions of individual words to an incremental discourse representation (Poesio and Traum 1995; Poesio 1995; Poesio and Traum 1997, 1998). Update rules specify how the informational components change as a result of observation and processing of dialogue moves. For example, once a question has been answered, the question is no longer a motivating factor requiring an answer. Update strategy is a method for deciding which rules to apply when.

In addition to these dialogue modeling components, there are parallel methods for selecting system utterances for dialogue management. Selection rules specify when a dialogue move could be performed, given the current information state, while a selection algorithm decides how to prioritize the selection rules. 9.3.2.4  Statistical Reward Models

With the increasing amount of linguistic dialogue corpus data available, as well as increasing pools of users for systems, statistical approaches toward dialogue modeling and management have become more popular. These models generally represent dialogue as either a Markov decision process (MDP) (Levin et al. 2000), where the resulting state after an action follows a probability distribution, or a partially observable Markov decision process (POMDP) (Williams and Young 2007), where the system’s belief about the current state is also a probability distribution, rather than being known. Reinforcement learning can be used to optimize a policy for action given the current state or belief state, by using a reward function. In this way, the ‘rules’ used for dialogue model (or belief state) updating and dialogue policy selection can be learned from dialogue data so as to maximize expected reward. Often there is not enough existing data, so systems learn from interactions with users or simulated users (Georgila et al. 2006). Initially, these systems adapted an extension 150



Computational Approaches to Dialogue

of the structural approach, with transitions between absolute dialogue states. More recently, however, these systems have used an information state, involving a set of features that are either present or absent in a state, rather than a full model of all possible states.

9.4  Issues for Dialogue Systems In this section, we briefly summarize some important issues for dialogue systems, particularly those attempting to model dialogue that is similar to human–human dialogue. Many systems attempt to simplify or eliminate these issues by constraining the types of interactions, number of users, and dialogue activities that the systems can participate in; however this is usually at the cost of naturalness and flexibility of normal human– human dialogue.

9.4.1  Grounding, Feedback and Repair Grounding (Clark and Wilkes-Gibbs 1986; Clark and Schaefer 1989) is the process by which participants in a conversation establish new common ground. Many early computational accounts and systems assume that mutual belief results just from performing an utterance in a common, mutually perceptible space. By contrast, studies of human dialogue show that there is often explicit feedback (Allwood et al. 1992) given, indicating the degree and sometimes content of what has been understood. When a (perceived or assumed) lack of common ground exists, then participants often take the trouble to repair the (potential) misunderstanding (Schegloff et al. 1977). Grounding is thus a collaborative process (Clark & Wilkes-Gibbs 1986), in which the participants of a conversation jointly work to establish common ground, rather than separately performing tasks of encoding and decoding meaning to/from language. Grounding can be even more of a difficult issue for human–computer dialogue than for human–human dialogue, because of difficulties in speech recognition, language interpretation, common ontologies, and common-sense reasoning for computer dialogue systems. Many systems include explicit dialogue acts and sequences or principles to deal with grounding phenomena. For example, consider the system-initiated question–answer exchange in example 1, utterances 1 and 2. Each of the possible system follow-up moves in 3 is plausible, but has a different role in terms of grounding. (1) 1 System: Where do you want to go? 2 User: Boston. 3 System: a. Tell me more about your travel plans. b. When would you like to go? c. When would you like to go to Boston? d. ok e. Boston. f. Boston? g. Do you want to go to Boston? h. Did you say Boston? i. Boston or Austin? j. Where? k. Please repeat that. 151

David Traum

Responses 3a–c all indicate a desire to move forward with the dialogue, moving from the filled slot of destination to an open slot (e.g. time of departure). Response 3a gives initiative back to the user to select which slot to talk about next, while 3b and 3c specify the departure slot. Responses 3a and 3b do not give any explicit signal of what was understood; however, they implicitly provide evidence of understanding by passing up an opportunity to request repair and continuing with a next relevant contribution (Clark and Schaefer 1989). Response 3c both indicates that “Boston” was understood as the destination, as well as moving along to the departure time slot. Responses 3d and 3e acknowledge understanding of what was said, without either requiring further feedback or confirmation but also without moving on. It is thus unclear whether the next turn should be by system or user. 3d, like 3a and 3b just claims understanding (in this case more explicitly), while 3e displays that understanding. Utterances 3c and 3e–i all give some evidence of having understood “Boston” as the destination. 3c is what is often called (e.g., (Möller 2005)) an implicit confirmation, contrasted with 3f–h, which are termed explicit confirmation. Even though both kinds of utterance are equally explicit about what was heard, the difference is that ‘implicit confirmations’ have as a main purpose something other than confirmation, while the main purpose of an ‘explicit confirmation’ is to confirm the system’s expressed understanding. Implicit confirmations can be more efficient if understanding is correct; however, they can be harder to notice or produce a correction for, if needed. 3f–h are all ‘explicit confirmations’; however, they differ in terms of what aspect they are conforming. 3g focuses on “Boston” as being the destination, while 3h focuses on the surface expression rather than the meaning. 3f is ambiguous between whether it is confirming what was heard or meant, though is equally vague about the slot name. 3i–k indicate less confidence in having understood than the previous alternatives. 3i presents two options, and might be preferred if both are likely, given the overall context and what was heard. This also makes it more likely than previous utterances that an error will not go unnoticed. 3j and 3k are both more appropriate when it is more likely that the system did not understand correctly, and a confirmation would be more likely wrong and confusing than helpful. Like the contrast with 3g and 3h, 3j indicates the slot name that is at issue, while 3k focuses on the surface level of what was said rather than what was meant or the context in which it is to be understood. Systems may produce all or only a subset of these response types, as well as other variations. Any of the control mechanisms from Section 9.2 could be used to decide on a policy of when to use each response type. For example, a reinforcement learning model might use features such as the current confidence level of the interpretation, confidence levels for the next highest interpretation, user style preferences, and number of attempts made to elicit this information, and learn the ideal threshold values for deciding on one move or another, in terms of dialogue length and user satisfaction. Some systems go beyond simple confirmation and repair strategies, and model the state of grounding for different aspects of information. For example, the systems described in Allen and Schubert (1991), Matheson et al. (2000), Traum and Rickel (2002), and Traum et al. (2005a) all use versions of the speech acts model of grounding of Traum and Allen (1992) and Traum (1994), which makes use of local structures for each ‘discourse unit’ as well as principles for deciding on next moves. An extended grounding model, that takes into account degrees of grounding and different patterns of grounding, was presented by Roque and Traum (2008), and used in several systems (Roque et al. 2006; Roque and Traum 2009).

152



Computational Approaches to Dialogue

9.4.2  Non-cooperative Dialogue Many accounts of meaning in dialogue rely on cooperative principles (Grice 1975). This is not a problem for the majority of dialogue systems that play a role that is either a tool for the human user to accomplish her goals, or at least a collaborative partner that shares goals with the user. However, many dialogue roles are not purely cooperative: goals fail to align exactly, or may even be diametrically opposed. In a commercial transaction, both parties may have commonality of goals in coming to a mutually satisfactory deal, but the buyer will want the price to be as low as possible, while the seller would like the price as high as possible. Even in cases in which overall goals converge, such as for a tutoring system, local goals may still be such that ideal systems would not be perfectly cooperative. For example, a student might ask for the answer to a difficult problem, while the tutor might think it would be better for the student to figure it out himself rather than be given the full answer. Dialogue systems that play a non-cooperative role, such as in a negotiation–commercial transaction, tutoring–coaching session, or zero-sum game, must make some adjustments compared to tool and helper systems. Understanding and complying with user intent is not sufficient, and in some cases not necessary, for deciding on system behaviors. Other mechanisms, such as discourse obligations (Traum and Allen 1994), are needed to motivate coherence, in the absence of adopting the goals of a collaborative partner. A certain amount of deliberation is also needed to optimally balance concerns of the user with the system’s own goals, to decide when to choose one over the other. The PRACMA system (Jameson et al. 1994; Jameson and Weis 1995) was one of the first to explore principle-based non-cooperative agents in a sales context. Traum (2012) summaries some of the issues involved, and briefly describes systems for training people in negotiation (Traum et al. 2005b; Plüss et al. 2011) and tactical questioning (Roque and Traum 2007; Gandhe et al. 2008). Other recent work on dialogue models for negotiation and non-cooperative dialogue have been presented by (Guhe and Lascarides 2012; Asher and Lascarides 2013; Efstathiou and Lemon 2014).

9.4.3  Multimodal Dialogue While most dialogue systems focus only on text or speech, natural dialogue has a lot of information coming through other channels as well. Multimodal dialogue systems allow one to communicate also through other channels. Some systems that are on a computer, or other device with a screen, allow the system to display graphical or tabular information. Some of these systems also allow users to use pointer input as well as speech or text. For example, Johnston et al. (2002) allow users to select a region on a map to ask about. Other systems allow more human-like natural gestures in communication. These include complementary deictic gestures that point out an object or location, as well as gestures for emphasis, iconic gestures, and signals of addressee selection, turn-taking, and various kinds of feedback. Cassell et al. (2000) present embodied conversational agents that give the system an animated body as well as speech to express conversational signals. These systems are also called virtual humans (Gratch et al. 2002). Some researchers also focus on dialogue with robots (Burke et al. 2002; Matsusaka 2008) that have physical bodies, and in some cases the ability to move around the environment, change points of view and perspective. There are currently several toolkits that allow people to associate specific physical behaviors with communicative intent (Cassell et al. 2001; Lee and Marsella 2006), and which can automatically apply rules to select these behaviors. 153

David Traum

There are also active research efforts on understanding communicative behaviors from the visual or multimodal channels, and integrating these channels in dialogue models (Traum and Morency 2010). Often using multiple channels can provide not just more and different kinds of information, but better accuracy or confidence in information that is observable in different modalities. For instance, visual lip reading could enhance speech recognition, or using gaze as well as lexical information could more accurately allow recognition of addressees. In this case, there is a need for multimodal fusion of independent input signals. Late fusion is more common, in which recognition is made via each modality, and then the separate pieces of information are merged into a unified whole. Early fusion means that the raw signals or intermediate features from different modalities are combined in order to select a joint hypothesis.

9.4.4  Multiparty Dialogue Most dialogue systems look only at dyadic interaction, primarily between the system and one human user. However, much of human dialogue occurs in a context of small group interaction involving more than two entities. Multiparty conversation brings additional issues that must be modeled. First, there are multiple conversation participant roles, rather than just speaker and hearer (Goffman 1981; Levinson 1987). Addressee identification (Martinovski et al. 2003; Jovanovic and op den Akker 2004; op den Akker and Traum 2009) is an important problem with more than one non-speaker present, so the system can decide whether or not it must respond. Traum (2004) discusses how other issues such as turn-taking, grounding, interaction management and obligations are impacted by multiple participants. Several systems simulate multiparty interaction (Padilha and Carletta 2002; Patel et al. 2004) including variations for different cultures (Jan et al. 2007). Other systems allow interactions with multiple people and an artificial system (Matsusaka 2008) or one user and multiple agents (Traum et al. 2008; Plüss et al. 2011). Bohus and Horvitz (2009) also consider the case of multiple users who may have side conversations, and how to detect engagement in open-world dialogue, not constrained to a laboratory experiment or single task.

9.5 Conclusions In this chapter we have briefly surveyed some of the diversity of approaches to computational modeling dialogue and constructing systems that can engage in dialogue with users. While they all have in common a certain degree of formalization, in order to allow computers to process dialogue, they all equally fall short of full expression of the range of mechanisms and uses to which human dialogue is put. Some approaches tend more toward engineering solutions to engage in dialogue for a specific task, while other approaches try to capture more of the essence of human-dialogue skills, including some of the ways that dialogue is inextricably connected with cognitive and social skills. What is common in dialogue systems is a focus on a model of dialogue that is not strictly language understanding coupled with language production, but each in support of units containing input from the self and other entities, that must be linked for coherent dialogue interaction. The formalization creates a kind of linguistic competence that, when coupled with the interactions with real people, creates a kind of “competence-in-performance” (Weigand 2006), in which the systems must cope with novel interactive behaviors. At present, these skills are extremely limited, much of it occurring through human analysis of data and

154



Computational Approaches to Dialogue

updating of models, rather than the system developing new understanding and capabilities through dialogue use. However, we see rapid growth in the area, not just in fundamental algorithms and vastly increased amounts of data, but also in the more widespread use of such systems and diversity and complexity of the interactions that are being achieved.

Acknowledgements This work was supported by the U.S. Army. Any opinion, content or information presented does not necessarily reflect the position or the policy of the United States Government, and no official endorsement should be inferred.

References Alexandersson, J., Buschbeck-Wolf, B., Fujinami, T., Kipp, M., Koch, S., Maier, E., Reithinger, N., Schmitz, B. & Siegel, M. (1998), Dialogue Acts in VERBMOBIL-2 – Second Edition, VerbmobilReport 226, DFKI Saarbrücken, Universität Stuttgart, Technische Universität Berlin, Universität des Saarlandes. Allen, J. F. (1983), Recognizing intentions from natural language utterances. In M. Brady & R. C. Berwick, eds, Computational Models of Discourse, MIT Press, Cambridge. Allen, J. F. & Perrault, C. R. (1980), Analyzing intention in utterances. Artificial Intelligence 15(3), 143–178. Allen, J. F. & Schubert, L. K. (1991), The TRAINS project, TRAINS Technical Note 91-1, Computer Science Dept. University of Rochester. Allen, J. F., Hendler, J. & Tate, A., eds (1990), Readings in Planning. Morgan Kaufmann, San Francisco. Allwood, J., Nivre, J. & Ahlsen, E. (1992), On the semantics and pragmatics of linguistic feedback. Journal of Semantics 9(1), 1–26. Asher, N. & Lascarides, A. (2013), Strategic conversation. Semantics and Pragmatics 6(2), 1–62. Austin, J. A. (1962), How to Do Things with Words. Harvard University Press, Cambridge. Bohus, D. & Horvitz, E. (2009), Models for multiparty engagement in open-world dialog. In Proceedings of the SIGDIAL 2009 Conference: The 10th Annual Meeting of the Special Interest Group on Discourse and Dialogue, Association for Computational Linguistics, 225–234. Bos, J., Bohlin, P., Larsson, S., Lewin, I. & Matheson, C. (1999), Evaluation of the model with respect to restricted dialogue systems, Technical Report Deliverable D3.2, Trindi. Bretier, P. & Sadek, M. D. (1996), A rational agent as the kernel of a cooperative spoken dialogue system: Implementing a logical theory of interaction. In J. P. Müller, M. J. Wooldridge & N. R. Jennings, eds, Intelligent Agents III—Proceedings of the Third International Workshop on Agent Theories, Architectures, and Languages (ATAL-96), Lecture Notes in Artificial Intelligence, Springer-Verlag, Heidelberg. Burke, C., Harper L. & Loehr, D. (2002), A dialogue architecture for multimodal control of robots. In Proceedings of the International CLASS Workshop on Natural, Intelligent and Effective Interaction in Multimodal Dialogue Systems, Edmonton. Carberry, S., Chu-Carroll, J. & Elzer, S. (1999), Constructing and utilizing a model of user preferences in collaborative consultation dialogues. Computational Intelligence 15, 185–217. Carletta, J., Isard, A., Isard, S., Kowtko, J., Doherty-Sneddon, G. & Anderson, A. (1996), HCRC dialogue structure coding manual, Technical Report 82, HCRC. Cassell, J., Vilhjálmsson, H. & Bickmore, T. (2001), Beat: The behavior expression animation toolkit. In Proceedings of ACM SIGGRAPH, ACM Press, New York, 477–486. Cassell, J., Sullivan, J., Prevost, S. & Churchill, E., eds (2000), Embodied Conversational Agents. MIT Press, Cambridge.

155

David Traum

Cassell, J., Stocky, T., Bickmore, T., Gao, Y., Nakano, Y., Ryokai, K., Tversky, D., Vaucelle, C. & Vilhjálmsson, H. (2002), Mack: Media lab autonomous conversational kiosk. In ‘Proceedings of Imagina 2002’, Monte Carlo. Chu-Carroll, J. & Carpenter, B. (1999), Vector-based natural language call routing. Journal of Computational Linguistics 25(30), 361–388. Clark, H. H. (1992), Arenas of Language Use. University of Chicago Press, Chicago. Clark, H. H. & Wilkes-Gibbs, D. (1986), Referring as a collaborative process. Cognition 22, 1–39. (Also appears as Chapter 4 in Clark (1992).) Clark, H. H. & Schaefer, E. F. (1989), Contributing to discourse, Cognitive Science 13, 259–294. Cohen, P. R. (1978), On knowing what to say: Planning speech acts. PhD thesis, University of Toronto. (Reproduced as TR 118 Department of Computer Science, University of Toronto.) Cohen, P. R. & Perrault, C. R. (1979), Elements of a plan-based theory of speech acts. Cognitive Science 3(3), 177–212. Cohen, P. R. & Levesque, H. J. (1990), Rational interaction as the basis for communication. In P. R. Cohen, J. Morgan & M. E. Pollack, eds, Intentions in Communication, MIT Press, Cambridge. Cohen, P. R. & Levesque, H. J. (1991), Teamwork, Nous 25(4): 487–512. Cooper, R. & Larsson, S. (1999), Dialogue moves and information states. In H. Bunt & E. C. G. Thijsse, eds, Proceedings of the Third International Workshop on Computational Semantics, Tilburg. Dahlbäck, N. & Jönsson, A. (1998), A coding manual for the linköping dialogue model. Unpublished manuscript. DeVault, D., Georgila, K., Artstein, R., Morbini, F., Traum, D., Scherer, S., Rizzo, A. S. & Morency, L.-P. (2013), Verbal indicators of psychological distress in interactive dialogue with a virtual human. In Proceedings of SIGdial 2013 Conference, Metz. Efstathiou, I. & Lemon, O. (2014), Learning non-cooperative dialogue behaviours. In Proceedings of the SIGdial 2014 Conference, Philadelphia. Elsner, M. & Charniak, E. (2008), You talking to me? A corpus and algorithm for conversation disentanglement. In Proceedings of ACL-08: HLT, Association for Computational Linguistics, Columbus, 834–842. Fernández, R., Frampton, M., Ehlen, P., Purver, M. & Peters, S. (2008), Modelling and detecting decisions in multi-party dialogue. In Proceedings of the 9th SIGdial Workshop on Discourse and Dialogue, SIGdial ‘08, Association for Computational Linguistics, Stroudsburg, 156–163. Retrieved from: http://dl.acm.org/citation.cfm?id=1622064.1622095 Gandhe, S., DeVault, D., Roque, A., Martinovski, B., Artstein, R., Leuski, A., Gerten, J. & Traum, D. (2008), From domain specification to virtual humans: An integrated approach to authoring tactical questioning characters. In Proceedings of Interspeech Conference. Georgila, K., Henderson, J. & Lemon, O. (2006), User simulation for spoken dialogue systems: Learning and evaluation. In Proceedings of Interspeech Conference. Goddeau, D., Meng, H., Polifroni, J., Seneff, S. & Busayapongchai, S. (1996), A form-based dialogue manager for spoken language applications. Proceedings of Fourth International Conference on Spoken Language Processing, ICSLP 96. 2, 701–704. Goffman, E. (1981), Footing, in E. Goffman and D. Hymes, eds, Forms of Talk, University of Pennsylvania Press, Philadelphia, 124–159. Graesser, A., Chipman, P., Haynes, B. & Olney, A. (2005), Autotutor: An intelligent tutoring system with mixed-initiative dialogue. IEEE Transactions on Education 48(4), 612–618. Gratch, J., Hartholt, A., Dehghani, M. & Marsella, S. C. (2013), Virtual humans: A new toolkit for cognitive science research. In Cognitive Science. Gratch, J., Rickel, J., Andre, E., Cassell, J., Petajan, E. & Badler, N. (2002), Creating interactive virtual humans: Some assembly required. IEEE Intelligent Systems 54–63. Grice, J. P. (1975), Logic and conversation. In P. Cole & J. L. Morgan, eds, Syntax and Semantics, Vol. 3: Speech Acts. Academic Press, Amsterdam, 41–58. Grosz, B. J. & Sidner, C. L. (1986), Attention, intention, and the structure of discourse. Computational Linguistics 12(3), 175–204. 156



Computational Approaches to Dialogue

Grosz, B. J. & Sidner, C. L. (1990), Plans for discourse. In P. R. Cohen, J. Morgan & M. E. Pollack, eds, Intentions in Communication, MIT Press, Cambridge. Guhe, M. & Lascarides, A. (2012), Trading in a multiplayer board game: Towards an analysis of noncooperative dialogue. In Proceedings of the 34th Annual Meeting of the Cognitive Science Society, Cognitive Science Society, Austin. Hajdinjak, M. & Mihelic, F. (2004), Information-providing dialogue management. In P. Sojka, I. Kopecek & K. Pala, eds, TSD, Vol. 3206 of Lecture Notes in Computer Science, Springer, New York, 595–602. Hays, M., Campbell, J., Trimmer, M., Poore, J., Webb, A., Stark, C. & King, T. (2012), Can role-play with virtual humans teach interpersonal skills? In Interservice/Industry Training, Simulation and Education Conference (I/ITSEC), Orlando. Hill, Jr., R., Gratch, J., Marsella, S., Rickel, J., Swartout, W. & Traum, D. (2003), Virtual humans in the mission rehearsal exercise system. KI Embodied Conversational Agents 17, 32–38. Hintikka, J. (1971), Semantics and propositional attitudes. In L. Linsky, ed., Reference and Modality, Oxford University Press, Oxford, 145–167. Howes, C., Purver, M., McCabe, R., Healey, P. G. T. & Lavelle, M. (2012), Predicting adherence to treatment for schizophrenia from dialogue transcripts. In Proceedings of the 13th Annual Meeting of the Special Interest Group on Discourse and Dialogue, Association for Computational Linguistics, Seoul, 79–83. Retrieved from: http://www.aclweb.org/anthology/W12-1610 Jameson, A. & Weis, T. (1995), How to juggle discourse obligations. In Proceedings of the symposium on Conceptual and Semantic Knowledge in Language Generation. Jameson, A., Kipper, B., Ndiaye, A., Schäfer, R., Simons, J., Weis, T. & Zimmermann, D. (1994), Cooperating to be noncooperative: The dialog system PRACMA. In B. Nebel & L. DreschlerFischer, eds, Proceedings of the Eighteenth German Conference on Artificial Intelligence (KI-94), Springer, 106–117. Jan, D., Herrera, D., Martinovski, B., Novick, D. G. & Traum, D. R. (2007), A computational model of culture-specific conversational behavior. In Pelachaud, C., Martin, J.-C., André, E., Chollet, G., Karpouzis, K. & Pelé, D., eds, Intelligent Virtual Agents, 7th International Conference, IVA 2007, Paris, France, September 17–19, 2007, Proceedings. Vol. 4722 of Lecture Notes in Computer Science, Springer, 45–56. Johnston, M., Bangalore, S., Vasireddy, G., Stent, A., Ehlen, P., Walker, M., Whittaker, S. & Maloor, P. (2002), Match: An architecture for multimodal dialogue systems. In Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics (ACL), Philedelphia, Pennsylvania, 376–383. Johnson, W. L., Beal, C. R., Fowles-Winkler, A., Lauper, U., Marsella, S., Narayanan, S., Papachristou, D. & Vilhjálmsson, H. H. (2004), Tactical language training system: An interim report. In J. C. Lester, R. M. Vicari & F. Paraguaçu, eds, Intelligent Tutoring Systems, Vol. 3220 of Lecture Notes in Computer Science, Springer, New York, 336–345. Jovanovic, N. & op den Akker, R. (2004), Towards automatic addressee identification in multi-party dialogues. In Proceedings of 5th SIGdial Workshop on Discourse and Dialogue, Cambridge. Kay, M., Norvig, P. & Gawron, M. (1994), Verbmobil: A Translation System for Face-to-Face Dialog, University of Chicago Press, Chicago. Kenny, P. G., Parsons, T. D., Gratch, J., Leuski, A. & Rizzo, A. A. (2007), Virtual patients for clinical therapist skills training. In Pelachaud, C., Martin, J.-C., André, E., Chollet, G., Karpouzis, K. & Pelé, D., eds, Intelligent Virtual Agents, 7th International Conference, IVA 2007, Paris, France, September 17-19, 2007, Proceedings. Vol. 4722 of Lecture Notes in Computer Science, Springer, New York, 197–210. Larsson, S. (2005), Dialogue systems: Simulations or interfaces? In Proceedings of the Ninth Workshop on the Semantics and Pragmatics of Dialogue. Lee, J. & Marsella, S. (2006), Nonverbal behavior generator for embodied conversational agents. In Proceedings of the 6th International Conference on Intelligent Virtual Agents, Springer, Heidelburg, 243–255. 157

David Traum

Lehr, M., Prud’hommeaux, E., Shafran, I. & Roark, B. (2012), Fully automated neuropsychological assessment for detecting mild cognitive impairment. In Interspeech 2012: 13th Annual Conference of the International Speech Communication Association, Portland. Leuski, A., Patel, R., Traum, D. & Kennedy, B. (2006), Building effective question answering characters. In Proceedings of the 7th SIGdial Workshop on Discourse and Dialogue, Sydney, 18–27. Levin, E., Pieraccini, R. & Eckert, W. (2000), A stochastic model of human–machine interaction for learning dialog strategies. In Proceedings of the 5th European Conference on Speech Technology (Eurospeech ’97), Rhodes, 1883–1886. Levinson, S. C. (1987), Putting linguistics on a proper footing: Explorations in Goffman’s participation framework. In P. Drew & A. Wootton, eds, Erving Goffman: Exploring the Interaction Order. Polity Press, Oxford, 161–227. Lewin, I. (1998), The autoroute dialogue demonstrator, Technical Report CRC-073, SRI Cambridge Computer Science Research Centre. Litman, D. J. & Silliman, S. (2004), Itspoke: An intelligent tutoring spoken dialogue system. In Demonstration papers at HLT-NAACL 2004, Association for Computational Linguistics, Boston, 5–8. Loebner Prize in Artificial Intelligence (2015), The First Turing Test. Retrieved from: http://www. loebner.net/Prizef/loebner-prize.html Martinovski, B., Traum, D., Robinson, S. & Garg, S. (2003), Functions and patterns of speaker and addressee identifications in distributed complex organizational tasks over radio. In Diabruck: Seventh workshop on semantics and pragmatics of dialogue, Scheidburg. Matheson, C., Poesio, M. & Traum, D. (2000), Modelling grounding and discourse obligations using update rules. In Proceedings of the First Conference of the North American Chapter of the Association for Computational Linguistics. Matsusaka, Y. (2008), History and current researches on building a human interface for humanoid robots. In I. Wachsmuth & G. Knoblich, eds, Modeling Communication with Robots and Virtual Humans, Springer, New York, 109–124. Meng, H., Busayapongchai, S., Giass, J., Goddeau, D., Hethetingron, L., Hurley, E., Pao, C., Polifroni, J., Seneff, S. & Zue, V. (1996), Wheels: A conversational system in the automobile classifieds domain. In Spoken Language, 1996. ICSLP 96. Proceedings, Fourth International Conference on Spoken Language Processing, Vol. 1, IEEE, Philadelphia, 542–545. Möller, S. (2005), Quality of Telephone-Based Spoken Dialogue Systems, Springer Science & Business Media, New York. Murray, G. & Renals, S. (2008), Detecting action items in meetings. In A. Popescu-Belis & R. Stiefelhagen, eds, Machine Learning for Multimodal Interaction, Vol. 5237 of Lecture Notes in Computer Science, Springer, Berlin, 208–213. Nakano, M., Sato, S., Komatani, K., Matsuyama, K., Funakoshi, K. & Okuno, H. G. (2011), A two-stage domain selection framework for extensible multidomain spoken dialogue systems. In Proceedings of the SIGDIAL 2011 Conference, Association for Computational Linguistics, Association for Computational Linguistics, Portland, 18–29. op den Akker, R. & Traum, D. (2009), A comparison of addressee detection methods for multiparty conversations. In Proceedings of DIAHOLMIA’09, the 13th Semdial Workshop on the Semantics and Pragmatics of Dialogue, Stockholm, 99–106. Padilha, E. & Carletta, J. (2002), A simulation of small group discussion. In Proceedings of EDILOG 2002: Sixth Workshop on the Semantics and Pragmatics of Dialogue, 117–124. Patel, J., Parker, R. & Traum, D. (2004), Simulation of small group discussions for middle level of detail crowds. In Proceedings of the Army Science Conference, Orlando. Pelachaud, C., Martin, J.-C., André, E., Chollet, G., Karpouzis, K. & Pelé, D., eds (2007), Intelligent Virtual Agents, 7th International Conference, IVA 2007. Paris, France, September 17–19, 2007, Proceedings, Vol. 4722 of Lecture Notes in Computer Science, Springer, New York. Plüss, B., DeVault, D. & Traum, D. (2011), Toward rapid development of multiparty virtual human negotiation scenarios. In Proceedings of SemDial 2011, the 15th Workshop on the Semantics and Pragmatics of Dialogue, Los Angeles. 158



Computational Approaches to Dialogue

Poesio, M. (1995), A model of conversation processing based on micro conversational events. In Proceedings of the Annual Meeting of the Cognitive Science Society, Pittsburgh. Poesio, M. & Traum, D. R. (1995), A multi-purpose model of conversational context. In IJCAI-95 Workshop on Context in Natural Language Processing, Montreal. Poesio, M. & Traum, D. R. (1997), Conversational actions and discourse situations. Computational Intelligence 13(3), 309–347. Poesio, M. & Traum, D. R. (1998), Towards an axiomatization of dialogue acts. In Proceedings of Twendial’98, 13th TwenteWorkshop on Language Technology: Formal Semantics and Pragmatics of Dialogue. Raux, A., Langner, B., Bohus, D., Black, A. W. & Eskenazi, M. (2005), Let’s go public! Taking a spoken dialog system to the real world. Proceeding of the International Speech Communication Association. Rich, C., Sidner, C. L. & Lesh, N. (2001), Collagen: Applying collaborative discourse theory to human-computer interaction. Artificial Intelligence Magazine 22(4), 15–25. Roque, A. & Traum, D. (2007), A model of compliance and emotion for potentially adversarial dialogue agents. In Proceedings of the 8th SIGdial Workshop on Discourse and Dialogue, Belgium. Roque, A. & Traum, D. (2008), Degrees of groundedness based on evidence of understanding. In The 9th SIGdial Workshop on Discourse and Dialogue, Columbus. Roque, A. & Traum, D. (2009), Improving a virtual human using a model of degrees of grounding. In Proceedings of International Joint Conference on Artificial Intelligence IJCAI-09, Pasadena. Roque, A., Leuski, A., Rangarajan, V., Robinson, S., Vaswani, A., Narayanan, S. & Traum, D. (2006), Radiobot-cff: A spoken dialogue system for military training. In Proceedings of the 9th International Interspeech Conference, Pittsburgh. Sadek, M. D., Ferrieux, A. & Cozannet, A. (1994), Towards an artificial agent as the kernel of a spoken dialogue system: A progress report. In Proceedings of the AAAI Workshop on Integration of Natural Language and Speech Processing, Seattle. Schegloff, E. A. & Sacks, H. (1973), Opening up closings. Semiotica 7, 289–327. Schegloff, E. A., Jefferson, G. & Sacks, H. (1977), The preference for self correction in the organization of repair in conversation. Language 53, 361–382. Searle, J. R. (1969), Speech Acts. Cambridge University Press, New York. Sinclair, J. M. & Coulthard, R. M. (1975), Towards an Analysis of Discourse: The English Used by Teachers and Pupils. Oxford University Press, Oxford. Smith, R. W. & Hipp, D. R. (1994), Spoken Natural Language Dialog Systems: A Practical Approach. Oxford University Press, Oxford. Smith, R. W., Hipp, D. R. & Biermann, A. W. (1995), An architecture for voice dialogue systems based on prolog-style theorem proving. Computational Linguistics 21(3), 281–320. Stevens, A., Hernandez, J., Johnsen, K., Dickerson, R., Raij, A., Harrison, C., DiPietro, M., Allen, B., Ferdig, R., Foti, S., Jackson, J., Shin, M., Cendan, J., Watson, R., Duerson, M., Lok, B., Cohen, M., Wagner, P. & Lind, D. (2006), The use of virtual patients to teach medical students history taking and communication skills. American Journal of Surgery 191(6), 806–811. Sutton, S., Novick, D. G., Cole, R. A. & Fanty., M. (1996), Building 10,000 spoken-dialogue systems. In Proceedings of the 4th International Conference on Spoken Language Processing (ICSLP-96), Pittsburgh. Swartout, W., Artstein, R., Forbell, E., Foutz, S., Lane, H. C., Lange, B., Morie, J., Noren, D., Rizzo, A. & Traum, D. (2013), Virtual humans for learning. AI magazine; Special issue on Intelligent Learning Technologies 34(4), 13–30. Traum, D. R. (1994), A computational theory of grounding in natural language conversation. PhD thesis, Department of Computer Science, University of Rochester. (Also available as TR 545, Department of Computer Science, University of Rochester.) Traum, D. (2004), Issues in multi-party dialogues. In F. Dignum, ed., Advances in Agent Communication, LNAI 2922, Springer Verlag, 201–211. Traum, D. (2012), Non-cooperative and deceptive virtual agents. IEEE Intelligent Systems: Trends and Controversies: Computational Deception and Noncooperation 27(6), 66–69. 159

David Traum

Traum, D. R. & Allen, J. F. (1992), A speech acts approach to grounding in conversation. In Proceedings 2nd International Conference on Spoken Language Processing (ICSLP-92), Banff, 137–140. Traum, D. R. & Allen, J. F. (1994), Discourse obligations in dialogue processing. In Proceedings of the 32nd Annual Meeting of the Association for Computational Linguistics, New Mexico, 1–8. Traum, D. R. & Rickel, J. (2002), Embodied agents for multi-party dialogue in immersive virtual worlds. In Proceedings of the First International Joint conference on Autonomous Agents and Multiagent systems, Bologna, 766–773. Traum, D. & Larsson, S. (2003), The information state approach to dialogue management. In J. van Kuppevelt & R. Smith, eds, Current and New Directions in Discourse and Dialogue, Kluwer, The Netherlands, 325–353. Traum, D. R. & Morency, L. P. (2010), Integration of visual perception in dialogue understanding for virtual humans in multi-party interaction. In AAMAS International Workshop on Interacting with ECAs as Virtual Characters, Toronto. Traum, D. R., Robinson, S. & Stephan, J. (2004), Evaluation of multi-party virtual reality dialogue interaction. In Proceedings of Fourth International Conference on Language Resources and Evaluation (LREC 2004)8888, Lisbon, 1699–1702. Traum, D., Swartout, W., Marsella, S. & Gratch, J. (2005a), Fight, flight, or negotiate: Believable strategies for conversing under crisis. In Proceedings of the Intelligent Virtual Agents Conference (IVA), Springer-Verlag Lecture Notes in Computer Science, 52–64. Traum, D., Swartout, W., Marsella, S. & Gratch, J. (2005b), Virtual humans for non-team interaction training. In Proceedings of the AAMAS Workshop on Creating Bonds with Embodied Conversational Agents, Utrecht. Traum, D. R., Marsella, S., Gratch, J., Lee, J. & Hartholt, A. (2008), Multiparty, multi-issue, multistrategy negotiation for multi-modal virtual agents. In H. Prendinger, J. C. Lester & M. Ishizuka, eds, IVA, Vol. 5208 of Lecture Notes in Computer Science, Springer, New York, 117–130. Traum, D., Roque, A., Leuski, A., Georgiou, P., Gerten, J., Martinovski, B., Narayanan, S., Robinson, S. & Vaswani, A. (2007), Hassan: A virtual human for tactical questioning. In The 8th SIGdial Workshop on Discourse and Dialogue, Belgium. Traum, D., Jones, A., Hays, K., Maio, H., Alexander, O., Artstein, R., Debevec, P., Gainer, A., Georgila, K., Haase, K., Jungblut, K., Leuski, A., Smith, S. & Swartout, W. (2015), New dimensions in testimony: Digitally preserving a holocaust survivor’s interactive storytelling. In H. Schoenau-Fog, L. E. Bruni, S. Louchart & S. Baceviciute, eds, Interactive Storytelling, Vol. 9445 of Lecture Notes in Computer Science. Springer, New York, 269–281. Tur, G., Stolcke, A., Voss, L., Peters, S., Hakkani-Tur, D., Dowding, J., Favre, B., Fernandez, R., Frampton, M., Frandsen, M., Frederickson, C., Graciarena, M., Kintzing, D., Leveque, K., Mason, S., Niekrasz, J., Purver, M., Riedhammer, K., Shriberg, E., Tien, J., Vergyri, D. & Yang, F. (2010), The calo meeting assistant system. In IEE Transactions on Audio, Speech, and Language Processing 18(6), 1601–1611. Turing, A. M. (1950), Computing machinery and intelligence. Mind 59, 433–460. Retrieved from: http://cogprints.org/499/ Walker, M. A., Fromer, J. C. & Narayanan, S. (1998), Learning optimal dialogue strategies: A case study of a spoken dialogue agent for email. In Proceedings of the 36th Annual Meeting of the Association for Computational Linguistics and the 17th International Conference on Computational Linguistics, Montreal. Walker, M. A., Passonneau, R. & Boland, J. E. (2001), Quantitative and qualitative evaluation of darpa communicator spoken dialogue systems. In ACL ’01: Proceedings of the 39th Annual Meeting on Association for Computational Linguistics. Association for Computational Linguistics, Morristown, 515–522. Wallace, R. S. (2009), The anatomy of a.l.i.c.e. In R. Epstein, G. Roberts & G. Beber, eds, Parsing the Turing Test. Springer, Netherlands, 181–210.

160



Computational Approaches to Dialogue

Weigand, E. (2006), Argumentation: The mixed game, Argumentation 20(1), 59–87. Weizenbaum, J. (1966), ELIZA—a computer program for the study of natural language communication between man and machine. Communications of the Association for Computing Machinery 9(1), 36–45. Williams, J. D. & Young, S. (2007), Partially observable Markov decision processes for spoken dialog systems. Computer Speech and Language 21(2), 393–422. Zinn, C., Moore, J. D. & Core, M. G. (2002), A 3-tier planning architecture for managing tutorial dialogue. In S. A. Cerri, G. Gouardéres & F. Paraguaçu, eds, Intelligent Tutoring Systems, Vol. 2363 of Lecture Notes in Computer Science, Springer, New York, 574–584.

161

10 From Speech Act Theory to Dialog Dialog Grammar Sebastian Feller

10.1 Introduction For the better part of the twentieth century, linguists predominantly engaged in the study of a presupposed, artificial linguistic system. Linguists viewed language as strings of words arranged in accordance with a quasi-logical order. In this kind of Positivist spirit, they, likewise, analyzed word meaning by breaking down words into artificially constructed building blocks which they called semantic primitives. They ascribed truth values to sentences, and represented the syntactical make-up of a sentence along chains of computer code-ish symbolic representations. The basic idea behind all this was the view that language were a rule-governed system that was describable in terms of clear-cut categories, quasi-logical mechanisms and pre-defined unitary boundaries. Toward the second half of the twentieth century the system view of language increasingly came under heavy fire. A new force emerged in the linguistic world that took a very clear stand against the system view and its theoretical implications. This movement came under a new name: pragmatism. In the late 1960s, early 1970s, the pragmatic turn brought a fresh start to the study of language. The object of study was no longer a system; in contrast, linguists were more and more interested in how speakers use language in real life. Questions like ‘What does the speaker intend by saying something?’, ‘What kind of situational conditions should be met by a speech act to be sound?’, and ‘What kind of effects does a speech act cause in the hearer?’ opened up a new path of linguistic enquiry. The rule concept, which had been in full bloom under the system view, was now increasingly replaced by a comparatively more “natural” concept: action. Austin’s How to Do Things with Words (1962) is seminal in this context. One path of linguistic enquiry under the pragmatic turn consisted in the insight that people use language as a means to achieve specific goals; in other words, by and through language, people do things. And the larger setting for communicative action is dialog (see also Chapter 5, this volume). Under this new linguistic program an entirely new linguistic camp emerged: Speech Act Theory. It is, however, somewhat misleading to talk about Speech Act Theory in the singular. Considering the developments in the field, Speech Act Theories is certainly a much better way to put it. The theory has undergone a few significant changes from its beginnings 162



Dialog Grammar

to the current state-of-the-art. To put it briefly, it has changed from a monologic theory of isolated speech acts to a dialogic theory of language games. We will see later on what this means exactly. This chapter is mostly about selected developments of Speech Act Theory. I will highlight a number of pivotal theoretical contributions and paradigmatic changes in the field. Special focus will be on the turning point from the monologic toward the dialogic study of speech acts. As we will see later on, Hundsnurscher’s Dialog Grammar played a special part in this. It was Dialog Grammar which paved the way toward a genuinely dialogic version of Speech Act Theory. This said, Dialog Grammar has a special place in this chapter. But before we dive into this journey, let me first of all provide an overview of what lies ahead of us in the sections to come. Section 10.2 deals with traditional Speech Act Theory. Searle’s work is seminal for this part. The main focus will be on two of his book publications: Speech Acts (1969) and Expression and Meaning (1985). This is followed by an overview of Hundsnurscher’s (e.g., 1980, 2000) linguistic program in Section 10.3: Dialog Grammar. As I will discuss in more detail later, Hundsnurscher takes an important first step toward a dialogic version of Speech Act Theory by putting Searle’s monologic speech acts into a dialogic framework. Section 10.4 introduces a further significant change in the treatment of language: that from a linguistic model of speaker intentions toward a model of social interactions. It is Habermas’ sociological model of language use and argumentation (e.g., 1984, 1987) combined with Habermas and Fultner’s (2002) approach to pragmatics. Section 10.5 deals with the development of a dialogic Speech Act Theory. In this context, I will briefly trace the developments of Weigand’s linguistic theory. The discussion focuses on two of her books: Sprache als Dialog (2003 [1989]), in which she lays out her original conventional theory, and the volume, Dialogue: Mixed Game Model (2010), her revised theory of dialog based on principles of probability.

10.2 The Traditional Speech Act Theory: Language as a System of Constitutive Rules Before we deal with language, we will first have a look at games. This is because playing a game is in many ways very similar to how we use language. Although an oversimplification, the game analogy helps us understand some of the core aspects of language use. Let us take soccer as an example. The players in a soccer game act in very specific ways. For example, they move within the confines of the soccer pitch, they dribble and pass the ball with their feet, they shoot corner kicks, free kicks and penalties from specific location on the soccer pitch, and they shoot goals by kicking the ball into the goal with their foot moving within the confines of the pitch. It becomes immediately obvious that the behavior we observe on the soccer pitch is regulated in some way. In other words, soccer has rules and each player needs to act in accordance with these rules in order to participate in the game. In the case of rule violation, sanctions are enacted which regulate the manner and degree of punishment. Austin’s (1962) and, later, Searle’s (1969) theory of speech acts build on exactly this analogy. To be precise, in Speech Acts (1969: 52) Searle refers to American football, not soccer. One of the basic assumptions of traditional Speech Act Theory is that language use, just like games in general, is built on a set of constitutive rules (Searle, 1969: 33ff.). These rules are defined, sanctioned and applied by a group of speakers. As a consequence, the actions the speakers are allowed to carry out in the ‘language game’ are rule-governed, just like the actions the players carry out in a soccer game. To give an example, consider the 163

Sebastian Feller

action

action conditions

action means

action goals

Figure 10.1  Action Structure.

speech act of promising: when making a promise the speaker acts along a very specific set of rules that constitute her social behavior in a typical promising situation. This also applies to the way she uses language. The promiser obliges herself to the carrying out of a certain action by uttering a specific sequence of words in a situation that is deemed adequate as the social context of a promise. The rules and conditions under which speech acts are carried out are regarded as “institutional concepts” (Searle, 1969: 52). It is, at the same time, important to note that the institutional concepts imply conditions that lie beyond language per se. As with the speech act of promising, most speech acts imply a certain state of affairs that should be met in the specific speech situation: for instance, for some speech acts, the dialog partners must possess the right status or authority. In other cases, the soundness of a speech act is bound to the right location. For example, a verdict should be uttered by a judge in a court house, not by a salesperson in a shopping center. In comparison to other theories of language use which hold the constitutive rules to be of an a priori nature (e.g. Reinach, 1913), Austin and Searle consider them social constructs or conventions. The members of a speaker community do certain things in certain ways. Over a period of time habits are formed and expectations on how people act in certain situations arise from the re-occurring behavior. This has nothing to do with logical absoluteness or a priori givens. In contrast, an act like promising is analyzed into a set of socially formed, man-made rules (see Searle, 1969: 57) based on habitual social behavior: • • • •

The speaker expresses the proposition that p in the utterance T. In expressing that p, S predicates a future act A of S. S intends to do A. S intends that the utterance of T will place him under an obligation to do A.

One of the defining characteristics of the traditional Speech Act Theory is the monologic treatment of speech acts. Austin (1962) hinted at a dialogic conception of the theory by his concept of ‘take up’, that is, the receiving of a speaker’s speech act. However, this concept has not been followed up by Searle. In contrast, Searle treats speech acts as isolated actions of an individual speaker. The main question under the Searlian program is how linguistic form relates to communicative function; in other words, what expressions do speakers use in order to instantiate a specific speech act. As will be discussed in the next section, this view was increasingly called into question later on during the pragmatic turn. In response to traditional Speech Act Theory and also the sociolinguistic program under conversation analysis (Sacks et al., 1974), Franz Hundsnurscher developed a new theory of language: Dialog Grammar. Dialog Grammar starts off from the speech act as part of a larger speech act sequence. 164



Dialog Grammar

10.3  Dialog Grammar Before we dive into Dialog Grammar, we take a short detour through conversation analysis (CA). The reason for this is that Hundsnurscher introduced Dialog Grammar as a sort of counter-program to the heavily data-driven research program of CA. The basic assumption of CA is that the study of language should concentrate on the study of communicative performance exclusively, that is, the linguistic data produced by speakers in and by communication. Following this strictly empirical doctrine, authentic conversations are transcribed into transcripts that provide an overview of who said what to whom and when. Detailed analyses include International Phonetic Alphabet (IPA) transcriptions and even notations for paralinguistic features, among others. For example, diacritics might be used to indicate features like word stress, speaker intonation and unintelligible speaker output. Conversation analysis wants to account for the expression side of language use displayed in communicative interaction. In the end, its methodology delivers meticulous snap-shots of how selected interlocutors use linguistic expressions in specific conversations. As Psathas (1995: 3) puts it: The ethnomethodological character of conversation analysis should be clear. Social actions in the world of everyday life are practical actions and are to be examined as ongoing practical accomplishments. The logic or organization of such actions is a practical logic, an achieved organization, locally produced, in situ, in the “there and then” and the “here and now”. Later versions of CA attempted to draw generalizations emerging from the in situ data. The turn-taking ‘theorem’ is an example in point. The study of turn taking forms a branch of CA in its own right, with linguists forming generalized language- and culture-specific turn-taking models derived from observed turn-taking behavior in singular conversational instances. Although, at first sight, it might make sense to take a close look at the expression side, CA is trapped in its own methodology. As Hundsnurscher contends: Although it might be very interesting and sometimes even entertaining to investigate the various forms and ways in which people talk to each other, as linguists we cannot content ourselves with just picking out pieces of authentic verbal interaction, describing and analyzing individual aspects to a certain point, and then going on to other pieces of authentic talk in different situations and pointing out yet other coincidental phenomena. (Hundsnurscher, 2007: 61) As a consequence, in his countermove to this type of performance-oriented linguistics, Hundsnurscher shifts the focus back toward communicative competence. In Section 10.2, we have already talked about rules and institutional concepts that underlie traditional Speech Act Theory. As we have seen, the rules and conventions serve to give some kind of order to the seemingly overwhelming complexity of linguistic data. Interestingly, Hundsnurscher applied similar assumptions to his linguistic program and went even a bit further by leaning on assumptions that stem from the competency-driven research program of Generativism. It is especially the distinction between competence and performance which Dialog Grammar adopts. Hundsnurscher’s main assumption is that language use is governed by a formal system of speech act sequences (e.g., 2000: 945; 2007: 62). The speech act 165

Sebastian Feller ISPA

ReSPA

Figure 10.2  Minimal Speech Act Sequence.

sequences are organized on the level of deep structure modeling the speaker’s communicative competence. The deep structure predicts the legitimate organization of speech acts into well-formed dialogs that are realized through the speakers’ performance on the surface level. Following Austin’s action-centered view of language use, Hundsnurscher (2007: 62) builds his own theory around the action concept. In communication, the speaker carries out communicative actions that, according to Hundsnurscher, consist of three basic building blocks (see Figure 10.1): Hundsnurscher’s action conditions resemble the soundness conditions we find in traditional Speech Act Theory. Similarly, [means] and [goals] are old acquaintances that were already introduced by Austin and Searle. Hundsnurscher (2007: 62) points to the parallelism himself: “Austin’s (1962) and Searle’s (1969) concept of a speech act can be modeled along these lines [of the triple structure] and this will yield the general structure of speech acts or verbal actions in general […]”. However, Dialog Grammar and Speech Act Theory differ in one major aspect: the basic unit of investigation. While Speech Act Theory investigates single sentences in isolation as instantiations of speech acts, Dialog Grammar starts off from a minimal speech act sequence (see Figure 10.2) (Hundsnurscher, 2000: 945). In the minimal speech act sequence, speaker 1 (SP1) performs an initial speech act (ISPA), in response to which speaker 2 (SP2) carries out the reactive speech act (ReSPA). Importantly, in Dialog Grammar, the ReSPA is not a functionally separate speech act. It is solely structurally defined, namely by its position in the speech act sequence. Hundsnurscher argues that there are two basic types of speech act sequences: either SP2 uses a positive ReSPA with respect to SP1’s initiation, or she rejects the initial speech act by a negative ReSPA. While the former sequence is considered closed, the latter is called open. In an open sequence there is no consensus between the speakers. As a result, the sequence may carry on until a consensus is finally in place or one of the speakers backs out. Hundsnurscher argues that the speech act sequences are specific to the illocution of the initial speech act ISPA. In this sense, “we can construct illocution-specific dialogue sequences” (Hundsnurscher, 2007: 68). In the following sections, we will discuss two selected examples of such sequences with a view to arriving at a better understanding of how the speech act sequences unfold on the level of deep structure.

10.3.1 Argumentation An argumentation sequence requires the following two conditions to be met (see Hundsnurscher, 2007: 68): 1. The ISPA is a representative speech act, i.e. the speaker makes a claim to truth. 2. The ReSPA is a negative answer (NA), i.e. the dialog partner rejects the ISPA.

Consider the following dialog which I witnessed the other day in a classroom discussion between two students: 166



Dialog Grammar SP1

SP2

pos. R.

SP1

#

ISPA

Resign

#

neg. R.

INSIST

Figure 10.3  INSIST Structure. Based on Hundsnurscher, 2007: 69.

SP1: “Electric cars help save the environment.” [CLAIM] SP2: “That’s not true.” [NA] SP1: “Yes, it is true. They make no noise and give off no exhaust fumes.” [INSIST]

In this type of conversation the dialog partners act along the so-called INSIST-structure: As Figure 10.3 shows, the ReSPA is the starting point for a new minimal sequence: the NA by SP2 serves as a new ISPA, which is taken up by SP1 through the uttering of the subsequent ReSPA. The ReSPA may, at the same time, initiate another response by SP2. The exchange usually continues until either both speakers reach a consensus or one speaker backs out, thus breaking the conversation off. Hundsnurscher (2007: 70) argues that, next to the INSIST pattern, other patterns do exist. ISPAs with illocutions other than the representative lay the starting point for distinct types of speech act sequences. The following section provides an example in point.

10.3.2  Bargaining Dialog A typical bargaining speech act sequence deviates from the argumentation-specific INSIST pattern. A bargaining dialog takes off from an initial speech act by which the speaker presents an offer. In the ideal case, the offer is immediately accepted by the dialog partner, thus concluding the bargaining dialog. The ideal or, as Hundsnurscher (2001: 78) names it, the well-formed bargaining dialog, consists of a minimal speech act pair (see Figure 10.4). Here an example, again from my own recollection: SP1: “I give you 3 roses for 4 Pounds.” SP2: “No, I give you 3 Pounds, no more.” SP1: “How about 3.50?” SP2: “Okay, deal.” 167

Sebastian Feller ISPA

PA (positive answer)

Figure 10.4  Ideal Bargaining Dialog.

It is important to note that both ISPA and PA (Positive Answer) can be instantiated in ­multiple different ways: as for the PA, for example, besides uttering a verbal response, Dialog Grammar accounts for the fact that the dialog partner might express her acceptance also by nonverbal means including, for instance, body language (e.g., nodding or hand shaking) or a direct action like handing over money (Hundsnurscher, 2001: 78). The opening up of the set of communicative means that is beyond verbal means of expression is thereby an important advancement compared to traditional Speech Act Theory. The comprehensive integration of communicative means in Dialog Grammar is facilitated by the dialogic point of view: many nonverbal communicative means of the ReSPA become interpretable once considered from a dialogic point of view. Take a handshake, for example. By itself a handshake can have many different meanings: it might be used for greeting, saying goodbye, accepting, or promising, among others. In the light of a preceding offer, however, it becomes clear that the handshake instantiates a PA to an offer. It is the ISPA which specifies the communicative meaning of the handshake as the accepting of the speaker’s preceding offer. In the case that the dialog partner does not accept the speaker’s offer, the dialog moves into a new phase. Now SP1 can either accept or reject the dialog partner’s NA. By accepting the rejection, the speaker brings the dialog to an end. She gives in to the dialog partner’s negative reaction, signaling that her offer has failed and that the outcome she desires cannot be reached. On the other hand, the speaker might likewise react negatively toward the rejection: she might then put forward a modified offer in her attempt to persuade the dialog partner to give in. From this point onwards, the dialog sequence runs along a recursive path, where both dialog partners might cling on either until an agreement has been reached or one of the dialog partners finally gives in or backs out ultimately.

10.3.3  The Issue of a Dialog Typology The illocution-specific speech act sequences provide a starting point for developing a dialog typology (see Hundsnurscher, 1994). As discussed earlier, according to Hundsnurscher, it is the initial speech act of the speech act sequence which determines the exact type of that sequence. For example, • • • • •

an assertion might give rise to an argumentation sequence, an offer to a bargaining sequence, a suggestion to a checking sequence, a request to an enforcing sequence, advice seeking to a problem-solving sequence (Hundsnurscher, 1994: 220f.).

Besides this illocution-specific typology, Hundsnurscher (1994: 221) offers an alternate model which takes the level of cooperation between the dialog partners as its starting point for the sequence classification. Hundsnurscher distinguishes here between dialogs with convergent versus divergent speaker interests at the most basic level of differentiation. The typology branches out into dialogic subtypes including supplementary and contributory dialogs under convergent, that is dialogs with the purpose of providing information 168



Dialog Grammar

and facilitating problem solving (Hundsnurscher, 1994: 223). Divergent dialogs are further ­analyzed into conciliatory and intransigent dialogs. The former are characterized by both dialog partners having different expectations at the onset of the dialog, while trying to arrive at a consensus toward the end. In contrast, intransigent dialogs exclude consensus from the very beginning. This might be the case in a situation where one dialog partner cannot give in under any circumstances. Hundsnurscher refers to an officer or principal “who has to see to it that orders or refutations are being strictly obeyed and adhered to” (Hundsnurscher, 2007: 74) as an example in point. With his Dialog Grammar Hundsnurscher paves the way toward a theory of interactive language use. Looking at generalizable dialog patterns beneath the empirical conversational data, he explicitly rejects a purely empirically based linguistic analysis. At the same time, the patterns he is after are not strictly defined patterns of logic; in contrast, Dialog Grammar is based on the idea of conventions and rules that speakers apply in authentic language use. Hundsnurscher is caught between two stools: the logic-driven hardliner Generativist and the data-driven conversation analyst.

10.4 From Intentions to Interactions: Habermas’ Theory of Communicative Action Jürgen Habermas’ (e.g., 1984, 2002) theory of communicative action is a sociological theory of social interaction. The central concept for Habermas is cooperation, which itself is based on two concepts: voluntary action and equality in communication. The basic question of his linguistic program is how communication might enable people to live together in ways that ensure maximum fairness and equality in society. For Habermas (1984: 19), the key to an answer lies in the concept of understanding. In conversation, the speakers should communicate in comprehensive and unrestricted ways that enable them to reflect on the living conditions and living environments they are in. They should call their living conditions into question and, by and through argumentation, refine the values and norms they live by. Habermas suggests that speakers act along the following four conditions with a view to arriving at an understanding (see Forester, 1987: 209): 1. 2. 3. 4.

speak comprehensively, speak truthfully with regard to the factual states in the world, communicate adequately with respect to the world of social relations, and speak truthfully with regard to one’s intentions and emotions.

For Habermas, understanding is the result of negotiation processes between the dialog partners that take place during argumentative discourse. Argumentative discourse, on the other hand, runs along the following lines (see Colclasure, 2011: 8): 1. process: the context and the nature of the action; the goal is to persuade/convince the dp (dialog partner), 2. procedure: rules of interaction; the goal is to reach a consensus, and 3. product: the structure of arguments; the goal is to have one’s arguments accepted.

The ultimate aim of the argumentative discourse is the rationalizing of the interlocutors’ living environments. In other words, the more people communicate about the world they live 169

Sebastian Feller

in including the given state-of-affairs, norms and, not to forget, their own inner states, the more rational the world they live in becomes. According to Habermas, communication is the continuous testing, probing and negotiating of the validity of the status quo. As mentioned above, in Habermas’ view (1987, VI/1E.), communication is, in general, cooperative. Every speech act presupposes a set of validity claims that imply a certain reaction on behalf of the dialog partner. This interactive claim is a fundamental change away from traditional intentional Speech Act Theory in the fashion of Austin and Searle. Against this background, the treatment of speech acts as isolated units of communication does not make much sense anymore. The kind of cooperative rationalization of the interlocutors’ living worlds boils down to a competition between arguments. All dialog partners ought to be able to put forward their arguments and, likewise, react to what has been said by the other dialog partners. The competing arguments are not judged primarily with regard to their logical soundness but in the light of the rhetorical impact they have. In order for a judgment to be sound, Habermas (2008) suggests that the argumentation runs in accordance with the following criteria: 1. no one capable of making a relevant contribution has been excluded, 2. participants have equal voice, 3. they are internally free to speak their honest opinion without deception or self-deception, and 4. there are no sources of coercion built into the process and procedures of discourse.

It is obvious that these criteria describe an ideal speech situation, which will hardly ever come about in real-life language use. In real-life conversations, it is quite often the case that the dialog partners are in an unequal relationship where some kind of social or communicative coercion is in place. Nonetheless, Habermas’ criteria provide a set of guidelines or, in a way of speaking, communicative commandments which, if followed, make the communicative interaction more rational and objective. Zooming in on Habermas’ theory of discourse, we find three basic types of communicative action he distinguishes (Habermas, 1984: 275ff.) 1. constative: making a claim about the world, 2. regulative: ordering, requesting, commanding, 3. expressive: making a claim about one’s inner states.

Taking these three types of action as a first starting point, it shows that speakers use language with respect to three partly integrated discourse contexts: the outer world, the inner world, and the normative world. In addition, the liability of the speech act depends on its level of institutionalization. An institutionally bound speech act derives its liability from the norms and values it inherits from the institution, while institution is roughly defined here along Habermas’ (2008) criteria (1) to (4) mentioned earlier. On the other hand, a speech act that is per se not bound to an institutional context receives its liability from the interlocutors themselves. The interlocutors should ground their communication on a rational foundation which enables them to become cooperative partners in the communication. Habermas’ interactive view of communication is an important step on the way toward a genuinely dialogic perspective of language-in-use in linguistics. However, as most of his 170



Dialog Grammar

theorizing is based on the idea of an idealized version of communication, some of the realities of authentic language use are still ignored.

10.5  Tying Up Loose Ends: Weigand’s Theory of Dialog The preceding sections have drawn a clear picture: Speech Act Theory has come a long way from the study of isolated speech acts towards the analysis of larger speech act sequences and interactive discourse structures. Hundsnurscher’s theoretical program plays an important part in this: Dialog Grammar can only be fully understood in the light of the empirical entrenchment of conversation analysis. One way of seeing it is that Dialog Grammar restored the balance in a time where the empirical study of linguistic data became somewhat of a fetish in linguistics. Instead of scrutinizing each data point with a magnifying glass, Hundsnurscher sought to uncover the blueprints of a set of basic dialogs beneath the surface phenomena of authentic language use. Nonetheless, competency-driven theories like Dialog Grammar raise some doubt: does this underlying system of rules really exist? In other words, are dialogs actually arranged after generalizable patterns of communication stored in the hidden deep structure of linguistic competence? If so, how can the gap between the idealized output of communicative competence and the relatively chaotic authentic language use be bridged? In her theory of dialog, Edda Weigand tackles these and other questions from an entirely new perspective (see Weigand, 2010; also Weigand, 2000, 2003): ‘competence-in-performance’. Her theory is partly grounded in Dialog Grammar; however, she introduces a number of concepts which point in a different direction. Weigand merges communicative competence and communicative performance into one. In her opinion, the gap between these two opposites “turns out to be a methodological pseudo-problem” (2010: 5) stemming from a basic misconception of linguistic theorizing: the object of study is cut out after methodological assumptions. Instead, Weigand considers the object of study, that is, natural and authentic language-in-use, to be the correct starting point for linguistic theorizing. While in her early work, which is laid out in detail in Sprache als Dialog (2003 [1989]), Weigand still applied the idea of well-formed speech act sequences, she later revised this point: the methodology should actually be derived from the object, that is language use, not the other way around. Accordingly, she argues that, in language use, there simply is no distinction between an idealized competence and a seemingly chaotic performance. The human mind operates along principles of probability, granting us leeway in how we apply the norms and conventions of language use (Weigand, 2010: 71ff.). The probability assumption is underscored by Weigand’s shift from convention to expectations. While in Sprache als Dialog she sees language use grounded in social conventions, she later holds that the dialog partners operate on expectations about what should be happening in communication. The probability assumption of Weigand’s theorizing touches on another essential distinction of the theory. While Dialog Grammar holds that the sets of communicative means the speakers can draw on are limited, Weigand argues for open-ended sets of means (e.g. Weigand, 2009: 147, 355). In language use, the speakers can deviate from the norm at any time. They can invent entirely new ways of communicating, while the goal of coming to an understanding might still be reached. Given the complexity of natural language use, determining the set of possible correlations of communicative goals and means results in

171

Sebastian Feller

combinatorial explosion. From a theoretical point of view, assuming open-ended sets of means is thus a plausible move. Another important assumption of Weigand’s Mixed Game Model is that speakers integrate different abilities and means in communication. They use verbal, perceptual and cognitive means and combine them in meaningful ways in order to construct meaning and arrive at an understanding (Weigand, 2010: 87). Weigand’s redefinition of linguistic theorizing has paved the way for new theoretical endeavors in linguistics in general and Dialog Analysis in particular. These include Feller’s Meaning-in-Use (2010), which redefines meaning in terms of an action-theoretical view of language, as well as new, dialog-based explorations in the area of teacher-learner-interaction-research (e.g. Feller, 2015, 2014, 2013).

References Austin J (1962) How to Do Things with Words. Oxford: Oxford University Press. Colclasure D (2011) Habermas and Literary Rationality. New York: Routledge. Feller S (2015) Dialogic knowledge building in learning communities: Discovering an electric circuit through deep reasoning. In Koike D and Blyth C (eds) Dialogue in Multilingual and Multimodal Communities. Amsterdam: Benjamins, 195–220. Feller S (2014) Designing a cognitive speech act taxonomy for dialogic teaching and learning: Explorative action games for conceptual change learning. Journal of Language Teaching and Research 5(3): 524–535. Feller S (2013) Language for deep learning: Cognitive explorative action games for teacher-learner interactions. Journal of Cognitive Science 14(4): 361–378. Feller S (2010) Lexical Meaning in Dialogic Language Use. Amsterdam: Benjamins. Forester J (1987) Critical theory and planning practice. In Forester, J (ed) Critical Theory and Public Life. Cambridge, MA: MIT Press, 202–230. Habermas J (1984) The Theory of Communicative Action, Vol. I. Reason and the Rationalization of Society. McCarthy T (trans.). Boston: Beacon. Habermas J (1987) The Theory of Communicative Action. Vol. 2. Lifeworld and System: A Critique of Functionalist Reason. Cambridge: Polity Press. Habermas J (2008) Between Naturalism and Religion. Cambridge: Polity Press. Hundsnurscher F (1980) Konversationsanalyse versus Dialoggrammatik. In Rupp H and Roloff HG (eds) Akten des VI. Internationalen Germanisten-Kongresses, Basel 1980, Teil 2. Bern, Frankfurt, 89–95. Hundsnurscher F (1994) Dialog-Typologie. In Fritz G and Hundsnurscher F (eds) Handbuch der Dialoganalyse. Tübingen: de Gruyter, 203–238. Hundsnurscher F (2000) Das Konzept der Dialoggrammatik. In Brinker K (ed) Linguistics and Conversation. Berlin: de Gruyter, 945–951. Hundsnurscher F (2001) The grammar of bargaining. In Weigand E and Dascal M (eds) Negotiation and Power in Dialogic Interaction. Amsterdam: Benjamins, 77–90. Hundsnurscher F (2007) The principles of dialogue grammar. In Berlin L (ed) Theoretical Approaches to Dialogue Analysis. Selected Papers from the IADA Chicago 2004 Conference. Tübingen: de Gruyter, 61–76. Psathas G (1995) Conversation Analysis: The Study of Talk-in-Interaction. Thousand Oaks, CA: Sage. Reinach A (1913) The a priori foundations of civil law. Crosby, J (trans.). Aletheia, 3 (1983): 1–142. Sacks H, Schegloff EA, and Jefferson G (1974) A simplest systematics for the organization of turntaking for conversation. Language, 50: 696–735. Searle JR (1969) Speech Acts: An Essay in the Philosophy of Language. Cambridge: Cambridge University Press.

172



Dialog Grammar

Searle JR (1985) Expression and Meaning: Studies in the Theory of Speech Acts. Cambridge: Cambridge University Press. Weigand E (2000) The dialogic action game. In Coulthard M, Cotterill J and Rock F (eds) Dialogue Analysis VII: Working with Dialogue. Tübingen: Niemeyer, 1–18. Weigand E (2003 [1989]) Sprache als Dialog. Tübingen: Niemeyer. Weigand E (2009) Language as Dialogue: From Rules to Principals of Probability, Feller S (ed) Amsterdam: Benjamins. Weigand E (2010) Dialogue: The Mixed Game Model. Amsterdam: Benjamins.

173

11 The Mixed Game Model A Holistic Theory Edda Weigand

11.1  The Issue In the preceding chapters of this Handbook we traced the development in modern linguistics from the beginning with de Saussure’s model of language as a system of signs to the concept of language as dialogue. Various turning points have marked this journey, the most decisive ones being those from the structural sign to the generative sentence and from the sentence to pragmatic language use. However, pragmatics still struggles to clarify its object ‘language use’. Can it be grasped by neo-Gricean principles, that is, by the addition of rule-governed patterns and individual inferences? The flight from logic and rationality to vague terms such as the ‘pragmatic perspective’ can be seen as an attempt to cover a field beyond patterns and rules. Recently a new term has even been created, the ‘pragmeme’ (e.g., Mey, 2001), which imitates structural terms such as ‘phoneme’ and ‘morpheme’ and promises some certainty other than the term ‘perspective’. However, this promise proved to be empty. The pragmeme only represents a label for classifying a construct which is defined by adding Searle’s abstract speech act (1969) to the context. Despite such a definition, it is then applied to any element of language use whatsoever (Allan et al., 2017), be it the speech act, the utterance, conversation or the dialogic language game. Language use cannot be equated with language put into context. Context is not an autonomous, separate object. Instead of acknowledging that ‘language use’ constitutes a new object-of-study which requires a new way of theorizing, it is again constructed artificially, implanting the error already at the very outset. When confronted with a complex object-of-study, science cannot start by setting up models and postulating that the object fits the model. In post-Cartesian times we intend to address complexity without distorting the object. Pure observation will not do nor will abstract constructions. Science starts with specific scientific interests which are different in different disciplines. If our scientific interest in linguistics goes beyond documenting and aims at explaining what language is about, observation needs to be goal-­oriented (Feynman, 2001) towards giving an answer to crucial questions such as why we use language, and how language use works, and whether there is an autonomous object language at all. 174



The Mixed Game Model: A Holistic Theory

What we first need when being confronted with complexity is to find a key to opening it up. This issue is especially difficult to settle if we ourselves are a component of the complex object. There is no language as such, it is human beings who use it. Anything we do is dependent on our own abilities. It is therefore only reasonable to look for the key in our own nature. What are we doing when we use language, and why are we using it? Due to our nature as dialogic individuals, we use language in order to communicate, and in communicating we aim at coming to an understanding with our fellow beings in ever-changing surroundings. Communication means dialogue, which results in a concept of language as dialogue. Dialogue is more than a verbal sequence of moves or turns, dialogue is a human need and implies a specific methodology of action and reaction as already Von Humboldt (1963 [1827]) emphasized. Dialogue therefore provides the methodological key with which to structure the ‘uncircumscribed perspective’ of pragmatics. We have thus arrived at the turning point from pragmatics to dialogue (see Chapter 5, this volume). Even strict pragmatists at last seem to acknowledge that language use does not stop at the limit of one utterance nor that Cartesian methods are apposite to it (Chomsky, 1966). After decades of going ahead separately, there is hope for a joint venture of pragmatics and dialogue analysis. Progress in research is not an issue of labels or terms but an issue of argumentation and verification. Finding out what language use is about necessitates breaking with Cartesian methodology of patterns of competence. Natural language use is a matter of performance. The issue is how to address performance in theory, how to deal with individuality and chance, with cognition and perception which obviously play a role in dialogue. Questions of this sort have to be taken into account. A first step, restricted to the empirical level, has been made by multimodal approaches (see Chapter 2, this volume). The empirical level of authentic language use is, however, only the tip of the iceberg. There is no direct correlation between individual empirical elements and dialogic meaning. The role of elements is revealed in the whole of the object. Meaning is in the end a matter of the insiders of the game insofar as it depends on their knowledge and individual worldview. In performance the correlation of expression and meaning can only be established on the basis of expectations. We as interlocutors expect certain events to happen and certain utterances to occur with a certain probability. When using the term ‘probability’ I am not referring to any mathematical calculation. I simply mean that human beings estimate the occurrence of certain events by calculating some degree of probability. The methodological basis of a theory of performance can therefore only be Principles of Probability. The change from the simple to the complex (Gell-Mann, 1994), from the Cartesian pattern view to post-Cartesian thought, poses a huge challenge to science. We can find some orientation in the natural sciences which have already passed this turning point in physics by making the change from classical physics to the modern physics of quantum theory. Recently in the humanities and social sciences such a turning point also seems to be gaining ground. Herbert Simon’s article on the ‘architecture of complexity’ of 1962 can be seen as a pilot study. Going beyond patterns and addressing complexity means changing theorizing from the classical view of reducing empirical variability to a rule-governed pattern to the modern holistic view of starting with the complex natural object and deriving components from it by specialization (Weigand, 2011). The notion of complexity depends on the object which is characterized as complex. Complexity might come about randomly or by the laws of nature, for instance, in a whirlpool. Complex systems of this type are the subject matter of William A. Kretschmar’s book on Language and Complex Systems (2015) which does not really deal with language but with speech as a variety of dialects. To my mind, the term ‘complex system’ presupposes 175

Edda Weigand

autonomy of the object. Speech, however, does not exist on its own. It is used by human beings according to their intentions and goals in a complex whole of dialogic interaction which is completely different from what Kretschmar calls a ‘complex system’. Whereas Cartesian reduction and abstraction led to searching in the ‘underground’, coming to grips with the complexity of performance requires a totally different procedure. As can be seen in the chapters of Part I, the issue has mainly been addressed by empiristic approaches or by artificially adding models of competence to those of performance. Both procedures contradict what we as human beings do in performance. At the level of heterogeneous models which, by definition, have nothing in common, the gap between rules of competence and the chaos of performance cannot be bridged. Human beings are, however, able to come to grips with performance by bridging the gap in their minds. They have an extraordinary ability which I have called competence-in-performance (Weigand, 2009) that enables them to adapt to ever-changing surroundings and to find ‘the laws’ of what might, at first glance, seem to be chaos (Prigogine, 1994). Empiristic approaches ignore the fact that pure observation will not decipher the chaos of authentic language use. Language cannot be equated with empirical details of spoken language. Human beings do not register any and every element but, on the contrary, pass over many empirical elements which are not really relevant for arriving at an understanding. The dogma of ‘data’ has already been rejected half a century ago by Chomsky (1959) in his brilliant critique of Skinner. Besides the dogma of ‘data’ and the dogma of ‘language as an abstract system’ there is a third dogma which determined modern linguistics, the dogma of the plurality of models. Accepting any model whatsoever means nothing other than rephrasing as strength what in reality is weakness. The plurality of models can be accepted as long as they are translatable or complementary. Science finally seems to have become aware of having relied on ‘dogmata’ and is searching for a new reasoned start.

11.2  What Constitutes a Theory In recent decades the term theory has been pervasively used in the literature without any reflection upon the conditions of its legitimate use. The reason might be that extending the object-of-study from an abstract concept to language-in-use resisted traditional methodology. The simplest way of avoiding the problem was to declare the ‘goal of a homogeneous discipline with a unified theory’ as ‘pointless’ and even to construct a whole sub-discipline, discourse analysis, on such fragile ground (e.g., Tannen, 1989: 7f.; Van Dijk, 1985). The first question we have to clarify is therefore the question of: ‘What constitutes a theory in general, and of performance in particular?’. A theory is first and foremost a theory about an object. In going beyond abstraction and focusing on language-in-use we have arrived ‘at the end of certainty’ (Toulmin, 2001). We are facing complexity, which does not mean chaos but replacing mathematical rationality by rationality of life (Weigand 2014). Pure rationality works in systems which are from the very outset based on logic. Such systems have pervaded modern linguistics from structuralism and generativism and have even reached out to pragmatics. What are pragmatic ‘theorists’ doing other than starting from abstract rational models and adding non-conventional inferences in ‘theories’ which as a whole are inconsistent and of little worth for human beings’ lives? Life is not logic but performance. Facing a complex object such as ‘language-in-use’ we cannot simply start with pure observation either and hope to be somehow enlightened by what we can record. Our 176



The Mixed Game Model: A Holistic Theory

object-of-study has to be a minimal autonomous object if we want to describe how it works. A theory is, however, expected not only to describe but also to explain its object-of-study. Explaining an object in the end means uncovering what it is used for. Why do human beings use language? In the literature we find different responses, among them the generative view, that we use language in order to express our thoughts or even in order to think. Finally, we seem to agree upon the view that language is primarily used for dialogic purposes. Coming to an understanding in dialogue is different from understanding insofar as it goes beyond ‘interpreting’ the speaker’s utterances and includes reacting to them, thus changing the hearer to an interlocutor. Action and reaction are not the minimal autonomous object yet. They are carried out by human beings by means of their abilities, in the end by their competence-in-performance, an innate ability which gets individually shaped by experience. Competence-in-performance allows human beings to go beyond rules and to adapt to the probabilities of performance. It is the challenge of life which has shaped our abilities, and it is the challenge of science to describe them and explain how they work. Fortunately there has been another turning point in science which helps us move ahead in the adventure of complexity: the brain is no longer a black box. We can finally verify our assumptions, at least to some extent, by reference to neurological experiments. Neuroscience has also detected the relationship between genetic evolution and the evolution of cultures by pointing to the coevolution of genes, mind and culture (Lumsden and Wilson, 2005). Evolution allows for individual differences in evaluating and learning in different social surroundings. Beside neurological research we need anthropological empirical research throughout the world in order to describe the particularities of multiple languages and cultures (Marchand, 2010). In the end, it is by means of sociobiology that we can verify our assumptions about human action and behaviour. The following figure sums up the components and their relationships in a post-Cartesian theory of human competence-in-performance. theory [descriptive and explanatory power, verified by sociobiology]

dialogue as natural object-of-study

®



methodology: principles of probability

scientific interest in how human beings can come to an understanding

Figure 11.1  Components of a Post-Cartesian Theory.

11.3  How to Cope with Complexity Having delineated in general the conditions a theory of performance has to meet, we can now address the issue of creating a holistic model which is capable of describing how human beings proceed in performance. The model I developed in recent years, the Mixed Game Model (MGM), is—as far as I see—the only genuinely holistic model available which aims to describe and explain the whole of language use in dialogic performance as it happens in different languages and cultures (Weigand, 2010). It is a model which outlines in theory how human competence-in-performance works. Cases which have traditionally been called ‘practice’, for instance, teaching, are included insofar as they represent a specific case of human competence-in-performance. 177

Edda Weigand

Why is the model called the Mixed Game Model? The feature ‘mixed’ characterizes the unit of the ‘action game’ which is based on the general assumption that language-inuse means action. The name ‘mixed game’ goes back to a discussion with members of the Consiglio Nazionale Forense in Rome in 2002, who got to the heart of the issue by calling the dialogic action game ‘un gioco misto’. The feature of all components being mixed in the complex—not separable or manageable one after the other—is indeed a basic feature of dialogic interaction in performance and a crucial feature of a holistic theory. Complexity does not mean the addition of multiple parts; complexity is based on the integration of components. Integration is the name of the game. Considering ‘language as dialogue’ rests on this premise: language is not separable but is always integrated with other human abilities. There is no language as such, there is only the human ability of speaking which is integrated with thinking, perceiving and feeling. Fortunately, this insight into the integration of human abilities which can be comprehended by an unbiased mind has finally been verified by experiments in neuroscience (e.g., Damasio, 1994). Abilities as such are not autonomous objects, they are dependent on the human brain, which not only encompasses what is innate but also what can be learned. We cannot go beyond the capacities of our brain. Absolute insights or absolute truths are not within the reach of human cognition. Learning, of course, depends on the cultural and social environment but in the end it depends on the innate ability of ‘prepared learning’ which is a prerequisite for the coevolution of genes, mind and culture (Seligman, 1970; Lumsden and Wilson, 2005). But not even human beings are independent beings. We are living in the world, not outside the world. There is no context to be added as assumed by pragmatic contextualists, for instance, Fetzer (2004), Finkbeiner et al. (2012) and others. The context, including the situation or genre, is always there from the very outset and affects our view of the world and the way we express ourselves. The unit is the action game, a mixed game, with human beings at its centre acting and reacting according to the conditions of the game and to their own individual worldviews. The term ‘game’, in the mixed game or action game, means a game like children’s games, a game of performance, not a chess game. The ‘architecture of complexity’ (Simon, 1962) is based on two fundamental premises. Besides the premise of the integration of the components there is another crucial premise: the whole is more than the sum of the parts. To be clear: there are no parts at all in the sense of separate parts, there are components which are integrated and thus create the complex which is more than the sum of its constituents. The components have to be derived from the complex whole. Derivation means specialization, which presupposes that the components are ‘nearly decomposable’. Complex hierarchies can thus be derived which are different from simple mathematical hierarchies, such as those introduced by Chomsky. The relation between constituents of different levels is not the relation of ‘being part of’, as in generative grammar, but of ‘being an integrated component’, that is, ‘derivative and nearly decomposable’. By this process of specialization all types of action games can be derived from the complex whole of ‘dialogue in the stream of life’, to use Wittgenstein’s metaphorical phrase (1981 [1967]). Specialization became necessary when communities grew and their tasks and goals required special knowledge by experts. Institutions were founded on a legal basis by division of labour. As Figure 11.2 illustrates, dialogue as a human need is a common object-of-study for all the disciplines of the humanities and social sciences. In this sense they are called to cooperate and contribute to the unity of knowledge or consilience (Wilson, 1999). We do not start 178



The Mixed Game Model: A Holistic Theory the complex whole: dialogue in the stream of life | specialization hierarchy of complex interacting subsystems

everyday life business law instruction

institutional life health

the media governance ….

Figure 11.2  Specialization of Subsystems in a Complex Hierarchy.

from a specific language or specific cultural conditions, we start from the complex whole which is, to some extent, universal and the basis of all possible parallel derivations in different languages and cultures. Moreover, institutional dialogue sets particular conditions by the fact that it presupposes a legal basis which determines the status function of their members (see Searle, 1999: 125 and Weigand, 2010: 233; also Chapters 20 and 21, this volume). This condition, too, is a condition of the type ‘institution’ which is then differently applied in different cultures, for instance, different business cultures. Scientists can learn from observing how human beings tackle the uncertainties of life. They look for rules and standard cases but when rules come to an end they orient themselves towards probabilities and even take account of chance when trying to find out about a particular case. Human beings are complex adaptive systems. Even rules depend on the basic principle of performance which is a principle of probability. In doing so, human beings demonstrate another interesting feature which is included in their competence-in-performance. They know that not every empirical detail is relevant for coming to grips with the challenge of a complex object and therefore do not scrutinize any individual ‘data’ but focus their attention on as many variables as possible which seem relevant. In this way they tentatively and adaptively proceed in the complexity of life. The basic premises of integration and specialization also allow us to include culture and even to explain why there are so many different cultures. Culture is only in part an external, separately describable component, for instance, of a specific value system. Essentially it is an innate human feature. Human beings are cultural beings, as a consequence of the co-evolution of genes, mind and culture. Culture affects our way of thinking, especially how we evaluate the relationship between the members of a community, and consequently it basically affects our dialogic devices. A very clear example of different evaluation in different cultures are compliments (see Weigand, 2016). Evaluating is an inborn human feature: we not only listen, we react, and we react according to how we evaluate what happens around us. Evaluation includes taking a position in balancing our double nature as dialogic individuals, which is an ethical issue and inevitably influences our actions. Any action theory of dialogue has therefore to take account of evaluation and culture.

11.4  Language as Dialogic Action The view of ‘language as dialogue’ is based on a view of human beings as dialogic individuals. We do not speak because we like speaking or because we need speaking in order to 179

Edda Weigand

think, we speak because we have goals which need to be negotiated with our fellow beings. Human beings are goal-directed, social beings who are living in the world together with other human beings. Language use, therefore, means dialogic interaction. Acting and reacting is not the same as turn taking. Turn taking remains at the empirical level; adjacency pairs are mainly formally identified by their position and by frequency (Sacks et al., 1978). Even if terms like action and interaction are nowadays frequently used in conversational analysis, they are mostly not really justified as actions at the level of meaning. Reactions as a special type of action which is functionally dependent on the initiative action were first highlighted by Hundnurscher (1980) in his article on “Konversationsanalyse versus Dialoggrammatik” (see Chapter 10, this volume). In this way monologic speech act theory changed from a theory of speakers to a theory of interlocutors who act and react in dialogue. In order to achieve our goals we do not simply speak, we simultaneously think and perceive and have emotions. The integration of abilities is a human condition. Utterances are therefore not only units of speaking, they result from the mix of different human abilities. The problem of separating utterances can only be settled at the level of meaning. Utterances are used as devices for speech acts. The utterance ends where the speech act ends. Speech acts are not the only type of human action. We not only have communicative purposes but also practical and even cognitive purposes. We therefore also act practically, for instance, when felling a tree, and we also act cognitively, for instance, when we need to take a decision. In general, actions can be defined by the correlation of having a purpose or goal and using appropriate means, verbal, cognitive, practical, to achieve it. Having a purpose includes intention. Action in this sense is restricted to human beings and some animals, for instance, dogs and dolphins (see, for instance, Anscombe, 1957). Recently another concept of action was proposed according to which inanimate things, such as words and texts, can also act without any author being involved (see, for instance, Cooren, 2010). In the MGM words and texts are communicative means used by their authors even if in some cases the author cannot be identified. In the literature, we also find the term ‘activity’ as well as ‘action’, unfortunately without a clear definition. Actions are defined by the correlation of purposes and means, that is, they presuppose intentions and are related to a specific purpose of their own. On the other hand, activities simply mean doing, which might even intentionally and consciously have been carried out but lack a purpose constitutive to it, for instance, dancing with the goal of having fun. Mostly activities are carried out as auxiliary or routinized parts of action, for instance, approaching the window in order to carry out the action of closing the window. As mentioned above, the course of interaction in performance is not exclusively a matter of rules and patterns but basically a matter of probabilities. We orient ourselves in dialogue according to speech acts which might be carried out and to strategies which might be taken. Principles of probability are therefore the methodological basis of the MGM. We can distinguish between three different types: constitutive, regulative, and executive. I can only briefly introduce the main principles (for details see Weigand, 2010).

11.4.1  Constitutive Principles Constitutive principles are the principles of action, dialogue and coherence. They constitute dialogue in its fundamentals.

180



The Mixed Game Model: A Holistic Theory

The Principle of Action defines action, as mentioned above, as correlation of communicative purposes, which are the driving force, and appropriate communicative means: communicative purposes

↔

appropriate communicative means

Figure 11.3  The Action Principle.

Actions are defined by their meaning. Empiristic approaches take the opposite direction and try to arrive at action functions by starting from expressions. Such a direction, of necessity, ends in an impasse: there is no direct correlation between specific means and action. The means are rather infinite and continuously changing and include means which are not empirically observable. The purposes are to a great extent quasi-universal and related to human nature. Cultural differences do not touch upon the basic human needs but can shape them differently, for instance, by differently evaluating how to address our interlocutors. Searle’s formula F(p) indicates the functional constituents of a speech act from the perspective of classical philosophical speech act theory (Searle, 1969). F action function

(p) (proposition) reference

predication

Figure 11.4  The Classical Speech Act Formula.

In order to achieve a formula for the speech act which can be the basis of a theory of competence-in-performance we have to be aware of the fact that not everything is said explicitly. Interests which remain concealed have to be included in the formula as they influence the sequencing of actions. I therefore extended the classical formula by introducing a superordinate predicate interest (e.g., Weigand, 2006): INTEREST [F(p)]

Figure 11.5  The Extended Speech Act Formula.

Communicative actions are always dialogic actions, insofar as human beings are dialogic individuals who have to find some balance between their self-interest and social concerns. The Action Principle is therefore intrinsically combined with the Dialogic Principle proper. Dialogue not only means being in a dialogic relationship with someone else, dialogue means action and reaction. Orthodox speech act theory has to be changed to a dialogic speech act theory. Any speech act is either an initiative or a reactive speech act, which is not only a distinction of position but a distinction of function. This means we not only have one type of speech act, Searle’s illocutionary act, but two types, the initiative ‘illocutionary’ speech act and the reactive ‘perlocutionary’ one. In clear cases we need not know the position; the utterance as such, for instance, you are right, indicates whether it can be used initiatively or represents a reaction.

181

Edda Weigand

Calling the moves ‘turns’ in conversation analysis is not really appropriate for speech act theory insofar as ‘turns’ remain units of the empirical surface. Turns and ‘data’ do not allow safe access to the meaning of speech acts. ‘Initiative’ and ‘reactive speech acts’ can be functionally defined by the criteria of making a claim and fulfilling this very claim. ‘Fulfilling’ does not mean accepting the claim of the preceding initiative speech act but relating to it in some way, positively, negatively, or negotiating the decision. Only in this way can we arrive at some understanding which in performance will always be gradual understanding, not exclusively based on rules and conventions but on expectancy. coming to an understanding

↔

action

reaction

making a claim

fulfilling the same claim

Figure 11.6  The Dialogic Principle Proper.

To my mind, there is an interesting cultural difference in the meaning of reaction. In some cultures we might be confronted with a reaction which is the negation of any reaction. We might, for instance, understand the Italian reactive speech act ne parleremo as ‘we’ll talk about it later’, whereas it simply means ‘I don’t want to react, not now nor in the future, I refuse to react’. The actual meaning becomes clear if we insist and ask ‘when will we talk about it’ and again get the reply ne parleremo. Cases like this demonstrate that the fundamentals of action and reaction are universal, but specific social habits might evolve in different cultures and these might cause misunderstandings. Extended sequences of speech acts are also structured according to the basic two-part sequence (Weigand, 2010: 113f.). The three-part sequence which we sometimes find in the literature, especially in the field of education, can be restructured as two two-part sequences. I take an example by Mehan (1985) from educational discourse: ‘elicit’

→

‘reply’

→

‘evaluate’

initiative

→

reactive initiative

→

reactive

→

two-thirty

→

very good, Denise.

what time is it?

Figure 11.7  Restructuring the Three-Part Analysis.

The two-part sequence of action and reaction which can secondarily be initiative determines any dialogue, from minimal dialogues to extended sequences. What at the empirical surface seems to be sequences of an initiative speech act and a series of reactions is at the level of function structured as follows: action

→

reaction action

→

reaction action

→

reaction action

Figure 11.8  The Basic Two-Part Sequencing Principle. 182

→

reaction



The Mixed Game Model: A Holistic Theory

The Principle of Coherence is the third constitutive principle (Weigand, 2010: 86f.). Coherence is no longer established at the verbal level of text or discourse but results from the adequate integration of verbal, cognitive and perceptual means due to the integration of our abilities. In this sense coherence is established in the minds of the interlocutors. dialogic purposes

↔

dialogic means integration verbal

cognitive

perceptual means

Figure 11.9  The Principle of Coherence.

Two decisive authentic examples will demonstrate how coherence works: (1)  One for her. – Eve, too. This piece of text is the verbal component of an action game which can only be understood by including what can be perceived: A woman and a man—we assume them to be mother and father—are walking around a lake with a little boy and a little girl—we assume them to be their children. The father cuts willow branches for the boy. Traditionally, such examples might be described as ellipses (see Chapter 12, this volume). In authentic dialogues we are, however, not restricting ourselves to the verbal level. We cannot do other than use our abilities in integration, in this case the ability of speaking integrated with perceiving. Coherence comes about by the attempt of the interlocutor to understand the utterance which is more than understanding what has been verbally said. In principle, it is not always a simple task for an outsider of the game to find out which perceptual features are relevant. The other example was given by Brown and Yule (1983: 196): (2)  There’s the doorbell. – I’m in the bath. Examples of this type very clearly demonstrate that we mean more than we say, speaking and thinking are integrated. In contrast to Cognitive Pragmatics we need not call such an example a ‘problematic case’ (Yule, 1996: 4) but can explain it as a quite normal case of language use. We draw conclusions of different kinds and expect our interlocutors to do the same. Conclusions can relate to conventional habits of our life as in our example. They can however also be individual, non-conventional conclusions which require that the interlocutor is an insider of the game. In any case, in both examples, it is not only the integration of different dialogic abilities but also the Dialogic Principle proper which establishes the coherence between the two utterances insofar as both relate to the same claim, a claim to volition, in (1) directly expressed, in (2) indirectly expressed and fulfilled positively in (1) and—as it seems—negatively in (2).

183

Edda Weigand

Both examples clearly demonstrate that an empiristic approach will not suffice to describe and explain authentic language use which is more than spoken language. The relevant perceptual features and cognitive means are not indicated in a corpus nor can they be secondarily added in an abstract multilevel model of language use. They can only be dealt with consistently in a holistic action-based theory.

11.4.2  Regulative Principles Regulative principles can be considered as corollary principles that operate on the basis of constitutive principles. As they are rooted in human nature, they exert a permanent influence on human interaction and behaviour in the mixed game. Whereas the coherence principle is primarily an issue of understanding how different abilities interact, regulative principles mediate between possibly opposing abilities and interests of the speaker such as the balance between the speaker’s selfish and social interests or the interplay between reason and emotion (see also Chapters 17, 18, and 22, this volume). Human beings are social individuals by their nature. They therefore need some principles to deal with the balance between sometimes opposing interests. Self-interest in social interaction is basically directed towards achieving acceptance. In this sense human beings are persuasive beings. Persuasion in the broad sense of achieving acceptance might even include some sort of power. Since we live in a community we are at the same time faced with social concerns and the necessity of respecting the other human being. Respect is usually dealt with under the heading of politeness, mostly as a negative phenomenon in the sense of ‘avoiding face-threatening acts’. At its core however, politeness represents a positive concept, the concept of the respect to be paid to our fellow beings and expected for ourselves. Again we encounter a very interesting and striking example of how culture affects the universal basis. In different cultures the balance between the self and the other is dealt with very differently as a consequence of different evaluation. Roughly speaking, Western cultures very much emphasize individual freedom whereas Eastern cultures evaluate the individual according to their role in the community (Weigand, 2010: 95). We have to be cautious, though, because such a rough-and-ready distinction will have to be more clearly differentiated by future studies. Human beings’ double nature as social individuals will inevitably imply both cooperative as well as confrontational tendencies. [universal]

achieving acceptance individual the self

[culture]

↔ ↔

community the other

biased evaluation West the self → the other bias on self-interest confrontation

East the self ← the other bias on respect harmony

Figure 11.10  The Balance of Self-interest and Respect in Different Cultures. 184



The Mixed Game Model: A Holistic Theory

Dialogic action always calls for taking a position in balancing our double nature. Do we intend to push through our selfish interests or do we want to include social concerns? Dialogic action therefore represents a rhetorical and ethical issue (see Chapters 18 and 19, this volume). Considering human beings as persuasive beings means giving up the traditional distinction between rhetoric and grammar and integrating rhetoric in a holistic model (Weigand, 2010: 63, 72ff.; 2015a, 2015b). In the same way, due to human beings’ double nature, human action and behaviour are intrinsically ethical action and behaviour. Ethical principles which mediate between possibly opposing interests have to be included as a component of regulative principles. The decision taken in each individual case will be an individual decision which might be affected by cultural value systems. Regulative Principles not only shape the balance between self-interest and respect but also tell us how to mediate between reason and emotion. In our post-Cartesian times, eventually neurology experimentally proved that emotion and reason do not represent separate areas but interact with each other and with other human abilities (Damasio, 2000; Pinker, 1997: 370; Chapter 17, this volume). Emotions have the extraordinary feature that they can express themselves as physiological processes even against the intention of the speaker. Physiological processes which are not prompted by intention cannot count as communicative means. On the other hand, emotions can intentionally and explicitly be expressed by various lexical and grammatical means, and they can be expressed for various purposes, among them in order to persuasively move other people. ‘Expressing emotions’ can even be the action function of speech acts, for instance, in emotive speech acts (see Weigand, 2010: 167; 2004). The balance between reason and emotion is again a case where culture takes effect. There are, in principle, three types of internal biological mediation between emotion and reason which also affect the dialogic sequence (see Chapter 17, this volume): Emotions are kept under control. Emotions can no longer be controlled and become dominating. Emotion and reason are in conflict and lead to quarrels. Rhetorical Principles tell us how to deal with these kinds of internal regulation in dialogic interaction. They are dependent on different cultural evaluations. For the culture of northern Europe, we can consider the following Principles of Emotion as conventional or normative rhetorical guidelines in public dialogue: Act rationally and hide or control emotions. Keep quiet, do not be provoked. Learn to take criticism. Do not overreact. For a more detailed description of emotions in different cultures you are referred to Grein and Weigand (2007) and Weigand (2004). There you will also find authentic examples which illustrate the principles (see also Chapter 22, this volume.)

11.4.3  Executive Principles Besides constitutive and regulative principles there are executive principles which go beyond simple sequences of action and reaction and determine longer sequences of negotiating the 185

Edda Weigand

goals and interests of the interlocutors. Executive principles can be rhetorical strategies in order to be successful in dialogue. There are various types of strategies, among them co-operative strategies of persuasion versus confrontational strategies of power, as is illustrated by the following authentic examples (Weigand, 2010: 123f.). The first example is a dialogue between a girl and her mother. The girl tries to persuade her mother to allow her to take part in a very attractive but also expensive school exchange programme with America. The mother has not decided yet and is more inclined to decide for the cheaper programme. (3)  Girl We have to talk about America. I thought I could save, I could earn money by giving lessons and I would contribute the money which I have in my account.

Mother Well, I’ll think about it. Be quiet, I tend to agree.

The girl’s strategy of cooperation is successful. It is just the appropriate way of persuading her mother in the particular action game. The other example can be considered as a counter-example, if at all, of failed persuasion because it uses the strategy of confrontation and power: (4)

A B A (angry) B A (very angry) B    

We have asked you to present a joint working plan. Here it is. This won’t do. You cannot start with your position only. But I expressed it as you wanted it, some time ago. We won’t discuss with you. OK, then I might as well go.

In this specific action game power is the wrong means to use because it is vacuous, only simulated. Interlocutor A who uses it does not really possess it. His claim to volition, intransigently expressed by we won’t discuss with you and even strengthened by putting his foot down, must therefore fail. The strategy of power might have been successful with a shy and timid interlocutor, but in the actual action game it leads to the dialogue breaking off. These examples demonstrate the trivial fact that any action game is an action game with individual interlocutors in a specific situation. The individuality of the interlocutors as well as the uniqueness of the situation are inherent components of the game. Human competence-in-performance does not work by means of the addition of constructs but by facing the complex whole from the very outset and adapting to it. Dialogue is the key to pragmatics.

11.5  A Dialogic Speech Act Taxonomy The issue of a speech act taxonomy has been a basic issue of speech act theory from its beginnings (see Chapter 10, this volume). There are many approaches which try to define the types of speech acts we are carrying out in language use (see Weigand, 2010, 1996). Mostly they start from Searle’s monologic speech act theory, but there are also approaches which claim they can find the way to action functions by starting from empirical data or even assume that speech act types can be equated with speech act verbs. The challenge is to develop a dialogic comprehensive speech act taxonomy which starts from the complex whole and derives the particular types as components from it. 186



The Mixed Game Model: A Holistic Theory

Again it is the principle of specialization which allows us to tackle this challenge and to derive dialogic action types from the general purpose of dialogue, which is coming to an understanding. As we start from the dialogic unit of action and reaction (Figure 11.6 above), the taxonomy not only derives initiative actions but at the same time reactive actions which are defined by their preceding initiative action. Defining reactive speech acts by their dependence on an initiative claim does not mean that we as interlocutors are restricted to react in this way. In performance the correlation is a matter of expectancy. Insofar as a dialogic speech act taxonomy starts from the sequence of action and reaction, it will inevitably contain speech act types which are missing in Searle’s taxonomy and will also describe and assign some of Searle’s speech act classes differently. Specialization proceeds in two steps: first by specializing the dialogic claim, that is, the action function, and second by differentiating the proposition. Human beings’ mental states are the origin of human action, the long-sought bridge between cognition and speech acts. If we consider ‘intention’ to be included in the purpose of an action, then the mental states of belief and desire will form the basis of the fundamental claims to truth and volition on which the whole action typology is based. Let me outline the first basic steps of such a comprehensive dialogic speech act taxonomy (for details see Weigand, 2010: 129–207). dialogue aiming at coming to an understanding action ↔ reaction [reaction necessary] claim to truth REPRESENTATIVE

[reaction not necessary]

claim to volition ↔ ACCEPTANCE [directed to knowledge] EXPLORATIVE ↔ RESPONSE

claim to volition, made and fulfilled [directed to reliability]

[not directed to knowledge] ↔ CONSENT

DECLARATIVE

DIRECTIVE

Figure 11.11  Dialogic Typology of Basic Action Games.

Dialogue cannot be reduced to information exchange. There is no truth as such, only the human claim to truth which ranges between a simple claim ‘that it is so’ and a modal claim ‘that it would, could, should be so’ in representative action games. There is also the human claim to volition which can be a claim directed to knowledge in explorative action games or to practical action in directive action games. In addition, living with uncertainty requires us to create some reliability by specific action conditions in declarative action games. In the end any action we classify in authentic dialogues, initiative as well as reactive, needs to find its place in the complex hierarchy of the typology. The taxonomy I developed in the Mixed Game Model (2010) relies, to a great extent, on the taxonomy I had already developed in my book Sprache als Dialog (first edition in 1989 and revised in 2003). The decisive difference is the difference between a taxonomy of dialogic competence based on conventions (see Chapter 10, this volume) and a taxonomy of dialogic competence-in-performance based on expectancy or principles of probability. Following the holistic procedure of specializing we can derive sub-types by specializing the dialogic claim (Weigand, 2010: 155ff.). The claim to volition, which defines directives insofar as they are not directed to knowledge (see Figure 11.6), can be differentiated towards a claim which is backed by some commitment or duty and a claim which can only 187

Edda Weigand

appeal to the kindness of the interlocutor. Specializing the initiative speech act also presupposes some specialization of the corresponding reaction. In this way the following directive subtypes can be derived. ↔

DIRECTIVE

CONSENT

claim to volition

[claim to be fulfilled]

[+sanction] ORDER

↔ OBEDIENCE

[no claim to be fulfilled]

[− sanction] REQUEST

↔ CONSENT

PLEA

↔ GRANT

Figure 11.12  Derived Types of Directives.

Examples: (5) order (father to young daughter) A You’re home at 8 o’clock! B O.K. (6) request (among friends) A Can you bring a glass of milk, please. B Yes, certainly. (7) plea (between customer and salesperson) A Would you please put the book aside for me for one day? B Yes, for one day, but no longer. Deriving subtypes of the representative claim to truth according to different positions on the scale of certainty is particularly differentiated insofar as we not only have the basic claim to truth of an assertion ‘that it is so’ (Figure 11.13) but also the modal claim ‘that it could/ would/should be so’ by which we can express our beliefs, conditions of action and norms of behaviour (Figure 11.14). As can be seen in Figure 11.13, Searle’s class of expressives is included in the representative class of ‘emotives’. Expressing our feelings means a representative speech act which relates to our inner world and also contains some emotional feeling. There is no need to classify expressives as a separate basic speech act class. Having derived subtypes by specializing the dialogic claim or the action function of the speech act, we can now continue specializing the speech act type by including propositional features (Weigand, 2010: 173ff.). Again, with representatives the derivation is particularly manifold. I can only illustrate the derivation by means of propositional features by an example. In the literature speech acts of ‘advising’, ‘warning’ and ‘proposing’ are often presented without any definition. The difference can be made explicit by specifying the proposition (see Figure 11.15).

188



The Mixed Game Model: A Holistic Theory

REPRESENTATIVES

to take as true

↔

ACCEPTANCE

[simple claim to truth ‘that it is so’]

[to be proved]

ASSERTIVES

[modal claim]

[not to be proved]

[rationality]

[definition]

↔ ACCEPT

IDENTIFIERS

↔ ACCEPT [new]

INFORMATIVES

↔ COMMENT [obvious]

CONSTATIVES

[emotional]

↔ CONFIRM/BELIEVE

EMOTIVE

↔ EMPATHY

Figure 11.13  Derived Representative Types of the Simple Claim to Truth. REPRESENTATIVES

[simple claim to truth ‘that it is so’]

CONDITIONAL

↔ ACCEPTANCE [modal claim]

[‘that it would be so’]

↔ ACCEPTANCE/BELIEF [‘that it could be so’]

DELIBERATIVE

[‘that it should be so’]

↔ ACCEPTANCE [desire]

DESIDERATIVE

[norm]

↔ BELIEF/ACCEPTANCE

NORMATIVE

↔ ACCEPTANCE

Figure 11.14  Derived Representative Types of the Modal Claim to Truth.

‘advice’ ‘warning’ ‘proposal’

FUNCTION

(PROPOSITION)

ASSERTIVE

the action by the interlocutor is useful the action by the interlocutor is dangerous the action by the speaker and interlocutor could be useful

ASSERTIVE DELIBERATIVE

Figure 11.15  Differentiating the Proposition. 189

Edda Weigand

The difference between these types of speech acts can also be demonstrated by the use of different modal devices, for instance: (8)

I suggest we could go to Paris. I advise you to go to Paris. *You could go to Paris, I advise you.

“Wir könnten nach Paris fahren, schlage ich vor.” “Du solltest nach Paris fahren, rate ich dir.” *Du könntest nach Paris fahren, rate ich dir.

Two striking differences between this dialogic speech act taxonomy and Searle’s monologic taxonomy should be pointed out: the class of commissives is missing, and the class of exploratives has been added. ‘Commissives’ confront us with the problem of whether they can be initiative at all, as we would expect from all the speech act types in Searle’s taxonomy. Can we really open up a dialogue by promising? (9) I promise to attend the lecture. On the contrary, the usual sequence would be: (10) A B

Let’s attend the lecture. OK, I’ll be there.

The problem, however, is not solved by accepting the reactive nature of promises. The problem also relates to the difference between simple consents and promises. What does it mean to ‘commit oneself to a future action’? Where is its place in the sequence? Speech acts of ‘commissives’ underwrite the consent. What distinguishes them is a difference in the sequence position: DIRECTIVE – CONSENT – ASKING FOR COMMITMENT – ‘COMMISSIVE’

Figure 11.16  The ‘Commissive’ Sequence.

Searle’s commissives actually have their place in the fourth move of the sequence. They take up the reaction of consent and strengthen it by a feature of commitment: (11) A I invite you to my party. B Thanks, I’ll come. A Can I rely on it? B Sure. Making a commitment constitutes a declarative speech act. Committing oneself to a future action does not represent an autonomous speech act but strengthens the preceding speech act of consent by adding a declarative component (Weigand, 2010: 150f.). Now let me explain why I introduced the basic class of exploratives. Exploratives serve to fulfil a basic need of human beings: the need to be informed or to get the information necessary for action. It is all the more astonishing that in Searle’s taxonomy such an essential type of interaction is not included as a type of its own but is subordinated to directives. Exploratives have something in common with directives: the claim to volition. They are, however, fundamentally different with respect to the object of this claim to volition: exploratives aim at knowledge, directives refer to practical actions by the interlocutors. This crucial difference becomes manifest in a dialogic speech act taxonomy through the reactive act: the reaction after an explorative act is not the same as the reaction after a directive act: 190



The Mixed Game Model: A Holistic Theory

EXPLORATIVE

↔

± RESPONSE

↔

± CONSENT

claim to volition

knowledge

DIRECTIVE

claim to volitionpractical action

Figure 11.17  Exploratives versus Directives.

Examples: (12)  Do you know ‘Citizen Kane’? – No, I’m sorry. (13)  Come and see me tomorrow! – I’m sorry, I already have an appointment. To sum up: The basic issue of an action theory of language is the issue of a dialogic speech act taxonomy. Actions are functional concepts to be defined at the level of meaning. The dialogic speech act taxonomy described above demonstrates that it is possible in a holistic theory to derive all the particular speech act types we are confronted with in performance from the general purposes which defines the whole of dialogic action. The fact that particular speech act types are not always explicitly or clearly expressed is an issue of the dialogic means, that is, an issue of an utterance grammar, not of the concepts of meaning (see Chapter 13, this volume).

11.6  Minimal versus Complex Games In performance the minimal two-part action game of an initiative speech act and the corresponding reaction is extended in various ways. I can only briefly deal with them in order to show that even ‘dialogue in the stream of life’ (Wittgenstein 1981 [1967]) is structured according to the ‘architecture of complexity’. For a more detailed description you are referred to Weigand (2010: 209ff.). There are basically two ways of extending the minimal game: either within the phase of action and reaction or going beyond it. Problems of understanding or even misunderstandings might arise and require clarification (Weigand, 1999) or the interlocutor may need more information before deciding how to react. The minimal phase is therefore extended by embedded sequences, as, for instance (Weigand, 2010: 214): (14)

Let’s go to the cinema this evening. Is there an interesting film? There is Benigni’s “La vita è bella.” Oh, very fine, let’s go. directive →

← consent

explorative ↔ response

Figure 11.18  Embedded Clarifying Sequence.

Dialogues of argumentation can also be considered as extending the first sequence of thesis and antithesis by putting forth arguments pro and contra (Weigand, 2010: 215ff.): Arguments can also strengthen the positions in a directive action game as we have seen with example 3. 191

Edda Weigand representative type of argumentation ↔

thesis ↑ ARGUMENTS

- acceptance, anti-thesis ↑ ARGUMENTS

↓ argumentation ↓ thesis, anti-thesis, compromise

Figure 11.19  Argumentation in a Representative Game.

Whereas one-phase games, minimal as well as extended games, are to some extent within our grasp as they can be directly traced back to speech act functions, for multi-phase games the situation is different. We are confronted with complex purposes related to complex states of affairs, that is, with a network of subcomplexes which needs to be elaborated. In such multi-phase games we are faced with a twofold challenge: to analyse the complex purpose, that is, to clarify what is at stake, for instance, in ‘planning’ or in ‘stocktaking’, and to decide on how to proceed in putting the subpurposes in linear order. To give an example for the structure of ‘planning’ (Weigand, 2010: 246): PLANNING

FIXING THE GOAL

→

PROPOSALS

→

for achieving the goal with specific reference to fields and measures

EVALUATING

the proposals by arguments

→

DECIDING

the course of the action

Figure 11.20  The Action Game of Planning.

For a detailed analysis of such a ‘dialogue in the stream of life’ you are referred to Weigand (2015b).

11.7 Conclusion The Mixed Game Model claims to be comprehensive and to lay the basis for language both in general and in particular insofar as it allows us to differentiate any particular component of any language or culture within the complex whole. In this way all the issues raised by the various turning points in the history of modern linguistics can be consistently settled and also be verified by neurobiology and anthropology in a comparative theory of different languages and cultures. The crucial point is not only a change in the object ‘language’ from a system of signs to ‘language as dialogue’ but also the change in theorizing from reductionism to holism (Weigand, 2011), which is necessary for consistently analysing a complex object-of-study. The complex whole has its own architecture which basically rests on two premises: the whole is more than the sum of the parts, and integration is the name of the game. Fields which traditionally have been dealt with as separate parts become integrated as components of the complex human ability of competence-in-performance. Also ethics is included insofar 192



The Mixed Game Model: A Holistic Theory

as dialogic action requires evaluating and balancing our double nature as dialogic individuals. Evaluating is an inborn feature which is highly affected by cultural criteria, social and ideological ones (see Chapters 13, 19 and 22, this volume). Addressing the complex object of human beings’ actions and behaviour turns out to be an endeavour of various disciplines which can no longer search for their own truth restricted by the artificial limits of their discipline. On the whole they all should aim to contribute to the unity of knowledge or consilience (Wilson, 1999) in a genuine sense of interdisciplinarity. In post-Cartesian times sociobiology, or the co-evolution of genes, mind, and culture, represents the basis of science which unites verifications by neurobiological experiments and by empirical observations of anthropology worldwide (Lumsden and Wilson, 2005; Damasio, 1994; Marchand, 2010). What makes the human species special is not a single ability but the integrated use of our abilities which defines the concept of ‘language as dialogue’ and reveals the human species to be a dialogic species.

References Allan K, Capone A and Kecskes I (eds) (2016) Pragmemes and Theories of Language Use. Cham: Springer. Anscombe GE (1957) Intention. Oxford: Blackwell. Brown G and Yule G (1983) Discourse Analysis. Cambridge: Cambridge University Press. Chomsky N (1959) Review: B. F. Skinner: Verbal Behaviour. New York 1957. Language 35: 26–58. Chomsky N (1966) Cartesian Linguistics: A Chapter in the History of Rational Thought. New York: Harper & Row. Cooren F (2010) Action and Agency in Dialogue: Passion, Incarnation and Ventriloquism. Amsterdam: Benjamins. Damasio A (1994) Descartes’ Error: Emotion, Reason, and the Human Brain. New York: Putnam. Damasio A (2000) The Feeling of What Happens: Body, Emotion and the Making of Consciousness. London: Vintage. Fetzer A (2004) Recontextualizing Context: Grammaticality Meets Appropriateness. Amsterdam: Benjamins. Feynman RP (2001) The Pleasure of Finding Things Out. London: Penguin Books. Finkbeiner R, Meibauer J and Schumacher PB (eds) (2012) What Is a Context? Linguistic Approaches and Challenges. Amsterdam: Benjamins. Gell-Mann M (1994) The Quark and the Jaguar: Adventures in the Simple and the Complex. London: Abacus. Grein M and Weigand E (eds) (2007) Dialogue and Culture. Amsterdam: Benjamins. Hundsnurscher F (1980) Konversationsanalyse versus Dialoggrammatik. In Rupp H and Roloff H-G (eds) Akten des VI. Internationalen Germanisten-Kongresses. Basel 1980, Part 2. Bern: Lang, 89–95. Kretschmar WA, Jr (2015) Language and Complex Systems. Cambridge: Cambridge University Press. Lumsden CJ and Wilson EO (2005) Genes, Mind, and Culture: The Coevolutionary Process. New Jersey: World Scientific. Marchand THJ (2010) Making knowledge: Explorations of the indissoluble relation between minds, bodies and environment. Journal of the Royal Anthropological Institute 16(1): 1–21. Mehan H (1985) The structure of classroom discourse. In Van Dijk TA (ed.) Handbook of Discourse Analysis, Vol. III. London: Academic Press, 120–131. Mey JL (2001) Pragmatics: An Introduction, 2nd edn. Malden, MA: Blackwell. Pinker S (1997) How the Mind Works. London: Penguin. Prigogine I (1994) Les lois du chaos. Paris: Flammarion. Sacks H, Schegloff EA and Jefferson G (1978) A simplest systematics for the organization of turn taking for conversation. In Schenkein J (ed.) Studies in the Organization of Conversational Interaction. New York: Academic Press, 7–55. 193

Edda Weigand

Searle JR (1969) Speech Acts: An Essay in the Philosophy of Language. Cambridge: Cambridge University Press. Searle JR (1999) Mind, Language and Society: Philosophy in the Real World. London: Phoenix. Seligman MEP (1970) On the generality of the laws of learning. Psychological Review 77: 406–418. Simon HA (1962) The architecture of complexity: Hierarchic systems. Proceedings of the American Philosophical Society 106: 467–482. Tannen D (1989) Talking Voices: Repetition, Dialogue, and Imagery in Conversational Discourse. Cambridge: Cambridge University Press. Toulmin S (2001) Return to Reason. Cambridge: Harvard University Press. Van Dijk TA (ed.) (1985) Handbook of Discourse Analysis, Vols 1–4. London: Academic Press. Von Humboldt W (1827 [1963]) Ueber den Dualis. In Flitner A and Giel K (eds) W. von Humboldt: Schriften zur Sprachphilosophie, Vol. 3. Darmstadt: Wissenschaftliche Buchgesellschaft, 113–143. Weigand E (1989 [2003]) Sprache als Dialog. Sprechakttaxonomie und kommunikative Grammatik, 2nd rev. edn. Tübingen: Niemeyer. Weigand E (1996) The state of the art in speech act theory: Review article on: Tsohatzidis SL (ed) (1994) Foundations of Speech Act Theory. Pragmatics & Cognition 4: 367–405. Weigand E (1999) Misunderstanding: The standard case. Journal of Pragmatics 31: 763–785. Weigand E (2004) Emotions: The simple and the complex. In Weigand E (ed) Emotion in Dialogic Interaction: Advances in the Complex. Amsterdam: Benjamins, 3–22. Weigand E (2006) Argumentation: The mixed game. Argumentation 20(1): 59–87. Weigand E (2009) Language as Dialogue. Sebastian Feller (ed). Amsterdam: Benjamins. Weigand E (2010) Dialogue: The Mixed Game. Amsterdam: Benjamins. Weigand E (2011) Paradigm changes in linguistics: From reductionism to holism. Language Sciences 33(4): 544–549. Weigand E (2014) Rationality of Performance. Philosophia Scientiæ 18(3): 247–268. Weigand E (2015a) Persuasion or the integration of grammar and rhetoric. In Săftoiu R, Neagu M-I and Măda S (eds) Persuasive Games in Political and Professional Dialogue. Amsterdam: Benjamins, 1–18. Weigand E (2015b) Dialogue in the stream of life. Language and Dialogue 5(2): 197–223. Weigand E (2016) How to verify a theory of dialogue. Language and Dialogue 6(3): 349–369. Wilson EO (1999) Consilience: The Unity of Knowledge. New York: Vintage Books. Wittgenstein L (1981 [1967]) Zettel, 2nd edn. Oxford: Blackwell. Yule G (1996) Pragmatics. Oxford: Oxford University Press.

194

Part II

Theoretical Key Issues and Open Debates

Part II of the Handbook deals with theoretical key issues and open debates which c­ haracterize the current state of dialogue research. There is first the concept of language which underwent many shifts from de Saussure’s system of signs to Weigand’s mixed game. Ruth Kempson focuses on the change from language as an abstract system to authentic language use which constitutes the pragmatic turning point. She deals with it at the level of syntax as shift from generative syntax to interactive syntax. Language interaction, however, means more than syntax and requires a reshaping of the whole concept of grammar. The issue is how a grammar of performance can be designed. Various approaches were proposed without critically reflecting on what constitutes a grammar and, in particular, a grammar of performance. Edda Weigand takes up this challenge and proposes a holistic model of an utterance grammar which fits the Mixed Game Model. An utterance grammar aims at describing human beings’ competence-in-performance in expressing and negotiating their mutual goals of action by means of utterances. As has become evident in some chapters of Part I, especially Chapter 2 (Mondada), the utterance does not only consist of verbal means. The integration of speaking, thinking and perceiving raises a crucial issue for an utterance grammar. Weigand’s chapter can only give a sketch of how to proceed. Elaborating a comprehensive utterance grammar will be a challenge in the future, especially if we take account of language comparison and the goal of a comparative grammar. The issue of the concept of language is connected with the issue of theorizing and the fundamental question of what constitutes a theory. The relationship between the object-ofstudy and methodology is central to it. The question is where to begin, with the model or with the object. André Martinet (1975) already emphasized decades ago that we should not distort our object-of-study by methodology, as was, for instance, the case with generative grammar. Arto Mustajoki deals with this issue and takes account of the perspective of various approaches. The distinction between theory and practice, or between theoretical and applied linguistics, represents another difficult issue. As soon as the object-of-study is language use in general, language use in areas of practice also belongs to it and should be covered in a theory of dialogue. There are, of course, special questions of practice which go beyond theory and should be focused on in applied linguistics. Language teaching, for instance, not 195

Edda Weigand

only means informing about the structure of foreign languages but includes motivating the learners and preparing the object so that it is suitable for their age. Dale A. Koike deals with this issue of theory and practice and asks what ‘application’ really means. There remains the question of how a theory can be verified. This is the issue of Marco Iacoboni’s chapter on the sociobiology of language. Reference to sociobiology means that the use of language in the end is dependent on human nature and on influences of the environment, the cultural as well as natural surroundings. Iacoboni focuses on the issue of what mirror neurons can tell us. Recent insights into the way our brain works combined with extended anthropological observations of different languages and cultures allow us to verify our assumptions about human abilities and about cultural influences which shape our creation of a theory of language. Extending the object language to human action and behaviour inevitably means going beyond disciplinary boundaries and taking account of the integration and interaction of various components. Genuine collaboration of different disciplines is needed in order to proceed on our way towards ‘consilience’ or the unity of knowledge.

Reference Martinet F (1975) Functional linguistics: La linguistique fonctionnelle. In Martinet F (ed.) Studies in Functional Syntax: Études de syntaxe fonctionnelle. München: Fink, 9–81.

196

12 Shifting Concepts of Language Meeting the Challenge of Modelling Interactive Syntax1 Ruth Kempson

12.1  The Point of Departure: Grammar versus Language Processing For the past one hundred years and more, languages have been seen as grounded in an abstract system of rules governing the structural properties of sentences of the language; and human beings’ capacity for language is seen in these terms, this being their underlying language ‘competence’. How languages are used from this perspective is another matter, involving a composite of language-internal, cognitive, and social factors, which together determine the data of ‘performance’. The methodological consequence of adopting this stance is, however, severe: in this view, there is strict separation between competence and performance. The task of explaining a speaker’s ‘competence’ cannot make reference to these ‘performance’ data because to do so would be to face the charge of circularity; and linguists impose a strict methodology that attributes of performance have no place within the grammar-internal explanations they provide, which is restricted to characterising phenomena within the boundaries of the sentence. Yet this restriction confronts the problem that languages display pervasive context-dependence at every level of analysis. At the semantic level, for example, there is immense variety in possible interpretations of both simple and composite linguistic expressions depending on the context in which they are used; and there has been a continuing swathe of cross-disciplinary research on the dynamics of how languages are processed in context and what this means for the nature of word and sentence meaning (Clark and Brennan 1991; Clark 1996; Sperber and Wilson 1995, 2002; Kamp and Reyle 1993; Carston 2002; Recanati 2003; Stainton 2006; Asher and Lascarides 2003; Barr 2004; Keysar 2007; Gregoromichelaki et al. 2011, 2013). Defining these concepts in a way commensurate with this over-arching methodology is highly problematic, and this chapter will argue that the nature of linguistic competence needs to be rethought. The idealisation of language-as-system was introduced by de Saussure (1916) with his distinction between langue, which he took to be a body of generalisations shared by a community of people that speak it, and parole, language performance, which he thought too variable to be amenable to formal study. It was taken up with a new psychological twist by Chomsky (1965) with the distinction between competence and performance; and with this shift, syntax came to be seen as the core of grammars for natural language, the competence model being a system taken to constitute the idealised speaker–hearer’s knowledge of 197

Ruth Kempson

language comprising rules underpinning structural properties of sentences of the language, along with phonological and semantic components: all these were defined without reference either to context or to the vagaries of performance. The assumption of a competence-performance divide is shared by otherwise conflicting frameworks. The major alternative of Montague Grammar (see Thomason 1974 for Montague’s collected papers), for example, was developed with the goal of formally modelling meaning in a language as an articulation of the language-world correspondence via the notion of truth conditions for its sentences (reflecting insights of Russell 1905, and Frege, translated by Geach and Black 1970). Though a departure from the Chomskian paradigm, Montague Grammar was equally sententialist, espousing a methodology restricted to modelling sentences of languages and their interpretation, making no reference to performance. Other paradigms have emerged, but the methodology of defining language as an abstract system remains standard.

12.2 Shifting Perspectives: Problems in Modelling Natural-Language Semantics 12.2.1  The Context-Relativity of Natural Language Understanding From early on there were problems confronting this perspective on the human language capacity, in particular the phenomenon of context-dependence. This was initially treated by many as a peripheral problem not threatening the basic sententialist methodology of grammar articulation, but there has been increasing recognition that dependence on context is systemic in natural languages. Furthermore, the concept of context required to explain such context-dependency is much broader than that provided within the boundary of a sentence in ways that are nonetheless relevant to articulating the intrinsic property of expressions of language. Pronouns are the core case, signalling that their interpretation has to be retrieved from something other than themselves within the context in which they occur. This source of their construal may be recovered from some previous part of the same sentence (as in example 1), from previous sentences as in example 2, from subsequent material (as in examples 3 and 4, labelled cataphoric), or indeed from some nonlinguistic source (as in so-called indexicals: example 5) (pronoun and antecedent, if there is one, are in bold): (1) If a colleague is arriving late from New York, he will be tired by the end of the day. (2) A: A colleague is arriving late from New York. B: He will be tired by the end of the day. (3) Knowing that he might get left behind, John tried to hurry. (4) It is obvious that Mary is sick. (5) A (seeing Bill coughing) He still smokes, I’m afraid. And as example 2 shows, whether it is the same or a different speaker that has provided that antecedent source for the pronoun’s construal is of no relevance to the success of the pronoun-antecedent correlation. Switch of speaker–hearer roles does not jeopardise the pronoun-antecedent relation at all. Modelling this intrinsic dynamic to the understanding of pronouns in language was a challenge first taken up by Kamp (1981), who developed Discourse Representation Theory (DRT) as a basis for modelling the context relative to which utterances are understood, the 198



Modelling Interactive Syntax

concept of partial representations of meaning encoded in natural language expressions, and the process whereby such encoded input is updated to yield their content in that context. Despite a wealth of insights, the problem which such a model confronts is that modelling a concept of context that is sufficiently rich to incorporate the various uses of pronouns—anaphoric, cataphoric, and indexical—does not sit well with the assumption that grammars characterise only properties internal to sentences, and those working in DRT adopted conventional assumptions about language from syntacticians. The only way to sidestep this problem while retaining the conventional methodology is to argue that it is only those pronoun construals that can be articulated within sentence boundaries which fall within the remit of the grammar: others are taken to be different in kind. Conventionally we refer to pronouns with a sentence-internal antecedent as bound pronouns or as involving c­ o-indexation. We refer to cross-sentential antecedent-pronoun relations as involving a discrete concept of discourse anaphora; and it was these which were Kamp’s main focus, hence the term Discourse Representation Theory (Kamp and Reyle 1993). Anticipatory pronouns as in example 4 are labelled as expletive pronouns; and pronouns as in example 5 are said to be indexical pronouns. Note the pseudo-technical way in which these are labelled as discrete ‘pronouns’ warranting wholly distinct characterisations within the grammar device, rather than different forms of construal for the various morphological forms. This leads to a further problem; all pronouns get analysed as ambiguous, solely by virtue of the methodology, and the effect is an implicit refusal of any commitment to explain the systemic contextdependence of natural language as displayed by anaphora, despite this being arguably the diagnostic of a natural language. All that has been achieved is the division of each pronoun into different types, each one assigned a set of different analyses. An additional problem is the lack of any attempt at modelling the interactive effect achievable by pronouns and other anaphoric devices. They are an overt trigger for interaction effects in conversational exchanges. Pronouns are used because some contributor has already talked of some person or entity; and the pronoun serves to signal that this entity remains pertinent to the ongoing dialogue. This phenomenon has been given considerable attention with the concept of anaphoric chain (Chastain 1975; Brandom 1995 among others). What has been given much less emphasis is that these pairings are signally unaffected by who is responsible for uttering either pronoun or antecedent (see examples 1 and 2): establishing the anaphor-antecedent relation is successful irrespective of any switch in speaker–hearer roles. This form of interaction is at best unexpected from a single-speaker perspective on language ability articulated exclusively through sentence-internal properties: it is monologues which would constitute the default mode of language use, not dialogues with their to and fro interaction. Pronoun-antecedent pairings are standardly taken to encode an agreement of content presupposed by all parties to the communicative event, involving mutual knowledge (Grice 1975), common ground (Clark 1996), and so forth. Yet, ever since the first airing of the concept of mutual knowledge (Grice 1975), it has been recognised as being overly strong (see Schiffer 1972; Sperber and Wilson 1995, amongst others); and the constraint that communication is successful only if grounded in the goal of achieving shared knowledge does not match many successful communicative exchanges (see Linell, Chapter 7, this volume; Keysar 2007; Barr 2004). What is uncontentious is that the use of anaphoric processes to pick up on some previously presented content invariably establishes an interactive effect between the various participants in the conversation irrespective of the content itself. As we shall see, it is this which underpins context-dependent effects rather than content recognised as shared (a view which matches that of Linell, this volume). 199

Ruth Kempson

12.2.2  Ellipsis and the Concept of Context Before turning to this issue, it is notable that the phenomenon of context-dependence and its resulting interactive effect is broader than the mere subset of anaphoric devices which a language happens to encode. Ellipsis follows the same pattern, ellipsis being where words seem not to need to be said because what is meant is obvious from the context. Like the construal of anaphora, the relation between ellipsis site and the expression from which it picks up its interpretation, characteristically but not necessarily straddles sentence boundaries (examples 6 and 7); can be established across a switch of speaker/hearer roles (example 8); allows for cataphoric uses (example 9); and even indexical construals (example 10): (6) John is thinking of going to London, despite the fact that Bill is not. (7) If John is coming up to London, then Bill will be too (8) A: John is writing a paper. B: Bill is too. (9) Knowing that he shouldn’t have, John nevertheless copied what Mary had written. (10) (said by a mother of a teenager standing looking longingly out to huge waves with a surfboard in hand) I wouldn’t if I were you: the flag’s flying. So the general dynamic of how interpretation is built up for ellipsis, and the effect of interaction it gives rise to, directly parallels that of anaphora. As with pronouns, there have been debates as to what representation of content needs to be invoked for such elliptical fragments. However, unlike anaphora, for which invoking an underspecified content intrinsic to the encoded anaphoric expression is uncontroversial, elliptical forms are standardly modelled as having a content assigned by grammar-internal rule, either syntactic or semantic. For syntacticians, what is missing are words and syntactic structure, a stance apparently confirmed by the fact that in some cases, particularly in languages with productive case morphology (such as Greek), the morphological form of the fragment has to match what would be appropriate if spelled out overtly: (11)

I Maria den egrapse to grama? Oxi, ego/*emena [modern Greek] Didn’t Maria write the letter? No, INom/*meAcc

On this syntactic view, the structure underpinning the elliptical form is as complex as that associated with the interpretation of the form of the preceding sentence of which it is a covert copy, a form subject to phonological deletion at some late stage of the derivation. This stance has the disadvantage of leading to an explosion of ambiguity, each fragment being assigned as many structures as the interpretations to be recovered. Semanticists, on the other hand, argued that it is meaning that is reconstructed at the ellipsis site by a process of abstraction from some previous sentence’s content, with the advantage of this approach being not to need to postulate ambiguity unnecessarily (Dalrymple et al. 1991, and others following). Yet neither of these is sufficient. There are fragments which resist any characterisation in terms of some sentence source providing the required interpretation (Stainton 2006): (12) A (coming out of the lift in an office block addressing an employee): McWhirter’s? B Second on the left. So, at best, grammar-internal accounts of ellipsis have to admit to incompleteness of their accounts, with no overall explanation of the phenomenon itself. 200



Modelling Interactive Syntax

It is clear that something is missing in these accounts of ellipsis. Strikingly, no notice is taken of the process of setting up the interpretation. Protagonists merely disagree on how to capture string-interpretation pairings, simply assuming that all elliptical forms are incomplete sentences in order to see the phenomenon as falling within the sentence-based methodology. Yet the insight Kamp brought to the study of anaphora was that it is the process of building up interpretations that lies at the heart of context-dependence in natural language, and this applies equally to ellipsis, as we see in turning to the occurrence of ellipsis in conversation.

12.2.3  Split Utterances in Dialogue: The Challenge of Incrementality The first thing which strikes one from even a cursory glance at conversational dialogues is how prevalent the use of elliptical fragments is in conversation. Very commonly, we do not speak in complete sentences. Well-formed sentence structures emerge through a sequence of fragmentary contributions, each participant adding some fragment to a partial structure, with fluent switching of speaker–hearer roles, not only across but also within sentential structures. The effect is rich interactivity between participants in a dialogue: (13)

A: B:

I need a a mattock. For breaking up clods of earth

(14)

Conversation from A and B, to C: A: We’re going to B: Burbage C: Burbage? A: where Auntie Ann lives. C: with the dogs? B: in the Honda, where they’ll have plenty of room.

The problem for the grammar-writer is that though such speaker-switching may take place at what might be construed as sentence boundaries, this isn’t always necessary. People can interrupt each other at any point in a clausal sequence, even very early on, as in example 15, and by no means only collaboratively with respect to this content, as in examples 16 and 17: (15) A: I had to go back to the hospital for a follow-up appointment. The doctor B: Chorlton? A: Mhm. He said I had a shadow on my lungs. [BNC] (16)

A: B:

Did we… embarrass her? I hope not.

(17)

A: B:

It’s obvious from what he says that you are wrong

Split dependencies can, moreover, be arbitrarily complex: (18)

A: B: A:

Has every student handed in their homework? or even any assignments?

201

Ruth Kempson

The overall phenomenon is that, in language use, hearers in a discourse can take a­ rbitrary complete or partial structures as context, and use these as the point of departure from which to switch into speaking and so overtly contribute to the ongoing utterance. Conversely, a speaker can take where they have got to in producing some utterance as the background context relative to which they can switch into understanding what is then said to them. The seamless fluency with which individuals take on or hand over utterance responsibility presents a formidable challenge to current grammar formalisms. The problem is that these points of switch can separate sub-parts of a structure, splitting dependencies of every sort, whether syntactic or semantic; yet the sententialist methodology relative to which linguists explain these dependencies precludes any straightforward modelling of the data, as the fragments which each individual speaker offers may not even be what the grammar licenses as independently well formed. How can an individual speaker control such fragments if the individual parts which they offer as their contribution to conversational exchanges are not licensed as part of the output of the grammar? Adding to the puzzle, the various parts which speakers contribute in such exchanges do not always go together straightforwardly to yield a well formed sentence string: (19) (Ruth emerging from a smoking kitchen) Ruth: I’ve burnt the kitchen rather badly. Mike: Have you burnt Ruth: Myself? No. In example 19, the relevant subparts, put back together, yield Have you burnt myself, an ungrammatical string, yet this is a perfectly well formed exchange in which the reflexive pronoun establishes an antecedent that is appropriately local. What such data indicate is that these exchanges are not about putting word sequences together to form some single sentence string. In 19, A takes over the projection of the second part of the utterance, extending its content by the words she chooses in a way that enables her to project a dependency of the object on the subject term provided by B, while simultaneously reflecting the fact that it is she that has now become the speaker. So the projection of structure has not only to represent content, but also context in just the same terms as that content: so both content and context are partial structures, and both evolve. This indicates the extent of the challenge these data pose. The dependencies have to be defined over semantically transparent partial structures, with whatever imposed context-relativity is determined by the words chosen, but nonetheless echoing the required level of granularity to match whatever constraints the individual language imposes. Such data are a challenge for pragmatics and the study of the underpinnings of communication. The assumption shared by pragmatists is that successful communication necessarily involves the hearer grasping the proposition which the speaker either has or could have had in mind in uttering a given sentence, with success in communication supposedly depending on the progressive build up of such propositional knowledge which all parties to the exchange can assume is shared “mutual knowledge” (Grice, 1975) or “mutually manifest” (Sperber and Wilson 1995). However, the assumption that it is some intended propositional content that needs to be grasped is not sustainable as a prerequisite for utterance understanding for, in many cases, there may be as yet no identifiable propositional content constructed by the hearer and yet they have understood sufficiently to warrant their own contribution, leading to a resulting exchange which may nonetheless be successful. The many examples above illustrate this; however, consider also example 20 in which the son is resisting all attempts to get him to be helpful. He is 202



Modelling Interactive Syntax

certainly not waiting for the third of the commands despite the fact that its emergence is clearly signalled, even if he has bothered to process the second, but nevertheless he has responded wholly appropriately: (20) (A mother, B son) A:  This afternoon first you’ll do your homework, then wash the dishes and then B:  you’ll give me 10 pounds? And it is not merely that it may be unnecessary to achieve agreed content in all successful communicative exchanges. Even the presumption of there being some fixed single intended speech act may be in question, since a single fragment may be able to serve more than one such function in some sense simultaneously with the success of the exchange turning on the uncertainty of the speech act being achieved: (21) Lawyer: Client:

Will you choose your son as your attorney or My wife

Stepping back from the details, what we seem to need, if these various puzzles are to be resolved, is a language system that can model evolving contents, evolving contexts, and even evolving intentions, all characterized in ways that allow for incremental word by word accumulation and a corresponding growth of interpretation without necessarily having access to some complete thought that anyone has in mind to say.2

12.3  Dynamic Syntax Though the dynamics of ongoing dialogue has received little attention from the core linguistics community, those working at the interface between linguistics and sociology in the conversational analysis paradigm have long been sceptical of the dichotomy between competence and performance, and the attendant commitment to an over-deterministic methodology. Their insights have recently been taken up by those developing computational models of dialogue (Poesio and Rieser 2010; Ginzburg 2013). One paradigm which seeks to directly model a more flexible perspective is that of Dynamic Syntax.3 Dynamic Syntax (DS) is a grammar framework that defines representationalist model of interpretation of which the core structural notion is incremental word-by-word interpretation growth relative to context. On this view, the setting out of such dynamics constitute the grammar: syntax just is a set of principles for inducing growth of such structures, procedures that define how parts of representations of content can be introduced and updated, all such growth being relative to context. Moreover, context is as structural and dynamic as the concept of content with which it is twinned, a record of the emergent (partial) structures that represent the unfolding of structures for content plus the actions used to develop this incremental process. So the overall picture is a shifting one: shifting goals, shifting contents, and shifting contexts, as each established content becomes part of the context for the next step in the process. These goal-driven sequences of action are taken to involve building partial trees, with, as output, a tree whose nodes reflect the content of some uttered formula. We start with the idealised case of a sentence uttered in isolation, as in example 22, whose processing is diagrammatically presented with input to the construction process and its endpoint in example 23: 203

Ruth Kempson

(22) Jo fainted The input point of departure is some tree whose only decoration gives a skeletal indication of the goal which has to be achieved, a tree of type t, otherwise unspecified ?Ty(t), given on the left of the → in Figure 12.1. What has to be built is a proposition matching this type on the basis of what the context already provides, here null, plus information provided by words used, as achieved by the formula on the right of → in Figure 12.1:4

Faint ¢ (Jo¢)(SPAST ) : t,

?T y(t),

SPAST : es

Faint ¢ (Jo¢) : es Jo¢ : e

Faint ¢ : e

t (es

t)

Figure 12.1  Processing Example 23: Input and Output.

The steps of transition from that starting point to the achievement of that goal will involve constructing partial trees which step by step, following the words in the order selected, lead to the output tree. This minimal display gives barely even a hint of the dynamics of the processes, but nevertheless the pattern is general. Each node in a complete tree has a content formula to the left of “:” a type specification to its right (e for entity, es for event, e → (es → t) for predicates). The trees are binary, with by convention the functor node (the one indicating what is to be combined with what) on the right node, and the argument (the simple entitytype nodes) on the left. Individual terms are invariably of type e in the system, presuming on an account of quantification, not given here. There is a particularised type es for event terms, only schematically represented here. The dynamics of growth is partly top down, as general strategies open up possibilities, and partly bottom up, as actions induced by the words provide terminal decorations which lead to bottom-up compositionality for the resultant tree. Such trees are built both in parsing and production. The only difference between the two is that, in parsing, there is only the weak goal of trying to build some propositional formula, whereas, in production, there is some more particular goal to achieve, so every update associated with a putative word must be extendable towards that goal tree to allow that word to be produced.

12.3.1  Formal Properties of Trees To flesh out this concept of growth, we need vocabularies for defining trees, concepts of tree growth, and actions that induce such growth. Dynamic Syntax (DS) adopts a modal logic (Blackburn and Meyer-Viol 1994), with two basic modalities: and . α holds at a node if α holds at its daughter (one node down), with variants and for argument and functor daughter relations respectively. α: ‘α holds at a mother node, one node up, if α holds at its mother’, equally with argument and functor variants indicative of the status of the daughter-mother relation so identified. There are so-called Kleene star operators which yield concepts of dominate and be dominated by. 204



Modelling Interactive Syntax

Based on these, there are defined Kleene star operators which yield concepts of dominate and be dominated by: Tn(n) holds at a node when a node Tn(n) is somewhere below it, Tn(n) holds at a node when a node Tn(n) is somewhere above n. There are then the analogue Tn(n) and Tn(n) defining a functor spine within a tree, and these will give us concepts of locally dominate and be locally dominated by, which we can use to define structures projected from a single clause (see Section 4.3.1). The core pair of concepts underpinning the tree growth process is that of underspecification. There are three principal different types. Trees may be partially specified because one or more nodes remain yet to be constructed. Formulae decorating the nodes may be underspecified for content, classified with a semantic type but as yet without assigned formula: this is the way pronouns and predicate ellipsis are modelled, assigning them some place-holding meta-variable of given type, for example, U : e for pronouns, U : e → (es → t) for predicate ellipsis. Even relations between nodes may be incompletely specified giving rise to a so-called ‘unfixed’ node. Any node so constructed will remain in need of subsequent updating (this is the mechanism taken to underpin so-called long-distance dependency), the primary motivation for transformational rules in Chomskian paradigms, and bare argument fragments. All such underspecified parameters are twinned with requirements, ?X, for some annotation X. Requirements are essential for reflecting the time-linearity involved in progressively building up (partial) trees, as they allow the construction process to continue incrementally while yet not providing a fixed value at that point to fully satisfy that requirement. This is provided often substantially later, driven by the requirements which have to be satisfied to achieve a successful outcome. For example, for some initial goal, formulated as ?Ty(t) in example 23, there will be a number of available strategies, each of which will open up subgoals, the satisfaction of which will involve a number of intermediate steps, and it is only in the final step of the derivation that the initial requirement will be satisfied. Furthermore, context evolves along with such emergent structural representations of content, and every step of update, once achieved, becomes part of the context relative to which the next update takes place. To model how words map onto these partial trees, a language is defined to express tree growth actions. There are four primitive actions: make(X), go(X) and put(Y) operations, where X and Y are tree relations and node-decorations respectively, plus an ABORT action, which terminates an action sequence. The simplest way of seeing the effect of these is to see how words induce a set of conditional actions which directly yield partial growth of a tree. Verbs are the best illustration, as they induce the most structure, a sequence of actions, stored as a macro, that induces a tree by constructing all the necessary tree-node relations in the minimal predicate-argument structure which its predicate is to inhabit, doing this node by node—building an argument node, moving to it, decorating it, going back to the mother, building a functor node, moving to it, decorating it, building from it a further argument daughter and so on—repeating the subsequence of actions until an entire propositional template emerges with appropriate nodes for further development, each node decorated with whatever information the verb form happens to encode. The effect, spider-like, is to create a composite local predicate-argument structure from a sequence of strictly local manoeuvres. Thus faint induces in a step-wise fashion a structure dominated by a node decorated ? Ty(t), building two argument nodes, one of which is an event-term node, its details to be provided by tense specifications, and a functor chain along which there is the predicate node decorated by the predicate Faint’:5

205

Ruth Kempson

fainted IF Tn(a), ?Ty(t) THEN make((↓0)) : go((↓0)); put(Ty(es), Fo(UPAST ), ?∃xFo(x)); go (24) make; go; put(?Ty(e → (es → t))); make((↓1)); go((↓1)); put(Fo(Faint’), Ty(e → (es → t))) go; make; go; put(?Ty(e), Tn(a)); go ELSE Abort Languages vary as to what array of options for argument decorations they allow. English finite verb forms, like other Germanic languages, specify argument nodes merely with a requirement of the form ?Ty(e), since further lexical processing is needed to provide appropriate decorations for these nodes.6 It is notable that the only node induced by this sequence which lacks any requirements for later update is the node decorated by that predicate term. However, even this is a lexical idiosyncrasy. Some words project rather less than a representation of content, specifically pronouns, VP-anaphoric expressions and auxiliaries. What they encode, as formula decoration, is just a typed place-holder, their value in that context having to be provided by something else. Thus this vocabulary provides a tool for specifying not merely the way words induce partial structures, but also how they may underdetermine content within that structure. The value for such place-holders will then come either from context already available or from some later step in the construction process, an update needed for successful completion of the tree. The effect of such lexical specifications is a pairing of a word with a partial structure. However, there is more to syntax in this framework than just word-structure pairings. Opening stages in the tree construction process involve strategies for early tree-growth development; in the closing stages, strategies compile up content on nonterminal nodes to yield completed trees. The simplest of the early construction steps, and central to the system, is the license to construct a node characterised solely as being unfixed within a given tree domain, the third type of underspecification. This is easily expressed using the operator. A transition from the point of departure to some ‘unfixed’ node is licensed within this domain by introducing a node along a relation characterised solely with the treenode identifier indicating that ‘somewhere above me is the Tree node a’, represented as Tn(a), with its attendant requirement for a fixed tree node position. There is a condition on this action that it can only take place if no other node from this point of departure already exists within this newly emergent tree:7 IF ?Ty(t), Tn(a) ?Ty(t), Tn(0) THEN IF ⊤ THEN Abort ELSE make; go; Ty(e), put(Tn(a), ?Ty(e),) ?∃x.Tn(x,) put(?∃xTn(x)) Tn(0) ELSE Abort ♦ Figure 12.2  Building Unfixed Nodes : *Adjunction.

206



Modelling Interactive Syntax

Such a conditional format is not an idiosyncrasy of this particular type of construction-action. All actions are given in this conditional format: indeed, the force of this account is that the system is action-based, defined as a series of conditional actions able to reflect, and so license, the incremental process of building structural representations of content. We can now see how the process of building structured interpretation unfolds in the simplest of cases, example 22. The first step is to induce an unfixed node, which is then decorated by the actions given by the word Jo, which we assume here is just the concept of picking out someone with that name (no details given). The second step is to build the structure as driven by the lexical specification of fainted which updates that emergent tree with its skeletal propositional template. All that remains is to combine the functor term with its arguments progressively up the tree to yield the output specified in example 23. This then becomes part of the context for what may come next. Of course, there is much more to be said, in particular because these actions build relative to context, which may include partial trees, but this sets the pattern. All structure is built by progressive pairing of words and a representation of interpretation built from them, with each such induced structure becoming part of the context for establishing the next update. Though we have only taken the simplest possible derivation, we have sufficient to show how locality constraints can be defined over such binary-branching trees and the way they unfold. In any local tree for a simple clause, for example, there will be an unbroken chain of functor steps along its right edge (as in example 23). This is because the first non-­functor node above any node decorated by the predicate will be the type t node. Constraints invoking concepts of locality can then be identified as a constraint on the term to which some locality restriction applies having to be resolvable within the domain characterisable as somewhere above, the point of characterisation along a functor chain. For example, for reflexive pronouns, the antecedent for a local anaphor has to meet the condition Tn(a) (‘T n(a) holds up one argument-relation plus a possibly empty sequence of function-path relations plus one argument relation down’). Contrarily, pronouns are definable relative to a constraint that any putative antecedent may not occur in this domain.

12.3.2  Split Utterances: An English Case-Study Even with such simple structures, we have sufficient to see an immediate bonus. The phenomenon of split utterances is no longer a puzzle, because the possibility of splitting any dependency across more than one speaker is an immediate consequence of the general dynamics presumed by the model. Being a grammar formalism, production is taken to make use of the same tree growth mechanisms as parsing, hence equally incremental, with just one additional constraint. Every putative construction step in production has to be checked against some richer partial goal tree representing what the speaker is intending to say, to make sure that the actions encoded as the next step are extendable into that tree. In all other respects, the transitions induced by parsing and production are identical. This commensurability of speaker and hearer mechanisms means that, in all exchanges, speaker and hearer will be severally building up a tree to serve as the basis for whichever process they are engaged in, the one as parser, the other as producer: each will be incrementally updating their own semantic representations. In consequence, the speaker who midway has to shift into becoming hearer can effectively continue to mirror the same series of partial trees as

207

Ruth Kempson

her interrupting co-participant after the switch, albeit no longer knowing the precise content of what is to come. And, for both of them, whatever structure the producer or parser will have just constructed will become part of their context relative to which they can continue the utterance construction process, despite a shift of roles. Let’s take this up step by step for example 25, a simplified variant of example 19: (25) Mary: Did you burn Bob: myself? No. (26) displays the partial tree at the point of Mary’s utterance having been processed, involving a substitution of the meta-variable projected by you with the name of the interlocutor/parser (setting aside how to capture questionhood). At this point, Bob can complete the utterance commensurate with his own goal using the reflexive, because this expression, by definition, copies a formula from a local co-argument node onto the current node just in case that formula satisfies the conditions set by the person and number of the uttered reflexive, and, in this case, that it names the current speaker.

Mary:Did you burn

?Ty(t), Q SPAST

?Ty(es Ty(e), Bob¢

t) ?Ty(e

?Ty(e),

(es Ty(e

Bob:myself?

t)) (e (es Burn¢

t))),

?Ty(t), Q ?Ty(es

SPAST Ty(e), Bob¢

t) ?Ty(e

Ty(e), Bob¢,

(es Ty(e

t)) (e (es Burn¢

t))) ,

Figure 12.3  Modelling an Utterance Split.

The fact that the structure may have been induced in part by someone else is irrelevant: that partial tree still under construction will be part of his own context. This phenomenon, as here modelled, is specific to the construal of a reflexive pronoun, said to be subject to a Principle A constraint (Chomsky 1981), and the associated Principle B constraint imposed on a nonreflexive pronoun debarring it from picking up an antecedent from this domain is equally simple to express. In both cases, the restrictions on interpretation are expressible as a restriction definable over the emergent semantic representation, hence a restriction on

208



Modelling Interactive Syntax

processing. Though this illustrates the simplest type of split-utterance ellipsis, the point is general. This seamlessly achieved splitting of utterances can separate off any expression from the syntactic environment it needs for its well formedness because both speaker and hearer use the same mechanisms. The same individual whether as speaker or as hearer will invariably have a partial structure on which to rely at the point of participant switch without need of any recourse to reasoning over what their interlocutor might have had in mind. The resulting effect of interaction is secured by the mechanisms of the grammar directly. We can now see how these split utterances are no more than one sub-type of ellipsis. Characterised in general terms, ellipsis is where contributors can so reliably rely on context to establish the intended construal that nothing further needs to be added. The cases addressed in the literature (see Section 3.2) were cases involving a complete sentence as the antecedent context, following the sententialist methodology: (8)

A:  John is writing a paper. B:  Bill is too.

But as we would now expect, such a restriction is not warranted. All operations can take place relative to partial contexts, so the conventional accounts of ellipsis, at best, were only providing a partial picture; split utterances merely round out the ellipsis picture by showing how elliptical fragments can rely for their interpretation on partial structures as well, adding to the effect of interaction which, like anaphora, they equally give rise to.

12.3.3 Discontinuities as Underspecification and Update We now turn to the phenomenon of discontinuity, notorious as needing an encapsulated language-particular formulation, in which the initially placed constituent may be arbitrarily far away from its point of resolution: (27) Mary, John trusts. (28) John, I suspect Mary is much less sure of. ?Ty(t), Tn(0)

?Ty(t) Mary ¢,

Ty(e), ? x.Tn(x), ( )Tn(0),

Mary,

Ty(e), ? x.Tn(x), ( )Tn(0),

SPAST

?Ty(es

John¢,Ty(e)

t)

?Ty(e ?Ty(e), Ty(e

Building the unfixed node

(es

t))

Trust (e (es

t)))

Integrating the unfixed node into the emergent tree step (i)

step (ii)

Figure 12.4  Building and Resolving an Unfixed Node: Long-Distance Dependency. 209

Ruth Kempson

Though this structural dependency is standardly taken to be an irreducibly syntactic phenomenon, it can be straightforwardly expressed in this tree-growth perspective. Example 27 is modelled along lines diagrammatically sketched in example 29. First the computational action of constructing an unfixed node is taken up which the left-peripheral expression decorates. This ‘unfixed’ node then has to have its hierarchical position within the tree fixed, and this will be established once a putative update becomes available. In such simple cases as these, with only the subject node intervening, this is upon the processing of the verb, with only the subject node intervening when the open requirement for its object value, and the requirement of a fixed relation for the unfixed node, can both be satisfied by unifying the two nodes in question. However, given the restriction expressed in the characterization of the unfixed node as Tn(0), ‘this is a node somewhere within the tree developed from Tn(0)’, all that is required in the subsequent tree development, is that this unfixed tree relation be resolved somewhere in that emergent tree, in example 29, this being at the object node. Subsequent to this unification, the general process of applying the functor term at each nonterminal stage to its argument will lead to full decoration of the tree.8 As expected, these structures can be split across parties: (30) A:  Mary, John... B: trusts? A:  Well, she clearly likes him. The structure with which A starts off is the building of an unfixed node, identified weakly as Tn(0) and then decorated by processing Mary. It is only when the structure emerges through the parsing of John plus the verb trusts (which B provides) that this initially underspecified tree relation can be resolved, so giving rise to the effect of interaction across parties. As elsewhere, neither party needs to be reading the mind of the other in order to achieve this effect. Thus the pattern of distributing discontinuities across more than one participant extends even to these supposed hard-core cases of syntactic dependencies. The phenomenon of what looks like high-level decision making to enable their various utterances to dovetail into each other’s contribution is resolved by the assumption that each party to such exchanges is in fact building relative to the partial tree they are themselves in the midst of constructing whether as speaker or hearer. The perspective of tree-growth for the articulation of grammars for natural language is thus providing a direct action-based explanation of the systemic interaction which individuals can achieve in conversational dialogue.

12.3.4  Extending the Parallelism of Resolving Underspecification to Syntax Finally, we can see that this mechanism of building unfixed nodes displays the very same parallelism as is shared by anaphora and ellipsis, that it can be either enriched from the construction process, or from previous utterance context and even indexically from the general scenario in which the utterance occurs. We have already just seen the forward resolution of such an unfixed node from within the construction process, and interactively, in example 30. This is the effect of long-distance dependency. The building of such an unfixed node may also, however, rely on previously constructed context to provide an update interpretation: this is what underpins a common form of ellipsis called stripping/bare-argument ellipsis in which the fragment is just a bare argument expression (possibly with too following):

210



Modelling Interactive Syntax

(31) A:  John admires Mary enormously. B:  Sue too. These fragments allow a construal recovering any action sequence just used in context that will match the type requirements of combining with an entity-denoting expression; example 31, for example, can be understood with Sue either as the object of John’s admiration or as the one who also admires Mary (see Kempson et al. 2015 for details). Like long-distance dependency, all that matters is that the interpretation of the fragment unify with some node during the construction of the overall tree emerging from the construction process. And the effect can indeed be one of interaction: one person can set up some pattern of interpretation, which the other then makes use of. And, rounding out the parallelism with anaphora, most strikingly with very young children, who in the one word utterance phase, present an expression to be taken as a trigger for initiating some containing structure with only the information that a term constructed from the expression must play some role within it. Structurally this amounts to inducing an unfixed node, decorating it with the term associated with the expression they have uttered, and yet without any further containing structure for it to combine with. This they leave to the person they are talking to. (32) Eliot (grandson on back of mother’s bike waving at empty mooring on the other side of the canal): Daddy. Mother:   That’s right dear, you were here yesterday with Daddy clearing out the boat. Here we have a transparently indexical use of the fragment. The child is entertaining some predicate to be applied to his fragment, relying on the routine of interactivity between him and his carer to be sufficient to ensure a wholly successful communicative exchange. Details aside, what is significant here is the demonstration that the mechanism underpinning long-distance dependencies—that of building an unfixed node and resolving that underspecification—can be established across different participants in all the ways that other context-depending uses of language do. Though regularly said to be an impoverished form of language, such one-expression utterances are common in adult life too, as witness example 33 when said halfway to the office, upon realising that one has left one’s phone behind or the car lights on, or such like: (33) My phone/my lights/...

12.4 Reflections The significance of this resolution of unfixed nodes constructed in anticipation of resolution either forwards, backwards, or indexically, in all cases allowing distribution across more than one interlocutor, should not be missed. Long-distance dependency, in particular, has been the core of syntactic arguments over decades as an irreducibly syntactic phenomenon justifying the necessity of positing an encapsulated syntax sealed off from other cognitive processes, and this insistence by the linguistics community has led to other cognitive science research having to posit high levels of inference to address the phenomena which such a stance provided no basis for explaining. But with the shift into a perspective where the emphasis is on processing actions, structural properties of language can be explicated within the general perspective of modelling the process of constructing word-concept pairings 211

Ruth Kempson

across the time-span of interactive exchange, a task now, by assumption, for the grammar. And with this shift, the potential which language provides for interaction inexorably emerges because the interaction comes from mechanisms of grammar, without any need to argue for superimposed higher-order reasoning, or mind-reading, or composite simultaneous duplication of production and parsing resources (as advocated by Pickering and Garrod 2013). I note in closing that such a characterisation of the language system has much in common with new developments falling under the broad umbrella of embodied cognition (see in particular Pezzulo 2011; Clark 2013). The horizon which thus opens up is the promise that, by articulating a view of language as a set of mechanisms for directly determining interactive behaviour, we are at last able to propose models of language that are naturally embeddable within an overall dynamic cognitive architecture.

Notes 1 This chapter is grounded in work done on The Dynamics of Conversational Dialogue (ESRCRES-062-23-0962). I am immensely grateful to my co-workers on this project and too many others to thank who have contributed to the fine landscape which this work has opened up. 2 For a related perspective on communication see Linell, Chapter 7 this volume, who argues that intersubjectivity is too strong a grounding of natural language communication, and that what it needs to be replaced by is due recognition of interactivity. 3 In seeking to sustain conservative, sententialist assumptions for grammar design, Ginzburg (2012) and Poesio and Rieser (2010) face all the problems attributable to that perspective. 4 ♦ is a so-called pointer which at each stage indicates which node is to be the next one under ­development. 5 The account is simplified here in ignoring the restriction that finite (non-passive) verb forms in English cannot be clause initial unless in imperatives, and that whatever argument node is already in the emergent tree has to be construed as the logical subject. 6 Verbs in so-called pro-drop language may license their argument nodes with a place-holding metavariable, indicating their value can be identified directly from context. 7 The tree beside this rule formulation displays diagrammatically, using a dashed line, the partial tree relation that results. Such tree diagrams record only the effect of the rule, and not conditions on its application. 8 To articulate limits within which this underspecified relation has to be updated is to analyse the various environments within which long distance cannot be resolved as independent trees: see Cann et al. (2005) for accounts of relatives in these terms.

References Asher, Nicholas and Lascarides, Alex (2003) Logics of Conversation. Cambridge: Cambridge University Press. Barr, Dale (2004) Establishing conventional communication systems: Is common knowledge necessary? Cognitive Science, 28(6), pp. 937–962. Blackburn, Patrick and Meyer-viol, Wilfried (1994) Linguistics, logic and finite trees. Logic Journal of the Interest Group of Pure and Applied Logics, 2(1), pp. 3–29. Brandom, Robert (1994) Making It Explicit. Harvard: Harvard University Press. Cann, Ronnie, Kempson, Ruth and Marten, Lutz (2005) The Dynamics of Language. Oxford: Elsevier. Carston, Robyn (2002) Thoughts and Utterances. Oxford: Blackwell. Chastain, Charles (1975) Reference and context. In Gunderson, Keith (ed.) Language, Mind and Knowledge: Minnesota Studies in the Philosophy of Science Vol. 7, pp. 194–269. Minneapolis: University of Minnesota Press. Chomsky, Noam (1965) Aspects of the Theory of Syntax. Cambridge: MIT Press. Chomsky, Noam (1981) Lectures on Government and Binding. Dordrecht: Foris. 212



Modelling Interactive Syntax

Clark, Andy (2013) Whatever next? Predictive brains, situated agents, and the future of cognitive science. Behavioural and Brain Sciences, 36(3), pp. 181–204. Clark, Herbert (1996) Using Language. Cambridge: Cambridge University Press. Clark, Herbert and Brennan, Susan (1991) Grounding in communication. In Resnick, Lauren Levine, John and Teasley, Stephanie (eds.) Perspectives on Socially Shared Cognition. American Psychological Association 13, pp. 127–149. Dalrymple, Mary, Shieber, Stuart and Pereira, Fernando (1991) Ellipsis and higher-order unification. Linguistics and Philosophy, 14, pp. 399–452. De Saussure, Ferdinand (1916) Course in General Linguistics. Frege, Gottlob. (1970) Translations from the Philosophical Writings of Gottlob Frege. Geach, Peter and Black, Max (eds.). Oxford: Blackwell. Ginzburg, Jonathan (2012) The Interactive Stance: Meaning for Conversation. Oxford: Oxford University Press. Gregoromichelaki, Eleni, Cann, Ronnie and Kempson, Ruth (2013) Coordination in dialogue: subsentential speech and its implications. In Goldstein, Laurence (ed.) Brevity, pp. 53–73. Oxford: Oxford University Press. Gregoromichelaki, Eleni, Kempson, Ruth, Purver, Matthew, Mills, Gregory, Cann, Ronnie, MeyerViol, Wilfried and Healey, Patrick (2011) Incrementality and intention-recognition in utterance processing. Dialogue and Discourse, 2(1), pp. 199–233. Grice, Paul (1975) Logic and conversation. In Cole, Peter, and Morgan, Jerry (eds.) Syntax and Semantics 3, pp. 41–58. Chicago: Academic Press. Kamp, Hans (1981) A theory of truth and semantic representation. In Groenendijk, Jeroen Janssen, Theo and Stokhof, Martin (eds.) Truth, Interpretation, and Information, pp. 1–32. Dordrecht: Foris. Kamp, Hans and Reyle, Uwe (1993) From Discourse to Logic. Dordrecht: Kluwer Academic Publishers. Kempson, Ruth, Gregoromichelaki, Eleni and Howes, Christine (eds.) (2011) Dynamics of Lexical Interfaces. Chicago: CSLI Press. Kempson, Ruth, Cann, Ronnie, Eshghi, Arash, Gregoromichelaki, Eleni and Purver, Matthew (2015) Ellipsis. In Lappin, Shalom and Fox, Chris (eds.) Handbook of Contemporary Semantics (2nd edition). Oxford: Blackwell Publishing. Keysar, Boaz (2007) Communication and miscommunication: The role of egocentric processes. Intercultural Pragmatics 4, pp. 71–84. Pezzulo, Giovanni (2011) Shared representations as coordination tools for interaction. Review of Philosophy and Psychology, 2(2), pp. 303–333. Pickering, Martin and Garrod, Simon (2013) Forward models and their implications for production, comprehension, and dialogue. Behavioral and Brain Sciences, 36(4), pp. 377–392. Poesio, Massimo and Rieser, Hannes (2010) Completions, co-ordination and alignment in dialogue. Dialogue and Discourse 1: pp. 1–89. Recanati, Franҫois ( 2003) Literal Meaning. Oxford: Blackwell. Russell, Bertrand (1905) On denoting. Mind, 114(456), pp. 73–87. Schiffer, Stephen (1972) Meaning. Oxford: Clarendon Press. Sperber, Dan and Wilson, Deirdre (1995) Relevance: Communication and Cognition (2nd edition). Oxford: Blackwell. Sperber, Dan and Wilson, Deirdre (2002) Pragmatics, modularity and mind-reading. Mind and Language, 17, pp. 3–23. Stainton, Robert (2006) Words and Thoughts. Oxford: Oxford University Press. Thomason, Richmond (ed.) (1974) Formal Philosophy: Selected Papers of Richard Montague. New Haven: Yale University Press.

213

13 The Concept of Language in an Utterance Grammar Edda Weigand

13.1  The Issue: Shifts in the Concept of Language What is language? The responses to this central question of linguistics and related disciplines have changed and reflect the changes linguistics went through in the twentieth and at the beginning of the twenty-first centuries. The preceding chapter by Ruth Kempson focuses on an important shift which mainly became manifest after the pragmatic turn and demonstrated that language is much more than an abstract system of signs or the product of language competence. The analysis of authentic examples made clear that the terms and techniques of structural and generative linguistics do not suffice to address language-in-use. The shift in the concept of language from a system of signs to authentic language use, of necessity, caused a shift in the type of approach, in the case of Ruth Kempson’s chapter a shift from sentence syntax to interactive syntax. The new object of authentic language use does not, however, only relate to syntax or to verbal means because language cannot be isolated as an autonomous object-of-study. Cognition and perception are inevitably included when human beings use language in dialogic interaction. The question is how language is integrated within the totality of the dialogic means that are used to achieve the purpose of dialogue. What is at stake is the concept of language in a grammar of language use which requires a shift from sentence grammar to utterance grammar or from the Cartesian grammar of a pattern to a grammar that goes beyond patterns.

13.2  Grammar and Performance A grammar beyond patterns can no longer be restricted to a limited set of utterances (Weigand, 1984). Before reflecting upon its structure we should clarify what a grammar is about. What do we expect from an object called ‘grammar’, and what consequences have to be drawn from such an object for methodology? The relation between object and methodology is crucial for any scientific approach. Before the pragmatic turn we were convinced that our starting point had to be the model, that is, methodology; after the pragmatic turn we no longer wanted to distort our natural object by methodology (Martinet, 1975). We accepted the complexity of natural language use and this resulted in new concepts of language as 214



Language in an Utterance Grammar

dialogue and of dialogue as a mixed game (see Chapter 11, this volume). The methodological consequences which ensued meant a turning point in science from reductionism to holism (Weigand, 2011).

13.2.1  The Concept of Grammar: From the Simple to the Complex Returning to our question of what a grammar is about, an utterance grammar is not a separate part but a component of the complex whole of the mixed game. The focus is on the utterance as correlation of meaning and expression. The terms ‘meaning’ and ‘expression’ that were coined in structuralistic times have to be redefined and broadly extended. There are no signs as expressions which have by definition meaning; it is human beings who assign meaning to the dialogic devices they use in utterances. Utterances are the carriers of action. Human beings are at the centre of the game and therefore also at the centre of an utterance grammar. Human beings play the game of dialogic interaction in order to achieve their purposes, goals and desires related to certain events and they use devices they have at their disposal due to their abilities and to the conditions of the game. In this sense an utterance grammar focuses on the correlation of the purposes human beings have in dialogic interaction and the dialogic means they use to achieve them: utterance grammar

meaning purposes (events)

expression dialogic means

Figure 13.1  Correlation of Meaning and Expression in an Utterance Grammar.

In the same way as actions are dialogically oriented in the mixed game, the utterance is also dialogically oriented, either as initiative or as reactive. The problem that utterances cannot be separated into distinct units at the empirical level represents a pseudoproblem. The utterance is not a unit of verbal means but a unit of correlation and includes cognitive as well as perceptual means. Utterances can only be delimited at the action level. Defining the utterance as carrier of the action means the utterance begins and ends with the speech act. An utterance grammar which aims to demonstrate how human beings use language in order to achieve their goals cannot start with picking out verbal elements and calling them data. ‘Data’ presupposes that they have meaning as Chomsky (1959) brilliantly demonstrated in his criticism of Skinner. Meaning in performance is not a matter of single verbal elements but a matter of the role they play in the whole. An utterance grammar as a grammar of performance has to tackle the issue of individual language use which is not an issue of ‘anything goes’. Meaning has to meet the condition of understanding. Understanding relies on the Principle of Coherence (see Chapter 11, this volume) which is a crucial premise of an utterance grammar. Coherence is differently shaped in different languages. The utterance grammar I’ll delineate in this chapter has therefore to meet the conditions of a comparative grammar, that is, it has to start from the complex, quasi-universal whole and derive the particularities of individual languages and cultures by specialization. It goes without saying that this chapter can only describe the nuts and bolts of such a complex endeavour. 215

Edda Weigand

13.2.2  The State of the Art Recently many different types of approach have been put forth which claim to be a grammar of use. After the pragmatic turn the challenge was to come to grips with authentic language use. Authentic language use means individual use, spoken as well as written. Theorizing on spoken language cannot start by examining random cases. Spoken language only represents the tip of the iceberg. We cannot hope to be enlightened by ‘data’. The first question to be posed is why we use language. We use language in order to come to an understanding in dialogue. The purpose circumscribes the whole and determines the means—not vice versa. However, new types of so-called ‘grammars’ have been created on the basis of authentic spoken language, among them corpus-based grammars, discourse grammars, ‘construction grammars’ and ‘emerging grammars’. Such empiricist approaches ignore the basic premises of a grammar and of science in general. The notion of an ‘emerging grammar’ (Hopper, 1987) converts the object into the opposite: a ‘grammar’ cannot ‘emerge’ from actual utterances which having been uttered, immediately vanish, and are lost forever in this precise form. The hypothesis of an ‘emerging grammar’ cannot be compared with historical change or with language evolution as has happened over thousands of years. An utterance grammar, on the one hand, presupposes a conventional core in order to achieve understanding, and on the other hand allows for individual variations. Only these variations might be described as ‘emerging’ on the spot but they cannot be the object of a grammar. In order to come to grips with the chaos of performance, that is, with particular cases, we need to rely on standard cases which are, as it were, subject to the ‘laws of chaos’. In the same way, what is called ‘construction grammar’ cannot be any construction which catches the eye but needs to meet the conditions of the whole. Contrary to ‘construction grammars’ on the spot, corpus-analytic studies aim to provide comprehensive corpora of different languages, spoken and written, which demonstrate ways-of-use of verbal elements that go beyond our competence as native speakers. Corpus-analytic insights confirm that it is not single words but a network of phrases which determines the structure of the lexicon and the use of verbal means in general (see below). Besides empiricist types of grammar there is the neo-Gricean type which is based on the addition of orthodox abstract models and individual inferences. How could such an inconsistent step be the basis of a grammar of performance? In performance there is no ‘interface’ between truth-conditional semantics and individual action. Such artificial multi-level constructions (e.g., Levinson, 2000) remain games played in the ivory tower, speculations which cannot bridge the gap between competence and performance. Recently, models of multimodality have been gaining ground. They rightly demonstrate that empirical means are not only verbal means but include perceptual means such as gestures and movements. Like corpus-analytic studies they represent complementary approaches which provide useful information about the multiple variations on the expression side (see Chapter 3, this volume, and Mondada, 2016). Other types of grammar which are often referred to in the literature are, for example, ‘cognitive grammar’ (e.g., Langacker, 1987, 1991) and Halliday’s so-called ‘functional grammar’ (Thompson, 1996). Langacker’s cognitive grammar considers language to be a ‘configuration of concepts’, which is closer to generative grammar than to a grammar of language use. Halliday’s model claims to be a functional grammar but—despite the term— starts from the expression side of language. 216



Language in an Utterance Grammar

None of these attempts really succeeded in tackling the issue of a grammar of language use. The feeling is growing that reductive models or models based on the addition of parts are obsolete. More and more approaches are being introduced as holistic approaches. We should, however, not let ourselves be misguided by a construct which cannot do justice to genuine complexity and simply uses the term ‘holistic’ as a catchword. If, for instance, Simon Kasper (2015) calls his ‘Instruction Grammar’ a ‘non-reductionist’ programme, it is not really a holistic programme but a competence approach based on the addition of traditional parts such as syntax and semantics. The approach by Günthner et al. (2014) in Grammar and Dialogism also aims at performance but—incomprehensibly—uses a ‘pattern’ approach which claims to combine an empiricist construction grammar with Bakhtin’s ‘dialogism’. Dialogue is reduced to a ‘perspective’ which can be applied to various ‘constructions’. ‘Patterns’ are considered to arise between ‘emergence and sedimentation’. Even if in this way some useful observations of empirical details can be elaborated, the result on the whole is by no means convincing. It only represents a conglomeration of multiple terms which cannot conceal their vagueness and lack of justification. It goes without saying, that complexity depends on its object. Consequently, there is not only one type of holism. A genuinely holistic approach of language-in-use is quite different from a type of holism which we can find in nature, such as a whirlpool. Moreover, addressing the whole presupposes that the whole represents a largely autonomous object. Kretschmar’s book Language and Complex Systems (2015) deals with speech and the variety of dialects, by no means with the whole of language use. To sum up: The issue remains of how to conceive of a grammar of performance. How can language use be delineated in an utterance grammar? Actual language use is a matter of the moment and will never appear again. The conclusion we take from this cannot be to trace any piece of actual use as it ‘emerges’ and construct from it some sort of grammar which will, in fact, be no grammar at all. We have to acknowledge that the object of study has changed from a non-existing reductionist construct of language as a system of signs to a complex object of performance which requires a change in theorizing. The structure of an utterance grammar is dependent on the complex whole of dialogue as a mixed game (see Chapter 11, this volume). Complexity is a matter of simultaneously balancing the components, among them utterances as carriers of action.

13.3  Fundamentals of an Utterance Grammar At first sight it might seem that an utterance grammar aims to describe spoken language. However, describing language use goes beyond recording and transcribing spoken language and means describing how language use works. We should start with goal-oriented observation and take as our guideline how we ourselves proceed when we are confronted with the vicissitudes of life. We proceed by adaptation, that is, we first focus our attention on as many variables as possible and try to identify standard cases. Then, if necessary, we look for particular features in order to identify particular cases. I will introduce some fundamentals of an utterance grammar which need to be clarified. • Language in a grammar of use

Language is used in dialogic action games (see Chapter 11, this volume). The games are not chess games but mixed games based on the integration and interaction of the components (Weigand, 2010). The use of language is a basic component which is integrated with 217

Edda Weigand

perception and cognition. There is no language as such, only the ability to speak, which does not represent an autonomous object. The question of how language use works is dependent on the purpose we assign to language use. We do not use language for expressing our thoughts (e.g., Chomsky, 1988: 38); we use language in dialogue in order to come to an understanding about our mutual purposes (Weigand, 2009). As dialogic individuals we have to negotiate our goals and desires with our fellow beings. In order to arrive at some understanding, there must be some common ground (Clark, 1996), a more or less conventional core. In performance, however, we are faced with ‘the end of certainty’ (Toulmin, 2001), and even conventions are dependent on probabilities of use. In the end, when searching for an autonomous object we arrive at human beings’ minds and the ability to come to grips with performance. We orient ourselves according to rules as far as they go. Where conventions and regularities come to an end, we go beyond them and look for particularities according to Principles of Probability. It is our competence-in-performance which allows us to bridge the gap between competence and performance by adaptation and specialization in ever-changing environments. A grammar of use has to describe and explain this extraordinary ability of competence-inperformance which requires a change in theorising, from addition to integration and from separating parts to analysing them as components of the whole. The challenge of complexity can only be tackled by taking account of fundamental premises such as: The whole is more than the sum of the parts. Integration of components is the name of the game. • The unit of an utterance grammar

Any scientific approach needs to be oriented towards an object which should to some extent be an autonomous object. In this sense the object of an utterance grammar is human beings’ ability of being-competent-in-performance. Analysing such a complex object we need a key to opening it up. The key to human beings’ action and behaviour is their drive to interact which is rooted in their nature as goal-oriented beings and as dialogic individuals. Utterances as carriers of action can be taken as units of analysis. They are not exclusively empirical units but units of correlation, which implies that they need to be accessed from the level of action. The purpose of the speech act determines the utterances–not vice-versa. In a model of ‘language as dialogue’, any speech act is dialogically oriented, either as an initiative or as a reactive utterance, which is a feature of the action function and only secondarily of the dialogic means. At the level of the dialogic means the set of utterances which can be used for a specific speech act is not limited, mainly because perceptive and cognitive means are integrated. The limit is set by the Principle of Coherence, that is, by the condition of understanding what is meant in actual use. The ‘utterance’ means the authentic, spoken utterance, which implies we are dealing with the difference between oral and written language use, an issue which is not properly accounted for in the literature yet. The difference is not only a difference of the means but a difference of structure. Obviously in written texts we feel obliged to correct phrases which we do not consider appropriate, literary texts excluded, whereas we dismiss irregularities in oral language use. Nonetheless we can speak of authentic written texts if they emerge from actual written use in contrast to texts simulated by researchers. I cannot delve into this issue in this chapter, which is on the issue of an utterance grammar of oral language. I only want to point to the fact that some inner normative grammatical system seems to be accepted in written use (see Chapter 14, this volume), whereas oral language use represents a process 218



Language in an Utterance Grammar

on the spot which will never occur again in this form. In addition, written language use is restricted to verbal means and has to transform what can be perceived and thought, as far as it is relevant, to verbal means. Obviously a grammar of written language use represents a separate issue and would require its own comprehensive investigation. • Meaning as starting point

To develop a grammar of performance we must start from meaning. Meaning determines the selection of means which are multiple and, in principle, not limited. It is mainly due to cognitive means, conclusions and inferences, that the individual selection of means is not restricted. We do not produce utterances in order to speak but in order to act. An utterance grammar, conceived of as grammar of action, has to start from what we want to achieve in dialogue, that is, from the meaning of action or from purposes, and to correlate meaning with appropriate means. I take the extended formula I developed in the Mixed Game Model (MGM) as the meaning structure of the speech act (see Chapter 11, this volume): Interest [purpose (proposition)] Figure 13.2  Meaning Structure of the Speech Act.

The correlation of purposes and means can be schematized as follows: grammar of action

purposes, interest related to states of affairs

means (verbal, cognitive, perceptual)

Figure 13.3  Correlation of Meaning and Means.

Due to human beings’ double nature as dialogic individuals, any speech act is dialogically oriented and shaped by conditions of the natural and cultural environment. The speech act is therefore, by its very function, either initiative, making a dialogic claim, or reactive, fulfilling this very claim, positively, negatively or postponing the decision. The basic types of action, defined by the dialogic claims to truth and volition, can be derived from the basic mental states of belief and desire (Weigand, 1991). The action function dominates the proposition. Propositions relate to events by reference to objects and by predicating on them. Meaning aims at understanding. Insofar as the dialogic means also include cognitive means, such as conclusions and inferences, in the end only insiders of the game can arrive at an adequate understanding. Misunderstandings are not excluded. • The correlation between meaning and means

A grammar of competence describes the correlation between meaning and expression on the methodological basis of rules and conventions. A grammar of performance aims to find out the ‘laws of chaos’ which can correlate components of meaning and dialogic means on the basis of Principles of Probability (Chapter 11, this volume). Coherence becomes a matter of what we can expect according to some rationality of performance (Weigand, 2014). 219

Edda Weigand

We proceed by looking for rules as far as they go and try to identify particular cases by adaptation to specific conditions of action. • Integration of means

The communicative means are dependent on human abilities and on how the brain works. As has fortunately been verified by neuroscience (e.g., Damasio, 1994), these abilities, speaking, thinking, perceiving, are integrated and cannot be used separately or one after the other. Coherence results from appropriate integration and interaction of different types of means. The Principle of Coherence represents the crucial condition for understanding and can be considered as the basic principle of an utterance grammar (Weigand, 2010: 90ff.). The integration of speaking and thinking has in part been dealt with by means of terms such as ‘direct’, ‘indirect’ and ‘idiomatic’ speech acts (Weigand, 1989 [2003]), which relate to how the action function is expressed: explicitly or implicitly. We have, however, to question whether these terms, introduced in a model of dialogic competence, represent a workable basis for a model of performance. • Specialization or how to derive components of an utterance grammar

The grammar as a whole represents a complex hierarchy of components which are not separate parts but derived from the whole by specialization (Simon, 1962). The whole and its components therefore have to be ‘nearly decomposable’. The central theme that runs through this complex hierarchy is the purpose of the components, starting with the purpose of the whole and deriving from it the subpurposes of the components. There is not only one type of hierarchy. Simple hierarchies arise from the relation ‘being part of’, that is, from the division and addition of constituents, for instance, the generative hierarchy. Complex hierarchies arise from derivation by specialisation of the complex whole. Human competence-in-performance definitely belongs to the type of hierarchy of the complex. • Grammar as comparative grammar

The utterance grammar starts from a universal meaning structure. As we know, translation cannot proceed word for word but has first to find out the meaning of the original utterance and then look for an adequate expression in the target language: Universal meaning structure

Language 1 Language 2 ………….

Figure 13.4  Comparative Grammar.

Slight differences in meaning have to be accepted. They cannot be taken as a justification for constructing different language-specific types of grammar. The differences are due to different specialization of the universal meaning structure resulting from conditions in the environment. In principle, it is the co-evolution of genes, mind and culture which affects how human beings recognize and express their needs (Lumsden and Wilson, 2005). The environment includes the natural environment, for instance, differences in the climate, and the social and cultural environment. These influences can, for instance, result in different grammatical constructions, to mention only the grammatical categories of honorifica, but they do not produce different types of grammar from the outset. Cultural demands may have 220



Language in an Utterance Grammar

evolved differently in the evolution of humankind from rather simply structured communities to complex globalized societies. • Verification

An utterance grammar as a coherent complex whole needs verification if it claims to be more than a set of assumptions and speculations and more than a set of simple observations. Verification, in the end, comes from human nature and how it is shaped by the environment. These are the criteria which are fundamental to grammar and meaning. Fortunately, it seems that evolutionary and sociobiological criteria will eventually be able to verify the basic guidelines of human beings’ action and behaviour. The distinctive feature of the human species is not a single ability, be it the ability of thinking or speaking, but the double nature of dialogic individuals who are able to communicate in order to survive by mediating between their individual interests and social concerns. To sum up: These fundamentals of an utterance grammar imply that there are multiple variants of dialogic means at our disposal to express what we want to communicate. These variants are not arbitrary but are dependent on the structure of the complex whole and on the multiple conditions of the environment. An utterance grammar thus represents a system of all possible, that is, parallel, choices that might be taken in order to achieve understanding in ever-changing environments.

13.4  The Structure of Meaning One of the most difficult and in the end futile questions is the question of the nature of meaning. Why are we enquiring into the ontology of meaning? The ontology will forever be concealed from human beings because it goes beyond human reach. There is no meaning as such but meaning in the sense of meaningfulness to our minds and meaning in the sense of purpose. The meaning of an object is its purpose, the use we can make of it. In this sense, meaning in an utterance grammar is what we want to express by means of utterances in order to achieve the purpose of dialogue and what we need to express due to our nature and the social and natural environment. Meaning is dependent on our abilities. What we cannot perceive or think, we do not want to express. Meaning is thus rooted in human nature and dependent on the co-evolution of genes, mind, and culture. In structural, generative and semiotic linguistics we set up semantic features or atomic predicates as building blocks of meaning. The discussion about atomic predicates went on for a long time. There are still models in circulation which artificially construct such units by speculation, for instance, Wierzbicka’s ‘natural semantics’ (1972) which is the opposite of natural. Instead of constructing such artificial systems we should start with our human abilities that are the precondition of meaning. Our abilities not only include the traditional five senses but are much more differentiated (Weigand, 1998): universe of meaning | human abilities

consciousness perceiving knowing thinking

moving acting speaking feeling learning

social behaviour …

Figure 13.5  The Universe of Meaning. 221

Edda Weigand

These abilities do not work in isolation but are, to a great extent, integrated and interact with each other. The meaning components we set up in the structure of meaning are dependent on them and as multiple as human nature and the conditions of our world demand. Human beings are goal-oriented beings. As social individuals they have to negotiate their mutual purposes and interests with their fellow beings in order to come to an understanding. The basic structure of meaning therefore represents a structure of action based on initiative as well as reactive actions. Action means having a purpose and applying appropriate means to achieve it. The sequence of action and reaction constitutes dialogue and characterises each utterance as dialogically directed. Being initiative versus reactive is a feature of meaning based on the same dialogic claim which is either made or fulfilled. In the MGM this basic meaning structure is theoretically dealt with by the Action Principle and the Dialogic Principle proper. Fortunately, other approaches, for instance, Enfield (2013: 28), also acknowledge that the sequence of action and subsequent reaction represents “a unit, not a conjunct”. However, grasping this unit as a ‘relationship’ does not do justice to the nature of action. Specific action types can be derived by specialisation from the general purpose of dialogic interaction in a dialogic speech act taxonomy (see Chapter 11, this volume). In the end it is two fundamental dialogic claims, the claim to truth and the claim to volition, from which all the particular action types can be derived. These fundamental claims correspond to the basic mental states of belief and desire (Weigand, 1991). The meaning structure of action is complemented by many other components of meaning, for instance, evaluation and ethics, including respect. Most of them clearly manifest the individual stance of the speaker. In the MGM they are dealt with by Regulative and Executive Principles. Regulative Principles, for instance, describe how human beings as persuasive beings mediate between self-interest and social concern. In order to be successful, that is, to achieve their goals, they use different strategies ranging from cooperation to confrontation and power. Strategies relate to extended sequences of action and can be described as Executive Principles. Evaluation, either individual or cultural, is an important component and greatly affects certain speech acts, for instance, compliments (Weigand, 2016). We necessarily have to take a position in balancing our double nature as dialogic individuals, which is an ethical question. Ethics as well as rhetoric are components of an action theory. All these components of meaning are reflected in the utterance. Action, that is, interests and purposes, relate to events, indicated as proposition p in the formula of the speech act (see Figure 13.2 above). The proposition consists of reference and predication. We refer to objects, animate and inanimate, real and fictive, and we predicate on them. Referring implies grasping the object as an individual, as a group, as definite or indefinite. These differences make up the construction of noun phrases. Noun phrases, which include reference, correspond to arguments. Even if we refer by means of pronouns, such as he or she, some predicates about the object are included. Objects may be present in the actual speech situation, or belong to the past or future. They can also be referred to by perceptual means, mainly gestures, or by cognitive means such as memory of the past or thinking about the future. The action functions of interests and purposes can be considered to be higher predicates on the proposition which consists of arguments and predicates. The big questions of an utterance grammar will be: how does reference come about? How is predicating structured? I cannot delve into this question but only want to emphasise one point: in trying to explain natural language use we must not forget that it is human beings who give meaning to the means they use, not an artificial system of ‘semantic primitives’ and their combinations constructed by scientists. 222



Language in an Utterance Grammar

13.5  Dialogic Means of Expression Dialogic means must not be equated with ‘data’. Dialogic means can be ‘data’ which have meaning. Meaning is in the end determined by dialogic action. Meaning is what the speaker wants to express in dialogic interaction. Expression need not be explicit and need not be verbal expression. Meaning as well as means are dependent on human abilities, not only of speaking but of thinking and perceiving as well. In the same way as we tried to find components or types of meaning, let us now try to find types of dialogic means. Speaking proceeds by sounds and their sequencing which often is based on some hierarchical syntactic order. Suprasegmentals are superimposed. Speaking is integrated with perceiving and thinking. We do not speak and then think or perceive. It is the integration of human abilities which results in the premise: We mean more than we say. An utterance grammar has to describe conventional correlations, such as rational conclusions, but also has to go beyond such patterns by allowing for individual non-conventional inferences. If we want to delineate a grammar of the complex, we are confronted with an open system of parallel possibilities of choice. Parallel choices are not totally equivalent in meaning. A grammar of performance should be capable of elaborating differences between parallel variants as far as possible. Verbal means embrace a large variety of different types. Among the segmental means there are words, inflectional and derivational affixes, and interjections. In addition, there are suprasegmental means such as intonation or accent. Traditional grammar is divided into syntax and the lexicon. In a grammar of performance these fields have to be re-defined. Syntax has become flexible, can be broken, has meaning in integration with other devices, especially with suprasegmentals. The lexicon in use does not consist of single words but of phrases or collocations. This insight has important consequences for traditional lexical relations such as synonomy, antonomy and polysemy, as can be clearly seen in comparative analyses, for instance (Weigand, 1998: 35; and Weigand, 2009: 106f.): (1)

thick forest thick crowd heavy drops heavy traffic swollen ankle swollen cheek ……….

dichter Wald dichte Menge dicke Tropfen dichter Verkehr dicker Knöchel dicke Backe …………..

forêt épaisse foule serrée gouttes lourdes circulation dense genou enflé bouche enflée ………..

bosco fitto grande/molta folla goccie grosse traffic pesante caviglia gonfia guancia gonfia …………

Words are used in phrases which cannot be built by rules but are arbitrary groupings and completely different in different languages. The whole of language use, especially the interplay of syntax and words, is strongly dependent on the system of phrases. In order to find out these multiple phrases we need corpus linguistic methods. Not even the native speaker knows every possible phrase. Our ability of perceiving also results in different types of perceptual means, for instance, body movements, gestures, and facial expressions. Fortunately this type of expression has recently been focused on by approaches to multimodality (see Mondada, 2016, and Chapter 2, this volume). Perceiving not only relates to intentional perceptual means but also represents a fundamental component without explicit expression in face-to-face dialogue. What can be seen in the environment of the actual situation 223

Edda Weigand

is presupposed and heavily affects how the utterance is verbally shaped (examples see below Section 13.7). These different types of dialogic means, verbal, cognitive and perceptual, and their integration represent subsystems in the complex whole of human competence-in-performance. They have to be investigated in their multiple interactions and diversity in different languages and cultures. Integration and interaction can finally be verified by sociobiology.

13.6  Correlation of Meaning and Expression: The Core of Grammar Describing the correlation of subsystems in the complex whole is an especially difficult endeavour. I have to restrict myself to mentioning a few guidelines and to illustrate basic correlations by a few typical examples. • An utterance counts as utterance if it is the carrier of an action and can be understood as such. The speakers can decide how to express the meaning components of action. Guided by our self-interests we do not comply with Grice’s principle of expressing honestly and clearly what we want to say (Grice 1975). The devices can range between direct and indirect expression or deliberately vague and unclear expression. • Grammar is based on expectation, that is, we look for conventions but are prepared to go beyond these. What is described in other approaches by the term ‘emergence’, in the end, relates to human beings’ adaptation to changing demands which also affect the choice of devices. • There are subsystems of grammar in which the necessity to comply with rules is stronger than in others. In morphology and elementary syntax, and for instance, in the construction of noun phrases, violations are less tolerated than in other areas.

13.6.1  Utterance Syntax Dialogue as a sequence of utterances is a dynamic object which proceeds in time. The order of verbal means is therefore of fundamental importance even if the order on its own represents an abstract relation. In Chapter 12 Ruth Kempson focused on the shift from generative syntax to interactive authentic syntax. However, much more is involved. The basic issue of an utterance syntax is to elaborate how syntax is related to meaning, basically to the action function of the speech act. The position of the utterance obviously plays a role in expressing whether the utterance represents an initiative or reactive speech act; however, it is not the decisive role. The features of initiative versus reactive are not features of the position but features of meaning and can be grasped by the distinction of ‘making a dialogic claim’ versus ‘fulfilling this very claim’ (Weigand, 2010: 81f.). The specific type of the reactive speech act can be specialized in that it is dependent on the specific type of the corresponding initiative speech act. For instance, after a representative speech act we can expect a speech act of acceptance or rejection or negotiation of the initiative claim. The two-part sequence of action and reaction constitutes the minimal dialogic unit which is a unit of interaction and not only a sequence of turns. Syntax not only means the linear order or position of utterances but also the internal structure of verbal components within the utterance. The internal syntax of an utterance not only relates to the manifest position of the constituents but also to the concealed hierarchical structure, as we have known since Chomsky’s generative grammar (e.g., Chomsky, 1965). It is, however, difficult to describe how syntax contributes to expressing the specific type of initiative speech 224



Language in an Utterance Grammar

act. Syntax is only one component among other components. Grammatical items such as mood or modal verbs play a role but also lexical items such as particles or adverbs: (2) He is (certainly) at home. representative (assertion) He could/might be at home. He is possibly at home. representative (deliberative) Er ist wohl zu Hause. The grammatical category of sentence type also plays a role; however, only a very restricted one (Weigand, 1994). The category itself is not only a syntactic one but includes other devices such as mood and intonation. There are specific sentence types which can be correlated to specific types of action, mostly declaratives, directives and exploratives: (3) Representative Directive Explorative

declarative sentence, mood, modal verbs Europe is on the edge of a precipice. – Maybe. imperative clause, grammatical category Take it. Let’s go for a walk question sentence, word order, intonation Will you participate in the game?

However, such a direct correlation can easily be broken by other dialogic devices, mainly cognitive devices, for instance, in so-called indirect speech acts. The category of the sentence type is only one feature among many others to be taken into account (Weigand, 1989 [2003]: 198). In performance we do not use sentences but utterances.

13.6.2 Suprasegmentals Syntactic means interact with suprasegmental means. Mostly suprasegmentals are not clear enough to enable us to decide on their action function. However, there are a few cases where they dominate and rule out or at least question other verbal means. For example, an utterance in the syntactic form of a declarative sentence becomes a directive speech act by intonation and accent: (4) You ‘leave now. ↓ Intonation can also weaken the assertive force of an utterance: (5) Ok, then we’ve agreed upon going on holiday in August. ↑

13.6.3  The Lexicon The structure of the lexicon has always posed a big problem in lexicology (e.g., Pustejovsky, 1995). There are so many different types of words, not only lexical and grammatical words but also different parts of speech: nouns, verbs, adjectives, particles and so forth. In an abstract grammar of language as a system of signs, words could be taken as lexical units even if they had to be subdivided into yet smaller units, the morphemes. However, in a 225

Edda Weigand

grammar of language use words can no longer be considered to be lexical units. In language use it is not single words but the use of words in phrases which has meaning (see above example 1). The problem of the lexicon turns out to be the problem of phrases or so-called collocations. Lexical phrases determine the use of language to such an extent that it is methodologically advisable to structure the lexicon as a lexicon of phrases and keep the word only as a formal entry of alphabetical order (for details see Weigand, 2009: 95ff.). If we want to avoid the artificial view of quasi-unrestricted polysemy and disambiguation of single words, we have to accept that language as a whole represents a network of phrases. Phrases are internally syntactically structured, but their structure is to a great extent fixed. The speaker has to know the pre-determined structure of phrases and has to know when and to what extent he/she can change it. A special class of words are so-called utterance words. They stand for whole utterances and are therefore carriers of action, for instance: (6) Will you participate in the game? – Of course. Nimmst du an dem Spiel teil? – Klar. Utterance words can be particles or adverbs. Traditionally their use is explained as elliptic use which, however, does not say anything about their function. They mostly represent reactive speech acts which are determined by the initiative utterance, that is, by the Dialogic Principle proper, as in our example. Of course expresses a positive answer and additionally comments ‘that it goes without saying’. The enormous complexity of a phrasal lexicon becomes manifest when we try to describe the meaning of words-in-phrases by means of a corpus-based dictionary such as the Collins Cobuild English Dictionary. Let us take the example of fall and indicate a few ways-of-use (Weigand, 2009: 226f.): (7)

Big drops of rain fell. His wavy, reddish hair falls to his shoulders. This year Easter Day falls on March 30th. My work really falls into three parts. The party was falling into the hands of extremists. When he complained that it was unfair, they fell about laughing. The house was cheap because it was falling down. …………

It becomes clear that describing the meanings of words by a list of features, as was the case not only in structural and generative linguistics but in cognitive linguistics as well (Weigand, 2009: 95ff.), represents an artificial and futile endeavour. The meanings of words change continuously dependent on the phrases they are used in. The multiplicity of possible uses is so vast that it goes beyond the native speaker’s competence. As native speakers we are often surprised when we are confronted with yet another meaning of a word in another phrase. Language acquisition from the very outset in early infancy means acquisition of phrases and even whole utterances as they are used in specific action games. We need large corpora and techniques of corpus linguistics to detect the individual phrases which the use of language is based on.

226



Language in an Utterance Grammar

The multiplicity of ways-of-use unmasks a rule-governed compositional description as a myth (Baker and Hacker, 1984; Harris, 1981). The single item fall is characterized by an extreme degree of meaning indeterminacy that goes so far that it is difficult to outline. We would have to accept a quasi-infinite polysemy of the single word which is then disambiguated by context. As speakers who are competent-in-performance we know how to handle this issue not by knowing multiple different readings of the single word but by knowing how to use ‘fall’ in phrases or collocations in different situations. Language comparison confirms the view of language as a system of phrases, not only for predicating but also for referring and for whole utterances. Beside well-known phrases indicated in the dictionary there are subtle differences you only become aware of when using a foreign language. A clear comparative example of the phrasal character of the lexicon has been given above with example 1. The same is often true of whole utterances. You have to know the language-specific routines and the way they are used in specific situations. Slight differences in different languages do not change meaning but are conventional ways-of-use a competent speaker has to know, for instance: (8)

Du kennst keinen Maßstab. You know no limits. Non hai il senso della misura. Tu n’a pas le sens de la mesure.



Nichts für ungut. No offence. Non se l’abbia male/non avertela male.



Wie schmeckt es dir? Wie schmeckt dir der Fisch?

Ti piace?   How do you like it? Ti piace il pesce?   How do you like the fish?



Mach leiser!

Abbassa il volume!  Turn it down.

(polite apologies)

These examples demonstrate that translation cannot proceed literally, word for word, but has to be oriented towards the system of phrases in different languages. Such language-­ specific particularities of phrases are difficult to find out even in a corpus. Wandruszka (1969) offers a considerable number of examples. Interesting comparative particularities also show up in referential phrases, for instance: (9) Ein erstes Argument dagegen one main argument against einen Lauf machen to go for a run fare la corsa wie überall as everywhere come da altre parti etc. 227

Edda Weigand

The use of pronouns and articles goes far beyond the rules established by the category of determiners in generative grammar (Weigand, 1996). Rules such as that third-person pronouns have to be preceded by proper nouns are turned upside down in authentic language use: (10) Pass auf sie auf ! – Doris ist schon vorsichtig.   └    ┘ Take care of her. – Doris is careful. Stai attenta a lei. – Doris è attenta. Example 10 is also interesting regarding the Dialogic Principle proper which connects the initiative and the reactive speech act. It is in this case based on inferences such as: ‘I’ll pay attention but she/Doris really is cautious’. This inference mainly rests on the particle schon which clearly represents a dialogic particle. The particle schon has dialogic meaning and cuts short the inferences included in the Dialogic Principle proper. Interestingly, what in German is expressed by the particle schon, is expressed in English and in Italian by putting the stress on the predicate is and è. Particles are an interesting class of expression. The terms ‘discourse particles’ or ‘pragmatic particles’ remain vague and empty regarding their meaning. Recently an interesting study was published which unmasked ‘pragmatic particles’ as ‘dialogue particles’ (EngbergPedersen and Boeg Thomson, 2016). In their study on particles in Danish they described the meaning of these particles as the ‘dialogic viewpoint of the speaker’. It is, however, not only particles but any element in the action game which contributes to the dialogic nature of language use. The ‘dialogic viewpoint’ expressed by particles is also manifest in the following example which demonstrates that the general concept of ‘dialogic viewpoint’ needs to be approached in a more differentiated way: (11) Frank macht jetzt Sport. – Immerhin etwas. − Almeno qualcosa. Frank is now doing sport. – At least that’s something. Hai mangiato oggi? – Magari. Adverbs are often also assigned to the broad class of particles. Their dialogic meaning is especially manifest if they have the status of adverbials or sentence adverbs and predicate on the whole proposition. They can thus express the dialogic action function of the speech act: (12)

Unfortunately I cannot come because I have to work. Leider muss ich arbeiten. I regret, I cannot come. Dispiaciuto, non posso venire.

13.6.4  Grammatical Categories The universe of meaning comprehends many other types of meaning and ways of expression. I want only to point to grammatical categories, a vast area which is very differently dealt with in different languages and cultures (Grein and Weigand, 2007). Grammatical categories such as mood can affect the action function of the speech act, and for instance, distinguish between assertions and suppositions (Weigand, 2010: 172), or can add important

228



Language in an Utterance Grammar

meaning distinctions such as time. Grammatical categories can also express relational aspects of meaning, for instance, ‘respect’ expressed by the category of honorifica in east Asian languages (e.g., Grein, 2007). Honorifica represent ethical categories which regulate the double nature of human beings as dialogic individuals. In the MGM ethical attitudes are dealt with by Regulative Principles (see Chapter 11, this volume).

13.7  Speaking and Perceiving As mentioned above, integration is the crucial feature of an utterance grammar as a complex whole. Integration is not arbitrary but has to comply with coherence in order to achieve understanding. Faced with rather innumerable variables of ever-changing action conditions there are similarly unlimited possibilities of coherent utterances. We mean more than we say. Everyday face-to-face dialogue relies on the integration of perceptual and verbal means. A clear authentic example was analysed in Chapter 11 (this volume): (13) One for her. – Eve, too. In order to understand the sequence of these two utterances, you need to be an insider in the game or be able to observe precisely what the interlocutors are doing (Weigand, 2010: 87). Integration of speaking and perceiving not only relates to elements of the physical environment. Perceptual means are also intentional visible means, for instance, gestures, movements or face expressions, which go far beyond nodding one’s head as a perceptual means of indicating a positive reaction. Reference and especially deixis is a clear field which demonstrates the inclusion of perception. There are verbal means such as pronouns which relate to components of the speech situation, the speaker, the interlocutor, and the place and the time of the conversation. To some degree this correlation follows rules but on the whole it is dependent on performance variables and individual use of the means (Weigand, 1996). Deixis not only refers to speaker and hearer but also to the objects and events of the situation. An interesting example is the following (see the analysis of the whole action game in Weigand, 2015: 208): (14) R Hm, hm [on seeing the problematic result] E It can’t stay like that. The question is why the door was put in there at all. ………. Beside the deictic pronoun that which refers to what can be seen in the actual situation, there is another remarkable point: the dialogue starts with an interesting use of ‘hm, hm’, which in this case expresses the action of critically evaluating the problematic result. The fact that speaking and perceiving are integrated has already been emphasised in psycholinguistic studies for some time (e.g., McNeill, 2005, 2000; Kendon, 2004; Nash, 2007). Nonetheless, it remains a demanding field to be studied more precisely as a component of a comprehensive utterance grammar.

13.8  Speaking and Thinking Integration also includes the ability of thinking. Thinking is itself a complex field. We can of course think without speaking, for instance, when trying to find out in our minds how 229

Edda Weigand

the pieces of a puzzle are to be put together in order to form a meaningful picture. Related to language use, thinking or cognition mainly refer to the components of having appropriate knowledge and of making the right inferences. In neo-Gricean approaches inferences are added to an orthodox basis of semantics in order to cope with the fact that we mean more than we say. However, the addition of models which have nothing in common cannot settle the issue. Language use is pervaded by inferences which are only in part conventional and again demonstrate that it is the individual human being and their competence-in-performance that are where we have to start. The fact that a great part of meaning is left to individual thought represents a major source for misunderstandings. Language use aims at coming to an understanding which will, in principle, remain gradual understanding. The risk of misunderstandings cannot be excluded (Weigand, 1999). The interaction of speaking and thinking is at the heart of ‘indirect speech acts’ (see Weigand, 2010: 89f.). To some degree it can be conventionally described, based on events which normally are to be interpreted in a certain way, for instance, (15) Doris is waiting at the airport. – I’ll fetch her. (16) The doorbell is ringing. – I am in the bathroom. However, meaning and understanding often go beyond techniques of so-called indirect speech acts and cover a lot of completely individual interplays between speaking and thinking, as the following authentic example demonstrates: (17)

H E H F E

Don’t let yourself get infected! Are you ill? Didn’t you see the water? Everyone’s got a hobby I’d never do that when we pay so much just for the cleaning. Ah, now I understand. You’re right. No, I won’t let myself get infected.

Such sequences of dialogue can only be understood by insiders of the game as they presuppose common knowledge of past events and of the actual action conditions of the game. ‘Get infected’ in this case means ‘infected by a mania of cleaning’ which could be observed a few days earlier when a neighbour cleaned up the water in front of the door (for a more detailed description see Weigand, 2009: 297). There are many other interesting cases based on the integration of speaking and thinking, among them the persuasive use of magical words as, for instance, by George Bush in his well-known phrase (Weigand, 2009: 367): (18) If this is not evil, evil has no name. The so-called ‘power’ of words mostly derives from such inferences which are culturespecific and specific to certain action games. Bush uses the word ‘evil’ in order to achieve acceptance for a pre-emptive war. Rapaille’s study (2007) on the ‘culture code’ is interesting in this respect insofar as it demonstrates how unconscious cultural preferences are used to persuade the addressees in the field of marketing. 230



Language in an Utterance Grammar

13.9 Conclusion Turning back to the basic question of the nature of language, this chapter could shed only some light on the great challenge of an utterance grammar and provide guidelines for future work. An utterance grammar as a grammar of performance cannot be conceived of as a stable rule-governed system but represents a dynamic balance of multiple variables mediated by human beings in specific dialogic action games. Insofar as an utterance grammar includes individuality, it cannot aim at describing any possible occurrence of an utterance. The possibilities are, in principle, innumerable. It seems, therefore, reasonable to elaborate standard cases and indicate ways of going beyond these in particular cases. In the history of linguistics the concept of language has undergone a series of changes and in the end reveals itself to be an integrated component of human action and behaviour. What is needed is not only stocktaking in linguistics but in science in general. We have to ask ourselves what we want to do: construct artificial theories which are of interest only in the ivory tower, or create theories of performance which can help people understand and improve their use of language. If we decide for performance then the core of an utterance grammar will relate to the Principle of Coherence, which is a prerequisite for understanding. In an utterance grammar coherence is based on the integration of components of which only a few are empirically manifest. The components do not exist on their own as separate parts but are derived from the complex whole by specialisation. From stocktaking in this way we can derive the following guidelines for future research: • Language is not a static semiotic system. The concept of signs belongs to a linguistics based on an artificial system, not to a linguistics based on performance. Language use means speaking in integration with other human abilities. It is not signs which tell ­human beings how they can proceed, it is human beings who are at the centre of the game and, instead of signs, they use dialogic means which can be adaptated to individual action conditions in order to achieve their goals. Instead of proliferating artificial terms and structures, we should observe what human beings do with language in performance. Do we scrutinize any empirical element? Do we add parts in order to arrive at an understanding? We adapt to the whole from the very outset and try to come to an understanding with our fellow beings. • Language use represents an ‘indissoluble’ whole of components (Marchand, 2010) which interact in complex hierarchies dependent on ever-changing action conditions. Complex hierarchies are different from simple hierarchies based on the division into parts. The architecture of complexity (Simon, 1962) is guided by hierarchies of integration and interaction. • An utterance grammar needs to comply with recent insights of sociobiology. It is not an isolated object of study but grammar used by human beings for their purposes and goals and these are determined by human nature and made distinct by cultural influences. • Again we arrive at the old question of why human beings use language. Orthodox responses remain hypotheses without verification. Why should we use language for thinking? Language is not a means for acquiring knowledge either. It is true that we need knowledge for dialogic interaction, that is, for negotiating our purposes and desires with our fellow beings. It is, however, our human nature as dialogic individuals that determines how we apply this knowledge in the use of language. Syntax and the lexicon have up until now been the pillars of traditional grammar, but it is human beings who are at the centre of an utterance grammar. 231

Edda Weigand

Verification of our assumptions represents an important component of any theoretical approach. Verification of an utterance grammar is provided by sociobiological results, that is, by neurological experiments and anthropological goal-oriented observations of various languages and cultures. Different disciplines have to collaborate in the joint attempt to shed more light on how human beings succeed in coming to an understanding about their mutual purposes. They do this by balancing the multiple variables of their nature and this is dependent on ever-changing action conditions of the environment.

References Baker GP and Hacker PMS (1984) Language, Sense and Nonsense: A Critical Investigation in Modern Theories of Language. Oxford: Blackwell. Chomsky N (1959) Review: B. F. Skinner: Verbal Behavior. New York 1957. Language 35: 26–58. Chomsky N (1965) Aspects of the Theory of Syntax. Cambridge: MIT Press. Chomsky N (1988) Language and Problems of Knowledge: The Managua Lectures. Cambridge: MIT Press. Clark HH (1996) Using Language. Cambridge: Cambridge University Press. Damasio A (1994) Descartes’ Error: Emotion, Reason, and the Human Brain. New York: Putnam. Enfield NJ (2013) Relationship Thinking: Agency, Enchrony, and Human Sociality. New York: Oxford University Press. Engberg-Pedersen E and Boeg Thomsen D (2016) The socio-cognitive foundation of Danish perspective-mixing dialogue particles. In Dancygier B, Lu W-l and Verhagen A (eds) Viewpoint and the Fabric of Meaning: Form and Use of Viewpoint Tools across Languages and Modalities. Berlin: De Gruyter Mouton, 125–142. Grein M (2007) Kommunikative Grammatik im Sprachvergleich. Die Sprechaktsequenz Direktiv und Ablehnung im Deutschen und Japanischen. Berlin: De Gruyter. Grein M and Weigand E (eds) (2007) Dialogue and Culture. Amsterdam: Benjamins. Grice HP (1975) Logic and conversation. In Cole P and Morgan JL (eds) Syntax and Semantics, vol. 3: Speech Acts. New York: Academic Press, 41–58. Günthner S, Imo W and Bücker J (eds) (2014) Grammar and Dialogism: Sequential, Syntactic, and Prosodic Patterns between Emergence and Sedimentation. Berlin: De Gruyter. Harris R (1981) The Language Myth. London: Duckworth. Hopper P (1987) Emergent grammar. Berkeley Linguistics Society 13: 139–157. Kasper S (2015) Instruction Grammar: From Perception via Grammar to Action. Berlin: De Gruyter Mouton. Kendon A (2004) Gesture: Visible Action as Utterance. Cambridge: Cambridge University Press. Kretschmar WA, Jr (2015) Language and Complex Systems. Cambridge: Cambridge University Press. Langacker RW (1987) Foundations of Cognitive Grammar, vol. 1: Theoretical Prerequisites. Stanford: Stanford University Press. Langacker RW (1991) Foundations of Cognitive Grammar, vol. 2: Descriptive Application. Stanford: Stanford University Press. Levinson SC (2000) Presumptive Meanings: The Theory of Generalized Conversational Implicature. Cambridge: MIT Press. Lumsden CJ and Wilson EO (2005) Genes, Mind, and Culture: The Coevolutionary Process. New Jersey: World Scientific. Marchand THJ (2010) Making knowledge: Explorations of the indissoluble relation between minds, bodies and environment. Journal of the Royal Anthropological Institute 16(1): 1–21. Martinet F (1975) Functional linguistics: La linguistique fonctionnelle. In Martinet F (ed.) Studies in Functional Syntax: Études de syntaxe fonctionnelle. München: Fink. McNeill D (2005) Gesture and Thought. Chicago: Chicago University Press. McNeill D (ed.) (2000) Language and Gesture. Chicago: Chicago University Press. 232



Language in an Utterance Grammar

Mondada L (ed.) (2016) Writing in interaction. Language and Dialogue 6(1), special issue. Nash CE (2007) Gestural regulators in French, Japanese and American English dialogues. In Grein M and Weigand E (eds) Dialogue and Culture. Amsterdam: Benjamins, 115–140. Pustejovsky James. (1995) The Generative Lexicon. Cambridge: MIT Press. Rapaille C (2007) The Culture Code: An Ingenious Way to Understand Why People around the World Buy and Live as They Do. New York: Broadway Books. Simon HA (1962) The architecture of complexity: Hierarchic systems. Proceedings of the American Philosophical Society 106: 467–482. Thompson G (1996) Introducing Functional Grammar. London: Arnold. Toulmin S (2001) Return to Reason. Cambridge: Harvard University Press. Wandruszka M (1969) Sprachen, vergleichbar und unvergleichlich. München: Piper. Weigand E (1984) Lassen sich Sprechakte grammatisch definieren? In Stickel G (ed.) Pragmatik in der Grammatik. Jahrbuch 1983 des Instituts für deutsche Sprache. Düsseldorf: Schwann, 65–91. Weigand E (1989 [2003]) Sprache als Dialog: Sprechakttaxonomie und kommunikative Grammatik, 2nd, rev. edn. Tübingen: Niemeyer. Weigand E (1991) The Dialogic Principle revisited: Speech acts and mental states. In Stati S, Weigand E and Hundsnurscher F (eds) Dialoganalyse III: Referate der 3. Arbeitstagung, Bologna 1990, vol. 1. Tübingen: Niemeyer, 75–104. Reprinted in Weigand E (2009), 21–44; reprinted in Capone A and Mey JL (eds) (2016) Interdisciplinary Studies in Pragmatics, Culture and Society. Cham: Springer, 209–232. Weigand E (1994) Satztypen, Satzarten, Satzmodi und ihre Relevanz in einer kommunikativen Grammatik. In König P-P and Wiegers H (eds) Satz – Text – Diskurs. Akten des 27. Linguistischen Kolloquiums, Münster 1992, vol. 2. Tübingen: Niemeyer, 125–136. Weigand E (1996) Review of Barbara A. Fox: Discourse and Anaphora. Studies in Language 20: 236–243. Weigand E (1998) Contrastive lexical semantics. In Weigand E (ed.) Contrastive Lexical Semantics. Amsterdam: Benjamins, 25–44. Weigand E (1999) Misunderstanding: The standard case. Journal of Pragmatics 31: 763–785. Weigand E (2009) Language as Dialogue: From Rules to Principles of Probability. Feller S (ed.). Amsterdam: Benjamins. Weigand E (2010) Dialogue: The Mixed Game. Amsterdam: Benjamins. Weigand E (2011) Paradigm changes in linguistics: From reductionism to holism. Language Sciences 33(4): 544–549. Weigand E (2014) Rationality of performance. Philosophia Scientiæ 18(3): 247–268. Weigand E (2015) Dialogue in the stream of life. Language and Dialogue 5(2): 197–223. Weigand E (2016) How to verify a theory of dialogue. Language and Dialogue 6(3): 349–369. Wierzbicka A (1972) Semantic Primitives. Frankfurt/M.: Athenäum.

233

14 The Issue of Theorising Object-of-Study and Methodology Arto Mustajoki

14.1 Introduction The twentieth century gave us prominent linguists with an enormous influence on our understanding of language and scientific research on it, such as Ferdinand de Saussure, Noam Chomsky, and William Labov. However, a more characteristic feature of linguistics, especially during the last decades of the century, was the appearance of manifold approaches to language. William Kretzschmar (2009: 6) speaks of a “marketplace, in which ideas about language are promoted and accepted, bought and sold”. Some of these directions have a common research ideology gathering scholars under a large umbrella, such as functional and cognitive linguistics. Some are scientific schools with a shared theoretical foundation such as generative grammar (and the minimalist programme), constructive grammar, and relevance theory. Some call on researchers to share a certain methodological tool, such as conversational analysis, discourse analysis, or corpus linguistics. Some pay attention to various neglected sides of language, such as variation studies or pragmatics. Linguistics has also faced a shift of its object. There have always been linguists with different interests and orientations, but if we look at mainstream research it is easy to see a certain transition in research objects. The development has gone from small units to larger ones, first from phonetics and phonology to morphology and then to syntax. During the last two or three decades, more and more attention has been paid to units larger than the sentence (or utterance), to texts and dialogues. Research on oral discourse and its prototypical format, dialogue, has been the main area of expansion of linguistics during the last thirty years or so. We know much more about the rules governing the conduct of dialogue now than a generation ago. Despite this intensified interest in dialogical speech, one can still claim that linguistics suffers from a ‘written language bias’, as Per Linell puts it (Linell, 1998, 2012; see also Marková, 1982). Various features of dialogical speech have been described with a high degree of precision, but this has had very little influence on our common understanding of what language is. Grammar and other general descriptions of languages are, as a rule, still based on written texts or on our notion of what a language should be. Despite (or because of?) intensified language studies, linguistics still lacks a shared understanding of the object of research (Kretzschmar, 2009: 6; Weigand, 2011: 544). There 234



Object-of-Study and Methodology

are numerous competing camps with their own approaches and theories, and their adherents seem hardly to understand each other. This can be seen as richness and as a sign of the vitality of the field, but on the other hand the methodological and theoretical pluralism has had negative consequences for linguistics, which has proceeded slowly as a research field. It also makes it more difficult for linguists to find a common voice when this is needed for advocating language studies as a most relevant area of research. The poor visibility of oral discourse in our understanding of language and the insufficiency of research tools in this field is paradoxical, because according to the famous Russian language philosopher Mikhail Bakhtin, everyday conversation is a primary genre while novels, scientific texts, and other ‘organized cultural communication’ are secondary genres (Bakhtin, 1986). For the epithet primary one can see two explanations. In human history the primary genres have developed first; as a matter of fact, they were for millennia the only format of language that human communities had. From the contemporary point of view, one may argue that dialogue still represents the natural form of a language, while written genres are artificial constructions. Nevertheless, grammar books, which are commonly conceived as ‘official’ descriptions of languages, are generally based on the genres named by Bakhtin as secondary. Another relevant point is that every child learns his or her first language at home from parents. Then at school the child has a subject called ‘the mother tongue’ which is, in practical terms, a foreign (non-native) language for him or her (Zemskaja, 1987: 4). It takes years of intensive work to learn this language form, which differs substantially from the real mother tongue the child has learnt without deliberate instruction. The aim of this chapter is to elaborate the mystery of the essence of a natural language: what is language, and what kinds of evidence can we apply in describing language(s)? The first question is philosophical or ideological, the second methodological. We will start by considering the dichotomous nature of language.

14.2  Dichotomy of Language Even for a layman it is easy to understand that language has two states, two forms of existence. On one side, it is something that is ‘located’ in our minds as a potential, an inventory and an ability. On the other side, it can be heard and seen as a product of human activity. Using the vocabulary of contemporary technology, one can speak of offline language and online language (Kibrik and Koshelev, 2015). These two states of language are closely interlinked. A child creates offline language on the basis of his or her experience of online language. Then, in learning the standard form of the mother tongue, he or she acts in the opposite way by trying to apply a set of memorised rules to speech production. The offline language is useless without its online implementation, and the online language makes sense only in interaction between people who have mastered a similar enough offline language. Despite the close relationship between offline and online forms of language, they differ from each other in many essential ways. Offline language can be regarded as mental, abstract, and more or less stable; by its nature it is а code. Online language can be regarded as physical, concrete, and different in each communicative situation; it is а message. There have been several attempts to conceptualise the dichotomous character of language. Three of these have had the strongest influence on linguistic ideology. Wilhelm von Humboldt saw in language two sides, ergon (product, Werk) and energeia (activity, Tätigkeit) (Von Humboldt, 1836: LVII). Ferdinand de Saussure launched the concepts of 235

Arto Mustajoki

langue and parole (De Saussure, 1995). Noam Chomsky made the distinction between (linguistic) competence and performance and articulated very clearly that linguistics is about studying competence (Chomsky, 1957). The discussion around these concepts still continues, but it is not necessary to go into its details here. Instead, it is important to point out that the idea of seeing dialogue as the crux of language started with von Humboldt but then disappeared for several decades (compare Habermas, 1991; McLuskie, 2003). Now we are witnessing a rebirth of this kind of thinking.

14.3  The Question of the Norm ‘Norm’ is not a common term and concept in theoretical linguistics. Many researchers associate the word with the standardisation (codification) of a language and place it beyond the scope of their interest, among the concerns of applied linguistics. However, as will be shown below, the whole existence of language is based on the idea of conventions and norms within a certain community. A speaker will not be understood if the phrases he or she produces are not in concord with the linguistic tradition of that community. Some deviations from the norm are possible, but they cannot totally break down the frames given by the norm. What we usually mean by speaking of the norm of a language is the official norm which is presented to native speakers on a ‘top-down’ basis and concerns mainly the written form of language. It is connected to the notions of ‘standard language’, ‘literary language’, and ‘codification of a language’. The idea for linguistic normalisation is closely linked to the idea of a nation state, the establishment of which led to the need to rule and educate citizens by distributing information to them (Taylor, 1990; Gal and Irvine, 1995). Most languages of the world have no literary (standard) form; thus, they do not have an official norm. Nevertheless, they have a collective norm which is built ‘democratically’ by unconscious agreement among the people of the linguistic community. The notion of a collective norm is relevant even in those countries where people receive intensive instruction in the standard language at school, because native speakers’ impressions of the correctness of linguistic units are not equal to the official norm (Mustajoki, 1988, 2013a). Linguists may be involved in the codification of a standard language and they may study the history of this process, but it is the collective norm, people’s common understanding of language, which is a reasonable source for evidence in studies where the correctness of linguistic units is a focal criterion. (Collective) linguistic norm is a general context-free notion. Its main components are grammatical, lexical, and phonetic norms. The ability to determine that ‘cows’ and ‘sheep’ are correct plural forms, but ‘sheeps’ is not, is an important constituent of the grammatical competence of a native speaker of English. There is nothing but the collective norm to determine that we should say ‘How do people comment on this?’ instead of saying ‘How people comment about it?’ Every language has created its grammatical regularities and pronunciational traditions through a centuries-long, complicated process. However, as noted by Dell Hymes (1972), and then repeated and expanded by other researchers (reviewed by Bagarić and Djigunović, 2007), other competences are also needed in human interaction. Constituents of communicative competence have been labelled by various names: sociolinguistic, discourse, strategic, organisational, pragmatic, and other competences. All of these abilities of speakers (and recipients) deal with socio-pragmatic norms of the linguistic community concerned. We usually notice the importance of these factors when something in communication goes wrong. A child may break a sociolinguistic norm by saying ‘Why does that woman have such a big tummy?’ In order to cover such 236



Object-of-Study and Methodology

Table 14.1  Basic elements of language as a research object

langue/ PATTERN COMMAND potential of non-statistical competence language approach language in context-free realisation of COMMAND OF offline language patterns and BASIC LINGUISTIC mode potential for NORM (grammatical, the use lexical, phonetic) COMMAND OF language in statistical SITUATIONAL context approach NORM realisation of parole/ USUS language language performance in use (online LINGUISTIC concretization mode) MATERIAL

features of well-formed conversation, we use the concept of situational norm. Here again the concept of a norm is not typical of linguistic discourse, but by using this term we emphasise the societal aspect of the phenomenon: it is the speech community that determines which kind of language usage is acceptable, appropriate, and fitting in various communication situations. In order to launch a further element to our understanding of the essence of language, let us apply the term and concept usus put forward by Louis Hjelmslev (1942). It is based on what is common and frequent in speech. Talking about usus, we do not care about acceptability or correctness; it is a statistical generalisation of linguistic material produced by speakers and thus reflects the communication habits of people. So, usus is the statistical generalisation of online language. Usus can be retrieved only by analysing linguistic material produced by speakers. For understanding the way that language works, one more constituent is needed. We call it pattern. It helps us to create new linguistic units on the basis of the intuitive knowledge of language that we have acquired. When we heard the English word ‘vlog’ for the first time, we could immediately build the plural form ‘vlogs’ due to the analogy pattern we have in our minds. In a similar way, we use our pragmatic knowledge when we find ourselves in an unfamiliar communication situation. An inexperienced car buyer tries to cope with the situation by comparing it with the patterns (scripts, scenarios) he or she already knows about other purchase actions. Although pragmatic patterns differ substantially from those in grammar, the basic mechanism is the same: speakers creatively expand their linguistic capacities in order to be effective and successful in communication. These patterns are realised in communication situations, but they exist as a hidden potential and reservoir all the time. Table 14.1 highlights the basic features of the concepts presented above.

14.4  Linguistic Methodology: General Questions A focal and distinctive feature of any kind of scientific research is the use of methods recognised and accepted by the research community. The methods employed count as evidence for the findings and arguments that researchers want to distribute to their colleagues. Natural sciences strive for exact methods, which are objective in the sense that every researcher who uses them gets the same results. An example of this is the double blind method used in drug 237

Arto Mustajoki

research. This means that the nurse giving a drug and the patient taking it do not know whether it is a real drug or an empty one. In this way, it is possible to exclude subjective elements and concentrate on the real effect of chemical constituents of the drug. This kind of hard evidence can be seen as an ideal of proper scientific research. However, even in medical research a strict demand for such methodology leads to situations where relevant treatments are ignored. It is impossible to use the double blind method when psychiatry or physical exercise are used as а treatment. In the humanities and social sciences, the use of strict methods is even more problematic. In history studies, researchers use historical documents as evidence, but at the same time they have to perform source criticism that brings subjective elements into the picture. This means that researchers may arrive at different conclusions based on the same material. Among sociologists there are two camps as to the methodology of their field of science. Some swear by quantitative methods as the only acceptable way to conduct research, while others argue that the essence of societal phenomena may only be captured if qualitative methods are employed. Thus, methodological pluralism is quite a normal situation in the Social Sciences and Humanities. In linguistics, the question of evidence-based methodology and an endeavour to imitate the natural sciences can lead to extreme statements, such as this one made by Noam Chomsky in an interview: “Corpus linguistics doesn’t mean anything. It’s like saying suppose physics and chemistry decide that instead of relying on experiments, what they’re going to do is [to] take videotapes of things happening in the world” (quoted in Andor, 2004: 97). Chomsky adds that sciences do not do this. As a matter of fact, Chomsky is mistaken in two things (Kopotev and Mustajoki, 2008: 9–10). First, physicists also go outside their laboratories and measure, for example, aerosol particles in the air. Second, in physics researchers making experiments reach the same results regardless of who is conducting the experiment. When ‘armchair’ linguists make experiments based on their linguistic intuitions, they may end up with different conclusions. Although one may argue that linguistics is somehow closer to the natural sciences than some other humanities fields, it still lacks a common agreement on the methodology dilemma. It is therefore worth stopping to consider why this question is so difficult to solve in linguistics. One reason for this is the circumstance described above: the research interests of linguists are divided between the apparatus which stands behind speech and communication (offline language) and the actual use of language (online language). If there is disagreement on the object of study, it is evident that it is hard to reach a consensus in methodological questions. This discrepancy becomes even less solvable due to the fundamental difference between these objects of study: one of them is by its nature mental (psychological), while the other is social. These two areas of research require very different approaches and methods.

14.5  Linguistic Methodology: Language-Oriented Research Now we turn to the concrete question on linguistic research methodology by differentiating two approaches. First we consider this issue in the framework of what we call languageoriented research; then we launch an alternative approach, communication-oriented research, and examine methodology from that point of view. Strictly speaking, the methodology of language-oriented research should be a rather simple issue. If you are interested in the correctness and acceptability of linguistic units, as many linguists still are, you should turn to the intuitions of native speakers. If you are interested in language use, you should observe, collect, and analyse linguistic material produced 238



Object-of-Study and Methodology

by people (full lines in Figure 14.1). However, in practice the situation is much more complicated because there are also indirect ways of obtaining evidence on these matters (dotted lines in Figure 14.1) (see Penke and Rosenbach, 2004: 487–488). The objects and aims of research are expressed in the form of two questions. One of these reflects the idea of langue, while the other is connected with the concept of usus, which is a statistical generalisation of parole. A linguist aiming to describe what is potential, correct, or acceptable in language relies on the intuitions of native speakers (line 1). Correctness and/or acceptability may concern linguistic units of any level: pronunciation, semantics, morphological forms, syntactic structures, components of a dialogue, and so forth. The person taking a stand on this can be the linguist him- or herself, or a larger group of informants who are asked to take part in this kind of test. When the informant is the researcher him- or herself, the process of making assessments about linguistic units may take place intuitively without conscious reflection. When turning to informants, the questions should be well formulated. If a linguist is interested in the government of English nouns, he or she may ask: ‘Which of the following options do you regard as correct?’. An alternative formulation is ‘Which of the following options do you regard as acceptable?’. These questions are context-free, which means a basic linguistic norm approach. Similar questions may be linked to a certain genre, situation, or action game, for example: In your opinion, which of the following options is correct/the most appropriate: (a) in a written document of the department; or (b) in an oral presentation by the head of the department? disagreement over the strategy of the department disagreement on the strategy of the department disagreement about the strategy of the department For a linguist interested in usus (language usage), the main source of information is naturally linguistic material (line 6 in Figure 14.1). Electronic corpora are nowadays widely used in linguistic studies with various objectives (see Chapter 3 in this volume). As a matter

Intuition of native speakers 1

2

Linguistic materials

Experiments 3

Question A: What is potential, correct, and acceptable in language?

4

5

6

Question B: What is typical and frequent in language usage?

Figure 14.1  S  ource of information in linguistic research (the two states of language, langue and parole). 239

Arto Mustajoki

of fact, corpus methods appeared in linguistics much earlier than it was possible to employ electronic corpora (McEnery and Wilson, 2001: 2–4; Dash, 2009). Despite the rapid growth of linguistic corpora, the linguistic data that researchers have at their disposal covers only a tiny fraction of the whole existing language production. Oral discourse is particularly poorly represented. I don’t think there can be any corpora, however large, that contain information about all of the areas of English lexicon and grammar that I want to explore. ... [but] every corpus I have had the chance to examine, however small, has taught me facts I couldn’t imagine finding out any other way. (Fillmore, 1992: 35) After considering the primary sources in linguistic research, we now turn to secondary ones, described in Figure 14.1 by dotted lines. Line 5 denotes the case when linguistic material is used in answering Question A. This method is sometimes used, although usus does not directly reflect what is correct in language. Traditionally, researchers try to eliminate this discrepancy by using carefully selected linguistic sources, mostly written texts by wellknown writers. Because the language of any writer also contains some incorrect linguistic units, researchers apply special ‘filtering’, eliminating linguistic units which are incorrect from the point of view of their intuitions. Line 2 demonstrates the opposite case where tests based on intuition are used in exploring how people speak. Here informants are not asked to reveal their opinions about the correctness of linguistic units, but about their potential uses, e.g. ‘Which preposition would you prefer when using the following construction: disagreement ____ the strategy of the department?’ Informants are capable of answering such questions, but it is quite another thing whether their answers reflect the real usage of these constructions in their authentic speech. Lines 3 and 4 refer to the use of experiments in linguistic research. Referring to intuition is sometimes also regarded as an experiment. It is true that linguists may test hypotheses in this way. However, in most cases such processes take place automatically without strict design or arrangement. We understand by experiment specially organised conditions where informants conduct previously planned tasks and/or are subjected to neurophysiological measurements.

14.6  Communication-Oriented Research Now we turn to a wholly different approach to language where attention is paid to the actions and processes that take place in concrete communication situations. The communicants bring into the situation their whole personalities with certain communicative capacities and goals of action. The environment where the communication takes place is an important factor as well. When having such a communication-oriented approach, a linguist is interested in quite different phenomena from those that the language-oriented approach focuses on. For understanding the constituents of a communication situation, we have to take into account the motivational stances of interlocutors. Evidently, when taking part in a dialogue, they have some communicative aims, intentions, and needs. In order to cope with these communicative needs, they apply their (basic) linguistic competence, system competence, and situational (norm) competence. A good term for the ability to do this is competence-inperformance, launched by Edda Weigand (2003, 2004, 2010). 240



Object-of-Study and Methodology

We examine the communication situation in Figure 14.2 from the point of view of the speaker. The recipient (listener) takes an active role in communication as well, but for reasons of simplicity we concentrate here on the role of the speaker. Any communication process consists of choices. The first choice of the speaker is to decide whether to take part in communication or just to leave the situation. After deciding to participate in interaction, the next step is to find a linguistic (or other) way to do so. Our thoughts are usually rather sketchy ideas and it is not always obvious how to transfer them into words. In most cases, the speaker may choose between various expressions denoting the same state of affairs. Sometimes the options available reflect a scale of directness or politeness, compare ‘Wash the car’, ‘Would you please wash the car’, ‘The car needs washing’, ‘The car is dirty’. In addition to deciding what to say, the speaker also needs to decide how to say it—with a nasty or friendly tone of voice, shouting or whispering, with determination or with caution. As a rule, the speaker makes the choices about what and how to say automatically, unconsciously. However, we are not dealing with an engine, but with a complicated and complex human process. One could think that this is too demanding an object of research to be approached through scientific methods. It is, however, worth comparing this with cancer research. Scientists will never uncover all the secrets of the mechanisms causing cancer. Nevertheless, it is reasonable to try, step by step, to understand some of the factors influencing the process. The same concerns communication as an extremely complex object of research. On a larger scale, the factors playing a greater or lesser role in the choice of utterances are: the speaker’s linguistic capacity, his or her mental world, communicative needs, other possible needs, and the circumstances where communication takes place. Let us look briefly at these factors. At a general level the linguistic capacity of communicants varies a great deal. There are substantial differences in, among others, the amount and content of vocabulary, the ability to use various syntactic structures, and the ability to implement different communication tactics.

Linguistic capacity

External circumstances

Mental world Choices of the speaker

Other

(non- communicative)

needs

Communicative needs

Figure 14.2  Main constituents influencing the course of communication. 241

Arto Mustajoki

Much more variation is seen when we contemplate different communication situations. Some people excel at delivering talks, others are good at small talk and social interaction. We all have some subject areas (profession, interests, hobbies) where we feel comfortable and others where our ability to conduct normal communication falls almost to zero. Basic linguistic competence (see Table 14.1) forms the foundation of our linguistic capacity. In all communication situations and action games, we have to master, at least to some extent, the main constituents of speech: knowledge of a certain amount of lexical units adjusted to the context, ability to build larger linguistic units by organising lexical units in a line, ability to pronounce words and utterances, and an understanding of the rules of constructing texts and dialogues. From the communicative point of view, in most types of action games a full command of these competences is not needed. Extracts from people’s normal everyday conversation and research on English as a lingua franca show that substantial deviations from the correct (normative) language do not harm the smooth course of conversation in any way; they sometimes even make it more effective. Pattern knowledge and ability to use it in communication vitally expands our communicative opportunities. It adds to our capabilities a creative component, which enables us to produce and comprehend utterances that we have never heard or seen before. One may think that this is a capacity which is essential for writers or public speakers, but it is also a relevant part of everyday conversation. Situational (norm) competence is a crucial element in successful communication. It includes appropriate genre and register design in order to be effective and understandable in conversation. To write a blog is quite a different task from writing a scientific article. The norms of conducting dialogue in everyday conversation at home differ very much from those in business negotiations or public political debate. These differences concern general features of building utterances (syntax), rules of turn-taking in dialogue, feedback mechanisms, and the need and possibility of using repairs and asking for clarification. A salient constituent of situational competence is the speaker’s ability to conduct recipient design by adapting their speech to the recipient (audience) present in the communication situation. The importance of recipient design in the course of communication has been shown by Sacks and Schegloff (1979) and subsequently by many others. Linguistic capacity is one part of our mental world (Figure 14.2) (Mustajoki, 2012, 2013b). Our cultural backgrounds, like our linguistic capacity, are based on our personal histories—what we have seen, heard, and experienced until the moment of communication. This part of the mental world has been a popular topic in research on intercultural communication (e.g. Banks et al., 1991; Bürig and ten Thije, 2006), but it is an important factor in any kind of communication. This information is manifested in presuppositions, stereotypes, scripts, mental sets, scenes, schemas, and mental models. There are also some situational constituents of the mental world which affect speech production and comprehension: what the communicants have just experienced, what their mental and physiological states are (whether they are tired or fresh, calm or overheated) and what their attitudes are towards other communicants. Mental world is an essential background factor in communication because it affects the way the communicants outline the communication situation and the themes that are touched upon during the conversation. When we take part in a conversation, we have some communicative needs including conscious or unconscious intentions, hidden agendas, and overt or covert aims. We may just want to say something about the surrounding world or about our feelings, or talk about our future plans. We may aim to compliment, flatter, thank, hurt, insult, or humiliate the person(s) we are talking to. Or we simply want to make contact with another human being. 242



Object-of-Study and Methodology

Besides primary communicative needs, the communicants may also have other aims in taking part in communication. When people do something in company, they often have an unconscious need for self-presentation: they want to be regarded as smart, intelligent, and humorous. This may lead to language use which, from the communicative point of view, is impractical. The speaker may succumb to trendy or sophisticated vocabulary although he or she knows that the recipient may not be familiar with it (see Mustajoki, 2015: 554–555). One more factor that has a significant influence on people’s behaviour in communication is the need to avoid cognitive efforts. This is difficult to recognise in the normal course of a conversation and may therefore seem a meaningless and irrelevant consideration. However, several experiments have shown that this factor dictates human activities in many ways (Bargh and Chartrand, 1999; Shintel and Keysar, 2009). The capacity of the brain to handle information unconsciously is huge, but for conscious operations it is very limited. This is why we spare our cognitive efforts whenever possible. In communication this means a diminished level of concentration on speaking and listening. In speaking it leads to careless choice of utterances and unclear pronunciation; in listening, it is even possible to switch off totally. This is an evident risk for misunderstanding and failures in communication. External circumstances sometimes make interaction onerous. A vacuum cleaner at home, music in a disco, or traffic noise in the street may be serious hindrances for understanding speech. The speaker could overcome these problems by speaking more loudly, but the volume needed for proper interaction is often so high that the speaker underestimates the voice level needed. We have considered the main factors influencing the course of communication. Now it is time to look at the methods that are needed in studying them.

14.7  Approaching Communication with Different Tools In Section 14.5, we considered methods that can be applied in language-oriented research when a linguist tackles the questions: ‘What is potential, correct, and acceptable in language?’, on one hand, and ‘What is typical and frequent in language usage?’, on the other. If a researcher (you may call them a linguist or something else) tries to understand and describe communication by answering the questions: ‘What is the communicative capacity of the speaker?’ and ‘What happens in the interaction?’, a different methodological package is necessary. The main difference between these orientations is the general level of the study. A language-oriented researcher tries to identify features and constituents of a communication tool that is ‘owned’ by a certain community of speakers, while a communication-oriented researcher focuses on a situation of communication and aims to unravel the secrets of that complicated process. As discussed in Section 14.6, this, as a rule, requires crossing the boundaries of traditional linguistics. Let us first define the objects of research in a more concrete way. As we are interested in communication, it is evident that what happens in the interaction is a focal theme here (Question D in Figure 14.3). If one wants to find an approximate equivalent for this in the history of linguistics, the closest candidate is Humboldt’s energeia. As demonstrated in Figure 14.2, there are many more influencing factors than in Humboldt’s concept, but as a whole the idea is the same. A further object of research to be tackled is the communicative capacity which the speaker and other interlocutors bring into the situation, as described in Figure 14.2. The two issues researchers try to study, the communicative capacity of the speaker and the course of interaction, have an analogous (but not totally equal) relationship, as between 243

Arto Mustajoki

Experiments

1

2

Observation 3

Question C: What is the communicative capacity of the speaker?

4

Interpretation 5

6

Question D: What happens in the interaction?

Figure 14.3  Source of information in communication-oriented research.

Questions A and B in Figure 14.1. Question C represents the potential which is available for fulfilling communicative tasks and Question D the realisation of this potential. We start by examining Question C. It is important to understand that the focus here is not on a certain language per se, but on the communicative capacity of an individual speaker. Consequently, every communication encounter is different in this respect: speakers have different vocabularies, different sets of favourite syntactic structures, and different knowledge of culture and environment. To be precise, one may even argue that in every particular instance the same speaker has a capacity which differs slightly from previous situations because the speaker’s most recent experiences are at the forefront of her or his mind. There are only indirect ways to study people’s communicative capacities. One option is to turn to native speakers and ask them whether they know certain linguistic units or not. The results of this kind of study are based on speakers’ subjective impressions about their linguistic capacity. It is clear that informants’ opinions do not reflect reality on a full scale. Other options for studying linguistic capacity include tracing the reactions of the informants’ brains to verbal stimuli by means of magnetoencephalography (MEG), eye tracking, or other technical devices. Similar techniques could be utilised in studying other parts of our communicative capacity, for example, to find out what people know about the world and how they see it—both factors influencing interaction. Our knowledge reservoir consists of thousands of facts, and each of us has a personal package of them: the capital of Nigeria, the latest hit of Lady Gaga, the name of a particular species of tree, and essential features of nature, the human body, or society. It has been shown that there is also cultural variation in seeing the world. According to dozens of studies, the Western type of thinking is more analytic and the East Asian type more holistic (see Nisbett et al., 2001 and literature there). These differences influence, among others, what people see in pictures: the holistic approach pays greater attention to the background, while the analytic approach concentrates on active objects and their movements. This affects the way people understand each other when they meet. Quite another method of tackling these questions would involve observation and recordings of authentic communication situations. It is obviously a ‘mission impossible’ to record all the linguistic production of a single speaker. It is an enormous task even to record a speaker’s daily production, as has been done in the ‘One speaker’s day’ project (Sherstinova, 2010). 244



Object-of-Study and Methodology

Such recordings of real speech reveal to us important features of human interaction, but they still cover only a small part of a person’s communicative capacity. In applying such an approach, we are dealing with a potential which can be realised in communication but has not necessarily been realised yet. One can compare this with the situation in language-­ oriented research: corpora can never encompass the complete linguistic reservoir that people possess. Question D, to understand and describe all the factors influencing the course of interaction, is a huge task for the whole research community. Within the framework of the action game theory, researchers approach the question as a unitary phenomenon which consists of various elements but makes up a single object of research. A possible approach to Question D is to scrutinize causes of misunderstanding and failures of communication, that is, situations where the recipient understands the message in a different way from what the speaker meant (Ryan and Barnard, 2009: 45). For researchers tackling this problem, it is not enough to describe authentic conversation extracts of this kind; they must also try to explain why communication failure takes place. In studying this question, one may apply explanations of various levels. To illustrate such an interpretative approach, let us consider one of the basic characteristics of language, namely ambiguity. The first observation is that cases of syntactic ambiguity, as in the classic example ‘Visiting relatives can be boring’ (Tyler and Marslen-Wilson, 1977), are in most cases resolved by two factors: (1) the mental worlds of the communicants give the same interpretation for the phrase; (2) in the context, there are certain deconflicters which help to eliminate wrong alternative interpretations (Kibrik, 2011). The same is true with regard to referential ambiguity, as in ‘Tony Blair shook hands with George Bush in the White House. He wanted to discuss the situation in Iraq’ (Järvikivi et al., 2005). In studying such cases, we encounter the limits of experiments. In real communication situations, occurrences of this kind do exist but are usually harmless. From the point of view of successful interaction, other types of ambiguity are more important, such as those referring to something that is not mentioned in the context (‘Bring me the document tomorrow’, ‘It was really hot in there’). Indirect speech, producing phrases which can be interpreted in various ways, is also a permanent risk for successful communication: is the phrase ‘The car is dirty’ a statement or a request? In everyday interaction people often meet situations where the real intention of the speaker is open. What does the speaker want to say by telling us that ‘Linda went with John to the football match’ if Linda hates football but fancies John? Ambiguity as such is seldom an explanation of communication failure. Even when it seems to be the triggering factor, the real cause of misunderstanding is in fact improper recipient design, which leads us to move on to the second level of explanation, which is any more purely linguistic. By conducting proper recipient design, the speaker may avoid any kind of miscommunication (Mustajoki, 2012: 230). The next level of explaining communication failures is to try to find out why recipient design fails. There are several possible reasons for this. First, people are often not aware of how they actually speak (Epley, 2008; Keysar and Henly, 2002). Second, because of egocentric thinking, people do not always realize that recipient design is needed (this is known as the common ground fallacy) (Keysar, 2007; Mustajoki, 2012). Third, the speaker may be in a state of strong emotion, tiredness, or such, which hinders a normal course of conversation. Fourth, the speaker may be thinking of something else rather than being fully present in the situation. And finally, speakers sometimes try to conduct recipient design but are not skilful enough to do so. Another important background factor is avoidance of cognitive efforts; a mechanism of that was described above. 245

Arto Mustajoki

14.8 Methodology-Driven Research Compared to Phenomenon-Driven Research Finally, we try to take a broader look at the development of linguistics in recent years. Edda Weigand has claimed that linguistics moves from searching for ‘the simple’ towards challenging ‘the complex’ (Weigand, 2004: 3), or from ‘reductionism’ to ‘holism’ (Weigand, 2011). Approximately the same idea is introduced in Istvan Kecskes’s ‘socio-cognitive approach’ (Kecskes, 2010). As a matter of fact, what we see in linguistics is part of a larger development in the academic community. There are two big lines which lead to a rethinking of knowledge production in research. One of them is connected to the need for an increased societal impact to be made by science. There are two major driving forces behind this. There is huge pressure from decision-makers who finance research to see a closer link between the work done by academia and economic growth. In this context multifarious concepts are used (see details in Hessels and van Lente, 2008): among others, ‘Mode 2’ research, ‘Academic Capitalism’ and ‘Enterprise University’. Another trend in science policy is a new approach to priority areas of research. For a long time, they were determined on the basis of disciplines such as nanotechnology, cancer research, or material sciences. According to the new doctrine, priorities are determined as grand challenges or wicked problems. Usually the list of these consists of such phenomena as climate change, energy supply, and the ageing of the population. The list compiled by the Academy of Finland (2011) also includes the themes of ‘knowledge and know-how in the media society’ and ‘dialogue of cultures’. As a matter of fact, linguistics and communication studies should be one of the priority areas of research because contemporary societies and international trade and relations are totally based on the ability to understand each other. One could claim that the most serious problem in the world is that people do not understand each other. We, as linguists, should take part in solving this problem. This and other grand challenges cannot be met solely by the efforts of researchers in one discipline. Therefore, a multidisciplinary approach is often necessary. The factors mentioned above concern general developments within science policy. However, the need to rethink research objects and methodology can also be seen as a consequence of the stage that the academic world has reached by intensive research. Research is more and more fragmented into small subfields. Knowledge has become deeper and more detailed, and the risk of losing a larger picture of complex phenomena has increased. We will now consider how these trends influence the general ideology of research. In this connection, it useful to differentiate theory- or method-driven research on one side, and phenomenon-driven on the other. In a theory- or method-driven approach the starting point of the researcher is a certain theory or method which, in his or her opinion, is the best way to find out something important and new in the object of research. When working in such an environment, a researcher applies the chosen theory or method to more and more objects. Such a subfield has a strict vocabulary and commonly accepted rules for doing research and reporting its results. The theory or method itself may be developed little by little, but its basic elements remain untouched. The working mechanism of the research community applying this approach seems very science-like. The results that are reached are exact and easily recognised by other researchers subscribing to this theory or method. For researchers of other schools and directions, it may be difficult to follow the chain of argumentation. There are, of course, certain observations by researchers which have stimulated the development of the theory. However, after its creation the theory places strong limitations on the results that can be achieved. Important elements of the essence of the phenomenon 246



Object-of-Study and Methodology

may remain outside the interest of researchers. Usually, studies like these try to answer the question ‘what’; the ‘why’-question does not interest researchers, or is answered strictly in the framework of that particular theory. An alternative approach is phenomenon-driven research. In its extreme version, a researcher tries to find out and describe the essence of a given phenomenon, but has no clear understanding of the method to be used. Naturally, a certain methodological toolkit is needed, but it is determined by the phenomenon rather than being selected beforehand. Because various methods are needed, they often (but not invariably) lead to a multidisciplinary approach. The essence of a phenomenon can hardly be clarified without ‘why’questions. Because the methodological basis of research is diffuse, the results may seem somehow speculative or even unscientific. The phenomenon-driven approach works also in traditional language-oriented linguistics. We may take the case selection in Russian negative clauses as an illustration of this. The choice between the genitive and the accusative cases seems to follow some principles, but these make up a complex bundle of factors, so that the case usage seems to be rather unpredictable and unsystematic. It is possible to discover some regularities by using corpus methods (Mustajoki and Heino, 1991), but they usually reveal only some correlations between the case and various factors (of which there are about fifty). An alternative possibility is to use experiments. They can be organised in such a way that it is possible to determine the influence of one single factor (Mustajoki, 1985). There are further options for approaching the same problem, such as by working with informants and using the reaction time method mentioned above. As an example of methodology-driven research, let us consider conversational analysis. It examines real authentic dialogues. The space of research it strictly limited. The linguist applying this approach takes into account only the dialogue as such and avoids going beyond it. The main aim is to describe the course of dialogue by using the precise terminology developed in this linguistic school. The ‘why’-question may arise concerning the characteristics of the dialogue itself, but not concerning the background or history of the participants. There is definitely serious theoretical thinking behind conversational analysis, but as a whole, I would regard it as a methodology-driven approach to language. Chomsky’s generative grammar may be seen as an example of theory-driven research. The theory itself has developed, having as its latest modification minimalist theory. In its original form, generative grammar tries to describe the syntax of various languages from the point of view of an ideal native speaker. It aims to show dependency structures of sentences. The theory itself does not try to answer ‘why’-questions, but the research results can be applied to explaining for example, the time the human brain needs for understanding and producing sentences of varying complexity. As mentioned above, a generativist is a typical armchair linguist who denies the necessity of paying attention to language use in real situations of communication. Grice’s theory on communication stands very far from the ideas of generative grammar because it deals with language use, more specifically with dialogical speech. At the same time, there are some similarities between the two approaches in that they both view language as an ideal construction. Chomsky states this explicitly by referring to an ideal native speaker; Grice speaks of well-structured dialogues. Indeed, Grice’s maxims of conversation are hardly ever followed in real communicative situations (Sarangi and Slembrouck, 1992). In the case of Grice, intuition is used to describe correctly built dialogues rather than syntactically correct sentences. Nevertheless, the general approach is the same. Grice’s theory is based on the norms of building dialogues in a linguistic community, rather than on real usus. 247

Arto Mustajoki

14.9 Conclusion Dialogue is a very complex phenomenon and it is quite understandable that there are plenty of different approaches and methods in studying it. It is unrealistic and impractical to try to reach total consensus on this issue. Instead, it is important that researchers have an understanding of the diversity of the field. Knowledge of other possible approaches to the object fertilises research and opens up novel, sometimes unexpected perspectives. It helps us to obtain a fuller picture of the phenomenon that we are interested in. The Mixed Game Model developed by Edda Weigand over the last two decades is an attempt to demonstrate a more holistic approach to language than previous theories (Chapter 13, this volume). It abandons the idea of describing the very complex phenomenon of human communication by using methods that see language and its use as a strict entity with separate constituents. Instead, she proposes a probabilistic approach. There are various action forces which determine the sequence of speech acts in dialogic action games: representatives–acceptance, directives–consent, exploratives–responses, and declaratives–­ confirmation. Using these characteristics, it is possible to create a dialogic speech act taxonomy. This approach could be characterised as theory-based, phenomenon-driven research. Weigand puts it this way: “A holistic theory of a natural phenomenon needs to start from the complex object and to derive methodology from it without sacrificing the integrity of the object” (Weigand, 2011: 546).

References Academy of Finland. (2011) Grand challenges. Available at: http://www.aka.fi/en/about-us/media/pressreleases/2011/Academy-of-Finland-identifies-grand-challenges-and-priorities-for-research-policy Andor J (2004) The master and his performance: An interview with Noam Chomsky. Intercultural Pragmatics 1(1): 93–111. Bagarić V and Djigunović JM (2007) Defining communicative competence. Metotika 8(1): 94–103. Bakhtin MM (1986) Speech genres and other late essays. (Translation of ‘Estetika slovesnogo tvorčestva’). University of Texas Press Slavic series 8. Banks SP, Ge G and Baker J (1991) Intercultural encounters and miscommunication. In Coupland N, Giles H and Wiemann JM (eds) “Miscommunication” and problematic talk. Newbury Park: Sage, 103–120. Bargh JA and Chartrand TL (1999) The unbearable automaticity of being. American Psychologist 34(7): 462–479. Bürig K and ten Thije JD (eds) (2006) Beyond Misunderstanding: Linguistic Analyses of Intercultural Communication. Amsterdam: John Benjamins. Dash, NS (2009) Language Corpora: Past, Present and Future. New Delhi: Mittal Publications. Chomsky N (1957) Syntactic Structures. De Haag: Mouton. de Saussure F (1992/1995) Cours de linguistique générale. Édition critique préparée par Tullio De Mauro. Paris: Payot. Epley N (2008) Solving the (real) other minds problem. Social and Personality Psychology Compass 2(3): 1455–1474. Fillmore CJ (1992) Corpus linguistics or computer-aided armchair linguistics. In Svartvik, J (ed.) Directions in Corpus Linguistics: Proceedings of Nobel Symposium 82 Stockholm, 4–8 August, 1991. Berlin: De Gruyter Mouton. Gal S and Irvine J (1995) The boundaries of languages and disciplines: How ideologies construct difference. Social Research 62(4): 967–1001. Habermas J (1991). A reply. In Honneth A and Joas H (eds.) Communicative Action: Essays on Jurgen Habermas’s ‘The Theory of Communicative Action’. Cambridge, MA: MIT Press, 214–264. 248



Object-of-Study and Methodology

Hessels LK and van Lente H (2008) Re-thinking new knowledge production: A literature review and a research agenda. Research Policy 37: 740–760. Hjelmslev L (1942) Langue et parole. Cahiers Ferdinand de Saussure 2: 29–44. Hymes D (1972) On communicative competence. In Pride JB and Holmes J (eds) Sociolinguistics. Harmondsworth, UK: Penguin Books, 269–293. Järvikivi J, van Gompel RPG, Hyönä J and Bertram R (2005) Ambiguous pronoun resolution. Psychological Science 16(4): 260–264. Kecskes I (2010) The paradox of communication: Socio-cognitive approach to pragmatics. Pragmatics and Society 1(1): 50–73. Keysar B (2007) Communication and miscommunication: The role of egocentric processes. Intercultural Pragmatics 4(4): 71–85. Keysar B and Henly AS (2002) Speakers’ overestimation of their effectiveness. Psychological Science 13: 207–212. Kibrik AA (2011) Reference in Discourse. Oxford: Oxford University Press. Kibrik AA and Koshelev AD (2015) Kognitivnaja lingvistika—v poiskah edinstva [Cognitive linguistics—in search of unity]. In Kibrik AA and Koshelev AD (eds) Jazyk i mysl’: sovremennaja kognitivnaja lingvistika [Language and Mind: Contemporary cognitive linguistics]. Moscow: Jazyki slavajanskoj kul’tury, 21–25. Kopotev M and Mustajoki A (2008) Sovremennaja korpusnaja rusistika [Contemporary Russian corpus linguistics]. In A. Mustajoki et al. (eds.) Instrumentarij rusistiki: korpusnye podhody (Slavica Helsingiensia 34). Helsinki, 7–24. Kretzschmar Jr WA (2009) The Linguistics of Speech. Cambridge: Cambridge University Press. Linell P (1998) Approaching Dialogue: Talk, Interaction and Contexts in Dialogical Perspectives. Amsterdam: John Benjamins. Linell P (2012) On the nature of language: Formal written-language biased linguistics vs. dialogical language sciences. In Kravchenko A (ed.) Cognitive Dynamics in Linguistic Interactions. Cambridge: Cambridge Scholar Publ., 107–124 Marková I (1982) Paradigms, Thought, and Language. Chichester, NY: Wiley. McEnery T and Wilson A (2001) Corpus Linguistics. Edinburgh: Edinburgh University Press. McLuskie E (2003) Reading Humboldt through the theory of communicative action: The democratic potential of symbolic interaction. In Trudeau LJ (ed.) Nineteenth-Century Literature Criticism. Detroit: Gale, Cengage Learning, 178–192. Mustajoki A (1985). Padež dopolnenija pri otricanii v russkom jazyke: poiski novyh metodologičeskih priëmov v izučenii staroj problemy [Object Case in Russian Negative Clauses: Searching for new methodological ways to study an old problem]. (Slavica Helsingiensia 2). Helsinki. Mustajoki A (2012) A speaker-oriented multidimensional approach to risks and causes of miscommunication. Language and Dialogue 2: 216–243. Mustajoki A (2013a) Raznovidnosti russkogo jazyka: analiz i klassifikacija [Varieties of Russian: Analysis and classification]. Voprosy jazykoznanija 5: 3–17. Mustajoki A (2013b) Risks of miscommunication in various speech genres. In Borisova E and Souleimanova O (eds) Understanding by Communication. Cambridge: Cambridge Scholar Publ., 33–53. Mustajoki A (2015) Kommunikativnye neudači skvoz’ prizmu potrebnostej govorjaščego [Communication failures through the prism of the needs of the speaker]. In Kibrik AA and Koshelev AD (eds) Jazyk i mysl’: sovremennaja kognitivnaja lingvistika [Language and Mind: Contemporary cognitive linguistics]. Moscow: Jazyki slavjanskoj kul’tury, 543–551. Mustajoki A and Heino H (1991) Case Selection for the Direct Object in Russian Negative Clauses. Part II: Report on a Statistical Analysis. (Slavica Helsingiensia 9). Helsinki. Nisbett RE, Choi I, Peng K and Orenzayan A (2001) Culture and systems of thought: Holistic versus analytic cognition. Psychological Review 108(2): 291–210. Penke M and Rosenbach A (2004) What counts as evidence in linguistics? Studies in Language 28(3): 480–526. 249

Arto Mustajoki

Ryan J and Barnard R (2009) ‘Who do you mean?’ Investigating miscommunication in paired interactions. The TESOLANS Journal 17: 44–61. Sack H and Schegloff EA (1979) Two preferences in the organization of reference to persons in conversation and their interaction. In Psathas G (ed.) Everyday Language: Studies in Ethnomethodology. New York: Irvington, 15–21. Sarangi SK and Slembrouck S (1992) Non-cooperation in communication: A reassessment of Gricean pragmatics. Journal of Pragmatics 17: 117–154. Sherstinova T (2010) Quantitative data processing in the ORD speech corpus of Russian everyday communication. In Grzybek P, Kelih E and Mačutek J (eds) Text and Language: Structures, Functions, Interrelations, Quantitative Perspective. Wien: Praesens Verlag, 195–205. Shintel H and Keysar B (2009) Less is more: A minimalist account of joint action in communication. Topics in Cognitive Science 1: 260–273. Taylor T (1990) Which is to be master? The institutionalization of authority in the science of language. In Joseph J and Taylor T (eds) Ideologies of Language. London: Routledge, 9–26. Tyler LL and Marslen-Wilson WD (1977) On-line effects of semantic context on syntactic processing. Journal of Verbal Learning and Verbal Behavior 16: 683–692. von Humboldt W (1836) Über die Verschiedenheit des menschlichen Sprachbaues und ihren Einfluss auf die geistige Entwickelung des Menschengeschlechts. Berlin: F. Dümmler. Weigand E (2003) Sprache als Dialog. Sprechakttaxonomie und kommunikative Grammatik. Tübingen: Max Niemeyer. (2nd ed.) Weigand E (2004) Emotions: The simple and the complex. In Weigand E (ed.) Emotions in Dialogic Interaction. Amsterdam: Benjamins, 3–31. Weigand E (2010) Dialogue: The Mixed Game. Amsterdam: Benjamins. Weigand E (2011) Paradigm changes in linguistics: From reductionism to holism. Language Sciences 33: 544–549. Zemskaja EA (1987) Russkaja razgovornaja reč’: lingvističeskij analiz i problemy obučenija [Russian Colloquial Language: Linguistic analysis and problems of teaching]. Moscow: Russkij jazyk.

250

15 Theory and Practice Dale A. Koike

15.1 Introduction This chapter presents a look at a change of focus from theories of dialogue that attempt to explain dialogue in the abstract, to theories that directly address a dialogic ‘performance’, a topic that Weigand (2010) addresses in her Mixed Game Model (MGM) theory. Like Weigand, I consider ‘performance’ here as the concrete, situated realization of a ‘practice’, based on Chomsky’s (1965) terms of ‘performance’ in contrast with ‘competence’. ‘Practice’ denotes some contextualized activity, such as teaching, while ‘competence’ represents one’s idealized, abstract knowledge of the language, and ‘performance’ denotes language use, along with all the imperfections and ungrammaticalities that occur when people communicate. For example, if the teaching of a class is a ‘practice’, one can look at the performance of the participants (e.g. teacher and students) at a given point in time in an activity. First, after discussing some issues with regard to abstract and holistic theories, I argue in favor of a change in theoretical focus from dialogue in the abstract to a theory of ‘performance’, as does Weigand (2010). I then point to the need for a more particularized approach to a theory of dialogue; specifically, by addressing performance directly, since dialogue is given meaning only in its situated context. I present some examples of performance in the classroom language learning context, and how a theory of dialogic performance might (or might not) be applied to them. Finally, I include some suggestions for future areas of dialogic research in relation to theory and performance.

15.2  Theories of Dialogue To begin this discussion of theories, I first describe theories as they are traditionally conceived to address language in the abstract, followed by a discussion of theories that describe in a holistic sense.

15.2.1  Theory and Practice: Traditional Stances A formal linguistic theory is an attempt to explain by means of a rule-bound system why some linguistic or communication-related phenomena occur, or to describe the 251

Dale A. Koike

phenomena themselves (see Weigand 2015). Theories abound in linguistics literature; not all of them meet with success in adequately explaining actual data, either of the same research in which they occur, or in subsequent studies that try to apply them to explain their own data. In such cases, the theory must be modified to find greater explanatory power or must be abandoned. As an example of a formal linguistic theory that attempts to capture language by means of formal rules, one could point to Chomsky’s generative grammar (e.g., 1965), which comprises a defined set of components and rules in an effort to describe an idealized grammatical competence. Such theories explain how sentences are parsed and represented as well-formed strings. However, in reality human beings do not always exchange well-formed strings in interactions with one another. Therefore, the usefulness of such a formal theory is limited when describing what happens in real-world language use, as Weigand (2010: 545) states: “… describing language use as rule-governed communication meant reduction to underlying well-formed patterns”. Thus, using rules to describe everyday communication as an ‘imperfect’ system according to an idealized competence becomes an endeavor riddled with imperfections and exceptions. Dialogue in normal language use is ‘performance’, requiring knowledge and skills put into use by interlocutors in a particular context, with emergent meaning as it is constructed at the moment. Since it is commonly imperfect when measured against formal rules and idealized language, if one studies dialogue directly, describing the performance itself, one then asks, “Why do we need a theory of dialogue, and if we need one, what should it describe or explain?” I briefly address these basic questions here. Assuming that the most generalizable theories are ‘best’ for all areas of language study, the problem that often occurs is that the more abstract and formal the theory, when necessary to apply it to specific, contextualized data, there is usually some factor that confounds use of the theory, such as social factors or individual variation. Although there is no theory that covers all the phenomena, fits all the facts, and is found to be adequate by a majority of those working in the linguistics field, one would hope for a theory that would address as many issues as possible. This would require a theory that would be general, powerful and specific enough at the same time, which is nearly impossible to achieve. The next best option would be to propose a theory of a linguistic performance in context. In so doing, theorists would directly provide an explanation of the data itself, and generalize its applicability to a broader context. Such a proposal has already been put forth by Weigand (2009, 2010, 2015), which I discuss below.

15.2.2  Theory and Practice: Holistic Stances Weigand (2009, 2010, 2015) offers her ‘theory of competence-in-performance’, which she states is not an abstract theory of dialogue that describes abstract language but rather a holistic one that explains ‘competence-in-performance’. She states that it is not limited to any specific context but applies to language use in any area and in areas of specific practice. The author proposes a theory based on the concept of ‘Action ¬ → Reaction’, whereby one interlocutor does an action (such as a speech act) and the other interlocutor reacts in some way. All of this forms what she calls the ‘Dialogic Principle Proper’, and is considered the cornerstone of dialogue functions. However, beyond this principle, it is also necessary to add a component to account for all the other variables that interact in the dialogue, such as cognition, emotion, social characteristics, culture, past experiences (with the particular interlocutor of the particular dialogue, and with others) and a host of other complexities 252



Theory and Practice

that are involved in the production of dialogue between interlocutors. Thus Weigand (2010) describes principles of probability to account for the way a dialogue might transpire. To elaborate further, Weigand proposes that performance must be “based essentially on principles which can be seen as techniques of our communicative competence for tackling the problem that social interaction is mainly based on cases of probability” (2009: 182, italics as cited), meaning that language use should be described through linguistic and sociocognitive principles that are applied as the individual needs them. Briefly, in her model, there are three types of principles that operate in dialogue: (a) Constitutive Principles directed to human interaction; (b)  Regulative Principles involving aspects of human nature, such as emotions and ‘social concerns’; and (c)  Executive Principles which guide interlocutors to use specific strategies to achieve their communicative goals. These have been discussed by others (e.g. Feller 2015; Grein, Weigand, this volume). Thus Weigand’s Mixed Game Model addresses broadly the actions and reactions that occur in dialogue, not as a fixed set of rules but rather as principles that apply by a law of ‘probability and chance’. As Kecskes (Chapter 5 in this volume) points out, the fact that dialogue must be described via principles rather than by formal rules is the very reason that Weigand and others consider dialogue as an ‘action game’: dialogue derives not from given rules that are laid out for interlocutors to execute, but rather from ‘behavior games’. Besides the emotive and social aspects of talk mentioned earlier, the model also addresses communicative strategies that speakers and hearers use in their interactions, changing from a ‘logical rationality’ of abstract models to a rationality that reflects the changing nature of language in performance (see Weigand 2014). However, she also emphasizes that one needs, for example, a ‘Principle of Coherence’ to derive the meanings and understandings in dialogue as well. I agree that a direct application of theory to explain ‘dialogic performance’ or ‘action game’ is more useful for purposes of elucidating what occurs in dialogue, especially in the case of applied linguistics fields such as second language acquisition and sociolinguistics where contextual factors are all important. That is, in these fields, the researcher is concerned with dialogue situated in a given practice or other parameters, with given participants (e.g. non-native learners of Spanish in the university setting in first-year courses, who have had no previous experience with the language; or in sociolinguistics, lower-class, female native speakers ages 15–30 who have only studied up to the sixth grade and never traveled outside of their city in a work setting). Each dialogue reflects individual participant variables and the turn-by-turn co-construction of meaning that occurs in the current dialogic interaction, making formal generalities difficult to ascertain.1 By looking at theory as linked to a given performance, one then changes from an abstract, rule-governed theory of dialogue to what is in essence a theory of dialogue in ‘performance’, grounded in its particularized circumstances.

15.3 Applied Linguistics: Dialogue in the Practice of Teaching and Learning The learning and teaching of languages is only one of the areas of the broad, interdisciplinary field of applied linguistics, which generally seeks to study and suggest solutions to linguistic contexts and problems of real life. Other areas include, for example, applications to psychology, communications and computer science. I focus here on language teaching, in 253

Dale A. Koike

particular foreign-language teaching and learning, because the dialogic structure in typical teacher-centered classrooms is interesting given the typically unbalanced power structure inherent in it. I refer specifically to the way the teacher is normally the one who controls the action-reaction exchange, and how the exchange frequently is not encouraged to continue much past the initial action-reaction. We examine this issue below. According to Weigand (e.g. 2010), a theoretical framework behind approaches to the study of dialogue, as an evolving entity, must be different from those that study a stable, fixed product. Instead, the theories must reflect not only the verbalized elements of dialogic exchanges, such as the deictic orientations used in utterances (e.g. the significance of the use of inclusive ‘our’ instead of ‘their’), but must also account for the dynamics of interaction, such as requests for clarification of meaning between the participants that occurs in dialogue. These kinds of requests are especially of importance in interactions between native and non-native speakers of a language because, perhaps barring those non-native speakers who are near-native in ability, non-natives do not usually share a complete set of linguistic and cultural knowledge of the target language with native speakers. Such knowledge greatly facilitates communication between them. As an example of clarifying meaning (in her terms, ‘negotiation’ of meaning), in example 1 drawn from Mackey (1999: 558–559), the non-native English speaker (NNS) does not understand the term ‘reading glasses’ and resolves this problem through the exchanges with the native English speaker (NS). They discuss the problem to a satisfactory resolution of understanding, marked with an arrow. (1) Clarification of meaning (Mackey 1999: 558–559) NS: There’s there’s a pair of reading glasses above the plant. NNS: A what? NS: Glasses reading glasses to see the newspaper? NNS: Glassi? NS: You wear them to see with, if you can’t see. Reading glasses. → NNS: Ahh ahh glasses to read you say reading glasses. NS: Yeah. The request for clarification is carried forward by the NNS asking the NS to repeat the focal words ‘reading glasses’. After the NS does so with more elaboration to clarify the meaning, the NNS asks for confirmation of the word with ‘glassi’. This prompts the NS to elaborate on the term ‘reading glasses’ and to repeat it. It is all part of a dialogue, but it temporarily suspends the exchanges of information on content to focus on lexical clarification. Turning to the context of a teacher-centered classroom, much has been written on the Socratic dialogic exchange in the classroom between teacher and student to promote learning, in which the teacher (T) poses a question, the student (S) responds, and the teacher evaluates the response. Originally it was called the ‘Initiation-Response-Evaluation’ (IRE; Mehan 1979), in which the teacher’s evaluation stated if the student’s response was correct or not. It is considered the default structure in classroom discourse, as in examples 2 and 3. (2) T: What is this?[pointing to a clock] S: It’s a clock. T: Yes! Good!

[Initiation] [Response] [Evaluation]

If the student makes an error of any kind, the third turn may be in the form of a different kind of follow-up, such as a correction, as in (3): 254



Theory and Practice

(3) T: What is this? [Initiation] S: It’s clock. [Response] T: It’s A clock. It’s a clock. [Evaluation] Both examples reflect the dialogic structure of Action-Reaction espoused by Weigand (2010). That is, the teacher’s initiation is an action, followed by a student’s response, which is the reaction. The third evaluative turn could be considered as another action; however, exchanges in the classroom context typically stop here with the evaluation and no verbal response from the student; the teacher then continues on in the fourth turn with some other action (e.g. another question) to which a student responds. For this reason, I believe the third turn could also be seen as another reaction (by the teacher) to the student’s reaction, to which a response is not expected. In any case, this behavior is expected in the teaching and learning context, and Weigand’s Principle of Coherence would account for those expectations. Van Lier (1995), Wells (1999), Nassaji and Wells (2000), and Cazden (2001) also point out that this evaluation in the third turn shuts down the dialogue with the student, allowing the teacher to continue to initiate a new question or action. Prior to the IRE, the same kind of exchange was labeled more broadly the ‘Initiation-Response-Feedback’ dialogic structure (IRF; Sinclair and Coulthard 1975), suggesting that the third turn does not have to be evaluative alone and that the exchanges can be encouraged to continue via other kinds of responses by the teacher. An invented example is as follows, where the teacher (T) initiates and gives the follow-up, as in example 4: (4) T: Can anyone tell me why we should eat less red meat? [Initiation] S: It’s expensive. [Response] T: But also think about how it might affect our bodies. [Feedback] In example 4, the third turn consists of a subtle or partial acceptance of the student’s response with the word ‘also’ but continues on to indicate that the teacher is looking for either more or different information. In this way, the teacher initiates another set of exchanges, either from this particular student or from anyone in the class who elects to answer. Thus, the belief is that the third turn in the IRF sequence can open the classroom exchanges to more extended dialogue, leading to more opportunities for students to communicate in meaningful ways. As seen in example 4, teachers can ask for further clarification or probe for more information instead of closing the sequence with an evaluation. In this way, the IRF more closely reflects Action ¬ ® Reaction and a default dialogic structure as Weigand describes in the MGM. I discuss other examples of this in the next section. But before proceeding, I want to mention another issue related to the IRF structure that has been raised in applied linguistics literature concerning second language learning and teaching in the classroom. In the practice of teaching foreign languages, for example, researchers have focused on the kinds of dialogic exchanges between teachers and students that appear to maximize opportunities for learning. One proposal is found in the Interaction Hypothesis of Long (1983), and elaborated upon in Pica, Young, and Doughty (1987). The authors affirm that comprehensible input is necessary for learning, and they propose that certain modifications in the dialogic structure when learners are ‘negotiating meaning’ (as it is called in the second language acquisition research, referring to the clarification of meaning) or some communication difficulty are helpful for making input more comprehensible to the learners involved. Hence, following work by Long (1983) on input and interactional features, they 255

Dale A. Koike

propose that certain types of questions promote language learning, such as the clarification request made by B that elicits clarification of the previous utterance seen in example 5: (5) A:  She is on welfare. B:  What do you mean by welfare? (from Ellis 1990: 108, based on Pica et al. 1987) These kinds of questions in the context of clarifying meaning are proposed to be most beneficial when they are asked by learners who are interacting with those who know more than they do. These ‘more knowledgeable others’ may be other more advanced learners, instructors or native speakers, and their questions are assumed to promote learning because they call attention to certain linguistic aspects, such as vocabulary or the correct form of words. The principle belief is that when there are interactional modifications brought about because of these types of questions, the input becomes comprehensible and meaningful to the learner and thus aids learning of the words, structures and meanings called into question. The points focalized through the interaction should be more easily learned because they occurred in the context of meaning-making in the dialogue, and because the learner’s attention was drawn to them. Thus, the dialogic interaction espoused by the Interaction Hypothesis centers on the types of questions or statements that are hypothesized to be more helpful to language learning. Returning again to the teaching context in which a teacher is leading the class during instruction, one of the central questions of importance is not if the teacher and student follow an IRF pattern in their dialogue, but rather what kinds of questions and reactions are being asked and offered, and which of those kinds seem to promote learning to the greatest extent.2 The Interaction Hypothesis describes several types of questions and comments that are supposed to promote learning, such as ‘confirmation checks’ (e.g. “A: Mexican food have a lot of ulcers? B: Mexicans have a lot of ulcers? Because of the food” from Ellis 1992: 108), and the ‘clarification request’ shown in example 5. While it is acknowledged that the dialogic questions and problem-solving serve some positive benefit for learning, the issue is still open as to whether the dialogue actually causes learning or if it “often serves as a priming device, setting the stage for learning rather than being a forum for actual learning” (from Winke et al. 2010: 80, citing Gass 1997). In the next section I discuss some examples and other proposals for changes to the IRF to promote language learning, and examine these ideas in light of the MGM.

15.4  Examples That Question the IRF Weigand (2015) draws a distinction between a theory that provides explanatory power and the presentation of a ‘case study’ that provides a scenario for descriptive power. In this section I examine some examples of how a holistic theory can be applied to performance in a ‘case study’ to test its explanatory power. Focusing on performance, a model must be able to address a range of language uses and cover as many aspects of communication as possible. A theory must address a real practice in order to be useful and practical, instead of serving as an explanation for a generalized, abstract practice. The recognition of these dialogic aspects in performance has led to a turning point in dialogue theorizing, which is that researchers have come to recognize the need to theorize about the dialogic performances themselves. In the practice of teaching, as discussed earlier, the traditional dialogic structure (at least, in most of the Western world) is “Initiation-Reaction-Evaluation” or IRE, a tradition following the Socratic Method, reborn as the IRF. Other modifications to the IRF have been 256



Theory and Practice

proposed in an effort to maximize student learning, which of course is the driving goal in the institutional discourse of the classroom (see, for example, Miao and Heining-Boynton 2011). This IRF structure has also been widely used in language classes. Exploring other aspects of the IRF, Hellermann (2003, 2005) examines the IRF sequence in other evidence from the classroom context, and notes the important role of prosody in these exchanges for teaching and learning. Focusing on the third turn in the IRF by the teacher, the author looks at the function of the prosodic cues that occur in the sequences. The cues serve to shape the “recurring social practices of the discourse within which teaching and learning are accomplished” (Hellermann 2003: 80). For example, by means of an acoustic analysis of an IRF using digitized sound samples with PRAAT software (Boersma and Weenink 2001), the author found that teachers often repeat in the third turn (Feedback) what the student says in the second turn (Response), using the same prosody that the student used. Regarding the prosody itself, the author notes that teachers indicate a positive evaluation of what was just said by means of: (a) rhythmical placement synchronized with the student response in the previous turn; (b) falling pitch contour; (c) mid-level pitch; and (d) a longer duration than the student response (Hellermann 2003: 88). For example, in example 6, the teacher initiates the IRF in line 10 with a question: “When the buffalo’s gone, what didn’t they have?” Faith offers the answer “food”, which the teacher repeats in the third turn in line 17 with what Hellermann considers to illustrate an affiliative quality. (6)

Example of repetition, reproduced from Hellermann (2003, Ex. 3) 10 T: right. when the 11 /buffalo’s 12 /gone 13 /what didn’t they 14 /have. 15 Jill: /(I know) 16 Faith: /food. 17 T: /food. (Hellermann 2003: 89)

Thus the teacher provides a positive evaluation of the student’s previous turn in line 17 through the use of prosodic cues and repetition. He also points out that the teacher’s prosody is “middle pitch level with a falling contour” (Hellermann 2003: 89–90) and that the length of the utterance is slightly longer than that of the student. He claims that such cues indicate to the students that the teacher is not assessing the previous turn but rather showing approval through affiliation, without closing down the exchanges. In other examples, Hellermann illustrates that different prosodic patterns with the same repetition pattern can present different meanings, sometimes signaling a desire for additional dialogue with the students. Students do not appear to have difficulty understanding these cues but the author acknowledges that they also form part of the IRF dialogic patterns. As the author points out, they form part of triadic sequences that are relevant to institutional discourse such as that found in classroom teaching, and not that of everyday conversation. For example, repetition in the third slot in a non-teaching sequence seems to indicate a communicative difficulty, as in example 7: (7) Exchange from a non-classroom context (from Hellermann 2003: 100) 1 A: Where’s the car? 257

Dale A. Koike

2 B: on Johnson. ® 3 A: on Johnson. 4 B: What’s wrong? Thus the teacher’s repetition in the third turn as in example 6 seems particular to contexts of institutional discourse, such as that found in the practice of classroom teaching, and is used more as a second reaction to the second turn, sometimes implying another response is needed from the student. I note that in a classroom in which the teacher elects not to be the center of the activity, which is currently a more common format than a teacher-centered class in most institutions of higher learning, students are asked to do dialogic tasks in pairs or small groups, usually to solve some kind of problem. This format also entails dialogue, but among the learners themselves. However, when the task is too challenging for the student’s level of competence in the target language, students in pairs or groups may also begin to concentrate so much on the linguistic expression of the task that they stop creating a true exchange of meaning between themselves, as in example 8. (8) Difficult dialogue in dealing with a task in a language class; two male students in a fourth-semester Spanish class

1 R: Sí sí. (reading aloud) “¿Has participado alguna vez en una manifes tación o una huelga para protestar contra algo?” No.  Yes yes. “Have you participated sometime in a demonstration or a strike to protest against something?” No.

2 M: No. No no um, igualmente. Um um porq- pero es muy importe No. No no um, too. Um um becau—but it is very import

258



3 R: Importante? Important?



4 M: Importante um= Important um=



5 R: A partici[par= To partici[patê=



6 M:



7 R: =[a participar= =[to participate=



8 M: =participar= =to participate=



9 R: Participé. I participated.

=[to let’s see [par- = = [to let’s see[par- =



Theory and Practice



10 M: [partic— [partic—



11 R: [Something comes after the ( ) period. [Something comes after the ( ) period.



12 M: It is important. It is important.



13 R: ( ) [simple subject? ( ) [simple subject?



14 M: [Sí ( ). Es importante it’s important to, [Yes ( ). It is important it’s important to,





es importante to participate. Y’all emphasize the es it is important to participate. Y’all emphasize the it is



es importante partici[pé it is important I partici[pated

15 R: [or you put that es importante [or you put that it’s important ((talk continues))

In this example, student M concentrates only on the form of his own utterances and stops listening for a time to R even though R provides the correct utterance in turn 7. This example would seem to present problems for the Action ® ¬ Reaction pattern, not only for the lack of a true reaction to the interlocutor’s actions, but also for a suspension of the Principle of Coherence. The basic problem stems from the fact that in a context of teaching (and in other contexts as well) and the imbalance in power, the multifunctionality of utterances is often used to mask true intentions. In example 4, in which the teacher in her feedback to the student asks a question that both signals only partial acceptance (with ‘also’), but also that more information is needed (with ‘but), it is not clear if this utterance is a reaction (of feedback) or an action (initiating another exchange with the student), or both. This multifunctional issue would seem to confound the classification of action and reaction in the dialogue. But perhaps we cannot really consider example 8 a ‘dialogue’, even though the two people are exchanging utterances, but they are completely focused on forms instead of communicating meaning in a broad sense. Or perhaps in the context of the language classroom, some of the principles can be suspended as students pursue their own individual goals for institutional demands. These issues are left for future discussion.3 There is another issue of importance in the study of dialogue in the language teaching and learning class environment: the influence of affect and motivation. That is, in a language class, one assumes that at least the majority of the students and the teacher are interested in the subject material and that they share a common goal for the students to learn the new language. At the same time, it is a foregone conclusion that there may be students who are not genuinely interested in learning the new language and are not motivated by a desire to learn skills and content in that language. Since efforts to create language in dialogue with 259

Dale A. Koike

a partner or the teacher in the language class are driven by such goals, a lack of interest and motivation would impact the quality of the dialogue. In addition, it is entirely possible that student participants in a dialogue have different goals, either at a local level or in general, as may be illustrated in example 8 above. In such cases, the dialogue may vary in its structure and principles followed from those described by the MGM.

15.5  Conclusion: Applying the Holistic Theory to the Practice When examining the same dialogues in examples 5 and 6 with Weigand’s holistic MGM, the basic pattern of Action ® ¬Reaction (the Dialogic Principle Proper) does apply to describe its structure. However, by its nature one could argue that the third turn in the IRE/IRF could be viewed as a reaction to the second turn, or perhaps both an action and a reaction at once. Even though the IRF has always been considered a triadic turn structure, one could consider the teacher’s turns to be actions (initiative and evaluative) and the student’s turn as a reaction. However, in doing so, the third turn seems to lose its qualitative connection to the role that the teacher plays in the practice of teaching with this Socratic method; that is, providing information, checking on comprehension or some other function to draw the student into the thought processes, and then reacting to what the student offered in the second turn. These roles interact with the kind of questioning and reacting that the teacher and student offer in this teaching and learning context, in effect, creating a reaction by the student and another reaction by the teacher in the third turn that somehow assesses the student’s reaction. Also, considering the repetition and intonation issue raised by Hellermann, one could ask if Weigand’s Principle of Coherence would guide the learner to understand that such devices imply either a kind of acceptance or a type of repetition to signal the need for more elaboration or reformulation by the learner. Since these repetition and intonation patterns do not occur in everyday speech, and may even be culture-specific, how do non-native speakers understand these signals in the teacher–student dialogue? Obviously, as the students learn classroom discourse, they come to expect these patterns. But what initiates this understanding? And how do we explain those complex, seemingly ‘non-dialogic’ sequences that sometimes occur when learners are working with others in an activity, as in example 8? These questions do not detract from the utility of a holistic theory like the MGM for describing dialogue in the institutional language-learning context. The model serves to describe the performance by students and teachers. The few questions I have raised may well extend beyond the scope of the model itself and the type of dialogue in the language classroom context. Perhaps the MGM does not easily capture these examples because these institutional dialogues in examples 4, 6 and 8 do not represent exchanges that function to communicate meaning in the same sense that one has in conversational dialogues, as Hellermann points out. That is, one may question if these instructional exchanges between students and teacher in the classroom are truly representative of ‘dialogue’ in the sense of meaning-making in a broader view (instead of, say, a focus on linguistic forms), and ‘realistic’ outside of the context of a classroom, since the goal of this dialogue is to instruct and learn. In applied linguistics, analysts aim to take these particular, contextualized cases and study them profoundly. So it is understandable that a holistic dialogue theory does not always match the practice. Finally, I return to the question asked at the beginning of this chapter: Why do we need an abstract theory of dialogue? After all I have discussed, and considering again that dialogue itself is a situated performance, the logical conclusion seems to be that for applied

260



Theory and Practice

linguists it is more profitable to address dialogue in the performance of a practice, which, using Weigand’s (2015) terms, could also be seen as ‘case studies’. The case studies should support the general model, which the MGM provides.

15.5.1  Areas for Future Research Other areas in which one could examine dialogic performance in a holistic perspective to include interfaces of practices are: • Dialogue and technology: With recent technological innovations, researchers can capture more aspects of interaction than before. Beyond recording an interaction via a simple audio or a video recording, the continuous improvements in technology afford the researcher greater insights into the dialogue. For example, Koike and Blyth (2014) use a split screen video recording software to allow the researcher to show the data immediately after recording to the informants. Such a method encourages the use of retrospective questioning of the participants to understand their interior thoughts during the recording, and can also show data to the participants themselves for instructional purposes. • Dialogue and dynamic systems: Larsen-Freeman (2007) discusses the complexity of systems that change over time according to conditions that affect them, a description that certainly can fit what happens in dialogue and the co-construction of meaning. While the point of Dynamic Systems Theory (Larsen-Freeman and Cameron 2008) is not to predict what will happen in such a change but rather to offer an explanation for what happened, the theory holds promise for future studies of dialogue as an evolving and modifiable entity, susceptible to variables that can affect it as it occurs. While the MGM as a model of performance is dependent on variables that change over time, such as in the action game of negotiation (e.g. Weigand 2001), it would be interesting to see how the MGM yields insight into specific aspects of those changes. • Dialogue and emotion: Closely tied to the evolving aspects of dialogue is the issue of changing states, which of course include emotions (Grein, this volume; Lyle 2015). Now that the technology of speech recognition has evolved to the point that it can capture some form of recognition and adaptation to emotions (Pittermann et al. 2010), we can imagine that in the near future there will be experimental research that can advance our knowledge of emotion in dialogue. This interface between the psychological and affective aspects of talk will help to better understand the aspect of ‘mental states’ and their role in dynamic changes, something that the MGM addresses in general principles. • Dialogue as ‘polylogue’: As interactions become increasingly multi-party (e.g. in online chats, in any group event), and as researchers are able to capture the interaction among the participants with the aid of technology, a focus of study can be broadened to what Kristeva (1977) terms ‘polylogue’. Kristeva refers to multi-party interactions such as multiple speeches. Others have studied multi-party interactions (e.g. Szymanski et al. 2005), and when such ‘polylogues’ become the object of study, they have shown how the group dynamics change everything about the dialogue. Until recently, most studies have focused on dialogue among pairs, probably because it is easier to separate the voices, but having the capability to do a more micro analysis of a polylogue, through video, should shed more light into aspects such as how the talk changes from moment to moment, how common ground is sought and achieved, and how the individual background factors impact the group dynamics, such as in alignment.

261

Dale A. Koike

Notes 1 I use the term ‘interaction’ and ‘talk’ at times in place of dialogue, but for me, both terms are broader in scope. In my view, ‘interaction’ and ‘talk’ denote not only the oral language used and the c­ o-­construction of meaning through words but also include other means of expression such as gesture and facial expressions. However, ‘interaction’ and ‘talk’ are not clearly related to the dialogic methodology of action and reaction. 2 Feller (2015) also discussed this issue in the language learning context but his objective was to discover which questions elicited ‘deep learning’, in the sense of “learning based on deep reasoning and active knowledge building (e.g. Krathwohl 2002; Sfard 2001)” (Feller 2015: 197). He claims this ‘deep learning’, or what he calls “Dialogic Knowledge Building”, is achieved through questions that “scaffold deep reasoning along higher level cognitive processes like ‘evaluating’ and ‘creating’” (195). 3 Kempson (Chapter 12 in this volume) describes a variation of the Action-Reaction structure in her chapter. Kempson discusses the case in which Speaker A begins a sentence and it is then completed by Speaker B, in a type of linguistic ‘projection’. In such cases, there is again a multifunctionality at work in the sense that Speaker B not only extends the content initiated by Speaker A but also redefines the context by taking over the turn as the primary speaker, and the dialogue then evolves. This is not the same as the multifunctional utterances that can occur in the third turn by the teacher as in my example 4, but it also illustrates how the notion of Action-Reaction can be seen as complex and not easily defined in all instances.

References Boersma, Paul and David Weenink. 2001. Praat: A system for doing phonetics by computer. Glot International 5(9/10): 341–345. Cazden, Courtney. 2001. Classroom Discourse: The Language of Learning and Teaching, 2nd ed. Portsmouth, NH: Heinemann. Chomsky, Noam. 1965. Aspects of the Theory of Syntax. Cambridge, MA: MIT Press. Ellis, Rod. 1990. Instructed Second Language Acquisition. Oxford, UK: Blackwell. Feller, Sebastian. 2015. Dialogic knowledge building in learning communities: Discovering an electric circuit through collaborative learning. In Dialogue in Multilingual and Multimodal Communities, ed. by Dale Koike and Carl Blyth, 195–219. Amsterdam: John Benjamins. Gass S (1997) Input, Interaction, and the Second Language Learner. Mahwah, NJ: Lawrence Erlbaum Associates. Hellermann, John. 2003. The interactive work of prosody in the IRF exchange: Teacher repetition in feedback moves. Language in Society 32: 79–104. Hellermann, John. 2005. Syntactic and prosodic practices for cohesion in series of three-part sequences in classroom talk. Research on Language and Social Interaction 38(1): 105–130. Koike, Dale and Carl Blyth. 2014. A metapragmatic approach to intercultural awareness: Gaining a dialogic perspective. Paper given at the American Pragmatics Association 2014, UCLA. Krathwohl, David. 2002. A revision of Bloom’s Taxonomy: An overview. Theory into Practice 41(4): 212–218. Kristeva, Julia. 1977. Polylogue. Paris: Seuill. Larsen-Freeman, Diane. 2007. Complex systems and applied linguistics. International Journal of Applied Linguistics 7: 226–240. Larsen-Freeman, Diane and Lynne Cameron. 2008. Complex Systems and Applied Linguistics. Oxford, UK: Oxford University Press. Long, Michael. 1983. Native speaker/non-native speaker conversation in the second language classroom. In On TESOL ’82: Pacific Perspectives on Language Learning and Teaching, ed. Mark Clark and Jean Handscombe, 207–225. Washington, DC: TESOL. Lyle, Cory. 2015. The complex nature of language-related episodes. In Dialogue in Multilingual and Multimodal Communities, ed. Dale Koike and Carl Blyth, 25–48. Amsterdam: John Benjamins: 25–48. 262



Theory and Practice

Mackey, Alison. 1999. Input, interaction, and second language development: An empirical study of question formation in ESL. Studies in Second Language Acquisition 21(4): 557–587. Mehan, H. (1979) Learning Lessons: Social Organization in the Classroom. Cambridge, MA: Harvard University Press. Miao, Pei and Audrey Heining-Boynton. 2011. Initiation/response/follow-up, and response to intervention: Combining two models to improve teacher and student performance. Foreign Language Annals 44(1): 65–79. Nassaji, Hossein and Gordon Wells. 2000. What’s the use of ‘Triadic Dialogue’? An investigation of teacher-student interaction. Applied Linguistics 21: 376–406. Pica, Teresa, Richard Young and Catherine Doughty. 1987. The impact of interaction on comprehension. TESOL Quarterly 21: 737–758. Pittermann, Johannes, Angela Pittermann and Wolfgang Minker. 2010. Emotion recognition and adaptation in spoken dialogue systems. International Journal of Speech Technology 13: 49–60. Sfard, Anna. 2001. There is more to discourse than meets the ears: Looking at thinking as communicating to learn more about mathematical learning. Educational Studies in Mathematics 46(1/3): 13–57. Sinclair, John and Malcolm Coulthard. 1975. Towards an Analysis of Discourse. The English Used by Teachers and Pupils. London: Oxford University Press. Szymanski, Margaret, Paul Aoki, Luke Plurkowski, James D. Thorntonand Allison Woodruff. 2005. Organizing multi-party talk: Orienting to multiple simultaneous conversations. 9th International Pragmatics Conference (IPrA) 2005, Riva del Garda, Italy. van Lier, Leo. 1996. Interaction in the Language Curriculum: Awareness, Autonomy, and Authenticity. London: Longman. Weigand, Edda. 2001. Games of power. In Negotiation and Power in Dialogic Interaction, ed. Edda Weigand and Marcelo Dascal, 63–76. Amsterdam: John Benjamins. Weigand, Edda. 2009. Language as Dialogue: From Rules to Principles of Probability. Amsterdam: John Benjamins. Weigand, Edda. 2010. Dialogue: The Mixed Game. Amsterdam: John Benjamins. (Dialogue Studies 10) Weigand, Edda. 2014. Rationality of performance. Philosophia Scientiae 18(3): 247–268. Weigand, Edda. 2015. Dialogue in the stream of life. Language and Dialogue 5(2): 197–223. Wells, Gordon. 1999. Dialogic Inquiry: Toward a Sociocultural Practice and Theory of Education. Cambridge, UK: Cambridge University Press. Winke, Paula, Susan Gass and Tetyana Sydorenko. The effects of captioning videos used for foreign language listening activities. Language Learning and Technology 14(1): 65–86.

263

16 The Sociobiology of Language What Mirror Neurons Can Tell Us Marco Iacoboni

16.1 Introduction The scientific study of the biological, ecological, and evolutionary aspects of human social interactions through language can take many forms (Wilson, 1975). This chapter takes the perspective opened up by a fairly recent neuroscientific discovery that has generated broad interests in many academic fields. About twenty-five years ago, single cell recordings in the ventral sectors of the premotor cortex in the macaque brain revealed a population of neurons whose spiking activity had unexpected properties (di Pellegrino et al., 1992). Recordings in the same region from previous studies had already suggested that neurons in the ventral premotor cortex had rather complex properties (Rizzolatti et al., 1988). The premotor cortex had been previously associated with the selection and preparation of movements (Wise et al., 1997). Other studies had also suggested a broad differentiation between the dorsal and ventral sector of the primate premotor cortex. The dorsal premotor cortex seemed specialized for ‘non-standard’ or arbitrary sensory-motor associations (Wise et al., 1996), as when we stop the car at a red traffic light. The ventral premotor cortex, on the other hand, seemed more specialized for ‘standard’ or non-arbitrary sensory-motor associations, as when we shape our fingers around an object we want to grasp, for instance an apple (Jeannerod et al., 1995). While there are in principle many different ways of grasping an apple, the overwhelmingly used grasping action to grab an apple is what is called whole-hand prehension. This type of grip shapes all fingers and the palm around the apple. It is a ‘standard’ or nonarbitrary sensory-motor association because the type of grasping action is the best fit for the shape of the object to be grasped (whereas we could stop the car at a traffic light for any arbitrarily agreed upon color). In the anterior sector of the ventral premotor cortex, neurons code for goal-oriented actions, like grasping an apple. Most of these neurons seem not concerned at all with the detailed motor plans necessary to achieve the goal. The clearest examples are grasping cells that fire for grasping the same object (say, an apple), when the monkey uses the right hand, the left hand, and even the mouth. Given that there is not a single muscle in common between the three actions performed with these three body parts, it is clear that what is coded by the firing of these neurons is not a specific motor plan but rather the goal of the action. Ventral premotor neurons show further evidence of goal-oriented coding by firing 264



What Mirror Neurons Can Tell Us

selectively for specific kinds of actions that require activations of similar muscles and muscle groups. For instance, some neurons fire for grasping an object, while others for holding the same object, and others for bringing the object to the mouth (say, a peanut). Grasping, holding, and bringing to the mouth actions all require the contraction of identical hand muscles, and yet neurons in ventral premotor cortex activate selectively for one type of action but not the others (Rizzolatti et al., 1988). Some of the motor neurons in the anterior sector of the ventral premotor cortex (that is, neurons that activate in association with a goal-oriented action of the macaque) also have rather surprising sensory properties. Indeed, a population of cells that has been called ‘canonical’ neurons, fire not only when the monkey performs a goal-oriented action like grasping an apple, but also when the monkey is observing a graspable object without grasping it. Here, the response of the neuron is clearly a ‘visual’ response. Why do we have visual responses in neurons located in a brain area devoted to the control of movement? To better answer this question, it is useful to consider some other properties of these cells. There are two main types of grasping neurons in the anterior sector of the ventral premotor cortex, and they code for the two main types of grasping actions that we can perform. To grasp an object we generally use either all fingers and the palm of our hand (the whole hand prehension we perform to grasp something like an apple), or we use the thumb and the index finger to grasp a much smaller object, like a pill. This type of grasp is called precision grip. Whole-hand prehension and precision grips are dictated for the most part by the size of the objects we intend to grasp. Canonical neurons also code either for whole-hand prehension or precision grip, when the monkey is grasping. Visually, canonical neurons respond only to object whose size is ‘congruent’ with the type of grasping action that is coded from a movement standpoint. That is, if a canonical neuron fires while the monkey performs a precision grip, it will also fire when the monkey is simply observing a small graspable object, for instance a raisin. The same neuron will not fire at the sight of an apple. If, on the other hand, a canonical neuron fires while the monkey performs a whole-hand prehension, it will also fire when the monkey is simply observing a large graspable object, for instance an apple. In this case the same neuron will not fire at the sight of a small object like a raisin (Raos et al., 2006). Clearly, these visual responses seem to code ‘how to grasp’ the observed object, rather than visual properties that have little to do with the kind of action that one can perform on the object (for instance, the color of the apple: whether the apple is green or red is irrelevant to the type of grasping action one can perform on the apple). That is, while these responses are visual (because the monkey is not making an action directed at the object), they retain ‘motor’ characteristics, because they are selective for the kind of action that the object affords. In other words, they represent some form of intentionality that is tightly associated with the neurons in the anterior sector of the ventral premotor cortex.

16.2  Mirror Neurons While recording from the anterior sector of the ventral premotor cortex of the macaque, a group of neurophysiologists made an unexpected discovery. Some of the cells coding for goal-oriented actions also fired when the monkey was simply observing someone else making a goal-oriented action (di Pellegrino et al., 1992). After a rather understandable initial disbelief, further experiments confirmed the phenomenon (Gallese et al., 1996). These neurons were called mirror neurons because it was as if the monkey was watching her own actions reflected by a mirror while watching someone else performing the action. Mirror neurons were initially divided into two main categories: strictly and broadly congruent mirror cells. 265

Marco Iacoboni

Strictly congruent mirror neurons fired for the same action, whether performed or observed. Broadly congruent mirror neurons, which tend to be about two-thirds of all mirror neurons recorded so far, fired for actions that either achieved the same goal or belonged to a chain of actions meant to achieve the same intention. For instance, a broadly congruent mirror cell would fire while the monkey grasps a piece of food, and while the monkey observes someone else bringing the food to the mouth (Gallese et al., 1996). Grasping the food and bringing it to the mouth belong to the same chain of actions that leads to eating, thus sharing an intention that is highly relevant to the agent and/or observer. Mirror neurons also fire for actions that are partially hidden (Umiltà et al., 2001). In the relevant experiment, the monkey observes a human experimenter grasp a piece of food and mirror neurons obviously fire. In other trials, the monkey sees a piece of food on a table, but then a screen occludes the direct sight of the food. After that, the monkey sees the human experimenter reaching behind the screen. The monkey knows that there is a piece of food behind the screen but can not obviously see the actual grasp, since sight is occluded by the screen. Yet, mirror neurons fire, as if the grasping action was actually visible. Two other control conditions establish that this firing actually codes for the partially occluded grasping action. In one condition, the human experimenter pantomimes grasping something on the table, since there is nothing to be grasped on the table. Here, mirror neurons do not fire because mirror neurons in macaques do not fire at the sight of a pantomime (macaques do not pantomime, so there is not much to be mirrored in this situation). In another condition, the macaque sees that there is nothing on the table, and then a screen partially occludes the sight of the table, and a human experimenter reaches beyond the screen, as if to pantomime a grasp (since there is nothing to be grasped). In this case, mirror neurons also do not fire. Note that this experimental condition is visually identical to the condition of occluded grasping described above, and that led to the firing of mirror neurons. The only difference between the two experimental conditions is the prior knowledge of ‘something graspable’ (the piece of food) behind the screen in the condition that triggers the firing of mirror cells. This suggests that, in the condition of partially occluded grasping, the firing of the cell does code the unseen grasping action. Another important property of mirror cells is that they respond to action sounds, even when the action is not visible (Kohler et al., 2002). For instance, a mirror cell with auditory properties will fire when the monkey breaks a peanut, when the monkey sees someone else breaking a peanut, but also when the monkey simply hears the sound of someone breaking a peanut, even when the action is not seen at all. This is conceptually important for two reasons. First, it shows that mirror cells can be multimodal cells that code conceptually similar events (the breaking of a peanut) in multiple modalities (movement, vision, sound). Second, it shows that this population of cells does have access to the auditory channel. Both properties are relevant to theories regarding the involvement of mirror neurons in language, as we will see in later sections of the chapter. Two more properties of mirror cells are worth discussing with regard to the sociobiology of language. The first one is related to selective coding for observed actions that occur in peri-personal space (defined as the sector of space that can be reached by the person) or extra-personal space (defined as the sector of space beyond the reach of the person) (Caggiano et al., 2009). With regard to the sector of space in which observed actions occur, mirror neurons can be classified in three subpopulations. Some cells will fire regardless of whether the action occurs in peri-personal or extra-personal space. Some other mirror cells will fire when the action occurs in peri-personal but not extra-personal space. Finally, some mirror cells will not fire when the action occurs in peri-personal space but will fire when the 266



What Mirror Neurons Can Tell Us

action occurs in extra-personal space. Why these different responses according to the location in space of the actions of others and with respect to the body of the observer? The main difference between peri-personal and extra-personal space is that in the former it is possible to act, whereas in the latter our actions are at best mediated by objects (I can throw a stone in a far space location, but I cannot grasp something that is located outside my reach). We can definitely interact with agents performing actions in our peri-personal space, but we cannot do that when the action is in far space, outside our reach—which means that an important component of the coding of the actions of other people that is provided by mirror neurons has to do with coding the possibility of interacting with others. As language is an essential medium for social interactions, mirror neurons may provide some general coding mechanisms that are foundational for language mediated social interactions. Finally, mirror neurons do exist for specific mouth actions (Ferrari et al., 2003). Both ingestive (eating, drinking) and communicative actions through facial expressions are coded by mirror cells. These properties of mirror cells make them an ideal neuronal population for affective communication through facial expressions, a type of non-linguistic communication that is central to human conversation but also empathy.

16.3  The Anatomo-Functional Hypothesis As we already mentioned, mirror neurons were originally discovered in the anterior sector of the ventral premotor cortex of the macaque brain. This area is called area F5 and contains neurons that prevalently code for hand actions, mouth actions, and hand-to-mouth actions. A comparative neuroanatomy theory suggests that area F5 of the monkey brain may be the homologue of human Brodmann area 44, an area considered the posterior part of Broca’s area, one of the key language areas in the human brain (Rizzolatti and Arbib, 1998). The theory is based both on anatomical evolution of sulci (a furrow or groove) in the lateral wall of the frontal lobe and microscopic similarities between the two areas. With regard to sulci evolution, area F5 is situated immediately posterior to the ventral sector of the arcuate sulcus, that divides premotor from prefrontal cortex in the macaque brain. The evolutionary theory suggests that the arcuate sulcus of the macaque evolved into the two major horizontal sulci of the human frontal lobe. The dorsal sector of the arcuate sulcus would have evolved into the superior frontal sulcus in humans, whereas the ventral sector of the arcuate sulcus would have evolved into the inferior frontal sulcus in humans. If this is correct, then whatever region was located immediately behind the ventral sector of the arcuate sulcus (like area F5) should have been pushed downward, to occupy the posterior part of the inferior frontal gyrus in humans. This is exactly where Brodmann area 44 is located. If we also consider that there are microscopic anatomic similarities between area F5 and Brodmann area 44, in that they are both fairly dysgranular areas, then the hypothesis that area F5 of the macaque is somewhat a precursor of Brodmann area 44 in humans becomes even more plausible. Beyond the anatomical similarities and correspondences, functional considerations reinforced the hypothesis that mirror neurons may have provided mechanisms that are important for the evolution of language and protolanguage. A linguistic theory that preceded by some decades the discovery of mirror neurons posited that at functional level, for any communication to occur, there must be a common code between the sender and the receiver of a message (Liberman and Mattingly, 1985). The firing of mirror cells both while performing and observing an action clearly provides that common code. The linguistic theory posited that the common code for speech was actually the articulatory motor plan to produce the speech sounds. This theory, called the motor theory of speech perception, suggested that whenever 267

Marco Iacoboni

we hear a speech sound, we internally retrieve the motor plan to emit those sounds, even though we are obviously silent while listening. A rather amusing anecdote regarding mirror neurons and the motor theory of speech perception is the following. The group of neurophysiologists who discovered mirror cells were initially and understandably skeptical about the phenomenon. At some point, however, after repeated testing, the skepticism had to recede due to the reproducibility of the phenomenon under increasingly controlled experimental conditions. The critical tests were done when the head of the lab was away, traveling. When he got back, he was briefed about the latest studies and the inevitable conclusions that they suggested. At that point, even though the experiments conducted so far had to do mostly with grasping actions, he said that the behavior of these cells reminded him of the motor theory of speech perception! Another aspect of the properties of mirror neurons that provides links to evolutionary theories regarding the origins of language, is the gestural one. Theories on the origins of language have proposed that gestural communication was a key factor in the acquisition of language in humans (Corballis, 2010). Even now, gestural communication plays a major role in human communication and conversation. The coding of hand actions by mirror neurons may have facilitated in the observer the understanding of the gestures of other people. Finally, if one looks at language not so much as a sign system, but rather as the expression of social agents who are interacting through dialogue, who are acting and reacting to each other (Weigand, 2002), then mirror neurons and their properties are ideally suited to implement this behavior. Indeed, mirror neuron properties are much better suited to this view of language as dialogical interaction than to the more classical view of a sign system. These considerations, regarding the potential role of mirror neurons in language evolution, were made fairly early on after the initial discovery. Since they made a rather convincing case, they inspired a number of empirical studies that further tested the role of neural mirroring in speech perception and language.

16.4  Neuroscience Studies on Neural Mirroring and Speech The earliest human study on neural mirroring was performed by using transcranial magnetic stimulation (TMS) (Fadiga et al., 1995). The logic of this study was as follows: if humans have mirror neurons, these neurons should activate while watching actions of other people. Since there are mirror neurons in premotor cortex, and the premotor cortex is tightly connected with the primary motor cortex, the activation of premotor mirror neurons should make the motor cortex more excitable. This predicts higher excitability of the motor cortex while watching actions compared to watching other kinds of stimuli, a phenomenon often called ‘motor resonance’ (since the motor system of the observer seems to ‘resonate’ with the motor system of the agent performing the observed action). TMS can measure excitability of the motor corticospinal system with an excellent readout, that is, the contractions of muscles controlled by the motor cortex itself. Indeed, a stimulating TMS pulse over the motor cortex induces an involuntary contraction of a body part controlled by the sector of the motor cortex stimulated by TMS. Generally, levels of motor excitability in this literature are measured with motor evoked potentials (MEP), that is, the electrical activity associated with the muscle contraction. An early human study demonstrated that MEPs had bigger amplitude when subjects were watching actions of other people, compared to control conditions, as predicted by the hypothesis that the human premotor cortex contains mirror neurons. This early study also demonstrated that the motor resonance observed with 268



What Mirror Neurons Can Tell Us

TMS-induced MEPs was ‘muscle specific’, that is, it was stronger in the muscles involved in the perceived, observed action (Fadiga et al., 1995). Muscle specificity of motor resonance inspired the first TMS study of speech perception. In this study, subjects listened to words and pseudo-words while their motor cortex was stimulated (Fadiga et al., 2002). Subjects were listening to words and pseudo-words that required—when pronounced—either strong tongue movements or slight tongue movement. Subjects’ MEPs from tongue muscles were higher when listening to words and pseudowords involving, when uttered, strong tongue movements, thus supporting the hypothesis that listening to spoken words produces the activation of specific speech motor plans necessary to produce the perceived speech sounds. Subsequently, imaging studies supported this early TMS study by showing activation of motor areas for speech while subjects were also listening to speech sounds (Wilson et al., 2004). These activations showed topographic specificity and correspondence between speech production and perception (Pulvermüller and Fadiga, 2010; Pulvermüller et al., 2006), and activation levels that discriminated between speech and non-speech sounds during perception (Wilson et al., 2004). These speech motor areas demonstrated also different levels of activation for native and non-native speech sounds during speech perception (Wilson and Iacoboni, 2006). Furthermore, functional connectivity studies revealed strong connectivity between the speech motor areas also activated during speech perception (mirroring areas for speech) and the superior temporal cortex, the primary cortical system in the human brain to process sounds (Wilson and Iacoboni, 2006).

16.5  Correlation and Causation: TBS and the Simulation Model While all this evidence is rather compelling with regard to the hypothesis that speech motor areas are activated during speech perception, it still says nothing with regard to the functional role of these activations in speech perception. In principle, the muscle-specific increased excitability, the motor activations, and the functional connectivity between motor areas and auditory cortex may be entirely epiphenomenal, or, at the very least, not crucial for the ability to perceive speech sounds. To test whether all these neural markers of motor activation during speech perception are essential to the ability of perceiving speech sounds, it is necessary to perturb the activity of the speech motor areas during speech perception and test whether this perturbation induces a functional impairment in speech perception. This can be done non-invasively with specific TMS protocols. A study using low frequency repetitive TMS, a type of TMS protocol that produces cortical inhibition in the stimulated area, demonstrated that indeed speech motor areas are essential to speech perception (Meister et al., 2007). In this study, subjects were stimulated using image guided TMS. With image guided TMS, it is possible to visualize the stimulated area when the MRI of the stimulated subject is available. In a first experiment, subjects performed a speech perception task and a color discrimination task and were stimulated over the motor cortex and the auditory cortex. Task performance had been equated for each subjects with adaptive staircase procedures before stimulation. Performance after stimulation—compared to before stimulation—demonstrated a selective impairment in speech perception after motor stimulation only. This result was the first example of induced perceptual deficit due to interference with activity in a motor cortical area. The result was also rather surprising, especially considering the lack of a TMS effect on speech perception when the auditory cortex was stimulated. A second experiment followed the same design as the first one, except that the color discrimination task was replaced by a tone discrimination task. 269

Marco Iacoboni

This second study demonstrated again a speech perception deficit after stimulation of the motor cortex, compared to before stimulation, but also a tone discrimination deficit after stimulation of the auditory cortex, compared to before stimulation. The lack of a task by stimulated site interaction, however, suggested that both motor cortex and auditory cortex are somewhat involved in speech perception. But how? One possibility, which fits the TMS data well, is that the auditory cortex provides an initial acoustic analysis of speech sounds and then relays this information to the motor cortex. The motor cortex, on the basis of this acoustic analysis, would simulate the motor plan necessary to emit those speech sounds. The predicted sensory consequences of this motor plan get sent back to the auditory cortex, which compares its previous acoustic analysis with these predicted sensory consequences. If there is a mismatch, an error signal is sent from the auditory cortex back to the motor cortex for a corrected simulated motor plan (Iacoboni, 2008). This model suggests that the neurobiological underpinnings of speech perception are neither purely perceptual nor purely motor, but rather sensory-motor in nature. A key question is why a functional process that seems on the face of it purely perceptual, turns out to be sensory-motor in nature. What is the adaptive advantage of having such sensory-motor functional architecture? One strong possibility is that the interactive nature of language as it unfolds in the real world would benefit from such sensory-motor functional architecture. More evidence in support of this hypothesis comes from studies on human conversation.

16.6  Conversation, Imitation, Empathy, and Embodied Concepts There are two striking aspects of human conversation that are relevant to the adaptive advantage of a sensory-motor architecture of human language, at least as it unfolds in real life interactions. First, turn-taking during conversation is very fast (Gambi and Pickering, 2011). Second, speakers engaged in conversation demonstrate a pervasive imitative behavior (Garrod and Pickering, 2004). Let us look in detail at each one of these issues. How fast is turn taking? The gap between turns in conversation is about 200 microseconds (ms), which equals human performance in simple ‘go’ tasks, where there is no response selection and no signal discrimination (the type of task associated with the fastest response) (Levinson, 2016). Considering that the production of a simple syllable requires approximately 200 milliseconds and that human languages have very large vocabularies, such rapid turns are almost astonishing. A potential account for such fast turns may be that during conversation hearers predict the moment when the speaker will stop talking and plan to come in (Pickering and Clark, 2014; Pickering and Garrod, 2007). As we have seen for the case of speech perception, these predictive abilities are typical of neural and cognitive structures that plan actions, because skillful action planning does require predicting the sensory consequences of the motor plan. This hypothesis, however, generates some puzzling questions. For instance, this scenario suggests that cognitive and neural resources dedicated to language production are also heavily involved in language perception (as we have seen in the case of speech perception). How can they do both jobs at the same time without some detrimental effects on performance? Here is where and when mirroring and imitation may come in. If neural and cognitive structures mirror during conversation, their activity will code the same event during perception and production of language. Leveraging heavily on this feature, mirroring structures may get ‘primed’ in their activation during perception and be more readily accessible during production. If this—admittedly rather vague—account nears the truth, one would predict that 270



What Mirror Neurons Can Tell Us

a lot of imitation goes on during human conversation (because the mirroring cognitive and neural structures for, say, the word ‘sofa’, would be active during perception and this activation in turn would prime them to be active also for the production of the word). Indeed, evidence suggests that this may be the case. Humans tend to imitate each other while engaged in conversation, and they do so at multiple levels, even at the level of syntactic structures (Pickering and Garrod, 2013). One aspect of human imitation that is well known from social psychology studies is that imitation and empathy go together. For instance, it is well known that humans tend to imitate each other subconsciously during social interactions. This subconscious imitation, however, is related to conscious behavior. Subjects who tend to subconsciously imitate more during social interactions also tend to report to be highly empathic individuals. Furthermore, the amount of subconscious imitation that goes on during a social interaction predicts the amount of liking between the people involved in the interactions. The more imitation is going on between them, the more they like each other (Chartrand and Bargh, 1999). In studies of human conversation, the focus tends to be on the linguistic features that get imitated. However, a parsimonious account of neural and cognitive mechanisms of human interactions in which there are prominent similarities across seemingly different domains (as bodily postures, facial expressions and gestures with regard to bodily imitation, and linguistic features with regard to conversation), should attempt to identify common mechanisms at play. Obviously, mirroring is a plausible candidate in both neural and cognitive terms. The question is: is there something else that is shared between bodily imitation during social interactions and linguistic imitation during conversation? The issue of whether higher ­levels of linguistic imitation also correlate with higher levels of empathy and prosociality as for bodily imitation during social interactions is still open, since most studies on imitation during conversation have not focused on it. Yet, a shared aspect of bodily imitation and linguistic imitation may rest on the grounding of linguistically mediated concepts onto perceptual and motor experiences. Studies investigating the relationships between mirroring, bodily representations and language, have demonstrated somatotopically organized activity in premotor areas for action observation and linguistic representations of actions, supporting the hypothesis that linguistically mediated concepts are grounded in sensorymotor representations (Aziz-Zadeh et al., 2006). While recent, compelling evidence for embodied cognition goes well beyond the motor system, encompassing systems associated with multiple modalities (Fernandino et al., 2016), the association between motor and linguistic representation makes us see human social interaction through a different light. The parallel bodily and linguistic imitation that humans display while interacting includes another, hidden form of bodily imitation, the mimicry of motor representations through the imitation of bodily-based linguistic representations.

16.7 The Need for Implicit Control and the Study of Stimulus-Driven, Synchronized Brains The discovery of the phenomenon of mirroring has generated quite a bit of incredulous reactions. The phenomenon is indeed at first sight rather puzzling because even though—as we have seen above—there is plenty of often unnoticed mimicry going on between people engaged in social interactions and specifically in conversations, it is not that people imitate each other all the time and repeat what the other person said all the time! That level of parroting would be highly dysfunctional. At first sight, there is somewhat of an incongruence between the mechanism of mirroring and the absence of continuous imitation. Nevertheless, 271

Marco Iacoboni

this seeming incongruence is very easily explained by the existence of control mechanisms. What is emerging in recent studies, however, is that the type of control at play with mirroring is somewhat different from a traditional view of control. Indeed, this recent work may invite us to reconsider how we think about control. When we think about control, we tend to think about deliberate control. I am hearing you talking to me and I do not mindlessly repeat what you are saying. And yet, it does not feel we need to deliberately make an effort not to repeat what other people are telling us. Still, as briefly discussed earlier, we do tend to imitate quite a bit of what other people are telling us. If you talk to me about your living room and that old sofa that needs to be replaced, whenever I mention that thing that you want to replace, I am more likely to use the word ‘sofa’ than the word ‘couch’, even though they both signify the thing you want to replace in your living room. Even syntactical structures get imitated, even though they tend to be fairly abstract and complex. How can we account for this behavior? One way to think about it is to consider exactly the two main starting points of this discussion, which seem to be so much at odds with each other. On one hand we do have a neurobiological machinery that leads us to imitate others when we interact with them. On the other hand, continuous imitation of each other would be highly dysfunctional and needs to be modulated. These two opposite functional processes are most likely at play continuously. It is highly adaptive to change the control of imitation in a non-deliberate, implicit process. If we were to allocate resources, energies, and attention to the need not to imitate, we would do very little else. While this idea makes intuitive sense, it is challenging to provide evidence in support of such a concept. Perhaps the best empirical evidence comes from brain imaging studies that look at how activity across brains of subjects watching the same, long, continuous input (such as a movie) gets synchronized (Hasson et al., 2004). Synchronized brain activity among free viewing and listening subjects while watching a movie (or listening to a voice telling a story) must come from a stimulus-driven brain response (Wilson et al., 2008). Brain imaging data from subjects watching (and listening to) the video (or the audio only) of someone retelling them the plot of a Bugs Bunny cartoon show many areas of synchronized brain activity. Some areas are easily ‘explained away’ as processing the incoming sensory stimulus (both visual and auditory areas). Some other areas of synchronized activity across subjects are premotor areas and may reflect the covert verbalization of the speech sounds subjects are listening to, in line with the findings discussed above on this issue. However, other cortical areas of synchronized activity are quite surprising. These are very rostral areas in the prefrontal cortex, and areas in the medial prefrontal cortex. These areas can hardly be conceived as ‘sensory processing areas’. They are high level, ‘cognitive’ areas. How can they possibly show synchronized activity across subjects simply watching and listening to someone else telling them a story? If this synchronized activity does not reflect sensory processing, what is it? It is unlikely that it represents between subjects synchronization of high level semantic information, since that information can take many forms, hence resulting in lack of synchronization. But what if the need for an implicit, effortless control that mirroring requires shapes dedicated control systems for specific sensory information, so that whenever that information is processed at sensory and motor level, a dedicated, automatic control mechanisms is also triggered? This hypothesis may readily explain the inter-subjects synchronized activity of prefrontal areas while subjects are passively watching and listening to someone telling a story (Wilson et al., 2008). Importantly, large patches of inter-subject synchronization occur in the medial prefrontal cortex, an area that recent evidence suggests is critical for imitation control (Cross et al., 2013), even in fairly highly complex scenarios (Holbrook et al., 2016; Klucharev et al., 2011). 272



What Mirror Neurons Can Tell Us

16.8  Future Directions and Concluding Remarks The study of the sociobiological mechanisms of language is probably at the beginning of a surge of newly devised methods and techniques that attempt to investigate neural mechanisms at play during communicative interactions (Schippers et al., 2010). For decades neuroscience has been ill equipped to study complex human interactions. However, advances in devices and analytical techniques are making it more and more plausible to dare studying the neurobiological underpinnings of language as it unfolds in real-life conversations. For instance, a promising source of data comes also from neurosurgical patients who are chronically implanted for medical reasons, and with implants in their brains that can both record brain activity and stimulate brain regions. Furthermore, advances in virtual reality technology will make it more and more likely that experiments will be set up that mimic real-life interactions but that can also be controlled from the top down by the experimenters. When it comes to mirroring and its control, a better understanding of the dynamic interactions between the bottom-up mirroring processing and the top-down dedicated control will shed light on foundational mechanisms of the sociobiology of language.

References Aziz-Zadeh L, Wilson SM, Rizzolatti G, Iacoboni M (2006) Congruent embodied representations for visually presented actions and linguistic phrases describing actions. Current Biology 16:1818–23 Caggiano V, Fogassi L, Rizzolatti G, Thier P, Casile A (2009) Mirror neurons differentially encode the peripersonal and extrapersonal space of monkeys. Science 324:403–6 Chartrand TL, Bargh JA (1999) The chameleon effect: The perception-behavior link and social interaction. Journal of Personality and Social Psychology 76:893–910 Corballis MC (2010) Mirror neurons and the evolution of language. Brain and Language 112:25–35 Cross KA, Torrisi S, Losin EA, Iacoboni M (2013) Controlling automatic imitative tendencies: Interactions between mirror neuron and cognitive control systems. Neuroimage 83:493–504 di Pellegrino G, Fadiga L, Fogassi L, Gallese V, Rizzolatti G (1992) Understanding motor events: A neurophysiological study. Experimental Brain Research 91:176–80 Fadiga L, Craighero L, Buccino G, Rizzolatti G (2002) Speech listening specifically modulates the excitability of tongue muscles: A TMS study. European Journal of Neuroscience 15:399–402 Fadiga L, Fogassi L, Pavesi G, Rizzolatti G (1995) Motor facilitation during action observation: A magnetic stimulation study. Journal of Neurophysiology 73:2608–11 Fernandino L, Binder JR, Desai RH, Pendl SL, Humphries CJ, et al. (2016) Concept representation reflects multimodal abstraction: A framework for embodied semantics. Cerebral Cortex May; 26(5):2018–34 Ferrari PF, Gallese V, Rizzolatti G, Fogassi L (2003) Mirror neurons responding to the observation of ingestive and communicative mouth actions in the monkey ventral premotor cortex. European Journal of Neuroscience 17:1703–14 Gallese V, Fadiga L, Fogassi L, Rizzolatti G (1996) Action recognition in the premotor cortex. Brain 119:593 Gambi C, Pickering MJ (2011) A cognitive architecture for the coordination of utterances. Frontiers in Psychology 2:275 Garrod S, Pickering MJ (2004) Why is conversation so easy? Trends in Cognitive Sciences 8:8–11 Hasson U, Nir Y, Levy I, Fuhrmann G, Malach R (2004) Intersubject synchronization of cortical activity during natural vision. Science 303:1634–40 Holbrook C, Izuma K, Deblieck C, Fessler DM, Iacoboni M (2016) Neuromodulation of group prejudice and religious belief. SCAN Mar; 11(3):387–94 Iacoboni M (2008) The role of premotor cortex in speech perception: Evidence from fMRI and rTMS. Journal of Physiology Paris 102:31–4 273

Marco Iacoboni

Jeannerod M, Arbib MA, Rizzolatti G, Sakata H (1995) Grasping objects: The cortical mechanisms of visuomotor transformation. Trends in Neuroscience 18:314–20 Klucharev V, Munneke MA, Smidts A, Fernández G (2011) Downregulation of the posterior medial frontal cortex prevents social conformity. Journal of Neuroscience 31:11934–40 Kohler E, Keysers C, Umiltà MA, Fogassi L, Gallese V, Rizzolatti G (2002) Hearing sounds, understanding actions: Action representation in mirror neurons. Science 297:846–8 Levinson SC (2016) Turn-taking in human communication: Origins and implications for language processing. Trends in Cognitive Science 20:6–14 Liberman AM, Mattingly IG (1985) The motor theory of speech perception revised. Cognition 21:1–36 Meister IG, Wilson SM, Deblieck C, Wu AD, Iacoboni M (2007) The essential role of premotor cortex in speech perception. Current Biology 17:1692–6 Pickering MJ, Clark A (2014) Getting ahead: Forward models and their place in cognitive architecture. Trends Cogn Sci 18(9):451–6 Pickering MJ, Garrod S (2007) Do people use language production to make predictions during comprehension? Trends in Cognitive Science 11:105–10 Pickering MJ, Garrod S (2013) An integrated theory of language production and comprehension. Behavioral and Brain Sciences 36:329–47 Pulvermüller F, Fadiga L (2010) Active perception: Sensorimotor circuits as a cortical basis for language. Nature Reviews Neuroscience 11:351–60 Pulvermüller F, Huss M, Kherif F, Moscoso del Prado Martin F, Hauk O, Shtyrov Y (2006) Motor cortex maps articulatory features of speech sounds. PNAS 103:7865–70 Raos V, Umiltà MA, Murata A, Fogassi L, Gallese V (2006) Functional properties of grasping-related neurons in the ventral premotor area F5 of the macaque monkey. Journal of Neurophysiology 95:709–29 Rizzolatti G, Arbib MA (1998) Language within our grasp. Trends in Neuroscience 21:188–94 Rizzolatti G, Camarda R, Fogassi L, Gentilucci M, Luppino G, Matelli M (1988) Functional organization of inferior area 6 in the macaque monkey. II. Area F5 and the control of distal movements. Experimental Brain Research 71:491–507 Schippers MB, Roebroeck A, Renken R, Nanetti L, Keysers C (2010) Mapping the information flow from one brain to another during gestural communication. PNAS 107:9388–93 Umiltà MA, Kohler E, Gallese V, Fogassi L, Fadiga L, et al. (2001) I know what you are doing: A neurophysiological study. Neuron 31:155–65 Weigand, Edda (2002) Constitutive features of human dialogic interaction: Mirror neurons and what they tell us about human abilities. In Mirror Neurons and the Evolution of Brain and Language, Maxim I. Stamenov and Vittorio Gallese (eds), 229–48. Amsterdam: Benjamins. Wilson, Edward O (1975) Sociobiology: The New Synthesis. Cambridge, MA: Belknap Press of Harvard University Press. Wilson SM, Iacoboni M (2006) Neural responses to non-native phonemes varying in producibility: Evidence for the sensorimotor nature of speech perception. Neuroimage 33:316–25 Wilson SM, Molnar-Szakacs I, Iacoboni M (2008) Beyond superior temporal cortex: Intersubject correlations in narrative speech comprehension. Cerebral Cortex 18:230–42 Wilson SM, Saygin AP, Sereno MI, Iacoboni M (2004) Listening to speech activates motor areas involved in speech production. Nature Neuroscience 7:701–2 Wise SP, Boussaoud D, Johnson PB, Caminiti R (1997) Premotor and parietal cortex: Corticocortical connectivity and combinatorial computations. Annual Review of Neuroscience 20:25–42 Wise SP, di Pellegrino G, Boussaoud D (1996) The premotor cortex and nonstandard sensorimotor mapping. Canadian Journal of Physiology and Pharmacology 74:469–82

274

Part III

Components of Dialogic Interaction Human Nature, Institutions and Cultures

Part III deals with the issue of integration, which is central to a holistic theory of dialogic interaction. The complex whole of human action and behaviour is more than the sum of its parts. There are no parts at all, there are components which are derived from the complex whole by specialization. These components relate to human nature, the environment and culture. There is no autonomous level of verbal means but speaking, thinking, perceiving and having emotions are integrated and interact with each other in the dialogic action game. Marion Grein focuses on the integration of emotion and reason which demonstrates Descartes’ error and heralds the post-Cartesian thought. The integration of different human abilities has eventually been verified by the experimental results of neuroscience, to mention only Damasio’s studies (e.g. 1994). Human beings’ actions and behaviour are dependent on human nature and the coevolution of genes, mind and culture. Human beings’ nature as dialogic beings has been verified by mirror neurons as described by Iacoboni (Chapter 16). Human beings are, however, also individual beings who try to achieve their own interests. It is their double nature as dialogic individuals which needs to be balanced by strategies mediating between selfinterest and social concerns. Strategies are an essential component of human competencein-performance and derive from human beings’ nature as goal-oriented, persuasive beings. Jennifer Adams conceptualizes a dialogic rhetoric of effective language use. She focuses on implications for persuasion and argumentation and how they can be dealt with in the ‘Mixed Game Theory’. Human double nature is also at the heart of ethics and, in principle, characterizes any human action and behaviour which includes taking a position between the self and the other as ethical action and behaviour. Language thus becomes the ‘originative house of dialogic ethics’ in Ronald C. Arnett’s chapter, which sheds light on ethics as a component of human competence-in-performance. Human competence-in-performance is not only shaped by human nature but also depends on the environment. We do not live outside the environment, natural or cultural, we live in the environment. The environment is there from the very beginning and cannot adequately be dealt with by the addition of context. The environment includes the natural environment of inanimate surroundings but also the institutional environment created by human beings 275

Edda Weigand

themselves. Human beings are at the centre of this ‘indissoluble’ whole of nature and culture (Marchand, 2010). Sebastian Feller focuses on the component of ‘institutions’. True to principles of the ‘Mixed Game Model’, institutions can be derived from the complex whole by the necessity of ‘division of labour’. The law is a prominent example. Fritjof Haft demonstrates how jurisdiction developed and changed over the centuries as a dialogic institution. Finally, we must not forget that we are not confronted with one universal language but with multiple specific languages all over the world. We might assume that human beings are born with a universal ability of competence-in-performance from which specific languages have evolved and have been shaped by the influence of different cultural environments. Marion Grein deals with this issue by focusing on the question of how culture affects language and dialogue.

References Damasio A (1994) Descartes’ Error: Emotion, Reason, and the Human Brain. New York: Putnam. Marchand THJ (2010) Making knowledge: Explorations of the indissoluble relation between minds, bodies and environment. Journal of the Royal Anthropological Institute 16 (1): 1–21.

276

17 Emotion, Reason and the Human Brain Marion Grein

17.1 Introduction Language and emotion are fundamental to human nature. The relationship between language, cognition and emotion is of major importance if we wish to grasp that nature. Emotions are decisive in our thinking and acting, and determine our interpretation and evaluation of interactants and situations in any communicative setting. In any conversation, debate, commercial or entertainment program on TV, we find emotions or emotional strategies – most of them conveyed through verbal (and non- and para-verbal) means. In linguistics, the study of emotion was neglected for many centuries because mind (cognition) and the body were considered to be separate units (so-called ‘Cartesian dualism’). Whereas early brain studies found evidence that cognition and emotion were autonomous units, it has now been proven that there is a complex interaction between cognition and emotion (cf. Schwarz-Friesel, 2013: 8). This chapter, which actually carries the same title as the last part of Damasio’s famous book Descartes’ Error: Emotion, Reason and the Human Brain, will emphasize that emotion, cognition (or reason) are integrally rooted in the brain. In fact, the publication of Damasio’s book was the starting point for the ‘emotional turn’ in language studies. Neurology has proven that emotion and reason do not represent separate fields but interact with each other, together with further human abilities (Damasio, 2004; Pinker, 1997: 370). In fact, reasoning requires emotions. Thus, Descartes who (seemingly) separated mind from body, and reason from emotion, is proven to be wrong. A strict separation of cognition and emotion is impossible. According to an enduring and still often dominant view, language is the product of an innate, universal, domain-specific and encapsulated module (Chomsky, 1980; Fodor, 1983; Pinker, 1994). In recent years, however, and when taking a closer look at the role of emotions, the idea of a single underlying linguistic system differing only in surface realization seems increasingly unlikely. Yet, still many linguistic approaches do not integrate emotion into their language analysis and stick to the investigation of syntactic structures.

277

Marion Grein

17.2  Historical Research (Seventeenth and Eighteenth Century) Emotion studies began as a field of study in philosophy. The term ‘emotion’ itself entered the field in the nineteenth century. Seventeenth-century philosophers talked about ‘passion’ and ‘affect’, while eighteenth-century academics preferred to write about ‘sentiments’. In the very beginning, passion, affect and sentiments were considered to be something negative, as something that had to be severely suppressed. Early modern philosophy operated with strictly opposed fields, dividing mind and body, perception and will, reason and desire, nature and convention. Early sources were Aristotle, Cicero, Seneca, Aquinas and Scholastic approaches, followed by approaches of the humanists Machiavelli and Montaigne (cf. Schmitter, 2014). Descartes and his Passion of the Soul (1649) used the term ‘passion’ and in the beginning considered passions as receptive, passive states and a species of perception. Pascal in his Pensées (1670) introduced the terms ‘feelings’ and ‘sentiments’. In some philosophical treatises, the term ‘sentiment’ was used interchangeably with ‘passion’, in others as a contrast to ‘passion’. Seventeenth-century philosophers continued by setting up long lists of emotions, for example, Hobbes offered a list of about thirty in the Leviathan (1651). Descartes distinguished only six ‘basic passions’ (basic or primitive passions are: wonder, which he calls “the first of all the passions”, love and hatred, desire, joy and sadness). Like many of his contemporaries, Descartes tried to systemize his passions, initially distinguishing between movements coming from the body and those originating in the soul. The distinction was soon neglected and the simple categorization of good and bad emotions was taken as a basis (affective quality of emotions, including neutral emotion). Passions were then organized into simple and complex forms, the simple forms based on their evaluative character. One of the major concerns of the philosophical debate was where to locate emotions. Following Descartes, emotions were subsumed as a part of our perception, the perception, however, belonging to human intellect and thus being a basis for judgments (see Section 17.3). Hume, and eighteenth-century theorists, identified emotions specifically with sense-perceptions, maintaining that a distinctive feature of ‘passion’ or ‘sentiment’ is that they touch the mind more forcibly than other perceptions. Another important – yet often neglected subject – was the physiology of emotions and their implications for medicine. Descartes considered emotions central to the treatment of mental and bodily illness. Emotions were generally considered to be motivating, and, thus, Descartes called for the managing and cultivation of emotions. Furthermore, emotions were discussed in political and social theory (cf. Kahn et al., 2006). So-called ‘how-to’ books with techniques for governing others by managing their emotions became popular (political philosophy), just as the role of emotions in social order maintenance was investigated. Descartes actually maintained that emotions could facilitate social interaction in his analysis of ‘love’ and ‘generosity’. Malebranche (1997) went even further, arguing that the communication of emotions is crucial to social organization. Further, eighteenth-century philosophers differentiated between self- or other-directed emotions (cf. e.g. Edmund Burke, Hume, Rousseau).

17.3  Descartes’ Error – Descartes View of the Brain So, what was Descartes’ actual error? Damasio (1994: 249) cites Descartes: From that I knew that I was a substance, the whole essence or nature of which is to think, and that for its existence there is no need of any place, nor does it depend on any 278



Emotion, Reason and the Human Brain

material thing; so that this ‘me,’ that is to say, the soul by which I am what I am, is entirely distinct from body, and is even more easy to know than is the latter; and even if body were not, the soul would not cease to be what it is (Descartes, 1649, The Passions of the Soul). and comments: “This is Descartes’ error: the abyssal separation between body and mind.” In fact, Descartes began discussing the so-called ‘pineal gland’, and the separation of body and soul, in his Treatise of Man (written before 1637, but published posthumously, first in a Latin translation in 1662, and then in the original French in 1664). First mention of the pineal gland is to be found in his Dioptrics (1637). It contained the thesis that “a certain small gland in the middle of the ventricles” is the seat of the sensus communis, the general faculty of sense (cf. Cottingham et al., 1985, I: 100).1 In The Treatise of Man, Descartes described the conceptual model of a human being, created by God, consisting of a body and a soul. He intended to elaborate on the two ‘parts’ of a human being; yet, he focused on the body, hardly providing any information on the soul and its interaction with the body. The bodies of Descartes’ hypothetical humans are presented as machines. I suppose the body to be nothing but a statue or machine made of earth, which God forms with the explicit intention of making it as much as possible like us. (Cottingham et al., 1985, I: 99) Indeed, the pineal gland played a major role in Descartes approach. The pineal gland is a very small endocrine gland in the vertebrate brain. Problematic, however, was that the basic anatomical and physiological assumptions of Descartes were completely mistaken. First, Descartes believed, fallaciously, that the pineal gland was suspended in the middle of the ventricles. Second, Descartes believed the pineal gland to be full of “animal spirits”. Third, Descartes described these “animal spirits” as “a very fine wind, or rather a very lively and pure flame” (Cottingham et al., 1985, I: 100). The pineal gland was his crucial notion for explaining perception. The gland was understood as the seat of perception, imagination and common sense. The nerves, he believed, were filled with “animal spirits” and contained small fibers which connected the sense organs to the brain. These were the basis for ideas. Descartes explained: And note that I say ‘imagines or perceives by the senses’. For I wish to apply the term ‘idea’ generally to all the impressions which the spirits can receive as they leave gland H. These are to be attributed to the ‘common’ sense when they depend on the presence of objects; but they may also proceed from many other causes (as I shall explain later), and they should then be attributed to the imagination. (Cottingham et al., 1985, I: 106) Descartes even explained the learning process or rather the function of memory. The gaps between the fibers of the substance of the brain become wider as a result of the flow of the animal spirits through them. This, he believed, changed the pattern in which the spirits would continue to flow through the brain and thereby figures may be “preserved in such a way that the ideas which were previously on the gland can be formed again long afterwards without requiring the presence of the objects to which they correspond. And this is what memory consists in” (Cottingham et al., 1985: I: 107). 279

Marion Grein

In 1640, Descartes wrote several letters to answer various questions of several people and added further explanations. My view is that this gland is the principal seat of the soul, and the place in which all our thoughts are formed. The reason I believe this is that I cannot find any part of the brain, except this, which is not double. Since we see only one thing with two eyes, and hear only one voice with two ears, and in short have never more than one thought at a time, it must necessarily be the case that the impressions which enter by the two eyes or by the two ears, and so on, unite with each other in some part of the body before being considered by the soul. Now it is impossible to find any such place in the whole head except this gland; moreover, it is situated in the most suitable possible place for this purpose, in the middle of all the concavities; and it is supported and surrounded by the little branches of the carotid arteries which bring the spirits into the brain. (Cottingham et al., 1991: 143) And in December 1640 he added: Since it is the only solid part in the whole brain which is single, it must necessarily be the seat of the common sense, i.e., of thought, and consequently of the soul; for one cannot be separated from the other. The only alternative is to say that the soul is not joined immediately to any solid part of the body, but only to the animal spirits which are in its concavities, and which enter it and leave it continually like the water of river. That would certainly be thought too absurd. (Cottingham et al., 1991: 162) In his The Passions of the Soul (1649) – a continuation of Treatise of Man – Descartes then sets off by dividing human beings into a body and a soul. Let us again take a closer look at his explanation and his criteria for determining which functions belong to the body and which to the soul. Anything we experience as being in us, and which we see can also exist in wholly inanimate bodies, must be attributed only to our body. On the other hand, anything in us which we cannot conceive in any way as capable of belonging to a body must be attributed to our soul. Thus, because we have no conception of the body as thinking in any way at all, we have reason to believe that every kind of thought present in us belongs to the soul. And since we do not doubt that there are inanimate bodies which can move in as many different ways as our bodies, if not more, and which have as much heat or more […], we must believe that all the heat and all the movements present in us, in so far as they do not depend on thought, belong solely to the body. (Cottingham et al., 1985, I: 329) According to Descartes then, the part of the body in which the soul functions is not the heart, but the pineal gland. Thus, body and soul – according to Descartes – are linked in the human brain. To Descartes, the soul was not the principle of life, but the principle of thought. Actually, even though wrong, his explanations of the retrieval of data from memory are not as far off as often suggested. He explained: Thus, when the soul wants to remember something, this volition makes the gland lean first to one side and then to another, thus driving the spirits towards different regions 280



Emotion, Reason and the Human Brain

of the brain until they come upon the one containing traces left by the object we want to remember. These traces consist simply in the fact that the pores of the brain through which the spirits previously made their way owing to the presence of this object have thereby become more apt than the others to be opened in the same way when the spirits again flow towards them. And so the spirits enter into these pores more easily when they come upon them, thereby producing in the gland that special movement which represents the same object to the soul, and makes it recognize the object as the one it wanted to remember. (Cottingham et al., 1985, I: 343) Emotions (or passions, according to Descartes) are also explained with the help of the pineal gland: “the ultimate and most proximate cause of the passions of the soul is simply the agitation by which the spirits move the little gland in the middle of the brain” (Cottingham et al., 1985, I: 349). Taking a closer look at his original writing, we can actually see that his views can and were interpreted quite differently. The established and often quoted separation of body and soul (termed as the Cartesian thesis or Cartesian dualism) is not as clear-cut as often anticipated.

17.4  Damasio’s Approach (1994) – Emotions and Reason in the Brain The starting point of Damasio’s approach was the aforementioned ‘Cartesian dualism’, which he assigned to Descartes even though – as hopefully made clear – this was just one possible interpretation of Descartes’ approach. The book, however, was an important milestone against any Cartesian dualism, making it evident that body and soul are by no means separate parts of the human being. Damasio (1994: 87) informs us that “the brain and the body are indissociably integrated by mutually targeted biochemical and neural circuits”. In fact, serious accidents with damage in various brain regions (e.g. neurological patients with brain lesions) gave the impetus for realizing that reason and emotions are deeply localized in our brain. Damages within the prefrontal cortex, for instance, impaired the patient’s abilities to plan and to make decisions (cf. Damasio, 1994: Sections 17.1 and 17.2 on Phineas Gage). It was found that damage to the brain’s prefrontal cortex prevented any decision making due to a lack of the necessary emotional machinery (cf. Damasio, 1994: XIII). Damasio (2004: 61) explains “if the ventromedial sector is included in the lesion, bilateral damage to prefrontal cortices is consistently associated with impairments of reasoning/decision making and emotion/feeling”. Thus, reason and emotion are inseparable. The prefrontal cortex sits at the front of the brain. Signals from all the brain’s sensory regions go to the prefrontal cortex. Furthermore, the brain’s bio-regulatory areas send signals to the prefrontal cortex, such as the areas that regulate the neurotransmitters, serotonin (stress) and dopamine (happiness) (cf. Grein, 2013a: 24). The prefrontal cortex, next to the limbic system, is involved in classifying experiences as either “relevant and positive” or “irrelevant and negative”. The prefrontal cortex is ideally constructed for reasoning about personal situations, organizing our thoughts and is in charge of our self-control. In the prefrontal cortex, the so-called ‘somatic markers’ are produced.

17.4.1  The Somatic Marker Hypothesis ‘Somatic’ stems from the Greek word soma (body). All emotions are ‘bodily or rather have a ‘bodily state’. Somatic markers are a necessary criterion for making a decision, directing 281

Marion Grein

attention to alternatives and biasing the brain towards one of the alternative decisions. Again, this proves that a strict separation between emotion and reason is impossible. Damasio (1994: XIII) sums up: “Emotion, feeling, and biological regulation all play a role in human reason.” Further lesion studies have shown that impairments in other regions than the prefrontal cortex also have an effect on reason and emotions simultaneously (cf. Damasio, 1994: 61–79 on anosognosia and the somatosensory system). In fact, in current neuro-didactics, one focus is on the limbic system (amygdala): every piece of information (stimulus) is evaluated by the limbic system (cf. Grein, 2013a: 10–13). Further emphasis is put on the neurotransmitters (cf. Grein, 2013a: 23–25): specific neurotransmitters motivate learners, enforce their creativity and increase their learning capability (e.g. dopamine), whereas other neurotransmitters (e.g. serotonin) make them depressive and prevent effective learning. Knowledge and experiences are neuronal connections in the cortex, activated when a new stimulus is perceived in the sensory cortices. The evaluation process includes the activation of the emotional as well as the declarative or knowledgebased ‘stored data’ (cf. Grein, 2013a: 10). Currently, there is a huge number of functional magnetic resonance imaging (fMRI) studies on emotion and cognition. Most of the studies report empirical findings that bridge the divide between mind and brain, showing high correlations between measures of individual differences relating to individuality, emotion and cognition and brain activity. Even though Vul et al. (2008) found that some of these studies used defective research methods, the direct correlation is still observable. Davidson and Sutton (1995) were among the first to investigate the neural basis of emotions and introduced the field of so-called ‘affective neurosciences’. Since then, especially the emotion of fear has been studied in detail. It has been proven that the amygdala is a key neural pathway to fear (cf. LeDoux, 1996, 2001). The current state-of-the-art consensus is: emotions are taken to be crucial for any decision making. Furthermore, it has been proven that irregularities in the functioning of specific brain regions are responsible for emotional disorders, including social phobias (cf. Reiman, 1997) and depression (cf. Drevets et al., 1997). In all cases, decision making is strongly affected. Schultz (2000), taking a look at these brain regions and reward processing, indicated that neurons adapt their activity according to ongoing emotional feelings and experiences. The prefrontal cortex is considered to be the area that plays a major function in the experience of emotions. Based on these findings, Damasio (1994) developed the aforementioned ‘somatic marker hypothesis’. In 2011, Damasio published ‘Neural basis of emotions’, arguing that emotions are “complex programs of actions triggered by the presence of certain stimuli, external to the body or from within the body, when such stimuli activate certain neural systems” (Damasio, 2011: 1804). Here, Damasio argues that “emotions” and “feelings of emotions” are “distinct aspects of a functional sequence that begins when an object or situation triggers a specific behaviour – the emotion – which is followed rapidly by the perception of the changes related to the behaviour – the feeling of emotion.” The range of emotions is finite. According to Damasio, human beings have different programs of emotions. Basic (or primary) emotions being fear, disgust, sadness, joy, anger and surprise, alongside simpler programs such as enthusiasm or discouragement (known as background emotions) and complex programs, often called “social emotions”, such as embarrassment, shame, guilt, contempt, compassion and admiration (cf. Damasio, 2011: 1804). The basic emotions can often – depending on cultural principles of displaying or hiding emotions – be observed by explicit facial expressions. Background emotions (well-being, calmness, tension, discouragement etc.), however, can only be detected “by subtle details of body posture, speed and 282



Emotion, Reason and the Human Brain

contour of movement, minimal changes in the amount and speed of eye movements, and in the degree of contraction of facial muscles” (Damasio, 1999: 52). In a nutshell, emotions are thus, biologically determined processes, that is, the organism’s reaction to conscious feelings that are to some degree influenced by culture. As mentioned, the triggering requires a stimulus initiating the execution of the emotion which is then followed by the respective feeling state. As an example, Damasio describes the well-researched emotion of fear: the amygdala receives signals (auditive, visual, tactile, and olfactory). When those signals are emotionally-competent, they activate nuclei in the hypothalamus and in the periaqueductal grey (PAG). Those brain sites execute the basic emotional actions – release of cortisol, adjustment of heart rate, respiratory rate, degree of gut contraction and specific behavior such as facial expressions and posture. The interaction of these “actions” constitutes the emotional state of fear. In 2014, Jack et al. discovered that there might only be four instead of six basic emotions (happiness, surprise, fear, disgust, anger and sadness). Anger, disgust, fear and surprise can hardly be distinguished and they conclude that these emotions might not be basic ones, but emotions that developed for social reasons. Biologically they, therefore, distinguish only four basic emotions: happy, sad, afraid/surprised, and angry/disgusted. Irrespective of whether there are four or six basic emotions, language analysis needs to take the human being into account when analyzing verbal data. In Section 17.5, I will present different approaches to language analysis and emotions, leading to the final holistic approach of Edda Weigand.

17.5  Language Analysis and Emotions According to Lewis and Haviland-Jones (2004: 19) emotions as a research area began in the 1980s (cf. Schwarz-Friesel, 2013: 12). A probable reason for the delay of scientific research on emotions, next to the predominance of the system-theoretical approach, could be that the term emotion itself is not sufficiently defined. There are at least four different meanings: emotions can either be understood as subjective feelings (how do I feel when I am sad), expressive behavior (outward signs like facial expressions and body language), physical reactions (e.g. crying, trembling, increase of heart rate2) (cf. Izard, 1981), and verbal representations (on word, sentence, text level). Emotions are understood as states, reactions, representations, processes and categories of experience, and researchers either argue in favor of an innate or acquired status of emotions (cf. Schwarz-Friesel, 2013: 56). In fact, emotions seem to be both, innate and acquired or learned. Wilce (2009: 3) argues that “nearly every dimension of every language at least potentially encodes emotion”. On the one hand, emotions can directly be encoded with linguistic means (“I feel sad”; talking about emotions or emotional experiences; emotion talk), on the other, language-use can be seen as performed “emotive”, implying that every utterance, or rather the speaker, carries emotions, which are displayed verbally or non- and/or paraverbally – often unintentionally (emotional talk). In Section 17.5.1, I will give a short overview of different fields of study, which often merge.

17.5.1  The Linguistic Fields of Analysis 17.5.1.1  Emotions and Prosody (Speech Emotion Analysis)

In this field of study, vocal behavior is analyzed as a marker of emotion or affect. Interactants raise their voice or lengthen vowels to signal a particular emotion or emotional intensity 283

Marion Grein

(cf. Goodwin and Goodwin, 2001; Wilce, 2009). Voice qualities indicate emotional states (cf. Gobl and Ní Chasaide, 2003). The basic assumption is the existence of objectively measurable voice parameters (like respiration, pitch) that can reflect the affective or emotional state of a speaker. 17.5.1.2  Emotions and Words (Lexical Models)

Studies focusing on the lexicon and emotions are most salient. All languages seem to have literal and figurative expressions for emotions. Some studies focus on words denoting or referring to emotions. Most notably Wierzbicka (cf. Enfield and Wierzbicka, 2002; Harkins and Wierzbicka, 2001; Wierzbicka, 1996, 1999) and her search for a natural semantic metalanguage consisting of semantic primitives (say, think, know, want, feel, good, bad) as “innate human concepts” (Wierzbicka, 1996: 16) has to be mentioned here. Within her meta-language concept, we find both, studies focusing on specific words and studies comparing whole lexicons (cf. Toivonen et al., 2012; for criticism of the concept see Weigand, 1996: 164). Within these emotional lexicon studies, the use of concrete emotions (e.g. angry, happy, sad – nouns, verbs, adjectives) is analyzed. Several studies focus on interjections (e.g. wow, oh, oops), which express emotional or mental states. Interjections vary cross-culturally (cf. Wierzbicka, 1992). Japanese has many ideophones (giseigo, gitaigo, sound symbolism), which often express inner states (odo means being shy and timid) and are analyzed as a means of expressing emotions (cf. Oda, 2000; Imai et al., 2008). In German, especially, modal particles are analyzed. Kövecses (2003) assumes that words and phrases (for vocabulary as a set of phrases cf. Weigand, 1996, 1998a and b) do not exist in isolated entities but form supra-lexical structures, and focusses on the analysis of metaphors and further figurative expressions (e.g. hit the roof, to be on cloud nine [cloud seven in German!]). His approach is mostly cited as “the lexical approach”, introducing a more empirical element in analyzing these conventionalized expressions. Data are either elicited or collected from lexicographic sources. His basic idea is that “language, particularly its lexicon, is a reflection of our conceptual system” (Kövecses, 1990: 41). By analyzing conventionalized expressions (i.e. metaphors, metonymies, idioms, proverbs and collocations), he intends to come to a description of the complex structure of emotion concepts. Alongside analyzing words that overtly mark emotions, researchers also have taken a closer look at connotations (cf. Schwarz-Friesel, 2013: 164–172). In recent years, especially within the field of metaphor research, studies have taken corpus-based approaches (cf. Stefanowitsch, 2006; Stefanowitsch and Gries, 2006). Corpus analysis yields data in the form of co-occurrence, usually using large corpora, like the Corpus of Contemporary American English, the DWDS (Digital Lexicon of German in the 20th Century) or the Collins Cobuild Dictionary. Few studies have yet examined emotional language use at dialogue level (‘emotion as discourse’; cf. Lewis and Haviland-Jones, 2004: 39). Here, mostly debates, disputes, discussions and consulting meetings are analyzed. Lüdtke (2006: 20) summarizes the areas of emotion research taken into account: • • • • 284

Phonetics: sound symbolism, alliteration, vowel lengthening Morphology: diminutives, prefix reinforcement (i.e. super-) Lexicon: metaphors pejoratives Pragmatics: use of sociolect, irony



Emotion, Reason and the Human Brain

17.5.1.3  Emotions and Nonverbal Means

Next to verbal means, nonverbal and paralinguistic mechanisms (facial expressions, gesture, posture, voice, etc.) are analyzed (cf. Ekman and Davidson, 1994). Yet, most of these studies do not give us insights that are helpful for language analysis. Seeing someone cry, tells us that he or she is sad. Combined with his or her facial expression and posture, we might assume how sad the person might be. Yet, this seems not to be of primary importance for a linguistic analysis, especially as the display of emotions is very individual. A study conducted with Dutch and Japanese, for instance, showed that the Japanese rather seek for information about the emotional state of interactants by focusing on prosody, while Dutch considered facial expressions more (cf. Tanaka et al., 2010). 17.5.1.4  Emotions and Morpho-Syntax

Some researchers focus on morpho-syntactical analysis, like diminutive morphemes or exclamative clauses (cf. Fries, 1996; Foolen, 1997). Some languages have bound morphemes, mostly suffixes or clitics, which additionally convey a range of meanings, including sympathy, endearment, intimacy, and so forth. In Japanese, clitics convey meaning, like no as a sentence final downtoner or yo to put emphasis on the speaker’s belief in the utterance. In some languages (like Kham or Magar), surprise is coded grammatically (‘mirative’) (cf. DeLancey, 2012: 535, 547–548). Again in Japanese, there are grammatical restrictions concerning the verbalization of emotional states: only the speaker’s psychological state (feeling, desires) can be reported directly. I can say “watashi wa kanashii desu” (I am sad) but it is not grammatical to say “kodomo wa kanashii desu” (The children are sad) (cf. Hasegawa and Hirose, 2005). There are different approaches to the study of emotions, which have already been mentioned in the previous paragraphs. However, for reasons of clarity they will be recalled in Section 17.5.2.

17.5.2 Major Approaches to Emotion Studies 17.5.2.1  The Universalism–Relativism Dispute (in Emotion Studies)

Many researchers, like Damasio, adhere to the evolutionary approach, which mainly derives from neuroscience and psychology. They argue – as we have seen in Section 17.4 – that emotions are biologically determined processes grounded in the brain, using the body, especially the face, as their stage. By adopting the distinction between basic or primary, secondary or psychosocial, and background emotions, they argue that the primary emotions are innate and, consequently, universal. Approaches that are mostly grounded in anthropology and psychology adopt a social or rather cultural perspective and argue that emotions are culture specific. Emotions are therefore not considered innate, but acquired and learned during the socialization process. Evolutionary approaches mostly study emotion not directly in relation to language use, for example, Joseph LeDoux explores the neuronal system of fear and Paul Ekman investigates facial expressions. And, actually, their approaches do not refute cultural diversity. Ekman (1980: 87–89; 1984: 320) talks about universal facial expressions and acquired bodily movements which are culture specific and a product of cultural learning. And it was Ekman who introduced the term “display rules” to show how social conventions interfere with the expression of basic emotions. The cultural factor is not denied, only the focus 285

Marion Grein

remains on what happens in the brain. Damasio (1999: 49) argues “we can educate our emotions but not suppress them entirely”. Comparative lexical approaches here are interested in ‘emotion terms’ that seem to have no equivalent in other languages, like the Japanese amae (a special type of dependency or a childlike desire for indulgence) or the German schadenfreude (malicious pleasure at an outgroup’s misfortune). Cross-cultural approaches are also interested in different nonverbal means conveying emotions. Belgians for instance, when expressing condolence and mourning at a funeral procession, clap their hands (like applauding), something very strange for many other cultures (cf. Schwarz-Friesel, 2013: 60). There are a large number of scientific publications focusing on innate versus learned and, thus, culture-specific emotions and display of emotions (cf. Wierzbicka, 1986, 2009). 17.5.2.2  Emotion Talk vs. Emotional Talk

As mentioned, studies can either focus on how language is used when we talk about emotions (emotion talk, use of emotion words) or on the means speakers use to display emotions, that is, the choice of words, metaphors, voice quality or raise of voice (emotional talk; cf. Wilson and Lewandowska-Tomaszczyk, 2012). Any content or issue can be expressed in different ways (cf. Schwarz-Friesel, 2013: 31–32): The functionally more or less equivalent statement “the old man died” can be emotionally encoded in various forms, like “he finally passed away, he kicked the bucket, he deceased, he is gone to be with the Lord, he is worm food now, he bit the dust, he hopped the twig” and so on. As the examples show, metaphors (euphemism and dysphemism) are a good demonstration of this. Again, prosody makes a difference, depending on the accentuation or stress “he finally passed away” can display different emotions. 17.5.2.3.  Emotion as Event Schema

Schema models represent emotion as a process, embedded in social and psychological scenarios (Lewis and Haviland-Jones, 2004: 38). In the most simplified form, emotion schemas consist of: event ® feeling ® thought/action response, a schema used by Wierzbicka, Lakoff and Kövecses. Emotional schemas are internal templates from which emotional experiences are processed. They are developed from early childhood when the child learns to interpret and regulate emotions through interaction with parents and mostly other close family members. Emotion schemas, then, are based on numerous experiences. They receive further elaboration throughout a person’s life span (cf. Bucci, 1997). An emotion schema could be conceived as an emotion interacting dynamically with perceptual and cognitive processes, which then influence the human being’s behavior. Most studies take a closer look at the culture-specific differences of these event schemes. 17.5.2.4  Studies on Cultural Display Rules

The expressions of emotions vary depending on social and cultural identity in accordance to the afore mentioned display rules (Ekman and Friesen, 1975). Display rules are defined as culturally shared norms about when, where, to whom and how to express emotions. “They are assessed as cognitive representations that guide individual’s expressive behavior” (Safdar et al., 2009: 1). The study conducted by Safdar et al. (2009), for instance, compared US Americans, Canadians and Japanese, as examples of individualistic versus collectivistic 286



Emotion, Reason and the Human Brain

cultures. In fact, they integrated gender and in- and out-group differences into their study and could prove that display rules (concerning powerful emotions) were culture-specific (cf. also Koopmann-Holm and Matsumoto, 2011 on display rules comparing Americans and Germans). In fact, there is a vast amount of studies focusing on display rules. 17.5.2.5  Emotion Studies and Bilingual/Bi-Cultural Speakers

Research within this field compares emotion concepts within two native languages (cf. Chin and Wigglesworth, 2007: 3). Wierzbicka (2004) emphasizes that the perspective on bilingual’s perspective on emotions enriches both, research on multilingualism and on emotions. Differences are found between ‘balanced’ and ‘late’ bilinguals (cf. Pavlenko, 2005). Perunovic and Perunovic (2012) validate that bicultural individuals have two sets of emotion concepts. Depending on the language they use, their thought, behavior and emotional experience is congruent with this specific culture. When they switch to their other language contradictory thoughts and emotion concepts can occur. Instead of one culture replacing the other, both cultures are retained and each is activated when communicating in the corresponding language (cf. Grein, 2013b). 17.5.2.6  Emotion and Gender Studies

The common assumption is that women are more emotionally expressive than men (cf. Kring and Gordon, 1998). Emotions of happiness, sadness and fear are mostly linked to women, whereas men are linked to anger (cf. Kelly and Hutson-Comeaux, 1999). Differences are partly innate, yet, mostly learned or acquired during the socialization process. Display of “real” emotions and those just simulated can be differentiated by analyzing facial muscles and activity patterns in the brain (cf. Damasio, 2010: 205). Analyses are conducted in all fields of emotion studies. Parkins (2012), for instance, takes a look at prosodic features, while Koch et al. (2007) apply brain studies, using an fMRI approach. 17.5.2.7  Quantitative vs. Qualitative Studies

Depending on the research question, numerous qualitative and quantitative research methods can be found: observations, introspection, interview studies, questionnaires, corpus analysis (face-to-face, TV, social networks, etc.), recordings, and measurements of processes in the brain. As Schwarz-Friesel (2013: 15) says, we are still, regardless of the approach, in need of an integrative theory which will incorporate and interrelate all components of language and links language, cognition and emotion. She (2013: 41) adds that when human beings communicate, they always draw on world knowledge and previous experiences, stored as conceptual schemes in their brains. These are always stored with specific emotions. This leads to Edda Weigand’s approach of the so-called ‘Mixed Game Model’ (MGM).

17.6  The Holistic Model MGM By Edda Weigand Weigand’s approach is based on Damasio’s (1994, 1999) arguments and brain research showing that emotions are biologically determined processes. Emotions are strongly connected to all mental capabilities. As depicted in Chapter 11, Weigand’s focus is on human beings acting and reacting in the ever-changing world on the basis of their perception, cognition, emotions as reflected in language-use and as described in the MGM (2010). The key concept is human beings’ 287

Marion Grein

nature as goal-oriented beings and as social individuals who have to negotiate their mutual purposes and interests. This innate drive is influenced by human beings’ social and cultural surroundings. Thus, all social, individual and cultural variables that have the power to influence human actions have to be included when investigating communicative behavior. Whenever two human beings communicate (or interact), the two interlocutors have their very own perceptions, cognitions, individual experiences, language skills, social and cultural imprints and emotions. Dialogic action, thus, is inevitably affected by emotions, which – depending on culture – are either displayed or hidden (to some extent). Weigand (2004: 9) differentiates between conditions of emotion (including the specific situation) and emotions as an ability. Human beings’ perception and reasoning are preset by their abilities, one ability being that we have emotions. Meaning, then, is structured and understood through the filter of our abilities. The following human abilities are assumed (Weigand 2004: 9): 1. 2. 3. 4. 5. 6. 7. 8. 9.

Consciousness or relating-the-object-to-the-self Perceiving through the five senses Knowing the object Thinking Moving Acting Speaking Feeling/emotions Learning social behavior.

In sum, emotions are internal states based on needs (including survival needs) in relation to the world and expectations in relation to other human beings and the world as a whole (cf. Weigand, 2004: 10). Furthermore, emotions themselves are an indeterminate concept that means that they are moderately describable, yet not accurately definable. Additionally, emotions are invariably present and integrated into the other mentioned abilities. They are not totally controllable by reason. And, as mentioned, they are in part innate and universal, in part shaped by previous social or cultural experiences, values and norms. Refined and sophisticated concepts of emotion, thus, have to interconnect various facets, like biology, sociology, bodily expressions. Thus, having emotions and talking about emotions are two different things. How we talk about emotions will not necessarily give us essential insights. Instead, we have to analyze how emotions are expressed by human beings within the ‘Dialogic Action Game’. Any speaker always carries emotions when he or she speaks (or nods, or raises his or her eye brows etc.). Yet, how, when and where does the emotion enter the utterance (verbal and nonverbal)? Is the emotion explicitly or implicitly expressed? Can a word alone convey an insight into the emotional state? Is it the phrase (cf. Weigand, 1996, 1998a)? The whole utterance, the speech act? (for several examples and a detailed analysis see Weigand, 2004). Weigand (2004) differentiates two roles of emotions. On the one hand, emotions are part of any communicative act: (1) You didn’t bring down the garbage again! In (1), a speech act of reproach, we have an accompanying emotional component of anger. 288



Emotion, Reason and the Human Brain

Next to this, three speech act types refer to emotions (for a detailed presentation of the different speech act types see Weigand, 2004: 16–17). In Dialogue: The Mixed Game (2010) Weigand renamed expressives to emotives, the term I will use here. We can explicitly state the emotion we feel in constatives: (2) I don’t feel well today. We can emphasize an emotion in emotives: (3) What a wonderful view! And we can create social relationship by declaring emotions in declaratives (must be differentiated from statements): (4) I love you! (5) I regret it (at court) vs. statement in everyday talk. In general, it is the grammatical structure and the intonation that constitutes the action. Besides these three basic speech acts for declaring, stating and emphasizing an emotion, there are further clues, often represented by a single word, so called ‘accompanying features’, again often dependent on the intonation. (6) You are playing the piano again. Depending on the intonation, the speaker’s emotion could be positive (joy ® encouragement) or negative (anger ® accusation). For analysis, examples (1) and (6) are of major concern, showing the role of emotions in many or rather every communicative action game. In general, emotions and reason are interrelated in dialogue. In the Mixed Game Model, principles of emotion, therefore, belong to regulative principles, which regulate reason and emotion (Weigand, 2010: 105ff.). In accordance with Weigand (2009: 172) three situations are to be found: 1. Emotions are completely controlled by rationality / reason. 2. Emotions dominate rationality. 3. Emotions and rationality are in conflict.

Depending on the culture and culture-based display rules, we keep our emotions under control, trying to act rationally and hiding our emotions. The Japanese, for instance, hardly ever display their emotions to people out of their close in-group (maxim ‘hide emotions in the public sphere’), whereas Italians display much more of their emotions, even to strangers. The emotional principle ‘do not overact’ is more frequently neglected. In real life situations, emotions are often displayed by gesture and facial expression. Further means are irony, sarcasm, voice-raising (loudness), rapid speaking or giving rather short replies. The risk of a dialogue break is especially strong when interlocutors of different cultures communicate. Communicators of different cultural worlds often interpret utterances completely differently. 289

Marion Grein

All in all, Weigand’s MGM is a multi-disciplinary approach, integrating all facets of human cognition and other human abilities. Her basic ideas or underlying principles are: 1. The focus is on human beings’ acting and reacting on the basis of their perception and cognition. 2. The key concept is human beings’ nature as goal-oriented beings and dialogic individuals. 3. The innate drive is deeply influenced by human beings’ social and cultural surroundings. 4. Emotions are in part innate and universal, in part socially and culturally shaped. 5. She understands emotions as one human ability, next to many others. 6. She claims that any emotion is an indeterminate concept that is not accurately definable. 7. Considering point 6, it is of minor interest how we talk about emotions vis-à-vis how we actually express the emotions we experience. 8. Next to focusing on accompanying emotional components, we have to take a closer look at the three speech act types: constatives, emotives and declaratives.

Further studies on emotion should take the MGM as the basis for the analysis and understand emotions as an integral part of any communicative setting.

17.7 Conclusions Language analysis means analysis of language-in-use. We use language to act and react in dialogic action games. The dialogic means we use to achieve our dialogic purposes are based on our integrated abilities of speech, cognition, perception and emotions. Emotions, thus, are an integral part of any dialogic action game and cannot be analyzed separately. We need to overcome the Cartesian restriction to rules and accept the indeterminance of emotions which are influenced by our genes, mind and culture (cf. Lumsden and Wilson, 2005). In fact, cognitive studies prove that every stimulus is emotionally rated by the limbic system. Thus, reason, brain and emotions are inseparable.

Notes 1 All Cottingham (1985) quotations are Copyright © 1985 Cambridge University Press. Reprinted with the permission of Cambridge University Press. 2 Yet, there is no clear correlation between one emotion and physical reaction. While one person might tremble when being extremely happy, the other might only react similarly when extremely frightened.

References Adam C. and Tannery P. (eds.) (1964–1974). Oeuvres de Descartes, 13 vols., Paris: Vrin/CNRS. Bucci W. (1997). Psychoanalysis and Cognitive Science: A Multiple Code Theory. New York: Guilford Press. Chin B. and Wigglesworth G. (2007). Bilingualism: An Advanced Source Book. London & New York: Routledge. Chomsky N. (1980). Rules and Representations. New York, NY: Columbia Press; Oxford, UK: Basil Blackwell Publisher. Cottingham J., Stoothoff R. and Murdoch D. (1985). The Philosophical Writings of Descartes, 2 vols., Cambridge, UK: Cambridge University Press. Copyright © 1985 Cambridge University Press. Reprinted with the permission of Cambridge University Press. Cottingham J., Stoothoff R. and Murdoch D. (1991). The Philosophical Writings of Descartes: Volume 3, The Correspondence von René Descartes, Cambridge, UK: Cambridge University Press. 290



Emotion, Reason and the Human Brain

Damasio A. (1994). Descartes’ Error: Emotion, Reason, and the Human Brain. New York, NY: GP Putnam’s Sons. (1995 in GB Picador). Damasio A. (1999). The Feeling of What Happens: Body and Emotion in the Making of Consciousness. New York, NY: Harcourt. Damasio A. (2004). Emotions and feelings: A neurobiological perspective. In: Manstead ASR, Fruda N and Fischer, A (eds.) Feelings and Emotions: The Amsterdam Symposium. Studies in Emotion and Social Interaction. New York, NY: Cambridge University Press, 49–57. Damasio, A. (2010). Self Comes to Mind: Constructing the Conscious Brain. New York, NY: Pantheon Books. Damasio A. (2011). The neuronal basis of emotions. Scholarpedia 6(3): 1804. Davidson R. J. and Sutton S. K. (1995) Affective neuroscience: The emergence of a discipline. Current Opinion in Neurobiology, 5(2): 217–224. Descartes R. (1637). La Dioptrique (Dioptrics) [Online] Available at: http://science.larouchepac.com/ fermat/ Descartes%20--%20Dioptrique.pdf. [Accessed: 25 January 2015]. Descartes R. (1649). Les Passions de l’âme. (Passion of the Soul). Translation by Jonathan Bennett. [Online] Available at: www.earlymoderntexts.com/authors/descartes [Accessed: 20 September 2016]. DeLancey S. (2012). Still mirative after all these years. Linguistic Typology 16: 529–564. Drevets W. C., Price J. L., Simpson J. R. et al. (1997). Subgenual prefrontal cortex abnormalities in mood disorders. Nature 386: 824–827. Ekman P. (1980). Face of Man: Universal Expression in a New Guinea Village. New York: Garland. Ekman P. (1984). Expression and the nature of emotion. In: Scherer K and Ekman P (eds.) Approaches to Emotion. Hillsdale, NJ: Erlbaum, 319–344. Ekman P. and Davidson R. J. (eds.) (1994). The Nature of Emotion: Fundamental Questions. New York, NY: Oxford University Press. Ekman P. and Friesen W. (1975). Unmasking the Face. Englewood Cliffs, NJ: Prentice-Hall. Enfield N. J. and Wierzbicka A. (eds.) (2002). The Body in Description of Emotion. Cross-linguistic Studies. Amsterdam/Philadelphia: Benjamins. Fodor J. A. (1983). The Modularity of Mind. Cambridge, MA: MIT Press. Foolen A. (1997). The expressive function of language: Towards a cognitive semantic approach. In: Niemeier S and Dirven R (eds.) The Language of Emotions. Amsterdam, Netherlands: Benjamins, 15–31. Fries N. (1996). Grammatik und Emotionen. LiLi 1996/26 (101): 37–69. Gobl C. and Ní Chasaide A. (2003). The role of voice quality in communicating emotion, mood and attitude. Speech Communication, 40: 189–212. Goodwin M. H. and Goodwin C. (2001). Emotion within situated activity. In: Duranti A (ed.) Linguistic Anthropology: A Reader. Malden, MA: Blackwell Publishing, 239–257. Grein M. (2013a). Neurodidaktik. Grundlagen für Sprachlehrende. München, Germany: Hueber Verlag. Grein M. (2013b). Review article “Emotions reconsidered”. Language and Dialogue, 3 (2): 299–311. Harkins J. and Wierzbicka A. (eds.) (2001). Emotions in Crosslinguistic Perspective. Berlin, Germany: Walter de Gruyter. Hasegawa Y. and Hirose Y. (2005). What the Japanese language tells us about the alleged Japanese relational self. Australian Journal of Linguistics, 25: 219–251. Hobbes (1651). Leviathan (complete title: Leviathan or the Matter, Form and Power of a Common Wealth Ecclesiastical and Civil [Online] Available at: (www.gutenberg.org/ebooks/3207). [Accessed: 20 September 2016]. Imai M., Kita S., Nagumo M. and Okada H. (2008). Sound symbolism facilitates early verb learning. Cognition, 109: 54–65. Izard C. E. (1981). Die Emotionen des Menschen. Eine Einführung in die Grundlagen der Emotionspsychologie. Weinheim u.a.: Beltz. (Original: Izard CE (1977) Human Emotions. New York, NY: Plenum) 291

Marion Grein

Jack R. E., Garrod O. G. B. and Schyns P. G. (2014) Dynamic facial expressions of emotion transmit an evolving hierarchy of signals over time. Current Biology 24 (2): 187–192. Kahn V., Saccamano N. and Coli D. (2006). Politics and the Passions, 1500-1850. Princeton, NJ: Princeton University Press. Kelly J. R. and Hutson-Comeaux S. L. (1999). Gender emotion stereotypes are context specific. Sex Roles 40 (1–2): 107–120. Koch K., Pauly K., Kellermann T., Seiferth N.Y., Reske M. et al. (2007). Gender differences in the cognitive control of emotion: An fMRI study. Neuropsychologia 45 (12): 2744–2754. Kövecses Z. (1990). Emotion Concepts. Berlin/New York: Springer Verlag. Kövecses Z. (2003). Metaphor and Emotion: Language, Culture, and Body in Human Feeling, second edition. Cambridge, UK: Cambridge University Press. Koopmann-Holm B. and Matsumoto D. (2011) Values and display rules for specific emotions. Journal of Cross-Cultural Psychology 42 (3): 355–371. Kring A. M. and Gordon A. H. (1998) Gender differences in emotion: Expressions, experience and physiology. Journal of Personality and Social Psychology 74: 686–703. LeDoux J. E. (1996) The Emotional Brain. New York, NY: Simon & Schuster. LeDoux J. E. (2001) Synaptic Self. New York, NY: Viking. Lewis M. and Haviland-Jones J. M. (eds.) (2004). Handbook of Emotions. New York, NY: The Guilford Press. Lüdtke U. M. (2006). Sprache und Emotion: Neurowissenschaftliche und linguistische Zusammenhänge. In: Bahr R. and Iven C. (eds.) Sprache – Emotion – Bewusstheit. Beiträge zur Sprachtherapie in Schule, Praxis, Klinik. Idstein: Schulz-Kirchner Verlag, 17–26. Lumsden C.J. and Wilson E.O. (2005) Genes, Mind, and Culture. The Coevolutionary Process, 25th Anniversary Edition (Original 1981). Singapore: World Scientific Publishing Co. Malebranche N. (1997) The Search after Truth, translated and edited by Lennon T.M. and Olscamp P.J. Cambridge, UK: Cambridge University Press. Oda H. (2000) An embodied semantic mechanism for mimetic words in Japanese. Unpublished PhD thesis. Indiana University, Bloomington, USA. Parkins R. (2012) Gender and emotional expressiveness: An analysis of prosodic features in emotional expression. Griffith Working Papers in Pragmatics and Intercultural Communication 5 (1): 46–54. Pascal B. (1670) Pensées. Paris. [Online] Available at: www.gutenberg.org/ebooks/18269. [Accessed: 20 September 2016] Pavlenko A. (2005) Emotions and Multilingualism. Cambridge, UK: Cambridge University Press. Perunovic E. and Perunovic M. (2012) Language and emotion: The case of bicultural individuals. In: Wilson P.A. (ed.) Dynamicity in Emotion Concepts. Lodz Studies in Language. Frankfurt: Peter Lang, 359–372. Pinker S. (1994). The Language Instinct. New York, NY: Harper Perennial Modern Classics. Pinker S. (1997). How the Mind Works. New York, NY: W. W. Norton & Company. Reiman E. M. (1997). The application of positron emission tomography to the study of normal and pathologic emotions. J Clin Psychiatry 58 (suppl 16): 4–12. Safdar S., Matsumoto D., Kwantes C. et al. (2009) Variations of emotional display rules within and across cultures: A comparison between Canada, USA, and Japan. Canadian Journal of Behavioural Science 41 (1): 1–10. Schmitter A. M. (2014). 17th and 18th century theories of emotions. In: Zalta EN (ed.) The Stanford Encyclopedia of Philosophy. Oxford: Oxford University Press. Schultz W. (2000). Multiple reward signals in the brain. Nature Reviews Neuroscience 1: 199–207. Schwarz-Friesel M. (2013) Sprache und Emotion, 2nd edn. Tübingen and Basel: A. Francke Verlag. Stefanowitsch A. (2006) Words and their metaphors: A corpus-based approach. Trends in Linguistics Studies and Monographs, 63–105. [Online] Available at: www.academia.edu/5548002/Words_ and_their_metaphors_A_corpus-based_approach [Accessed: 20 September 2016]. Stefanowitsch A. and Gries S. T. (eds.) (2006). Corpus-based Approaches to Metaphor and Metonymy. Berlin and New York: Mouton de Gruyter. 292



Emotion, Reason and the Human Brain

Tanaka A., Koizumi A., Imai H. et al. (2010) I feel your voice: Cultural differences in the multisensory perception of emotion. Psychological Science 21 (9): 1259–1262. Toivonen R., Kivelä M., Saramäki J. et al. (2012). Networks of emotion concepts. PLoS ONE 7(1), e28883. [Online] Available at: www.becs.aalto.fi/bml/pdf/toivonen_etal_2012.pdf. [Accessed: 20 September 2016]. Vul E., Harris C., Winkielman P. and Pashler H. (2008). Puzzling high correlations in fMRI studies of emotion, personality, and social cognition. Perspectives on Psychological Science, 4 (3) San Diego, CA: University of California and Massachusetts Institute of Technology, 274–290. Weigand E. (1996). Words and their role in language use. In: Weigand E and Hundsnurscher F (eds.) Lexical Structures and Language Use, Vol. 1. Tübingen: Niemeyer, 151–167. Weigand E. (1998a). Contrastive lexical semantics. In: Weigand E (ed.) Contrastive Lexical Semantics, Amsterdam/Philadelphia: Benjamins, 25–44. Weigand E. (1998b). The vocabulary of emotion. A contrastive analysis of anger in GermanItalian-English. In: Weigand E (ed.) Contrastive Lexical Semantics. Amsterdam/Philadelphia: Benjamins, 45–66. Weigand E. (2004). Emotion in Dialogic Interaction. Amsterdam/Philadelphia: Benjamins. Weigand E. (2009). Language as Dialogue. Amsterdam/Philadelphia: Benjamins. Weigand E. (2010). Dialogue: The Mixed Game. Amsterdam/Philadelphia: Benjamins. Wierzbicka A. (1986). Human emotions: Universal or culture-specific? American Anthropologist, New Series 88 (3): 584–594. Wierzbicka A. (1992). The semantics of interjection. Journal of Pragmatics 18: 159–192. Wierzbicka A. (1996). Semantics: Primes and Universals. Oxford, UK: Oxford University Press. Wierzbicka A. (1999). Emotions Across Languages and Cultures: Diversity and Universals. Cambridge, UK: Cambridge University Press. Wierzbicka A. (2004). Bilingual lives, bilingual experience. Journal of Multilingual and Multicultural Development 25 (2/3): 94–104. Wierzbicka A. (2009). Language and metalanguage: Key issues in emotion research. Emotion Review 1: 3–14. Wilce J. M. (2009). Language and Emotion. Cambridge, UK: Cambridge University Press. Wilson P. A. and Lewandowska-Tomaszczyk B. (2012). The nature of emotions. In: Wilson PA (ed.) Dynamicity in Emotion Concepts. Lodz Studies in Language, Vol. 27. Frankfurt am Main: Peter Lang, 7–30.

293

18 Self-interest and Social Concerns Jennifer L. Adams

18.1 Introduction As is true of many of the disciplines that are involved in the study of the role of language in the production of meaning, scholars who study communication between social actors in terms of self-interest and persuasion have long relied on theories of rhetoric and social influence to explain interaction. As defined by Aristotle in the fourth century BCE, rhetoric is the counterpart of the dialectic and more precisely, is the art of discovering in any given situation the available means of persuasion (Aristotle, The Rhetoric). His framing of rhetoric as effective language use continues to be a focus of many scholars of rhetoric today, yet despite thousands of years of theorizing, no single explanation for how rhetoric as persuasion works has gained popularity (Foss, Foss, and Trapp, 1991). Often, disagreements emerge as scholars suggest and revise various rule-based theories that help explain how and why a given persuasive message is impactful. However, recent research in the area of human cognition and language suggest that the search for abstract rules that explain human language may be in vain. For example, proponents of the mixed-game theory of language abandoned such a quest for a universal explanation of how persuasion works and argue that language cannot be separated from its use by human beings in the moment of the utterance (Weigand, 2010). These theorists call for a more holistic and interdisciplinary explanation for understanding meaning as dialogue. In this chapter, I will explore the application of the Mixed Game Model (MGM) of language to the topic of rhetoric in an attempt to review and explore its implications (see Chapter 11, this volume). Toward this end, I will review the dialogic mixed-game theory as conceptualized by Edda Weigand (2010) prior to exploring its implications for persuasion (including social concerns such as respect and politeness) and for argumentation, two often distinct branches of modern rhetoric studies that are brought together in the MGM (Weigand, 2010: 93; see also Adams, 2012). Additionally, I will review the concept of audience in rhetorical theory, focusing on the implications of a dialogic rhetoric for understanding the relationship between rhetor and audience. Through this discussion, self-interest and social concern, along with issues relating to negotiation, power, politeness, and respect, will be examined within the frame of the MGM. 294



Self-interest and Social Concerns

18.2  Mixed Game Theory Within the realm of traditional rhetorical theory, language has been treated as a clear and separate object in itself. Maintaining this view, scholars have long sought to understand the rules and principles that guide competent language use. However, Edda Weigand (2010) argues that perspective has led to needless theoretical abstractions that take researchers further away from the object (language) they hope to better understand. In a critique notable for its centrality to much contemporary theorizing about communication including rhetoric and influence, Weigand (2010) highlights the fundamental abstraction committed by De Saussure in his definition of language as a sign system by noting the lack of a speaker in his model: “No reference is made to human beings who create and use signs; the sign itself is the agent, carries meaning” (Weigand, 2010: 11). Recognizing that theoretical abstractions can never answer questions about language use in the lives of real people, Weigand advances an alternative, pragmatic theory of language use – the MGM. Radically dialogic, Weigand (2009) argues that the study of language cannot be removed from its use by human beings, who are social beings, always oriented toward their own needs and purposes as separate, biological beings. Furthermore, each communicative utterance is a part of the human world, and the linguistic components of that utterance are only part of the meaning within a dialogic speech situation. In other words, not all communicative exchanges are literal exchanges between participants. Conversational meaning is drawn from all aspects of what is said, how it is said, by whom, and in what context (Weigand, 2009: 26). Given these premises, Weigand proposes a holistic theory, outlined in its entirety in Dialogue: The Mixed Game (2010), that considers language only as it is used, by human beings who are competent, complex users. MGM not only provides reference to the context in which an utterance happens, but also highlights the fact that utterances are dialogic actions. Thus, “dialogic action is not action by verbal means but action by the integrated use of communicative means, verbal, perceptual, and cognitive” (Weigand, 2010: 43). Importantly, the study of language-in-use also assumes competency, in that human beings are already competent, adept users of language. Weigand’s MGM includes several features that she suggests are rooted in basic human nature. First, human abilities are parts of a large whole in that they are integrated with one another. Second, human beings have goals and desires, meaning that their actions are driven by a purpose. Third, human beings are social beings by nature, which implies that they have a dual interest in self and other. Finally, humans are persuasive beings who, because of their social nature, want to be accepted by others (see Weigand, 2009; Weigand, 2010). Here, those with an interest in rhetoric will readily see that persuasion is central in the MGM and forms the basis by which we dialogue with others, with varying levels of directness. The MGM model, as advanced by Weigand (2010), can be said to have three primary constitutive principles: the principles of action, dialogue, and coherence. Giving further theoretical depth to these three constitutive principles are a regulative principle and an executive principle. Understanding the constitutive, regulative and executive principles is compulsory to understanding rhetoric and social concerns through the MGM model. As a constitutive principle, the principle of action recognizes that people are self-­ motivated actors who are aware of and act towards their own self-interest. The principle of dialogue recognizes that people are also social actors who use communication to either initiate or react to dialogue. As Weigand (2010: 74) explains: “There is no single pattern underlying interaction; meaning and understanding are different for different individuals and open for negotiation.” Finally, the principle of coherence recognizes that the ability to 295

Jennifer L. Adams

speak is but one of many related abilities innate to humans, and that perceptual and cognitive patterns of people impact dialogue. These principles serve as the constitutive basis for the MGM and have important implications for rhetoric as persuasion and as argumentation, explicated later in this chapter. Equally important to the MGM theory’s constitutive principles are the regulative and executive principles. Regulative principles are required simply because humans are social beings who engage one another, necessitating a balance between their own self-interest and their social concerns (Weigand, 2010: 75). One such regulative principle, central to the concerns of this chapter, is the politeness principle. For Weigand, “Politeness can mean respect for the other human being or can be used in the self-interest of the speaker in order to make things easier” (2010: 75). In either case, politeness is used by human actors who are competent communicators with rhetorical experience engaging others. Politeness is not the only means through which individual speakers regulate dialogue when explained by the MGM, and is complemented by the Regulatory Principles of Emotion. Emotion has long been recognized as a human ability that interacts with other abilities, such as rationality, coloring the experience of dialogue. Thus, the principle of emotion derives from cultural sources that create expectations and conventions for its use in conversation (Weigand, 2010: 76). Importantly, then, culture is not an addition to the MGM because it is inherent to one’s expression of emotion and politeness. The regulative principles are both rhetorical and cultural in nature, and are central to any consideration of persuasion, social concerns, or argumentation. Finally, when human beings participate in dialogue with others, taking turns in speaking and listening requires a complex understanding of linguistic sequencing. In the MGM, the ‘executive principles’ guide reactions in extended dialogues that assist people in obtaining goals. Executive principles can be viewed as “strategies” that are considered to be effective by users in attempting to fulfill their own desires (Weigand, 2010: 76). Thus, the constitutive, regulative, and executive principles all have important implications for those wishing to understand rhetoric, especially as it relates to dialogue in use. Persuasion lies at the very heart of the MGM, which seems quite natural given the complex social situations that all human beings must negotiate with others. Balance between self and other require skill and understanding in dialogue, and this chapter will explore the roles of persuasion and argumentation as means of negotiation available for goal-directed human beings. However, before moving toward a closer examination of persuasion, self-interest and social concern, a brief explanation of some types of different action games will help one gain an appreciation for the varied communicative actions of human beings. An “action game” is the term used by Weigand to describe dialogue. Acknowledging that dialogue is complex, realistically, humans are faced with an “ever-changing performance, i.e. with innumerable games” (Weigand, 2010: 129). Weigand distinguishes between two general types of action games: minimal games that involve only action and reaction, and more complex games that may involve an extended minimal game or else a language game that involves multiple phases (Weigand, 2010: 130). Weigand addresses several types of minimal action games including the following: declaratives (games that create); directives (games that change); representatives (games that express) and explorative (games that question) (see Weigand, 2010: 126–154 for detailed descriptions and taxonomy of each). Even in these minimal action games, one will note that persuasion is at work here, in the basic desire for recognition and response. Of course, the dialogue engaged by humans may extend beyond minimal action games, as when there are problems in understanding, when sequences must be used for clarification, when providing arguments, when using forms of power, and when simply extending propositions (see Weigand, 2010: 211–226). When 296



Self-interest and Social Concerns

dialogue-in-use becomes even more difficult for simple sequences, one beholds the notion of the “complex” action game, one which is fully engaged in the stream of human life and full with linguistic and paralinguistic meaning. In sum, Weigand’s MGM offers a new way of understanding the complex language games engaged by human beings who are both self-interested and social. This fundamental tension requires action games that mediate them in ways that help people come to understanding with one another through communication. That the MGM clearly intersects with the concerns of those who study the traditional art of rhetoric and persuasion is clear even in this necessarily brief summary. In the remainder of this chapter, I will address issues of traditional concern to rhetoricians, including persuasion, argumentation, and audience as they are implicated by Weigand’s MGM.

18.3 Persuasion Those who study persuasion have long searched for techniques that might be employed to effectively convince others of some position through the skilled use of language, emotion, and rationality. Others, however, have argued that persuasion involves more than mastering techniques, suggesting that language, in a fundamental way, is persuasive. For example, I. A. Richards was one of the first to attempt to rescue the semiotic triangle from abstraction by adding the function of meaning and context to understanding (Richards and Ogden, 1929). Meaning, according to Richards and Ogden, is necessarily rhetorical because the function of language is “as an instrument for the promotion of purposes” that is prone to complications resultant from intentions of the speaker (1929: 16). Richard Weaver also highlighted the persuasive function of all language when he famously noted, “Rhetoric is cognate with language” (Weaver, 1985: 221). Weaver argued that all use of speech, both written and oral, is persuasive and he used the term “sermonic” to explain the suasory component of all language: “We are all of us preachers in private or public capacities. We have no sooner uttered words than we have given impulse to other people to look at the world, or some small part of it, in our way” (1985: 224). In this way language, as rhetoric, is never neutral and it is also embodied with the subjectivity of the speaker and context of the speaking situation. Finally, and perhaps most clearly, Kenneth Burke also theorized that all language is rhetorical. However, Burke (1950) introduced a new term, “identification” rather than persuasion, to describe the ultimate function of language. According to Burke, identification works to join interests of human actors, who are necessarily independent, isolated agents that join with others. In this way, human actors are consubstantial with one another: “in being identified with B, A is substantially one with a person other than himself. Yet at the same time he remains unique, an individual locus of motives. Thus, he is both joined and separate, at once a distinct substance and consubstantial with another” (Burke, 1950: 21). Importantly, identification cannot exist in human consciousness without division, because “if men were not apart from one another, there would be no need for the rhetorician to proclaim their unity” (1950: 21). Thus, identification is central to the way that Burke conceptualizes language, which he calls “symbolic action”, and all human communication works toward the goals of identification, making it an elemental function of all language. Identification as a function of discourse happens because interlocutors are both individual and socially-attuned creatures. Any language use, any dialogue, is necessarily rhetorical in that it is used by competent actors to impact others. This is of primary importance in understanding the MGM as advanced by Weigand (2008), who argues, “rhetoric is inherent 297

Jennifer L. Adams

to dialogue. Texts are not just rhetorical texts if they contain rhetorical figures, they are always produced by human beings who are attempting to achieve more or less effectively certain purposes in dialogic interaction” (2008: 3). Human beings are always interested parties in the events of their own lives, and interact in dialogue with others in ways that they feel will be rhetorically effective. Indeed, Weigand (2001: 65) aptly notes: When we enter into dialogue and try to negotiate our position, it is obvious that we will do it in a way that is intended to be more or less effective. It is simply misguided to assume we would consciously damage our communicative purpose by not trying to achieve it in a more or less effective manner even if in some cases we are not particularly concerned about it. In the context of the MGM, this rhetorical aspect of dialogue is called competence-inperformance, the extraordinary human ability to communicate and understand through language in complex and competent ways. Weigand’s arguments constitute a pragmatic view of language that recognizes the skill of the human beings who use it, and so rhetoric is naturally a part of the process of dialogue. If the goal in normal human communication is to achieve a mutual understanding involving both speaker and listener, then everyone shares in their desire to be effective. Therefore, all language use is persuasive. Rhetoric, then, is part of the larger process of dialogic action games, and a regular part of human interaction. Importantly, this perspective does not diminish the role of rhetoric for understanding the effective use of communication; in fact, within the MGM, rhetoric is an inherent part of language-in-use. Retracing our steps back to Aristotle’s original formulation of rhetoric, one might note that as yet, I have addressed primarily those issues related to logos, which is but one of three artistic proofs he identified in Aristotle’s The Rhetoric. Yet, importantly, the MGM is not limited when explaining issues of effectiveness or mutual understanding between interlocutors; the artistic proofs of ethos and pathos are accounted for and continue to remain significant in Weigand’s explanation of dialogue. Recalling the regulative principles discussed in Section 18.2 outlining the MGM, the principle of politeness and the principle of emotion are both embedded in dialogue, and both speak directly to concerns of pathos and ethos in the traditional formulation of rhetoric’s artistic proofs. First, I will consider the issue of regulation between the individual self-interest and the competing desire to be a participating member of a community. According to the MGM, human beings are naturally persuaders, given that they have goals and interests that are directed by their own individual wants and needs as mediated by cultural values. Weigand (2010: 93) explains: “On the one hand, we want to pursue our own purposes and interests, on the other hand, we are at the same time the other social being whose interests need to be respected.” In many ways, this concern parallels concerns of ethos, which can be called the audience’s perception of the character of the speaker. Defined in the literature of rhetoric studies and speech, ethos is variously understood as good will, credibility, or disposition (see Rosenthal, 1966; McCroskey and Young, 1981; McCroskey and Teven, 1999; and Lucas, 2011). Within the MGM, these slight variations in interpretation become somewhat insignificant in that all are subsumed by the notion of the principle of politeness, a form of social regulation that mediates self-interest and social interests (Weigand, 2010: 94). Weigand (2010: 94) defines politeness as “a positive concept, the concept of the respect to be paid to our fellow human beings and expected for ourselves”. Actions that are considered polite vary across cultures as does the role of politeness in any given interaction, but 298



Self-interest and Social Concerns

all humans seem to use it as part of their own self-interested nature (see Weigand, 2007: 45–56). There are several possible strategies for politeness that can be used by individuals across its different functions and in different contexts; thus, human attempts to achieve acceptance from another vary across the different types of action games utilized. When engaging in Directive Action Games, individuals use rhetorical means to accomplish their directives, such as offers or threats. Most scholars of rhetoric balk at the notion that a threat is a form of persuasion and define it outside the realm of ethical rhetorical practice (for example, see Woodward and Denton, 2013; Perloff, 2013). Similarly, in the MGM, threats may be said to be the complete removal of politeness, and as such, generally ineffective in maintaining a balance between self and other. Offers can include modest gifts of time or tokens of appreciation to outright bribes or extortions, and so may also vary in their engagement of the principle of politeness and ultimately in their effectiveness (Weigand, 2010: 96). Additionally, directive action games can also involve rhetorical strategies that are indirect, using techniques of suggestion or implication that trigger emotions and indirectly work toward a directive goal. Weigand (2010: 98) notes the frequent use of these indirect strategies which have the effect of increasing the volition of a directive in advertising copy. For example, the American insurance company State Farm has long had a slogan, “State Farm is on your side”. While this may not appear to have any immediate persuasive content, State Farm is only on your side if you have paid insurance premiums for the security of your home, car, or other property. Thus, this simple statement is directive in that it indirectly requires one to pay for State Farm to be on their side (for other examples, see Weigand, 2010: 98–99). Rhetorical means are also used in Representative Action Games, used when one desires to strengthen a truth claim. Given that there are no absolute truths, the construction of a reified truth requires support and rhetorical expertise, and Weigand (2010: 99–100) notes that rationality and the selection of appropriate arguments are central to these types of action games. When considered as persuasive discourse, these action games require some balance between the speaker and the listener and fulfil the needs of the regulative principle. While I will reserve prolonged discussion of argumentation within the MGM until Section 18.4, I will explain the role of rationality as it functions within the principle of politeness in the MGM. Rationality has long been understood as logical conclusiveness, but that is not the meaning of rationality-in-performance within the MGM. Instead, rationality-in-performance refers to “so-called reality or what is going on in life” (Weigand, 2014: 248). This is a complex state of affairs with which humans must necessarily engage, and from the perspective of the MGM, humans always act in a way that is rational to them given their self-directed interests and experiences with others. This notion is fundamental: “Rationality-in-performance is the central component of what I call competence in performance, an extraordinary human ability which enables us to meet the vicissitudes of life and to negotiate our interests and purposes in dialogue with our fellow beings” (Weigand, 2014: 249). Given this skill, humans understand the need for some strategy of politeness that will be selected based upon the unique goals and sequences of the action game(s) engaged in any given dialogue. Rationality is not, then, something that can be removed from emotion, requiring any method of analysis of language use must account for it. Traditionally, the study of emotion in rhetoric has been understood as a form of pathos, in which the use of emotion is a conscious choice on the part of the speaker for the purpose of attaining a specific feeling or reaction from the listener or audience (Lee, 1939; Hyde, 1984). In this sense, pathos as emotion is something used concurrently with reason and logic to achieve the goal of the 299

Jennifer L. Adams

speaker. It is simply one of several means of persuasion available to a speaker. Although this approach does acknowledge that emotion impacts and shapes human action, it does not acknowledge the ubiquitous nature of emotion as embedded in dialogue as does the MGM. Attempting to move beyond Aristotelian conceptions of pathos, some scholars of rhetoric suggested consideration of Heidegger’s mood in understanding emotion and language, suggesting that mood consistently underlies human being-in-the-world and as such shapes the nature of all human experience (Crosswhite, 1989; Ramsey, 1993). Although these attempts to conceptualize emotion as central and inseparable from meaning went largely unheeded, they do complement the more centralized placement emotion has within the MCM. A basic assumption within the MGM is that human beings use language in their own interest or benefit. The principle of emotion acknowledges the social nature of humans, who necessarily trigger a variety of emotions within any dialogue. The study of emotion is relevant to the MGM because it necessarily impacts dialogue (see Chapters 17 and 22, this volume). As Weigand notes, emotions are “on the one hand dependent upon the interlocutor, and on how he or she sees and evaluates the world, and on the other hand, they influence the sequence of actions” (2009: 167). As such, emotion cannot be isolated from dialogue and it requires attention from scholars attempting to understand any language-in-use. Additionally, much of what is communicated between individuals occurs through nonverbal communication and paralanguage. Indeed, communication scholars suggest we impart a majority of our interpretations in conversation with others based upon such nonverbal indicators (Floyd, 2013.) Often, these metacommunicative devices carry the emotion of an expression. Taking into consideration all these factors, Walrod (2004) outlines the impact of emotions on normative discourse, or discourse used to influence or persuade, especially as it relates to the self-directed goal of social engagement and control. Scholarly consensus, then, suggests that emotion and the meaning of a dialogue, including its individual sequences, are interdependent and must be considered as such in any linguistic analysis. Before concluding this review of rhetoric as persuasion in dialogue, I believe it is important to reflect upon the role of power in human relations that necessarily impact action games between people. If we accept that all language is rhetorical in that people attempt to use effective communication to achieve their goals, then we must also recognize that the act of negotiation is also inherent to authentic dialogue. Weigand (2001: 64) has written about the role of human negotiations as dialogue, importantly noting specifically: In language-in-use, we do not send and receive packages of defined meanings whose understanding is presupposed. On the contrary, meaning and understanding are negotiated by interlocutors. In this sense, negotiation means a general methodological concept for every type of language-in-use, namely that human beings interact in order to negotiate their respective positions. Extending this idea to its natural conclusion, she notes that persuasion in negotiation will always employ the use of power, of which there are multiple types that can usefully be categorized as positive or negative uses of power. Positive power is utilized when one person has the ability to offer a reward, by making an offer or some concession, to the other. Negative power, which according to Weigand (2001) is the more commonly employed, is the ability to punish or harm the other through a varied array of means. Identifying “selfassertion” as an “inherently human strategy which easily leads to the negative use of power” (2001: 67), Weigand (2001) also identifies cognitive principles of negative power that include intimidating a conversational partner by simulating power even if one does not have 300



Self-interest and Social Concerns

it, using subordinates or others who have little choice to support one’s position, threatening the partner with any number of real or imagined harms, and using deceit, lies or confusion against a partner. Persuasion, then, as a function of rhetoric that focuses on the effective use of language is not something that exists outside of dialogue. Instead, according to the MGM, rhetoric is a component of dialogue and effectiveness in dialogue that is a natural skill that every interlocutor intends to use, but which in practice can happen in more or less skillful ways. Competence-in-performance within the MGM includes the rhetorical skill that is used by interlocutors as they engage in action games through dialogue with one another.

18.4 Argumentation In the contemporary academic study of rhetoric, argumentation is often considered to be a separate concern from persuasion with its focus upon issues of reason and rationality (Weigand, 2010). Yet, within the frame of the MGM, rationality in performance cannot be considered to be apart from persuasion; indeed, persuasion is a required skill in the performance. Instead, both are necessary to the understanding of a complex human actor in dialogue. The most studied argumentation theorist of the twentieth century, Stephen Toulmin, outlined a theory of practical arguments used in everyday contexts in his book The Uses of Argument (1958). After rejecting theoretical, analytical arguments as formal processes for discovering universal truths as unpractical, Toulmin put forth a theory of “argumentfields” that included a six-part model for understanding and analyzing arguments, including the claim, the grounds, the warrant, the backing, the qualifier, and the rebuttal. However, despite rejecting overly formulaic and rule-based models of argument, Toulmin’s model for argument analysis still remains unrecognizable as real dialogues, in which interlocutors go back and forth, using elements of the speaking situation, silence and other non-textual cues to communicate and affect their conversations (Weigand, 2009). Whereas Toulmin ultimately suggests norms for reasoned argument in practical use, Weigand makes no normative claims in the MGM. Instead, MGM is a descriptive model that rejects prescriptive models that might suggest valuations. Weigand (2006: 63) explains, a descriptive model “distinguishes between different ways of behavior, among them what could be called civilized behavior which can be expected and even demanded by everyday and institutional conventions. Conventions however are not yet norms”. In this way, the MGM abandons the unnecessary abstractions of normative behaviors and instead positions argumentation as an action game played by those with both self-interest and desire for effectiveness. According to the MGM, an argument is a structured action game that is “not simply a reason but a justification for an action” (Weigand, 2006: 66). In the MGM, every speech act is sequential; it is either an action or a reaction in a larger sequence of speech. As already described, this sequential interaction is known as the “dialogical principle proper”, and its function is ultimately coming to an understanding. Recall that there are four basic speech acts in the minimalist form identified by Weigand: representatives, directives, exploratives, and declaratives (see Weigand, 2006; Weigand 2010). These four fundamental game types “correspond to the basic mental states of belief and desire” (Weigand 2006: 63), and are themselves sequential actions and reactions. The link between the actions and reactions in dialogue is known as coherence, which is “no longer given in the text, [but] given by interlocutors” (Weigand, 2009: 160). 301

Jennifer L. Adams

Thus, representative speech acts are those that make a claim to truth (the action) and require acceptance or rejection (the reaction). Directives are those that make a claim of volition that relates to the performance of a physical action, and require consent in the performance or not. Exploratives are those that make a claim of volition that relates to knowledge, and requires a response related to sharing that knowledge or not. Finally, declaratives are statements that contain both the making and fulfilling of a claim, and thus require no reaction aside from possibly confirmation (for a more thorough description, please see Weigand, 2009). Importantly, for each type of speech action, the reaction is not prescribed or necessary in any way based upon rules of reason or logic. Instead, we must think of all reactions as “conceived of as expectation based on probability” (Weigand, 2006). No formal skill in argumentation or persuasion in the action game necessitates the reaction desired, and all forms of action games can be considered arguments. However, Weigand (2006) notes that representatives and directives are most specifically relevant for understanding argument in human negotiations. In the case of representative speech acts, interlocutors are engaging in a dialogue regarding issues of truth, and arguments ideally support the acceptance of the act by another party. For example, I may state, “President Obama is a citizen of the United States” as a reason to justify a legal vote for him in an election. This statement is a representative speech act that I can support by arguments including “President Obama holds a legal birth certificate from the United States” and “President Obama has an American Social Security Card”. These statements, which function as assertions, impact the possibility of acceptance on the part of my conversational partner, and are thus considered justifications in support of my original representative statement. My partner may accept this reasoning, or may present a thesis that counters mine. Of vital importance to this sequence, however, are the probabilities of any possible reaction. Interlocutors engaged in argumentation utilize this probability, and the end result may be that my conversational partner and I agree to my thesis, or her thesis, or some compromise between them. This is evidenced, in my representative example above, by the active participation of many Americans in the “birther” movement to prove that President Obama is indeed not a legal citizen of the United States (Howell, 2012). Furthermore, recall that more than the textual elements of the speech act are at work in any given dialogical sequence, making acceptance dependent upon more than the mere words but also nonverbal and contextual elements. In the case of directives, which are claims of volition for practical action, arguments do not support the claim directly, as with representative speech acts, but instead justify the “preconditions underlying the claim” (Weigand, 2006: 67). According to Weigand, these types of justifications are typically assertions about “practicality, usefulness and motivation which are used to reject a directive claim” (2006: 67). For example, I may make a directive argument: “Please recycle all the plastic bags you receive at the market.” I may justify this by using arguments about the practicality of this action: “Most stores have recycling stations to recycle plastic bags.” I may also justify this through arguments about usefulness, so that “Recycling plastic bags really will help the environment by directly reducing pollution.” Finally, I might justify recycling of plastic bags through statements about motivation or desirability: “Recycling plastic bags saves limited natural resources.” My conversational partner, in response, has a range of reactions possible for all these justifications, and might consent to the act, refuse to engage in the act, or even postpone the act. Thus, the arguments of motive, usefulness, and practicality impact whether my partner in dialogue consents to the act of recycling plastic bags, but they are all preconditional in the act of recycling itself. 302



Self-interest and Social Concerns

The example of a directive speech act, however, is a good reminder that the MGM does not see rhetoric as two distinct actions, one persuasive and the other rational or reasonable. Indeed, the directive speech act is ultimately an act of persuasion itself. Furthermore, Weigand (2009) reminds us that a directive claim may be disguised by an interlocutor who uses the form of a representative speech act, as when a volition is presented as a thesis with arguments for and contra. However, these dialogues often reveal themselves as directives through the use of emotion or evaluative reasoning. As dialogic action games, representative and directive function as means of negotiation in conversations between individuals who have competing needs, goals or desires. Weigand (2006: 75) explains this concept eloquently by noting, “Persuasion on the meaning side of language corresponds to rhetoric as the art of effective language use on the side of the communicative means”. Rhetoric and argument are thus integrated in the MGM.

18.5 Audience Rhetorical, persuasive discourse is always written or spoken for an audience composed of individuals who an orator would like to change in mind or deed. Because rhetoric, as an academic discipline, has long been concerned with the means of persuasion and the concept of the audience has figured as a central component in rhetorical theory since the earliest conceptions. To the extent that the point of rhetoric is persuasion, the audience must be considered a primary concern for those concerned with rhetoric. Cicero notes, “Now nothing in oratory … is more important than to win for the orator the favor of his hearer, and to have the latter so affected as to be swayed by something resembling a mental impulse of emotion, rather than by judgment or deliberation” (De Oratore: 328). Given this primary concern, I conclude my review of the MGM as it relates to rhetoric by exploring its implications for audience. There are at least three clearly different conceptualizations of audience that have emerged from rhetorical theory. First, there is the most basic sense that an audience includes all those who are exposed to a given rhetorical act, or in other words, those who are actually present to hear the arguments. Often called the “empirical audience”, this audience may include individuals immediately present in a face-to-face situation or those who are removed in time or space but hear the message over some medium (Campbell and Huxman, 2003). This audience is necessarily limited by the reach of the given message. A second type of audience outlined by theorists is the “ideal” or target audience. Generally, most rhetorical acts are constructed with an audience already in mind that includes those individuals who are most suited psychologically and ideologically to accept a persuasive message. Perelman and Olbrechts-Tyteca (1960: 19) define audience as “the ensemble of those whom the speaker wishes to influence by his argumentation”. Perelman and Olbrechts-Tyteca further divide the notion of an ideal audience into two smaller typologies including the “universal” audience and the “particular audience”, useful to those constructing persuasive messages. While the particular audience is much like a basic audience that includes all listeners, the “universal” audience is one that is somewhat idealistic in that it is a mental construct invented by the rhetor and used to imagine the type of audience that is most likely to be receptive to the rhetorical message. The concept of the universal audience serves two purposes for the speaker: it serves to aid in the choice of appeals used in persuasion, and it serves as a standard that audiences can use to determine good audiences from bad (Foss, Foss, and Trapp, 1991). Finally, there is a notion of the audience as “agents of change” (Campbell and Huxman, 2003). This audience is limited to those hearers who are prepared and able to act upon a 303

Jennifer L. Adams

given message. Theorized by Lloyd Bitzer (1968: 8), this audience is considered a part of the rhetorical situation: “a rhetorical audience consists only of those persons who are capable of being influenced by discourse and of being mediators of change.” While anyone who hears a message might be a member of a generic audience, Bitzer’s rhetorical audiences must be able to become an “agent of change” (1968: 8). Despite rhetorical theorists’ appreciation for these different types of audiences in a rhetorical situation, the audience is never conceived as a participant in a conversation or dialogue in any of these models. While Bitzer’s notion of the audience as agents of change does make the audience active, the only action of concern is their ability to affect the change desired by the rhetor. Thus, although rhetorical discourse has always been considered participatory by the theorists and practitioners of the art, not all have accepted this interaction as dialogic. M. M. Bahktin (1988) famously labeled rhetoric as a monologic, rather than a truly dialogic, enterprise. Bakhtin considers rhetoric monologic, authoritarian, and polemic because it is ultimately representative of one voice or perspective, including only the voice of the orator and appropriating or using the heteroglot voices of other in the social arena (see Murphey, 2001; Halasec, 1998; and Jasinski, 1997). Despite the work of many rhetoric scholars to re-imagine the subject as a more dialogic enterprise, these theorists most often accept Bahktin’s critiques as sound and reformulate or reconceptualize rhetoric so it is more in line with the multi-voiced novel. Yet, from the perspective of the MGM as conceived by Weigand, rhetoric (like all discourse) is already dialogic. Rhetoric as language simply cannot be separated from the participants who speak it and hear it. As Weigand (2010: 41) explains: “Dialogic action not only means face-to-face dialogue with another interlocutor ... the term “dialogue” comprises any form of communicative language use, including the monologic form without turn taking.” The distinction between monologue and dialogue is a formal distinction, but “at the functional level, both forms are dialogically directed according to the general dialogic principle of any communicative language use” (Weigand, 2010: 41). Ultimately, even discourse without formal turn-taking can be a dialogue because “the response can also be a mental action and need not be verbally expressed” (2010: 41). In this way, then, all human discourse is fundamentally dialogic regardless of the activity level of an audience or listener. Rhetoric, then, is not dialogic simply because it is spoken with audience compliance in mind. Simply understanding that there is an audience who hears, processes, and complies (or not) with the speaker does not make rhetoric automatically dialogic, and Bakhtin’s criticisms of monologism are legitimate when one considers that the primary question asked of most rhetoric is whether or not the audience can accept the demands of the rhetor. Ultimately, this allows for a very limited understanding of dialogue. As Weigand (2009: 131) notes: “In a strict sense, dialogue begins by acting and reacting in the action game. These reactions are not restricted to adherence but might bring in another type of argumentation and thus result in a negotiation of different views between different interlocutors.” That rhetoric is dialogic stems not from how individual orators use language more or less effectively. Instead, rhetoric is dialogic simply because it is discourse used by human actors. Thus, in addition to the action principle, the coherence principle, and the dialogic principle in the MGM, we can speak of “Rhetorical Principles”, which teach us how to use language effectively and persuasively. As Weigand (2009: 140) defines them: “Rhetorical principles refer on the one hand to strategies for pushing claims and arguments and on the other to strategies for the use of specific evaluative expressions.” These rhetorical principles include the concerns of rationality and logic, emerging from the study of rhetoric as argument, but they also include the issues of power and emotion that emerge from the study of 304



Self-interest and Social Concerns

persuasion (see also Weigand, 2010: 118ff; Weigand, 2014). In this way, rhetoric can be conceived as a holistic communicative interaction featuring human interlocutors. In this chapter, I have reviewed the implications of the MGM of language for rhetoric. Clearly, from the frame of the MGM, the study of rhetoric cannot be divided into distinct subcategories like persuasion or argumentation. Instead, the MGM encapsulates both negotiation and reason into a coherent theory of dialogue that considers interlocutors as central to understanding meaning. Without doubt, this only makes clearer the necessity to understand culture, ideology, and context as it relates to speakers and audience in any given rhetorical act. Furthermore, I have considered the implications of the MGM for understanding audience, which must move into a completely dialogic view in contrast to the limited monologic understandings that proliferate in the literature even today.

References Adams J. (2012). Book review of the book Dialogue: The Mixed Game by Edda Weigand. Journal of Language and Social Psychology 31(2), 226–229. Aristotle. (2007). On Rhetoric: A Theory of Civic Discourse, 2nd edition. Trans. Kennedy G., New York: Oxford University Press. Bakhtin M. (1981). Discourse in the novel. In: Holquist M. (ed. & trans.), The Dialogic Imagination: Four Essays. Austin, TX: University of Texas Press. Burke K. (1950). A Grammar of Rhetoric. Berkeley, CA: University of California Press. Campbell K. K. and Huxman S. S. (2003). The Rhetorical Act: Thinking, Speaking and Writing Critically, 3rd edition. Belmont, CA: Wadsworth/Thomson Learning. Cicero, M. (2001). De Oratore. In Bizell, Herzberg B. (eds.) The Rhetorical Tradition: Readings from Classical Times to the Present. Boston, MA: St. Bedford, 326–341. Crosswhite J. (1989). Mood in argumentation: Heidegger and the Exordium. Philosophy and Rhetoric 22(1): 28–42. Floyd K. (2011). Interpersonal Communication, 2nd edition. New York: McGraw Hill. Foss S., Foss K., and Trapp R. (1991). Contemporary Perspectives on Rhetoric. Prospect Heights, IL: Waveland. Halasek K. (1998). Starting the dialogue: What can we do about Bakhtin’s ambivalence toward rhetoric? In Farmer F. (ed.) Landmark Essays on Bakhtin, Rhetoric and Writing. Mahwah, NJ: Hermagoras, 97–105. Howell J. (2010). Not just crazy: An explanation for the resonance of the birther narrative. Communication Monographs 79(4): 428–447. Hyde M. (1984). Emotion and human communication: a rhetorical, scientific, and philosophical picture. Communication Quarterly 32(2): 120–132. Jasinski J. (1997). Heteroglossia, polyphony, and the Federalist Papers. Rhetoric Society Quarterly 27(1): 23–46. Lee I. (2009). Some conceptions of emotional appeal in rhetorical theory. Speech Monographs 6(1): 66–86. Lucas S. (2011). The Art of Public Speaking, 11th edition. New York: McGraw Hill. McCroskey J. C. and Young T. J. (1981). Ethos and credibility: The construct and its measurement after three decades. Central States Speech Journal 32(1): 24–34. McCroskey J. C. and Teven J. (1999). Goodwill: A reexamination of the construct and its measurement. Communication Monographs 66(1): 93–103. Murphy J. M. (2001). Mikhail Bakhtin and the rhetorical tradition. Quarterly Journal of Speech 87(3): 259–277. Perelman C. and Olbrechts-Tyteca L. (1960). The New Rhetoric: A Treatise on Argumentation. South Bend, IN: Notre Dame.

305

Jennifer L. Adams

Perloff R. M. (2013). The Dynamics of Persuasion: Communication and Attitudes in the 21st Century. New York: Routledge. Richards I. A. and Ogden C. K. (1923). The Meaning of Meaning: A Study of the Influence Upon Thought and the Science of Symbolism. Reprint, New York: Mariner Books, 1988. Rosenthal P. I. (1966). The concept of ethos and the structure of persuasion. Speech Monographs 33(2): 114–126. Toulmin S. E. (1958). The Uses of Argument. Reprint, New York: Cambridge University Press, 2003. Walrod M. R. (2004). The role of emotion in normative discourse and persuasion. In: Weigand E. (ed.) Emotion in Dialogue: Advances in the Complex. Amsterdam, Netherlands: John Benjamins Publishing, 207–220. Weaver R. (1970). Language Is Sermonic: Richard M. Weaver on the Nature of Language. Reprint, Johannesen R., Strickland R. and Eubanks R. T. (eds.) Baton Rouge, LA: Louisiana State University Press, 1985. Weigand E. (2001). Games of power. In: Weigand E. and Dascal M. (eds.) Negotiation and Power in Dialogic Interaction. Amsterdam, Netherlands: John Benjamins Publishing, 63–76. Weigand E. (2006). Argumentation: The mixed game. Argumentation 20(1): 59–87. Weigand E. (2007). The sociobiology of language. In: Grein M. and Weigand E. (eds.) Dialogue and Culture. Amsterdam/Philadelphia: John Benjamins Publishing, 27–49. Weigand E. (2008). Rhetoric in the mixed game. In: Weigand E (ed.) Dialogue & Rhetoric. Amsterdam, Netherlands: John Benjamins Publishing, 3–21. Weigand E. (2009). Language as Dialogue: From Rules to Principles of Probability. Amsterdam, Netherlands: John Benjamins Publishing. Weigand E. (2010). Dialogue: The Mixed Game. Amsterdam: John Benjamins Publishing. Weigand E. (2014). Rationality of performance. Philosophia Scientae 18(3): 247–268. Woodard G. C. and Denton R. E. (2014). Persuasion and Influence in American Life, 7th edition. Prospect Heights, IL: Waveland.

306

19 Language as the Originative House of Dialogic Ethics Ronald C. Arnett

19.1 Introduction Acknowledging the dialogic, ethical, and social nature of linguistic life turns communicative inquiry from an informational and predictive activity to action attentively responsive to revelatory signs embodied within a priori linguistic social meaning that shapes the parameters, limits, and scope of the significance of face-to-face encounters. Linguistically, dialogic ethics renders two basic functions: (1) insight into what matters and its relative degree of importance, and (2) a call for responsive action. Together, alertness to importance and action are sign sensitive, with the form tied to sign significance and the latter generating a social pathway composed of touchstones of signification—dialogic ethics is linguistically a communicative action of attentive responsiveness. To explicate dialogic ethics as communicative actions of attentive responsiveness, I elucidate the intersection of ethics and linguist inquiry, framing the importance of ethical considerations within a pragmatic social linguistic tradition. This first section, “Ethics and Linguistic Inquiry”, provides a context for understanding creative insights within a linguistic tradition, attempting to do such research from an ethical perspective attentive to both macro- and micro-considerations. I then examine Weigand’s pioneering work on language and dialogue, ending the second section, “Ethics in Dialogue”, with a review of the early work by Craig and Tracy (1995) on grounded theory in discourse analysis. Their work offers a descriptive picture of how one can discern responses called forth by language embodied in contextual significance. The final section frames dialogic ethics as a discernment/response fulcrum upon which things that matter pivot into social significance. The final section, “Discerning What Matters: Linguistic Weight and Height,” frames dialogic ethics as the linguistically pragmatic key for understanding the interaction of Scout and a mob in To Kill a Mockingbird.1 The macro importance of this essay is a reminder of the socially embedded nature of what matters. Long before we begin to speak with another, signification is present, only to be contextually carried forth in conversation with another. This perspective on dialogic ethics has kinship with the work of Michel Foucault (1969/2002), who describes communicative interaction as situated within the a priori of a discourse background that informs the interpretive meaning of a given foreground discourse. This perspective in discourse 307

Ronald C. Arnett

analysis resides in the writing of François Cooren (2010, 2015), Linda Putnam (Fairhurst and Putnam, 2004; Putnam and Fairhurst, 2015), and Gail Fairhurst (2007, 2011). The distinctness of a dialogic ethics perspective is its reminder that within each discourse there is a tradition of what matters in a given context that shifts the social significance of a given discourse. Background and foreground are co-present, while an ethical alert pragmatically calls forth response. This essay outlines the linguistic reality of dialogic ethics as unceasing and happening prior, during, and/or after a given conversation in action—declaring what matters, while demanding responsive action.

19.2  Ethics and Linguistic Inquiry One can turn to a number of essays stressing the importance of linguistic weight and height tied to ethics (Arnett et al., 2014; Cooren, 2010; Taylor, 1989/2006). This perspective is central to Magdalena Kubanyiova’s (2008) “Rethinking Research Ethics in Contemporary Applied Linguistics: The Tension Between Macroethical and Microethical Perspectives in Situated Research.” Kubanyiova’s work exemplifies the shift from abstract linguist inquiry attentiveness to micro-social dimensions of ethical signification. Kubanyiova relies on two ethical frameworks discussed by Beth E. Haverkamp (2005): an ethics of virtue and an ethics of care. An ethics of virtue assumes that signification of importance rests within a public domain. This perspective contrasts with personal ethics, termed values. Public ethics are virtues called forth by agreement or rebellion, while registering meaning within a social group. An ethics of care is relationally attentive; what matters takes on relational consequences between persons. An ethics of virtue has public/macro importance and an ethics of care has relational/micro significance. Kubanyiova asserts that both the macro and the micro perspectives are essential guides for doing linguistic inquiry that has ethical stature. Her contention is that the macro without micro considerations transfers into an unethical realm that is abstract and inconsiderate of the reality of real persons within a distinct communicative environment. Kubanyiova asserts that ethics is far too often a low priority in the actual doing of applied linguistic research. She contends that as linguistic exploration increasingly responds to questions of contextualized signification, queries about ethics must become a commonplace necessity. Kubanyiova’s stress on ethics is twofold: (1) without ethical discernment, the research results are contextually dubious; and (2) with this research recognition of contextual ethics, one understands the fundamental nature of ethics in applied linguistics as tied to the social and the contextual. Responding to the linguistic implications of a given context requires navigation about what is and is not important, which propels research into the interactive realm of macro- and micro-ethics. Kubanyiova’s analysis of macro- and micro-implications of ethics in applied linguistics renders appropriate attention to contextualized environments. Kubanyiova’s (2008: 505) essay shifts sole consideration of a “bird’s eye perspective” or “macro-ethics” to ethical considerations responsive to a particular place, time, and persons; micro-ethics is responsive action that assists applied linguistic research. From the vantage point of an ethics of virtue, Kubanyiova (2008: 505) outlines three macro-principles that function as procedural ethics in applied linguistic research and operate as fundamental considerations for Institutional Review Board (IRB) committees: 1. Respect for persons. 2. Beneficence that yields good, not harm. 3. Justice requiring a fair dissemination of the results of the inquiry. 308



The Originative House of Dialogic Ethics

These three macro-ethics require an ethics of care responsive to the micro or the particular in action. Kubanyiova (2008: 506) states that the above three principles may contrast with a utilitarian perspective of the “greater good”. To address this macro-ethics conflict, one must explore the micro-ethical implications of applied linguistic research that necessarily must be responsive to “ethically important moments” (Kybanyiova, 2008: 506). An ethics of care incorporates the relational bonds between the researcher and those who are part of a given study. The macro-principles of such inquiry find support from a particular/ micro-ethical alert discerned when a researcher feels a specific discomfort about a given aspect of the inquiry. Both the initial and the later research responses witness an ethics of virtue in action. Ethically significant moments rely upon micro-insights in the doing of applied linguistic research, finding identity through the interplay of an ethics of care and an ethics of virtue. Within the context of applied linguistic inquiry, Kubanyiova (2008: 508) uncovers three critical ethical episodes: research design, research treatment, and research relationships. Situating applied linguistic research in ethics requires attentiveness to the particular participants; the researcher refuses to permit an abstract research design to eclipse the distinct concerns of those involved in the study. Macro-engagement of research design necessitates micro-adjustments within the context of implementing a given applied linguistic investigative environment. Kubanyiova’s (2008: 515) appraisal of high quality research involves methodological rigor, epistemological soundness, and constant adjustment to “micro-­ethical priorities” within a specific context of enquiry. Research treatment involves a natural conflict between macro- and micro-ethical coordinates in research. This tension permits an invitational perspective of research participant responses to morph into a perceived invasion into another’s private world. For example, requesting participants to employ self-reflection during the inquiry becomes an offensive act in the eyes of some subjects, requiring the researcher to reject an initial research treatment expectation and make adjustments. As stated, a supposed ethical relationship framed by openness of self-reflection can unwittingly invite coercion when there is a perception that imposition of disclosure guides the interaction. Research treatment requires macro- and micro-ethical responsibilities, propelled by a “heightened sensitivity” to conflicts between abstract and particular research treatments (Kybanyiova, 2008: 512); such research requires attentiveness to macro- and micro-concerns for ethical coherence in design, treatment and relationship development with participants. Kubanyiova’s proposal for ethics in linguistics inquiry suggests a framework akin to dialogic ethics: 1. One must engage dialogically with research design, treatment, and relationships. 2. One must dialogically attend to the distinctiveness of a research context. 3. One must acknowledge that a research context and its participants constitute a dialogic realm of inquiry that assumes linguistic height about what is and is not ethical in a given place.

The imposition of non-contextual ethics in research ignores a linguistic inquiry dialogue about what matters before, during, and after the research experience. Kubanyiova reminds us about the subtleties of ethics—abstract good intentions can lead to the perception of imposition; such a realization calls for discerning an increasing commitment to holistic inquiry in linguistics, which centers the work of Edda Weigand (2011) and Robert T. Craig and Karen Tracy (1995). 309

Ronald C. Arnett

19.3  Ethics in Dialogue I begin with Weigand’s sensible discernment into everyday human communication, as she shifts language study from the empirical to the abstractly distant, moving into the realm of the pragmatic and the concrete. Weigand elucidates the manner in which human beings linguistically discover what is and is not of importance, which then calls forth response and action. Weigand (2011: 545) articulates a fundamental shift from reductionist to a holistic conception of linguistic inquiry. Weigand understands the difficulty in defining language within the discipline of linguistics, which takes her to the notion of “generative linguistics”, which offers insight into the process of semiosis within sign systems. Weigand (2011: 545) stresses “integration” as a way of uniting differing perspectives. She cites Edward O. Wilson’s use of integrative terms “sociobiology” and “epigenetic rules” as an example of integrative thinking that yokes biology and environment (Weigand, 2011: 545). The world no longer can be understood from a provincial standard of a single discipline; we live in a time of “infinite mix of parameters” (Weigand, 2011: 546). A holistic linguistic theory must address changes in understanding and explore “competency in performance” (Weigand, 2011: 546). Weigand (2011: 546) demonstrates this orientation in the Model of Dialogic Action Games or the Mixed Game Model (MGM). Her model offers a holistic architecture that embraces the complexity of linguistics in action, underscoring the performative characteristics of linguistic competence in action, which are ever attentive to complex objects resistant to compartmentalized into parts. The multifaceted whole is greater than the sum of interactions; components are only recognizable via their specific function within the performative action of the unabridged event or experience. This conception of “competence-in-performance” is a reflection of the extraordinary activity within the human brain. Weigand differentiates three approaches to competence-in-performance: “human nature, culture, and the external environment not created by human beings” (Weigand, 2011: 546). These distinctive and simultaneously interactive dimensions of human life shape “competence in performance” (Weigand, 2011: 546). Weigand then articulates the premises that propel each of these determiners in action. The human nature perception assumes that human beings are the pivotal reference point for reasoning; the world is only as big as the scope of human abilities in their reach and understanding of the temporal limits of existence. Human abilities emerge through the interacting of integrated abilities. These capacities arise in a social world where goals and persuasion shape engagement. The second determiner of competence in performance is culture. Weigand cites “epigenetic rules” where culture is not merely external, but over time becomes internalized and shapes human behavior. Humans are social creatures with culture evolving in communal and external interaction, giving rise to a “system of values and habits” (Weigand, 2011: 546). The third component of competence in performance emerges as humans deal with an external environment, not of their own making. Humans must adapt; they begin with routine cases, which give insight into the particular. Specialization emerges as one gradually becomes more proficient in negotiating a given external subsystem. The model of Dialogic Action Games addresses complex objects propelled by the three singular determiners of competence in performance (Weigand, 2011: 546); together these determiners permit one to navigate uncertainties as one gathers insights that suggest probabilities that can direct action. A linguistic methodology responsive to complex objects in a holistic fashion must emphasize probabilities. A holistic linguistic theory contrasts with a conventional theory based upon rules. Weigand (2011: 547) stresses the value of a 310



The Originative House of Dialogic Ethics

“coherence principle,” as she differentiates from a rule-base concept of theory. She then describes three principles: 1. The “constitutive,” which weaves human interaction. 2. The “regulative,” which mediates self-interest and social life. 3. The “executive,” which frames the importance/significance behind actions and discourse (Weigand, 2011: 547).

These principles arise from a basic premise that “our abilities are integrated” (Weigand, 2011: 547) and require constitutive principles. Weigand emphasizes two major constitutive principles of integration: action and the dialogic. Actions can be communicative, mental, or practical. Dialogue unites action and reaction, which are central to the dialogic nature of each speech act. Communication does not originate from autonomous speech acts, but from speech acts in dialogue that give rise to a collective sense of meaning. This dialogic principle assumes that speech acts are in meaningful association and connection with other speech acts. From a linguistic orientation, an initiative speech act is not an act of monologue, but rather a claim to truth that requires affirmation or rejection. Action initiates reaction, which opens the door to dialogue and understanding. However, one cannot equate action with meaning and reaction with understanding; dialogue goes beyond this correlation in intersubjective meeting of interactive engagement. Dialogic speech act taxonomy includes intersubjective understanding that begins with action that functions as a fundamental claim for truth and/or a claim for volition that necessitates response. Reaction fulfills the claim of action. There are claims and sub-claims or “assertions and statements but also presumptions and suggestions” (Weigand, 2011: 548). Sub-claims undergird multiple speech acts; they permit linguistic clarity to emerge in an integrative fashion with multiple claims framing identity for a particular claim. This dialogic model goes beyond traditional linguistics and speech act theory; it incorporates multiple claims and responses as integrative to dialogic identity of a complex object. For example, from this perspective, a phrase such as “Toyota does not break down” is not a functional monological statement, but rather an action claim that calls forth reaction (Weigand, 2011: 547). “The Dialogic Principle proper of action and reaction represents the basic unit of language use” (Weigand, 2011: 548). This conception of a dialogic principle manifests itself in action—“initiative action” and “reactive action” (Weigand, 2011: 548). This view of action is different from what J. L. Austin and John Searle called “illocution” (the real-life consequences of speech such as what occurs with a marriage announcement) and “locution” (the vocalizing of speech) in their Speech Act Theory (Weigand, 2011: 549). Weigand’s analysis positions dialogue as either an initiative action that prompts responsive action. She deems Austin and Searle’s distinction between illocution and locution as an important first step, but not sufficient in recognizing “the communicative means” that linguistically “negotiate meaning and understanding” (Weigand, 2011: 549). Weigand’s (2011: 549) model suggests that each action, whether practical, verbal, or communicative, finds definition via “correlation of some purpose and corresponding means,” suggesting the importance of a Principle of Coherence. Weigand concludes with an overview of the linguistic landscape in the twentyfirst century; she contends that holistic reactions to complex objects must guide contemporary linguistic inquiry. Her dialogic model offers insight through the “tracing” of linguistic practices in everyday human activity. Formal scientific rigor cannot render a concrete picture of human behavior. Object language does not live in an autonomous and abstract space; 311

Ronald C. Arnett

it dwells in an ongoing mixed dialogic game of action and reaction of persons within social arenas. Dialogic Action games emphasize multiplicity of sub-claims that interact with a particular claim at the forefront of discussion; action evokes reaction. Dialogue is central to understanding a complex object. Dialogue is the natural manner in which complex objects invite understanding in the give and take of action and reaction. Weigand (2015) continues this line of inquiry in “Identity as a Dialogic Concept,” where she stresses the notion of complexity when the object under examination is society, itself. She states that identity in a society finds structure in bureaucratic systems. The complex construct of identity on a personal level dwells within the performance of “a dynamic sequence of events” (Weigand, 2015: 7). A complex understanding of identity requires linguistic examination from holistic theory. Socially, identities emerge through “communities of practice” (Weigand, 2015: 9). Weigand explores the manner in which identity finds construction in language use. Dialogue is the pragmatic communicative feature that permits multiple identities to emerge in public communication. Weigand’s (2010) book, Dialogue: The Mixed Game, announces dialogue as a holistic framing that integrates multiple constituent parts in discourse. Dialogue emerges in action and reaction, never limited to immediate language alone; it involves the meeting and clashing of opinions, which constitute interests and reflect cultural feeling and frame the expression of identity in demonstrations/public displays that can fuel arguments over identity. In each case of linguistic consideration of identity, dialogue involves a multiplicity of speech acts that form a horizon of meaningfulness. The holistic gathering of perspectives renders a social and/or personal identity reshaped with further dialogue that enriches the complex of identity differently. Weigand’s position finds additive support through the insights of Roger D. Sell (2011), as he links her dialogic model to the notion of ethics. He underscores a position on dialogue that is otherwise than face-to-face and turn-taking conceptions of the concept. He stresses Weigand’s linguistic dialogic understanding tied to linguistics, emphasizing dialogue that is present long before specific interlocutors enter the arena of conversation. Sell follows the action framework of dialogue and examines ethics tied to address and response. He points to dialogue dependent upon recognizing linguistic importance present within the language engaged. The dialogue of address and response has inherent ethical implications. Ethics enters as one is increasingly attentive to address and response. One must listen to and acknowledge dialogic richness within language as multiple signs giving credence to particular signs; dialogue before a given meeting makes signs visible prior to a conversation gathering. Discerning what matters has ethical consequences in that it directs one’s attention in one manner and/or location at the exclusion of another. Dialogue between persons necessitates respect for the other to discern what is of importance; such attentiveness requires respect for the textured nuances present within a linguistic sign. Language and language use is inherently dialogic with attentiveness to what matters as an ethical sign. Sell illustrates this conception of dialogic ethics in his analysis of the writer-reader relationship, which he understands as fundamentally a dialogic activity. Sell separates his work from Hans-Georg Gadamer (1900–2002); Sell contends that Gadamer is too attentive to the aesthetic dimension and misses the ethical implications of literature. As an alternative, Sell (2012) outlines the dialogic contribution of Mikhail Bakhtin (1895–1975). According to Bakhtin, each text involves “heteroglossia” or “varied-tonguedness” in which linguistic contributions within a text announce difference and tension (Sell, 2012: 13). The perceived differences are akin to a carnival of multiple perspectives gathered together with ongoing noise and activity that provides the performative function of letting off steam; yet, according 312



The Originative House of Dialogic Ethics

to Sell, not all literary works are embedded in sociopolitical implications. Sell (2011: 85) rejects that assumption that “distorted communication” is the only outcome in literary communication. He returns to dialogue between writer and reader, stressing a dual fact: human beings are social beings and simultaneously require independence of imagination, intellect, and will. The interplay of the social and independence of action makes the task of writing for an implied reader a meeting of historical plurality, not a single or specific audience. The autonomy of the individual results in multiplicity of perspectives; Sell (2011) stresses the ethical limitations of literature when it imposes upon a reader by insisting upon one design alone. Great writers have ethical respect for the reader, permitting engagement with the text via what John Keats (1795–1821) termed as “negative capability” (Sell, 2011: 91), which advocates mystery and uncertainty. Such writing has ethical implications as the reader walks into the unknown; this interpretive journey announces ethical height within the writer-reader relationship. The ethical dimension of negative capability decrees an increasing importance of discernment within writing that is attentive to both personal dimensions and historical representativeness. Respect for a historical moment and persons void of imposition propels action resistive to prescribed response. Sell (2011) claims that Winston Churchill (1874–1965) was an exemplar of responsiveness to the historical moment as he engaged it from a personal standpoint. In 1953, Churchill received the Nobel Prize for literature with many wondering why. For Sell, the reasons for the award rest in Churchill’s writing of negative capability and self-irony in which he mediated between “the values of an already distant ‘then’ and those of an urgent ‘now,’ he takes for granted no status quo can ever be fully restored” (Sell, 2011: 93). In contrast, Sell states that the poetry of William Wordsworth (1790–1850) bullies the reader into a given position. Sell cites Keats’s assertion that poetry with “palpable design” limits imagination and, in the words of Sell (2011: 91), invites the unethical. Sell states that Wordsworth’s best writing uplifts the ethical with a dialogue between the personal and the historical moment within which negative capability guides without offering undue assurance. Sell also cited Charles Dickens (1812–1870) as an exemplar of this dialogic mode of engagement that embraces the personal and the historical moment without dictating what another should think. The writing engages dialogue and, in addition, the writer and reader continue the enterprise of discovering insights. Negative capability dwells in demands of the historical moment and the personal with a reluctance to limit the choices of another, no matter how important the task. Like Shakespeare taking us to a world of “wound,” negative capability reminds us of a reality that does not attempt to “sew up or re-shape” the lesion with the artificial use of clarity and direction (Sell, 2011: 98). Art offers force of perspective without palpable design, respecting the autonomy of other’s independence and intelligence. Sell’s insights bid another entrance into Weigand’s model of dialogic action, suggesting that linguistic life is contextual, historical, and personal; such a perspective rejects abstract certainty that imposes fictional assurance. Yet, another way of detailing the interplay between linguistic life, dialogue, and ethics dwells within Robert T. Craig and Karen Tracy’s description of grounded theory (1995). This essay continues to offer pragmatic power for the linguistic and discursive study of ethical implications in conversation.

19.4  Linguistic Contextual Weight as Call of Responsive Action Together, Weigand and Sell stress the pragmatic reality of dialogue before, during, and after the awakening of signs of importance that evoke subsequent action/response. Language as signs of contextual importance yields responsive action, as humans engage in an unending 313

Ronald C. Arnett

series of judgments about what matters. Craig and Tracy (1995: 253) connect theory and practice through what they term “grounded practical theory.” They unite normative theory and practical disciplines that yield grounded practical theory as a mode of problem-centered inquiry. Normative theory employs what might be, replacing an empirical “is” with an “ought” (Craig and Tracy, 1995: 249). Craig and Tracy place dialogue within normative theory and then submit that dialogue must address practical problems to assist communication theory in action. They consider communication a practical discipline with an obligation to address genuine problems; communication inquiry requires a commitment to usefulness, which necessitates grounding theory within practices. Craig and Tracy (1995: 253) offer three ways to conceptualize a practice; they texture the linguistic implications of term, practice. First, on the “technical level”, a practice is a gathering in action of “communication strategies and techniques”. Second, on the “problem level”, a practice is a “strategic response” to a given dilemma (Craig and Tracy, 1995: 253). Third, on the “philosophical level”, a practice is the carrier of normative ideals (Craig and Tracy, 1995: 253). Grounded practical theory reconstructs communication dilemmas via reflection on communication techniques, strategic response, and normative ideas that constituted a practice or set of practices that emerge in discourse. Craig and Tracy offer a contribution to communication in their description of grounded practical theory that is linguistically attentive in everyday understanding of discourse. They stress the importance of ongoing reflection in the doing of conversational analysis through examination of dissimilar communicative practices composed of distinctive problems, techniques, and ideals that shape communication practices. Craig and Tracy do not propose a social science theory of prediction, but rather a reflective theory that reconstructs practices within a discourse to facilitate understanding of differences. Grounded practical theory unites moral argument and empirical observation; this objective is distinct from much discourse analysis that does not emphasize the moral dimension of communication. Craig and Tracy contend that grounded practical theory in action examines questions of acceptance and rejection of the legitimacy of a given normative theory. Practical judgments in grounded theory examine discourses with contrasting communication techniques, strategies, and ideals in rendering adjudication on the problem of legitimacy. Craig and Tracy detail communication research as objectively charged with understanding differences that carry ethically divergent forms of insight into what matters. Such research opens the door to reflection and understanding of communication dilemmas that publicly emerge in conversation. Such discoveries often emerge in punctuations of exasperated utterances of “Why? How could this be?” and “Who could think like that?” These punctuated moments assist in understanding the normative dimensions of the conversation, and they announce a practical dialogic ethic capable of illuminating moments defined as quandaries resistive to clarity of answer. Craig and Tracy ground practical theory as methodological investigation into dialogic understanding about what matters in a given contexts defined as a quandary. They point to communicative junctures that suggest a dialogic ethic.

19.5  Discerning What Matters: Linguistic Weight and Height The work of Weigand (2010, 2011, 2015), Kubanyiova (2008), Sell (2011, 2012), and Craig and Tracy (1995) is clearly distinct; yet, they share a common linguistic objective—investigation and understanding of a linguistic world that is contextual, personal, historical, and social within an objective of inviting temporal holistic insight into quandaries within discourse. Their scholarship yields argument over what matters, while revealing the pragmatic 314



The Originative House of Dialogic Ethics

reality of dialogue in action. Language matters; it houses ethics in dialogic tension in the meeting and understanding of how and what matters. Dialogic ethics assumes that linguistic pluralism lives within the very notion of dialogic ethic that is historical, contextual, personal, and social in its ability to offer temporal insight. To illustrate the temporal linguistic power of dialogic ethical clarity, I review the discourse that constitutes the scene in To Kill a Mockingbird between Scout and a “hanging mob.” Harper Lee’s (1960/2006) world-renowned work, To Kill a Mockingbird, continues to generate attention (Murphy, 2010). She received the Pulitzer Prize in 1961 and now another version of the manuscript has surfaced, igniting more conversation with the publication of Go Set a Watchman (Lee, 2015). In Lee’s (1960/2006) notable manuscript, in To Kill a Mockingbird, the exchange between Scout and an angry mob elucidates the power of linguistic weight of dialogic ethics in action and response. Scout at the young age of six lives in Maycomb, Alabama, with her older brother Jem and her widowed father, Atticus Finch. Atticus is a town attorney who defends Tom Robinson, an African American man charged with rape. A mob of local men emerges in contention with Atticus’s controversial defense of Robinson. This major scene in the novel begins with Tom Robinson being moved to the Maycomb jail. The narrator, Scout, reminds us that this is Sunday, the day of “fake peace” (Lee, 1960/2006: 168). Shooting is not lawful on Sunday. That evening as many were going to sleep Atticus Finch was at the Maycomb jail; he sat with his feet propped up against the door of the jail, which many considered the most “hideous” structure in the county (Lee, 1960/2006: 170). The jail had no light on its front door; Atticus required an extension cord to light a bare bulb for reading of the newspaper. In the shadows of the building, one observed men getting out of their cars and beginning to gather. Then a voice protruded from the darkness, “‘He in there, Mr. Finch?’” (Lee, 1960/2006: 171), referring to Tom Robinson. Finch stated, yes, while reminding people that Robinson was asleep. As men gathered with the smell of whiskey on their breath, Scout came running to her father’s side, only to find him displeased by her presence. To the group, Atticus stated that they should go home. As the men began to get out of hand, Scout attempted to kick one, only to miss and swing her leg into the air. Scout noticed that in the shadows the men wore clothing with hats over their ears that concealed their identity. She looked for familiar faces. Then the power of language in familial terms connects Scout to Mr. Cunningham. Scout recounts the following: 1. “Don’t you remember me, Mr. Cunningham? I’m Jean Louise Finch [Scout].” (Lee, 1960/2006: 174). 2. “I go to school with Walter … He’s your boy, ain’t he? Ain’t he sir?” (Lee, 1960/ 2006: 174). 3. “He’s in my grade … He’s a good boy …” (Lee, 1960/2006: 174).

At that point, Mr. Cunningham has no response and Scout continues to exert much effort at making him “feel at home” (Lee, 1960/2006: 174). 4. Scout then said, “Entailments are bad,” as she found herself addressing the entire group of men (Lee, 1960/2006: 174). 5. Then a linguistic miracle happens, as we witness dialogic ethics in action and response. Mr. Cunningham squats down and takes Scout by her shoulders. He then said: “I’ll tell him you said hey, little lady …” [he then called out] “Let’s clear out … Let’s get going, boys” (Lee, 1960/2006: 174–175). 315

Ronald C. Arnett

6. Finally, Atticus turns off the light and picks up his chair with Dill, Scout and Jem’s neighborhood friend, asking if he can carry it. Scout wondered if Atticus would be angry because they did not go home when Atticus first asked, only to watch her father reach out and massage Jem’s hair—“his one gesture of affection” (Lee, 1960/2006: 176).

The crowd had dispersed, and it was time for sleep and rest … . Scout and this experience offer evidence to the power of dialogic ethics lived in initiation and response with macro-reminders about the importance of a home connection. Dialogic ethics does not live in abstract, but in the concrete experience of humans engaged together. Weigand’s (2010) book, Dialogue: The Mixed Game, reveals dialogue in its holistic trappings with dialogic ethics emerging in action and reaction, never limited to the immediate language alone, as gestures from Mr. Cunningham and Atticus suggest. Dialogic ethics naturally embodies clashing of opinions reflective of different interests and cultural feelings. Dialogic ethics gives us identity, found through the complex everyday life of action and response that constitute an identity that composes a horizon of meaningfulness. As Weigand and others remind us, dialogue is a holistic gathering of perspectives, shaped by social and personal identity in a given historical context, ever engaged in a constant reshaping of a complex of identity with each meeting and response. Dialogic ethics is linguistically the weight and height that we discover in the ongoing engagement of dialogue in action and response. Words carry signification. Gestures carry signification. Language houses what matters in verbal and nonverbal form, telling us in action and response via dialogic ethics of importance—such as, Mr. Cunningham, I know your boy and he is a good boy… and then the men within the shadows stood upright for a moment and returned to their homes.

Note 1 This scene appears in Chapter 15 of Harper Lee’s 1960 classic novel, To Kill a Mockingbird.

References Arnett R. C., McManus L. M. B. and McKendree A. G. (2014) Conflict Between Persons: The Origins of Leadership. Dubuque, IA: Kendall Hunt Publishing Company. Cooren F. (2010). Action and Agency in Dialogue: Passion, Incarnation, and Ventriloquism. Amsterdam, Netherlands: John Benjamins Publishing Company. Cooren F. (2015). Organizational Discourse: Communication and Constitution. Cambridge, UK: Polity. Craig R. T. and Tracy K. (1995). Grounded practical theory: The case of intellectual discussion. Communication Theory 5(3): 248–272. Fairhurst G. T. (2007). Discursive Leadership: In Conversation with Leadership Psychology. Thousand Oaks, CA: Sage Publications. Fairhurst G. T. (2011). The Power of Framing: Creating the Language of Leadership. San Francisco, CA: John Wiley & Sons. Fairhurst G. T. and Putnam L. (2004). Organizations as discursive constructions. Communication Theory 14(1): 5–26. Foucault M. (2002). The Archaeology of Knowledge. New York: Routledge. First published 1969. Haverkamp B. E. (2005). Ethical perspectives on qualitative research in applied psychology. Journal of Counseling Psychology 52(2): 146–155. Kubanyiova M. (2008). Rethinking research ethics in contemporary applied linguistics: The tension between macroethical and microethical perspectives in situated research. The Modern Language Journal 92(vi): 503–518. 316



The Originative House of Dialogic Ethics

Lee H. (2006). To Kill a Mockingbird. New York: HarperCollins. First published 1960. Lee H. (2015). Go Set a Watchman. New York: HarperCollins. Murphy M. M. (2010). Scout, Atticus, and Boo: A Celebration of Fifty Years of  To Kill a Mockingbird. New York: HarperCollins. Putnam L. and Fairhurst G. T. (2015). Revisiting “Organizations as Discursive Constructions”: 10 years later. Communication Theory 25(4): 375–392. Sell R. D. (2011). Dialogicality and ethics: Four cases of literary address. Language and Dialogue 1(1): 79–104. Sell R. D. (2012). Introduction. In: Sell R. D. (ed.) Literary Community-Making: The Dialogicality of English Texts from the Seventeenth Century to the Present. Amsterdam, Netherlands: John Benjamins Publishing Company, 1–16. Taylor C. (2006). Sources of the Self: The Making of the Modern Identity. Cambridge, UK: Cambridge University Press. First published 1989. Weigand E. (2010). Dialogue: The Mixed Game. Amsterdam, Netherlands: John Benjamins Publishing Company. Weigand E. (2011). Paradigm changes in linguistic: From reductionism to holism. Language Studies 33: 544–549. Weigand E. (2015). Identity as a dialogic concept. Language and Dialogue 5(1): 7–22. Wilson E. O. (1975). Sociobiology: The New Synthesis. Cambridge, UK: Belknap Press of Harvard University Press.

317

20 Dialogue in Institutions Sebastian Feller

20.1 Introduction We live in an increasingly complex world. Information continuously hails down on us in an almost uncontrollable stream. To cope with the complexity of our modern-day world, special forms of societal organization have emerged over time (Armitage et al., 2008). The fastpaced growth of society and the increasingly complex nature of the activities and knowledge therein have led to the formation of special groups of experts with specific knowledge in and about selected parts of society (e.g. Room, 2011). In short, we are talking about the rise of “institutions”. Institutions are cultural entities with specific goals and a legal or, at least, traditional foundation in society (Margolis, 2009; Rogoff, 2003). The general goal of any institution is to establish order. Institutions form distinct sub-collectives of society, each with a particular specialization in selected traits and operations. Law, for example, regulates social conduct, while medicine is about treating and curing disease. Institutions also operate under certain value and belief systems and adhere to selected conventions and habits. In a way of speaking, one could say that they possess their own world view or ideology. Law, for example, operates from a cause-and-effect-point of view where intentions and actions are considered to cause a particular set of consequences or circumstances. For example, if someone takes somebody else’s property without permission to do so, he or she is committing theft. Religion, on the other hand, is based on the belief in a higher power and grounds in notions that usually lie beyond the constraints of logic of the secular world. Despite these differences, both law and religion share the same overall institutional goal: they regulate people’s conduct. However, most important for this chapter is the fact that institutions do communicate (e.g. Hardy, 2011; Lammers, 2011). Institutional communication is enacted by and through the people that are part of them. As we will get to understand throughout this chapter, institutional communication is necessary with a view to achieving the institutional goals and purposes. In a way of speaking, communication is the fuel that keeps the institutional engine running.

318



Dialogue in Institutions

This said, it seems little surprising that coming to a better understanding of why and how people communicate within institutions is key to arriving at a better understanding of the institution itself. Next to the institutional goals per se, communication is one of the major components of any institution. It is through and by communication that the institution creates and maintains an identity (Schnurr, 2012: 110). The what for and how of communication reflect the institutional structure including things like the roles of its members as well as the activities and processes the members engage in. For example, how a person understands her role and how it is understood by the other members in an institutional setting is, next to the legal demarcation of her role, largely reflected by the communication (Schnurr, 2012: 109ff.). To give an example, a judge is addressed as Your Honor in court, a surgeon Doctor in the operating room, and a military commander Sir when in action. The forms of address already signal the status and power that their bearers possess. Along similar lines, the communication gives life to the particular activities and processes that are characteristic for a particular institution (Knill and Lenschow, 2005). It is the “what for” and the “how” of the communication that reflect how things are done. Complex processes like decision making, stocktaking, planning, and creating assurance, which make up the main components of institutional operations, are also shaped by how the institutional members communicate (e.g. Weigand, 2010: 129ff.). To return to our previous example, in law hierarchy is of paramount importance. The institutional hierarchy is, next to the forms of address, reflected by the turn taking behavior and the types of speech acts the interactants use in communication. Someone high up in the hierarchy might take the floor without much ado, while someone of lower status or privilege usually has to wait a longer time before they can add to the communication. In an operating room, for example, it is usually the surgeon who guides and regulates the communication. Other staff such as a nurse, for instance, have to follow the surgeon’s lead and ought to speak up only if permissioned to do so. Obviously, processes like decision making are done in very specific ways by selected institutional members in accordance with institution-specific communication protocols. The strong link between the institutional communication and the institution-specific processes is further highlighted by a contrastive view. If we compare the communication in hierarchical institutions like law and medicine to what we can find in the business world, for example, we can immediately make out the difference: compared to medicine and law, many organizations in the business world operate in far more democratic and participatory environments. People think and communicate in relatively unrestricted ways to make a profit (Sopazi, 2014). Stiff turn-taking protocols and rigid forms of address would rather be in the way. As a consequence, the communication becomes more flexible overall. Before we go further into medias res, let us take a look at how the remainder of the chapter is organized: Following the introduction, Section 20.2 presents the basics of Weigand’s action game theory, which she calls the Mixed Game Model (e.g. 2000, 2010). Similar to Weigand, I consider the dialogic action game as the basic unit of communication in general and, consequently, as a basic unit of communication in the institutional context. The minimal form of the action game is a dialogic exchange between dialog partners consisting of the communicative action and the communicative reaction. Action and reaction include speech acts like requesting information, making a statement about the world, asking somebody to carry out a practical action, as well as complying and agreeing, among others. Starting off from the action game model, I classify institutional communication into distinct types. The classification is done on the basis of a type continuum with the two end points “impersonal/ 319

Sebastian Feller

pre-defined/formal” (Guffy, Rhoddes, and Rogin, 2005) and “personal/undefined/informal” (Ogaard, Marnburg, and Larsen, 2008: 2). The main questions for the classification task are: what are the main institutional goals and the intentions and interests that drive them and how do the goals impact the institutional communication with respect to the institution’s position on the type continuum? Section 20.3 discusses the main functions of institutional communication. The leading questions here are: why do the members of an institution communicate in the first place? And: what are the communicative purposes behind institutional communication? In Section 20.4, I shift the focus to institutional knowledge and the institutional culture. Here, I apply Hall’s (1976) framework of low-context versus high-context cultures. My argument is that the nature of the institutional knowledge as well as the position of the institutional culture on the low-context versus high-context axis determines the way and extent the knowledge is communicated within and across the institution. Section 20.5 condenses the previous sections into an action-theoretical framework for institution-specific communication on the basis of Weigand’s action game model. The framework presented here integrates institution-internal factors like the institutional structure, the institutional knowledge, and the institution-specific communication channels as well as culture as an influence factor both on the in- and outside of the institution. Section 20.6 sheds additional light on culture as an integrated component of institutional action and communication. In this context, I refer to Rapaille’s (2007) culture code with a view to arriving at a better understanding of the culture-specific mindsets that influence how the institutional members act and communicate. Finally, Section 20.7 concludes the chapter and provides a brief outlook on potential directions for future research.

20.2  Institutional Action Games: Formal and Informal Communication Describing and understanding institutional communication in a more detailed way requires a theoretical foundation that provides a set of basic assumptions. Following Weigand (e.g. 2000, 2003, 2010), I start off from the first such assumption that communication is always dialogic. In its minimal form, any communication thus consists of two communicative actions that are mutually related to each other: communicative action and reaction, or, to say it in a more linguistic parlor, the initiative and the reactive speech act. I, henceforth, take this minimal unit of communication or, as Weigand refers to it, the action game, as a minimal representation of institutional communication. Another important assumption that I share with Weigand (e.g. 2010: 6, 63ff., 270) is that the minimal action game is a cultural game. More precisely, dialog partners are always embedded in and act in the context of a larger cultural environment defined by culturespecific values, norms, and belief systems as well as traditions and habits. Culture naturally shapes the way the dialog partners interact and communicate with one another. While, for instance, in many Western cultures, speakers look their dialog partner straight in the eye, such behavior is usually considered offensive in the Eastern world. As for institutions, culture has a significant impact on important institutional practices including stocktaking, reporting and creating assurance, to name a few. From a general point of view, I furthermore hold that institutional communication lies on a type continuum between the two end points “impersonal/pre-defined/formal” and “personal/undefined/informal”. Impersonal language is usually marked by indirect expressions that are arranged along pre-defined and formal ways of use. Here, the speakers act along 320



Dialogue in Institutions

communicative patterns that are encoded in a relatively rigid communication protocol. Legal texts are an extreme case exemplifying this kind of communication. However, personal communication is normally comparatively loose and undefined, that is, the speakers make things up on the go and speak more freely instead of relying on pre-defined ways of speaking. Depending on the exact nature of the institution and the communicative situation, the institutional communication will lean toward one of these two end points. I hold that the exact nature of the institutional goals and the organizational structure of the institution are decisive factors in this matter. In the end, we are dealing here with a reciprocal relation between the institutional communication and the nature of the institution itself. Once again, a contrastive perspective helps to illustrate this point: while, for example, the main institutional goal of medicine is to save lives, the business world strives for profit. When it is about saving people’s lives, the margin of error should be as small as possible. As a consequence, medical staff communicate along pre-given, clear-cut communicative patterns that minimize misunderstandings and maximize communicative effectiveness. In contrast, the business world lives off innovation and creativity. These usually require open environments with as few restrictions to thinking and communicating as possible. In a business, people do not usually speak in pre-given phrases like medical staff do; in contrast, they are often required to break out of conventional thinking and talking with a view to arriving at innovative problem solutions that give them a chance to beat their competition. As mentioned, next to the goals, it is the institutional structure which impacts the institutional communication. For example, hierarchical institutions like the military usually apply communicative means that lie further to the “impersonal/pre-defined/formal” end of the type continuum. On the battlefield, specific, pre-given communicative patterns (e.g. the Dictionary of Military Terms, 2004) are in use ensuring that the command structure is enacted by all relevant staff. Again, room for interpretation means room for error, which may easily result in lost lives on the battle field. This kind of military parlor also shows in how members of the military ought to be addressed. For example, the correct form of address for military staff usually includes the rank. The communication reflects the status and military decorations of the dialog partner. In contrast, the modern business world is increasingly abandoning tall hierarchies. The work life is based on flat relations among staff. The communicative tone in such environments mirrors the fact that work is done without a great ado about status or rank of the people involved (e.g. Perry-Smith, 2006). Rigorous communication protocols would stand in the way of cultivating a work culture where incubating new ideas has become a top priority. As a consequence, people talk in a more personal and informal tone of voice. The contrast between the two end points on the type continuum, that is, the “impersonal/ pre-defined/formal” versus “personal/undefined/informal” type of communication, can be represented by the following break-down of institutional interests: a. gives a representation for medical communication, while b. stands for the business world: a. “impersonal/pre-defined/formal” b. “personal/undefined/informal”

(INTEREST (saving lives) [general goal: maintaining operations]) (INTEREST (making a profit) [general goal: maintaining operations])

While the general goal of any institution is to maintain operations, that is, its own and, subsequently, those of society, “interest” defines the institution-specific goal. By interest I 321

Sebastian Feller

mean a higher order goal. As for the given examples, saving lives is the institution-specific goal of medicine, while making a profit is what the business world is ultimately after.

20.3  The Functions of Institutional Communication In Section 20.2 I briefly touched on the institutional goals and interests as well as some selected aspects of the institutional structure like the hierarchy that govern the institutional communication. However, another important point is the immediate communicative purpose. In other words, we need to take a closer look at the communicative functions involved. The communicative functions are, at the same time, dependent on the particular institutional interests like saving lives or making a profit. As mentioned earlier, the main priority for institutional communication is keeping the institution up and running (cf. also Castells, 2013). Institutions usually consist of a large number of individuals who are often grouped into distinct sub-groups. Cross-group communication becomes mandatory for the institution to stay operational. For example, in law, we find judges, attorneys, and lawyers. In medicine, people are separated into groups like doctors, nurses, and patients, among others. All these different groups communicate within the larger framework of the institution. Furthermore, many institutions branch out into separate sub-institutions: In economic institutions like banks, for instance, we find staff that is specialized in investment banking, loans or foreign exchange trading, and so on. Medical institutions are usually separated into distinct departments such as cardiology, neurology, and dermatology. The different institutional groups must work together to carry out the institutional operations and ultimately fulfill the institution’s interests and goals. A trial in court is an example in point. The representatives of the distinct sub-groups like the prosecution, the defense, and the judge, come together in a shared effort with a view to deciding on the guilt or innocence of the accused. They weigh the evidence, construct arguments from a variety of standpoints and interview witnesses. Later, the jury and the judge combine the communications of all the sub-groups to the verdict. In the medical world, a similar picture emerges. For example, in a patient–doctor dialog the patient tells the doctor about her condition. After the medical examination, the doctor provides the diagnosis and a treatment plan for the patient. In addition, we find corollary dialogs such as, for instance, a doctor–nurse interaction, in which the doctor instructs the nurse to carry out selected medical tests and where the nurse provides test readings and other relevant information for the medical treatment. In both cases the importance of the cross-group communication is immediately evident: the distinct institutional sub-groups need to communicate with one another to function as a consistent whole. The institutional communication clearly has a linking function in these cases. It is via communication that the distinct sub-groups create coherence and meaning. The construction of coherence and meaning-making is related to another key concept: professionalization. The following department store metaphor helps us understand how professionalization works in this context: the things we can buy in a department store are arranged into distinct sections. The sections cater to different needs. We can buy sports equipment in the sports department, stationary in the office department, clothing in the clothes department, and cooking ware in the household department. Staff are distributed across the departments in accordance with their specific expertise. This kind of knowledge distribution helps optimize the provision of information and service; the customer can tap 322



Dialogue in Institutions

on profound expert knowledge in her field of interest simply by asking the staff in the respective department. In a similar fashion, the general institutional function of maintaining the institution’s operations breaks down into several sub-functions with more immediate institutional goals. These sub-functions include but are not limited to reporting, stock taking and decision making, to name a few (see also Weigand, 2010). Similar to the department store, where the different sections add up to the store as a whole, the sub-functions combine to the workings of the institution at large. The sub-functions are representable in term of Weigand’s dialogic action games (e.g. 2000, 2010). The dialogic action games that instantiate the sub-functions comprise speech acts like representative, that is, making a statement about the world, directive, in other words, expressing one’s will that someone else does something, as well as agreeing and complying on the reaction side of things. Weigand provides a detailed analysis of these action games and the speech acts involved in her books Dialogue: The Mixed Game (2010) and Sprache als Dialog (2003 [1989]). As mentioned earlier, institutions are a means to tackle the increasing complexity of the world. In this sense, professionalization is ultimately a question of practicality. Modern societies have grown into largely complex entities branching out into countless sub-communities. These sub-communities create their own worlds around their particular goals, value systems, and processes. Over time, it has become impossible to oversee and manage all the sub-­ communities in society that have emerged. Subsequently, experts who have specialized in the activities of specific societal fields have filled in the void and paved the way for the increasing professionalization that we witness in our modern-day societies. The institution’s specializing in a limited number of selected processes provides a working foundation on which society can operate and remain functional. Limited responsibilities and activities are assigned to selected actants, providing orientation and a sense of order for the world we live in. The question whether institutionalization is a natural or an evolutionary process, or whether it results from the cultural development of society is a question that has raised lots of controversy in recent years. Lumsden and Wilson’s (1985) work is seminal here. It opens up an integrative view that combines both worlds. The authors argue that culture and nature stand in a reciprocal relationship. In other words, both have their share in how we and the world, and therewith institutions, develop. Be that as it may, it is obvious that the breaking up of society into smaller operational units such as institutions is a practical necessity to facilitate our dealing with the complexities of modern-day life. It seems hardly thinkable that we could cope with the entirely overwhelming vastness of societal activities without institutions in place.

20.4  Institutional Knowledge Another essential aspect of the institution is the institutional knowledge. By institutional knowledge I mean all the bits and pieces of information that float around within the institution and which are relevant for its operations. This information includes but is not limited to knowledge about the institutional roles and functions, including institution-specific processes and procedures, as well as the institution-specific concepts and notions. Some of this information is shared across all institutional members such as, for example, the code of conduct forming a kind of common ground the institution operates on. Other information might be held by a selected group of members only. Classified information in the military is an example in point. 323

Sebastian Feller

Institutional knowledge spans an array of information of different sorts (cf. also Javernick-Will and Levitt, 2009). To put it simply, at the most basic level, we can classify institutional knowledge as expert knowledge with two integrated sub-types: procedural and domain knowledge. While procedural knowledge describes how things are done, that is the procedures and protocols that guide the institutional activities and processes, the domain knowledge refers to the objects, concepts and notions that are part of the institutional world. For example, a medical doctor should possess knowledge on how to treat diseases and how to communicate with patients, among others; at the same time, she or he should know about institution-specific concepts including human anatomy, medical drugs, and diseases. Both types of knowledge – procedural and conceptual – are essential building blocks of the institutional operations at large. The exact role of the knowledge in an institution impacts the way the knowledge is communicated. Following Hall’s (1976) theory of culture, we can, at the most basic level, distinguish between two types of knowledge distribution across the institutional members: low context and high context. Low-context institutions possess large amounts of hidden knowledge, that is, knowledge that is held by certain members only. However, in high-context institutions, the knowledge is shared across the majority of members. The institutional knowledge is here part of the common ground the institution operates on. As a consequence, in a low-context institution members have to communicate more extensively than in a high-context institution. Members must often “dig” for information. They need to get in touch with their peers with a view to finding things out. In comparison, a highcontext environment requires far less communicative effort. Since all relevant information is part of the common ground (cf. Clark, 1996) and thus easily accessible, there is comparatively less need to communicate. To give an example: medicine and art represent the opposite ends of the high-context versus low-context spectrum. On the high-context end, we find medicine, where most institutional members have a working knowledge of the basic concepts like human anatomy and disease, and possess important background information about treatment procedures and surgery protocols. These facts, concepts, and protocols are, at all times, present in the background of the institutional operations and are accessed by the institutional members “on the fly” if need be. In contrast, art requires an extensive amount of communication about its concepts and procedures. There is no existing common ground per se, as art is grounded in creative acts of the individual artist. The concepts and procedures have to be (re-)constructed, unveiled, and shared through huge communicative efforts across the institutional members. For example, artists are interviewed, art pieces compared, and artistry is documented in a variety of ways with a view to “getting to the bottom of things”. In contrast to a high-context institution like medicine, the common ground that is established in this way provides no more than a snap-shot view of a particular artist or, at the most, of an art genre. Similarly to institutional communication in general, institutional knowledge can be described as a point on a continuum between two end points. These end points are “shared/ formal” and “hidden/informal”. It shows that both continua, that is, the one for institutional communication and the one for institutional knowledge, correlate with each other; the knowledge on the shared/formal end of the continuum is usually communicated in a comparably impersonal, pre-defined, and formal manner, while the opposite is true for hidden/informal information. To be more precise, knowledge that is largely distributed across an institution often comes in some codified way. However, hidden knowledge comes in all shapes and sizes. Think about legal texts versus an art critique. While legal texts adhere to a very specific, pre-given patterns of language use, an art critique may be written in a unique and creative voice and express a largely personal tone. 324



Dialogue in Institutions

How the knowledge gets distributed within the institution is, to a large extent, dictated by the exact position of the knowledge on the knowledge continuum. For example, knowledge might be shared electronically via email. Email communication usually applies whenever the recipient base is large and the knowledge leans toward the shared/formal end of the continuum. Similarly, many organizations provide intranets that employees can log into with a view to tapping on shared/formal knowledge. More discrete channels for distributing hidden/informal knowledge are provided by data uploads, which direct the data to pre-defined addressees only. In some cases, knowledge regarded as sensitive may only be shared through in-person communication. For example, many organizations hold appraisal talks in which supervisors meet their employees personally to give them feedback on their work performance and prospects for their professional development. Likewise, business meetings create an in-person work environment, in which exclusive information on business strategies and the like can be shared among a selected group of “insiders”. In this context, it is important to note that while face-to-face communication, being used on the hidden/informal end of the knowledge continuum, integrates cognition, perception, and verbal expressions from the very beginning, information and communications technology (ICT) mediated communication on the formal/shared end can usually provide only a truncated version of natural dialog. For example, email and chat lack the perceptive/visual level. Virtual conferencing tries to bring the perceptive/visual level back in; however, haptic and olfactory features are usually left out. And, as research shows, the virtual representation of something or someone is perceived differently than the real thing. Our brain responds more actively to a real person that stands in front of us compared to an avatar on the computer screen. On the other hand, ICT can also enhance in-person communication. In meetings, for instance, computer-assisted presentations are commonly used to display information for the audience, which the speaker does not explicitly mention in her talk. Metacommunicative information can be added, for example, by stylistically highlighting selected words and inserting images on the presentation slides. Depending on how it is used and the context it is used in, ICT enhances the “impersonal/ pre-defined/formal” or the “personal/undefined/informal/” character of the knowledge shared. Email might in some cases create a sense of anonymity, which might evoke a more informal tone, as compared to a business presentation where the participants are all physically present and politeness and face are much more prominent. To conclude this section, the important point here is that the type of knowledge – shared/ formal versus hidden/informal – opts for a specific type of distribution in the institution. While shared/formal knowledge is more manageable via mass-distribution channels, hidden/informal knowledge requires in-person forms of communication like personal meetings and conversations. At the same time, the selected communication channel like email and business meeting amplifies the knowledge features “shared/formal” and “hidden/informal” respectively.

20.5  A Theoretical Framework for Institutional Action Games As discussed in the previous sections, my conceptual description of institutional communication is based on Weigand’s (2000, 2010) action game model. The basic cornerstones for my framework of institutional communication laid out so far are:

325

Sebastian Feller

1. 2. 3. 4. 5.

The institutional goals and interests. The institutional structure. The main functions of institutional communication. The institutional knowledge. The used communication channel(s).

These five factors form the heuristics for demarcating and analyzing specific institutional action games. It is important to note that all five factors stand in a reciprocal relationship with one another, with the institutional goals and interests being at the center of the framework. It is in the end the goals and interests that determine how the other factors are instantiated exactly. They determine which types of knowledge become the main currency of the institution as well as which institutional structure and which communication channels are installed. To illustrate the connections between [goals] and [structure], I consider an example from economy. In the business world, the main institutional interest is to make a profit. In response, businesses have created sub-groups like marketing and sales which are specialized in profit making. The groups possess expert knowledge on processes and techniques that match commercial products with the market demand. In contrast, in institutions where profit making is not the primary goal, these sub-groups usually do not exist. Law, for example, runs under a completely different structure, which is due to the fact that the primary interest is directing human conduct, not profit making. In the law context, we rather find sub-groups like legal defense and prosecution. Similarly, different institutional interests and goals correlate with different variants of [knowledge]. A business that is after profits must know about the market, its competitors, production, and the customer. In contrast, a social institution should possess detailed knowledge on social issues, collaborators, beneficiaries, and societal structures. In this way, the business world should create a customer- and profit-oriented environment, whereas nonprofit institutions like the law and politics ought to be built on developing a social conscience and striving for societal goals that serve “the greater good”. In a similar way, the goals influence the knowledge distribution channels. The business world usually runs advertisement campaigns spanning over a wide range of different media like newspapers, television, and social media with a view to attracting new customers. To outperform competitors and to increase sales, the communication channels should reach as many potential customers as possible. This kind of publicity is totally at odds with more “introverted” institutions like the law, for instance. In law, the communication channels are deliberately restricted. One reason for this is that making sensitive information public could have reverse consequences for an ongoing trial or investigation, for example. This said, it clearly shows that institutional action games integrate the above mentioned factors from the very beginning. As we will see in the next section, the complexity of the game is increased by another decisive factor: culture.

20.6  The Influence of Culture In addition to the factors mentioned above, culture plays a crucial part for any institution. Rapaille’s (2007) work is seminal in this context. According to Rapaille, the culture code or cultural bias that affects how we perceive the world around us determines to a significant extent the norms and values under which an institution operates. To be precise, Rapaille’s

326



Dialogue in Institutions

discussion mainly focuses on consumer expectations in a capitalist society. In his discussion, he focuses on a contrastive analysis of the North American and European cultures. For instance, Americans perceive food as fuel for the body. In contrast, food is encoded as pleasure in the French culture. Similarly, while Americans perceive their car as a representation of who they are, Germans regard it a piece of engineering. Institutions operate under the same cultural influences. For example, a particular business culture is shaped by how work, products, and markets are defined in the larger cultural context. For example, in the American culture, work stands for self-fulfillment and identity seeking. The American business world is filled with entrepreneurs who reinvent themselves day in and day out. At the same time, money works as a proof that one is performing well in one’s job. In many European countries, money is, on the contrary, viewed with a certain suspicion. Rich people are often accused of cheating and deceiving others. Institutions absorb such values and norms and integrate them into their code of conduct and operations. While in the United States huge bonuses are still in place and largely accepted for high-stakes managers in the business world, in Germany, many managers forfeit their bonus payments with a view to creating a better public image for themselves. Consequently, integrating culture into the action game model for institutional communication gives us the following picture, see Figure 20.1. As indicated by the lines between the boxes, the factors are all cross-linked. They form an integrated whole, influencing and shaping one another. Replacing the factor variables with concrete data points, we arrive at a specific institutional action game. Culture on the outside of the institution impacts the nature of the institutional culture on the inside. As a result, institutions might share the same set of internal organization, while their operations still differ significantly due to a difference in the cultural impact from the outside. This, in return, shapes the way the institution communicates. The important point is that the factors in the institutional action game stand in reciprocal relationships. This means that a change in one factor results in a change in the other factors. I have already discussed this point in Section 20.5 with regard to how different institutional goals and interests lead to the instantiation of distinct institutional structures. The model presented here is certainly not exhaustive. There are other factors I have not discussed including information on the individual level such as the level of motivation, work preferences, the level of experience, and emotional states of the individual members, to name a few. This information is certainly of paramount importance to arrive at a more

goals

culture

structure

Knowledge

channel

institutional culture

Figure 20.1  Components of the Institutional Action Game. 327

Sebastian Feller

complete picture of a specific institution. However, extended empirical research is needed to collect the relevant data.

20.7 Conclusion In this chapter, I have laid out an action–theoretical model of institutional communication along a set of institution-specific factors including the institutional goals, structure, knowledge, and culture, among others. These factors impact the institutional outline, shaping the way communication in the institution takes place. I have argued that institution-specific goals and interests are of paramount importance: they are the driving force behind how the other factors are instantiated. As for the institutional structure, I hold that tall hierarchies usually require more impersonal and formal communication along closed patterns of language use, while in flat hierarchies, institutional members communicate in comparatively personal and open ways. Knowledge, however, affects communication through how and what kind of information is distributed across the institutional members. I have distinguished between high-context and low-context institutional settings, that is, institutions where most knowledge is shared across the majority of members and those where knowledge is exclusively held by individual experts or smaller groups of experts. I have argued that within low-context institutions knowledge is usually conveyed via personal/undefined/informal communication, while the opposite is true for high-context institutions. Furthermore, I discussed the role of the communication channels. More anonymous channels like email tend to push the communication toward the “personal/undefined/informal” end of the communication continuum, while analogue forms, like an in-person business meeting, lean toward the “impersonal/pre-defined/formal” end of the communication continuum. In addition, which channel is used relies on the nature of the knowledge that is conveyed. Sensitive information calls for analogues channels of communication. On the contrary, insensitive information is likely to be conveyed through more impersonal channels of mass distribution using ICTs. Next to the internal factors and the connections they hold, the discussion was extended to culture, both as an external and an internal determinant of the institutional operations and communication. I have argued that culture shapes the way an institution operates on the inside, for example, in the form of the institutional culture, and, therewith, impacts how the institutional communication is shaped. On the outside of the institution, culture provides a framework of beliefs and values, that is, a specific culture code, as discussed in Rapaille (2007). The culture code shapes the institutional culture on the inside by providing culturespecific perceptions of institution-specific notions and concepts. This chapter has laid out a conceptual framework of institutional communication in terms of a multi-factor communicative action game. I have argued that influences both from in and outside of the institution shape the way the institutional communication is instantiated. As with language use in general, we need to take a holistic theoretical standpoint with a view to coming to grips with what is happening in institutional communication. Dialog analysis and Weigand’s action game theory provide valuable ideas in this area of research.

References Armitage D. R. et al. (2008). Adaptive co-management for social-ecological complexity. Frontiers in Ecology and the Environment 7(2): 95–102. 328



Dialogue in Institutions

Castells M. (2013). Communication Power. Oxford, UK: Oxford University Press. Clark H. H. (1996). Using Language. Cambridge, UK: Cambridge University Press. Dictionary of Military Terms: A Guide to the Language of Warfare and Military Institutions. Reference Reviews, 2004, 18. Jg., Nr. 2, S. 24–25. Guffy M. E., Rhoddes K. and Rogin P. (2005). Business Communication. Toronto: South-Western. Hall E. T. (1976). Beyond Culture. Garden City, NY: Anchor Press. Hardy C. (2011). How institutions communicate; or how does communicating institutionalize? Management Communication Quarterly 25(1), 191–199. Javernick-Will A. and Levitt R. E. (2009). Mobilizing institutional knowledge for international projects. Journal of Construction Engineering and Management 136(4): 430–441. Knill C. and Lenschow A. (2005). Compliance, competition and communication: Different approaches of European governance and their impact on national institutions. JCMS: Journal of Common Market Studies 43(3): 583–606. Lammers C. (2011). How institutions communicate: Institutional messages, institutional logics, and organizational communication. Management Communication Quarterly 25(1): 154–182. Lumsden C. J. and Wilson E. O. (1985). The relation between biological and cultural evolution. Journal of Social and Biological Structures 8(4): 343–359. Margolis J. (2009). Culture and Cultural Entities-Toward a New Unity of Science. Dordrecht: Springer Science and Business Media. Ogaard T., Marnburg E. and Larsen S. (2008). Perception of organizational structure in the hospitality industry: Consequences for commitment, job satisfaction and perceived performance. Tourism Management 29(4): 661–671. Perry-Smith J. E. (2006). Social yet creative: The role of social relationships in facilitating individual creativity. Academy of Management Journal 49(1): 85–101. Rapaille C. (2007). The Culture Code: An Ingenious Way to Understand Why People Around the World Live and Buy as They Do. New York: Crown Pub. Rogoff B. (2003). The Cultural Nature of Human Development. Oxford et al.: Oxford University Press. Room G. (2011). Complexity, Institutions and Public Policy: Agile Decision-Making in a Turbulent World. Cheltenham and Northampton: Edward Elgar Publishing. Schnurr S. (2012). Exploring Professional Communication: Language in Action. London and New York: Routledge. Sopazi P. N. (2014). The Role of Language and Culture in Technological Innovation (Thesis). Johannesburg: Johannesburg, University.  Weigand E. (2000). The dialogic action game. In Coulthard M., Cotterill J. and Rock F (eds.) Dialogue Analysis VII. Working with Dialogue. Tübingen, Germany: Niemeyer, 1–18. Weigand E. (2003 [1989]). Sprache als Dialog. Tübingen, Germany: Niemeyer. Weigand E. (2010). Dialogue: The Mixed Game Model. Amsterdam and Philadelphia: Benjamins.

329

21 Dialogue and the Law Fritjof Haft

21.1 Introduction Traditionally, the role of the institution of the law was regarded as that of resolving social conflicts. The parties carry on a regular dialogue in the trial before the court. The court listens to the parties, examines the legal situation, participates in the dialogue, refers to any evidence that may be required and then decides who is “in the right”. It supports this decision with reference to laws, precedence, opinions expressed in the literature and other legal material. The power of the state is available to enforce the judgement. The German jurist Rudolf von Ihering already described the disadvantage of this venerable institution in 1872 in an article with the resonant title “The Struggle for Law” as follows: [T]he law is significant only as an expression of conflict and represents mankind’s efforts to tame himself. Unfortunately however the law has tried to counter violence and injustice with methods that in a rational world will some time appear as alien as they are ignoble. Because the law has never really tried to resolve the conflicts in society but only to mitigate them in that it lays down rules according to which these conflicts are to be fought. (Ihering 2012: 121) It took a very long time until, “in a rational world”, serious efforts were undertaken to overcome this disadvantage. It was not until the 1970s that a worldwide movement called “Alternative Dispute Resolution” (ADR) began to take shape. The “alternative” to classical court proceedings consists in no longer having conflicts decided by a higher authority but by giving the parties institutional aids to resolving their conflicts on their own responsibility. The parties should independently conduct a dialogue in direct “negotiations”, the role of the judge being taken by a “mediator” who is not equipped with any higher authority but is a mere service provider who, contrary to a judge, has no sovereign authority whatsoever. The classical “judgement model” in the law is therefore matched by an alternative “consensus model”. This modern development must be borne in mind if one wishes to clarify what the institution of the law signifies today and how it has influenced the science of dialogue. 330



Dialogue and the Law

21.2  Language and Law Lawyers are the only profession that operates pure information processing. The medium in this context is overwhelmingly language. The knowledge of modern linguistics is reflected – albeit hesitantly – in a hyphenated discipline by the name of “legal informatics” (Müller 1989). As with another comparable hyphenated discipline, information-law, for example, this is not a legal discipline but a sub-discipline of the other specialist area concerned, in this case linguistics. The legal theorist who dares to enter this territory is first struck by the fact that modern linguistics has two major research directions: one sees language as a system of symbols. This broad context also includes the generative concept of language developed by Chomsky (1965); the other explores the use of language. If one now looks back into classical legal theory, the centre of which is dominated by the search for a true understanding of legal concepts in laws and of texts describing precedential cases, one finds only a trend towards the former area. Since the majority of lawyers are “xenophobic” about other disciplines, this classification is rarely challenged. Legal theory only invokes linguistics very occasionally. It is far more concerned with autonomous analyses of language units, especially the “concepts” already mentioned. An example is provided by German “methodology”, which has developed a body of “interpretation methods” for legal concepts (“characteristics of an offence”), which is based on the hermeneutics of the nineteenth century. The leading textbook on this subject has for many years been Larenz (2014). If linguistics and legal theory agree in the investigation of language as a system of symbols with regard to the subject of the investigation, this will only very rarely be the case when it comes to the issue of the use of language. That a change in linguistics on this subject currently taking place is only rarely recognised. Weigand describes this transformation as “the pragmatic turning point” (Weigand 2010, 16). This “pragmatic transformation” raises the problem of the smallest unit that does not exist with this poignancy in the “system of symbols” language. There are basic units in the “system of symbols” such as the phoneme and the morpheme, which in law correspond with the “concepts” referred to and further units, such as the “legal institutions”. On which unit, however, can the pragmatics of language be based? The representatives of the “speech act theory” have provided an answer to this question. Their founder, John Langshaw Austin, held a series of lectures with the programmatic title “How to Do Things with Words” at the University of Harvard in 1955 that was published posthumously in 1962 (Austin 1962). His pupil John Searle then developed the theory of the speech act further in 1969 (Searle 1969). He found in the speech acts the smallest unit of language pragmatics. Speech acts are speech utterances that are not satisfied with stating the contents but themselves carry out actions (“act”). That this approach is significant for the law is indicated merely by a glance at the court judgement that one may regard as a declaratory speech act par excellence (Weigand 2005). The “validity” of the judgement may have existential significance right down to the loss of the condemned offender’s freedom and life. But, as far as one can tell, only occasionally and more from a linguistic point of view have interdisciplinary investigations between linguistics and jurisprudence been conducted on this subject to date. An example is provided by a dissertation published at the University of Heidelberg, not in the legal faculty, however, but in the modern languages faculty (Marinkovic 2006). I suspect that the formal intellectual emphasis that characterises so many works in contemporary linguistics, namely the concept of the generative language skill, had a deterrent effect here. As soon as the “normal” lawyer sees the first formula, he or she is sure to switch off. 331

Fritjof Haft

As Weigand made very clear, speech acts cannot be investigated in isolation. Rather, partners in a dialogue inter-react in that they react to another speech act or evoke a speech act (“dialogical principle”). It is not the speech act as such but this inter-reaction that is the unit in the use of language. There exists in this situation open, complex room for manoeuvre in which the environment of the act is also involved. The term coined by Weigand for the behaviour of acting and reacting human beings in this environment is “action game”. Weigand makes a distinction between two types of action game, namely “minimal two-part games of action and reaction” and “complex games”. The former are “one-phase games” and the latter include two types, namely “extended minimal games” and such as include several phases (Weigand 2010, 129). If one transfers this distinction to the law, one meets two types of legal action game: “minimal games” corresponding with what I would like to call here “position games”, whilst the “complex games” correspond with what I term “consensus games”. In both kinds of game, the players have to surmount a problem that occurs in every legal case, even the simplest: I am talking about the problem of complexity.

21.3 Complexity The concept of “complexity” is inexact. One can interpret the concept of “complexity” tautologically, that is relate it to itself: the concept of complexity is itself complex. Complexity is used in very different ways by members of different disciplines. I am dealing with it here as the central methodological problem that is to be found in all kinds of action games. Complexity, according to this meaning of the word, is identified by eight characteristics. First, there are numerous “aspects”. These are aspects of the situation, legal aspects, economic aspects, intangible aspects, “base” aspects, social aspects, emotional aspects and many more. Not all aspects are known to the participants. When neighbours quarrel because the tree in one of their gardens throws shadows on the other’s garden, what lies behind it is possibly a greeting not returned by the wife. Second, the aspects form “aspect hierarchies” with sub-aspects, sub-sub-aspects et cetera. No single well proportioned “hierarchy tree” exists, but rather a large number of interrelated partial hierarchies in different “trees” that are construed differently by the players. Thus, a “contract” may consist of different complexes (“contract conclusion”, “fulfilment”) to which in turn sub-complexes (“preliminary discussions”, “negotiation”, “signing the contract”) belong. The plaintiff perhaps describes the origin of his claim and the defendant adds an expiry complex. One can distinguish relevant and irrelevant, disputed and undisputed facts; the latter are joined by facts as proof that are themselves sub-divided into evidence (documents, witnesses, expert appraisers). Third, the aspects are predominantly not “yes/no aspects” but rather gradual “moreor-less aspects”. They are, in other words, not “concepts” that can be defined but “types” that can ultimately only be explained by reference to examples (dealt with in more detail below). An apparently unambiguous legal concept as “matter” can become a problem at any time if one is, for example, dealing with electricity in criminal law that one can in fact “steal”, or with a long-distance ski run that can be “damaged”. The borderline between “damage” and “destroy” can only be drawn typologically here. And when one is dealing with highly abstract legal concepts such as “good faith” or “public interest”, any thought of a concept disappears from the outset. This becomes absolutely clear in the event, for example, of forecasts in court judgements.

332



Dialogue and the Law

Fourth, the aspects are interwoven; they represent “systems” that do no not always stay constant. Rather, different people produce different “systems” with regard to the same legal case. No one can master the totality of these systems. The lawyer can always only deal with sub-systems and requires the ability also to be able to navigate in alien systems. Since the time of Roman law, many lawyers in continental Europe have been concerned with “the” system of law. In the nineteenth century it was believed that it had been found with the aid of a so-called “conceptual jurisprudence” (Haft 1990, 32). This proved to be a mistaken belief. “The” system of law does not exist. Many areas of the law, in Germany tax law and social affairs law for example, became so chaotic under the influence of the lobby that they have become completely inaccessible to efforts to systematise them. Nevertheless, it is clear that no single legal aspect can ever be regarded in isolation. It is always embedded in systems made up of other aspects. These are themselves tied up in systems elsewhere – and so it goes on. Fifth, the systems have “their own inner dynamism”. There is a reason that court proceedings are called a “trial”. Things happen before the court and outside the courtroom. Those affected are faced permanently not only with the decision as to whether they intend to do something or not, but whether they are thereby intervening in processes that are already in motion and whether they wish to alter the course of the proceedings or wish to do nothing, which means to let things continue to drift in the same direction. Thus, the failure to answer a letter lodging a claim can trigger a writ. The failure by the other side to contest the presentation of facts can have fatal consequences. Lack of activity may entail time-expiry or judgement by default. Sixth, there is not a single goal in the solution but several, frequently contradictory, goals. Possible goals in litigation may be justice, economy, satisfaction, vanity, ambition, revenge. Anyone who concentrates only on the goal seen on the surface will run into truly proverbial risks: “fiat justitia, pereat mundus” (“Let there be justice, though the world perish”). Seventh, the individual sub-goals are frequently in “conflict” with each other, right through to the case in which one goal can only be achieved at the expense of another. When neighbours quarrel, even should one of them win, neighbourhood peace will have been destroyed and this loss may permanently embitter both neighbours’ lives. The eighth and indeed the most important characteristic of complexity: there are always “information deficits” that it is impossible to eliminate. However well one prepares oneself (something which one cannot do – simply due to the nature of the human memory, perfect preparation is impossible), it will always occur that problems and questions emerge for which one is not prepared. Complexity in the sense described here is not a characteristic of the modern world. It emerged eons ago as a methodological problem in efforts to resolve social problems. When, for example, our ancestors living in a community over used a field (“Tragedy of Common Property”) they had to deal with a problem that even the modern world has not been able to solve; one only has to think of the over-fishing of the world’s oceans.

21.4  Position Games Strategies to overcome the problems of complexity in legal disputes were developed very early on. The oldest of these strategies was the “palaver” practiced by African tribes. One could describe it as a “bloodless duel”. Today, the Western world understands a palaver as superficial and unstructured gossiping. The concept, however, had a positive connotation in African 333

Fritjof Haft

culture. It described the practice of discussing the community’s problems in an assembly in which the longer it lasted the more important was the issue and the more senior those participating. As far as the resolution of disputes was concerned, one finds here for the first time the structuring of social conflicts governed by rules. Procedural rules mark the beginning of the development of law. So-called “substantive law” followed later. Anyone wishing to influence litigants must stipulate formal rules. The most elementary of these rules is the allocation of statement and counter-statement, whereby attention must be paid to both being allotted the same amount of time. Modern research into negotiations has examined this and why such formalities are complied with (Haft 2000, 73ff.). Contrary to modern court proceedings, however, a palaver was governed to a large degree by freedom, which gave rise to a flexibility the equivalent of which can be experienced in modern cross-examination in the case of yes/no questions. An element of control was provided by the presence of each parties’ supporters, of the audience, but above all, of the elders who acted as moderators and prevented the tendency to empty gossip. A new action strategy in overcoming complexity that is still practiced today, developed as people’s courts were introduced together with democracy in classical Greece. Previously, the parties were required to present their case personally in their own words before such courts. They had to present the facts of the case, quote and interpret the relevant laws, subsume the case under these laws and submit certain pleas. The court did not receive any advice in reaching its verdict. On the contrary, the court decided spontaneously by agreement subject to the direct impression made by the parties’ testimony. In reaching its judgements, it was not permitted to deviate from the parties’ pleas. It had only the option of admitting or rejecting a plea. The proceedings took account of the fact that a people’s court, made up of several lay individuals, was unable to carry out a detailed investigation of the facts of the case or to weigh up different legal considerations. The key factor was solely the convincing speech. Everything depended on this. The sole assistance available to the participants lay in the possibility of having such a speech prepared by an expert and then learning this speech by heart. This gave rise to the profession of the logographer. Well known orators such as Isokrates (436–338 BC), a pupil of Gorgias, and Lysias of Syracuse (about 450–380 BC) practiced this profession. To what extent was this “progress” on a palaver, if it was progress at all? It was not in the public, nor in the verbal qualities, nor the lay judges, and not in the development of the proceedings as a “bloodless duel” – all these elements were present in the palaver. Progress lay in the clear and precise plea that the parties were required to formulate and which bound the court in its judgement. If it was the oratory that was the reason for the court’s conviction on right and wrong, it was the plea that programmed the decision that had to be made on the basis of this conviction. The plea is the rhetorical heir that has determined the fate of the law up to the present day (Haft 1999, 255). Right up to today, the plea enables complexity in the law to be overcome in the simplest conceivable way. But this only benefits the parties when quite simple issues are at stake, when, for example, the justification of a claim for money is in dispute. But this is not typical of most cases and then the plea frequently damages the parties. But it always benefits the lawyers. The parties do not notice this because the requirement to submit a plea provides them with a comfortable and simple opportunity to spell out an ideal future in which their conflict is resolved in a way that they would wish. Since the parties in dispute have different ideas of the future, each compares this ideal future with their own ideal future. And since both contrary versions cannot be transformed into reality, efforts to find a mutually agreeable solution to the conflict are doomed to failure. That also benefits the lawyers because 334



Dialogue and the Law

they will be called upon to decide on the conflict. Right up to the present time, the plea is the business foundation of lawyers’ professions. The first people who recognised and took account of this were the classical Roman lawyers. They created an action system in which only certain writs were admitted. Anyone who sought justice before the court had to lodge an admissible writ with the court and submit the plea contained in this writ. The court only had the option of accepting (“absolvo”) rejecting it (“condemno”). Should the person concerned be unable to formulate a permissible writ, however grave and incriminating his concern might have been, the appeal to the court was futile from the outset. Such a writ was “inadmissible”, in other words the court refused even to consider it. This “position thinking” that was established and refined in classical times, today dominates the whole of jurisprudence in the Western world. Whether the issue in civil law is to demand something from another person, whether there is an issue in administrative law concerning intervention by the state, or a demand by the citizen from the state, or whether in criminal law the presence of clear and precise evidence of an offence before the deed (“nulla poena sine lege” – no punishment without law) – everywhere the law has become a pure positions event. Anyone who cannot state clearly and precisely what he is demanding will not be admitted to any legal proceedings. Everywhere, thinking is expressed in terms of claims, demands, authority to intervene, assessments, legal acts, accusations, the circumstances surrounding a crime. No one can approach a court or a public authority with a “problem”. At all times and everywhere, he is required to take a position and to state unambiguously what he is demanding. Instead of asking questions, he must have his answers ready and fight to have them realised. The methodological origin of this event is quite forgotten. No one thinks any more of the fact that originally the plea served to enable a hopelessly over-burdened people’s court to arrive at a judgement. This aspect of making the work easier also played a major role with the Roman praetor. The writs admitted by the praetor were not those that best served the citizens of Rome but those that he knew and had mastered. If, for instance, a writ concerning property was lodged, the praetor knew what he had to do. He had to verify whether the plaintiff was the “owner” of the object and whether the defendant was “in possession of the object” who was not entitled to any “right of possession”. Should all three conditions be fulfilled, he admitted the writ and otherwise rejected it. Contrary to the Greek people’s court, he was guided in his work (at least in the later Roman era), no longer primarily by the ability of the speech to the court to convince, but by the legal commentary on the praetorian edict. Lawyers today do just the same. The development of conflict to a plea in no way necessarily moved in the way described. In many cultures of the world – from Far Eastern countries through India and the Islamic world to Africa – other traditional, more consensus-oriented models of conflict resolution existed and still exist, models in which no “positions” are taken (Haft 2014, 43). Many countries that like Japan adapted German civil law in the nineteenth century are currently becoming increasingly aware of their own traditions and are turning away from position thinking as it originated in the West. In Japan, apart from the out-of-court settlement (“Wakai”), there exists an arbitration procedure (“Chotei”), which is highly significant in family law for example. Approximately 90 percent of divorces are settled in this way in mutually acceptable proceedings (“Kyogi-Rikon”). In these countries’ opinion, the most serious disadvantage of a legal system fixated on a plea is that the parties frequently fail to see their true interests. If the only issue is to enforce a demand for money from a defaulting debtor, position thinking is quite in order. But it is rare that cases are so cut and dried. 335

Fritjof Haft

Frequently, the issue is not purely one of money but also and primarily about other interests. As a matter of principle, interests are something different from positions. They might even be completely contrary to them. An example from insurance practice may serve to make this clear. The interests of the injured victim of an accident are without question directed towards getting back to health. He is, however, unable to demand health from the insurance company of the party that caused the damage. He can only demand money. The more ill he is, the more money he may demand. He therefore has his doctor certify that he is unable to work and demands a pension for the rest of his life. Obviously, under these circumstances he cannot recover his health. Legal positions and interests are fully at odds with each other. One may therefore legitimately doubt whether the invention of the plea has brought progress compare with the archaic palaver. The supporters of ADR would also deny this question. They are convinced that the parties are frequently badly served by legal proceedings that are based on position thinking. They are endeavouring to find a better system that is focussed on the interests of the parties tied up in a conflict. This brings me to the contemporary attempt to overcome “position games” in the law and to move to consensus games.

21.5  Consensus Games During the 1970s and 1980s, the worldwide movement that was ADR arose. The “alternative” in this context referred to state court proceedings that, due to the compulsion to take and defend positions, frequently did not do justice to attempts to find a balance of interests between the parties to the dispute. Symptomatic of this situation is the fact that the German law governing the Code of Civil Procedure (ZPO) only mentions peace in a single paragraph of more than 1,000 paragraphs, namely in § 278 ZPO, and the remainder all deal with struggle. Paragraph 1 of the regulation reads as follows: “The court should endeavour at every stage of the proceedings to seek an amicable solution of the legal dispute or individual points therein.” The solution of conflicts through “settlements” or even such that, beyond the purely legal situation, balance the parties’ interests and might even lead to win-win solutions are just as difficult to achieve in “the struggle for justice” as in war, where the opponent first has to be exhausted by losses before peace is found. In 1978, Frank Sander, Professor of Law at Harvard University, delivered a talk on “Varieties of Dispute Processing” at the Pound Conference organised by the American Bar Association and thereby provided the stimulus for the development of ADR. At the beginning of this movement stood occupation with the subject of “negotiation”. That becomes evident simply from the names given to the research establishments that were established at that time. In the United States one could mention, for example, the “Harvard Negotiation Project”, the “Stanford Center on Conflict and Negotiation” and the “Consortium on Negotiation and Conflict Resolution” at Georgia State University – I could continue the list of projects with “Negotiation” in the title. In Germany, I founded the “Tübingen Negotiation Seminar” at Tübingen University in 1983 (Gottwald and Haft 1983). This origin also finds expression in the titles of the specialist journals that were established at that time. One should name, for example, Negotiation Journal established in the United States. In 1981, Roger Fisher and William Ury, law professors at Harvard University, published a bestseller on the subject of “negotiation” with their book Getting to Yes that achieved success throughout the world (Fisher 2011). Negotiation was the main subject. At the beginning of this movement, “mediation” was not even mentioned. It was by no means self-evident that researchers in Western industrialised countries would take up the subject of “negotiation”. Negotiating was (and still is) frequently understood as 336



Dialogue and the Law

“haggling”, in which the main issue is to successfully “pull the wool over the other’s eyes”. Especially since in Germany, this sort of behaviour had (and still has) a bad reputation. Germany is a country of fixed prices in which the “bazaar” is not loved. Popular books with promises such as “Nowhere else can you win so much with such a low stake by clever negotiating” contribute to this negative picture. People thought and still think of the ability to sell a used car at the highest price (or buying at the lowest price). One had and still has the image of the door-to-door salesman who uses the tricks of sales psychology to persuade customers to subscribe to a magazine or buy a vacuum cleaner. Even today, many lawyers believe that the subject of “negotiation” is almost unworthy of their profession. Anyone for instance who investigates the mechanisms of a bazaar in a paper, and in doing this deal with exact questions such as whether it is expedient to begin, as in a bazaar, by naming a figure in compensation proceedings, feels as if he is plumbing depths avoided by serious lawyers; no specialist law journal would print such a thing. Anyone, however, who writes a twentypage monograph on the question as to whether a limb (the side of a karate sportsman’s hand, for example) can be a dangerous tool within the meaning of the offence of causing grievous bodily harm, is conducting “serious” research; the monograph will be printed (even if not necessarily read). During the first seminar for students on the subject of negotiation that I held in the early 1980s, a colleague commented to me: “That will only teach you how to get a good price for your used car”. Incidentally, I believed (and still believe) that this is a perfectly worthwhile objective. Why did academics risk their reputations at that time by making “negotiation” a subject of research and teaching? They did it because they were pursuing an interest they believed to be important and that is in fact important. They were concerned to obtain scientifically founded knowledge on the manner in which human beings treat each other in conflict situations. Conflicts are an unavoidable part of humans living together. They can release positive forces. But they can also ruin people. Lawyers especially are familiar with the latter. Some of the potentially most ruinous occasions for conflict undoubtedly include litigation conducted before courts of law. People who have suffered this fate are driven to the edge of insanity (and not infrequently beyond it). One should avoid them. The old adage about a lawyer who represents himself in court has an ox as a client and a donkey as a lawyer, only confirms this finding. Obviously, there have also been trial deficiencies that led to ADR. In trials before a court the parties surrender procedural autonomy. It is not them but the court that determines how the case proceeds. Negotiations on a settlement do not take place in a trial until the conflict has reached a dead end due to the prescribed taking up of positions fixed in writing. The lack of the authority on the part of judges to enforce a proposed settlement moreover justifies the risk of the use of external and internal instruments of power (threats, manipulations) by the judge (Haft 2000, 170). Finally, the fact that the courts regularly issue judgements based on the principle of “everything or nothing” also plays a role. From the point of view of game theory, trials are zerosum games. One party’s gain is the other’s loss. The result is always zero. No economic value is created in this way. Creative solutions beyond the “pure legal situation” cannot be found in a judgement (and only with great difficulty in a court settlement). As a matter of principle, a trial cannot be transformed into a non-zero-sum game. But for all that, the origin of the dominance of position thinking ought also to be seen. As already said, this offers a simplification technique that can be comfortably employed by lawyers given the fact that human beings cannot adequately cope with complexity – and every conflict is complex. That was the case in the people’s court in classical Greece and 337

Fritjof Haft

is still the case today. Position thinking also offers the participants a strategy in overcoming complexity that is easy to use. It is always very easy to lodge claims, just as it is comparatively simple to decide on the “justification” of such claims. On the other hand, the participants frequently find it difficult to discover what complex interest situation really lies behind claims, that is what those involved really want. It is frequently the case that they do not know themselves. If one wishes to find that out and find solutions that take account of these interests, one has to part company with position thinking together with the associated venerable legal tradition. This is what one is trying to achieve with ADR. The reasons human beings are driven into litigation differ from country to country; the result is frequently equally deplorable. In Germany, reconstruction and the establishment of the rule of law after 1945, as well as the invention of legal protection insurance, gave rise to a veritable torrent of litigation amongst people. The Federal Constitutional Court, the highest court in Germany, once reported on the constitutional claim of a driver who drove on the hard shoulder during a traffic jam, had received a summons for a minor amount of money, and had therefore gone “as far as Karlsruhe” (to the constitutional court). The occupation with the issue of “negotiation” in ADR led to the search for answers to questions such as: what happens when the participants endeavour to resolve their conflicts directly – through negotiations? What hidden mechanisms are at work here? Why do negotiations frequently fail? What hindrances block the path to an agreement? Are these hindrances possibly erected by the participants themselves? Are they even erected by the jurists – lawyers and judges? What can those involved – the participants and the jurists – do to overcome these hindrances? And since the answers to these and other questions were without exception also researched at universities, this gave rise to an additional concern to make the subject of “negotiation” not only a subject of research but also a subject of teaching and not only in the law faculty but also in other disciplines. In addition to the relevant specialist knowledge, skills should be conveyed as part of the study course that are valuable and useful in practically every field of life – in one’s private life as in every profession. The ability to deal with conflicts rationally unquestionably belongs to those “soft skills” the teaching of which belongs to the elementary functions of a university in all disciplines. Jurisprudence is obviously of central importance in this context. Because the substance from which the tools of the law are made are indeed conflicts – that and nothing more. It is largely forgotten today that this had an idealistic starting point. Other than in the case of modern “mediation”, the issue at that time was not to widen access to an attractive market – the market of conflicts. No one was thinking at that time of establishing a new profession – that of the mediator. There was no talk at that time of struggles to divide the spoils, for example between lawyers and the so-called “psycho-socials”, or – in the legal profession – between lawyers, judges, notaries and others. Indeed, occupation with the issue of “negotiation” was in fact – seen professionally – counter-productive. Imagine, people were to be enabled to resolve their conflicts alone, without any help from lawyers. Lawyers and the courts would then become unemployed. Obviously, it will not come to that. There are hindrances along the path to an agreement, which the participants themselves – even with the assistance of lawyers – are not able to remove. The birth of mediation was heralded as this became clear. Its development from research into negotiations is the reason for the particular nature of mediation versus several earlier related forms of assistance from neutral third parties in the event of conflict. In balancing the participants’ interests in negotiations, they are once more confronted with the problem of complexity, because interests display all the characteristics of complexity covered in Section 21.3. To this must be added the search for creative solutions, a subject 338



Dialogue and the Law

with which the advertising industry has principally become involved (Rapaille 2007). It is for this purpose that the method of “brainstorming” has been developed. That is an event in which the collection of ideas in solving a problem is strictly separated from the evaluation of these ideas. No one may be criticised during the phase when ideas are collected because he has suggested an idea that although new might possibly be completely impracticable, indeed even completely nonsensical. Due to the strict separation of “collecting” and “evaluating”, one has the opportunity of possibly hitting on an idea that might appear completely absurd at first but, when examined more closely, is useful and helpful. Visual presentations on flipcharts are used as technical aids and have become known by such names as “mind-maps”. Seen from a methodological point of view, a second dimension is added to one-dimensional verbal language that operates over time. Visual presentations are coming increasingly to the attention of legal theory (Röhl and Ulbrich 2007) as well as of linguistics (multimode linguistics). The catchword used here is an “iconic turn”. Computer programmes enable such presentations to be changed and revised in a manner that is quite simple. Progress is to be found here which, however, due to the networks it creates, quickly exceeds people’s cognitive abilities. An authentic example should make interest focussed negotiating clearer. One normally understands a legal case as a controversial dispute between two parties before a court of law, see the case study in Weigand (2005). ADR, however, relates to the behaviour of parties acting on their own responsibility in which the rules of the game are most clearly recognisable in contractual negotiations. In this case, the key issue is precautionary conduct. This branch of the law is referred to as the jurisprudence of precautionary measures (iurisprudentia eurematica). In his standard work on negotiating, Howard Raiffa described such a case study that dealt with the negotiations on the sale of social home (Raiffa 1982, 35). The case was used by the Tübingen Seminar on Negotiating and re-enacted in both the training of students and in further training events for lawyers. The following extract is taken from my book on negotiating that has already been mentioned (Haft 2000, 11). “Elmtree House” was a home for handicapped young people in a suburb of Boston. A building developer represented by Mr Wilson wished to acquire the house in order to demolish it and build an apartment house in its place. The managing body of the home were not disinclined to sell because it felt that the current neighbourhood was not entirely suitable for the residents. It was believed that a new home in a different residential area would offer better therapeutic conditions. Steve, a professor, who was a member of “Elmtree House’s” advisory council, handled the negotiations on behalf of the home. Steve was an economist, but he did not have any special knowledge or experience in negotiating. Wilson on the other hand was obviously an experienced negotiator who knew all the tricks of the trade in the real estate business. In order to compensate for this deficiency, Steve received advice from Howard Raiffa. That obviously entailed certain surprises for Wilson. Steve began by collecting information. He calculated the home’s own “Reservation Price” (RP). This was the lowest amount the home required in the event of a sale. It corresponded with minimum cost of a different, more favourably located house. He calculated an amount of 220,000 dollars. Steve then looked for other prospective buyers of “Elmtree House”. These efforts were however unsuccessful. By questioning estate agents, he tried to establish the price that would have been achieved in an open-market sale. The estimates varied between only 110,000 and 145,000 dollars. Steve then tried to estimate Wilson’s “RP”, that is the maximum price that he was prepared to 339

Fritjof Haft

pay. Various experts were asked how they judged Wilson’s situation. The picture that emerged was highly varied and vague. Finally, Steve estimated that Wilson’s “RP” was somewhere between 250,000 and 475,000 dollars. Steve then lay down his negotiating strategy. Since Wilson’s presumed “RP” lay in a wide uncertain range, Steve was afraid of making a mistake if he were to open the negotiations by making a demand. He was afraid either of demanding a lower price than Wilson’s RP or of souring the negotiating atmosphere by making a demand that was considerably higher. He therefore made up his mind to try to persuade Wilson to make an initial offer. If this was not successful, he intended to make an initial demand of 750,000 dollars. He decided to introduce this figure in a nonchalant, provisional manner. Steve also thought out an alternative. It consisted in demanding 400,000 dollars and sticking to this amount for a time. He abandoned this idea however since he estimated that this amount was below Wilson’s “RP”. The first negotiating round developed into a disaster in Steve’s opinion. Wilson began with a request to Steve, presented in a humorous manner, to name his minimum demand; he would then try to add something on top. Steve replied that Wilson should instead name his maximum offer; he would then see whether he could grant him a discount. Following this introductory sparring, Wilson made an initial offer: 125,000 dollars. Steve matched this with a demand for 600,000 dollars; he chose this figure thinking that the mean between the two positions was 375,000 dollars and as such considerably higher than his “RP” of 220,000 dollars, and clearly within his comfort zone. Wilson replied that Steve’s demand was totally unacceptable. The negotiations were suspended. Steve decided to telephone Wilson a week later and to tell him that the managing body had decided to reduce its demand to 500,000 dollars. But two days later Wilson anticipated Steve’s call. He said that had given some thought to “Elmtree House’s” social work and wished to take into account that Elmtree House fulfilled a useful purpose for society. He would therefore increase his offer to 250,000 dollars. Steve thereupon reduced his demand to 475,000 dollars. During the following few days, Wilson increased his offer over 75,000 to 295,000 dollars and made a “final last offer” of 300,000 dollars. Steve replied by dropping his price from 475,000 over 25,000 dollars and 400,000 to 350,000 dollars. At this point the negotiations reached a dead-end. In order to enable Wilson to climb down from “final last offer” in a way that would save face, Steve proposed to Wilson that he should grant an additional tax-privileged donation of 25,000 dollars to “Elmtree House”. Wilson accepted this proposal. As it was then revealed that such a procedure presented legal obstacles, the purchase price was finally fixed at 325,000 dollars. Steve was satisfied with this result. He had achieved a purchase price that was nevertheless 105,000 dollars higher than his “RP”. The question as to how satisfied Wilson was remains open. The case demonstrated in an exemplary manner all the characteristics of complexity referred to in Section 21.3 (a multitude of aspects, hierarchies of aspects, more-or-less aspects, interlocking aspects, its own dynamism, a large number of contradictory goals, conflicts on objectives and information deficits). Steve’s game strategy illustrates certain essential elements of rational negotiating: the endeavour to obtain information, the “negotiation dance” around the initial figures, staking out the field of play and the attempt to explain concessions by rational sounding arguments. The latter is typical of negotiations in a bazaar. It can be explained by the desire to conceal the manipulation technique applied in the bazaar. 340



Dialogue and the Law

Psychologists talk in this connection of reciprocity and are referring to effective behavioural programmes in everyone to return gifts. In the bazaar, this is activated by apparent gifts so that the players motivate each other. The journey then ended more or less in the middle between the opening positions, to the extent that they were more or less realistically chosen (see Haft 2000, n. 2 p. 11). Also, the method of escaping from the dead end by “enlarging the cake” shows itself in the idea of a “donation”. The participants obviously knew that this route was legally not feasible but in this way a solution was found that saved face all round. The proximity of such negotiating games to linguistic action games is obvious. A highly promising inter-disciplinary field of research is being opened up here.

21.6  Structured Thinking 21.6.1 Summary It became clear very early in the development of the law in classical times that position thinking alone did not offer a satisfactory solution to the problems of complexity in the law. On the contrary, specific legal structures were required for this purpose. An initial intuitive development of such structures took place as statutory law replaced customary law in Rome in the middle of the fifth century BC. The Law of Twelve Tables contained a basic legal structure in which a distinction was made between two categories of objects, one in which title could only be transferred with considerable effort by means of “mancipatio” (“res mancipi”) and a category in which this was not the case (“res nec mancipi”). The “mancipatio” (from Latin “manus” meaning hand and “capere” meaning seize) was a ritualised and complicated transfer of title. It was required in the case of valuable objects, which included both real estate and slaves. The famous Roman lawyers who soon made an appearance did what lawyers right down to the present day enjoy doing: they developed the “sham sale” (“in iure cessio”) and in doing this established a tradition that has remained to the present day, namely, not to comply with the law but to avoid it wherever possible. To avoid the expense of a “mancipatio”, the purchaser lodged a claim to ownership before the praetor (“actio rei vindicatio”) and claimed, untruthfully, that a particular “res mancipi” belonged to him. The seller, when asked if he had any objections, replied that he did not, upon which, although transfer of ownership was not carried out, it was established with the force of law, which amounted to the same thing. What the praetor though of this method has not been passed down. The Law of Twelve Tables was, like all subsequent laws, abstract. It was therefore inevitable that cases of doubt arose and the custom became established of obtaining legal appraisals from respected lawyers (“responsa”). In the course of time, thousands and thousands of responsae were created in this way. To this were added ever more laws, court judgements and literary references. An early “legal information crisis” developed. How was one to find sources in this flood of material, particularly responsae, which were applicable to a particular case? And, possibly even more important, how should legal education be designed? The man who provided an answer to both questions was the Roman lawyer Gaius, author of the Institutions, the first legal textbook, around 160 AD. It may be briefly and tersely summarised as follows: first comes the method and then knowledge. And right up to the present day, one can identify the method as the concept of “structured thinking” (Haft 1993). In law, “structured thinking” stands for a game strategy in which the complex whole of a case is investigated – and every legal case, even the “minor” dispute between neighbours for example, is complex. In this context, even today the “concepts” as structural elements and 341

Fritjof Haft

their relational links in hierarchical “tree structures” and comparable “tabular structures” play a major role. The parallels with Saussure’s system concept of language as a system of symbols is obvious, but, as far as I can see, this basis of modern “structural” linguistics was not reflected in the law.

21.6.2  The “Concepts” When lawyers talk of concepts they are referring above all to the words in the laws. They are largely derived from common speech (e.g. “individual”, “object”, “destroy” etc.) or on the contrary they became established there as original legal terms (such as “property”, “possession”, “purchase”). They are the “elements of an offence” on the conditional side of the law (“If ...”), the occurrence of which triggers the legal consequences (“... then”) referred to in the law. Since laws are necessarily abstract, the concepts used in them must be given a concrete form. In Germany, this purpose is served by a judicial “hermeneutics” (from the Greek word “hermeneuein” meaning explain, interpret) developed in the nineteenth century. It occupies centre stage in so-called “legal methodology”. With the categories of “grammatical”, “historical”, “systematic” and “teleological” “interpretation” (to which has been added in our age the “constitutional interpretation”), abstract concepts are to be treated in such a way that it becomes possible to reach a correct decision on “problem cases” – such as the question as to whether trampling a cross-country trail for skiers constitutes criminal “damage to property”. The “subsummation”, that is, the subjugation of a precise set of facts to the situation of a law, one thinks of in this context as a syllogical conclusion within the meaning of Aristotelian logic (Haft 2014, 80). The “type” that has already been mentioned presents itself as an alternative to “concept”. Whilst a concept is “defined” by certain characteristics (in the classical meaning of the word), which leads to a “yes/no decision” (electricity is an object – yes or no; a third possibility does not exist) the type is not definable but can only be illustrated by examples. It has certainly a firm core but no fixed boundaries so that one or the other of the characteristic “traits” for a type may be absent without the typicality of a certain situation having to be called into question. Concept … is closed, type is open. The concept knows only strict “either/or”, the concept separates, conceptual thinking is separation thinking. The type … on the other hand is satisfied “more or less” with multiple reality, it unites, makes common feelings conscious, the Universal is displayed in it, understood holistically. The type cannot therefore be subsumed like the concept, rather only allocated to a concrete situation to a greater or lesser degree, introduce it to analogy. (Kaufmann 1982, 310) Thinking in concepts is traditionally focussed on the judicial side of legal activity. It is top heavy because the factual side is blotted out. Without additional legal theoretical considerations, one assumes that every legal concept can be allocated to a large number of situations. An example is provided by the offence “Road traffic accident”, which is to be found in the law and the facts of which are presented in many different ways, for example, police reports or reports to the insurance company. These reports must, however, also be structured, even if no judicial preliminary consideration has yet occurred. The police officer who takes down the details of a traffic accident is not yet guided by the law but uses an intuitive method that meets quite different possible legal requirements, whether these are in terms of civil law 342



Dialogue and the Law

(e.g. compensation), administrative law (e.g. withdrawal of the driver’s licence) or criminal law (e.g. personal injury). In doing this he works with “factual concepts” that also serve as structural elements. Traditional legal theory has not paid any attention to them. That is currently changing with “e-justice”. Thinking in concepts is an authoritarian way of thinking. It aims for knowledge and truth and corresponds with position thinking referred to above. It leaves no space for dialogue. The verdict of the judge, who has made and announced his decision, is an expression of this thinking. This is different from thinking in types. This thinking is open to dialogue. It offers room for comparative cases and arguments. It is the thinking that in an ideal case helps active citizens to resolve their conflicts themselves – without the intervention of outside authorities. There are many possibilities to create relations between concepts and to generate structures. In the simplest case, there is a linear structure. The elements are linked to each other in a row and there is some structure that gives it order, a chronology for example or a causal relationship that joins them together. One alternative of this is the circular structure that in classical times had already been developed as the “Circle of Learning”, the principle of which is still evident today in the word – “Encyclopaedia” (“Circle of Learning”). In the Western tradition of the law, two types of structure have achieved particular importance, hierarchical “tree structures” and visual tabular structures.

21.6.2  Tree Structures “Concept trees” have been well known since classical times. They correspond with the way in which human knowledge is stored in the brain and then recalled. In linguistics, they play a central role in Chomsky, for example. The Roman jurist Gaius was the first to apply this thinking in hierarchies for teaching purposes in his textbook. On this basis, “concept jurisprudence” was created in Germany in the nineteenth century. One wished to erect legal “concept trees” and discover hidden “legal sentences” and “legal institutions” in them by way of conceptual “design”. This was regarded as a responsibility of legal science that was explicitly seen thereby as a third source of law in addition to the statute and legal judgements. However, this teaching is now considered to have been theoretically overcome. Internationally, it never imposed itself anyway, not even in other countries of continental Europe, which, like Germany, stand in the tradition of Roman law. Conceptual thinking that is alien to dialogue continues, however, to dominate German legal practice.

21.6.3  Tabular Structures “Tabular structures” were already used in classical times. They play a major role in several areas today. They enable the comparison of different texts on the same subject. Since they are two-dimensional, they have to be regarded. They can only be understood with great difficulty or not understood at all with language that runs in a single dimension. One cannot “think” a table. Modern tables make a revival of visual presentations possible, as they were distributed earlier. Modern computer technology has opened up new opportunities in the use of tables. One should mention here the spreadsheet and the relational data banks. The model of a relational data bank, which today is an established standard in information technology, was developed in 1970, and has largely replaced older hierarchical data bank models. A relational data 343

Fritjof Haft

bank contains tables in which data cells are stored. The data bank language SQL (Structured Query Language) is largely used for enquiries and data bank manipulations. It is enlightening to look at the basic distinctions in legal theory between the legal systems in continental Europe and the Anglo-Saxon world under this technical aspect. Whilst there (in continental Europe?) – based on the tradition of Roman law – one developed the tree structure, legal activity amongst them (in the Anglo-Saxon world?) was dominated by the search for “precedents” and therefore the comparison of cases. A “dogmatism” based on real concepts as developed there could not gain a footing here, because tables – it should be repeated – cannot be “thought”; one has to regard them. Instead of searching for knowledge through conceptual design, the doctrine of “stare decisis”, connecting cases on which judgements have been made, is followed here. Obviously, here too the judge needs a benchmark, a “tertium comparationis”, to be able to compare different cases with each other. This purpose is served by the rules that are derived from the cases. They are used to establish a reference to new cases that, seen from a concrete point of view, are quite differently composed but that are decided in accordance with the same “ratio decidendi”. In considering a new case, the judge therefore looks for the decision-making rule of a precedent. Should he be unable to find a suitable precedent, he judges in accordance with the rule that he considers fair and expedient. A simple example might make this clearer: A is travelling. B takes A’s lawnmower without seeking permission and uses it to mow his lawn. The lawnmower is broken whilst he is doing this. A demands that B buy him a new lawnmower. B refuses. A sues. The judge decides and establishes the following rule: anyone who uses another’s property without permission and damages this property in the process, must pay for the repair. Should now a similar case occur in which X uses Y’s bicycle without permission and damages it, the court will quote “A vs. B” and pass its judgement accordingly. Contrary to real conceptual thinking, this kind of thinking is dialogical thinking.

21.6.4  Man and Computer in Dialogue Finally, an example drawn from criminal law should make clear that in the law the computer too can be a partner in a dialogue. According to German criminal law, anyone who leaves the scene of a road accident without waiting for a period of time “adequate given the circumstances” is committing a crime. It is agreed to a large extent that a legal concept such as “adequate” defies all attempts to achieve automatic understanding. Is that really the case? An evaluation of relevant judgements (Gerathewohl 1987) has revealed that a whole range of circumstances may be relevant in appraising “adequacy” (“nature and gravity of the accident”, “type of damage”, “location of the accident”, “traffic density”, “weather”, “likelihood of individuals able to record the incident appearing”, “danger to the person involved” and a lot more besides). Of these, only a few in differing combinations came into question in the cases investigated. They cannot all be treated as yes/no circumstances but were gradual more-or-less circumstances (“minor damage”, “considerable damage” etc.) They formed hierarchies (e.g. “type of damage -> persons- / damage to property”) and were interlocked with each other). Overall, a multi-dimensional decision area appeared that can no longer be mastered in one’s head but which contains a system of rules that can easily be applied to new cases (“Whenever the combination X, Y, Z is present, 30 minutes are (not) adequate”). This system of rules could no longer be described verbally but only mathematically. If a new case were now to be entered into the system, it was possible to determine precisely its location in the system of previous legal judgements so that a judgement consistent with 344



Dialogue and the Law

the system could be reached. This could either be simply adopted or adapted on account of the particular circumstances of the new case. In the latter case, these circumstances were entered into the system and changed it. Thus, a system capable of learning was created that became “wiser” with each new case. The surprising aspect in this situation was that the legal instinct of the experienced lawyer leads to comparable results, without our being able to explain this rationally. This also applies elsewhere, for example, in determining the punishment that is dependent in the same methodological way on numerous circumstances. During my time as defending counsel in criminal cases, I have repeatedly found that all the jurists involved – in cases of fraud for example in which the penalty may range from one month to ten year’s imprisonment – could predict with amazing accuracy the verdict that the court would come to in a particular case. Every experienced criminal lawyer can do that. Our brain is therefore cleverer than our purely rational competence when it comes to the challenge of addressing the complexity of performance. Allow the computer to make us conscious of this and to help us in the process.

References Austin J. L. (1962). How to Do Things with Words. Oxford: Oxford University Press. Chomsky N. (1965). Aspects of the Theory of Syntax. Cambridge, MA: The MIT Press. Fisher R. (2011). Getting to Yes: Negotiating Agreement Without Giving In (3rd ed., rev. ed.). New York: Penguin. Gerathewohl P. (1987). Erschließung unbestimmter Rechtsbegriffe mit Hilfe des Computers: ein Versuch am Beispiel der “angemessenen Wartezeit” bei §142 StGB. Tübingen, Germany: Universität Tübingen. Gottwald W. and Haft F. (eds.) (1992). Verhandeln und Vergleichen als juristische Fertigkeiten (2nd ed.). Tübingen, Germany: Attempto-Verlag. Haft F. (1990). Aus der Waagschale der Justitia – Eine Reise durch 4000 Jahre Rechtsgeschichte (4th ed.). München: C.H. Beck. Haft F. (1993). Strukturdenken. Der Schlüssel zum Erfolg (Japanese edition). Frankfurt / Main, Germany: Ullstein. Haft F. (1999). Mediation- Palaver oder neue Streitkultur. In: Geimer R (ed) Wege zur Globalisierung des Rechts: Festschrift für Rolf A Schütze zum 65. Geburtstag. München: C.H. Beck, 255–268. Haft F. (2000). Verhandlung und Mediation – Die Alternative zum Rechtsstreit (2nd ed.). München, Germany: C.H. Beck. Haft F. (2014). Juristische Methodenschule: IT-gestütztes Training juristischer Fertigkeiten. München, Germany: Normfall. Haft, F. and Schlieffen, K. (eds.) (2009). Handbuch Mediation (2. Aufl). München: C.H. Beck. Ihering, R. V. (2012). Der Kampf um’s Recht. Hamburg, Germany: Tradition Classics. Kaufmann, A. (1982). Analogie und “Natur der Sache”: zugleich ein Beitrag zur Lehre vom Typus. Heidelberg: Decker & Müller. Larenz, K. (2014). Methodenlehre der Rechtswissenschaft. Berlin, Germany: Springer Berlin. Marinkovic, D. F. (2006). Sprache-Geltung-Recht. Heidelberg, Germany: Universität Heidelberg, Heidelberg. Müller, F. (1989). Untersuchungen zur Rechtslinguistik: interdisziplinäre Studien zu praktischer Semantik und Strukturierender Rechtslehre in Grundfragen der juristischen Methodik (Vol. 133). Berlin, Germany: Duncker & Humblot. Raiffa, H. (1982). The Art and Science of Negotiation. Cambridge, MA: Belknap Press of Harvard University Press. Rapaille, C. (2007). The Culture Code: An Ingenious Way to Understand Why People Around the World Live and Buy as They Do. New York: Broadway. 345

Fritjof Haft

Röhl, K. F. and Ulbrich, S. (2007). Recht anschaulich: Visualisierung der Juristenausbildung. Köln, Germany: Halem. Searle, J. R. (1969). Speech Acts: An Essay in the Philosophy of Language. London: Cambridge University Press. Weigand, E. (2005). Conflict resolution in Court. Studies in Communication Sciences. Special Issue: Argumentation in Dialogic Interpretation, 193–202. Weigand, E. (2010). Dialogue: The Mixed Game. Amsterdam and Philadelphia: John Benjamins.

346

22 How Culture Affects Language and Dialogue Marion Grein

22.1  The Impact of Culture First, culture pervades what we are as human beings, how we act, how we think and, thus, how we talk and even how we listen. Human beings are socialized into their specific cultural context(s), and culture influences their interaction with each other. Second, cultural diversity is above all a fact: we are nowadays confronted with a wide range of distinct cultures. Awareness of this diversity is facilitated by increased cultural contacts. Whenever people of different cultures communicate (even in the lingua franca English), problems within these intercultural dialogues typically arise when the communicators have widely different values and concepts of language-in-use. Even if both communicators are genuinely interested in communicating with each other, it can be difficult to secure successful communication if they have different values or beliefs about the world. A real-life example will unveil what I mean. Since 2010, I regularly travel to Uzbekistan and invite colleagues from our partner University in Tashkent to Mainz (Germany). I have made friends with a young female researcher, Dilfuza Yusupova, and thus, felt the urgent need to address her with the German ‘du’ (address pronoun you) instead of the more formal and distant ‘Sie’ (formal address pronoun you). In German, if you want to express closeness and sympathy, you jointly agree to switch from the more distant ‘Sie’ to the solidarity pronoun ‘du’. Actually, the Uzbek language does also differentiate between these two addressing pronouns, yet, the underlying concept of sympathy and closeness connected with the address pronoun ‘du’ does not exist. Often, even one’s husband is addressed with ‘Sie’ in Uzbekistan. The concept of ‘du’ and ‘Sie’, thus, is culturally encoded. Our cognition of the underlying concept, thus, has to be taken into account in language study (cf. ‘du’ and ‘Sie’ in Weigand, 2010: 102). The more globalized the world becomes, the more we need to come to an understanding of the diverging concepts of language use. I am, for instance, in charge of the Masters programme “German as a foreign language”, a graduate master course for language teachers, which is – along with teaching didactics – designed to communicate intercultural skills, a task that needs a profound approach to language and culture. When trying to analyse (and to teach) intercultural communication, we need concepts that allow us to get a deeper 347

Marion Grein

understanding of both the differences and similarities within language-in-use and differences and similarities of the communicators. The study of language, culture and cognition, however, nowadays seems to be fragmented into separate approaches. Most of these approaches add culture as a separate module, instead of integrating it into a holistic model of language analysis. The Mixed Game Model (MGM), which I will sketch out in Section 22.3, has proven to be a convincing model within my courses on cross-cultural communication, integrating the vast amount of otherwise juxtaposed approaches, which (a choice – I have to admit) will be introduced in Section 22.2. Communication – in accordance with Edda Weigand and her MGM – is always functional and dialogic. Thus, any analysis of language-in-use or communication needs to take as its basis the dialogic interdependence of an initiative speech act, which makes a pragmatic claim, and the reactive speech act, which is expected to fulfil this very claim, considering that both or all interlocutors have a basic interest. Language-in-use can solely be described within a holistic model, integrating multiple human abilities such as cognition and perception. Cognition and perception – as well as conventions and principles of language usage – are determined by personal preferences, the specific situation, prevailing emotions and the cultural background of the communicators (as demonstrated by the example from the German-Uzbek interaction). I will describe the underlying concept of the MGM (MGM, Weigand, 2010; cf. chapter 11, this volume) in detail in Section 22.3, believing that the MGM is the most adequate concept or model to analyse the role of culture within language-in-use. But let us first take a look at further approaches to the concept of culture and dialogue, mostly referred to as “intercultural communication”.

22.2  Concepts of Culture, Language and Dialogue I will first try to give a chronological overview and then summarize some familiar approaches combining language analysis and culture. Research on the relationship between language and culture began in the eighteenth century. Wilhelm von Humboldt (1767–1835), Franz Boas (1858–1942), Edward Sapir (1884–1939), Benjamin Whorf (1897–1941) and Leo Weisgerber (1899–1985) were famous scholars, who first emphasized the relatedness between language, thought and culture. The most well-known, yet controversially discussed, view being the Sapir-Whorf Hypothesis (the principle of linguistic relativity), conducted on the Hopi language of America. Benjamin Lee Whorf (1936) argued that the Hopi would lack a concept of time because there were neither words referring to time, nor grammatical markers for tense (“the Hopi have no concept of time” (1936: 57)). He concluded that people of different cultures (and languages) would perceive the same world differently. Actually, in 1983, Ekkehart Malotki published a huge study on the grammar of the Hopi language, refuting the view of the “non-existing” concept of time. In 1969, the again well-known and again controversially discussed study by Brent Berlin and Paul Kay on colour terminology (Basic Color Terms: Their Universality and Evolution) re-opened the discussion of linguistic relativity, claiming that the language spoken by a human being has an active influence on that human being’s perception and view of the world as such. Further studies were conducted, some proving the principle of linguistic relativity, others refuting it. Among the advocates of linguistic relativism, some argue that language 348



How Culture Affects Language and Dialogue

determines cognition (strong version), whereas others claim an influence of language on cognition (weak version). Ludwig Wittgenstein (1953) in his Philosophical Investigations argued that language is organized through rules, which are based on cultural use. In fact, he introduced the idea of ‘language games’ to discuss the cultural dimension of communication and language use (1953: 4). The term ‘intercultural communication’, however, is commonly associated with Edward T. Hall and his publication The Silent Language (1959). Most research in the 1960s and 1970s within the field of intercultural communication were strongly influenced by Hall. Hall’s most well-known cultural continuum or value is the discussion of high-context and low-context cultures (Hall, 1976). Low-context cultures, like Germany, place their information in the verbal language, thus communication needs to be specific and explicit. In high-context cultures, like Japan, meaning is strongly embedded in the context. Thus, the listener has to understand the contextual clues to grasp the full meaning of any utterance. The information is not explicitly expressed (cf. Ting-Toomey, 1985, 1999). Furthermore, Hall focused on “non-verbal” communication (the so far “hidden dimension” of communication), including the study of proxemics (personal space and territory), haptic, kinesics and paralanguage. In 1961, Kluckhohn and Strodtbeck discussed further cultural value orientations which provided basic conceptual contributions to the field. Basic values, they proclaimed, were humans’ relation with time (focus on past, present or future), natural environment (mastery, submission or harmony), and further factors concerning the human being (hierarchically departed or equal; human nature: good, evil or a mixture, doing versus being). In the 1970s, the field of intercultural communication grew rapidly. The aforementioned milder version of the Sapir-Whorf Hypothesis arose, stating that language, thought and perception are interrelated. In 1972, American anthropological linguists, most notably Dell Hymes and John J. Gumperz (i.e. Gumperz and Hymes, 1972), established a framework for the ethnography of communication, or culturally distinctive means of communication. This called for a shift from descriptive linguistics to sociolinguistics in functional approaches, while simultaneously Chomsky deepened his formal transformational generative (1965) approach (D-structure, S-structure); an approach Hymes strongly rejected, arguing that speaking is more than constructing grammatically possible linguistic utterances. Hymes was one of the first to argue that ungrammatical utterances could socially be appropriate, just as grammatical utterances could be socially inappropriate (Hymes, 1974). Instead of looking for linguistic competence, he urged to look at communicative competence in language analysis (ethnography of speaking or communication). Analysing language has to “discover and explicate the competence that enables members of a community to conduct and interpret speech” (Hymes, 1974: 52). He introduced the term “speech event”, defining “the term speech event will be restricted to activities that are directly governed by rules and norms for the use of speech. An event may consist of a single speech act, but will often comprise several” (1974: 56). The natural unit for sociolinguistic description, according to Hymes, is not the language but the speech community. Hymes argues in favour of a “theory of language use” (Hymes, 1974: 52), and thus could be understood as a precursor of the MGM. Yet, instead of taking the individual speaker and his or her abilities into account, he focused on the speech community. In 1973, Indiana University awarded the first doctoral degree in intercultural communication and an extensive series of publications followed, among them Barnlund’s Public and Private Self in Japan and the United States (1975), Dodd’s Perspectives on Cross-Cultural 349

Marion Grein

Communication (1977), Weaver’s Crossing Cultural Barriers (1978) and Asante, Blake and Newmark’s The Handbook of Intercultural Communication (1979). Dodd introduced yet another value continuum framework: linear versus non-linear thought. The linear cultural pattern being more object than person or event oriented, focusing on the beginning and the end of events. Linearity favours the presentation of single themes, one following the other, facts are presented sequentially, and logic and coherence is predominant. Non-linear cultures, in contrast, simultaneously discuss multiple themes and nonverbal communication plays an important role. Time plays a minor role and the pattern is person and event oriented. Emotionally resonant Americans are considered to be very linear, whereas Arabs for instance are people and event oriented. These frameworks of values were most prominent in the 1980s and still are very en vogue in economic sciences today. People of different countries have different concepts of value and human beings tend to believe that their values (i.e. freedom, equality, modesty, or just the ways of greetings) are fundamental to all human beings (ethnocentrism). In 1977, the first issue of the journal International and Intercultural Communication Annual appeared. Condon and Yousef’s (1975) Introduction to Intercultural Communication and Samovar and Porter’s (1973) Intercultural Communication: A Reader were considered basic works and influenced the 1980s, especially because they were the first who tried to systematize the field of investigation. Samovar and Porter strongly emphasize the interrelationship between culture and communication, and Condon and Yousef combine research done in linguistics and anthropology, focusing on cultural value orientation, which was then taken up by Hofstede (1984) in his comprehensive and still most cited study on cultural values, following the value approach by Kluckhohn and Strodtbeck (1961). Hofstede operated (and still does) with six cultural values, which are: 1. Individualism versus collectivism. 2. Uncertainty avoidance – “the extent to which people feel threatened by uncertainty and ambiguity and try to avoid these situations” (Hofstede, 1991: 113). 3. Power distance (power inequality and authority relations). 4. Masculinity versus femininity (dominant values in masculine countries are achievement and success vis-à-vis feminine countries in which caring for others is dominant). 5. Long-term versus short term orientation/or Confucian Dynamism. 6. Indulgence versus restraint (Hofstede, 1991, 2010).

His original extensive research data (1976–1973) were collected in forty countries and included 116,000 employees at IBM. The data are constantly updated. The most influential volumes published in the 1980s, next to Hofstede (1984), were Gudykunst’s Intercultural Communication Theory: Current Perspectives (1983), Gudykunst and Kim’s Methods of Intercultural Research (1984) and Theories in Intercultural Communication (1988) and Asante and Gudykunst’s Handbook of International and Intercultural Communication (1989), focusing on the communicative process and operating with models of communicative cultural styles (i.e. direct versus indirect, elaborated versus succinct, personal versus contextual, instrumental versus affective communication styles). These styles vary depending on the context, that is, the situation, personal distance and relationship of the interactants. Erickson (1989) added a new perspective by arguing that we interpret any given piece of information according to our experiences – experiences which are (also) culturally imprinted. Culture is considered a reason why a given phenomenon can have different 350



How Culture Affects Language and Dialogue

meanings – dependent on the culturally based perception of the phenomenon. An example might be such a common word or rather concept as “pride”, which is more or less negatively connoted in Germany whereas it is a very positive concept in Arabian countries (cf. Roche and Roussy-Parent, 2006). Concerning research methodology, quantitative and rhetorical-interpretive research methods were predominant in the 1980s. In the 1990s, linguistic qualitative and ethnographic methods of research slowly entered the field of study. The journal International and Intercultural Communication Annual promoted the use of methodological pluralism and the greater consideration of qualitative methods. Depending on the research question, studies on intercultural communication were influenced by disciplines like philosophy, psychology, sociology and anthropology – and of course different approaches of linguistics. Common ground for all approaches was or rather is the interaction between interlocutors from different cultural backgrounds in real-life situations. In the following years, many different approaches to intercultural communication analysis occurred. Qualitative research, as mentioned, entered the field of intercultural communication research and the role of context became of primary importance, recognizing that research of real-life communication needs to implement the situation and the people involved in the communicative act. Questions relating language-in-use to power and control, class, gender, the individual and ethnicity arose. Frequent research methods were ethnographic techniques, systematic observations, content analysis and in-depth interviews. Increasingly, research on nonverbal communication (especially gestures and smile) – studied as early as 1969 by Ekman and Friesen – were integrated on a broader basis. And, increasingly, the influence of culture in cognitive processes (cognitive linguistics) was discussed. Resnick (1991) argued that the social, cultural and neurocognitive realms should be integrated into the language analysis because these factors would be necessary for determining the speaker’s thoughts, emotions and actions. The term “cultural linguistics” can be linked to the cognitive linguist Ronald Langacker, who maintained “the advent of cognitive linguistics can be heralded as a return to cultural linguistics. Cognitive linguistic theories recognize cultural knowledge as the foundation not just of lexicon, but central facets of grammar as well” (Langacker, 1994: 31). Mainstream cognitive linguists focused on exploring the relationship between language and conceptualization. Gary B. Palmer, in Toward a Theory of Cultural Linguistics (1996), argued that cognitive linguistics could be directly applied to the study of language and culture. He claimed that all conceptual structures are very likely to have a cultural basis. These cultural-specific conceptualizations were referred to as imagery or image schemes. Major tenets of cognitive linguistic research believe(d) that meaning is the conceptualization of experience and that every grammar reflects these conceptualizations. Language diversity is, thus, a result of the different conceptualizations of culture-specific experiences. Linguists like Wierzbicka (1997) continued to emphasize the relationship between language and culture and introduced the field of ethnolinguistics (actually most synonymous with cultural linguistics) to explore the relationship between language and culture. Wierzbicka is in search of language universals (not in a Chomskian tradition) by working empirically with several languages, whittling down concepts of languages to key concepts (i.e. good and bad, I and you, far and near). Cultural concepts may show a certain degree of compatibility; however, each is specific to a specific culture. “Cultures can be interpreted in part through their key words” (Wierzbicka, 1997: 17). 351

Marion Grein

Nordby (2008) argues that language is shaped by overall beliefs and socio-cultural contexts. He relies on philosophical discussions on language and communication, foremost Wittgenstein’s philosophy on language, arguing that members of different cultures have their very own culturally shaped values (Nordby, 2008). Nordby (2008: 2) set up three different understandings of value: (a) “values are properties we ascribe to actions we think of as ethically good or wrong”; (b) “general concepts people believe in” and values are (c) “intimately connected to the ways individuals wish to live their own lives”. He focuses on the individual interpretation of values. Cliff Goddard (2006) coined the term “ethnopragmatics” (Goddard 2006: Ethnopragmatics: Understanding Discourse in Cultural Context), focusing on cross-cultural speech practices. With pleasure, he observed the shift from the universalist interpretation to a culture-internal perspective. His ideas are based on Wierzbicka’s concept of “cultural scripts” – the key concept for ethnopragmatics. Cultural scripts are intended as descriptions of commonly held assumptions about how “people think” about social interactions (cf. Goddard and Wierzbicka, 2004). The following script, to give an example, is intended to capture the Japanese cultural norm of apology: many people think like this: when something bad happens to someone because I did something, I have to (= can’t not) say something like this to this someone: “I feel something bad because of this” I have to (= can’t not) do something because of this Alongside the mentioned approaches, there is the social anthropological or sociolinguistic approach in line with Michael Agar’s (1994) languaculture.1 In accordance with Agar, language is equated with discourse (verbal interaction) and language, or rather discourse, is understood as both, a social and a cultural phenomenon. The term “languaculture” refers to “the necessary tie between language and culture” (Agar, 1994: 60). Culture is mostly understood as something that directly belongs to the individual’s consciousness. “Culture starts when you realize that you’ve got a problem with language and the problem has to do with who you are” (1994: 20). Thus, Agar takes a deeper look at semantic and pragmatic variability and seeks to explore so-called “rich points” (1996) in communication, that is, points where communication fails. Spencer-Oatey (2000), then, added a new definition of culture: Culture is a fuzzy set of attitudes, beliefs, behavioural norms, and basic assumptions and values that are shared by a group of people, and that influence each member’s behaviour and his/her interpretations of the “meaning” of other people’s behaviour. (Spencer-Oatey, 2000: 4; cf. Spencer-Oatey et al., 2009) Furthermore, obviously on the basis of Hofstede, she distinguishes fundamental levels or layers at which culture manifests itself, ranging from surface-level observable artefacts, less visible beliefs and attitudes and inner core basic assumptions and values (so-called onionmodel of culture). Palmer’s paradigm of cultural linguistics was the starting point for some cognitive linguists to examine the possible grounding of language in culture. Palmer’s approach was further developed by Sharifian (2011) in a multidisciplinary framework, combining approaches of, for instance, cognitive linguistics, cognitive anthropology, cognitive psychology and 352



How Culture Affects Language and Dialogue

anthropological linguistics. Sharifian’s framework cultural conceptualization and language views language as grounded in group-level cognition that emerges from the interaction of the members of the specific cultural group. These cultural conceptualizations are heterogeneously distributed across the members of the specific cultural group. Thus, Sharifian (2010, 2011) is well aware of the fact that human beings are culturally shaped individuals. Language is understood as a primary means for encoding cultural conceptualizations. The focus, however, is on words and their diverging concepts (schemas). So, for instance, the concept or schema of the word “family” is completely different in Germany, Uzbekistan and Tonga (cf. Völkel, 2010). Next to the above-mentioned approaches to intercultural communication, further fields of linguistic research (mostly all functional approaches) include the factor culture into their research questions. Culture was for instance seen as a main factor within discourse analysis, speech act theory (see the cross-cultural speech act realization program; Blum-Kulka et al., 1989) and pragmatics (cross-cultural or contrastive pragmatics). Whereas traditional pragmatics focused on the relation of signs and their users, “modern” pragmatics includes the study of politeness (Goffman, 1967; Brown and Levinson, 1987), speech act theory (Searle, 1969), deixis (as verbalization of the relation between speaker and environment) and information structure (as the relation between speaker and cognitive items). The “intercultural” or “contrastive turn” enforced the comparison of these principles between different cultures. The approach of linguistic anthropology or cultural anthropology can be considered a direct precursor of the MGM, discussed in Section 22.3. Anthropology takes culture as its primary concept and offers a holistic view, considering culture, history, language and biology to be essential for a complete understanding of society. The basic aim, thus, is not to gain insight into language itself, but to consider language as an integral part of “humankind”. In other words, anthropology seeks to understand human beings as whole organisms who adapt to their environments through a complex interaction of biology and culture. Culture is understood to be learned or acquired, shared among the members, and transmitted from one generation to the next, subjected to change. Within the so-called “second generation” of cognitive linguists, language and cognition are basically explored in the context of culture (cf. Sharifian, 2011: XVI). Sharifian (2011: 3) claims: Human conceptualisation is as much a cultural as it is an individual phenomenon. Members of a cultural group constantly negotiate “templates” for their thought and behavior in exchanging their conceptual experiences. Often complex cognitive systems emerge out of somehow concerted conceptualisations that develop among the members of a cultural group over time. Such conceptualisations give rise to the notion of cultural cognition. Research on the relationship between language, culture and cognition from the theoretical perspective of cognitive linguistics is published within the book series Cognitive Linguistic Studies in Cultural Contexts by Benjamins, edited by Yu and Sharifian. Kecskes (2014: 6), in his Intercultural Pragmatics, joins the “individualistic intention-based cognitive-philosophical line, and the societal, context-based socio-cultural-­ interactional line”, leading to a socio-cognitive approach, based on the Gricean maxims of language use. He analyses encounters in which interlocutors of different first languages and cultures communicate in a common language. Kecskes operates with three “crucial factors” (2014: 128) in intercultural pragmatics: (1) context – comprising world knowledge in one’s mind and “out there in the world”; (2) common ground (information the interactants share) 353

Marion Grein

and (3) salience as being culture-specific. Thus, according to Kecskes, the important factors are individual prior experiences and the actual social experience in any given situation. They are the basis for the understanding and interpreting of any utterance (egocentrism). He (2014: 5) supports the weak version of linguistic relativism, believing that members of a speech community have common cultural expectations and that these are “the main variables that motivate the use of available linguistic means”. Edda Weigand’s MGM takes culture, the interlocutors and their situational factors, their emotions and cognition into account, integrating them into a holistic system, focusing on action games, thus, creating one holistic system with human beings as its basis. In Section 22.3, the MGM by Weigand will be introduced, taking several steps of its development into account. In Section 22.4, the role of culture – as an integral part – will be sketched.

22.3  The Mixed Game Model By Edda Weigand Language-in-use can only be described within the framework of a holistic model, integrating the multiple human abilities such as cognition, perception, personal preferences, emotions, rational behaviour and cultural factors (cf. Damasio, 2000). Thus, human beings are not restricted to rules, but can go beyond rules and patterns. Instead of rules there are communicative techniques such as communicative means and principles. There is no object “language” as such in performance; there is only the human ability of speaking which however cannot be separated from other abilities: speaking is integrated with thinking and perceiving. (Weigand, 2010: 2) In the dialogic view the speaker is not facing the world, but another speaker within the world (cf. Weigand, 2009: 79): monologic view: speaker/language Û world

dialogic view: speaker1Ûspeaker2 world

Figure 22.1  Monologic vs. Dialogic View.

Fundamental premises according to Weigand (2009: 271) are: 1. Language is used by human beings and cannot be separated from them. 2. Human beings are oriented towards purposes or needs. Purposes (later also including interests) are a key concept for explaining human behaviour. 3. Human beings are always different human beings and do not only take different roles as speaker and interlocutor. 4. Human beings are social beings. They use their language or communicate in order to come to an understanding with other human beings, that is, they have to correlate and to negotiate their positions, tasks and interests. 5. Language use accepts misunderstanding. 354



How Culture Affects Language and Dialogue

6. For human beings, there is no independent world, no reality as such, only a world perceived by them. 7. The minimal communicatively autonomous unit is the action game, a unit of our cultural world which comprises the different communicative worlds of the interlocutors. 8. The action game is not constituted as a type of situation, but determined by its interactive purpose. 9. The authentic text represents a component in the action game. Human beings use different abilities together as communicative means. They produce verbal texts and simultaneously, and not separable, they draw inference and rely on what can be perceived in the speech situation. Language use therefore can in most cases be described only incompletely from the observer perspective and is only in part represented by corpora of authentic texts. 10. Not everything is said explicitly, not everything can be said explicitly because of the complexity of meaning to be negotiated.

One basic origin of the MGM is the question to what extent biology and culture interact, taking socio-biology as a starting point for language-in-use analysis (Weigand, 2007). Body and mind are interconnected (“Descartes’ error”). According to Weigand (2010: 4) the central linguistic interest is directed at describing and explaining how human beings, as cultural beings, succeed in coming to an understanding in human affairs. Instead of believing in rules, Weigand sets up principles of probability. Major principles of Weigand’s model are the Action Principle (AP) and the Dialogic Principle proper (DP). The AP states that we communicate because we have a specific communicative purpose and in trying to fulfil our communicative purpose we use different communicative or rather dialogic means (verbal, cognitive and perceptual means). The communicative function can be expressed by a set of numerous forms of utterances: F(p) Û dialogic purpose

utterance1 utterance2 utterance3 .......

Figure 22.2  Communicative Function and Variation of Utterances.

An example in Weigand (2009: 28) for instance is: REQUEST (close (x,y)) Û

Close the window. Can you close the window? When are you going to close the window? .......

Figure 22.3  Communicative Function Request and Variation of Utterances.

The different functionally equivalent utterances are for example dependent on the communicative situation. Weigand correlates purposes and means, so specific purposes call for specific means. However, alongside the situation, there are many differences due to different cognitive abilities, personality, emotions and so forth. Human beings have different presumptions and associations (cf. Dascal, 1994). Weigand (2009: 80) claims: “The dialogic relation is not only a 355

Marion Grein

situational one. Our communicative means of language, perception and cognition cannot be considered separately”. There are on the one hand innate differences and on the other hand differences acquired due to different experiences between the interactants. “Every human being brings their own cognitive horizons, preferences and emotions” (Weigand, 2009: 147). F(p)

⇔

utterances (direct, indirect, idiomatic) dialogic or communicative means (linguistic, visual, cognitive means)

Figure 22.4  Communicative Function and Variation of Utterances (Direct, Indirect, Idiomatic).

In 2000 (cited in 2009), Weigand revised the Action Principle to: dialogic purposes (state of affairs) making and fulfilling pragmatic claims

⇔ dialogic means verbal, cognitive, perceptual means

Figure 22.5  Action Principle.

In more recent versions (original 2004), Weigand introduced a super-ordinate predicate “interest” (cf. Weigand, 2009: 250): INTEREST [PURPOSE (state of affairs)] ⇔

communicative means (speech, perception, cognition)

Figure 22.6  Revised Action Principle.

The DP claims that communicative actions are dialogic actions, mutually dependent on each other. The minimal autonomous communicative unit, thus, is the interdependence of an initiative speech act which makes a pragmatic claim and the reactive speech act which is expected to fulfil this very claim (see Weigand, 2009: 170). The communicative purpose of any dialogue is to come to an understanding. The link between the initiative and the reactive speech act does not need to be explicit, but can also be inferred. Or as Weigand (2009: 160) puts it: “coherence is no longer given in the text, it is given by the interlocutors”. A well-known example of coherence was given by Brown and Yule (1983: 196), yet, left without any explanation: A: There’s the doorbell. B: I’m in the bath. Coherence is established in the minds of the interlocutors. As interpreters, our cognitive background and knowledge enables us to understand the minimal action game. The analysis in accordance to Weigand’s MGM would look like this (Weigand, 2010: 89). All speech acts are dialogic acts in a dialogic speech act taxonomy (see Chapter 11, this volume). action making a claim

⇔

reaction fulfilling this very claim

Figure 22.7  Dialogic Principle. 356



How Culture Affects Language and Dialogue

Table 22.1  Example Analysis Action Principle

Sentence type Action Principle

Dialogic Principle

The doorbell is ringing

I am in the bathroom

Declarative Representative (doorbell ringing) Directive (indirect) open the door Directive Û

Declarative Representative (being in the bathroom) Representative (indirect) unable to open the door Consent (negative)

The basic dialogic patterns being the following: Representatives (claim to truth) Û ± Acceptance Representative speech acts make a claim to truth, the reactive speech act, thus, is the speech act of accepting (or not accepting). A: Angela Merkel is doing a good job. B: I agree / No, I don’t agree. The reactive speech act “fulfils” the very claim of the initiative act in an abstract sense which can be positive or negative. Directives (claim to volition – practical action) Û ± Consent We have a specific initiative action, here the directive speech act, which makes a pragmatic claim with regard to the interlocutor. Its specific quality is that of a claim to the performance of a practical action, that is, “give me the book”. Either the book could be handed over without any further speech act (practical action) or by a reactive speech act which fulfils the claim in positive or negative form, called “consent”. Even if the interlocutor refuses to perform the action, they have come to an understanding. The abstract notation Directives Û Consent contains the positive as well as the negative point. A: Don’t forget to use the vacuum cleaner! B: Ok, I will / No, not today! Exploratives (claim to volition – knowledge) Û ± Response An explorative speech act aims for a response. It makes a claim to volition which relates to knowledge. A: Where will you spend your summer holidays? B: In Spain. Declaratives (claim to volition – creating social structures) Û [Confirm]

357

Marion Grein

In declarative speech acts (like “I baptize you”), making and fulfilling a claim coincide. Uttering a declarative speech act implies that something is made to exist or be valid. A reaction, thus, is not necessary. Small talk with the communicative purpose of establishing and maintaining social relations can be considered to be declarative as well (Weigand, 2010: 226). Weigand (2010: 154) summarizes: Action types which • • • •

create a piece of world are declaratives change the world are directives express the world are representatives and ask questions about the world are exploratives.

The basic action games have further subtypes (see Chapter 11, this volume). Directive subtypes can be set up by differentiating the directive claim to volition (cf. Weigand, 2010: 160): • the claim of an Order (or Iussive) defined by the availability of sanctions • the claim of a Request (or Monitive) relying on mutual cooperation • the claim of a Plea (or Petitive) appealing to the helpfulness and kindness of the interlocutor.

Representatives have the following subtypes (cf. Weigand, 2010: 163ff, 172), that will not be elaborated in detail: Table 22.2  Subtypes of representatives

Assertives Û Accept Identifiers Û Accept Informatives Û [Comment] Constatives external world Û Confirm internal world Û Believe Emotive Û Empathy Conditional Û Acceptance/Belief Deliberative Û Acceptance Desiderative Û Belief / Acceptance Normative Û Acceptance

Claim to truth which is not immediately evident; basis of reason and arguments. Claim to truth, based on the knowledge which refers to definitions. Aim at being taken notice of. Claim refers to truth which appears obvious.

Announcing or expressing emotions, emotives (in earlier versions expressives) aim at empathy and compassion. Modal claim to truth; conditionals aim at a speech act of acceptance or belief. Modal claim to truth; express the possibility of a state of affairs. Modal claim to truth; express a desire and include emotional involvement. Modal claim to truth; aim at acceptance of norms.

Weigand (2009: 170, 250ff) differentiated the following preliminary principles which were then modified in the MGM (2010):

358



How Culture Affects Language and Dialogue

Constitutive Principles Action Principle Dialogic Principle Coherence Principle

having a purpose/interest and applying communicative means dependency of initiative and reactive speech act coherence is not a matter of the verbal text alone but is ultimately established in the minds of the interlocutors

Regulative Principles Principle of Emotion

Principle of Rhetoric

previously: EmP previously: RatP

emotion Û reason innate and dependent on cultural conditions (i.e. “act rational and hide or control your emotions”) effectiveness Û respect/politeness interest of the EGO interest of the ALTER EGO regulation depends on cultural and individual ideologies Principle of Emotions; every dialogic interaction has an emotional influence Rationality Principle; a principle that tells us to be rational

Further Principles SeP DWP OSP SuggP ConvP

Sequencing Principle; guiding the dialogic sequence of one-phase and multi-phase dialogues Principle of Different Worlds; accepting that all human beings have different communicative worlds Open-System Principle; infinite world knowledge and accepting the risk of communicative failure Principle of Suggestions; preferences or habits that are only probably valid Conventionality Principle; principle that tells us to use common conventional ways to come to a mutual understanding

22.4  Culture Within the MGM Language-in-use is “open, dynamic, energetic, constantly evolving and personal” (Shohamy, 2007: 5) and each language user uses his or her language(s) differently. According to Weigand (2009: 309) the interrelationship between verbal (and nonverbal) interaction and cultural background represents a central issue of dialogue analysis. Weigand (2010: 5) even argues that we need to know other cultures and languages to be able to recognize the peculiarities of our mother tongue. There are innate differences between the interlocutors, there are differences acquired by socialization, learning and experience. Every human being brings in her or his very own cognitive ability, her or his very own emotions and individual preferences. Human beings perceive the world, at least to some extent, in their own unique ways. It is not the situation itself which is multicultural, but the interlocutors who perceive and evaluate it from the point of view of different cultures (cf. Weigand 2009: 309). Weigand (2009: 168) speaks about the “Principle of Different Worlds” (DWP). Language is a social practice in which to participate (cf. Kramsch, 1994), and language-in-use constantly is used within a cultural context. 359

Marion Grein

Let me give an authentic example of an action game in Japan: A: I’m in trouble! B: What happened? A: Forgot the shoes! Hope I don’t have to move out. It is difficult to understand what is actually going on in this action game if you have never rented a house or flat in Japan. At some point, A, a foreigner in Japan, forgot to take off his shoes, when he entered his own home. In Japan, the rental agreement forbids the tenant to enter his house or flat wearing shoes. Forgetting to take them off could end in a termination of the rental agreement. Cultural values deeply influence cognition and preferences. In any communication, at least two human beings bring with them different views of the world. In cross-­cultural communication, the views of the world might be even more diverging. Weigand introduces the cultural unit of the minimal action game (2009: 147). In her later publications, Weigand (2009: 251f.) operates with the two regulative principles emotion versus reason and effectiveness versus respect/politeness, which are individually and culturally qualified. Whereas for instance a compliment for a German is mostly perceived as something positive, it may be embarrassing for a Japanese when expressed within a group. Yet, the value or rather the perception of the value of, for example, a compliment, is, as elaborated, not only culturally determined but depends on the individuality of the speaker. Another authentic example might be helpful. In 2011, we had a visiting lecturer from Rwanda at our University. Two weeks after his arrival, he approached me and said (in German): Marion, today I have the urgent need to tell you, that you have the walking movements of a pregnant cow. Actually, I had no idea what to reply and felt rather personally offended. I then had to learn that it was a very positive compliment. In Rwanda, and many African states, a cow is a beauty ideal and telling a woman that she walks like a cow is a sincere compliment (cf. Sundermeier, 1991: 133). Coherence problems can occur in intercultural communication situations. Take for instance the following dialogue: A: Did you finish your book in the bathtub? B: No, it became too cold after a while. A Japanese could hardly infer the link between the two utterances. Whereas in Germany, water is filled into the bathtub and then begins to cool down, the temperature remains constant in Japanese bathtubs (ofuro). The inference of the rapid cooling down of the bath water is totally unknown to the Japanese interlocutor. Another example on the formulaic level: In Japan, if you want to encourage your students to keep up the good work, you say “ganbatte kudasai” which literally means “hold on; go for it, work hard”. At my old University in Tokyo, the professor discouraged some foreign students translating it literally as to “you have to work harder”. Language analysis, thus, has to consider the interplay of various facets. Any utterance depends on individual evaluation. Evaluation, then, happens in the mind of the interactants and is, alongside individual experiences, dependent on cultural conditions, like for instance the evaluation of the German du and Sie mentioned at the very beginning of this article. The superficially identical concept is evaluated differently and thus not identical. 360



How Culture Affects Language and Dialogue

Emotions or rather the display of emotions and culture are again interdependent. The “Principles of Emotion” (EmP) are, according to Weigand (2009: 172), always in conflict with the “Principle of Rationality” (RatP). Seen as a continuum, there are the two poles of emotions totally under control by rationality and emotions totally dominating, with various cases in between with emotions and rationality in conflict. These principles are again differently manifested within each “cultural” human being. In Japan, depending on the social distance of the interlocutors, emotions are not displayed unless the interlocutors are very familiar, in Germany – I would argue – we tend to show emotions, yet, not as openly as for instance in Brazil. After a good oral exam for instance, our Brazilian students usually start crying for happiness, Germans show their happiness, whereas the Japanese smile very much the same as prior to the exam. Weigand (2010: 53) sets up the following holistic concept: human nature ∩ environment ∩ culture human action and behaviour

dialogic action action by communicative means dialogic purposes

practical action action by practical means

mental action action in the mind

practical purposes

cognitive purposes

Figure 22.8  Holistic Concept of the Action Principle.

So how can we use the MGM as a means of comparing language-in-use between two languages? The minimal autonomous unit is the action game, consisting of initiative and reactive speech act. We have to consider the aforementioned fundamental premises and take into account that human beings are individuals and individuality might play a more crucial role than the cultural imprint. We have to keep in mind that human beings use different abilities as communicative means. When analysing (written) authentic linguistic data in corpora, non- and para-verbal information might be missing as well as the profound knowledge of the situation. The context plays an important role and is dependent on personal and culture-specific factors. Culture can thus be grasped as part of inner human nature as well as of the external environment. It is however not only the cultural environment that shapes human beings’ behavior. The culture-independent natural environment is of enormous influence as well insofar as it imposes certain exigencies of human life. (Weigand, 2010: 55) Culture takes hold of human beings in the shape of internalized attitudes and preferences which are not easily changed. Activities and attitudes are perceived as culture specific (cf. Weigand, 2010: 69). In Japan, where I grew up, the value of me as an individual was considered with respect to the role I played within the group; in Germany, my home country, my very own independence was considered the most important part of my individuality. Both concepts are still alive in me, each depending on the specific situation (environment) I am in. 361

Marion Grein

To use the MGM for intercultural comparison, it is reasonable to compare action games that have a more or less equivalent point of origin, that is, the same dialogic purpose and a similar situational background and/or context. How could one proceed? One could analyse authentic action games in a first language and then compare them to authentic action games in a second language or take action games that have been recognized as miscommunication (due to cultural differences). As an example, I will choose the minimal action game of directives in German and Japanese (cf. Grein, 2007a). As mentioned, we have a specific initiative action, which makes a pragmatic claim to volition. Its specific quality is that of a claim to the performance of a practical action, that is, “get my clothes from the dry cleaner’s”. Either the clothes are picked up without any further speech act (practical action) or a reactive speech act follows which fulfils the claim in positive or negative form, called consent. Even if the interlocutor refuses to perform the action, they have come to an understanding. A first cultural difference is apparent in that positive consent is far more familiar in Japanese minimal action games than in German action games, depending, moreover on the interlocutors – in this case a relative. One could argue that it is more akin to a request (monitive) in Japan, whereas it might be a plea (petitive) in German. Refusing or declining to help a relative is far more uncommon in Japan than in Germany, relying more on mutual cooperation in Japan. The claim, in both cultures, can be uttered directly or indirectly, using several forms of toning down. There could be hints and allusions. F(p) Û dialogic purpose DIRECTIVE REQUEST/PLEA Û (fetch (x,y))

utterance1 utterance2 utterance3 ....... Fetch my clothes from the dry cleaner’s! Could you please fetch my clothes from the dry cleaner’s? Would you be so kind as to fetch my clothes from the dry cleaner’s? Would it be a great problem for you to fetch my clothes from the dry cleaner’s? Be a nice girl and get my clothes from the dry cleaner’s! I would very much appreciate if you could fetch my clothes from the dry cleaner’s! When could you fetch my clothes from the dry cleaner’s? It would be wonderful if someone could fetch my clothes from the dry cleaner’s Are you passing by the dry cleaner’s today? Perhaps you could fetch my clothes from the dry-cleaner’s …

CONSENT (positive or negative)

Figure 22.9  Example Dialogic Principle (Request).

On the one hand, one can compare the different variants of utterances. On the other, one can compare the various alternatives of reaction on the level of verbal means. And taking individual and cultural imprints into account, one can easily detect that the evaluation of the relationship is different in Germany and Japan. Declining the request or plea seems easier

362



How Culture Affects Language and Dialogue

in Germany. In Japan, the relative is part of the in-group and thus the positive fulfilment of the claim is most probable. One can and should compare the use of the various subtypes of the basic action games. In my research conducted in 2007, it was interesting to see that Germans and Japanese use different subtypes within the very same situation. While Germans frequently used assertives and constatives as a reaction to the request or plea, the Japanese refrained from using assertives especially because they could evoke contradiction, something that entails disharmony (cf. Grein, 2007b: 102). Taking further basic authentic minimal action games into account will make it evident that the choice for the reaction is not only culturally, but very often individually imprinted. To highlight the cultural differences, I will give two further examples of replies demonstrating the differences in evaluating the balance between self-interest and social concern (individuality versus respect). Situation A: A salesman is trying to sell you a new product. Both the German and the Japanese refuse to buy the product. A: (German): Nein, danke. Kein Interesse (No. Thank you. I am not interested.) B: (Japanese): Chotto kangaesasete kudasai (Let me think about it a moment.) Situation B: You come late to an official meeting. A: (German): Mein Auto ist liegen geblieben und dann musste ich den ADAC einschalten und der hat das Auto dann abgeschleppt. Sorry. (My car broke down and I had to call the automobile club to tow the car away. Sorry.) B: (Japanese): Osoku natte sumimasen. Gomeiwaku wo okakeshite moshiwakearimasen (I am very very sorry to be late. I very much apologize for the inconvenience and causing trouble.) Whereas the German minimal action game focuses on an explanation, the Japanese is using very polite versions of excuses – mostly avoiding any explanation.

22.5 Conclusions Within the MGM, language-in-use is described within the framework of a holistic model, integrating the multiple human abilities such as cognition, perception, personal preferences, emotions, rational behaviour and culture. The regulative principles of self-interest versus social concerns, reason versus emotion and thus, the differences of cognition and evaluation have to be taken into account. Instead of rules there are communicative techniques such as communicative means and principles. These are influenced by culture – however, individuality also plays a major role. In the future, we must further analyse authentic data applying the MGM model of Edda Weigand. The study of intercultural communication will then no longer be one separate module of language analysis, but an integral part. Furthermore, we must bear in mind that culture is only one facet affecting language and dialogue.

Note 1 The term ‘languaculture’ is borrowed from Friedrich (1989) who coined the term linguaculture.

363

Marion Grein

References Agar M. H. (1994). Language Shock: Understanding the Culture of Conversation. New York: Wm. Morrow. Agar M. H. (1996). The Professional Stranger: An Informal Introduction to Ethnography (2nd edn.) New York: Academic Press. Asante M. K. and Gudykunst W. (1989). Handbook of International and Intercultural Communication. Beverly Hills, CA: Sage Publication. Asante M. K., Blake C. A. and Newmark E. (1979). The Handbook of Intercultural Communication. Beverly Hills, CA: Sage Publication. Barnlund D. C. (1975). Public and Private Self in Japan and the United States. Tokyo, Japan: The Simul Press. Berlin B. and Paul K. (1969). Basic Color Terms: Their Universality and Evolution. Berkeley, CA: University of California Press. Blum-Kulka S., House J. and Kasper G. (1989). Cross-Cultural Pragmatics: Requests and Apologies (Appendix: the CCSARP coding manual.). Norwood, NJ: Ablex. Brown G. and Yule G. (1983). Discourse Analysis. Cambridge, UK: Cambridge University Press. Brown P. and Levinson S. (1987). Politeness: Some Universals in Language Usage. Cambridge, UK: Cambridge University Press. Chomsky N. (1965). Aspects of the Theory of Syntax. Cambridge, MA: MIT Press. Condon J. C. and Yousef F. S. (1975). An Introduction to Intercultural Communication. New York: Macmillan. Damasio A. (1999). The Feeling of What Happens: Body and Emotion in the Making of Consciousness. [2000: Ich fühle also bin ich. Die Entschlüsselung des Bewusstseins]. New York: Harcourt Brace. Dascal M. (1994). Speech act theory and Gricean pragmatics: Some differences of detail that make a difference. In: Savas LT (ed.) Foundations of Speech Act Theory: Philosophical and Linguistic Perspectives. London and New York: Routledge, 323–334. Dodd C. H. (1977). Perspectives on Cross-Cultural Communication. Dubuque, IA: Kendall Hunt. Dodd C. H. (1991). Dynamics of Intercultural Communication. Dubuque, IA: Wm. C. Brown Publishers. Ekman P. and Friesen W. (1969). The repertoire of non-verbal behaviour: Categories, origins, usage and coding. Semiotica 1: 49–98. Erickson F. (1989). Métodos cualitativos de investigación sobre la enseñanza. In: Wittrock M (ed.) La Investigación de la Enseñanza, II. Métodos Cualitativos y de Observación. Barcelona, Spain: Paidós/M.E.C., 195–301. Friedrich P. (1989). Language, ideology, and political economy. American Anthropologist 91: 295–312. Goddard C. (ed.) (2006). Ethnopragmatics: Understanding Discourse in Cultural Context. Berlin and New York: Mouton de Gruyter. Goddard C. and Wierzbicka A. (2004). Cultural scripts: What are they and what are they good for? Intercultural Pragmatics 1–2:153–166. Goffman E. (1967). Interactional Ritual: Essays on Face-to-Face Behaviour. Harmondsworth, UK: Penguin Books. Grein M. (2007a). Kommunikative Grammatik im Sprachvergleich: Die Sprechaktsequenz Direktiv und Ablehnung im Deutschen und Japanischen. Beiträge zur Dialogforschung. Tübingen, Germany: Niemeyer. Grein M. (2007b). The speech act of refusal within the Minimal Action Game. A comparative study of German and Japanese. In: Grein, M. and Weigand E. (eds.) Dialogue and Culture. Amsterdam and Philadelphia: John Benjamins, 95–113. Grein M. and Weigand E. (eds.) (2007). Dialogue and Culture. Amsterdam and Philadelphia: John Benjamins. Grice H. P. (1975). Logic and conversation. In: Cole P. and Morgan J. L. (eds.) Syntax and Semantics. Vol. 3: Speech Acts. New York: Academic Press, 41–58. 364



How Culture Affects Language and Dialogue

Gudykunst W. B. (1989). Intercultural Communication Theory. Current Perspectives. Beverly Hills, CA: Sage. Gudykunst W. B. (1993). Toward a theory of effective interpersonal and intergroup communication. In: Wiseman R. L. and Koester J. (eds.) Intercultural Communication Competence. London: Sage, 33–71. Gudykunst W. B. (1994). Bridging Differences: Effective Intergroup Communication, 2nd edn., Thousand Oaks, CA: Sage. Gudykunst W. B. and Kim Y. Y. (eds.) (1984). Methods for Intercultural Communication Research. Beverly Hills, CA: Sage. Gudykunst W. B. and Kim Y. Y. (1988). Theories in Intercultural Communication. Beverly Hills, CA: Sage. Gudykunst W. B. and Kim Y. Y. (1992). Communicating with Strangers: An Approach to Intercultural Communication. New York: McGraw-Hill. Gudykunst W. B. and Ting-Toomey S. (1988). Culture and Interpersonal Communication. London: Sage. Gumperz J. J. and Hymes D. (eds.) (1972). Directions in Sociolinguistics: The Ethnography of Communication. New York: Holt, Rinehart, Winston. Hall E. T. (1959). The Silent Language. New York: Anchor Books. Hall E. T. (1976). Beyond Culture. New York: Garden City. Hofstede G. (1984). Culture’s Consequences: International Differences in Work-Related Values. Newbury Park, CA: Sage. Hofstede G. (1991). Cultures and Organizations – Software of the Mind. New York: McGraw Hill. Hofstede G. (2010). Cultures and Organizations – Software of the Mind, 3rd edn. New York: McGraw Hill. Hymes D. (1972). Models of the interaction of language and social life. In: Gumperz J. J. and Hymes D. (eds.) Directions in Sociolinguistics: The Ethnography of Communication. New York: Holt, Rinehart & Winston, 35–71. Hymes D. (1974). Foundations in Sociolinguistics: An Ethnographic Approach. Philadelphia, PA: University of Pennsylvania Press. International and Intercultural Communication Annual (Journal). London: Sage Publication. Johnson M. (1987). The Body in the Mind: The Bodily Basis of Meaning, Imagination, and Reason. Chicago, IL: University of Chicago Press. Kecskes I. (2014). Intercultural Pragmatics. New York: Oxford University Press. Kluckhohn C. and Strodtbeck F. (1961). Variations in Value Orientations. Evanston, IL: Row Peterson. Kramsch C. (1994). Foreign languages for a global age. ADFL Bulletin vol. 25: 5–12. Lakoff G. and Johnson M. (1980). Metaphors We Live By. Chicago, IL: Chicago University Press. Langacker R (1994). Culture, cognition, and grammar. In: Pütz M (ed.) Language Contact and Language Conflict. Amsterdam & Philadelphia: John Benjamins, 25–53. Levine D. (1985). The Flight from Ambiguity. Chicago, IL: University of Chicago Press. Malotki E. (1983). Hopi Time. Trends in Linguistics: Studies and Monographs 20. The Hague, Netherlands: Mouton. Mathiot M. (ed.) (1979). Ethnolinguistics: Boas, Sapir and Whorf Revisited. The Hague, Netherlands: Mouton. Nordby H. (2008). Values, cultural identity and communication: A perspective from philosophy of language. Journal of Intercultural Communication, 17 (6), Available at: www.immi.se/intercultural/nr17/nordby.htm [Accessed 20 September 2016]. Palmer G. B. (1996). Toward a Theory of Cultural Linguistics. Austin, TX: University of Texas Press. Palmer G. B. (2003). Talking about thinking in Tagalog. Cognitive Linguistics 14/2,3: 251–280. Resnick L. B. (1991). Shared cognition: Thinking as social practice. In: Resnick L. B., Levine J. M. and Teasley S. D. (eds.) Socially Shared Cognition. Washington, DC: American Psychological Association, 1–20. Risager K. (2006). Language and Culture: Global Flows and Local Complexity. Clevedon, UK: Multilingual Matters. 365

Marion Grein

Roche J. and Roussy-Parent M. (2006). Zur Rolle der kontrastiven Semantik in interkultureller Kommunikation. FLUL 25, Sprachdidaktik – interkulturell. Tübingen, Germany: Gunter Narr, 228–250. Samovar L. and Porter R. (1973). Intercultural Communication: A Reader. Belmont, CA: Wadsworth Publishing Company. Samovar L., Porter R. and Jain N. (1981). Understanding Intercultural Communication. Belmont, CA: Wadsworth Publishing Company. Samovar L., Porter R. and Stefani L. (1998). Communication between Cultures. Belmont, CA: Wadsworth Publishing Company. Searle J. (1969). Speech Acts. Cambridge, UK: Cambridge University Press. Sharifian F. (2010). Cultural conceptualizations in intercultural communication: A study of Aboriginal and non-Aboriginal Australians. Journal of Pragmatics 42: 3367–3376. Sharifian F. (2011). Cultural Conceptualisations and Language: Theoretical Framework and Applications. Amsterdam and Philadelphia: John Benjamins. Sharifian F. (2012). Cultural linguistics. In: Chapelle C. A. (ed.) The Encyclopedia of Applied Linguistics. London: Wiley-Blackwell. Sharifian F. (2013). Cultural linguistics and intercultural communication. In: Sharifian F. and Jamarani M. (eds.) Language and Intercultural Communication in the New Era. Oxford, UK: Routledge. Shohamy E. (2007). Language Policy: Hidden Agendas and New Approaches. London: Routledge. Spencer-Oatey H. (ed.) (2000) Culturally Speaking: Managing Rapport through Talk across Cultures. London: Continuum. Spencer-Oatey H. and Franklin P. (2009). Intercultural Interaction. A Multidisciplinary Approach to Intercultural Communication. Basingstoke, UK: Palgrave Macmillan. Sundermeier T. (1991). Weisheit in den afrikanischen Religionen. In: Klimkeit H.-J. (ed.) Biblische und außerbiblische Spruchweisheiten. Studies in Oriental Religions. Wiesbaden, Germany: Harrassowitz, 117–134. Ting-Toomey S. (1985). Toward a theory of conflict and culture. In: Gudykunst W. B., Stewart L. B. and Ting-Toomey S. (eds.) Communication, Culture, and Organizational Processes. Beverly Hills, CA: Sage, 71–86. Ting-Toomey S. (1999). Communicating Across Cultures. New York: Guilford Press. Völkel S. (2010). Social Structure, Space and Possession in Tongan Culture and Language. An Ethnolinguistic Study. Culture and Language Use, 2. Amsterdam and Philadelphia: John Benjamins. Weaver G. (1978). Crossing Cultural Barriers. Dubuque, IA: W. B. Brown. Weigand E. (2000, 2009). Dialogue analysis 2000. Towards a human linguistics. In: Weigand E. (2009). Language as Dialogue. Amsterdam and Philadelphia: John Benjamins, 283–294. Weigand E. (2004, 2009). The end of certainty in dialogue analysis. Opening plenary lecture at the international IADA conference on “Theoretical Approaches in Dialogue Analysis” held at the Northeastern University of Chicago. Printed 2009 in Weigand E., Language as Dialogue. Amsterdam and Philadelphia: John Benjamins, 241–258. Weigand E. (2007). The sociobiology of language. In Grein M. and Weigand E. (eds.) Dialogue and Culture. Amsterdam and Philadelphia: John Benjamins, 27–49. Weigand E. (2009). Language as Dialogue. Amsterdam and Philadelphia: John Benjamins. Weigand E. (2010). Dialogue. The Mixed Game. Amsterdam and Philadelphia: John Benjamins. Whorf B. L. (1936). The punctual and segmentative aspects of Hopi. Language 12: 127–131. Wierzbicka A. (1997). Understanding Cultures through Their Key Words: English, Russian, Polish, German, and Japanese. New York: Oxford University Press. Wittgenstein L. (1953). Philosophical Investigations. Oxford, UK: Basil Blackwell. Yu N. and Sharifian F. (eds.) Cognitive Linguistic Studies in Cultural Contexts (book series). Amsterdam and Philadelphia: John Benjamins.

366

Outlook

Towards the Unity of Knowledge The three parts of the Handbook describe decisive steps in the history of modern linguistics, which changed our view of language and dialogue from an abstract model of competence to a holistic model of performance. It is interesting to remind ourselves that de Saussure, even if focusing on la langue, was already aware of the fact that there was not only the abstract system of la langue but also its application in la parole. However, he did not dare to bridge the gap between competence and performance but introduced the generic term “le langage” to cover both. The problem was thus already envisaged at the very outset of modern linguistics and hovered above the models which followed. The crux of the issue lies in the question of how to combine rules of competence with non-conventional individual moves of performance. The addition of orthodox rule-­governed models and models of individual inferences did not yield theoretically acceptable results. Complexity does not mean combining heterogeneous models or accepting chaos but finding out the “laws of chaos” (Prigogine, 1994). Among the basic laws are fundamental premises such as: • The object-of-study needs to be an autonomous object. In the end, we have to refer to the human mind as the centre of all our actions and behaviour. • The complex whole is more than the sum of parts. • Integration is the name of the game. • We have to start from the complex whole and derive all its components by specialization. • The key to opening up the complex whole of human action and behaviour is human beings’ nature as goal-oriented beings and as dialogic individuals.

As indicated in the first premise, it is human beings who bridge the gap between rules of competence and probabilities of performance in their minds by their extraordinary ability of competence-in-performance. Fortunately, the brain is no longer a black box, and neuroscience is now capable of verifying assumptions on human action and behaviour. By relating competence to performance, we have arrived at a fundamental turning point in science which forces us to go beyond rules and to address complexity. The natural sciences settled 367

Edda Weigand

this issue by the transition from classical physics to modern physics. For the humanities and social sciences this turning point means the change from Cartesian linguistics to a postCartesian approach. Eventually we have dared to address complexity and explain what we as human beings do when we use language to come to grips with the vicissitudes of life. We orient ourselves according to rules as far as they go and adapt to ever-changing particular cases by using principles of probability. The key to opening up the complex whole is not speaking but negotiating our mutual goals in dialogue with our fellow-beings. We do this by action and reaction to arrive at some mutual understanding. It goes without saying that the challenge of complexity and post-Cartesian thought does not represent a challenge for single disciplines such as physics or linguistics, but is a challenge for all the disciplines engaged in investigating human action and behaviour. This challenge can only be met by genuinely interdisciplinary work. In this sense, individual disciplines are considered to be integrated dimensions “of a single connected structure” as represented in Gell-Mann’s staircase model (1994: 111): ↑ the humanities and social sciences: psychology, linguistics, sociology, cultural ­studies, rhetoric, philosophy, ethics etc. ↑ anthropology ↑ ……… ↑ biology ↑ chemistry ↑ physics In the end, it is human beings’ minds which connect all the variables of such a complex whole but also set the limits for possible results. We must accept that there are boundaries to our human abilities of perception and thinking. The question of absolute truth belongs to the Cartesian period of the past.

References Gell-Mann M. (1994). The Quark and the Jaguar: Adventures in the Simple and the Complex. London: Abacus. Prigogine I. (1994). Les Lois du Chaos. Paris, France: Flammarion.

368

Index

action tradition 94 Adams, J. L. 294 affective neurosciences 282 Agar, M. H. 352 “agents of change,” audience as 303 ALICE chatbot system 147 Allen, J. F. 152 Allwood, J. 9, 15 alterity 109–11 Alternative Dispute Resolution (ADR) 330 anaphoric chain 199 anatomo-functional hypothesis 267–8 application program interface (API) 144 appreciative capacity 68 Arnett, R. C. 317 Asante, M. K. 350 Atkinson, J. M. 30 Austin, J. L. 164, 311, 331 Bach, K. 82, 87 background emotions 282 Bakhtin, M. M. 110, 130, 304 Barnlund, D. C. 349 Bitzer, L. 304 Blake, C. A. 350 Blyth, C. 261 Boas, F. 348 Bohus, D. 154 Bondi, M. 46 Brennan, S. E. 93, 103 Brigham Young University Corpus of Contemporary American English 48 British Academic Spoken English Corpus (BASE) 48 Burke, E. 131 Burke, K. 297 Carlson, L. 133 Cartesian dualism 277, 281 Cassell, J. 153 Cazden, C. 255 chatbot systems 147 Chomsky, N. 2, 176, 197, 234, 247, 252, 331

Churchill, W. 313 Clark, H. H. 93, 109 COBUILD project 51 cognitive linguistics 351 Cohen, P. R. 149 coherence 183, 311 COLT corpus 48 commissives 190 common ground fallacy 245 communicational criticism 138 communicative action, theory of (Habermas) 169–71 communicative competence 349 competence-in-performance 176, 298 components of dialogic interaction 275–6 computational approaches to dialogue 143–61; adjacency pairs 146; application program interface 144; applications for dialogue systems 145–6; chatbot systems 147; dialogue games 147; dialogue manager organizing principles 146–51; dialogue modeling 143; dialogue moves 150; dialogue system 143; early fusion 154; embodied conversational agents 153; explicit conformation 152; finite-state networks 148; formal representations 150; frame model of context 148; global and grammar-based structural approaches 147–8 grounding, feedback and repair 151–2; human computer dialogue 144–6; implicit conformation 152; Information State Approach 149–50; intelligent agents 145; interface systems 145; issues for dialogue systems 151–5; late fusion 154; logic and plan-based approaches 149; Markov decision process 150; medical corpus analysis 143; meeting summarization 143; missing axiom theory 149; mixed initiative 147; multimodal dialogue 153–4; natural language understanding task 147; non-cooperative dialogue 153; partially observable Markov decision process 150; selection algorithm 150; simulation type systems 145; speech to speech translation 143; statistical reward models 150–1; Turing Test 145; update rules 150; virtual humans 153

369

Index

concept trees 343 “conceptual jurisprudence” 333 Condon, J. C. 350 consilience 193 context-dependence 200 conversation analysis (CA) 26–45, 52, 165; forms of analysis 31–2; key notions 32–6; methodology 28–32; naturally occurring social interactions, recordings of 29–30; origins and history 26–8; overall organization 35–6; sequences, organization of 33–5; transcription 30–1; trends 36–42; turns-at-talk 32–3 Cooren, F. 308 corpus linguistics 46–61; “Flagstaff school” 55; interactivity in spoken and written discourse 52–4; and language in use 51–2; and language variation across speech and writing 54–6; organizational units vs. message-oriented units 52; semantic preference 51; spoken corpora, corpus annotation and corpus tools 47–50; Text Encoding Initiative 49 Cottingham, J. 279 Craig, R. T. 307, 313–14 cross-recurrent gaze analysis 104 cultural display rules 286–7 culture, effects on language and dialogue 347–66; cognitive linguistics 351; communicative competence 349; consent 362; culture, language and dialogue 348–54; culture within MGM 359–63; example 347; imagery 351; impact of culture 347–8; “intercultural communication” 348; languaculture 352; Mixed Game Model 354–8; onion-model of culture 352; “rich points” 352 Damasio, A. 281 DART (Dialogue Annotation and Research Tool) 49 de Saussure, F. 197, 234 Descartes, R. 278, 280 Dialog Grammar 162–73; Dialog Grammar (Hundsnurscher) 163; argumentation criteria 170; argumentation sequence 166–7; bargaining speech act sequence 167–8; conversation analysis 165; dialog typology 168–9; International Phonetic Alphabet transcriptions 165; semantic primitives 162; theory of communicative action (Habermas) 169–71; theory of dialog (Weigand) 171–2; traditional Speech Act theory 163–4 dialogical self theory (DST) 120 Dialogic Principle Proper 252, 301 dialog typology 168–9 dialogue, from pragmatics to 78–92; “collective intentionality” 80; dialogue approach 79–81; discourse analysis 78; need for dialogue approach 84–5; pragmatic theories 81–4; 370

salience effect in dialogic sequence 85–9; utterance-based inquiry 78 dialogue games 147 dialogue management 143 dialogue manager organizing principles 146–51; principle-based approaches 148–51; structurebased approaches 146–8 dialogue modeling 143 dialogue system 143 Dickens, C. 313 Dictionary of Military Terms 321 DIFA (double initial finite auxiliaries) 118 discourse analysis 62–77; appreciative capacity 68; background 62–3; conversation (extended sense) 68–9; design 67–8; extension of 64–7; face-to-face verbal conversations 69–71; hermeneutics 62; information unit 66; literacy, problem of 74; “probe cycle” 68; situating meaning task 64; things designed 75; Thomas Was Alone (game) 72; ‘unified theory’ of 63–4; and video games 71–4 Discourse Representation Theory (DRT) 198 Dodd, C. H. 349–50 Drew, P. 30 Du Bois, J. 116 Duranti, A. 121 Dynamic Syntax (DS) 203–11; discontinuities as underspecification and update 209–10; parallelism of resolving underspecification to syntax 210–11; split utterances 207–9; trees, formal properties of 204–7 Dynamic Systems Theory 261 embodied conversational agents 153 emotion, reason and the human brain 277–93; affective neurosciences 282; background emotions 282; bilingual/bi-cultural speakers, emotion studies and 287; Cartesian dualism 277, 281; cultural display rules 286–7; Damasio’s approach 281–3; Descartes’ view of the brain 278–81; emotion talk vs. emotional talk 286; event schema, emotion as 286; functional magnetic resonance imaging 282; gender studies 287; historical research 278; how-to books 278; language analysis and emotions 283–7; lexical models 284; linguistic approaches 277; MGM 287–90; morpho-syntax 285; nonverbal means, emotions and 285; “social emotions” 282; somatic markers 281; speech emotion analysis 283–4; universalism-relativism dispute 285–6 “empirical audience” 303 Enfield, N. J. 222 explicit conformation 152 exploratives 190 Fairhurst, G. 308 Feller, S. 162, 318



“Flagstaff school” 55 Foucault, M. 307 Fox Tree, J. E. 101 Fries, C. C. 29 functional grammar 216 functional magnetic resonance imaging (fMRI) 282 Gadamer, H.-G. 312 Garfinkel, H. 26 gaze tracking 99–100 Gee, J. P. 62 “generative semantics” 15 Goffman, E. 115 Goodin, G. 130 Goodwin, Candy 29 Goodwin, Charles 29 Gothenburg Spoken Language Corpus 21 grammar see Dialog Grammar; utterance grammar, concept of language in Grein, M. 277, 347 Grice, P. 17 Grosz, B. J. 149 grounded theory 313 grounding 96, 151 Günthner, S. 217 Habermas, J. 169 Haft, F. 330 Hall, E. T. 320, 324 Hanna, J. E. 93 Haverkamp, B. E. 308 Haviland-Jones, J. M. 283 Hellermann, J. 257 hermeneutics 62 heteroglossia 111 historical linguistics 9, 22 Hjelmslev, L. 237 holistic theory see Mixed Game Model Hong Kong Corpus of Spoken English (HKCSE) 48 Horn, L. R. 17, 82 Horvitz, E. 154 human brain see emotion, reason and the human brain human computer dialogue 144–6 human conversation, turn-taking during 270 Hundsnurscher, F. 133, 164, 169 Hymes, D. 236 Iacoboni, M. 264 “illocution” 311 implicit conformation 152 incrementality 201–3 indirect speech acts 225, 230 Information State Approach 149–50 information unit 66

Index

Initiation-Response-Evaluation (IRE) 254 Initiation-Response-Feedback (IRF) 255–60 institutions, dialogue in 318–29; controversy 323; culture, influence of 326–8; formal and informal communication 320–2; functions of institutional communication 322–3; hidden knowledge 324; hierarchical institutions 321; impersonal language 320; institutional action games 320–2, 325–6; institutional knowledge 323–5; medicine 322 intelligent agents 145 interactive syntax, modelling of see shifting concepts of language “intercultural communication” 348 interface systems 145 International Phonetic Alphabet (IPA) transcriptions 165 intersubjectivity in dialogue 109–26; and alterity 109–10; alterity, polyphony, heteroglossia 110–11; complementarity between perspectives 110; dialogical self theory, 120; DIFA (double initial finite auxiliaries) 118; forms 112–13; interactivities 120; and objectivity 122; “problem of other minds” 113–15; as property of interpersonal interaction 115–19; and subjectivity 120–1; “Theory of Mind” 113; third parties 112; and trust 111–12 Jack, R. E. 283 Jackendoff, R. 15 Jefferson, G. 26 jurisprudence of precautionary measures 339 Kamp, H. 198 Kant, I. 132 Kasper, S. 217 Kaufmann, A. 342 Keats, J. 313 Kecskes, I. 78, 81 Kempson, R. 197 Kinzel, T. 129 Kleene star operators 205 knowledge: hidden 324; institutional 323–5; mutual 199; pattern 242; unity of 367–8 Koch, K. 287 Koike, D. 251, 261 Kretschmar, W. A. 175 Kristeva, J. 261 Kubanyiova, M. 308 Kuhlen, A. K. 103 Labo, W. 234 Lakoff, G. 15 languaculture 352 “language acquisition device” (LAD) 13 language-as-action tradition 94 language-as-system 197

371

Index

law, dialogue and 330–46; complexity 332–3; computer 344–5; consensus games 336–41; language and law 331–2; position games 333–6; structured thinking 341–5; tabular structures 343–4; tree structures 343 LeDoux, J. 285 Lee, H. 315 Lejeune, G. 137 Levesque, H. J. 149 Levinson, S. C. 15, 17 Lewis, M. 283 lexical decision task 95 LINDSEI corpus 48 Linell, P. 109 literacy: activities presupposing forms of 113; notion of 55; problem of 74 literature, dialogue and 127–42; communicational ethics 136–40; dialogue in literature 128–31; literary dialogicality 136–40 Livnat, Z. 53 London-Lund Corpus (LLC) 47 Long, M. 255 Lumsden, C. J. 323 Mackey, A. 254 Malebranche, N. 278 Markov decision process (MDP) 150 Matheson, C. 152 McCawley, J. D. 15 medical corpus analysis 143 meeting summarization 143 Mehan, H. 182 message-oriented units 52 Mey, J. 16 Mildorf, J. 129 mirror neurons see sociobiology of language missing axiom theory 149 Mixed Game Model (MGM) 174–94, 219, 287–90; claim to volition 187; commissives 190; communicative actions 181; competencein-performance 176; complexity, coping with 177–9; consilience 193; constitutive principles 180–4; culture 354–8; ethical principles 185; executive principles 185–6; exploratives 190; goal-directed, social beings, human beings as 180; issue 174–6; language as dialogic action 179–86; minimal versus complex games 191–2; multi-phase games 192; natural language use 175; neo-Gricean principles 174; one-phase games 192; Originative House of Dialogic Ethics 310; pragmatics 174; Principle of Coherence 183; principles of probability 180; rationality-in-performance 299; regulative principles 184–5; Rhetorical Principles 185; science, specific scientific interests of 174; selfinterest, respect and 184; speech act taxonomy 186–91; utterances 180; what constitutes theory 176–7 372

mixed initiative 147 Mondada, L. 26 Montague, R. 15 morpho-syntax 285 motor evoked potentials (MEP) 268 motor theory of speech perception 267 multi-phase games 192 Mustajoki, A. 234 mutual knowledge 199 Nassaji, H. 255 natural language understanding (NLU) task 147 naturally occurring social interactions, recordings of 29–30 natural semantics 221 neural mirroring and speech 268–9 Newmark, E. 350 Nottingham MultiModal Corpus (NMMC) 50 offline language 235 Olbrechts-Tyteca, L. 303 one-phase games 192 onion-model of culture 352 online language 235 organizational units 52 Originative House of Dialogic Ethics, language as 307–17; call of responsive action, linguistic contextual weight as 313–14; dialogue and generative linguistics 310–13; ethics and linguistic inquiry 308–9; grounded theory 313; heteroglossia 312; “illocution” 311; integration, constitutive principles of 311; linguistic weight and height 314–16; Mixed Game Model 310; “negative capability” 313; Principle of Coherence 311; research design 309; Speech Act Theory 311 Page, N. 128 Palmer, A. 129 partially observable Markov decision process (POMDP) 150 Pascal, B. 278 pattern knowledge 242 Peirce, C. S. 13 Perelman, C. 303 periaqueductal grey (PAG) 283 point of disambiguation (POD) 99 politeness principle 296 polylogues 261 polyphony 111 PRAAT software 257 PRACMA system 153 pragmatics 9–25; communication 9; context, relationship of pragmatics to, 19–21; decontextualized view of language 13; “deep structures” 12; formal semantics, Peirce and 13–14; “generative semantics” 15; Gothenburg Spoken Language Corpus 21;



historical linguistics 9, 22; “ideal speaker– hearer,” model of 13; “language acquisition device” 13; non-structuralist linguistics 17–19; post-structuralist linguistics 21–2; structuralism 9–13; “universal grammar” 13; see also dialogue, from pragmatics to precision grip 265 Principle of Coherence 183, 260, 311 “problem of other minds” 113 product tradition 94 psycholinguistic approaches 93–108; action tradition 94; coordinating meaning, signals for 100–2; cross-recurrent gaze analysis 104; entrainment 96; gaze tracking 99–100; grounding 96; language-as-action tradition 94; lexical decision task 95; new psycholinguistics 98–104; online coordination between speakers and addressees 102–4; point of disambiguation 99; product tradition 94; traditional psycholinguistics 94–8; visual world paradigm 99–100 “psycho-socials” 338 Putnam, L. 308 Raiffa, H. 339 Rapaille, C. 230 reason see emotion, reason and the human brain Richards, I. A. 297 Rommetveit, R. 109–10 Rowley-Jolivet, E. 53 Sacks, H. 26, 242 Safdar, S. 286 Sapir, E. 348 Saul, J. 82 Scheff, T. 116 Schegloff, E. 26 Schegloff, E. A. 109, 242 Schiffrin, D. 53 Schubert, L. K. 152 science, specific scientific interests of 174 Scott, W. 128 Searle, J. 311, 331 Searle, J. R. 166, 181 selection algorithm 150 self-interest and social concerns 294–306; “agents of change,” audience as 303; argumentation 301–3; audience 303–5; competence-in-­ performance 298; “dialogical principle proper” 301; “empirical audience” 303; Mixed Game Theory 295–7; persuasion 297–301; politeness principle 296; rationality-in-performance 299; “symbolic action” 297 Sell, R. D. 127, 134, 312–13 semantic preference 51 SGML (Standard Generalized Markup Language) 49

Index

Sharifian, F. 352 shifting concepts of language 197–213; anaphoric chain 199; bound pronouns 199; contextdependence 200; Discourse Representation Theory 198; Dynamic Syntax 203–11; ellipsis 200; grammar versus language processing 197–8; incrementality 201–3; Kleene star operators 205; language-as-system 197; longdistance dependency 211; mutual knowledge 199; p­ roblems in modelling natural-language ­semantics 198–203; split utterances 201–3 Sidner, C. L. 149 simulation type systems 145 situational norm 237 Skinner B. F. 13 “social emotions” 282 sociobiology of language 264–74; anatomofunctional hypothesis 267–8; background emotions 282; ‘canonical’ neurons 265; conversation ­270–1; empathy 270–1; human conversation, turn-taking during 270; imitation 270–1; implicit control, need for 271–2; mirror neurons 265–7; motor evoked potentials 268; motor theory of speech perception 267; neuroscience studies on neural mirroring and speech 268–9; precision grip 265; “social emotions” 282; TMS and the simulation model 269–70; transcranial magnetic stimulation 268; ­whole-hand prehension 264 somatic markers 281 Speech Act theory 149, 162, 311; see also Dialog Grammar speech emotion analysis 283–4 speech to speech translation 143 Spencer-Oatey, H. 352 Sperber, D. 17 split utterances 201–3 state of the art 1–7 statistical reward models 150–1 structuralism 9–13, 17, 22 “substantive law” 334 Survey of English Usage (SEU) 47 Tardy, C. 54 Taylor, T. 110 TEI (Text Encoding Initiative) 49 theoretical key issues and open debates 195–6 theorising 234–50; common ground fallacy 245; communication, constituents influencing the course of 241; communication-oriented research 240–3; communicative capacities, ways to study 244; dichotomy of language 235–6; evidence-based methodology 238; linguistic methodology (general questions) 237–8; linguistic methodology (languageoriented research) 238–40; methodologyvs. phenomenon-driven research 246–7;

373

Index

norm 236–7; objects of research 243–5; offline language 235; online language 235; pattern knowledge 242; self-presentation, need for 243; situational competence 242; situational norm 237 “Theory of Mind” (ToM) 113 theory and practice 251–63; Action-Reaction 255; applied linguistics 253–6; applying holistic theory to practice 260–1; Dialogic Principle Proper 252; examples that question the IRF 256–60; future research areas 261; holistic stances 252–3; Initiation-Response-Evaluation 254; Initiation-Response-Feedback 255; Interaction Hypothesis 255; polylogues 261; teaching and learning of languages 253–6; theories of dialogue 251–3; traditional stances 251–2 Tollins, J. 101 Toulmin, S. 301 Tracy, K. 307, 313–14 transcranial magnetic stimulation (TMS) 268 transcription 30–1 transition-relevance place (TRP) 33 translation,Verbmobil system 143 Traum, D. 143, 152 trees, formal properties of 204–7 trust, intersubjectivity and 111–12 Turing Test 145 turn-constructional unit (TCU) 33 turns-at-talk 32–3 unity of knowledge 367–8 “universal grammar” 13 utterance grammar, concept of language in 214–33; adaptation 220; arguments 222;

374

collocations 226; correlation of meaning and expression 224–9; dialogic means of expression 223–4; functional grammar 216; fundamentals 217–21; grammar and performance 214–17; indirect speech acts 225, 230; language comparison 227; lexicon 225–8; Mixed Game Model 219; natural semantics 221; new types of grammars 216; particles 228; research guidelines 231; shifts in the concept of language 214; speaking and perceiving 229; speaking and thinking 229–30; structure of meaning 221–2; syntax 224–5; unit 218 van Lier, L. 255 Verbmobil translation system 143 video games, discourse analysis and 71–4 virtual humans 153 visual world paradigm 99–100 von Humboldt, W. 175, 348 Wandruszka, M. 227 Weaver, G. 350 Weaver, R. 297 Weigand, E. 133, 171, 254–5, 287, 295, 310 Weisgerber, L. 348 Wells, G. 255 whole-hand prehension 264 Whorf, B. 348 Wilce, J. M. 283 Wilson, D. 17 Wilson, E. O. 310, 323 Wittgenstein, L. 349 Wordsworth, W. 313 Yousef, F. S. 350