Building Coherence and Cohesion
Pragmatics & Beyond New Series Editor Andreas H. Jucker University of Zurich, English Department Plattenstrasse 47, CH-8032 Zurich, Switzerland e-mail:
[email protected] Associate Editors Jacob L. Mey University of Southern Denmark
Herman Parret Belgian National Science Foundation, Universities of Louvain and Antwerp
Jef Verschueren Belgian National Science Foundation, University of Antwerp
Editorial Board Shoshana Blum-Kulka Hebrew University of Jerusalem
Catherine Kerbrat-Orecchioni University of Lyon 2
Jean Caron Université de Poitiers
Claudia de Lemos University of Campinas, Brazil
Robyn Carston University College London
Marina Sbisà University of Trieste
Bruce Fraser Boston University
Emanuel Schegloff University of California at Los Angeles
Thorstein Fretheim University of Trondheim
Deborah Schiffrin Georgetown University
John Heritage University of California at Los Angeles
Paul O. Takahara Kansai Gaidai University
Susan Herring University of Texas at Arlington
Sandra Thompson University of California at Santa Barbara
Masako K. Hiraga St.Paul’s (Rikkyo) University
Teun A. Van Dijk Pompeu Fabra, Barcelona
David Holdcroft University of Leeds
Richard J. Watts University of Berne
Sachiko Ide Japan Women’s University
Volume 129 Building Coherence and Cohesion: Task-oriented dialogue in English and Spanish by María Teresa Taboada
Building Coherence and Cohesion Task-oriented dialogue in English and Spanish
María Teresa Taboada Simon Fraser University
John Benjamins Publishing Company Amsterdam/Philadelphia
8
TM
The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences – Permanence of Paper for Printed Library Materials, ansi z39.48-1984.
Library of Congress Cataloging-in-Publication Data Taboada, María Teresa Building coherence and cohesion : task-oriented dialogue in English and Spanish / María Teresa Taboada. p. cm. (Pragmatics & Beyond, New Series, issn 0922-842X ; v. 129) Includes bibliographical references and index. 1. English language--Discourse analysis. 2. English language-Grammar, Comparative--Spanish. 3. Spanish language--Grammar, Comparative--English. 4. Spanish language--Discourse analysis. 5. English language--Spoken English. 6. Spanish language--Spoken Spanish. 7. Conversation. I. Title. II. Pragmatics & beyond ; new ser. 129. PE1422.T33 2004 420.1’41-dc22 isbn 90 272 5372 2 (Eur.) / 1 58811 563 1 (US) (Hb; alk. paper)
2004057694
© 2004 – John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Co. · P.O. Box 36224 · 1020 me Amsterdam · The Netherlands John Benjamins North America · P.O. Box 27519 · Philadelphia pa 19118-0519 · usa
TSL[v.20020404] Prn:28/07/2004; 16:58
F: PB129CO.tex / p.1 (v)
Table of contents
List of figures ix List of tables xi Abbreviations and conventions xiii Preface xv Chapter 1 Introduction Chapter 2 A framework for the analysis of speech genres 2.1 Genre defined 7 2.1.1 Bakhtin’s speech genres 8 2.1.2 Register analysis, functional genre theory and generic potential 10 2.1.3 Genre and register 17 2.1.4 Genres, frames, scripts and schemata 20 2.1.5 Genres and prototype theory 23 2.1.6 A working definition of genre 25 2.2 Applications of generic analysis 29 2.3 Components in a generic analysis of conversation 32 2.4 Scheduling dialogues as genre 34 Chapter 3 Data description 3.1 The task 38 3.2 Recording 40 3.3 Transcription conventions 41 3.3.1 Human noises 42 3.3.2 Silence 42 3.3.3 Mispronunciations 42 3.3.4 False starts 43 3.3.5 Transcriber comments 43 3.3.6 Unit markers 44
1 7
37
TSL[v.20020404] Prn:28/07/2004; 16:58
F: PB129CO.tex / p.2 (vi)
Table of contents
3.3.7 Other conventions 44 3.3.8 Summary of transcription conventions 45 3.4 Speaker pairs and dialogues 45 3.5 Use of terms and translations 49 3.6 Other remarks on the corpus 51 Chapter 4 The thematic structure of dialogue 4.1 The many accounts of Theme-like concepts 55 4.2 Thematic realization in English and in Spanish 64 4.2.1 English 64 4.2.2 Spanish 67 4.3 Thematic realization in scheduling dialogues 72 4.3.1 Two examples 73 4.3.2 Thematic selection patterns 75 4.3.3 Thematic selection and Transitivity 80 4.4 Definition and articulation of thematic progression 84 4.5 Thematic progression in scheduling dialogues 93 4.6 Thematic progression and genre 100 4.7 Summary 102 Chapter 5 Rhetorical relations in dialogue 5.1 Rhetorical relations and text analysis 105 5.2 Mann & Thompson’s Rhetorical Structure Theory 108 5.3 Rhetorical relations in the present study 112 5.4 RST in conversation 114 5.5 Results: Turn-by-turn analysis 121 5.5.1 Background 121 5.5.2 Concession 123 5.5.3 Condition 125 5.5.4 Elaboration 126 5.5.5 Joint 127 5.5.6 Non-Volitional Cause 128 5.5.7 Non-Volitional Result 129 5.5.8 Restatement 130 5.5.9 Volitional Result 131 5.6 Results: Conversation-as-a-whole analysis 132 5.6.1 Evaluation 140
53
105
TSL[v.20020404] Prn:28/07/2004; 16:58
F: PB129CO.tex / p.3 (vii)
Table of contents
5.6.2 Solutionhood 142 5.6.3 Restatement 143 5.7 Discourse markers 144 5.8 Summary 151 Chapter 6 Cohesion in dialogue 6.1 A brief introduction to cohesion 155 6.1.1 Texture and structure, coherence and cohesion 156 6.1.2 Types of cohesive relations 159 6.1.3 Distance of cohesion and cohesive chains 166 6.2 Cohesion in scheduling dialogues 169 6.2.1 Cohesion types 169 6.2.2 Distances: Types and length 175 6.2.3 Chains: Types and length 176 6.3 Summary 179 Chapter 7 The generic structure of scheduling dialogues 7.1 Stages in scheduling dialogues 181 7.2 Speech acts 188 7.3 Development of stages through speech acts 7.4 Thematic structure and staging 196 7.5 Rhetorical relations and staging 199 7.6 Cohesion and staging 201 7.7 Characterization of stages 205 7.8 Summary 205 Chapter 8 Conclusions and consequences Appendix A Speech act inventory a.1 Accept 211 a.2 Ask-date 212 a.3 Backchannel 212 a.4 Channel 213 a.5 Filled-pause 213 a.6 Goodbye 215 a.7 Greeting 215
155
181
190
207 211
TSL[v.20020404] Prn:28/07/2004; 16:58
F: PB129CO.tex / p.4 (viii)
Table of contents
a.8 Inform 216 a.9 Inform-availability 216 a.10 Other 217 a.11 Politey 217 a.12 Propose-action 218 a.13 Propose-place 218 a.14 Reject-date 218 a.15 Repeat-confirm 219 a.16 Request-action 219 a.17 Request-confirmation 220 a.18 Request-date 220 a.19 Request-information 220 a.20 Request-meeting 221 a.21 Request-place 221 a.22 Self-introduction 221 a.23 Vocative 222 References
223
Name index
245
Subject index
251
TSL[v.20020404] Prn:27/07/2004; 9:59
F: PB129LF.tex / p.1 (ix)
List of figures
1.1 Summary of cohesion devices 2.1 Choices of mode 2.2 Relationship of genre and register according to Eggins and Martin (1997) 2.3 Hasan’s realization of genre and register 2.4 Relationship of schemata to genres and process of creation of both 2.5 Relationship of genre, register and language 2.6 Steps in a contrastive generic analysis of conversation 4.1 System network for textual meaning in the clause 4.2 Distribution of marked and unmarked Themes 4.3 Cline according to the presence of marked Themes 4.4 Theme realization and Transitivity (as in Table 4.7) 4.5 TP patterns according to Daneš 4.6 Split Rheme pattern 4.7 Thematic progression patterns 4.8 Thematic progression patterns (as in Table 4.8) 4.9 Percentages of derived versus not derived Themes 5.1 Examples of the five main schema types 5.2 Rhetorical relations in Example (78) 5.3 Two possible analyses for Example (89) 5.4 Rhetorical relations in Example (110) 5.5 Rhetorical relations in Example (111) 5.6 Rhetorical relations in Example (120) 5.7 Rhetorical relations in Example (121) 6.1 The three levels of coding in the language 6.2 Summary of cohesion types 6.3 Cohesion types, graphical representation 6.4 Cohesive chains in Example (154) 7.1 Sequencing of stages in scheduling dialogues 7.2 Flow of conversation in the Task-Performance stage
3 12 18 18 22 28 35 61 78 79 82 87 89 92 95 97 109 117 124 134 139 146 147 159 166 170 178 182 186
TSL[v.20020404] Prn:27/07/2004; 9:59
F: PB129LF.tex / p.2 (x)
List of figures
7.3 7.4 7.5 7.6 7.7
Sequence of speech acts in the Opening stage Sequence of speech acts in the Date-Proposal stage Sequence of speech acts in the Date-Proposal stage, continued Sequence of speech acts in the Place-Proposal stage Sequence of speech acts in the Closing stage
191 193 194 194 195
TSL[v.20020404] Prn:27/07/2004; 10:01
F: PB129LT.tex / p.1 (xi)
List of tables
2.1 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 3.11 4.1 4.2
A contextual configuration Intonation markers Summary of transcription conventions Dialogues in the English corpus Dialogues in the Spanish corpus Characteristics of the English female speakers Characteristics of the English male speakers Characteristics of the Spanish female speakers Characteristics of the Spanish male speakers Counts in the English and Spanish corpora Average length of English and Spanish dialogues Average length of English and Spanish dialogues by gender Summary of Theme-related concepts Unmarked Theme selection in Mood context. Fragment, from Matthiessen (1995) 4.3 Unmarked Theme selection in mood context in Spanish 4.4 Numbers of SDUs and clauses in English and Spanish 4.5 Simple and multiple Themes 4.6 Marked and unmarked Themes 4.7 Theme Realization and Transitivity. Percentages with respect to N 4.8 Thematic progression in the corpus 4.9 Linear versus constant patterns 4.10 Subclassification of the multiple type 5.1 Organization of the relation definitions in RST 5.2 Number and percentages in the turn-by-turn analysis 5.3 Number and percentages of the analysis of whole conversations 5.4 Occurrence of discourse markers 5.5 Equivalences of numbers and relations in Tables 5.8 and 5.9 5.6 Discourse markers in the English corpus 5.7 Discourse markers in the Spanish corpus
15 43 45 46 46 47 48 49 50 50 51 51 58 65 71 73 76 79 81 94 96 97 110 122 141 149 151 151 152
TSL[v.20020404] Prn:27/07/2004; 10:01
F: PB129LT.tex / p.2 (xii)
List of tables
6.1 6.2 6.3 6.4 6.5 7.1 7.2 7.3 7.4
Cohesion types in the corpus Ratios of cohesive links to words in the corpus Direction and distances of links in the corpus Average directions and distances per conversation Average number and length of chains Frequencies of speech acts in the corpus Ratio of speech acts to SDUs Cohesion chains in Example (177) Thematic, rhetorical and cohesive characteristics of stages
170 172 175 175 177 190 190 203 205
TSL[v.20020404] Prn:27/07/2004; 10:05
F: PB129ABR.tex / p.1 (xiii)
Abbreviations and conventions
| F M R/Rh RST R> SDU SEOS SFL T T/Th TP U
Boundary between SDUs; end of SEOS At the beginning of speaker’s identification (FBNT) – Female At the beginning of speaker’s identification (MJFG) – Male Rheme Rhetorical Structure Theory Register and Genre Theory Semantic Dialogue Unit Semantic End Of Segment (represented by |) Systemic Functional Linguistics In tables – total number of occurrences (raw count) Theme Thematic Progression Utterance
Theme Textual Theme Interpersonal Theme Ideational Theme Cohesive distance I M R K
Italics Bold Underlined
Immediate Mediated Remote Cataphoric
TSL[v.20020404] Prn:27/07/2004; 10:05
F: PB129ABR.tex / p.2 (xiv)
Abbreviations and conventions
Cohesive ties R1 Reference, personal R2 Reference, demonstrative R3 Reference, comparative S1 Substitution, nominal S2 Substitution, verbal S3 Substitution, clausal E1 Ellipsis, nominal E2 Ellipsis, verbal E3 Ellipsis, clausal L1a Lexical, same, identical L1b Lexical, same, rephrased L2 Lexical, synonym L3 Lexical, superordinate L4 Lexical, subordinate L5 Lexical, general word L6 Lexical, collocation Transcription Conventions . (period) Falling intonation , (comma) Slightly rising intonation, continuation of idea, and not a question ? (question mark) Marked rising intonation Human noises /. . . / slashes Non-human noises #. . . # hash marks/pound signs Silences *. . . * asterisks Mispronunciations [. . . ] square brackets (around whole word) (. . . ) parentheses (supply missing part of word or correct pronunciation of word, only inside square brackets) Transcriber comments {. . . } curly braces Accent |. . . | vertical bars/pipes False starts angled brackets
Preface
The goal of this book is to examine some of the resources that speakers employ when building conversations. These resources contribute to overall coherence and cohesion, constructed interactively as speakers build on each other’s contributions. The book furthers the study of coherence at all levels, from surface structure to internal organization. Typical dialogues are considered to be coherent by those participating, and often by those observing as well. However, it is difficult to pinpoint exactly what makes a dialogue or a conversation coherent, and how that coherence is achieved dynamically. Similarly, cohesion, as an internal property of togetherness in a text, is usually considered to be present in conversation. Coherence and cohesion together give us texts and conversations that we recognize as such. In this book, I present a study of coherence and cohesion. The study is cross linguistic, based on a corpus of task-oriented conversations between dyads of two speakers each. The corpus contains conversations between two native speakers of English on the one hand, and two native speakers of Spanish on the other. The interlocutors have conflicting agendas that cover a two to four week period; their task is to agree on an appointment within that time. The conversations are viewed as collaborative efforts, yielding texts built interactively by the speakers. My analysis is concerned with the discourse characteristics that make the dialogues hold together as texts. The framework of the study is the analysis of speech genres (Bakhtin 1986). These conversations are interpreted as instances of a genre, as a “staged, goaloriented, purposeful activity in which speakers engage as members of our culture” (Martin 1984: 25). The conversations can be divided into three main stages: Opening, Task Performance and Closing. Each of these stages contains different linguistic and discourse characteristics: thematic progression, rhetorical structure analysis and cohesion. These three are unified under the textual metafunction of language in the Hallidayan tradition (Halliday & Hasan 1976). The structural component of this metafunction is the thematic structure of the clause. The non-structural component is cohesion. A primary element of cohesion is conjunction, analyzed here according to Mann and Thompson’s (1988)
Rhetorical Structure Theory. The study investigates how two speakers create a text interactively, and sheds light onto the nature of text, textual metafunction and texture in a spoken genre. A second focus of this work is to find out whether, and how, the conversational process is different in English and in Spanish. Analyses were performed on both languages, following the same methods. The comparison allows us to examine how the task affects the dialogue organization for two different languages. The main contributions of the book are as follows: (1) corpus-based characterization of a dialogic genre; (2) compilation of a body of analysis tools for generic analysis; (3) application of English-based analyses to Spanish and comparisons between English and Spanish text-building strategies; and (4) a study of the characteristics of each stage in the dialogues, for both languages. This book is the result of a number of years of research, which started with my Ph.D. research at the Universidad Complutense de Madrid, in Spain. Parts of the book have appeared elsewhere: the English analysis of Chapters 4 and 5 in Functions of Language (Taboada & Lavid 2003); the contrastive analysis in Chapter 5 in Discourse across Languages and Cultures, edited by Carol Moder and Aida Martinovic-Zic (Taboada 2004); and parts of Chapter 6 in the Proceedings of the 26th LACUS Forum (Taboada 2000). Those materials are reproduced with permission of John Benjamins and the Linguistic Association of Canada and the United States. Thanks are due to a number of institutions and individuals for their contribution to the completion of the work. First of all, I would like to acknowledge the different sources of funding that allowed me to devote time to carrying out this research. In chronological order, I was first supported by a research assistantship in the Departamento de Filología Inglesa at the Universidad Complutense de Madrid, within the European Union project GIST (LRE 062-09). The research carried out within the GIST project first sparked my interest in discourse, under the supervision of Julia Lavid, my dissertation advisor, who was the team leader and principal investigator for GIST at the Universidad Complutense. A generous grant from “la Caixa”, awarded in 1995 for a period of two years, enabled me to continue my studies at Carnegie Mellon University. At Carnegie Mellon University I was also partly supported as a research assistant within the JANUS project, headed by Alex Waibel. I also express my gratitude to the Department of Linguistics at the University of Alberta. In the final stages of completion of my dissertation, the Department hosted me and offered a supportive academic environment.
While revising the manuscript at Simon Fraser University, I was supported by a grant from the Ministry of Science and Technology of Spain, under project MCYT-FEDER BFF2002-02441 (Ministerio de Ciencia y Tecnología / Fondo Europeo de Desarrollo Regional), co-financed by the Xunta de Galicia (PGIDIT03 PXIC 20403 PN), and whose Principal Investigator is María de los Ángeles Gómez-González. I was also supported by Simon Fraser University, under a small SSHRC grant and a Discovery Parks grant, and by the Natural Sciences and Engineering Research Council of Canada, under Discovery Grant 261104. Many thanks to my advisors: Julia Lavid, who supervised my Ph.D. dissertation, and Lori Levin and Alex Waibel, who were co-supervisors of my M.Sc. research. Thanks also go to mentors and colleagues in different places for their insights, advice and help of all kinds. Finally, thanks to Andreas Jucker and the two anonymous reviewers for John Benjamins. Any errors remaining are my responsibility. On the personal side, I want to thank family, friends and colleagues who have accompanied me during the last few years. Maite Taboada Vancouver, May 2004
TSL[v.20020404] Prn:29/07/2004; 11:38
F: PB12901.tex / p.1 (1)
Chapter 1
Introduction
My interest in the study of conversation was sparked some years ago. I had worked on (written) text analysis for a long time before I was entrusted with the analysis of dialogues similar to the ones found in this study. I surveyed the literature that analyzed conversation, discovering that a good deal of this work focused on its interactional character: how turns are organized; what kind of devices are available to the speaker and hearer in the process of turn construction; and how speakers introduce and maintain topics. In summary, there were many studies that used tools designed specifically for the study of conversation. I realized that it might be enlightening to take advantage of the tools that linguists and text analysts had developed for the study of written text. While acknowledging that conversation is fundamentally interaction, I decided to explore it with some of the resources found in the analysis of planned discourse. The choice of task-oriented dialogue had to do with the availability of the corpus. I was fortunate to have access to a large parallel corpus of English and Spanish task-oriented conversations. It is not every day that one comes across a corpus of significant size which has been collected using exactly the same methods for two different languages. This produced a constant context of situation – the same generic configuration – for the two languages of my interest. The next step corresponded to the choice of tools. The goal was to study connexity and coherence at all levels. Thematic progression came first to mind. Daneš (1974a: 114) considers thematic progression to be one of the representations of connexity in a text, one of the ways in which a text displays coherence. Next, I thought of describing coherence-building from a propositional point of view. Although the definition of coherence is still elusive (Bublitz 1999), my approach is to study it as logical coherence: the connection of propositions in different underlying relations between parts of text. For such purposes, I used Rhetorical Structure Theory (Mann and Thompson 1988). The third component in the analysis is the study of cohesion: how speakers introduce and point to elements in the conversation through reference, substitution, ellipsis and semantic relationships among lexical items.
TSL[v.20020404] Prn:29/07/2004; 11:38
F: PB12901.tex / p.2 (2)
Chapter 1
I considered the dialogues in the corpus to be instances of a particular genre. That is why I chose to include the three types of analysis in a larger framework: the analysis of speech genres. The generic characteristics of the dialogues will be described and explained in relation to the three main types of analysis. As instances of a genre, the dialogues can be described in terms of the sequence of stages the speakers introduce. Those stages are, in turn, described in terms of the sequencing of speech acts that they may contain. The three main types of analysis are unified under the textual metafunction of language (Halliday and Hasan 1976; Halliday 1994). Halliday and Hasan (1976: 29) divide the textual metafunction into structural and non-structural components. The structural component, at the clause level, is the thematic structure of the clause. The non-structural one is cohesion. One of the elements of cohesion is conjunction, which I represented here not through Halliday’s and Martin’s conjunctive relations, but through a different theory of text relations, Rhetorical Structure Theory. Halliday and Hasan, in the same work (1976: 325), refer to the three methods of creating texture: – – –
Cohesion Thematic structure Macrostructure (generic structure)
The study is one of text, textual metafunction and texture in a spoken genre. We shall explore, in the following chapters, how two speakers build a dialogic text interactively. The analyses chosen come together as types of cohesive devices, as shown in Figure 1.1, adapted from Halliday and Hasan (1985: 82). The elements in bold type represent the components considered for this study. Cohesion is divided into non-structural and structural. Within the former, two variables: the type of relation (componential or organic), and the type of device (grammatical or lexical), yield different types of cohesion, as represented in the figure. In terms of componential relations, I have studied all types, both grammatical and lexical. As far as organic relations are concerned, I have studied only the conjunctives, applying Rhetorical Structure Theory (Mann and Thompson 1988). The analysis of rhetorical relations included an examination of the discourse markers used to signal the relations. Finally, structural cohesion, which is realized through parallelism, Theme-Rheme development and the organization of Given and New information, is represented in this study in the analysis of Theme-Rheme development. The tools employed in the analysis stem from traditions in the analysis of English. Some of them have been applied to other languages, sometimes not
TSL[v.20020404] Prn:29/07/2004; 11:38
F: PB12901.tex / p.3 (3)
Introduction
Figure 1.1 Summary of cohesion devices
without controversy about their suitability. I have, consciously, avoided adaptations or reformulations of the original theories. One obvious reason is that the methods needed to be constant, if any rigorous comparison between the results for the two languages was to be carried out. The other main reason was that I decided to explore a possible English bias in those theories. My results show that the analysis of rhetorical relations is well represented by using RST. However, the analysis of phenomena that occur more on the surface of the language – thematic development and cohesion – needed fine-tuning for an accurate examination of Spanish. I have already mentioned a few research questions. In the rest of this introductory chapter, I shall outline those questions and discuss them in more detail. Here are the basic research questions that I set out to answer in this study.
TSL[v.20020404] Prn:29/07/2004; 11:38
F: PB12901.tex / p.4 (4)
Chapter 1
1. How can we characterize the texture of spoken language? 2. Are the tools developed for the analysis of written texts suitable to the study of spoken language? 3. What are the effects of staging and generic structure on texture and cohesion? 4. Are there significant differences in the realization of texture between English and Spanish, given a constant genre? 5. How can we best characterize the sequencing regularities found in the dialogues? The answer to the first question has already been outlined: through the analysis of thematic patterns, rhetorical structure and cohesion, we come to an understanding of the texture-creating devices available to the speakers. The second question will be answered throughout this work. In each chapter, and as I discuss the application of each type of analysis, I point out the difficulties faced when analyzing spoken data. A related question bears on the suitability of these tools for the analysis of Spanish. In like manner, that question finds an answer in each of the chapters that include textual analysis. The realization of the texture-creating devices is also studied with respect to the generic structure of the dialogues (research question number three). At the same time, question four is unravelled when the analyses are performed on both languages in parallel. Finally, the fifth question brings us to a model for the structure of the dialogues, based on the generic stages found in the dialogue, and their linguistic realization. The main body of the work starts out with an examination of the theory of speech genres, as the general framework of my study, in Chapter 2. I then provide a thorough description of the corpus, the data collection and the characteristics of the dialogues. This description is to be found in Chapter 3. Chapter 4 begins the description of texture-making devices, with the discussion of thematic realization and thematic progression in the dialogues, for both languages. Next, Chapter 5 examines the rhetorical structure of the corpus, again contrasting languages and taking into consideration their generic structure. The survey of texture is completed in Chapter 6, which unfolds the cohesive aspects of the corpus. Chapter 7 takes a different route, which leads to a specification of the stages, and how those stages are signaled by different thematic, rhetorical and cohesive devices. The study concludes with a discussion of results and consequences of this work, in Chapter 8. A few notes on form are in order now. I have limited my discussions of previous research to the research close to the areas I explore. Excellent summaries
TSL[v.20020404] Prn:29/07/2004; 11:38
F: PB12901.tex / p.5 (5)
Introduction
on the areas of discourse and conversation analysis, and Systemic Functional Linguistics are provided in, among others, Halliday (1994), Eggins and Slade (1997) and Schiffrin (1994). Chapter 2 delivers a review of the main framework for the study, the description of genre theory. In all other situations, I examine the main aspects of an analysis tool just before I discuss the analysis and the results obtained. That is the case in Chapters 4, 5, 6, and 7. Another caveat on form refers to intonation. The effects of intonation were not considered in the study, not because of disregard for the importance of intonation, but because of the complexity it would have added. In most instances, and when intonation seems relevant, I rely on the intonation markers and the segmentation as coded by the transcribers of the dialogues. I began this introduction by explaining the choice of task-oriented dialogue. I shall finish with a few words on the choice of title. I see the dialogues as instances of an interactive construction of coherence and cohesion. The coherence is reflected on the appropriateness of the genre to the purpose, and also in rhetorical relations, analyzed using Rhetorical Structure Theory. The cohesion is reflected in the texture of the conversations, their ‘hanging together’ as units with interconnected links. The conversations are a joint collaborative effort. Collaboration is defined as a process in which two or more participants coordinate their actions toward achieving a shared goal (Rich and Sidner 1997: 117). The speakers need to collaborate in order to achieve the goal of scheduling a meeting, and that collaboration takes place through talk itself. I perceive that collaboration as being constructed interactively. According to Gallardo Paúls (1998: 48), a conversation will always be guided by an interactive orientation, because each contribution establishes a relationship of mutual dependency with the previous contributions. By showing how the entirety of the contributions by two speakers in a conversation can be analyzed as a text, I intend to illustrate how a spoken text can be built interactively. The text is, then, an emergent product (Clark 1996: 22; Hopper 1988). In this view, language, in the form of a text, emerges as a result of each individual’s contribution.
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.1 (7)
Chapter 2
A framework for the analysis of speech genres
This chapter provides the framework that will lead us through the rest of the study. I consider scheduling dialogues as a genre, and devise a type of analysis based on what a generic analysis of conversation ought to consider. A genrebased analysis first requires an operational definition of what genre is, and how to perform an analysis based on genre. This chapter supplies the definition and delineates the type of analysis to be carried out. The first section provides some historical background, and a comparison of genre with other constructs that define background knowledge – including frames, scripts and schemata – to finish with a working definition of genre. Section 2.2 reviews some of the applications of genre analysis in curriculum design, second language teaching and discourse psychology, among others. I then consider what components we need to include in a generic analysis of conversation in Section 2.3. Section 2.4 explains why scheduling dialogues can be deemed a genre. We start, then, with a tour of the concept of genre.
. Genre defined There is no consensus on the definition of genre, except to refer to the work of Mikhail Bakhtin, especially the essay entitled ‘The Problem of Speech Genres’, written in the 1950s and published in English translation in 1986. A similar concept of register has been developed within Systemic Functional Linguistics following the proposal of Halliday (Halliday 1985) and the Generic Structure Potential structures that Hasan (Hasan 1977; Hasan 1984b) applied to the analysis of nursery tales and service encounters.1 Martin especially (Martin 1984, 1985b, 1992, 2001b; Martin et al. 1987; Eggins and Martin 1997) has worked on the definition front, trying to incorporate genre into the theory of register and delimiting the ground covered by each of the concepts. In relation to this school, but with a different emphasis, we can highlight the efforts of genre
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.2 (8)
Chapter 2
theorists who have applied genre to curriculum design and foreign language teaching. Most of the work reviewed here is within Systemic Functional Linguistics, because it is the framework of this study. There are other approaches to genre, most notably what has been dubbed ‘North American genre theory’ or ‘Rhetorical Genre Studies’ (Miller 1984; Freedman and Medway 1994a, 1994b; Freedman 1999). See also Moessner (2001) for a review of the terms genre, text type and register in different schools, and Hyon (1996) for a comparison of different schools. This section outlines the definitions by Bakhtin and others of genre and register, their applications, and will conclude with a working definition suitable for the purposes of this study. .. Bakhtin’s speech genres There seems to be agreement in pointing to Mikhail Bakhtin for early insights into an approach to language that calls for the concept of genre.2 For Bakhtin, language is realized through individual concrete utterances by participants in the various areas of human activity. These utterances reflect the specific conditions and goals of each area through different aspects – namely content, linguistic style (selection of lexicogrammatical resources) and compositional structure. All three aspects are linked to the whole of the utterance and determined by the nature of the particular sphere of communication where they are produced. “Each separate utterance is individual, of course, but each sphere in which language is used develops its own relatively stable types of these utterances. These we may call speech genres.” (Bakhtin 1986: 60). Speech genres are numerous and diverse, “because the various possibilities of human activity are inexhaustible, and because each sphere of activity contains an entire repertoire of speech genres that differentiate and grow as the particular sphere develops and becomes more complex” (Bakhtin 1986: 60). Given such functional diversity, one might be tempted to think that there are no common features of speech genres. Or, if there are, that those are abstract to a point that renders them devoid of meaning. This apparent need for excessive abstraction is what hindered, in Bakhtin’s opinion, the study of speech genres. Even in the prolific field of literary genres, the efforts have concentrated on their specific literary and artistic features, in terms of what makes them distinct from each other, “and not as specific types of utterances distinct from other types, but sharing with them a common verbal (language) nature” (Bakhtin 1986: 61).
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.3 (9)
A framework for the analysis of speech genres
We should not underestimate the difficulty of determining the general nature of the utterance, and the problems posed by the extreme heterogeneity of speech genres. However, the analysis of speech genres is indeed relevant, because we speak only in speech genres: Even in the most free, the most unconstrained conversation, we cast our speech in definite generic forms, sometimes rigid and trite ones, sometimes more flexible, plastic, and creative ones (everyday communication also has creative genres at its disposal). We are given these speech genres in almost the same way that we are given our native language, which we master fluently long before we begin to study grammar. [. . . ] If speech genres did not exist and we had not mastered them, if we had to originate them during the speech process and construct each utterance at will for the first time, speech communication would be almost impossible. (Bakhtin 1986: 78–79)
The study of speech genres is as relevant to the study of communication as the study of language itself. Genres shape our everyday communication as much as the language forms do. There is, however, one essential difference between language forms and generic forms: language forms are stable and compulsory for the speaker. They are normative, while generic forms are much more flexible and can be manipulated to suit the situation or the speaker’s purposes. “But to use a genre freely and creatively is not the same as to create a genre from the beginning; genres must be fully mastered in order to be manipulated freely” (Bakhtin 1986: 80). The process of selecting and using genres is directly related to their functional character. When we speak, we select each sentence given the whole utterance.3 The chosen genre will determine the type of utterances and the links between them. One of Bakhtin’s major contributions is to broaden the word “genre” to include, independently of the written/spoken distinction, everyday genres as well as literary ones. More precisely, he distinguishes between primary and secondary genres. Secondary, or complex speech genres are literary genres such as novels and dramas, but also genres related to scientific research and commentary. These arise in more developed and organized cultural communication, most of the time in the written mode. Secondary genres can absorb and contain primary or simple genres – for instance, pieces of everyday dialogue or letters found within a novel. Primary genres are directly connected to reality, and to the utterances of others. The interrelations between primary and secondary genres ought to shed light on the nature of the utterance. We come back to this subcategorization in Section 2.1.3, because it has been used to distinguish register from genre, and whole texts from sections of texts.
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.4 (10)
Chapter 2
The next few sections examine a number of theories and schools that have operated around the idea of genre as a product of social context. .. Register analysis, functional genre theory and generic structure potential The three labels in the heading of this section are grouped as one because they are all concerned with creating a functional theory of language. A functional theory of language seeks to establish a relationship between context and text. Context was first considered as a part of a model of language by Firth (1968). Firth was, in turn, influenced by the anthropologist Malinowski – see, for instance Malinowski (1923). Malinowski divided the notion of context into the immediate context of situation where the language is playing a part, and the more global context of culture. The meaning or meanings realized in a text never take place in a vacuum, but in a situation and a culture. The realization of meanings in context through text is not straightforward: [T]his relationship between context and text is theorized as probabilistic, not deterministic: an interactant setting out to achieve a particular cultural goal is most likely to initiate a text of a particular genre, and that text is most likely to unfold in a particular way – but the potential for alternatives is inherent in the dialogic relationship between language and context. (Eggins and Martin 1997: 236)
The idea of a relationship between context and text was first formalized in the concept of register. Halliday (Halliday et al. 1964), and later Ure and Ellis (1977), and Gregory and Carroll (1978) used register to refer to “a variety according to use in the sense that each speaker has a range of varieties and chooses between them at different times” (Halliday et al. 1964: 77). Register is thus different from dialect in that the latter is a variety according to speaker. Each speaker has only one dialect and uses it all the time. On the other hand, each speaker will very likely have more than one register, ready to use according to the situation. The description of register is reminiscent of Bakhtin’s description of speech genres as used in different spheres of communication: When we observe language activity in the various contexts in which it takes place, we find differences in the type of language selected as appropriate to different types of situation. (Halliday et al. 1964: 87)
A register is constituted by the linguistic features which are typically associated with a configuration of situational features, classified in values of the field, mode and tenor of the text’s context of situation.
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.5 (11)
A framework for the analysis of speech genres
Field refers to what is going on; the area of operation of the language activity. It describes the inherent features of the situation and the event taking place, with an emphasis on institutional areas of activity. Under this heading, registers are classified according to the nature of the whole event of which the language activity forms a part. In the type of situation in which the language activity accounts for practically the whole of the relevant activity, such as an essay, a discussion or an academic seminar, the field of discourse is the subject-matter. On this dimension of classification, we can recognize registers such as politics and personal relations, and technical registers like biology and mathematics. There are on the other hand situations in which the language activity rarely plays more than a minor part; here the field of discourse refers to the whole event. (Halliday et al. 1964: 90–91)
The field serves as a descriptor of the social institutions which determine the location and the actions that the participants engage in. Field focuses on the elements which serve as a starting point for the other two categories of tenor and mode. There is, however, a high degree of interaction among the three categories. The elements which constitute the field are not deterministic of tenor and mode. Tenor, formerly style of discourse, refers to the relations among the participants, to the extent that they affect and determine features of the language. While field refers to the inherent characteristics of the situation, tenor refers to the non-inherent features, to those elements that will vary according to the social interactions taking place. In the category of tenor we include degrees of formality, the roles played by the participants and the focus of the activity. Mode of discourse is the function of the text in the event. Halliday and Hasan include here “both the channel taken by the language – spoken or written, extempore or prepared – and its genre or rhetorical mode, as narrative, didactic, persuasive, ‘phatic communion’ and so on” (Halliday and Hasan 1976: 22). Most practitioners of register theory, however, like to place the genre or rhetorical mode one level above, as we shall see in Section 2.1.3. Mode is left to describe the degree of spontaneity between extempore and prepared, together with the amount and type of feedback possible. This yields different configurations, which could be represented as the choice of one of the cells in the matrix in Figure 2.1. Degree of spontaneity can be encoded as the distinction between planned and unplanned text (Ochs 1979), but also as a distinction between formal and informal modes of language. Somewhat simplistically, we could argue that the distinction spoken vs. written specifies whether there is immediate feedback. Although these differences are possibly best represented
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.6 (12)
Chapter 2
Figure 2.1 Choices of mode
as continua (see Eggins 1994: 45), because the boundaries between them are not clear-cut, the matrix in Figure 2.1 provides an illustration of the division of the space given the different factors. Each cell may contain different instantiations of those categories: post-it notes, e-mail messages, and grocery lists all fit in the ‘written, informal’ cell. In summary, register is the set of meanings determined by the situation – what is taking place, who is taking part and what part the language is playing – along with the words and structures used in the realization of those meanings. The three elements that realize context are linked to the linguistic system in the Hallidayan model. Field, tenor and mode have direct realizations through the metafunctions of language: ideational, interpersonal and textual (see Figure 2.5). Thus, the ideational metafunction is realized through field, the interpersonal one through tenor, and the textual one through mode (Halliday et al. 1964; Halliday and Hasan 1976; Halliday 1978, 1989, 1994). Register places emphasis on the context of situation, as defined by the field, tenor and mode variables. It does not account for the relationship of language to the context of culture. Functional genre theory fills that gap by providing an emphasis on social and cultural factors as the generating factors of all action, including linguistic action, through the concept of genre. The widely quoted definition by Martin is that genre is “a staged, goal-oriented, purposeful activity in which speakers engage as members of our culture” (Martin 1984: 25). Kress and Threadgold think of genres as “the socially ratified text-types in a community” (1988: 216). Genre is a recipe for producing a text. It is, however, not simply a schema or frame for action. It involves characteristic ways of “text-making”, and characteristic sets of interpersonal relationships and meanings. (Threadgold 1989: 96). Genre theorists consider that register theory focuses primarily on
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.7 (13)
A framework for the analysis of speech genres
linguistic structure. As a consequence, from the register point of view, the text is viewed as a synoptic product, leaving in the background the probabilistic, dynamic aspects of its performance. Genre, on the other hand, allows for a dual focus, both on the synoptic view of the text as product, and on the dynamic perspective of text as process. Genres are both “products” and “processes” – “systems” and “performances”. Each time a text is produced so as to realize and construct a situation-type it becomes the model for another text and another situation-type. As a model, it functions like a static, finished product or a system according to which new texts can be constructed. Once the constructing begins it becomes again a dynamic process, a “performance” which will inevitably change the model with which it begins. This means that we have to teach the interpersonal and textual characteristic of genres, the probabilistic, dynamic aspects of their performance as well as their schematic structures. (Threadgold 1989: 100)
The study of genre within Systemic Functional Linguistics has concentrated on structural characterizations through genre staging. This approach has been influenced by Mitchell’s study of the language of buying and selling in a Moroccan marketplace (Mitchell 1957). Part of the analysis involved a formula that describes the typical sequence of realization in market auctions and market transactions, that is, the staging of the activity. The following are the structures for each of those situations, where the symbol ^ stands for “followed by”, with some variability observed: (1) Market auction: Auctioneer’s Opening ^ Investigation of Object of Sale ^ Bidding ^ Conclusion (2) Market transaction: Salutation ^ Enquiry as to Object of Sale ^ Investigation of Object of Sale ^ Bargaining ^ Conclusion
Hasan (1977, 1984b) continued this work on genre staging, introducing the notion of Generic Structure Potential (GSP). GSP is the range of possible staging realizations of a given genre. She concentrated on nursery tales and service encounters, outlining the range of options open to the interactants as creators of the text, whether a written nursery tale or an ongoing conversation at a store. Associated with each genre of text is a generalized structural formula, which permits an array of actual structures. A complete text is a realization of a structure taken from such an array.
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.8 (14)
Chapter 2
Structure is one of two characteristics that define texts. In Hasan’s model, text is seen as something different from a random collection of sentences, different too from a super-sentence. Thus, she challenges the text grammar models – Harris (1963), Pike (1964), van Dijk (1972) – in favour of a view of text as an entity in its own right. The text-ness, the characteristics that make a text what it is, are texture and structure. As for texture: Texture is the technical term used to refer to the fact that the lexicogrammatical units representing a text hang together – that there exists linguistic cohesion within the passage. This cohesion is effected by the use of such linguistic devices as those of reference, substitution, ellipsis, conjunction and lexical organization (Halliday and Hasan 1976). (Hasan 1977: 228)
Hasan defines structure as “what allows us to distinguish between complete and incomplete texts on the one hand, and between different generic forms on the other” (Hasan 1977: 229). An incomplete text is one where only a part of an actual structure is perceived as realized in it. The structure can be represented in a structural formula, as we saw above for the auctions and transactions in a Moroccan marketplace. A structural formula is a well-defined configuration of the structural elements in a text. Each of those elements is, in turn, realized in some combination of lexicogrammatical units. The definition of the lexicogrammatical units themselves is a functional one, the functions being determined by the semiotics of the text genre. “A text is a social event whose primary mode of unfolding is linguistic.” (Hasan 1977: 229). A structural formula is not simply a list of the elements of structure, but also the permitted configuration of the elements. Hasan does not require all elements to be ordered; she suggests that there exists a partial ordering. There seem to be no genres where all elements of the structural formula are free to occur in any sequence. Partial ordering implies that the degree of mobility varies from pairs of elements to other pairs of elements. The ordering is usually imposed by the natural logic of the social event represented verbally by the text. I exemplify this with a contextual configuration, using the values represented in Table 2.1, from Hasan (1977). In everyday terms, this contextual configuration represents the situation in which a person (patient-applicant) calls a receptionist at a doctor’s clinic to set up an appointment to see the doctor. The structural formula associated with this contextual configuration will contain the following obligatory elements – that is, a text belonging to this genre will be perceived as incomplete or inappropriate should it be missing one of these:
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.9 (15)
A framework for the analysis of speech genres
Table 2.1 A contextual configuration Variables
Values of the variables
field
professional consultation: medical; application for appointment
tenor
client: patient-applicant, agent for consultant: receptionist; maximum social distance
mode
aural channel: – visual contact: telephone conversation; spoken medium
– – – –
Identification (I), divided into Initiator Identification (II) and Respondent Identification (RI) Application (A) Offer (O) Confirmation (C)
In addition, the formula includes some more elements which are considered optional: – – – – –
Greeting (G) Query (Q) Documentation (D) Summary (S) Finis (F)
The structural formula with only the obligatory elements will then be as follows: (3) RI ^ [A·II] ^ O ^ C
The dot between A and II shows that the two elements are not ordered with respect to each other. It is obvious that an II will follow an RI, and that the offer can only come after the application has been made. The confirmation will only take place once the offer has been accepted. But we can have different orderings for the II and the A. Example (4) shows the ordering A ^ II, whereas (5) would be an example of II ^ A. (4) Receptionist: Dr Scott’s clinic (RI) Patient: I wonder if I could see Dr. Scott today (A) the name is Mary Lee (II) (5) Receptionist: Dr Scott’s clinic (RI) Patient: This is Mrs Lee speaking (II) I wonder if I could see Dr. Scott today (A)
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.10 (16)
Chapter 2
The complete formula, including optional elements, is represented in (6). A parenthesis surrounds optional elements, and a subscript follows an element that can be iterated. (6) (G) ^ RI ^ (Q) ^ (G) ^ [A· II] ^ O ^ C ^ Dn ^ (S) ^ Fn
The conversation represented in (7) below is one possible instantiation of the formula, from Hasan (1977). “R” stands for “Receptionist” and “P” stands for “Patient”. (7)
R: good morning (G) Dr Scott’s clinic (RI) may I help you (Q) P: oh hello good morning (G) this is Mrs Lee speaking (II) I wonder if I could see Dr Scott today (A) R: um well let me see I’m afraid Mrs Lee I don’t have much choice of time today would 6:15 evening suit you (O) P: yes, yes, that’ll be fine (C) R: may I have your address and phone number please (D) P: 24 May Avenue, North Clyde and the number is 527 2755 R: thank you (D) so that’s Mrs Lee for Dr Scott at 6:15 this evening (S) P: mm yes thanks (F) R: thank you (F)
There are some evident connections between the Systemic Functional work in genre and Propp’s Morphology of the Folktale (Propp 1928). Propp treated fairy tales as instances of a genre, which could be defined in terms of stages, characters and props. The stages succeed each other in some predetermined order. The work of Firth, Mitchell and Hasan has inspired formalizations of staging in different genres. Sinclair and Coulthard (1975) developed a system to analyze classroom discourse in which generic structure is built incrementally from the speech act unit to larger units, such as moves, exchanges and transactions. One of the most important descriptions of staging is Ventola’s characterization of service encounters (Ventola 1987). She uses a flowchart notation to represent decisions made by the interactants in the construction of the social process. The flowchart represents the canonical sequence of elements in this genre, showing at the same time how variation in sequencing could be achieved. In addition, it allows for the inclusion of non-verbal activities as realizations of elements. Ventola herself, and many others – working within the Systemic Functional school or not – have worked with the notions of genre and register, devising a system where they are both integrated, but representing dif-
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.11 (17)
A framework for the analysis of speech genres
ferent aspects of the language and the situation. The next section discusses the relationship between genre and register. .. Genre and register The conceptualizations of the relationship between register and genre range from Eggins and Martin’s lumping of the two (at least in terms of labelling) into “Register and Genre Theory” (Eggins and Martin 1997), Hasan’s identification of register with genre (Hasan 1977), to the total separation, with genre one level above register (Ventola 1987; Martin 1992). Intermediate approaches apply either type of analysis according to the text type or length (Leckie-Tarry 1995). Martin (2001b) admits that “the relation of purpose, or what a speaker is trying to accomplish, to register has long been an uneasy one.” (Martin 2001b: 155). Eggins and Martin provide an introductory account of genre and register, under the heading “Register and Genre Theory” (R>). Under that label, they “seek to theorise how discourses, or texts, are like and unlike each other, and why” (Eggins and Martin 1997: 230). They establish two common themes in R> approaches: Firstly, they focus on the detailed analysis of variation in linguistic features of discourse: that is, there is explicit, ideally quantifiable, specification of lexical, grammatical and semantic patterns in text. Secondly, R> approaches seek to explain linguistic variation by reference to variation in context: that is, explicit links are made between features of the discourse and critical variables of the social and cultural context in which the discourse is enacted. Register and genre are the technical concepts employed to explain the meaning and function of variation between texts. (Eggins and Martin 1997: 234)
Register is “a theoretical explanation of the common-sense observation that we use language differently in different situations” (Eggins and Martin 1997: 234). It differs from dialect or sociolect, because it is a “variety according to use”, not a “variety according to user” (Halliday et al. 1964: 77). Register defines a probabilistic relationship between context and language. The relationship can be a weak one, where a high number of contextual factors have an impact on the text (Hymes 1974). Or it can be a strong one, as defended by Halliday (1978), in which case texts are realizations of a limited number of contextual dimensions, namely field, tenor and mode (see Section 2.1.2 above for a complete description of register). In addition to register variation, texts exhibit generic variation. This type of variation is determined by the purpose the text is achieving in the culture. We recognize a text’s genre by its purpose, but also through the way
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.12 (18)
Chapter 2
it unfolds dynamically: “the major linguistic reflex of differences in purpose is the staging structure by which a text unfolds” (Eggins and Martin 1997: 236). The relationship of register to genre is one of layering: “two layers of context are needed – with a new level of genre posited above and beyond the field, mode and tenor register variables” (Eggins and Martin 1997: 243). This is outlined in Figure 2.2, where the layer of genre comprises and extends the level of register with its three components. Register represents the context of situation, whereas genre refers to the context of culture. According to this description, register and genre stand in a clearly defined relationship to each other. Text analysis in this approach proceeds from the level of culture down to the level of situation, specifying the variables of field, tenor and mode. The purely linguistic analysis dynamically relates lexical, grammatical and semantic choices to the contextual variables. A different view is offered by Hasan, who believes that “[i]n the SF [Systemic Functional] model the concept of register is a ready-made link between context and generic structure, since for most material purposes register and genre are synonymous” (Hasan 1977: 230). For her, as mentioned above (Section 2.1.2), texts have texture (=cohesion) and structure, the latter determined by the text’s genre. The realization of the two categories is very different, as represented in Figure 2.3. It is surprising, then, that although their realizations are utterly different, she equates the two concepts “for most material purposes”. We have seen, in the previous section, how she used the concept of structural formula, and how
Figure 2.2 Relationship of genre and register according to Eggins and Martin (1997)
Figure 2.3 Hasan’s realization of genre and register
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.13 (19)
A framework for the analysis of speech genres
she defined Generic Structure Potential as a valid formalism of representation of a text’s genre. Ventola’s work is directly influenced by both Martin’s and Hasan’s. In her characterization of service encounters, she provides a definition of genre that could be also represented by Figure 2.2. We will be referring to her work in later sections. Gregory considers register, and knowledge of a register, as the linguistic part of the knowledge of a generic situation: I suggest that we communicatively cope with these ‘certain recurring situations’ because we have internalized a generalized knowledge of them. Any INSTANTIAL SITUATION we respond to in terms of this knowledge which is a knowledge of what I term GENERIC SITUATION, the micro-environment of our knowing and doing. This knowledge includes a knowledge of the meaning potential of the media of communication open to us. In the case of language as a kind of doing, the meaning potential associated with a generic situation is called REGISTER. (Gregory 1988: 303)
Leckie-Tarry associated the term “register” with primary or simple genres, and especially with sections of texts. More complex instances, full texts and what Bakhtin calls secondary genres are more commonly associated with the term “genre” (Leckie-Tarry 1995: 12). She also pointed to a level of genre that represents those events which have been culturally recognized. Registers then, “are free to mediate in any communicative event, socially identified or informal, complete or incomplete” (Leckie-Tarry 1995: 15). She sees register as explaining the social semiotics of texts – how they mean – together with their lexicogrammatical characteristics – what texts mean. In her analysis, genre is most appropriate for a socially-oriented analysis. Register is also used as a general-purpose label to describe variation according to situation and participants. Such is the use by Biber and colleagues (Biber et al. 1998; Biber and Conrad 2001). Register is related to text type – another general-purpose label – and different from dialect, which is associated with a particular speaker. “Registers are defined according to their situations of use (considering their purpose, topic, setting, interactiveness, mode, etc.)” (Biber et al. 1998: 135). Notice that the inclusion of purpose in the definition means that there is no distinction between what systemic functional linguists define as genre and register. Biber and others, however, distinguish genre from text type, the former being based on purpose and the latter on linguistic features (Biber 1988; Paltridge 1996).
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.14 (20)
Chapter 2
Before providing a final definition of genre, we will briefly consider the relationship of genres to other less socially-oriented theories, such as theories of background knowledge in the next section and prototype theory in Section 2.1.5. .. Genres, frames, scripts and schemata Frames, scripts and schemata are all different ways of representing and explaining background knowledge. Whether geared towards the explanation of human language processing or not, they all have been used in Artificial Intelligence to represent the way we store and use information about situations we have encountered or are likely to encounter in everyday life. At the risk of gross generalization, we can say that all three concepts hinge on the apparent capacity of the human brain to represent and reuse knowledge of the world. Minsky’s frames (Minsky 1975, 1977) are data structures stored in memory, which contain stereotyped situations. They represent any kind of knowledge, and since we possess knowledge of a language, there are also frames for linguistic facts. The structure of a frame contains slots which are filled with expressions called fillers. Thus, a HOUSE frame will have slots with labels such as “living room”, “bedroom”, “kitchen”, etc. A particular house will be an instance of the HOUSE frame, represented by filling the slots with the specific features of the house under consideration. Discourse understanding can be depicted as the processing of the incoming information to fit the relevant frames that we already possess: When one encounters a new situation (or makes a substantial change in one’s view of a problem) one selects from memory a structure called a frame. This is a remembered framework to be adapted to fit reality by changing details as necessary. (Minsky 1977: 355)
Scripts include not only objects, but also typical actions as slot labels. A RESTAURANT script will have “table”, “food”, “menu”, “waiter/waitress”, but also “order”, “eat”, “pay”, “tip” as some of its empty slots, with some optionality available, depending on the type of restaurant (e.g., no tipping in a fast food restaurant). Further detail on the concept of a script is to be found in Schank (1972, 1973), Schank and Abelson (1977) and Riesbeck and Schank (1978). Schemata are knowledge structures as well, for some even deterministic ways of conceptualizing the world. Racial and gender prejudice would be, in this view, the manifestation of a way of thinking about particular individuals on the basis of existing schema for members of the race or sex (Brown and Yule
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.15 (21)
A framework for the analysis of speech genres
1983: 247). A weaker, less deterministic view, is that a schema is an organized instance of background knowledge that helps us predict aspects of a discourse and its interpretation. Tannen, as well as Anderson and colleagues, describe the influence of different cultural background and different interests in the interpretation of texts (Tannen 1980; Anderson et al. 1977). The novelty of the schema concept is its active characteristic, found in the work of Bartlett (1932), where discourse is a constructive process. In the interpretation of discourse we use information from the present piece of text, but also knowledge from past experience, the schema. The schema is not a static representation ready to be filled with the current information, the way frames are; it is something which remains active and developing. Other descriptions of schemata de-emphasize the active aspect and define schemata as stereotypes of concepts (Rumelhart and Ortony 1977), approaching a prototype-theory description, as we shall see in the next section. In summary, we can conclude that all of these theories of background knowledge postulate some form of internal organization that helps us interpret and predict events in general, and linguistic events or discourses in particular. A more extensive, linguistic-oriented description, which also includes scenarios (Sanford and Garrod 1981) and mental models (Johnson-Laird 1980), can be found in Brown and Yule (1983). But how does all this relate to the concept of genre? I have described genres as socially-oriented and socially-enforced linguistic activities. Schemata (scripts, frames, scenarios; we can assign any of those names to the general concept) are representations of both general and linguistic activities. Whereas schemata include props and actors in a given situation, the description of a genre will include (see Section 2.1.6) information on how the particular piece of discourse is supposed to develop and what the appropriate language is, in general terms. Schema, genre and language interact with each other to determine the final outcome of the linguistic situation, and to help determine new situations. Each time a schema and a genre are used, they are modified to adapt to new variations. This view answers Kintsch’s criticism (Kintsch 1989) that script-like structures are too inflexible to adapt to the ever-changing context of the environment. For example, Scholtens’ (1991) model of planning in conversation involves knowledge of plan structures, knowledge of scripts and personal experience. The outcome of a conversation will be based on the co-operation of the participants when adjusting those three variables in their own planning to the status of the variables in the interactants’ current state, as inferred from the development of the conversation. Thus, the previously stored plans and scripts are modified according to the interaction as the conversation progresses.
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.16 (22)
Chapter 2
Figure 2.4 Relationship of schemata to genres and process of creation of both
Swales (1990: 84) proposes a model in which prior knowledge – both in terms of experience of the world and of previous texts – determines the shape of genres to come. Halliday mentions the intertextual context of any text, its relations with other texts. In the cycle of text and context the contexts of situation and culture both enable and restrict text interpretation and creation (Halliday and Hasan 1985: 49). My proposal is one where each new situation and each new instance of a genre that is created is incorporated into the model, as can be seen in Figure 2.4. Experience, both general and linguistic, contributes to the creation of schemata. Those include stereotyping, recipes for action, and knowledge of how the world works. But also recipes for linguistic action, i.e., genres. One does not need to participate directly in the situation in order to derive schemata and genres from it. An observation of the interaction (something that children do quite often before they participate) is enough to allow for the creation or modification of schemata and genres. This derives from a Bakhtinian dynamic conceptualization of genre, which leads to an understanding of the essentially intertextual, processual and always only probabilistic nature of genericity. This, of course, is what makes genre a category that is potentially both reality-maintaining and reality-changing: both conservative in maintaining the status quo, and subversive in always presenting the possibility of challenging it. (Kress and Threadgold 1988: 219)
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.17 (23)
A framework for the analysis of speech genres
Dorval (1990), a collection of papers on conversational organization, is an example of the agreement among researchers that coherence in talk is developmental. Hobbs describes how coherence (which can be considered a characteristic of genres) is probably influenced both by cognitive development and social pressure. If topical coherence does in fact increase with age, one question is whether this is a result of cognitive development or of imposed conventions. Does it happen because older children are capable of larger stretches of coherent thought, or does it happen because others come to demand more highly structured talk from them. It is likely that the answer is “Both”. Conventions are made available, or are imposed, as our cognitive development enables us to abide by them. (Hobbs 1990b: 21)
Figure 2.4 displays the process of creating schemata from previous world experience, direct or indirect. Genres are seen as a subset, more linguistic, of general schemata, which are also fed by the previous discourses in which we have taken part or which we have observed. Our current exposure to different media subjects us to countless discourses in written or spoken form. Although we are not able to take part in movies or television shows, the interactions that we observe in those provide us with generic expectations of a range of discourses. .. Genres and prototype theory Swales defines genre as a classificatory category, following Linnaeus. Genres are ideal types of language production, of which each text is an instance: “it is apparently common in this classificatory work to consider genres as ‘ideal types’ rather than as actual entities. Actual texts will deviate from the ideal in various kinds of ways” (Swales 1990: 34). Before we consider the kind of problems posed by an ‘ideal type of text’ view, I will expand some more on the concept of prototype, and how genres can be viewed as categories stored in the mind through prototypes. Rosch first provided a general perspective on problems of categorization already studied by other researchers.4 Her definition of a prototype is that cognitive reference points or prototypes are subcategories or category members that have a special cognitive status – that of being a best example. Prototypes are ways of storing and structuring categories: [. . . ] categories become definitively structured because, even when correlational structure in the world is only partial or when attributes are continuous, thus, producing categories which might tend to blend with other categories at
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.18 (24)
Chapter 2
the same level of abstraction, categories are maintained as discrete by being coded in cognition in terms of prototypes of the most characteristic members of the category. (Rosch 1977: 213)
Although Rosch’s theory of prototypes refers only to real-world objects, we could stretch it to include our perceptions of situations, including discourse situations. A prototypical description of a genre contains all of the most common features to be found in the linguistic phenomena realized in a given situation. The instances of that genre share then a family resemblance structure (Wittgenstein 1953; Rosch and Mervis 1975), the same way members of a family resemble each other, with both typical and atypical members. The link between genres and prototypes has been established before. Paltridge (1995c) proposes a framework to explain how people assign a particular text to a genre, even when the genre’s prototypical features are absent, based on Austin’s (1962) felicity conditions. He also discusses general aspects of the notion of genre as prototype in a different paper (Paltridge 1995b). De Geest and van Gorp (1999) apply the notion of prototype to the analysis of literary genres. This view of genres as prototypical categorizations poses a number of problems: what are the conditions for a borderline? When have we left one genre and started a new one? And, since genres are social creations, how do we abstract from instances to prototypes? Ventola tries to answer some of those questions. She ponders the idea of having the presence or absence of obligatory Generic Structure Potential elements as a criterion for classifying texts into a particular genre, following Hasan’s approach (see Section 2.1.2). According to Hasan, “[i]f a text does not include the obligatory elements, it is either incomplete or its generic membership cannot be determined, i.e., it is a non-text.” (Ventola 1987: 55). However, slight differences in the contextual configuration can lead to different realizations of obligatory generic elements. Thus, in a travel agency scenario, where “booking” would easily be considered as an obligatory element, we might be able to find a situation where the client enquires, but decides to take some more time before booking. Such a conversation would be perfectly functional. In Hasan’s view, then, there would be two different genres: one of travel agency inquiry, and another one of travel agency inquiry plus booking. The consequent explosion in the number of genres would render the concept unmanageable and useless. Ventola’s answer to the problem is the characterization of genres through flowcharts, where a high number of obligatory and optional elements are included, in order to capture the agnateness of different texts under the same genre.
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.19 (25)
A framework for the analysis of speech genres
The most intriguing question is the creation of prototypes from instances. Two different models have been postulated to explain prototype formation: the central tendency and the attribute-frequency models. In the central tendency model, a prototype represents the average or mean of a set of exemplars. The prototype is an abstraction stored in memory that represents the central tendency of the category (Posner and Keele 1968). The attribute-frequency model suggests that a prototype represents the mode, or most frequently experienced combination of attributes. The prototype itself is often unique because it does not represent any one instance, but a combination of attributes or features. The features, however, have been previously experienced. When presented with an instance, the observer stores both the features and the relationships between them in that situation. Because we will encounter the same features more often than the same type of relationships between features, the relationships are less well stored in memory (Solso 1988: 73). In genre terms, we could illustrate this with the change of mailing address discussed in the next section. I know I have to provide my new mailing address (feature), but I might be unsure about whether to provide my name first or not (relationship of order between features). As represented in Figure 2.4 above, the exposure to genres comes from both experienced and observed instances of language activity. .. A working definition of genre The definition of genre that I propose here is one where genre is primarily a structurally-determining characteristic of texts. A given text is perceived as belonging to a genre because of its structural characteristics, that is, its staging. For that recognition to happen, there must be established consensus that certain texts develop in a certain series of stages. The staging appropriate to a text is determined by the function of the text in a given situation. The broad term “function” encompasses two different aspects: the communicative purpose of the text and its social function.5 This is not a novel definition, nor a groundbreaking reformulation of the concept, but an operational definition, which will prove useful in approaching the dialogues in the corpus as instances of a genre. Martin and colleagues have repeatedly defined genre as a reflection of the goal-oriented aspect of the text (Martin 1984; Eggins and Martin 1997; Martin et al. 1987). Speakers have purposes of the most various kinds when engaging in verbal activities, spoken or written. For a maximally efficient satisfaction of those goals, texts tend to develop in a certain way: “[g]enre theory suggests that texts which are doing different jobs in the culture will unfold in different ways,
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.20 (26)
Chapter 2
working through different stages or steps” (Eggins and Martin 1997: 236). If I want to, for instance, change the mailing address attached to my credit card, I need to call the credit card company or bank. It is obvious that I first need to identify myself to the operator before he or she can proceed to change the information attached to my file. Thus, such an interaction will have at least three phases: identification of the caller (me), request (change my mailing address) and details of the request (new mailing address). The two first phases are interchangeable, but I always need to provide my identification (my name or my credit card number, or both) before I provide my new mailing address, so that the operator can retrieve my file. The two following invented examples provide alternative orderings where the request always follows an identification. (I=Identification; R=Request; D=Details). (8) Maite: Hi, my name is Maite Taboada, T-A-B-O-A-D-A. (I) I’d like to change my mailing address with you. (R) Operator: Sure, let me pull up your file. [pause, maybe further Identification questions] What’s the new address? Maite: It’s ... [here, a full address] (D) (9) Maite: Hi, I’d like to change my mailing address with you. (R) Operator: Sure, what’s the last name? Maite: Taboada, T-A-B-O-A-D-A (I) Operator: Okay. [pause, maybe further Identification questions] What’s the new address? Maite: It’s ... [here, a full address] (D)
On the other hand, such a conversation with a different ordering would be awkward. Example (10) would be infelicitous, simply because the operator could not possibly remember all the information contained in my first turn. Clark and Schaefer (1987) observe similar effects in directory assistance calls. In these calls, operators prompt for information in order of delicacy (e.g., city or town, address, person’s name). This allows for a more efficient search, without having to keep an excessive amount of information in short-term memory. (10) Maite: Hi, I’d like to change my mailing address with you. (R). The new address is 1234 Burrard St., apartment 101, the city is Vancouver, the postal code is V1B 2C3 (D), and my name is Maite Taboada, T-A-B-OA-D-A (I) Operator: ???
In addition to the satisfaction of purely communicative purposes, speakers need to satisfy social constraints. Indeed, some types of verbal activities are
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.21 (27)
A framework for the analysis of speech genres
intrinsically social. For example: conversations about the weather with neighbours, small talk between a customer and a store clerk while paying for merchandise, etc. Social and communicative purposes are woven together in the determination of the nature and order of the stages in any given text. When I call my travel agent, with whom I have worked in the past, I not only state the purpose of my call, but also greet her, ask her about her recent vacation, and, just before I hang up the phone, express my wish to see her again soon. The last three elements are there only for the purpose of maintaining a relationship between us enforced by our living in the same community. Murray (1991) divides discourse in two main types: conversation for action and conversation for social maintenance. Most conversations will have elements of both, but at any given point one of them will predominate. We will see, in the following chapters, how the conversations in this corpus can be divided in stages where one or the other type of communication is most prominent (Openings and Closings versus Task-Performance stages, in Chapter 7). The structures that constitute a genre seem to be learned only through socialization: “the stages we go through in language to achieve our goals are social in origin” (Rothery 1985, cited in Martin (1985a: 73)). This is the reason why we often feel anxious the first time we encounter a new genre: a doctor’s appointment, checking a book out of a new library, presenting a paper at a conference. Although we might have a vague idea of what the situation involves, we are unaware of exactly which stages we need to complete, and in which order. Command of a genre is higher among those who operate in it most often (see also Swales (1990)). Many people who have an excellent command of a language often feel quite helpless in certain spheres of communication precisely because they do not have a practical command of the generic forms used in the given spheres. Frequently a person who has an excellent command of speech in some areas of cultural communication, who is able to read a scholarly paper or engage in a scholarly discussion, who speaks very well on social questions, is silent or very awkward in social conversation. (Bakhtin 1986: 80)
Genre thus determines the elements – stages – of a text according to communicative and social purposes. Its influence on the language used is only indirect, mediated through register. Register is another contextual dimension, defined by the field, tenor and mode of the situation. The variation in register will always be a level below, and determined by, the variation in genre. We can easily imagine more or less well-defined genres, such as advisor-student consultations, or even service encounters, where the generic structure will be similar,
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.22 (28)
Chapter 2
Figure 2.5 Relationship of genre, register and language
but where variations in register – tenor, for instance – will make the language itself quite different. The relationship between the genre (organization) and the register (tenor, field and mode) in a text can be characterized in different ways. Figure 2.5 represents my own conceptualization of the relationship of genre and register to linguistic realization. Genre is a separate layer of context above everything else, but with elements that percolate into the language through the gap that separates them. The language is composed of a contextual level and a purely linguistic level. The contextual level of register is subdivided in values of field, tenor and mode. These values are realized linguistically in the three metafunctions of the language. Martin and others in the systemic functional tradition have represented this relationship of metafunctions, register and genre as concentric circles (Martin 1992; Eggins and Martin 1997). The representation proposed here is not fundamentally different. It only advances a looser connection between genre and register than the one between register and metafunctions. Register configurations are directly reflected in the language metafunctions. Generic configurations (especially genre staging), on the other hand, do not have such a close connection to lexicogrammatical patterns.
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.23 (29)
A framework for the analysis of speech genres
. Applications of generic analysis Genre theory, in its various forms, has become an area of stimulating theoretical research but, above all, of interesting applications. It has been primarily applied in the development of literacy programs in Australia. Kress (1982) suggested that learning about textual form – genre – was one major aspect of learning to write. Genre-based teaching (or learning) focuses on teaching as the conscious learning of structures, of socially-based language structures, but especially of curriculum genres. Classroom genres can be classified into written genres and curriculum genres, the latter being the genres produced orally together by the teacher and students (Samraj 1989). Martin, Christie and Rothery provide a summary of the origins and development of this research agenda (Martin et al. 1987), and Macken-Horarik (2002) a more recent illustration. Other sources of the theory and how it is applied in schools are: Christie (1985, 1994), Painter and Martin (1986), Rothery (1985), Cope and Kalantzis (1993), and Martin (1993, 2001b). Reid (1987) contains a collection of papers by major defenders of the genre-based approach to teaching, together with critiques made by other theorists. The debate over the concept of genre, and its integration into a more general theory of social semiotics, has a follow-up in Kress and Threadgold (1988) and Threadgold (1988). Gregory (1988) and Samraj (1989) also provide insights into applications and illustrations. The so-called North American genre school, or New Rhetoric, developed a slightly different concept of genre, also based on the ideas of Bakhtin (Miller 1984; Freedman and Medway 1994b). Although space precludes a detailed description of the theory itself, its applications are too in the area of language learning (Freedman and Medway 1994b). For a comparison of the ‘Sydney approach’ (Systemic Functional Linguistics), the New Rhetoric, and approaches in applied linguistics, see Hyon (1996), Paltridge (1997), Coe and Freedman (1998) and the chapters in Johns (2002), especially the introduction. Second language teaching has also benefited from the idea of teaching language through language situations. Mohan (1986), for instance, shows the application of generic structures represented in flowcharts to the teaching of English as a Second Language. Koike and Biron (1996) use genre for the development of oral proficiency in a Spanish conversation class. Dudley-Evans (1994), following Swales’ lead (Swales 1990), examines academic writing from a generic perspective, with a view towards applying genre analysis in writing classes and workshops geared towards those in the academic community. Ventola (1989) provides a summary of approaches to genre and issues related to its application in second language teaching. Hyon (1995) proposed a genre-based
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.24 (30)
Chapter 2
approach to teaching English as a second language. Paltridge has devoted a number of books and articles to the issue of genre and language education (Paltridge 1995a; Paltridge 2000; Paltridge 2001). A recent collection of essays summarizes the different perspectives of genre-based education (Johns 2002). One interesting result in genre-based second language teaching is that its effects are long-lasting. Hyon (2001) interviewed students in a genre-based course in English for Academic Purposes one year after the course. Students reported that they not only remembered some of the genre features discussed, but also that they applied the knowledge in their second language learning and writing. Zwaan (1994), from a psychological point of view, discusses how genre expectations affect text comprehension. His results prove that readers allocate their processing resources differently according to their expectations about the genre of a text. Students who were given texts either with a literary story or with a news story as purported genre activated the appropriate reading strategies for each discourse genre. The students reading under a literary perspective had longer reading times, better memory for surface information, and a poorer memory for situational information than those reading under a news perspective. An issue of World Englishes is entirely devoted to the topic of genre analysis (World Englishes 16 (3), 1997). In the introduction to the issue, Bhatia remarks on the influence of genre analysis in the teaching and learning of languages to learners in specialist disciplines like engineering, science, law or business (Bhatia 1997). Genre offers a dynamic explanation of the way expert users of language make use of generic conventions to achieve the goals associated with the discipline. Assuming that learners of language will be committed to the same goals, their language learning ought to involve an explanation of the generic mechanisms available to them. English for Specific Purposes and English for Academic Purposes have parceled language (English in most cases) into genres, focusing on those that are useful to students of the language. Business genres have been studied extensively (Bargiela-Chiappini and Nickerson 2002; Dudley-Evans and Henderson 1990). A collection of essays (Bargiela-Chiappini and Nickerson 1999) addresses aspects of business communication such as e-mail, letters, faxes, and billboard advertisements. Other studies examine business meetings (Bargiela-Chiappini and Harris 1997), letters of application (Upton and Connor 2001), letters of negotiation (Pinto dos Santos 2002), corporate home pages (Luzón Marco 2002), and other spoken genres, such as everyday conversation in the workplace (Brown and Lewis 2002). Christie and Martin (1997) present a collection of articles on genre in
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.25 (31)
A framework for the analysis of speech genres
schools, but also in the workplace, emphasizing the transition from one to the other. Bhatia (1993) presents applications of genre in business and legal settings, as well as in research articles. Other areas studied include the language of law reports (Badger 2003), the language of academic bureaucracy (Swales et al. 2000), and also on-line genres, or cybergenres (Shepherd and Watters 1998). Trosborg (2000) has edited a collection of essays that include business and academic genres, from genetic engineering and medicine to journalism and linguistics. Similarly, academic genres have been the subject of numerous studies, one of the most influential Swales’ (1990). Some scholars take genre as a starting point for contrastive analyses that reflect different discursive practices in different communities. Giannoni (2002) studies acknowledgments in research articles in English and Italian; Martín Martín (2002), abstracts of research papers in English and Spanish; Samraj (2002) studies introductions to research articles in different, but related, disciplines; and Paltridge (1993, 1997), the general characteristics of academic articles. Research on new media has made use of the concept of genre as social action, in the New Rhetoric tradition (Miller 1984) to understand the social interactions in on-line communities (Bregman and Haythornthwaite 2003). Traditional media have also been approached from a generic point of view: Meinhof and Smith’s (2000) collection of articles is organized around media genres, especially television (advertising, game shows, music shows, current affairs, political satire), but also print media. Fairclough (1992) studies the relationship between power and change, as part of the Critical Discourse Analysis school (Fairclough 1995; van Dijk 1998; van Dijk 2001; Weiss and Wodak 2003; Wodak and Meyer 2002). Computational linguistics has found genre analysis useful for a number of applications. Kessler and colleagues (Kessler et al. 1997) mention a few subfields that would benefit from the knowledge of a text’s genre. Parsing would increase in accuracy by taking genre into account. When parsing a transitive verb, the parser usually expects a direct object. But certain text types, such as recipes, often use transitive verbs with omitted direct objects. Part-of-speech tagging and word-sense disambiguation would be much easier if we knew the kind of text we are processing, because polysemous words show a higher frequency of a particular meaning in certain text types. The word tree will typically have the less common meaning of “structure display” in scientific articles in fields such as linguistics or computer science. In information retrieval, genre classification will help in the sorting of search results into categories. If I am searching for information on, for instance, the question of separatism in Québec, it will be
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.26 (32)
Chapter 2
useful to have that information sorted out into categories such as newspaper articles, newspaper editorials, political science research papers, opinion books, history books, etc. Finally, it is important to mention the wide-ranging work of Biber and colleagues in the study of genres (which they have called both registers and text types), emphasizing variation across spoken and written genres (Biber 1988), and across languages (Biber 1995), as well as the historical development of genres (Biber and Finegan 1989). Genre is also at the core of the grammar of the English language compiled by Biber, Johansson, Leech, Conrad and Finegan (1999).
. Components in a generic analysis of conversation The discussion so far has concentrated on the concept of genre, and its domain of action. Now we consider the application of genre theory to the analysis of texts – spoken texts in particular. Before we embark on a generic study, we first need to consider the kind of analysis and the tools that we will use. This section discusses two different characterizations of the components of a generic analysis, to conclude with a description of the components that will be used for the study. Eggins and Slade, in their analysis of casual conversation, propose six steps in a generic structure analysis (Eggins and Slade 1997: 231–235): 1. Recognizing a chunk. A “chunk” is a part, a segment or series of segments of a conversation that has a global or macro-structure. This is opposed to “chat” segments, a series of segments where the structure is managed locally, that is, turn by turn. The chat segments are amenable to a microanalysis, describing the move by move unfolding of talk. On the other hand, the chunk segments can be analyzed in terms of their predictable macro or global structure. Factors that might indicate the presence of a chunk are: a. a participant dominating the floor for an extended period of time, b. the presence of predictable stages in the stretch under consideration for chunk status. 2. Defining the social purpose of the chunk and labelling the genre. This step not only includes classifying the overall function of the genre and its name, but it also involves identifying the way the text type constructs social reality.
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.27 (33)
A framework for the analysis of speech genres
3.
4.
5.
6.
In this step we would define, then, the social practices the text refers to, and the attitudes and values formed by and reflected in it. Identifying and differentiating stages within a genre. Stages are constitutive elements of genres. They can be identified through the use of functional labels. Specifying obligatory and optional stages. The obligatory elements are defining of the genre and they are key elements in recognizing a genre. Optional elements are not defining features and can occur across genres. Devising a structural formula. This includes writing down the stages in a linear sequence, together with some notation that would show their ordering and whether they are obligatory or optional. In their notation – widely used after Hasan’s (1977) analysis of nursery tales and service encounters – the symbol ^ denotes order of the stages with respect to each other, and stages enclosed in brackets are optional (see Section 2.1.2). For example, a structural formula for narratives would read: (Abstract) ^ Orientation ^ Complication ^ Evaluation ^ Resolution ^ (Coda) Analyzing the semantic and lexicogrammatical features for each stage of a genre. Lexicogrammatical choices will differ across genres, but also across the different functional stages within a given genre.
In summary, Eggins and Slade’s system will have two basic steps: (1) finding a structural formula that will represent most instances of the genre and (2) analyzing the linguistic characteristic of each stage. I performed the first step – the first 5 steps described above – as a necessary one in the definition of a genre. As for the second step, number 6 above, a different route was taken, which did not explore features at the level of lexicogrammar, but at the textual level. I follow Leckie-Tarry (1995) in proposing that we should consider two approaches in decoding text: – –
Structural: Theme-Rheme, thematic progression and information structure Non-structural: elements of cohesion (reference, substitution, ellipsis, conjunction and lexical cohesion)
As for the first, I will focus on the thematic progression of texts in Chapter 4 and on information structure (rhetorical structure) in Chapter 5. Nonstructural elements will be studied in detail in Chapter 6. I will finish the present chapter with a section on how the corpus we are studying can be considered a genre.
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.28 (34)
Chapter 2
. Scheduling dialogues as genre The conversations described in Chapter 3 are considered instances of a genre, which we might denominate scheduling genre. They are instances of talk produced for a very specific purpose, that of setting up an appointment. If I were to play a recording, or show a transcript of one of these conversations to anyone, they would easily recognize what is happening in the conversation, and that the two people involved are arranging to meet. A taxonomy of speech genres would contain the scheduling genre under a more general type of task-oriented conversations. In these, the participants come together to complete a task. That is, they enter into a process of collaboration: “a process in which two or more participants coordinate their actions toward achieving shared goals” (Rich and Sidner 1997: 117). The defining characteristics of task-oriented dialogues are not collaboration and cooperation in and of themselves: Garrido Medina (1997: 224) counts cooperative conduct as a feature of any conversation. The most defining characteristic is the fact that the goal is a practical one, which speakers are committed to achieve. Other conversations also have goals, but those might be of a less practical nature, such as establishing rapport or maintaining social links. As a result of their practical purpose, the conversations are staged in particular ways: one speaker proposes a meeting, perhaps also a time; the other speaker replies either with a different time or with their availability for the time proposed. The conversation continues until a day and a time have been set. As a result of their social function, the conversations usually have Opening and Closing stages, and polite devices that will avoid face-threatening acts to the other speaker. Given their task-oriented character, we can expect them to show a structure more uniform and less deviant from an idealized model than casual conversation. Brown (1984) suggests that expectations about structure of a discourse, or discourse fragment, are more specific the higher the task orientation of the discourse in question. The reason for this is that the task usually has a clear structure, consisting of steps that have to be taken in a certain sequence in order to successfully complete the task. The conversations are the product of what Swales calls a sociorhetorical discourse community, one where the members have functional needs as primary determinants of linguistic behaviour (Swales 1990: 24). We can then agree with Bakhtin that this particular function or everyday purpose, together with the social conditions of speech communications have enabled this type of dialogue to develop into a genre of its own:
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.29 (35)
A framework for the analysis of speech genres
A particular function (scientific, technical, commentarial, business, everyday) and the particular conditions of speech communication specific for each sphere give rise to particular genres, that is, certain relatively stable thematic, compositional, and stylistic types of utterances. (Bakhtin 1986: 64)
My final goal in the analysis of these texts is to study the global similarities shared by the dialogues in each language, English and Spanish, independently. We will see how choices in the thematic, rhetorical and cohesive structure build the features of this scheduling genre. I also explore the deviations from the norm, where and when those appear. And finally, I compare both languages in order to reveal possible differences in the construction of the genre. I chose a generic framework because such an approach can help us “theorise how discourses, or texts, are like and unlike each other, and why” (Eggins and Martin 1997: 230) Figure 2.6 represents the analysis described in the rest of the book. The first step is to do an analysis of four different characteristics in the dialogues for each of the languages: – – – –
Thematic structure Rhetorical structure Cohesive structure Staging
The first three types of analysis will be taken into consideration as relating to staging. That is, the thematic, rhetorical and cohesive structures will be studied from the point of view of where they appear in the dialogues. Once I have studied those four elements for each language, I will establish similarities and differences in the realization of the four according to the language. Chapters 4
Figure 2.6 Steps in a contrastive generic analysis of conversation
TSL[v.20020404] Prn:7/10/2004; 9:05
F: PB12902.tex / p.30 (36)
Chapter 2
through 7 cover the characterization and comparison for each of the elements of analysis. Before describing the analysis, I provide, in Chapter 3, a detailed explanation of the corpus and how it was collected. Eggins and Martin (1997) propose a theory of genre and register as a theory of functional variation, allowing for “both textual prediction and contextual deduction” (Eggins and Martin 1997: 236). This study provides a description of a genre that maps contextual variables to text characteristics (textual prediction). Once we understand the characteristics, we will be able to recognize (contextual deduction) a text with those characteristics as belonging to the scheduling genre.
Notes . But see Hasan (1992) for a summary of the incompatibilities between Bakhtin’s views and those of Systemic Functional Linguistics. . Todorov (1990) is also a study of the origin and evolution of literary genres. . The concept of utterance is never clearly defined (Hasan 1992); Bakhtin seems to refer to a complete piece of speech. He sometimes equates it to genre, other times to what I call here a stage within an instance of a genre, but most often it is meant to be a turn in a conversation, which may consist of one or more sentences. Martin (1996), when referring to Bakhtin’s work, uses the word text, rather than utterance. Utterance is the term found in the English translation of Bakhtin’s 1986 essay. . See Lakoff (1987, Book I, Part I) for an excellent summary. . Askehave and Swales (2001) critique the notion of communicative purpose, concluding that it has become complex and difficult to define, and, as a consequence, not an ideal indicator of genre. Their proposed solution is the use of two procedures, one ‘text-driven’ and one ‘context-driven’, where structure, style, content and purpose all contribute to the identification of a genre. The identification is performed in steps, starting with either the text or the discourse community, depending on the procedure. The methods appear to be a combination of genre, register and lexicogrammatical characteristics, in SFL terms.
TSL[v.20020404] Prn:29/07/2004; 11:40
F: PB12903.tex / p.1 (37)
Chapter 3
Data description
The dialogues used in this study were collected by Carnegie Mellon University and The University of Pittsburgh as part of the JANUS project, a large speech-to-speech machine translation project within the Interactive Systems Lab (ISL) at Carnegie Mellon University. The goal of the project was to develop a speech recognizer that could take human voice input, convert it into text and feed it into a machine translation component that would translate into another language, then produce the synthesized speech for that translation. The situation in which such a system would be useful is one where two speakers that cannot communicate in a common language need to carry out a task together. In this case, the task was setting up an appointment. One speaker may be a native speaker of Spanish and the other of English. The system would take the first speaker’s utterance and translate it into English. Then the second speaker would understand what was said, respond to it in English, and have the system deliver the corresponding Spanish translation to the first speaker. In order to develop such a complex system, the project leaders decided to collect naturally-occurring conversations of this type. This effort provided, first of all, the large amount of raw speech needed to implement the speech recognizer, the JANUS Recognizer Toolkit. Secondly, it provided examples of the linguistic structures and vocabulary to help develop the corresponding computational grammars. And thirdly, it served as a model for the dialogue processors in charge of keeping track of the structure of the conversation. The JANUS project is composed of different subprojects funded jointly by the United States Department of Defense under the title Enthusiast Project, by the German Federal Ministry for Education, Science, Research and Technology (BMBF) under project VerbMobil, and by ATR Interpreting Laboratories of Japan. I will not go into detail here about the different destinations or sources of the dialogues, but will consider the corpus as a whole, differentiating only the two languages under consideration and the recording methods, as explained in Section 3.2. For an overview of the JANUS project, see Waibel (1996). The present description is based on different Technical Reports available from the Language Technologies Institute (Ahlén 1996, 1997; Ahlén and Vaidya
TSL[v.20020404] Prn:29/07/2004; 11:40
F: PB12903.tex / p.2 (38)
Chapter 3
1997), and from personal experience with the recording and transcription process. Sondra Ahlén provided valuable information and details when those were needed. My thanks to Alex Waibel and the Interactive Systems Laboratories (ISL) for permission to use the data.
. The task The dialogues involve two speakers who speak the same language, either English or Spanish, and who were recruited through posters, fliers, or word-ofmouth. They were given a small reward for their time. They were brought to a recording laboratory, where, in order to avoid non-verbal communication, they did not face each other. Each of them was equipped with headsets and a microphone. The dialogues are conversations between equals, as far as we know (no details about participants’ backgrounds, other than place of origin, were recorded). Studies in casual conversation usually consider relationships among equals, or equals in appearance – see Eggins and Slade (1997) for a study of the power relationships and social roles in casual conversation. The dialogues in this corpus do have an orientation towards a task (therefore they are not casual conversation), but the participants bring similar power baggage to the task. Neither of them called for the conversation; rather, they both volunteered for it and they both have the same interest in carrying the task to a successful end. The dialogues are, in Hasan’s terms, non-hierarchical dyads (Halliday and Hasan 1985: 57). The instructions explained that the participants have two conflicting agendas covering a period of two to four weeks,1 and that they needed to agree on an appointment lasting for at least two hours sometime within those two or four weeks. They were told to assume that they were already engaged in conversation and decided to make another appointment to meet. They were not given specific instructions as far as the reason for the appointment, the place for the meeting or any other particulars. If they were friends or acquaintances, they were asked to avoid unnecessary discussions not related to scheduling the appointment. The speakers also received brief instructions on how to use the software and two to three test sentences were recorded in order to allow the person in charge of recording to adjust the gain and to give the speakers a chance to get accustomed to the recording software. At that point, either speaker could begin the conversation.
TSL[v.20020404] Prn:29/07/2004; 11:40
F: PB12903.tex / p.3 (39)
Data description
The choice of strategies for data collection is always a first consideration in the study of natural language. Whether to use naturally-occurring or prompted data, for instance, is an issue that will affect the inferences to be drawn from the sample. Roger and Bull provide an excellent discussion of these matters in their 1989 book, specifically in the Introduction to Section 2, ‘Methods of Observation,’ wherein they argue that recordings of conversations occurring spontaneously are more appropriate for research purposes. Yet, it is costly, especially timewise, to collect a corpus of a large enough size by tape-recording conversations surreptitiously. Moreover, conversations recorded in that manner may be excluded from the final collection because the participants refused permission, or because the sound quality was not acceptable. On the other hand, prompted conversations or interviews are handicapped because of their non-natural production. A corpus of interviews is, for instance, the material collected for the study of educated speech in different Spanish-speaking cities (Esgueva and Cantarero 1981; Hernández et al. 1998). Although this is an excellent collection, the participants often refer to the task, even searching for topics of conversation so that they produce a sufficient amount of speech. Another pitfall of prompted conversations and interviews is the observer’s paradox, or “the problem of observing how people speak when they are not being observed” (Labov 1972: 256). Labov (1972), in his studies of inner-city speech, elicited data through interviews. He found that the interview setting might be excessively formal, not allowing the vernacular to emerge. He suggested the use of other methods so that the speakers do not focus on the formal interview setting. In the collection of the ISL corpus, the participants were allowed ample freedom. They were not asked questions or prompted while completing the task. It was hoped that the speakers’ involvement in the task would distract them from the non-natural situation. We cannot be certain about how successful the setting was, since the observer’s paradox may have been a factor in the recording of these conversations. It is impossible to know whether the setting we are observing as researchers is natural or not, simply because we cannot observe a natural setting without making it a non-natural one. The dialogues in the corpus are prompted only to the extent that the participants are given instructions before the conversation proceeds. There are no further prompts, and the speakers were free to carry out their task in whatever manner. The issue of naturalness still remains, but it is in part counterbalanced by the large amount of data that such a procedure allows (the corpus contains a total of 1,380 conversations). In addition, the use of the same situation and scenario for all participants rules out an important number of confounding
TSL[v.20020404] Prn:29/07/2004; 11:40
F: PB12903.tex / p.4 (40)
Chapter 3
variables. The English and Spanish dialogues are truly equivalent as far as the situation and task are concerned.
. Recording Recording was done in two different modes: push-to-talk and cross-talk. In the push-to-talk mode, speakers sit in front of a computer terminal. They receive a message and a beep noise from the system when it is ready to record. The speaker then hits the return key on a keyboard to begin and end recording. While this person is speaking the message on the terminal reads “TRANSMITTING...” until he or she hits the return key (‘Enter’ key) again. When the other person is speaking the terminal message changes to “RECEIVING. . . ” so that the two people cannot talk at the same time. Speakers hear each other through the headphones with only a small time delay. This type of recording was done with a Gradient Desklab Model 14, with a sampling rate of 16 kHz. Turn-taking is controlled by the speakers through the use of a keyboard. Goodwin (1981) discusses different hypotheses for turn-taking strategies, among them Jaffe and Feldstein’s theory that conversation is similar to shortwave radio in the presence of an “unambiguous end of message signal” (Jaffe and Feldstein 1970: 17). When this signal, whatever its nature, is produced, the direction of communication, and the transmitting and receiving roles are reversed. In short-wave radio, there is an explicit signal, such as the word “over”. In everyday communication, gaze, pause, or direct address provide the signal. The conversations found in the corpus are exactly like short-wave radio, in that the signal of end of turn is made explicit by pressing the ‘Enter’ key on a keyboard. In cross-talk there are no mechanical limitations and floor-taking happens in a more natural way. Those recordings are made with a portable DAT recorder. In this style of recording the two speakers can interrupt each other at any time since both record simultaneously, and there were no computers involved in the recording stage. All English data were recorded in the push-to-talk style. For Spanish, both types of recording were experimented. In this study, however, only push-totalk is used, in order to make the two languages comparable with regard to all recording parameters. The English scheduling dialogues were collected at the University of Pittsburgh, from English speakers with different backgrounds. They are all pushto-talk, a total of 881 dialogues.
TSL[v.20020404] Prn:29/07/2004; 11:40
F: PB12903.tex / p.5 (41)
Data description
The Spanish dialogues were mostly recorded in Pittsburgh, of Spanish speakers from several Spanish-speaking countries. Some of the dialogues were recorded by New Mexico State University and some others in a portable DAT recorder in Mexico. The Spanish dialogues total 499 (push-to-talk alone). It is plausible to assume that the Spanish speakers recorded in Pittsburgh were bilingual, or at least fluent in English, although we have no data pertaining to English skill levels.
. Transcription conventions Researchers have elaborated different transcription systems for different purposes: speech recognition, discourse analysis, conversation analysis, or grammar development in Natural Language Processing. Edwards and Lampert (1993) provide an overview of different transcription and coding systems, and their advantages and disadvantages. Based on the articles in Edwards and Lampert it is obvious that there is no agreement as to an ideal system. The ISL/CMU transcription system is influenced by different factors. For instance, the purpose of the transcripts was multifold: they were used to train the speech recognizer, to build the grammars that the parsers used, and to aid in the development of discourse processors for conversation. Because of their different uses, the transcripts contain a great deal of information that may be considered superfluous for any one purpose in particular. In the transcripts that I use for illustration, I decided to leave most of that information in place, even when the reader might find them verbose at times. The conventions described in this section are a set of codings that proved useful and reliable within the context of the JANUS project. Each utterance, or speaker turn, is marked with the initials of the speaker. Those consist of a set of four letters, the first one indicating the gender, “f ” for female and “m” for male, and the next three being the speaker’s first, middle and last name. Thus, “fmtt” would correspond to the female speaker María Teresa Taboada. The utterance marking also includes the scene number, which captures the type of calendar and the number of the dialogue in the case of speakers that record more than one conversation. Finally, the utterance number within this dialogue is included, all separated by underscore signs. The marking of a turn as “fmtt_01_02” would indicate the second turn (02) in the first conversation (01) that speaker FMTT recorded.
TSL[v.20020404] Prn:29/07/2004; 11:40
F: PB12903.tex / p.6 (42)
Chapter 3
The speech includes words, human noises, non-human noises, silences, mispronunciations, false starts, and transcriber comments, all described in the following sections. .. Human noises Human noises are delimited by slashes (/). A human noise is not a word if it does not appear in the dictionary. These are grouped under different categories: – – – –
Space fillers: /mm/, /nn/, /eh/, /ah/, /uh/, /hm/, /um/, /ehm/, /ehn/ Interjections: /m_hm/, /uh_uh/, /uh_huh/, /oh/, /oh_oh/, /oo/ (long [u]) Tag questions: /m_hm/, /hm/, /mm/, /eh/, /huh/, /uh/ Non-linguistic noises: – /lg/ laugh – /cg/ cough – /begin_laugh/ . . . /end_laugh/ laughing while talking
–
Uncommon speech sounds are spelled out, for instance /whistle/
.. Silence Periods of silence are marked with “silence” or “pause” between two asterisks. *pause* is 0.5 seconds to 2 seconds with no sounds at all. *silence* is more than 2 seconds with no sounds at all. .. Mispronunciations There are two kinds of mispronunciations: those which occur at the beginning or end of a word (at the word boundary), and those in the middle of a word (such as metathesis). These are marked in different ways. In the case of mispronunciations at word boundary, the entire word is enclosed within square brackets. The missing part is filled in if possible, within regular brackets. If a speaker meant to say “almuerzo” (“lunch”), but only said “alm”, the transcriber would write: [alm(uerzo)]. For mispronunciations inside a word, the word is typed as it sounds. The correct word is supplied after it with a space between the two words, and all of it surrounded by square brackets. For example, a speaker was trying to say the Spanish word for “twenty eight”, but mispronounced it, so it was transcribed as: [vienteocho (veintiocho)].
TSL[v.20020404] Prn:29/07/2004; 11:40
F: PB12903.tex / p.7 (43)
Data description
.. False starts Occasionally a speaker will start to say something, stop, and then start again. He or she may begin again with the same words, or change their mind and use a different wording. Such false starts are marked with angle brackets: (11) can we meet on Thursday instead of Tuesday
Angle brackets would also be used if a speaker made a mistake and started over with a new thought: (12) my number is /ah/ area code four one two
False starts are not nested. If a speaker has a false start within another false start, only one is marked: (13)
.. Transcriber comments Transcriber comments describe other things in the conversation beyond the level of words and sounds. The intonation comments follow orthographic conventions, marked with an orthographic symbol at the end of the corresponding section of speech (Table 3.1). These comments do not reflect, nor are influenced by, sentence structure. The speaker may have the intonation of a statement when he or she is, in fact, asking a question. He or she may have the falling intonation of the end of a sentence after a collection of words that do not, in any way, resemble a grammatically correct or complete sentence (Sacks et al. 1974). For that reason, the speech following a period or a question mark does not start with a capital letter. For the same reason, beginning question marks (¿) are not used in the Spanish transcripts. Although they are the orthographic norm, they are not related to the intonation at the end of the clause. Table 3.1 Intonation markers Orthographic marker
Intonation
. (period) , (comma)
Falling intonation Slightly rising intonation, continuation of idea, and not a question Marked rising intonation
? (question mark)
TSL[v.20020404] Prn:29/07/2004; 11:40
F: PB12903.tex / p.8 (44)
Chapter 3
Other transcriber comments include {begin_english} . . . {end_english} for words or stretches of talk that were spoken in English within a Spanish dialogue, often the case with names of people and places. In cross-talk conversations, speakers often talk at the same time, which is marked with {begin_simultaneous} . . . {end_simultaneous} in the corresponding utterances. .. Unit markers The original transcripts include what was named a SEOS marker (Semantic End Of Segment). The SEOS marker indicates units, i.e., portions of utterances which can stand alone. It usually coincides with a clause or sentence boundary, but is also used in other places, such as the marking of interjections and short stand-alone words and phrases (well, okay, let’s see, etc.). The stretches of talk occurring between a SEOS and the next one are called SDUs (Semantic Dialogue Units). In most cases, SDUs are comparable to Chafe’s information units (Chafe 1980), which, according to Halliday, are realized in tone groups: one information unit is realized as one tone group (Halliday 1967; Crystal 1969). A similar construct is the idea of turn-constructional units in Conversation Analysis (Sacks et al. 1974). In the transcriptions, sometimes the same information unit is broken down in different SDUs when there is a long pause or a restart. The original SEOS markers in the transcripts are represented with a vertical bar, or pipe (|). Most transcription systems call for a new line for each new unit of analysis (Du Bois et al. 1993). I chose instead to make the transcriptions continuous, in order to save space on the page, and to use the vertical bar to signal the end of a unit. SDUs were the starting point for the analysis, then further refined to result in clauses as units of analysis (further detail is provided in Chapter 4, Section 4.3). An SDU is any instance of talk between two bars, between the beginning of the turn and a bar, and between a bar and the end of turn. Bars are added even when other intonation markers (period, comma, question mark) are present, to clearly delimit the SDU. .. Other conventions All numbers are expanded into their spellings. Spellings in the dialogues appear in capital letters with a space between each one: my last name is Taboada, T A B O A D A.
TSL[v.20020404] Prn:29/07/2004; 11:40
F: PB12903.tex / p.9 (45)
Data description
Table 3.2 Summary of transcription conventions Category
Bracket
Human noises Non-human noises Silences Mispronunciations
/.../ slashes #...# hash marks/pound signs *...* asterisks [...] square brackets (around whole word) (...) parentheses (supply missing part of word or correct pronunciation of word, only inside square brackets) {...} curly braces |...| vertical bars/pipes angled brackets SEOS (Semantic End of Segment), unit boundary
Transcriber comments Accent False starts |
The following abbreviations are used for titles: Dr for doctor, Mr for mister, Ms for miz, Mrs for missus. A single quote can be used to denote missing sounds: ’em for them, ’cause for because. Vertical bars (or pipes) signal a word pronounced with a foreign accent. In the Spanish dialogues, speakers sometimes use English words, but pronounce them with a Spanish accent. The examples contains instances such as |s|okay, to indicate that the word okay is pronounced with a Spanish accent. These words are also enclosed by the {begin_english} . . . {end_english} markers. These vertical bars are different from the unit markers, in that they enclose only one letter (s for Spanish accent and e for English accent), and they are at the beginning of a word. .. Summary of transcription conventions A summary of the transcription conventions used in the corpus can be found in Table 3.2.
. Speaker pairs and dialogues For this analysis, I decided to extract an equal number of dialogues in English and Spanish. I wanted to have a balance of the two genders, so that there would be equal numbers of female-female, male-male, and female-male (or vice versa) conversations. Gender is not a variable in the study, and no generalizations are made with respect to gender. The balanced numbers in terms of
TSL[v.20020404] Prn:29/07/2004; 11:40
F: PB12903.tex / p.10 (46)
Chapter 3
Table 3.3 Dialogues in the English corpus FACR_FJYK_AU FAMS_FCLD_AU FBNT_MJFG_1 FCAD_FJAB_12 FCAE_FPAM_3 FCKA_FFMW_10 FCRE_FJSL_1 FEAS_MTMR_11
FECE_MNFH_10 FHKR_FDMR_JA FJBT_MSRH_11 FJLH_MCDE_1 FJMB_FMLZ_6 FKCF_MJCB_1 FKNB_MCJC_10 FNBS_MRJK_10
FRET_MRCT_1 FSEH_FPWJ_1 FRJP_MPMM_1 FSJB_FKDO_1 MAEM_MJAY_1 MAGH_MTNZ_11 MAJH_MRGM_1 MAKK_MKGD_10
MAWK_MANV_1 MBMS_MLAB_10 MGCT_MYAW_1 MGFR_MCBN_AP MJMB_MBRT_1 MTJH_MRIL_3
MACC_MRNN_01 MARC_MPHB_02 MBRP_MFJR_05 MEJH_MJJG_01 MJBP_MMBU_04 MJEG_MFAC_01 MJMG_MSNC_04 MJNM_FAMM_06
MJOS_FSPS_01 MOFC_MFMM_01 MRAP_FKAS_01 MRBZ_MCRA_03 MRMV_FFCS_06 MWMO_MMBU_09
Table 3.4 Dialogues in the Spanish corpus FCBA_MEBA_08 FFCS_FSNM_01 FJGC_FKAS_04 FKND_FKAS_01 FLCM_FKAS_08 FLNW_MPNE_08 FMBO_MENC_09 FMEM_MEOC_-02
FMGL_FMCS_01 FMNA_FSNM_06 FNBA_FCBA_04 FPDV_FSNM_04 FPLA_MLMG_10 FSMA_MENF_05 FVGC_FSNM_09 FYMM_FKAS_02
gender were sought merely in order to avoid possible gender-based differences. Ten conversations of each type were selected, which yields a total number of 30 conversations for each language. I decided that 60 should be the maximum number, given the intense analyses I intended to perform. Research manuals for applied linguistics advise on 30 as a minimum ideal number, if extraction to a general population is sought (Hatch and Lazaraton 1991). After the gender selection, the conversations were chosen from the pool following one requirement: that there should not be more than one conversation by the same speaker, or at least, that there should be as much variety of speakers as possible. This was feasible in English, since there was a larger pool from which to choose. In Spanish, I was forced to use conversations where a speaker or two were repeated. This section lists all the dialogues in the study, characteristics of the speakers, and statistics summarizing the data used. Tables 3.3 and 3.4 list the dialogues (speaker pairs) used in English and Spanish, respectively. Tables 3.5 and 3.6 list age, education, and place of origin for the English speakers, broken down in a table each, for female and male speakers. Tables 3.7 and 3.8 provide the same information for the Spanish pairs. Speakers with an asterisk after their name have some noticeable foreign accent, that is, an accent that cannot be considered one of the varieties of either English or Spanish. “N/A” stands for “not available”.
TSL[v.20020404] Prn:29/07/2004; 11:40
F: PB12903.tex / p.11 (47)
Data description
Table 3.5 Characteristics of the English female speakers Speaker
Age
Education
Raised In
FACR FAMS FBNT FCAD FCAE FCKA* FCLD FCRE FDMR FEAS FECE FFMW FHKR FJAB FJBT FJLH* FJMB FJSL FJYK* FKCF FKDO FKNB FMLZ FNBS FPAM FPWJ FRET FRJP FSEH FSJB*
23 19 21 21 23 22 19 N/A 23 20 N/A 19 N/A 21 49 24 33 26 21 23 48 19 33 18 35 34 23 N/A 31 35
Undergraduate Undergraduate Undergraduate Undergraduate Undergraduate Graduate Undergraduate N/A Undergraduate Undergraduate N/A Undergraduate N/A Undergraduate Undergraduate High School Graduate Graduate Undergraduate Graduate Undergraduate Undergraduate High School Undergraduate Undergraduate Graduate Undergraduate N/A Graduate High School
New Jersey Pennsylvania Pennsylvania Virginia Pennsylvania Canada, Egypt, Germany, Greece Pennsylvania N/A Pennsylvania Florida N/A Pennsylvania N/A Tennessee Toronto Texas Pennsylvania Washington Taiwan California Pennsylvania Utah Pennsylvania Pennsylvania Pennsylvania Virginia Texas, Connecticut N/A Illinois, Florida, Virginia Pennsylvania
As we can see from the tables, the pools of English and Spanish speakers are slightly different. Whereas the vast majority of English speakers come from Pennsylvania, and specifically from Pittsburgh, the Spanish speakers have varied origins. This poses a problem for generalizations from the Spanish data, because the sample is not homogeneous, whereas the English sample is. All claims are to be understood with respect to a standard Spanish, as far as it is represented in this sample of the Spanish-speaking population on the one hand, and a standard American English on the other.
TSL[v.20020404] Prn:29/07/2004; 11:40
F: PB12903.tex / p.12 (48)
Chapter 3
Table 3.6 Characteristics of the English male speakers Speaker
Age
Education
Raised In
MAEM MAGH MAJH MAKK MANV* MAWK MBMS MBRT MCBN* MCDE MCJC MGCT MGFR MJAY MJCB MJFG MJMB MKGD MLAB MNFH* MPMM MRCT MRGM MRIL MRJK MSRH MTJH MTMR MTNZ MYAW
25 34 21 18 18 20 19 19 20 31 26 N/A 46 30 23 17 N/A 22 18 20 20 25 23 23 21 20 29 17 26 19
Undergraduate Graduate Undergraduate Undergraduate Undergraduate Undergraduate Undergraduate Undergraduate Undergraduate High School Undergraduate N/A Graduate Graduate Graduate Undergraduate N/A Graduate Undergraduate Undergraduate Undergraduate Graduate Undergraduate Graduate Undergraduate Undergraduate Undergraduate Undergraduate Graduate Undergraduate
Arizona Germany, Lebanon, USA Pennsylvania New York Bombay, India Florida California, New York Pennsylvania England and Florida Pittsburgh Maryland N/A Wisconsin New York Minnesota California N/A Pennsylvania Pennsylvania Taiwan New York Texas and Connecticut Connecticut New Jersey New York Pennsylvania Indiana Pennsylvania Germany, Pennsylvania California
The next table, 3.9 summarizes the number of turns, SDUs, words, and the number of items (words plus human or non-human noises) for both languages. Average length of SDUs and words for each language is given in Table 3.10. The same figures are broken down by gender in Table 3.11. Spanish conversations, although slightly shorter in turns, pack more clauses and words in each turn than did the English ones. We can see, in Table 3.11, that English female-female conversations tend to be longer in all accounts: turns, SDUs, and words. The reverse is true for Span-
TSL[v.20020404] Prn:29/07/2004; 11:40
F: PB12903.tex / p.13 (49)
Data description
Table 3.7 Characteristics of the Spanish female speakers Speaker
Age
Education
Raised In
FAMM FCBA FFCS FJGC FKAS FKND FLCM FLNW FMBO FMCS FMEM FMGL FMNA FNBA FPDV FPLA FSMA FSNM FSPS FVGC FYMM
21 18 25 28 27 27 27 40 35 35 35 30 26 23 34 28 27 24 21 27 24
Undergraduate Less than High School Graduate Graduate Graduate Undergraduate Graduate Graduate Graduate Undergraduate Graduate Graduate Graduate Graduate Undergraduate Graduate N/A Undergraduate Undergraduate Graduate Undergraduate
Florida Peru Costa Rica New York Idaho Peru Spain Panama Argentina Argentina Italy Argentina Argentina Mexico Peru Argentina Argentina Pennsylvania Mexico Argentina Venezuela
ish: male-male conversations are longer among the Spanish speakers. Mixed conversations in both languages are within the average for turns taken by the speakers, but they are slightly shorter in terms of SDUs and words.
. Use of terms and translations Some clarifications are in order here with regard to the use of terms such as “utterance,” “turn,” or “speaker”. I use “utterance” and “turn” to refer to the same thing, namely, a speaker’s turn without interruption by the other speaker, marked in the transcripts by the name of the speaker and the number of the turn at the beginning of the contribution. “Conversation” and “dialogue” are also used synonymously throughout this work. Some authors (Weigand 1994) distinguish between discourse, dialogue, and conversation, and their respective analyses (discourse analysis, dialogue analysis, conversation analysis). I refer to the interactions as either conversations or dialogues, which are all ultimately discourse events.
TSL[v.20020404] Prn:29/07/2004; 11:40
F: PB12903.tex / p.14 (50)
Chapter 3
Table 3.8 Characteristics of the Spanish male speakers Speaker
Age
Education
Raised In
MACC MARC MBRP MCRA MEBA MEJH MENC MENF MEOC MFAC MFJR MFMM MJBP MJEG MJJG MJMG MJNM MJOS MLMG MMBU MOFC MPHB MPNE MRAP MRBZ MRMV MRNN MSNC MWMO
32 29 31 20 19 19 30 26 52 27 30 26 24 17 40 28 23 36 32 N/A 33 23 38 25 21 39 26 26 16
Undergraduate Undergraduate Graduate Undergraduate High School Undergraduate Undergraduate Graduate Undergraduate Undergraduate Undergraduate Undergraduate Undergraduate Undergraduate Graduate Graduate Graduate Graduate Graduate Undergraduate Graduate Undergraduate Graduate Graduate Undergraduate Graduate Undergraduate Undergraduate Graduate
Mexico Venezuela Peru Mexico Mexico Venezuela Argentina Argentina Colombia Mexico Peru Spain Dominican Republic Pennsylvania Mexico Spain Colombia Mexico Colombia Dominican Republic Mexico Uruguay Chile Salvador Mexico Chile Chile Argentina Peru
Table 3.9 Counts in the English and Spanish corpora
Turns SDUs Words Items
English
Spanish
249 784 6804 11056
248 1294 9112 12788
The words “speaker” and “hearer” refer to the current speaker and hearer in the moment of the description. Those roles switch in the next turn taken: the current speaker will become the hearer, or listener, and vice versa. When
TSL[v.20020404] Prn:29/07/2004; 11:40
F: PB12903.tex / p.15 (51)
Data description
Table 3.10 Average length of English and Spanish dialogues
Average length of dialogue – turns Average length of dialogue – SDUs Average length of dialogue – words
English
Spanish
8.30 26.13 225.93
8.27 43.10 302.43
Table 3.11 Average length of English and Spanish dialogues by gender f-f Av. length – turns Av. length – SDUs Av. length – words
8.9 32 285.8
English m-m 7.5 25.1 203.9
f-m
f-f
8.5 21.3 188.1
8.2 39.1 291.7
Spanish m-m 8.4 49.5 342.5
f-m 8.2 40.7 273.1
the distinction is not necessary, I simply refer to them as, for instance, speaker FMTT and speaker MONS to describe a conversation between these two participants. Given that we have access to the speakers’ sex through the first initial in their assigned identifiers, I use the appropriate pronoun whenever I have that information. The Spanish examples are translated in a style that lies between literal translation and perfect English. The translations are broken down according to the SEOS unit markers, represented in the transcripts with vertical bars (|). The unit markers are also included in the English translation of Spanish examples, in order to facilitate the mapping between the two languages. Whenever possible, I tried to preserve a word order that reflected the Spanish syntax, so that the correspondences were more transparent.
. Other remarks on the corpus The dialogues included in this study are a sample of a general population: the total 1,380 dialogues collected by the ISL. As I have suggested (see Section 3.1), these dialogues were themselves collected as representative of task-oriented dialogues, especially of scheduling dialogues. I have also mentioned the problems associated with data collection. Yet, unless I avail myself of a similarly sized set of conversations where no researcher and recording were obvious, I have no means of establishing whether our corpus is representative of naturallyoccurring dialogue. Thus, as far as the step from natural setting to corpus is
TSL[v.20020404] Prn:29/07/2004; 11:40
F: PB12903.tex / p.16 (52)
Chapter 3
concerned, I had to assume, with reservations, that the sample was representative enough. The second step consisted of selecting the dialogues to be included in the study. One method is random selection, with assignment of numbers then chosen according to a table of random numbers. Each dialogue should have, then, an exactly equal chance of being selected. However, I also wanted to include a balance in terms of gender, and as much speaker variety as possible. I decided on three sets of speaker pairs: both female, both male, and mixed. This decision lead to a choice of ten conversations in each group, which yielded a total of 30 for each language, 60 in total. Thirty is considered a large enough number as population representative, provided that the selection has been adequate (Hatch and Lazaraton 1991). I am confident that the dialogues selected for the study are representative of the phenomena present in the rest of the large-scale corpus. There still remained an issue of how much to generalize from the results obtained. Given the restrictive nature of the corpus, I would not make bold claims about the characteristics of the languages in general, but it remains my belief that this is an exhaustive study on the construction of task-oriented conversation. Generalizations with respect to other genres are speculative at this point. I will mention some of the observations in that regard as we come across them. One last point to raise is that this is a cross-linguistic study, not a crosscultural one. The Spanish speakers were not in their usual context of culture, for most of the Spanish recordings took place in Pittsburgh. In addition, the sample of Spanish speakers is also more varied than the English, since the Spanish speakers came from different parts of the Spanish-speaking world, where cultural and linguistic norms may be different. A number of studies point in that direction: Fant (1996), for instance, discusses differences in negotiation styles between Mexican and Peninsular speakers of Spanish; Márquez Reiter (2002) compares Peninsular and Uruguayan speakers of Spanish in their indirect requests; and Félix-Brasdefer (2003) discusses Latin American speakers speaking Spanish versus Americans speaking either English or Spanish. The focus in this book will be on the similarities across dialects of Spanish; not on the fine-grained differences.
Note . The instructions were provided in the language of recording, Spanish or English.
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.1 (53)
Chapter 4
The thematic structure of dialogue
In the clause, there is a transition from the first element towards the rest of the elements. At the same time, there exists a transition from the first few clauses in the discourse towards the rest of that discourse. These two types of transitions are reflected in the study of thematic realization and thematic progression. Discourse, whether in written or spoken form, takes place in a linear fashion. When writing we can only spell out one letter at a time – letters that will eventually become parts of words, sentences, and full texts. The same applies to reading: although we might chunk information and read not exactly one word at a time, but in saccades,1 the chunks succeed each other linearly. In spoken language roughly the same principles are at work: we can produce sounds only one after another, and we can process them only in a linear temporal sequence. The reader could argue that this is a simplistic view; we know that a great deal of information is pre-packaged in production, and understood in packages – most likely with the morpheme as the basic building block. Spoken language phenomena such as liaison and spoonerism2 would challenge an oversimplified account of processing one phoneme, or one word, at a time. Language is linear, nevertheless, for the most part, whether the items in succession are sounds, words, or semantic units. Given that speech is linear, the process of its unfolding is very important. We often discuss the importance of first impressions in the process of getting to know someone. The first impression that a piece of discourse produces on the listener is of foremost importance. It usually provides clues about the topic and register of the discourse to follow, and it also serves as a connection with the preceding discourse – or the preceding context, if we are considering the first utterance or utterances in a text. The linearity of information distribution is in fact related to the internal structure of the information unit, whereby the speaker organizes the components of the message block in such a way as to specify its relation to what has preceded. (Halliday 1967: 202)
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.2 (54)
Chapter 4
Perfetti and Goldman (1974) suggest that thematization (placing an element in first position) is instrumental in recall. In a series of experiments, they established that subjects were more likely to recall the element that appeared first in a sentence, regardless of its status as old or new information. MacWhinney remarks too on the importance of initial elements in language: “The speaker uses the first element in the English sentence as a starting point for the organization of the sentence as a whole. Similarly, the listener uses the first element in a sentence as a starting point in comprehension” (MacWhinney 1977: 152). It is not only the very first piece of the discourse that concerns us here. Once a stretch of talk has been produced, we find ourselves observing the next piece as if it were the first: we use it as an anchor to the previous discourse, and as a starting point for the following. Once we have processed this next piece – either as speakers or as listeners – another piece will follow. And so on with the rest of the discourse. This is the basis of the dynamic perspective in the study of language (Martin 1985b; Ravelli 1995). In a dynamic framework – which is opposed to a synoptic one – a text unfolds as a choice determined by the text leading up to the current point and by the context within which the text as a whole occurs. “A dynamic perspective means that options become available as a text unfolds: wherever a text is now provides the environment for what may follow” (Ravelli 1995: 203–204). Bakhtin applies the dynamic notion to every instance of speech that humans have ever produced. All discourse points towards the “already uttered”. Only a historical first speaker could have uttered something that was not related to any previous discourse. The dialogic orientation of discourse is a phenomenon that is, of course, a property of any discourse. On all its various routes toward the object, in all its directions, the word encounters an alien word and cannot help encountering it in a living, tension-filled interaction. Only the mythical Adam, who approached a virginal and as yet verbally unqualified world with the first word, could really have escaped from start to finish this dialogic inter-orientation with the alien word that occurs in the object. (Bakhtin 1981: 279)
Discourse in general, and every unit of discourse in particular, have backwardlooking links. But discourse also moves forward; otherwise nothing new would ever be uttered. That constitutes the most fundamental discourse articulation: a link to previous material, and a new contribution. That articulation is true of text, and also of the sentence, which we will consider for now a unit of discourse. Vallduví mentions the agreement – despite disagreements concerning the details – that “in the sentence there is some sort of informational split be-
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.3 (55)
The thematic structure of dialogue
tween a more informative part and a less informative part” (Vallduví 1990: 35). The split happens at the point where the part that is used as an anchor to the information, in order to guarantee that the information will enter the hearer’s knowledge store, gives way to the informative part.3 Names for the two “parts” of discourse are varied, and the exact definition of those two parts is also elusive. Vallduví and Engdahl (1996) revise the different names given to what they call focus and ground. For focus: new, NewInfo, rheme, dominant constituent. For ground: background, presupposition, openproposition, given, theme, topic (Vallduví and Engdahl 1996: 465). While most of the literature outside Systemic Functional Linguistics characterizes the split in terms of information, my emphasis here is on a linear split between the beginning and the end of the clause. The focus of this chapter is the exploration of how linearity and adjacency contribute to the development of a text that is being built in collaboration by two interlocutors, each using a common system of choices. Because texts are linear, and because they have to develop in time and/or space, I will consider what devices for the articulation and progression of information are available to the speakers of each of the two languages, English and Spanish. I will relate this surface progression to the underlying progression in terms of rhetorical relations that I explore in Chapter 5, and to the staging of the dialogues as described in Chapter 7. First we will examine discourse articulation in terms of Theme and Rheme.
. The many accounts of Theme-like concepts In this section I will outline a review of the most relevant literature on the concept of Theme. Comprehensive descriptions of the concept in general and its place in the theory of Systemic Functional Linguistics can be found in Berry (1996), Downing and Locke (2002), Eggins (1994), Halliday (1967, 1994), Jiménez Juliá (1986), Lavid (2000b), and Martin (1992), among others. Martin and colleagues (1997) provide an excellent summary, with exercises. Leong Ping (2000) proposes an alternative view of Theme, relating it to schema theory, and the cognitive processing of language. The School of Prague’s approach to Theme, within the theory of Functional Sentence Perspective, is covered in Daneš (1974b) and Firbas (1974). More recently, Hajiˇcová and colleagues have developed a new approach stemming from the Prague School tradition, which they denominate Topic-Focus Articulation (Hajiˇcová and Sgall 1988; Hajiˇcová 1991; Hajiˇcová et al. 1998a; Hajiˇcová et al. 1998b). A formal approach, in
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.4 (56)
Chapter 4
the concepts of topic and comment – a distinction first proposed by Hockett (1958) – can be found in Dahl (1969, 1974), Gundel (1977), and Kuno (1972). Finally, the following provide more specialized discussions, some of them contrasting and comparing different approaches: Downing (1990, 1991), Fries (1983), the papers in Ghadessy (1995), Gómez-González (2001), Hasan and Fries (1995a), Martin (1995), Rose (2001), Vallduví (1990), Vallduví and Engdahl (1996), and Williams (1988). I leave out of the discussion the concept of focus as non-presupposed element (Chomsky 1971; Jackendoff 1972), the many applications of that concept (Erteschik-Shir 1997; Zubizarreta 1998), and also the concept of focus in Artificial Intelligence (Grosz and Sidner 1986; Hajiˇcová 1987), concentrating on the dichotomies presented in Table 4.1. For an overview of the different types of focus, see Gundel (1999), and for the relationship between topic and focus, Gundel and Fretheim (2004). Büring (1997) summarizes the distinctions between background/topic and focus, and studies the intonational realization of focus. The starting point for understanding Theme/Rheme articulation and thematic progression (henceforth, TP) is the textual metafunction of language, as defined in Systemic Functional Linguistics. Theme is concerned with the information structure of the clause; with the status of the elements not as participants in extralinguistic processes but as components of a message; with the relation of what is being said to what has gone before in the discourse. (Halliday 1967: 199)
The textual metafunction is one of three in which language as discourse is structured to function as an organized message. The clause is the basic meaning-making tool, and it expresses three different kinds of meaning: – – –
Ideational, the linguistic representation of the world and how we experience it (clause as representation). Interpersonal, the relationship between speakers and hearers (clause as exchange). Textual, the organization of the text as a whole (clause as message).
These three meanings find their linguistic expression in different grammatical systems belonging to different functional regions of the grammar. The ideational meaning is expressed in the Transitivity structure of the clause, the interpersonal in the Mood, and the textual in the Theme. We will be concerned here with the textual metafunction and its expression in the Theme of the clause. Let us now look at one definition:
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.5 (57)
The thematic structure of dialogue
The Theme is the element which serves as the point of departure of the message; it is that with which the clause is concerned. The remainder of the message, the part in which the Theme is developed, is called [...] the Rheme. As a message structure, therefore, a clause consists of a Theme accompanied by a Rheme; and the structure is expressed by the order – whatever is chosen as the Theme is put first. (Halliday 1994: 37)
The preceding is a compact definition; it contains a great deal of information on what Theme is. And, mainly, it contains what some critics have pointed out is a double definition (Fries 1983; Downing 1990, 1991). The “element which serves as the point of departure of the message” and “that with which the clause is concerned” are, for some linguists, two quite different things. Fries calls those linguists the separators, as opposed to the combiners, who would like both ideas to define Theme (Fries 1983). The origin of the combining approach dates back as early as Mathesius’ 1939 work. Gómez-González (2001) also covers the differences between these two approaches exhaustively. She characterizes the dual definition as two different metaphors: a spatial metaphor (point of departure), and a matter metaphor (what the clause is about), which are often presented as different but equivalent glosses for Theme (Gómez-González 2001: 115). The second part of the definition, “that with which the clause is concerned”, refers to the relationship of Theme to information. Information structure, articulated in terms of Given and New, is not the same as thematic structure. Halliday again settles the matter: [W]hile ‘given’ means ‘what you were talking about’ (or ‘what I was talking about before’), ‘theme’ means ‘what I am talking about’ (or ‘what I am talking about now’); and, as any student of rhetoric knows, the two do not necessarily coincide. (Halliday 1967: 212)
That is, whereas there is a tendency for given information to come first in the sentence, and new information to appear later, it does not follow that Theme equals Given, or known information. Also, new information is usually identified with the focus, or most prominent information in the clause. At this point, we need a schematic identification of the terms we are examining. This is provided in Table 4.1. The table does not include other uses of the word theme to refer to, for instance, the main idea or subject topic4 of a text (Jones 1977). A more detailed summary is also provided by Östman and Virtanen (1999). As it will be evident by now, the literature on these issues is often contradictory. I will not contribute further confusion, but will instead gather a few sources from which I defined Theme for the purpose of the present study. The
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.6 (58)
Chapter 4
Table 4.1 Summary of Theme-related concepts Concept
Definition
School and Author(s)
Theme/Rheme
Theme – point of departure; what the clause is about Rheme – rest of the clause Given – recoverable information New – non-recoverable info.
Systemic Functional Linguistics (Halliday 1967, 1994)
Given/New Topic/Focus
Topic/Comment
Topic – items the speaker considers are shared with the hearer Focus – modifications or assignment of properties to the topic Topic – what the sentence is about Comment – what we say about the topic
Systemic Functional Linguistics (Halliday 1967, 1994) (New) Prague School (Hajiˇcová and Sgall 1988; Hajiˇcová et al. 1998b) (Generative) Linguistics Chomsky (1971), Jackendoff (1972), Kuno (1972), Gundel (1977)
first one of them is the work of Fries in the thematic organization of English discourse. For him, “[t]he theme-rheme organization of the sentence forms part of a larger pattern which governs the flow of information in any English discourse” (Fries 1983: 144). Information enters the picture again, but this time in the form of a tendency to correlate Theme and information. His definition makes it clearer. The theme of a clause or sentence is the point of departure of that clause or sentence as message. In English the theme is realized as the initial constituent of the clause or sentence. The information that is contained within the themes of the various sentences of a passage correlates with the method of development of the passage, and, if the passage is outlineable, the outline structure of the passage. The meaning of theme and its realization in English explains the general correlation of given information with the initial position in the sentence. (Fries 1983: 116)
Fries entertains a general correlation between Theme and given information, but not a necessary one. Likewise, he considers there is a correlation between N-Rheme – that part of the Rheme that contains only new information – and emphasis or contrastiveness. However, the correlation is not absolute, and therefore it should not be part of the definition of Theme. His analysis is based on the T-unit (Hunt 1977), an “independent conjoinable clause complex”, that is, an independent clause together with all the clauses which are dependent on it, in a relation of hypotaxis or embedding. He establishes the Theme of a
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.7 (59)
The thematic structure of dialogue
T-unit to be the element that provides a framework within which the Rheme of that T-unit can be interpreted (Fries 1994: 230). Downing also considers framework-setting as the main purpose of Theme in discourse. She discusses marked and unmarked Themes in different mood structures, and concludes that, although the thematic element in all of them is the point of departure, it is not always “what the clause is about”. Because Theme is identified with first position in the sentence, and because English, in the unmarked case, has a strict word order, the following will be Themes5 (Theme is italicized). (14) a. b. c. d.
Subject in a declarative clause – The Gauls sacked Rome. Finite + S in a polar interrogative – Did the Gauls sack Rome? Wh-element in a Wh-interrogative – What did the Gauls do? Predicator or Let’s in an imperative – Sack Rome! / Let’s sack Rome!
In the case of marked word order, that is, when we have a fronted Complement or Adjunct in the clause, the fronted element will be the Theme of the clause, as in the following examples. (15) a. b. c. d.
Fronted Adjunct – In 390 B.C. the Gauls sacked Rome. Fronted Object – Rome they sacked. Fronted Complement – Furious they were. Fronted Predicator + its complementation – Exulted the chief of the Gauls: Sack Rome we did.
Downing points out that some of those which are Themes by definition do not fill both parts of Halliday’s characterization of Theme. They might be the point of departure, but it is obvious that (15a) is not “what the message is about”. Although Theme always includes an ideational component in the case of multiple Themes, some ideational Themes do not necessarily coincide with the topic of the clause, which would be what the clause is about. Downing (1991) was the first to distinguish between the Theme of a clause and its topic, and the first one to point out that not all ideational Themes are topical Themes. Before we continue the discussion, we will explore the idea of multiple Theme just mentioned. I already remarked, in Chapter 2 and earlier in this section, on the classification of language in three metafunctions – ideational, interpersonal, and textual.6 The beauty of the classification is that it applies to all strata in the language, so that even Theme can be further classified into ideational, interpersonal, and textual. The reason is that sometimes the first element in the sentence is a conjunctive or modal element. Those are outside the experien-
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.8 (60)
Chapter 4
tial structure of the clause and, therefore, until a Participant, Circumstance or Process appears, “the clause still lacks an anchorage in the realm of experience” (Halliday 1994: 53). Thus, a thematic element (a continuative, a conjunction or a conjunctive adjunct), an interpersonal (a vocative, a modal or a mood-marking element) can be thematic elements as well. Example (16) is an invented example that contains the three types. Textual Theme is italicized, interpersonal Theme is in bold face, and ideational Theme is underlined. Section 4.2 will provide a more detailed explanation of what elements can appear as textual, interpersonal and ideational Themes. (16) Actually, Jennifer, I can’t meet on Mondays.
The order of the metafunctions in a multiple Theme is fixed (textual ^ interpersonal ^ topical/ideational), as shown in Figure 4.1, which also displays the realization of marked and unmarked Theme.7 The figure uses some of the Systemic Functional Grammar conventions for outlining choices in the language. Figure 4.1 is a system, which is read from left to right. The curly bracket on the left represents simultaneous choice. In the Theme system, there are three simultaneous choices, the first being whether the Theme will be single or multiple. That is an exclusive choice, represented by the square bracket after the entry point. A downward arrow indicates realization. If the Theme is single, it will be realized in the topical Theme. If, on the other hand, it is multiple, three exclusive choices are possible. If attitudinal, for instance, the realization will include both topical and interpersonal Theme. The next line in that choice indicates the order in which they have to be realized, with interpersonal preceding topical (the ^ symbol represents linear order). The rest of the figure follows similar conventions to represent choice. The figure represents the thematic options in English, especially as regards the realization of unmarked Theme in different moods (declarative, interrogative, etc.), which is different in Spanish. In the figure, we can see how multiple Theme always includes a topical element as the closing element for the thematic part of the sentence. Downing (1991), as I mentioned earlier, points out that, even in the case where we include everything (i.e., textual and interpersonal Themes) up to the first ideational element in the Theme of the clause, that ideational element might still not be what the clause is about (i.e., its topic). We should not, then, identify ideational and topical Themes. She distinguishes ideational Theme from topical Theme, an ideational Theme that coincides with the topic of the clause. Ideational Themes that do not coincide with the topic of the sentence include circumstantial Themes, as in Example (15a) above. Downing then explores the
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.9 (61)
The thematic structure of dialogue
single + topical Theme
attitudinal
+ topical Theme; + interpersonal Theme
interpersonal ^ topical multiple
conjunctive
+ textual Theme
textual ^ topical both
+ textual Theme; interpersonal Theme
clause
textual ^ interpersonal ^ topical unmarked Subject/Theme [declarative] Wh/Theme [wh-interrogative] Finite/Theme [interrogative] Process/Theme [imperative]
marked
other/Theme
predicated
Theme: it + be + ...
Rheme: that/who ... not predicated
Figure 4.1 System network for textual meaning in the clause
function of Theme as the element that helps establish a framework for the interpretation of the rest of the clause. For example, a participant Theme establishes an individual framework, and is likely to be the topic of the clause (what the clause is about). On the other hand, a spatial or temporal Theme establishes a circumstantial framework, and is not the topic of the clause. Once Theme and topic are clearly defined, each can be studied separately. For example, Downing (2000) and Moya Guijarro (2000) have analyzed the characteristics of topic. Moya Guijarro and Albentosa Hernández (2001) present a study, in which Theme and topic are examined separately. They found that there is no oneto-one correlation, and that the degree of correlation may be related to text genre. This leads to an organizational and semantic definition for Theme: the “point of departure of the message” on the one hand, and the element that es-
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.10 (62)
Chapter 4
tablishes a framework for the rest of the clause – although not necessarily “what the clause is about” – on the other. The idea of Theme as “starting-point”, or “peg on which the message is hung” has been criticized as being metaphorical and difficult to interpret (Fries 1995a). The conjunction of this metaphor together with a structural definition (first place in the clause) provides an objective method of analysis. My proposal is to place more emphasis on the organizational definition. It is a fact that Theme, topic, and Subject very often coincide in a clause, because they all are realized in first position. The problem, as has been well established in the literature, is what to do with each of those three concepts when they are not realized by the same element. I consider Theme as the point of departure for the message, the point of attachment for whatever comes next in the clause. Sometimes that element will also be what the clause is about, but we do not need an absolute identification of Theme and topic. Topic can appear anywhere in the clause, and it is always ideational. The first ideational element, on the other hand, is not always the topic (Downing 1990: 127). Gómez-González remarks that the definitions identifying Theme and topic involve a great deal of subjectivity (Gómez-González 2001: 16). Deciding on the topic of discourse, the topic of a paragraph or the topic of a clause is not straightforward. This is another reason to maintain topic outside our discussion. However, as Halliday points out, Theme needs an anchor in the ideational level of the language, and that is why I propose to keep his structural consideration of including every element in the clause up to the first ideational element as the Theme. The metafunction choices for Theme are usually described as textual, interpersonal and topical (see Figure 4.1), and realized in that order in the clause. I will avoid the term topical Theme in favour of ideational because I do not claim that Theme, under the definition considered for this work, coincides with the topic of the clause. It will always be true, however, that the Theme is the first ideational element (preceded or not by textual and interpersonal Themes). Downing has already established a dissociation of the Theme of the clause and the topic, whether clause- or discourse-based (Downing 1991; Downing and Locke 2002). In addition, the label ‘ideational’ follows the labels for the other two types of Themes, textual and interpersonal, since they represent the three language metafunctions. It could be argued that, if not defined semantically, Theme will not be useful in textual analysis. I defend that Theme, as a choice that the speaker has in constructing his or her discourse, is important in its own right.
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.11 (63)
The thematic structure of dialogue
The choice of Theme is important because it represents the angle from which the speaker projects his/her message, and partially conditions how the message develops. The initial element acts as a signal to the hearer, directing expectations regarding the structure that is likely to follow, or about the mental representation of what the message is likely to be. (Downing and Locke 2002: 223)
Thus, theme is the point of departure of the message, which is necessarily structurally realized in first position of the clause, since a point of departure in mid-sentence would simply be a point of continuation. Clauses have only one ideational Theme, which can be preceded by textual and/or interpersonal Themes. If a clause begins with an Adjunct (an ideational element), that would be the only Theme in the clause. This is contrary to the idea that the Subject is always the unmarked Theme, and if an Adjunct precedes it, then the clause has two Themes, one marked and the other one unmarked. This is the view of Taglicht (1983) or Berry, quoted in Hasan and Fries (1995b) as saying that “everything up to and including the element Subject is Theme”. For an illustration, let us look at (17). If Subject is always the unmarked Theme, this sentence would have they as unmarked Theme. Since the Adjunct in the kitchen precedes it, then the Adjunct would be the marked Theme, resulting in a clause with two Themes. (17) In the kitchen they would talk for hours.
The issue is long-standing, and there are defenders of both opinions. Halliday (1994) first established that the Theme ends with the first topical (ideational) element. Thompson (1996) and Fries (1995c) also take that view. On the other hand, Matthiessen (1992), Caffarel (2000), Foz (2000) and Martin and Rose (2003) include the Subject as unmarked Theme in cases where there is also a marked Theme. The argument is that both, marked and unmarked Themes, contribute to the development of the discourse. That is not the approach taken in this book. In my opinion, the clause should have only one anchor or point of departure: the first ideational element, the Adjunct in the kitchen in Example (17). The progression of Subjects throughout the text may well reveal other characteristics of a text. However, in this study I am concerned exclusively with Theme as point of departure.8 Another argument in favour of the single ideational Theme is that in Spanish the Subject is often elliptical. In that case, we could either revert to a single ideational Theme, or we could include the Process (since it contains Subject reference) in the Theme portion of the sentence. The Spanish equivalent of (17), provided in (18), would have either en la cocina as Theme, or en la cocina
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.12 (64)
Chapter 4
hablaban, whereas the Theme of the English translation would be in the kitchen, they. The single ideational Theme constraint results in parallel structures for that sentence in both languages, where only in the kitchen/en la cocina are Themes. (18) En la cocina hablaban durante horas. ‘In the kitchen they would talk for hours.’
The discussion so far has concentrated on the many views and definitions of the notion of Theme in general, illustrated with some examples from English. In the next section I provide an account of the possible realizations of Theme, both in English and in Spanish, arguing that Theme is realized through first position in the clause in Spanish as well.
. Thematic realization in English and in Spanish My account of thematic realization is one that proceeds from the semantics to the grammar. I argued for a semantic definition of Theme (the anchor for the ensuing discourse)9 which is realized structurally (first position in the clause). These two sides of the definition are what Berry calls ThemeM and ThemeF . ThemeM is Theme considered as meaning, whereas ThemeF is Theme considered as form (Berry 1996). In my account, ThemeM is the anchor, in any given clause, between what has gone before and the ensuing discourse. ThemeF is the first ideational element in the clause. From a semantic-structural starting point, I now advance towards a grammatical characterization of Theme, that is, a definition of what kind of elements can be in thematic position in the two languages under consideration. Accounts for other languages, such as Steiner and Ramm’s for German (1995), follow the opposite direction: they start with the grammatical realization, which eventually leads to a semantic characterization. The reason for my choice of direction is that I considered it appropriate to start with the same definition for both English and Spanish.10 That constraint is only satisfied when beginning with the semantics of Theme. .. English The status of Theme in English is well documented. Because English has quite a fixed word order, it is straightforward to provide an account of Theme, in terms of first position in the sentence. The summary that follows is taken pri-
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.13 (65)
The thematic structure of dialogue
Table 4.2 Unmarked Theme selection in Mood context. Fragment, from Matthiessen (1995) Theme textual ind.
Rheme interpersonal
declarative interrog.
yes-no wh-
Subject They
cooked curry.
Finite Did
Subject they
cook curry?
Wh
(/Subject, Complement, Adjunct) =
did they cook?
Process Cook
the curry!
Subject You
cook the curry!
What imp.
implicit Subj. explicit Subj.
topical
marily from systemic sources, such as Matthiessen (1992, 1995), Martin (1992), Halliday (1967, 1994) and Downing and Locke (2002). The main source of description is Matthiessen’s 1995 classification of Theme options in English, with a special attention to marked and unmarked contrast. This is mainly Halliday’s (1967, 1994) classification oriented to the building of Theme systems, and it has been implemented in NIGEL (Matthiessen and Bateman 1991), the computational systemic functional grammar component of the PENMAN text generation system – see Mann (1983a, 1983b) and Matthiessen and Bateman (1991). Table 4.2 displays an account of unmarked Theme in mood context, considering textual, interpersonal and ideational Themes.11 This includes only independent clauses and unmarked Themes. Matthiessen (1995) describes Theme as the resource for setting up the local context for the clause. This contextualization can be ideational, interpersonal, or textual. Interpersonal and textual Themes are selected through the ATTITUDE and CONJUNCTION systems. Let us remember that, as mentioned in Section 4.1, Theme can be multiple: a combination of ideational, interpersonal, and textual elements. The following can all be part of a multiple Theme: –
Textual – Continuative: discourse signalers such as yes, no, well, oh, now.
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.14 (66)
Chapter 4
– Structural: conjunctions (coordinating and subordinating), and whrelatives. – Conjunctive: a conjunctive adjunct preceding the topical Theme (that is, for instance, also, moreover, meanwhile, likewise, nevertheless). –
Interpersonal – Vocative – Modal: a modal adjunct preceding the topical Theme (probably, always, in my opinion, evidently, unfortunately, strictly speaking). – Mood-marking: Theme that expresses the Mood of the clause – such as a Finite or a Wh-element.
–
Ideational, or topical Theme: the first ideational element in the clause, typically a Participant, a Circumstance or a Process.
Examples (19) through (23) provide an illustration of different combinations of multiple Themes. These are simplified versions of corpus examples. Textual Theme is italicized, interpersonal Theme is in bold face, and ideational Theme is underlined. (19) I have to get home by five p.m. every day the next two weeks. (20) ...and Monday’s pretty bad for me. (21) Are you free on Wednesday the seventeenth? (22) What days are you free? (23) Oh, unfortunately I have a seminar all day on Friday the twelfth.
Table 4.2 exemplifies only unmarked ideational Themes. The marked/unmarked distinction hinges on the most typical realization of the structure of the clause. Because in English the most typical sentence structure in the indicative mood has a Subject as first element, the Subject is the unmarked Theme in the indicative. There is a strong tendency in English to make Subject and Theme coincide, as Tomlin reports with results from experimental studies (Tomlin 1995). This is true to the point that English allows passive sentences with Indirect Objects as Subjects (Jiménez Juliá 1986: 113). The volume edited by Hasan and Fries also contains a number of papers on the relationship of Theme and Subject (Hasan and Fries 1995a). Sometimes, however, we decide to front elements other than the Subject, thus making them thematic, as in example (24), where a circumstance with the function of Adjunct has been fronted.
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.15 (67)
The thematic structure of dialogue
(24) On the seventeenth I have a seminar from nine to four-thirty.
There is a question about what markedness really means: more complex, less frequent or something else (Dryer 1995). In the corpus, as we will see in Section 4.3, the cases of Circumstances as Themes, traditionally considered marked, are much less frequent. It is possible that Subject = Theme is the unmarked case in general; what we can establish from the English corpus analysis is that such configuration is much more frequent. .. Spanish The progression from a semantic definition to a structural one leads to a characterization of Theme in Spanish as the first element in the clause. Bernárdez (1995: 170) considers that the realization of Theme as first element in the clause is the result of a syntactic encoding of a strategy that could be expressed as “place in the first position the information that is necessary for the global understanding of the message”.12 Harold argues for first position in the sentence as a kind of meta-function that is common to all languages. He calls that metafunction organization, a function that is consistently used across languages, with a realization through first position (Harold 1995). Ocampo also agrees that Spanish follows a Theme-Rheme structure (Ocampo 1995). The first element in the sentence will be determined by word order constraints for Spanish. The literature on word order and constituent structure in Spanish – whether SVO or VSO – is extensive and varied. I will here simply summarize a small subset of it, with especial attention to the elements that are allowed in first position, as related to Mood and markedness. The descriptions are taken from: Bolinger (1955), Contreras (1978), Gutiérrez Ordóñez (1997), Hatcher (1956a, 1956b), Rivero (1980), Suñer (1982), Taboada (1995), Torrego (1984) and Zubizarreta (1998). I will first present a few considerations on Spanish word order, proceeding on to a list of the elements that can occupy thematic position. ... Spanish word order It has been argued that Spanish is an SVO language (Tomlin 1986), but many studies (Bentivoglio 1983), (Bolinger 1954, 1955), (Contreras 1978), (Givón 1983), (Hatcher 1956a, 1956b), (Ocampo 1990) show that there is a strong alternation between SV and VS, especially in certain structures determined by syntactic, semantic or pragmatic factors. Morris (1998) proposes a continuum in Spanish discourse between pragmatic and grammatical modes (Givón
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.16 (68)
Chapter 4
1979: 228), in which there is a tendency towards a pragmatic mode in impromptu speech. Givón, in his study on topic continuity in discourse, compares different languages, among them Spanish, and states that Spanish has a pragmaticallycontrolled flexible word order, but gravitates toward rigidification of SVO. He says that the alternance SV/VS is due to two main factors: topic-continuity pragmatics and presentative VS word order (Givón 1983: 28). In addition, Spanish is a pro-drop language, where the Subject is optionally present when it is recoverable from context and from the morphological markings on the Finite (verb). Therefore, the Subject will be explicit only when it is a Brand-new or Unused element (Prince 1981). In the cases where it is present, the Subject appears in VS ordering in presentational constructions, when a new Subject is introduced, or for reasons of end-weight. I will review each one briefly. Hatcher (1956b) proposes that Subject-Verb inversion occurs when the sentence is concerned with the existence of the Subject. This happens with presentational verbs, which introduce or denote the coming into existence of a Subject. She found nearly three hundred verbs to represent that the Subject is present or exists, is absent, begins, continues, is produced, occurs, appears, or arrives. In a vast majority of these constructions, the word order is VerbSubject. The most important argument in the sentence is the Subject (because it is new), and the Verb merely prepares the ground for the introduction of the Subject. In Suñer’s words, “[t]he verb serves to introduce the subject referent to the scene; it is like the tray on which a delicacy is presented.” (Suñer 1982: 126). For the cases where inversion does not take place, Hatcher offers no study, but hypothesizes that they are related to point of view, as in the alternance Entra la luz and La luz entra (‘The light comes in’). Examples (25) and (26) represent some of the cases Hatcher studied. (25) Porque entre nuestros espíritus existe una afinidad tan grande, because between our souls exists an affinity so big que... that... ‘Because there is such big affinity between our souls, that...’ (26) Pues ha llegado el momento de... so has arrived the moment of... ‘So, the moment of... has arrived.’
Silva-Corvalán (1983) also studied Subject-Verb inversion from the point of view of the contrast between old and new information, newer Subjects being introduced after the Verb. Hetzron (1971) argued that an NP that has just been
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.17 (69)
The thematic structure of dialogue
introduced and that will be part of the discourse is usually moved to a later position in the sentence. This case is different from the presentational cases in that the sentence does not express the coming into existence of the Subject. It simply represents the newness of the Subject.13 (27) Y cuando llegamos a Santa Mónica acá vivían muchos and when came.we to Santa Monica here lived.they many japoneses. japanese ‘And when we came to Santa Monica there lived many Japanese here.’
The third reason for SV inversion is end-weight. Long and complicated elements tend to move to the end of the sentence, in order to avoid burdening the hearer/reader with an undue amount of information from the beginning of the sentence. Bolinger (1955) uses too the notion of “longest element last”, but linking it up with the idea of new datum. The long element does not come last because it is long: it tends to be long for the same reason that it comes last, because it is a new datum, as in Example (28).14 (28) ¿Qué hace en la tienda a estas horas aquel señor de la what does in the shop at these hours that gentleman of the barba larga, los zapatos negros y el traje gris? beard long, the shoes black and the suit grey? ‘What is that gentleman with the long beard, black shoes and grey suit doing in the store at this time of the day?’
The fact that the Subject is long does not suffice to explain why there is inversion in presentative structures, because the inversion takes place with short Subjects, making us refer back to the known-new opposition. This is the case in (29), a typical introduction in radio and television programs, with a pause between Verb and Subject. (29) Dirige Luis Herrero. directs Luis Herrero ‘With Luis Herrero as director.’
In (29), the Subject is quite short, but still introduced later on. We could argue here that framework-setting (Downing 1991) might be the role that the Theme is playing. By providing the action (dirige) first, the speaker allows the hearer to establish a frame of reference. Once the action of directing has been activated, the hearer is able to receive and process the actor role of that action, which is the truly new element.
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.18 (70)
Chapter 4
When the sentence consists of a Subject, Verb and Object (Direct or Indirect) the possibilities are as represented in Example (30).15 Of course, not all the forms have the same probability of occurrence in normal discourse. The first two or three options are the most common. (30) Inés leyó el libro. (SVO) El libro lo leyó Inés. (O Clitic VS) El libro Inés lo leyó. (OS Clitic V) Inés el libro leyó. (SOV) Leyó Inés el libro. (VSO) Leyó el libro Inés. (VOS) ‘Inés read the book.’
We have, so far, only considered full sentences that consist of Subject, Verb and Object. Our task becomes much more complicated when we consider clitics and other Adjuncts. The morphological and grammatical richness of Spanish allows for a more flexible word order.16 I will not discuss all the possible combinations of word order here, but will only provide a listing of thematic elements in Spanish in the next section. ... Thematic elements in Spanish Mood, because of word order, determines the classification of Theme in Spanish. The classification in Table 4.3 is based on Taboada (1995). The distinctions between marked and unmarked constructions do not seem to be so clear-cut as in English. I propose the ones in Table 4.3 as a starting point. Frequency in spontaneous conversation, as we will see in Section 4.3.3, could determine the inclusion of other elements, such as Circumstance, in the unmarked category. In the indicative mood there are four types of elements that can occupy ideational Theme position, in addition to textual and interpersonal Themes. The first element is the Subject, when it is explicit in the sentence, as in Juan hizo paella (‘Juan cooked paella’). However, the Finite is most often the first element in the sentence, when the Subject is recoverable and thus omitted, as in Hizo paella (‘He/she cooked paella’). The Finite could also be the first element in the sentence as a result of a Subject-Verb inversion: Se nos secó la paella (‘The paella got dry (on us)’). In these cases, I analyze the Finite, preceded or not by clitics, as the unmarked Theme of the sentence.17 Clitics cannot be thematic elements by themselves, because they are not independent forms; they attach to the Verb. Vasconcellos (1992) also argues for the status of Finite as Theme in Portuguese, and she considers the clitic, including the impersonal se, a part of the Verb.
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.19 (71)
The thematic structure of dialogue
Table 4.3 Unmarked Theme selection in mood context in Spanish Theme textual interpersonal ind. declarative
interrog.
yes-no qu-
Subject Juan
hizo paella.
Finite Hizo
paella.
Non-subj. Part. A María
le gustó la paella.
No
Clitic(s) + Finite Nos hizo una paella estupenda. nos gustó el postre.
Finite ¿Trajo
Subject María
el vino?
Qu
(/Subject, Complement, Adjunct) =
trajo el vino?
¿Quién imp. implicit Subject
Rheme ideational
(Neg. marker) Finite ¡Pon ¡No pongas
la mesa, por favor! la mesa todavía!
(Neg. marker) Clitic(s) + Finite ¡No la pongas todavía! explicit Subject
Subject ¡Tú
Finite abre
el vino!
The third element that can be unmarked Theme in Spanish is a nonSubject Participant, most commonly an Experiencer in mental processes. Verbs such as gustar, encantar, agradar, all synonyms of like, code the Experiencer as an Indirect Object, with obligatory clitic doubling (Fernández Soriano 1999). In Example (31), the participant translated into English as the Subject resembles an Indirect Object, with the preposition a. This complement is repeated in the clitic le preceding the Verb. The Verb agrees with the Object of the liking process, as in Example (32), where the Verb is in third person plural, in agreement with its surface Subject, aceitunas. The clitic is co-referential with the Experiencer: in (33) it is plural, agreeing with a plural Experiencer. (31) A Oliver le encanta la paella. prep Oliver clitic.3sg love.3sg the paella ‘Oliver loves paella.’
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.20 (72)
Chapter 4
(32) A Maite le gustan las aceitunas. prep Maite clitic.3sg like.3pl the olives ‘Maite likes olives.’ (33) A los dos les gusta la sopa. prep the two clitic.3pl like.3sg the soup ‘They both like soup.’
In all these cases, although the Experiencer looks like a preposed Indirect Object, I consider it unmarked Theme, since no other word order with a full noun phrase Experiencer is less marked.18 It has also been argued that the Experiencer has Subject properties, particularly if the clitic is interpreted as an agreement marker (Suñer 1988; Carlson 2003). The fourth combination in the table displays the presence of clitics before the Verb. In such cases, both the clitic and the Finite are the unmarked Theme, since the clitic depends on the Verb for stress and meaning. In yes-no interrogatives, any of the options listed for declarative clauses is a valid thematic selection. With the addition of appropriate intonation, any declarative clause has the potential of becoming a closed interrogative. In addition, there is Subject-Verb inversion, as in ¿Trajo María el vino? (‘Did María bring the wine?’). For wh-interrogatives (or, more appropriately, qu-interrogatives), the situation is the same as in English: the interrogative word – which is at the same time a Subject, Complement or Adjunct – is the Theme. It is then both an interpersonal element, because it marks Mood, and a topical one, because it is part of the Transitivity structure. Finally, in imperative clauses, I make a distinction between those with an implicit Subject and those with an overt Subject. In the first case, the Finite, preceded or not by clitics, is the unmarked Theme. In the second case, the explicit Subject is not topical Theme, but interpersonal, closer to a vocative, as in ¡Tú abre el vino! (‘You open the wine!’).
. Thematic realization in scheduling dialogues This section explains the analysis of Theme realization in the corpus. For the analysis, I first took the SEOS markers (Semantic End of Segment, see Chapter 3 for a definition) as segmentation signals, considering SDUs (Semantic Dialogue Units, also defined in Chapter 3) as units of analysis. Some of these were minor clauses, therefore showing no Theme-Rheme articulation. Some oth-
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.21 (73)
The thematic structure of dialogue
Table 4.4 Numbers of SDUs and clauses in English and Spanish
SDUs Clauses in the thematic analysis
English
Spanish
784 799
1294 984
ers were, however, composed of more than one clause, usually embedded. In the cases of embedded clauses, I broke the main and subordinate clause into two clauses for Theme analysis. Therefore, I am disregarding Theme in the clause complex (the preposed subordinate clause being the Theme of the main clause), and will only consider Theme within the clause, whether main or embedded. This approach ensures that the units for the thematic analysis are the same as those for the rhetorical analysis (Chapter 5). Table 4.4 shows the number of clauses considered for the thematic analysis of each language, and the total number of clauses (SDUs) in the corpus. The discrepancies are due to the presence of minor clauses, disregarded for the thematic analysis, and to the split of some SDUs into two clauses (embedded and main). Spanish contained a higher number of minor clauses, resulting in a much lower number of units in the thematic analysis. In English, although a few minor clauses were excluded from the analysis, the split of complex clauses yielded more units for the thematic analysis. .. Two examples I provide two examples of Theme realization in context, one for each language. Neither is a full conversation, but they are contiguous turns in a conversation. As in the rest of the chapter, textual Theme is represented in italics, interpersonal Theme in bold face, and ideational Theme is underlined. (34) [FBNT_MJFG_1] fbnt_1_01: /oh/, would you like to meet, /uh/, for a two hour appointment, *pause* on, *pause* Monday? at, five P M, mjfg_1_02: /um/ I, have to get home by, five P M, every day, the next two weeks. | /uh/, and Monday’s pretty bad for me. | /um/, are you free on, Wednesday the seventeenth? | I’m free all day. fbnt_1_03: on the seventeenth, I am free. after four thirty. | so if you would like to meet, somewhere between, /uh/ five o’clock and nine P M, that would be alright with me. mjfg_1_04: no I think it’d be better if, it was scheduled during, normal business hours. | /um/, do you have any time before five? on the seven-
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.22 (74)
Chapter 4
teenth? fbnt_1_05: I’m sorry | on the seventeenth I have a seminar, from nine to four thirty. | so I won’t be available at all. | /uh/ /uh/, | /uh/, I think an evening would be better in fact. mjfg_1_06: let’s see. | /um/ *pause* since evenings would probably be bad for me, /um/, | are you free, the eighteenth? after, noon? ...
In (34) we can see that the textual Themes tend to be continuatives, in the form of filled pauses (oh, um, uh). Other textual Themes are structural elements (and, so, since). There is one conjunctive Adjunct, in fact. There are few interpersonal Themes, most of them required in interrogative clauses (would, are, do). Finally, the ideational Themes belong to two main categories: speaker or hearer, and dates, in a clear indication of the subject matter of the conversation. The Spanish example below (35) shows similar patterns with respect to textual and interpersonal Themes: continuatives (eh, ah, bueno) and structural elements (porque, pero, pues) under the textual elements; and wh-elements in interrogative clauses (cómo, cuándo, qué) as interpersonal Themes. The main difference with the English example is the presence of Processes as ideational Themes (necesitaría, podemos). Other ideational Themes are, as in English, Participants (in the conversation), and Circumstances. (35) [FMBO_MENC_09] fmbo_09_01: hola Ernesto, cómo le va? | /eh/ bueno. lo llamo porque tenemos que hacer una reunión nuevamente, para hablar de los temas que estuvimos hablando la semana pasada, | necesitaría reunirme con usted dos horas. | /eh/ /mm/ cuándo podríamos hacerlo? | *pause* yo estoy un poco ocupada esta semana | porque me voy de viaje. | pero *pause* qué le parece la semana que viene /eh/, <el trece?> el martes trece, a la mañana. menc_09_02: la verdad la verdad, /eh/ el trece no es el mejor día para reunirnos para mí. | me parece que /gl/ sería mejor /ah/ si fuera la semana entrante, | pues para mí sería mucho mejor <el> o el jueves a la tarde o el viernes en cualquier momento. el viernes dieciséis en cualquier momento. fmbo_09_03: el jueves /ehm/ quince estoy ocupada casi todo el día. | podría hacerlo o el lunes doce de abril, /ehm/ por la mañana | <después> desde las nueve, hasta las cuatro de la tarde tengo libre, | y si no el viernes dieciséis /eh/ a la mañana. | qué le parece a usted? menc_09_04: /ah/ pues el viernes a la mañana me viene bien. | /ah/ a las diez dijo usted? | porque a las diez podría hacer sí. <en [cua(lquier)].> o si
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.23 (75)
The thematic structure of dialogue
no en cualquier otra hora. estaré libre. | así que podemos reunirnos. ... TRANSLATION fmbo_09_01: hello Ernesto how are you | uh well I was calling because we need to meet again to talk about the issues we were discussing last week | I would need to meet with you for two hours | uh mm when could we do it | I’m somewhat busy this week because I’m going away | but what do you think about next week uh on Tuesday the thirteenth, in the morning menc_09_02: the truth the truth, uh the thirteenth is not the best day for us to meet for me | I think that it would be better uh if it were the coming week | so for me it would be much better either Thursday in the afternoon or Friday any time Friday the sixteenth any time fmbo_09_03: on Thursday um the fifteenth I am busy almost all day. | I could do it either on Monday April twelfth, um in the morning | from nine, until four in the afternoon I’m free, | and if not on Friday the sixteenth uh in the morning | what do you think? menc_09_04: uh well Friday in the morning is good for me | uh at ten did you say? | because at ten I could make it yes or if not any other time I’ll be free | so we can meet. ...
In the next few sections I expand on the differences observed in these two examples. I first describe the thematic selection patterns (whether single or multiple Themes, and what types), and then which Transitivity element in the clause is chosen as Theme. .. Thematic selection patterns All the clauses considered had at least one ideational Theme. In addition, clauses could have textual Themes, interpersonal Themes, or both. Table 4.5 shows the distribution of each.19 In both languages, there is a slight preference for clauses with multiple Themes (59.82% in English and 64.53% in Spanish). The English results are similar to Bäcklund’s (1992), who reports 57% multiple Themes in telephone conversations in English. The textual Themes are continuatives, conjunctions and conjunctive adverbs. The last two are straightforward, and they are similar to those found in written language. Conjunctions (coordinate and subordinate) introduce clauses in hypotactic and paratactic relations (and, but, or; because, although, if in English and y, pero, o; porque, si, para in Spanish). Conjunctive adverbs are not very frequent in the corpus, the majority being and then in English and en-
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.24 (76)
Chapter 4
Table 4.5 Simple and multiple Themes English
Spanish
T
%
T
%
Textual Themes Interpersonal Themes Ideational Themes
328 244 799
41.05 30.53 100.00
549 227 984
55.79 23.97 100.00
Total, simple Themes Total, multiple Themes
321 478
40.17 59.82
349 635
35.47 64.53
Total number of Themes
799
984
tonces, o sea que in Spanish. The use of conjunctive elements is most common in written language, where they are used to signal connections among parts of the text. Continuatives are the most common textual Themes in the corpus. They indicate that the speaker is holding the floor. (It will be recalled that speakers yield the floor mechanically, see Chapter 3.) Since speakers may sometimes have the floor when they are still not ready to respond to the previous turn, they often use continuative elements at the beginning of a turn as signals that they want to keep the floor. Continuatives are, for instance, well, so, oh, and, you know, yeah, okay in English, and bueno, y, este, a ver, es que, pues in Spanish. These are also used inside a turn, to indicate that the speaker has not finished talking. For instance, in (36), speaker MTJH uses yeah to indicate that he wants to elaborate on his acceptance of the date proposed. After evaluating the proposal (sure, that’s great), he hesitates to elaborate on when exactly the meeting is going to take place. The continuatives here are um, yeah and like I said. (36) [MTJH_MRIL_03] mril_3_04: okay, with this schedule, on the third, I could meet after two o’clock on the third, | *pause* but, that’s about the only time I have free that day ’cause I’m, tied up, from about ten in the morning, to about two in the afternoon, | is that okay by you? mtjh_3_05: sure, that’s great, | /um/ yeah, like I said, I’m out, at noon, | so, do you want to meet from, three to five? | we can, give that a shot then.
Example (37) shows a similar case for Spanish. There is a continuative, bueno ‘well/okay’, at the beginning of the turn, and another one, este (lit. ‘this one’, translated here as ‘well’), in the middle of the turn, which holds the floor while the speaker looks for more dates when she is available.
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.25 (77)
The thematic structure of dialogue
(37) [FCBA_MEBA_08] fcba_08_02: bueno. el ocho yo puedo, *pause* a partir de las dos. de la tarde. | del nueve al once voy a estar de viaje, | no tengo ninguna hora que puedo reunirme. | este, el doce voy a poder, [na(da)] más apenitas de once, a una. ... well. on the eighth I can, from two on. in the afternoon. | from the ninth to the eleventh I’m going to be away, | I don’t have any time that I can meet. | well, on the twelfth I’ll be able to, only from eleven, to one. ...
Other yes and no signals, such as yeah, yes, okay, no, nope and sí, bueno, okay, vale, no in Spanish are interpersonal in nature. These fulfill a different function from the turn holding continuatives. For instance, yeah in Example (38) constitutes an acceptance of the proposal, not a turn-holding device. (38) [FAMS_FCLD_AU] fams_au_11: okay, the tenth sounds great. | /uh/ how ’bout nine o’clock the, morning of the tenth? | /um/ if you have, any problems, just give me a call at my office. fcld_au_12: yeah, I can make it the tenth, at nine o’clock, ...
Interpersonal Themes usually belong to one of three categories: vocatives, modal adjuncts and mood-marking elements. Vocatives are present at the beginning of conversations, as a summons to the interlocutor, similar to the way they appear in the opening of telephone conversations (Schegloff 1968; ten Have 2002). Mood-marking elements are Finites in yes-no questions, as shown in Examples (39) and (40), and wh-elements in open-ended questions, which are illustrated in (41) and (42). Wh-elements conflate interpersonal and ideational in one Theme. (39) [FJBT_MSRH_11] fjbt_11_01: ... do you have anything free in your schedule, /uh/ from, day after tomorrow on? (40) [FKND_FKAS_01] fkas_01_01: ... /uh/ tiene usted /uh/ dos horas libres lunes? /uh/ have you /uh/ two hours free on Monday? (41) [FECE_MNFH_10] fece_10_03: well why don’t we try for another lunch meeting (42) [FFCS_FSNM_01] ffcs_01_01: ... cuándo es mejor para ti? ... when is it better for you?
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.26 (78)
Chapter 4
The most interesting interpersonal Themes are modal adjuncts. They are part of face-saving strategies, serving as hedges when a date is being rejected, or when it is being proposed. Perhaps, how about, I’m sorry, actually, I think that are used in this way in English. In Example (43), both how about and maybe act as interpersonal Themes, softening the potential face-threatening act of asking a direct question (Brown and Levinson 1978). In Spanish, qué tal, qué te parece, perdona, tal vez, quizás all act as interpersonal Themes. Example (44) shows such use of quizás (‘maybe’) (43) [MGCT_MYAW_1] myaw_1_04: so, how about maybe, a Tuesday. (44) [MJBP_MMBU_04] mjbp_04_04: así que quizás podríamos concertarla para el próximo lunes. so maybe we could make it for next Monday.
Within ideational Themes, we can establish a further distinction as to whether the Theme was unmarked or marked (Table 4.6). There seems to be, in both languages, a tendency to avoid marked Themes. We examine below how that relates to Transitivity, since most of the Themes represent Circumstances in the Transitivity structure. However, the percentage of marked ideational Themes is more than double in Spanish as compared to English. (Figure 4.2 provides a graphical representation of these distributions.) Both languages prefer unmarked to marked ideational Themes, but Spanish seems to favour marked Themes more than English. This could be due to more flexibility in Spanish word order. The following two examples are less marked than their exact-order translation into English, although note that in (46) there is a comma marker, English
Spanish
Marked 9%
Unmarked 91%
Marked 21%
Unmarked 79%
Figure 4.2 Distribution of marked and unmarked Themes
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.27 (79)
The thematic structure of dialogue
Table 4.6 Marked and unmarked Themes English Ideational Themes – unmarked Ideational Themes – marked
Spanish
T
%
T
%
729 70
91.24 8.76
778 206
79.06 20.94
a sign of rising intonation, between the fronted Adjunct and the rest of the clause. (45) [MJEG_MFAC_01] mfac_01_03: /ah/ para mí no me será posible el lunes, ... for me it won’t be possible on Monday ... (46) [MJMG_MSNC_04] msnc_04_04: ... así que entre las diez y las dos, no puedo. so between ten and two I can’t
Cummings, in a study of narrative in Old English (Cummings 1995), found that on average, narratives had 59.96% of marked Themes, much higher than non-narrative sections of the same texts, which averaged 42.5%. GómezGonzález (1995) describes a study of spoken texts from the Lancaster Corpus of Spoken English. She mentions that, despite being spoken, the language of these texts is not spontaneous, but planned and formal. Her percentages show 86% unmarked versus 14% marked. Although our methods might not be comparable, it is possible that there is a cline, from purely narrative written texts, through planned spoken language, to casual conversation, in degrees of Theme markedness. Since the conversations studied here are task-oriented, the end of the cline could be filled by casual conversation, with task-oriented dialogues preceding it (Figure 4.3). In the figure, the written language is represented only by narrative, but we could probably find different levels of markedness if we were to compare different written text types. Planned spoken language refers to language spoken aloud, but that has been planned beforehand (Ochs 1979), such as speeches and some forms of televised reports. Task-oriented conversation, although it
Figure 4.3 Cline according to the presence of marked Themes
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.28 (80)
Chapter 4
develops in predictable stages, is not planned. Finally, in casual conversation, there are no task-related constraints, the only constraints being social and communicative. .. Thematic selection and Transitivity I mentioned in the previous section that a good number of the ideational Themes are Circumstances in the Transitivity structure of the sentence. The common syntactic encoding of a Circumstance is by means of an Adjunct. Adjuncts are marked when in thematic position. However, the breakdown of marked and unmarked Themes in Table 4.6 shows very small numbers of marked Themes. This is realized through the encoding of Circumstance Themes as Subjects in a Carrier-Attribute relationship, thus resulting in a grammatical metaphor (Downing and Locke 2002: 147–152), (Halliday 1994: Ch. 5). One example will illustrate the realization patterns. (47) [FLCM_FKAS_08] fcka_10_03: well, /uh/ you’re available, after, /um/ the seventeenth? in the week, <seventeenth,> /eh/ seventeenth, the eighteenth, the nineteenth, the twentieth, the twenty first? | well, the seventeenth and the twentieth would be the most convenient for me | because I have, nothing planned for these two days. | so, what do you think. {crosstalk}
The speaker avoids making the seventeenth and the twentieth a marked Theme by encoding it as the Subject of the next clause. The Theme is picked up from the Rheme of the previous clause (ideational Themes are underlined), and incorporated into the Subject position. She uses the same strategy to make I the ideational Theme in the third clause, by picking it up from the rhematic position in the previous clause. Table 4.7 displays the presence of Themes as related to Transitivity. Here, the label Participant refers to the two interlocutors in the conversation.20 A Participant Subject will be typically an unmarked Theme (I am free). A Participant Adjunct is a marked Theme (For me that’s no good). Circumstances will occur in the same conditions: a Circumstance Subject is unmarked; it reflects the realization of a date or a time as the Subject of a relational process, attributive or circumstantial (Tuesday is good for me). A Circumstance Adjunct as Theme will be marked (On Tuesdays I’m busy all day). Finally, a Process is a Finite that occurs in first position in the sentence because the Subject has been omitted (Sounds good to me). The percentages are with respect to the total number of Themes in the corpus.
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.29 (81)
The thematic structure of dialogue
Table 4.7 Theme Realization and Transitivity. Percentages with respect to N English
Spanish
T
%
T
%
Participant Subject Participant Adjunct Circumstance Subject Circumstance Adjunct Process
356 3 220 73 33
44.55 0.37 27.53 9.14 4.13
85 19 79 191 523
8.64 1.93 8.03 19.41 53.15
n, Themes in this table N, Ideational Themes
685 799
897 984
The rest of the Themes not reflected in Table 4.7 are other Themes, such as wh-words that have a function in the Transitivity structure, but that cannot be considered marked, because they appear in first place in the unmarked cases (48). (48) [FJMB_FMLZ_6] fjmb_6_05: ... so, we could do something on the morning of the sixth, | how would that be for you.
The results are easier to compare in graph format, which can be found in Figure 4.4. There are a few striking differences between the two languages. First of all, English tends to make Participant Subject the Theme in a high number of cases, followed by Circumstance Subject. That is, the preferred Theme in English is, overall, the grammatical Subject of the sentence. In Spanish, on the other hand, the preferred Theme is a Process, followed by a Circumstance Adjunct. The next two preferences are Subjects, either Participant or Circumstance, both within the same range. Neither language tends to place Participant Adjuncts as Themes, although Spanish does so slightly more often. Processes (49) are, as we saw in Table 4.7, a common thematic realization even in English. When the Subject is elliptical in English, it is usually analyzed as an elliptical Theme (Halliday 1994). For instance, the Subject is elliptical in the second part of the coordinated sentence He came home and _ collapsed on the couch. Under this analysis, the first clause in (49) would have no explicit ideational Theme. I propose to analyze could as Theme, because it is what the speaker chose to start the sentence with. It is an instance of contextual, not textual ellipsis. Nariyama (2004) finds that subject ellipsis is a common phenomenon in conversation and casual letters in English.
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.30 (82)
Chapter 4 600 500 400 English Spanish
300 200 100 0 Part. Subj.
Part. Adj.
Circ. Subj.
Circ. Adj.
Process
Figure 4.4 Theme realization and Transitivity (as in Table 4.7)
(49) [FCAD_FJAB_12] fjab_12_08: couldn’t be better. | and then you know maybe if we’re, over a little bit ...
Fries mentions the preference, in English conversation, towards making humans the topic of the conversation (Fries 1995c). Berry also reports that speakers, in casual conversation, avoid grammatical Subjects that are not human (Berry 1987). This corpus confirms this tendency: 44.92% of the English Themes are Participants in the conversation.21 The fact that a good number are also Circumstances reflects, most likely, the genre of the corpus: the presentation and discussion of dates, times and places for a meeting. However, the tendencies are very different in Spanish: only 10.57% of the Themes are participants, and a total of 27.43% are Circumstances. It is clear that, in Spanish, it is acceptable to place Circumstances rather than Participants as Themes. Despite the validity of what I have just discussed, let me note that there is another factor at play. I have mentioned percentages, but if we look again at Table 4.7, we can see that we are not discussing the same rough numbers: in English, a total of 652 Themes were either Participants or Circumstances. In Spanish, only 374 fall into those categories, even though we started out with a higher number of ideational Themes in Spanish (see Table 4.6). The reason is that Spanish uses a much higher number of Finites as Themes. In fact, Processes account for more than half of the ideational Themes in Spanish.
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.31 (83)
The thematic structure of dialogue
And Processes contain, to a certain extent, Participant reference. We will see, in Section 4.5, how that also affects the thematic progression. Cummings notes that Finites are often thematic in Old English (Cummings 1995). Those Finites were either lexical verbs or verbs that would become auxiliaries in later stages of the language. He raises the question that in Halliday’s model (Halliday 1994), Finites can only be interpersonal Themes in interrogatives. He considered them marked Themes, and expresses uncertainty about whether they can be considered true topical Themes. I have, in my analysis, considered them unmarked ideational Themes, because they are the first element in the clause, often due to Subject omission. This is true of the numerous cases of Subject omission in Spanish, but also of the few found in English. I also believe that Finites as Theme are ideational, since they have a role in the Transitivity structure, as Processes.22 In Spanish, the case of Process being thematic is much more common, but we find the same types of occurrences, as in (50) and (51). (50) [MRBZ_MCRA_03] mrbz_03_03: /gl/ el martes, dos de febrero, acuérdate que es mi cumpleaños. | y aparte estoy fuera. | estoy de vacaciones. ... on Tuesday, February second, remember that it’s my birthday. | and besides I’m away. | I’m on vacation. ... (51) [MEJH_MJJG_01] mjjg_01_05: /eh/ de qué viernes estás hablando. ... uh which Friday are you talking about. ...
Let us explore a bit further the comparison of the two languages in terms of thematic selection: how Participants and Circumstances become marked or unmarked Themes. The numbers in Table 4.7 reveal strong tendencies that diverge in both languages. Whereas English tends to have, in 44.55% of the cases, a Subject as Theme, Spanish does that only in 8.64% of the time. This is due to the optionality of the Spanish Subject, which often makes the Process a Theme in the unmarked case. At the same time, Spanish uses a Participant Adjunct as Theme more often than English (52), sometimes with clitic doubling, as in (53). (52) [FMBO_MENC_09] menc_09_02: ... pues para mí sería mucho mejor <el> o el jueves a la tarde o el viernes en cualquier momento. | el viernes dieciséis en cualquier momento.
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.32 (84)
Chapter 4
... well for me it would be much better either on Thursday in the afternoon or Friday any time. | on Friday the sixteenth any time. (53) [MARC_MPHB_02] mphb_02_02: ... a mí me parece también, bueno de hacer una reunión, | pero estoy muy complicado... ... to me it also sounds, good to have a meeting, | but I’m very busy...
The reversal of the tendency can be seen in the numbers of Circumstances as Subjects and Adjuncts: both English and Spanish incorporate Circumstances as Subjects in the Transitivity structure. These are mainly the names of days or dates under consideration, as we can see in Examples (54) and (55). (54) [FRJP_MPMM_1] mpmm_1_06: /uh/ the sixteenth of March sounds good, ... (55) [FYMM_FKAS_02] fymm_02_02: *pause* el lunes veintinueve, en la mañana, es imposible para mí. ... Monday the twenty ninth, in the morning, is impossible for me ...
However, English uses this device much more often (27.53% of the time versus 8.02% in Spanish). On the other hand, Spanish prefers to code Circumstances as Adjuncts. We can see, then, that the general tendency is for English to prefer Participant Subjects as unmarked Themes. Spanish uses Circumstances Adjuncts as its most frequent unmarked Theme, after Process.
. Definition and articulation of thematic progression This part of the study deals with the development of Theme throughout the conversation. In the previous sections, we investigated what kind of elements the speakers choose as points of departure. Here, we are concerned with how those points of departure are picked up, elaborated on or abandoned in the following discourse. The connections of thematic elements in a text are what constitute its thematic progression (TP). TP is not to be confused with method of development (Fries 1983, 1995a). The method of development in a text is often correlated with its genre (see Section 4.6), and it tells us what the Themes of choice are, and how those change in the course of the text. The types of Themes will correlate with different stages in a text that belongs to a particular genre (Fries 1995c: 319).
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.33 (85)
The thematic structure of dialogue
The method of development of a text usually predicts the most typical Themes to be chosen according to its genre. Francis (1989) proposes to consider typicality from two angles. 1. Transitivity, the kinds of Participants and Processes selected as Theme (Halliday 1994: Ch. 5). 2. Lexical selection, whether the Themes are typically people, institutions, concrete or abstract entities, etc. Francis concludes that the most typical Themes in news articles are material and verbal processes, together with Sayers in verbal processes. This reflects the fact that news articles inform the reader about events and what people say about events. There is not a great deal of intervention by the writer; the relationships between clauses encode the progress of events in the world. On the other hand, editorials and letters, two other genres present in a newspaper, thematize relational processes, resulting in more nominalization. Nominalization in this case represents the writer’s purpose: it makes the events static, so that they can be evaluated. The texts are not any more about what has happened, but about the relationship of the events to some other events or the state of the world. Nwogu and Bloor (1991) also point to different Transitivity patterns as differences in genres, in their case between professional and popular medical texts. We saw, in Section 4.3.3, some features of the Transitivity structure of these dialogues. Dates and times are placed in subject position in CarrierAttribute relationships. This allows the speakers to evaluate those dates and give them attributes as to their suitability for a meeting. Thematic progression, on the other hand, considers how Themes evolve throughout the discourse, regardless of their realization in the Transitivity structure of the clause. That is, we consider the Participant or Circumstance that is realized as Theme in the clause, and examine how the speakers have linked it to the previous discourse. They might pick it up from the Theme of the previous clause, from its Rheme, or from either Theme or Rheme in some other clause in the discourse. In Example (56) we see how the Theme of FCRE’s utterance, any time after noon on Tuesday, is picked up by the other speaker and made the Theme of her utterance. (56) [FCRE_FJSL_1] fcre_1_03: ... I couldn’t do it until Tuesday, | any time after noon on Tuesday would be good. fjsl_1_04: ... after noon on Tuesday the sixteenth would be great. | what do you say, one o’clock?
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.34 (86)
Chapter 4
Francis (1989) summarizes the differences between method of development and thematic progression. Method of development refers to typical Themes. Thematic progression, on the other hand, is concerned with relatedness: where Themes come from, and how they relate to previous Themes and Rhemes. The first formalization of thematic progression was provided by Daneš (1974a). Daneš uses a somewhat more elegant (although more difficult to apply) theory of development in the clause. It is more elegant than bipartite accounts, such as the ones that postulate a strict Theme-Rheme opposition, in that the transition from thematic to rhematic elements does not happen at a boundary point. Communicative Dynamism (CD) perceives the clause/sentence as a continuum, a progression from given to new, or from thematic to rhematic elements (Firbas 1992). The communicative dynamism of an element in the sentence is its contribution to the development of the communication. New information has more CD, since it contributes more to the communication, but there are different degrees of newness. The theme, in the Prague School, is the element that contains the lowest degree of CD. In general, this element will be the first in the sentence, but not necessarily so. Despite the difference in definition for the concept Theme, I will use Daneš’ thematic progression, slightly modified, as we will see later. Daneš considers thematic progression to be one of the representations of connexity in a text: one of the ways in which a text displays coherence. A preliminary text analysis, mainly of scientific and professional texts, yielded the types of thematic progression represented in Figure 4.5. Each line represents an utterance, Un . The Ts and Rs refer to the Theme and Rheme of that utterance. Thus, in simple linear TP, the Theme of U2 is the same as the Rheme of U1 , and the same progression is repeated in U3 . This can be summarized in the formula: (57) T (Un ) = R (Un–1 )
Example (58) is a representation of simple linear progression across turns. The Rheme of FECE’s utterance, your office, is picked up as the Theme of the other speaker’s utterance. (58) [FECE_MNFH_10] fece_10_05: okay. | where’s your office. mnfh_10_06: /um/ my office is in Wean Hall, four six one six.
The types considered by Daneš are all simple, because the Themes derive from a single source, be it the Theme or the Rheme of the previous sentence. They are also contiguous because there are no other intervening types between them.
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.35 (87)
The thematic structure of dialogue
a. Simple linear TP (Simple contiguous linear)
b. TP with a constant theme (Simple contiguous constant) [T]
T1
R1 T1
R2
T1
R3
c. TP with derived themes
Figure 4.5 TP patterns according to Daneš
Thus, we can rename the first type of Theme discussed here as simple contiguous linear. The second type of TP that Daneš found is the TP with a continuous or constant Theme. The same Theme, or variations of it, is repeated throughout the text, appending new Rhemes to it. This is a simple contiguous constant Theme. A formula for this type is to be found in (59). Example (60) represents
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.36 (88)
Chapter 4
a short fragment of such progression. In this turn, the speaker uses repeatedly I as the Theme of her clauses (only ideational Themes are underlined). (59) T (Un ) = T (Un–1 ) (60) [FCRE_FJSL_1] fcre_1_03: Jennifer? I will be out of town until Thursday, | I am free on Friday the twelfth from eleven to one, | or, I’d have to go, into the next week. | /um/ /glottal/ I couldn’t do it until Tuesday, ...
Finally, the TP with derived Themes represents the case where the Theme of each utterance is derived from a hypertheme, the Theme of the whole text or paragraph. The Themes are all related to this hypertheme – not just among themselves, as is the case in the continuous Theme. (61) captures this type of progression. The symbol refers to the fact that the Theme of each utterance is related to the hypertheme, but not identical to it. (61) T (Un ) [ T ]
The issue of what constitutes a hypertheme is not at all clear. It could be a general topic, a title, or some higher-level concept that is present throughout the text. I had difficulty finding instances of hyperthemes in the conversations in the corpus, unless one considers “desire to arrange a meeting” a valid hypertheme. Another possibility is to consider a range of dates mentioned earlier in the conversation as the hypertheme that is developed in the rest of the conversation. This is the case in Example (62), where the month of August could be considered the hypertheme, with derived Themes on the seventh, eighth and the tenth and anytime after nine o’clock in the morning. The example is, however, somewhat constructed in that I deleted part of FACR’s utterance, which would have likely resulted in a linear progression. I show it here simply for illustration purposes. (62) [FACR_FJYK_AU] facr_au_01: in the month of August <would> /uh/ would you like to meet ...? fjyk_au_02: okay. | on the seventh, eighth, and the tenth. I can meet you, in the morning. | anytime after nine, o’clock in the morning would be good for me.
These patterns can be combined in different ways, and with variations such as the split Rheme, where a Rheme is divided in two or more different elements, each of which becomes a Theme in succession. Figure 4.6 represents this case, which we can also see in an example from the corpus, (63). This example was
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.37 (89)
The thematic structure of dialogue
Figure 4.6 Split Rheme pattern
modified in order to show the pattern in a straightforward way. The Rheme of the first turn, Thursday afternoon or anytime Friday, is divided in the next turn by speaker MTNZ. He uses on Thursday as the Theme of his first clause, and then Friday in the afternoon for the second clause. (63) [MAGH_MTNZ_11] magh_11_03: no, | /um/ that sounds like this week, doesn’t work. | next week, the earliest I could do it, would be again, Thursday afternoon, or, anytime Friday. mtnz_11_04: /sniff/ okay. | yeah on Thursday I can’t do it on, | but Friday, in the afternoon would be good, | how about /uh/, one to three?
Dubois (1987) performed an extensive analysis of a corpus of biomedical slide talks, and updated the typology to include a few new types, which reflect the fact that some progressions are gapped, that is, the Theme is picked up from an utterance that is not immediately preceding. Thus, the typology would look as follows, the first three types being the original ones from Daneš, and the last three, gapped variations of those. A gap takes place where there is intervening material between the Theme and its link to the previous discourse. – – – – – –
Simple contiguous linear Simple contiguous constant Simple derived from hypertheme Simple gapped linear Simple gapped constant Simple gapped derived from hypertheme
In addition, there are multiple realizations of all the above, where the origin of the Theme under consideration can be traced back to different links in the text, and through different patterns. A subtype of multiple is the integration of two
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.38 (90)
Chapter 4
Themes. Integration has an inverse in separation, which is equivalent to Daneš’ split. Dubois also points out that any Theme that is not new must be derived from previous Theme(s) or Rheme(s). In such a case, she proposes the labels themic and rhemic. A closer examination of Daneš’ types would reveal that the linear pattern is always rhemic, because it derives from previous Rhemes (see Figure 4.5 again). For the same reasons, the continuous is themic. She proposes to substitute those terms. I will, however, keep Daneš’ original terms, reminding the reader that they are obviously themic and rhemic, respectively. Dubois also considers the hypertheme to be underdefined. She quotes Daneš as saying that hyperthemes are section headlines of journal articles, and other elements that are given in the text. This givenness can be direct or indirect, derived via hyponymy, hyperonomy, association, and other semantic relations. She proposes to treat the hypertheme as a case of constant: a Theme that is derived from, or in some semantic relationship to, other Themes. It is certainly difficult to decide what a derived Theme could be. Anything that refers to the context of situation – i.e., a new Theme that has not been mentioned earlier in the text – could be a candidate for a derived Theme. Except for the clear cases of titles or headlines, most extra-textual derived Themes will be determined by the co(n)text. In the corpus I studied, there is no such thing as a title or headline, so I decided that all Themes the speakers had not mentioned earlier were new. Some of those are “newer” than others, in the sense that they refer to objects or concepts present in the speakers’ context. However, since my aim is not to establish a hierarchy of givenness or salience outside the text proper, I disregarded the category of progression derived from a hypertheme. The concept of derivation, however, I find useful: it would tell us when a Theme is repeated exactly as it was formulated earlier, or whether it is altered through some semantic process. Therefore, to the linear/constant and contiguous/gapped distinctions, I would like to add a further opposition, derived/not derived. In Figure 4.7 those are represented in the first four types, with subclasses for whether they are derived or not. We turn now to multiple Themes. Dubois points out that these can be the result of integration or separation. We can easily see that those could also be either contiguous or gapped. The integration could happen immediately after the mentioning of the elements to be integrated, or there could be intervening material. The same applies for separation. In Figure 4.7, both integration and separation are represented as being contiguous. In Example (64), an invented example, the two Themes Monday and Tuesday are integrated into a single Theme in the utterance immediately following (ideational Theme is un-
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.39 (91)
The thematic structure of dialogue
derlined). Note that this is also a derived Theme, since Monday and Tuesday are rephrased as those two days. (64) A: Monday is a bad day for me. Tuesday would be better. B: Those two days are already filled up for me.
In addition, both integration and separation can be themic or rhemic. In Example (64), the integration combines two previous Themes, making it themic. In Example (65), another invented example, the separation has as an origin the two previous Rhemes (this type is the one represented in Figure 4.7). This separation is not derived, since the dates are repeated verbatim. (65) A: I’m available anytime Thursday or Friday. B: Well, Thursday is no good. But on Friday I could see you after lunch.
Finally, integration could also be complex, the sum of previous Themes and Rhemes. The following example shows the Theme the eighteenth conjoined with the Rheme on the nineteenth, to form the Theme of B’s utterance. This would be an example of complex gapped integration. The gap occurs because there is another utterance before the integration, What do you think? All multiple are derived in some sense, because they do not repeat a previous Theme or Rheme exactly, but elaborate on them. (66) A: The eighteenth I’m booked all day. I could try to see you on the nineteenth. What do you think? B: The eighteenth and the nineteenth are completely out for me.
Many other alternatives and applications of thematic progression have been suggested. Cloran, for instance, proposes to examine Rheme more closely, and establishes a new pattern, which progresses from Rheme to Rheme (Cloran 1995). Time and space preclude a full treatment of Rheme. For this study, I decided to concentrate on Theme and thematic progression. Likewise, the analysis of thematic progression is restricted to the provenance of each thematic element. That is, for each Theme, I examine whether it is related to some other element in the discourse and, if so, I probe the relationship established – whether a progression from another Theme, from another Rheme, from a distant or an immediate one, etc. A different type of analysis could look at how those relations create thematic chains throughout the text. Such an enterprise is to be left as the subject of a different study.
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.40 (92)
Chapter 4
T1
R1
T1
R1
... R2
T2 (=R1)
T5 (=R1)
R5
... T3 (=R2)
R3
T8 (=R5)
1. Simple contiguous linear a. derived b. not derived
2. Simple gapped linear a. derived b. not derived
T1
T1
R1
R1
... T2 (=T1)
R2
T5 (=T1)
R5
... T3 (=T2)
R3
T8 (=T5)
R8
3. Simple contiguous constant a. derived b. not derived
4. Simple gapped constant a. derived b. not derived
T1
R1
T1
T2
R2
T2 (=R1')
R2
T3 (=R1'')
R3
T3(=T1 + T2)
R3
5. Multiple: Integration a. themic x. contiguous b. rhemic y. gapped c. complex
R1(=R1' + R1'')
6. Multiple: Separation a. themic x. contiguous b. rhemic y. gapped c. complex
Figure 4.7 Thematic progression patterns
R8
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.41 (93)
The thematic structure of dialogue
. Thematic progression in scheduling dialogues With a few exceptions, studies of thematic progression have been applied to written texts. In such analyses, we view the text as product,23 , examining how the Themes are related to other parts of the text. The same principles apply to the analysis of thematic progression in conversation. I am interested in finding out how speakers relate the anchor of their utterances to what their interlocutors, or themselves, have previously mentioned in the course of their interaction. Maynard conceives of thematic progression as happening across speakers’ turns. She presents a study of face-to-face interactions, where she studies the thematic progression patterns proposed by Daneš (Maynard 1986). She proves that the Functional Sentence Perspective is applicable to oral texts, and observes a continuum from interactional-based (oral) to content-based (written) texts. However, she identifies Theme more with topic, and studies thematic chunks, pieces of the conversation that refer to the same Theme: “It consists of minimally two turns; one representing thematic elements and the other providing rhematic information. This minimum chunk is the smallest unit of interaction and constitutes a minimal conversation” (Maynard 1986: 83). A similar approach is followed by Giora (Giora 1983), who studies linear progression of Themes as a segment-builder in novels. As we have seen in the previous sections, the formulation of Theme used in this study is much narrower, and does not include Theme as subject matter. The analysis of the corpus proceeded as follows: I considered only ideational Themes, whether marked or unmarked, and, for each one of them, examined whether it was related to other elements in the previous stretch of dialogue. If the Theme is not related to any linguistic material preceding it, then it is new.24 As we saw in the discussion of Themes derived from a hypertheme, those new Themes are usually related to the context of situation. However, since I will not consider extralinguistic material, the category “new” will include many different types. If the Theme under consideration is related, either through total identification or through some semantic transformation, to an element mentioned earlier in the conversation, then we can classify it according to the categories displayed in Figure 4.7. The other type not represented in the figure, apart from New, is the Syntactic type. This refers to empty, or dummy pronouns, such as it or there, when they act as the grammatical Subject, but have no referent in the text or in the real world. The results of this analysis can be found in Table 4.8. I have included a breakdown of each of the first four categories into derived or not derived
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.42 (94)
Chapter 4
Table 4.8 Thematic progression in the corpus English
Spanish
T
%
T
%
1. Simple contiguous linear a. derived b. not derived
97 69 28
12.14 8.64 3.50
34 14 20
3.46 1.42 2.03
2. Simple gapped linear a. derived b. not derived
44 22 22
5.51 2.75 2.75
59 17 42
6.00 1.73 4.27
3. Simple contiguous constant a. derived b. not derived
128 43 85
16.02 5.38 10.64
53 30 23
5.39 3.05 2.34
4. Simple gapped constant a. derived b. not derived
208 49 159
26.03 6.13 19.90
112 57 55
11.38 5.79 5.59
9
1.13
4
0.41
5. Multiple: integration 6. Multiple: separation 7. New 8. Syntactic n (Ideational Themes)
10
1.25
22
2.24
287
35.92
700
71.14
16
2.00
0
799
–
984
Themes. Thus, row 1, “simple contiguous linear”, is the addition of the two following rows, “a. derived” and “b. not derived”. These same results are more clearly comparable in Figure 4.8, where the horizontal numbers refer to the numbers on the table. In the figure I omitted the derived/not derived distinction. Figure 4.8 shows one striking difference between languages, namely the occurrence of New Themes. I will discuss that later on. Now I would like to concentrate on the other categories. The first category, Simple contiguous linear, is more numerous in English than in Spanish (12.14% in English, 3.46% in Spanish). One example of this is in (67), where the underlined Theme, Thursday the sixth, picks up the immediately preceding Rheme, sometime between the third and the fourteenth of May. It is derived, because it mentions one particular date in the range provided. (67) [FECE_MNFH_10] fece_10_01: so, let’s see, now we’re trying to meet? /um/, sometime between the third and the fourteenth of May. | /ah/ Thursday the sixth looks pretty good, and, so does Monday the tenth. | how ’bout for you.
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.43 (95)
The thematic structure of dialogue 700 600 500 400
English Spanish
300 200 100 0 1
2
3
4
5
6
7
8
Figure 4.8 Thematic progression patterns (as in Table 4.8)
Progression from a contiguous Rheme does not always take place within the same turn. In a Spanish example, (68), speaker FKAS selects the Rheme from the previous speaker, making it the first Theme of her utterance. (68) [FJGC_FKAS_01] fjgc_04_02: ... qué tal el miércoles veintiséis <de> a las diez de la mañana. | *pause* de diez a doce. *pause* fkas_04_03: de diez a doce, no puedo. | tengo una reunión a las diez y media ... ... how about Wednesday the twenty-sixth at ten a.m. | from ten to twelve. from ten to twelve I can’t. | I have a meeting until ten thirty ...
The trend for simple gapped linear is similar for both languages, with a total of 5.51% in English and 6% in Spanish. In these, the Theme of an utterance is in a previous Rheme, but not the immediately preceding one. This type may too be derived, through some semantic transformation of the material. One example will suffice to illustrate this pattern. In (69), speaker FKCF mentions in the next two weeks in the Rheme of her first utterance. After a turn by speaker MJCB, in which he refines the times when he can meet, speaker FKCF selects her previous Rheme, modified, to include only next week.
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.44 (96)
Chapter 4
Table 4.9 Linear versus constant patterns
Linear Constant
English
Spanish
17.65% 42.05%
9.46% 16.77%
(69) [FKCF_MJCB_1] fkcf_1_01: can I meet with you, some time in the next two weeks. | what days are good for you. mjcb_1_02: /um/ well I have some, free time on, *pause* almost every day. /um/ except for Fridays, | Fridays are bad. | /um/ so any day besides Friday we can probably work out a time. fkcf_1_03: well next week I’m out of town, Tuesday, Wednesday, and Thursday, | so, *pause* perhaps, Monday afternoon? *pause*
In general, speakers tend to use constant (themic) more often than linear (rhemic) thematic progression patterns. Table 4.9 summarizes the percentages of those for both languages. The percentage of constant Themes is higher in English due, as we shall see, to the repetition of Subject pronouns. As for constant patterns, gapped patterns are more frequent in both languages than contiguous ones. In both cases, the higher English percentages are a result of the repetition of Subject pronouns throughout the conversation, hence the higher number of non-derived types, as in Example (70). Speaker FRJP uses I in her first turn, which she repeats again when she holds the floor.25 Subject pronouns are not, however the only category in this pattern. In (71), speaker FSNM uses as the Theme of her first utterance a Theme derived from speaker FMNA’s Theme. (70) [FRJP_MPMM_1] frjp_1_01: /ah/ Peter. this is Renee. | I need, to schedule a two hour meeting with you, between March, eighth, and March nineteenth? mpmm_1_02: okay, I’m completely free, on the seventeenth of March. | so, let’s try that date, first. frjp_1_03: /um/ that day, I have a seminar, from nine to four thirty, | and I’m free after that, but not until, then. (71) [FMNA_FSNM_06] fmna_06_01: ... desde las doce hasta las cinco, cualquier horario sería posible para mí. | qué tal para vos? fsnm_06_02: bueno. | tengo toda la mañana libre, | pero entre la una y las cuatro tengo una reunión. | o sea que las horas que me dijiste no son buenas para mí. ...
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.45 (97)
The thematic structure of dialogue
... from twelve to five, any time would be possible for me. | how about you? well. | I have all morning free, | but between one and four I have a meeting. | so the times you told me are no good for me. ...
The differences in the first four types – i.e., contiguous or gapped, linear or constant – can also be categorized in terms of whether they are derived or not. Figure 4.9 summarizes the differences between those two for both languages.26 As we can see from the figure, both languages show similar percentages of derived and non-derived Themes. The non-derived Themes are slightly higher in English because the English speakers tend to repeat Subject pronouns. In Spanish Subject pronouns are frequently dropped. Multiple is divided into integration and separation of Themes. Those two types contribute a small percentage to the total distribution, and they are simplified in Table 4.8. Each type can be divided into contiguous or gapped, themic or rhemic, plus complex for the integration. I broke the results down in Table 4.10, which contains absolute counts. English
Spanish Derived 38%
Derived 46%
Not derived 62%
Not derived 54%
Figure 4.9 Percentages of derived versus not derived Themes Table 4.10 Subclassification of the multiple type English Contiguous Gapped
Spanish Contiguous Gapped
5. Multiple: integration a. themic b. rhemic c. complex
5 2 0 3
4 3 0 1
1 0 0 1
3 1 1 1
6. Multiple: separation a. themic b. rhemic
5 5 0
5 2 3
8 5 3
13 9 4
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.46 (98)
Chapter 4
Both integration and separation occur in two instances: participants and dates. In the case of participants, often the speakers mention I and you and, at some point in the conversation, integrate them into we, and vice versa: we is separated into one or two of the participants. With dates, the situation is very similar. In (72), speaker FCBA mentions a time period, from the fifteenth to the nineteenth. Speaker MEBA then breaks it down into each day of that period, starting a new clause always with a new date. This is an example of gapped separation. (72) [MEBA_FCBA_08] fcba_08_02: ... este qué tal para ti, *pause* del quince al diecinueve. | porque se me hace que el ocho al doce <está muy> no está muy bien para ninguno de los dos. *pause* meba_08_03: *pause* bueno. | el quince puedo cualquier tiempo en la mañana. | *pause* el dieciséis puedo cualquier tiempo en la tarde. | *pause* el diecisiete estoy libre todo el día, | *pause* el dieciocho puedo en la tarde, | *pause* y el diecinueve no tengo tiempo | cómo la ves? ... so how is it for you, from the fifteenth to the nineteenth. | because it seems to me that the eighth to the twelfth it’s not very good for either of us. well. | on the fifteenth I can any time in the morning. | on the sixteenth I can any time in the afternoon. | on the seventeenth I’m free all day, | on the eighteenth I can in the afternoon, | and on the nineteenth I don’t have time | what do you think?
The category ‘new’ is applied every time the speakers introduce a Theme in the conversation that has not been mentioned earlier – although some of those are part of the context of the situation. The percentages of this type are significantly different in the two languages. Whereas English uses 35.92% new Themes, Spanish resorts to it 71.14% of the time, about twice as much. The main reason for this effect lies in the presence of Finites as Themes. English, for the most part, uses Subjects as Themes; Spanish uses Finite verbs, as we can compare in the invented example in (73). English uses the personal pronoun I, but Spanish, in a direct translation of the same sentence, uses the verb estoy. (73) English: I’m busy on Thursday. Spanish: Estoy ocupada el jueves.
In English, except for the first time the pronoun is used in the conversation, the pronoun will always have a link to some other element. In Spanish the Finite verb is linked and contributes to the progression because it encodes person
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.47 (99)
The thematic structure of dialogue
and number, but it does not explicitly repeat the same material. The difference between English and Spanish in this respect poses a problem for the analysis as it has been carried out. Is it valid to argue that, since English uses a pronoun Subject and Spanish does not, we have different types of thematic progression? The crux of the problem is that English decomposes Subject and Finite into two words, whereas Spanish conflates them in one, the Finite plus its person markings. Perhaps an accurate analysis of Theme in Spanish should involve the breaking down of those two. Rose (2001) discusses languages that frequently thematize the Process, with affixes to indicate the participants. He proposes that those languages present the Participants (affixed to the Verb) as backgrounded points of departure. The Participants are implicitly the point of departure for each message, and they are only mentioned when they change or require foregrounding. I decided against an analysis that includes two types of Themes (Process and Participant reference) for two reasons. First of all, I have been working at the word level in all other cases, so resorting to the morpheme would make this analysis different from the other types of analyses (of rhetorical relations and cohesion). The second reason is that, were we to break the Finite into constituent morphemes, the thematic element would be the verb stem, as the first element. The person, number and tense marking is the second element in the Finite. If I adhere strictly to the equation Theme = first element, the results would not change. The number of new Themes would be exactly the same, because we would consider the verb stem as thematic. Another possibility is to break down the new category into Themes that are completely new, because they have just been introduced, and Themes that contain reference to Participants in the discourse context. New Themes could be sorted into the categories proposed by Prince (1981). Finally, the Syntactic type is only present in English in examples such as (74). There are no instances of this type in Spanish, because Spanish does not require an empty (‘dummy’) pronoun in Subject position. (74) [FSJB_FKDO_1] fsjb_1_03: well fortunately or unfortunately I have to go to a family <member> member’s wedding, /uh/ in the middle of the week. | and I’m not free at all Friday the twelfth. | so it looks like we’re into the week of the fifteenth,
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.48 (100)
Chapter 4
. Thematic progression and genre Researchers have tried to establish a link between both a text’s method of development and its thematic progression and its genre. Such work is reported, for instance, in Francis (1989), the two volumes edited by Ghadessy (1988, 1995), or the work of Fries (1983, 1994, 1995a, 1995b). More recent empirical studies (Lavid 1999, 2000a) have examined the relationship between text types, characterized by specific discourse strategies, and the phenomenon of thematization, both in English (Lavid 1999) and across different languages (Lavid 2000a). This body of work shows the existence of statistically significant correlations between those strategies and the semantic types of Theme selected to signal them. Let us remember that method of development refers to the elements chosen as Theme in terms of the Transitivity structure (whether Participants, Circumstances or Processes). On the other hand, thematic progression refers to how the Themes develop throughout the text. The relationship between Themes (thematic progression) will also vary dependent on the genre of the text. Hence it could be said that various registers on the cline of register will employ different methods of realization of Themes and discourse Topics. Different types of text will realize Themes in different ways. Different registers fulfil different language functions. The developmental processes (historical, ontogenetic and registerial) involve the extension of register range; that is, certain registers appear in the repertoire of a society or individual at particular stages in their development. [...] Hence the rhetorical structure, which involves both the content of the Theme, and the type of development of the global Topic can be expected to vary with the register, and can be expected to vary according to the stage of development. (Leckie-Tarry 1995: 147)
Thus, texts that are fulfilling different purposes will exhibit differences in both their thematic progression and method of development, even if the texts have the same subject matter. Accepting the basic position that all written discourse is organized in terms of thematic progression, we suggest that the thematic progression patterns which are manifest in a discourse are constrained by such factors as purpose, audience, and context, and that these factors may help to account for variations in thematic progression patterns which exist between different versions of the same subject matter. (Nwogu and Bloor 1991: 370)
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.49 (101)
The thematic structure of dialogue
There exists, certainly, a body of research that tries to establish taxonomies of genres, or text types, through thematic progression. Ghadessy (1997) proposes a consideration of all three Theme types (textual, interpersonal, ideational), to be added to Biber and Finegan’s cluster analysis (Biber and Finegan 1986). Fries studies Theme development by setting forth four different hypotheses (Fries 1995c: 319): 1. Different patterns of thematic progression correlate with different genres, i.e., patterns of thematic progression do not occur randomly but are sensitive to genre. 2. The experiential content of Themes correlates with what is perceived to be the method of development of a text or text segment. 3. The experiential content of Themes correlates with different genres. 4. The experiential content of Themes of a text correlates with different generic elements of structure within a text. The first hypothesis is difficult to establish from the corpus study results. There is a slight preference for constant (themic) progression in both languages, whether contiguous or linear (see Table 4.9, on page 96). But the most predominant type is the one that involves the introduction of a new Theme. I would like to tentatively propose this as a characteristic of spoken language: the introduction of new Themes that are elaborated in subsequent Themes. The proposal is tentative, since we would need a more extensive corpus analysis, and across other types of spoken language. As for the second and third hypotheses, a high number of Circumstances were thematic, either as Subjects or as Adjuncts. Here there are more marked differences between English and Spanish. Whereas English prefers Participants as Subjects above any other Theme, Spanish favours Circumstances as Adjuncts (after Processes). Nevertheless, very high numbers, in both languages, corresponded to Circumstances, a reflection of the subject matter being discussed. Finally, the fourth hypothesis found no bearing in the corpus. Except for the Openings, where a high number of new Themes were present (mostly Participant Subjects in English or Processes in Spanish), there was no correlation between stage of the conversation and experiential content or thematic progression.
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.50 (102)
Chapter 4
. Summary This chapter has addressed the structure of dialogues from the point of view of the Theme. Theme, defined as the first element in the clause, was studied first in relation to co-occurrence and to the breakdown of language functions in textual, interpersonal and ideational. In both languages, ideational Theme (the metafunction that is always present) is often accompanied by textual and interpersonal Themes. Within the ideational Themes, there is a strong preference for unmarked Themes, which I correlate with the spoken mode. Next was the study of Theme as an element in the Transitivity structure of the clause. I discovered a preference in English for Participant Subjects as Themes, followed by Circumstance Subjects. That is, English prefers Subjects, the unmarked element, as Themes. In Spanish, the highest number of Themes were Processes, because of the syntactic characteristics of Spanish, where the Subject can be ellipted. Apart from Processes, the next most frequent type of Themes were Circumstance Adjuncts, followed by Participant Subjects. Despite the differences in language, I believe that the subject matter of the conversations influenced the choice of Themes, thus producing overall a very high number of Circumstances in Subject position, and as Themes. Those Circumstances refer to the dates and times under discussion. The second part of the chapter was devoted to the study of thematic progression in scheduling dialogues. I proposed a classification of thematic progression patterns, which I applied to the corpus. This classification of thematic progression is based on Dubois’ reformulation of Daneš’ original typology, from which I eliminated the types that derive from a hypertheme. The results yielded a very high percentage of new Themes. I believe this can be attributed to the lack of planning inherent to spoken language. There is, after the new progression type, a slight preference for themic progression across the two languages. Finally, I studied the relationship between thematic selection and progression, on the one hand, and the generic elements of the structure of the texts studied, on the other. The results obtained did not show strong correlation between Theme selection or progression and the genre under study. This topic will be discussed again in Chapter 7.
Notes . See Paulson and Goodman (1999) for a historical overview of reading research.
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.51 (103)
The thematic structure of dialogue . A slip of the tongue attributed to Reverend William A. Spooner (1844–1930), where the initial consonants of words in a sentence are exchanged, often with humorous results. An example would be saying “You have hissed all my mystery lectures”, instead of “You have missed all my history lectures”. The peculiar characteristics of Dr. Spooner’s language are described by Potter (1980). . It has been argued that there exist sentences/clauses with no anchor, and thus with no split. These are called thetic judgments, monorhematic sentences (Gutiérrez Ordóñez 1997) or All-focus structures (Vallduví 1990). . ‘Subject’ and ‘topic’ being, together with ‘theme’, some more unfortunate terms, polysemous at best and vague at worst. . Examples 14 and 15 are Downing’s (1991: 123). . I am not making a distinction, within ideational, between the logical and experiential domains (Halliday 1994: 179). . Figure adapted from Eggins (1994: 274). . References about this issue were provided during a discussion in the Sysfling mail list (16–19 August, 2003). My thanks to the participants for the references and for clarifying some points. . By a semantic definition I mean that the concept of Theme has a content (anchor for the ensuing discourse), not that all realizations of such concept need to have a denotation or be topics (i.e., not all textual and interpersonal Themes have a referent in the real world). . Also, the meaning-to-grammar direction is the approach of Systemic Functional Linguistics in general. . I will use the labels ideational and topical interchangeably throughout the chapter when I am referring to work by other authors. Nonetheless, I do not consider Theme to always be identifiable with topic (see Section 4.1). . In the Spanish original: “situar en primer lugar la información necesaria para la comprensión global del mensaje”. Translation mine. . Spanish example and English translation from Silva-Corvalán (1983). . Spanish example from Bolinger (1955). . Example from Butt and Benjamin (2000: 524). . Morphological richness is only one factor that may enable a flexible word order. . McCabe-Hidalgo (1999) considers the se in this example as a topical Theme, similar to the English dummy it. I disagree, since se is in clitic position, and clitics are never Themes by themselves. English it, on the other hand, is in subject position, a thematic slot. . Markedness is still underdefined. In this case, marked means more awkward and less frequent. . Here, and in all other tables, ‘T’ stands for ‘Total’, the overall raw count of occurrences. . A more restrictive definition than the usual one for Participant. It refers exclusively to the two interlocutors.
TSL[v.20020404] Prn:29/07/2004; 9:28
F: PB12904.tex / p.52 (104)
Chapter 4 . I am, dangerously, comparing slightly different things: Fries refers to topic, Berry to grammatical Subject, and I am describing Theme. . Ideational, but not necessarily topical (see Section 4.1). . See Section 5.4 more a more detailed discussion of the product/process view of conversation. . Although Bäcklund (1992) and Romero Trillo (1994) have shown that extralinguistic, or non-verbal, elements, such as pauses, can be thematic. . Thematic I in successive utterances by different speakers is not treated as contiguous, since the reference is different. The second turn of this example contains another I in thematic position. That, however, refers to speaker MPMM, and is not part of the TP pattern initiated in the first turn by speaker FRJP. . I have not included multiple types here, because they are always derived.
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.1 (105)
Chapter 5
Rhetorical relations in dialogue
The present chapter furthers our study of scheduling dialogues through the analysis of rhetorical relations. The previous chapter examined the structural linkage and progression of elements on the surface of discourse. This one tackles underlying relations in the conversations. Martin, when describing conjunctive relations, points out that logicosemantic relations – what I will call rhetorical relations – are very sensitive to the staging structure of a genre. If we want to analyze text in context, the analysis of rhetorical relations is thus a useful place to start whenever the structure of whole texts is under consideration and an interpretation of their relationship to ideology, genre and register is what is required. (Martin 1992: 269)
My approach in analyzing rhetorical relations in dialogue was to take the conversations, first as sequences of turns that each speaker constructs, and second as whole texts that take shape through the interaction of both speakers at the same time. The main results in both types of analysis are described in Section 5.5. Before I present the results of the analysis, I provide, in the next section, a brief account of how rhetorical or coherence relations have been applied to the analysis of text, followed by a complete description of the theory I will be using, Rhetorical Structure Theory, in Section 5.2. Section 5.4 touches upon some of the problems in applying Rhetorical Structure Theory, initially conceived for written texts and monologic discourse in general, to the analysis of conversation. This is followed by the two sections on results, and by Section 5.7, which deals with the relationship of rhetorical relations to discourse markers.
. Rhetorical relations and text analysis For the analysis of underlying relations in the discourse, I will follow Rhetorical Structure Theory, henceforth RST. This section introduces the notion of
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.2 (106)
Chapter 5
rhetorical relations. Relations of this type have been observed for a long time,1 and they have received different names: rhetorical predicates (Grimes 1975), coherence relations (Hobbs 1979; Sanders et al. 1992; Kehler 2002), clause relations (Hoey 1983), and rhetorical relations (Mann and Thompson 1988). The more general term discourse relations (or discourse structure relations) is often used as a theory-neutral descriptor, even when relations are part of specific theories of discourse (Hovy and Maier 1992; Lascarides and Asher 1993; Stede et al. 1998; Webber et al. 1999). Knott and Sanders point out that coherence relations evolve out of a theory of discourse, and [t]he starting point for any theory of discourse is the oft-noted observation that what we call “a text” is more than just a collection of random sentences. (Knott and Sanders 1998: 135)
“More than just a collection of random sentences” means that texts show some internal coherence, that there are some established methods for linking one portion of text to another: The idea is that a choice amongst a finite set of alternatives must be made when juxtaposing two portions of text, and hence that a finite set of coherence relations will be sufficient to enable an analysis of every coherent text. (Knott and Sanders 1998: 136)
Note that coherence relations, in this account, are only supposed to enable the analysis. The assumption for this study has been – following Martin (1992), among others – that, whether recognized by the interactants or only by the analyst, coherence relations are underlying relations among the propositions in a text. Once we have agreed on the definition, the two most important issues remain whether coherence relations are analytical tools or psychological concepts; and how many coherence relations there are. As for the first question, accounts vary. Grosz and Sidner propose that readers do not need to understand what coherence relations exist in a text in order to understand it, and that relations are useful primarily as analytical tools, to describe text structure (Grosz and Sidner 1986). Traum (1993), also speaking on the number of relations, believes that it is not necessary for conversants to recognize a particular set of relations, as long as they can reach an agreement on the intended effects of those relations. Other accounts – (Hobbs 1979, 1990a), (Mann and Thompson 1988), (Knott and Sanders 1998), (Sanders et al. 1993) – prefer to think of coherence relations as cognitive entities. In this view, coherence rela-
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.3 (107)
Rhetorical relations in dialogue
tions are cognitive mechanisms that writers draw upon to join pieces of text together, and that readers recognize when interpreting those pieces. The claim for psychological validity of coherence relations in interpretation is as follows. When readers process a text, they construct a representation of the information it contains. A crucial property of this cognitive representation is that it integrates the individual propositions expressed in the text into a larger whole, just as our representation of an individual situation in the world draws together a number of separately perceived events and states [...]. The psychological claim about coherence relations is that they should be seen as modelling the different ways in which this integration between propositions can occur. According to this idea, determining the coherence relations in a text is part of the process of understanding it. (Knott and Sanders 1998: 138)
Knott and Dale (1994) provide some evidence from empirical studies to substantiate the previous claim, but note that the psychological basis of coherence relations is not as easily proven. Although coherence relations are used in automatic natural language generators – (McKeown 1985), (Hovy 1993), (Moore and Paris 1993) – it is not clear whether humans use them in text production. Sanders and Noordman (2000) suggest that there exist differences in the processing of different relations: in their experiments, problem-solution relations were processed faster than additive list relations. The processing is further aided by the use of discourse markers (e.g., because), added to causal relations. Here, I will assume that speakers and hearers share a notion of coherence, possibly represented in a number of rhetorical relations. The relations, from my point of view, help describe the structure of propositions in discourse, and how speakers build text that shows that structure. The next question in text analysis through coherence relations is how many of them we should use, or, in more general terms, how many of them there are. The taxonomies range from two (Grosz and Sidner 1986) to over 100 (Hovy and Maier 1992; Martin 1992). The answer is already clear in Mann and Thompson’s original accounts (1988 mostly). The number of relations is supposed to be flexible, but manageable, and they need to be attested in naturally-occurring discourse. Attempting to provide an exhaustive inventory of rhetorical relations may be compared to building an inventory of paratactic (subordinate) relations. Indeed, Matthiessen and Thompson (1988) propose that rhetorical relations at the lower level are grammaticalized into paratactic relations – likely then more frequently signaled by discourse markers, as we shall see in Section 5.7. Whereas an elementary set of rhetorical relations can be as easily described as with paratactic relations, a complete and exhaustive
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.4 (108)
Chapter 5
one is not necessary in most cases. Different researchers will use different sets for different purposes. For the analysis described here, I chose the original set proposed by Mann and Thompson (1988). Similarly, there exist different incarnations of the concept of rhetorical, discourse, or coherence relations. Bateman and Rondhuis (1997) provide a summary of three different approaches: Segmented Discourse Representation Theory (Asher 1993; Asher and Lascarides 2003), Conjunctive Relations (Martin 1992), and RST (Mann and Thompson 1988). Another approach, inspired by Hobbs (1979), is continued in the work of Kehler (2002). Researchers in The Netherlands, in collaboration with colleagues elsewhere, have developed an extensive body of research on how to classify coherence relations (Sanders et al. 1992; Sanders 1997), on their processing (Sanders et al. 1993; Sanders and Noordman 2000), and their linguistic marking (Knott and Sanders 1998). The analysis presented here draws mostly from the work of Mann and Thompson, as described in the next section.
. Mann & Thompson’s Rhetorical Structure Theory Rhetorical Structure Theory is defined as “a descriptive theory of a major aspect of the organization of natural text” (Mann and Thompson 1988: 243). RST constitutes one of the approaches devised to explain and analyze textual coherence and, in more general terms, text organization. According to the authors, in order to provide a system that analyzes the organization of a text, the theory should account for the kind of parts in the text, the arrangement of the parts, and the way they are connected to form a whole. But it should also provide a natural descriptive account of any particular text. As for the first three needs, RST explains and provides justification for them, which will be described further on. The only problem with the theory, as its authors acknowledge (Mann and Thompson 1987), is that the description has only been applied to short texts – although in another paper (Mann and Thompson 1988), they state that the framework is insensitive to text size – and that it is not exactly a neutral or objective description. The result of the analysis of a large number of texts showed that virtually every text has an RST analysis, with the exception of certain text types such as laws, contracts, reports “for the record” and language as art (including some poetry). The authors, nonetheless, find their study relevant because texts liable to be analyzed under RST – texts that are hierarchically structured and functionally organized – predominate “in our culture”.
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.5 (109)
Rhetorical relations in dialogue
circumstance
motivation enablement
joint contrast
sequence sequence
Figure 5.1 Examples of the five main schema types
There are two basic types of parts in a text: nuclei and satellites. A nucleus and a satellite are two non-overlapping text spans, which hold a relationship between each other, called a relation. The relation definition consists of four fields (Mann and Thompson 1988): 1. 2. 3. 4.
Constraints on the Nucleus, Constraints on the Satellite, Constraints on the combination of Nucleus and Satellite, The Effect.
To specify each field for the relations encountered, the analyst must make particular judgments, based on context and the intentions of the writer. These judgments are of plausibility more than certainty, that is, the analyst judges whether it is plausible that the writer had such intention or desired that Effect when creating the text. Besides nucleus and satellite, and the idea of relation, there are additional elements in RST that need to be explained – namely, schemas, schema applications and structures. Schemas define the relation between a small number of text spans. They represent all the possible RST structures. There are five kinds of schemas, shown in Figure 5.1, from Mann and Thompson (1988: 247). Straight lines represent the nuclear spans, and the curves denote the relations. The relations not represented here all follow the Circumstance pattern, and are named after the corresponding relation. Schema applications may differ slightly from the above representation, because of a different order in the text spans (e.g., nucleus preceding satellite in the first schema), or in cases where a relation is repeated within a schema.
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.6 (110)
Chapter 5
Table 5.1 Organization of the relation definitions in RST Circumstance Solutionhood Elaboration Background Enablement and Motivation Enablement Motivation Evidence and Justify Evidence Justify Relations of Cause Volitional Cause Non-Volitional Cause Volitional Result Non-Volitional Result Purpose
Antithesis and Concession Antithesis Concession Condition and Otherwise Condition Otherwise Interpretation and Evaluation Interpretation Evaluation Restatement and Summary Restatement Summary Other Relations Sequence Contrast
Structure in a text refers to the way schemas are applied. A text to be analyzed is divided into units with functional integrity, which are basically clauses, except for clausal Subjects and embedded clauses considered as part of their host clause. The relations do not hold only between clauses, but between spans of texts, in a recursive manner. Thus, it is possible to characterize whole texts using only a relatively small number of relations. The recursive application of relations characterizes the structure of a text. The defined relations are represented in Table 5.1, from Mann and Thompson (1988). Those relations are not to be considered the only ones, and different analysts may find the case for adding relations to the list, when dealing with other genres or cultural styles. Definitions are based on functional and semantic criteria, never on morphological or syntactic signals, because no reliable or unambiguous signal for any of the relations was found. This is, in my opinion, part of the success of RST – its basis on functional criteria, because this is what it makes it a language-independent theory. It has been applied to Chinese (Cui 1986; Kong 1998; Ramsay 2000), German (Rösner and Stede 1992; Reitter 2003), Dutch (Abelen et al. 1993) and French (Péry-Woodley 2001), among other languages. The fact that it is language independent also makes this theory useful in text generation systems. In order to understand the fields given above, and how they combine to define and characterize a relation, let us see one definition, for Non-Volitional Result (Mann and Thompson 1988):
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.7 (111)
Rhetorical relations in dialogue
relation name: NON-VOLITIONAL RESULT constraints on S: S presents a situation that is not a volitional action constraints on the N + S combination: N presents a situation that caused the situation presented in S; presentation of N is more central to the writer’s purposes in putting forth the N–S combination than is the presentation of S the effect: the reader recognizes that the situation presented in N could have caused the situation presented in S locus of the effect : N and S Text example: The blast, the worst industrial accident in Mexico’s history, destroyed the plant and most of the surrounding suburbs. Several thousand people were injured, and about 300 are still in hospital.
As we can see from this relation, the notions of satellite and nucleus are not necessarily related to any syntactic factors, and they do not correspond to main and subordinate clause. Besides, the text example provides a realization of the relation where no marker is present; it is only the logical and propositional transition from one sentence to another which signals the relation. Relations are divided into two main types, presentational and subject matter, depending on what the intended effect on the reader is. In subject matter relations, the intended effect is that the reader recognizes the relation in question. In presentational relations, on the other hand, the intended effect is to increase some inclination in the reader, such as the belief in a proposition or a positive inclination for a statement. Presentational relations are: Motivation, Antithesis, Background, Enablement, Evidence, Justify and Concession. All the other relations are of the subject matter type (Mann and Thompson 1988: 257). RST has not been free of criticism. One of the main objections to the theory, as we saw in the previous section, is the number of relations: Hovy and Maier (1992) propose a taxonomy that conflates the 400 or so relations proposed by different frameworks into a set of 70. Even when taking the orthodox RST description, researchers who applied it to specific domains have altered the original set. For example, Rösner and Stede (1992) use RST for the automatic generation of technical manuals and find the need for Precondition, Until, Alternative and Step-sequence relations. In my own analysis, I have often missed an Alternative relation, usually expressed by an “either... or” sequence, as in the example below. (75) [FEAS_MTMR_11] feas_11_01: okay, /uh/ it looks like the, earliest time I’d be able to see you would be, Thursday, June third, /uh/ maybe after lunch, or, we could meet for lunch or something.
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.8 (112)
Chapter 5
The other main objection to RST analyses stems from the subjectivity and uniqueness in the choice of relations. There ought to be only one relation holding between parts of a text. If multiple candidates are possible, the analyst is to choose the most plausible relation. Moore and Pollack (1992), Maier (1996b) and Redeker (2000) have noted that often there is more than one relation, but those relations belong in different levels in the structure of discourse. Moore and Pollack (1992) establish an informational and an intentional level, after Grosz and Sidner (1986). Maier (1996b) opts for Halliday’s metafunctions as levels in the discourse. In all cases, more than one relation is allowed, with relations taking place at the different levels of the structure of discourse. Moore and Pollack (1992) point out that such an observation is accounted for in the theory, in the split between presentational and informational relations. However, RST’s constraint about deciding for only one relation in each case overrides the potential benefits of the two types of relations. Mann and Matthiessen (1991) discuss the comparison of RST with Halliday’s language metafunctions, dividing the original rhetorical relations across the three metafunctions according to the locus of effect for the relation. However, they rule out the possibility of providing more than one RST relation for each nucleus-satellite link. Despite these shortcomings (apparent or real) I decided to use the taxonomy proposed by Mann and Thompson (1988) because my intention was to show how we could conduct an RST analysis of spoken language with the theory “as is”. The assumption was that, if RST was designed to account for functionally-organized text, then dialogues should be as suitable for a functional analysis as written language. A related issue is whether RST, which was originally developed and applied to English, would be suitable for the analysis of Spanish. The theory has been applied to languages other than English, but the objections made, for instance, to the taxonomy of relations by Rösner and Stede (1992) seem to come from the application to different domains, not to different languages. As with the different domain, I decided to keep the original taxonomy for the new language, so that the results for both could be comparable.
. Rhetorical relations in the present study This study presents some significant innovations in the application of rhetorical relations to text analysis. First of all, the mode is spoken language. Secondly, this is a cross-linguistic study of which there are few precedents, among them the work of Rösner and Stede (1992), who applied RST to the analysis of tech-
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.9 (113)
Rhetorical relations in dialogue
nical documents (written language) in English and German for the purposes of Natural Language Generation.2 And finally, it is applied to Spanish, which had not been, to my knowledge, analyzed employing RST either in the written or spoken form.3 With some exceptions, coherence relations, especially within the RST framework have always been applied to written text. RST has not yet been effectively related to dialogue. It must be expanded beyond written monologue to dialogue and multilogue in order to encompass a fully representative range of the functions of language. This will perhaps be the most significant modification of its present form. (Mann et al. 1992: 68)
Fawcett and Davies (1992) propose RST analyses of conversations that cover intra-turn relations, thus considering a turn as a monologue within a conversation. Daradoumis (1996) extends RST to relations across turns. He follows Berry’s (1981) and Martin’s (1992) exchange model and provides an extended version, Dialogic RST, with new relations to capture the exchange structure of conversation. Daradoumis’ work concentrates also on collaborative conversation, especially tutorial dialogues. Stent (2000) proposes a number of new relations to capture relations across turns (Question-response, Greetingacknowledgement). My work differs in that I did not attempt a modification of the original theory, although I do consider relations across turns. A combination of a coherence analysis with a generic analysis, such as the one carried out here (Section 7.5, in Chapter 7) can provide a full account of both local and global coherence. In this chapter I present the results of an extensive cross-linguistic analysis of a corpus of spoken language, in order to show the explanatory power of rhetorical relations. As argued for below, I only considered the initial set of relations described by Mann and Thompson in the central description of the theory (Mann and Thompson 1988). The extension, and sometimes reduction, of the number of useful rhetorical relations has been, as mentioned earlier, a very much debated question – see, e.g., Hovy and Maier (1992) and Sanders et al. (1992). An important reason for considering only the initial set of relations is that those relations have full definitions and examples in the literature – mostly in Mann and Thompson’s 1988 article. As with the analysis of cohesion in Chapter 6, the main purpose was to apply a well-established type of analysis in order to unveil its merits and limitations in a new domain and a new language. My intention is to show how a set of relations which could be enhanced (see above on the need for an Alternative relation), is still powerful for three different purposes:
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.10 (114)
Chapter 5
1. Description of spoken dialogue 2. Comparison between two languages 3. Application to a new language, Spanish There has been in the last few years an increased interest in coherence relations, especially in how they are signaled through discourse markers. An International Workshop on discourse markers produced two special editions of different journals, Discourse Processes (Discourse Processes 24 (1), 1997) and Journal of Pragmatics (Journal of Pragmatics 30 (2), 1998). The Association for Computational Linguistics held a workshop under the title “Discourse Relations and Discourse Markers” in their 1998 meeting (Stede et al. 1998). The interest in discourse markers has led to the idea that we can find and taxonomize coherence relations only through the lexical elements that signal them. Knott and colleagues, especially, have gone from taxonomizing the “cue phrases” used to mark coherence relations (Knott and Dale 1994) to proposing that a taxonomy of cue phrases is equivalent to a taxonomy of coherence relations (Knott and Sanders 1998). My results provide support for the hypothesis that, if we rely on discourse markers alone, we will never taxonomize the full range of relations that speakers employ. As we will see in Section 5.7, a good number of the rhetorical relations present in my corpus were not marked at all. A taxonomy that relies on external marking will miss the presence of those relations, which seem to be quite numerous in spoken dialogue.
. RST in conversation An RST analysis of a text implies that the text in question is functionally and hierarchically organized. That dialogue, and especially task-oriented dialogue, features a functional component is intuitively plausible. Hierarchical organization might be less plausible in an on-going effort of two participants who have not planned this particular conversation before, are not allowed to produce drafts and final versions, and do not know what the other speaker will say – not specifically, at least, although there is some predictability in a familiar genre or type of conversation. Considering functional and hierarchical organization, there are two different ways in which the analysis can proceed, according to two different points of view. The first point of view is that of the analyst, where the conversation is presented as a product of the interaction of the two speakers. In this view the conversation is an autonomous, independently-functioning piece of text.
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.11 (115)
Rhetorical relations in dialogue
The analyst is an observer, maybe even a recipient of this final product. We impose a second layer on the dialogue, a detachment from the context where it transpired to the context of the analysis. If this detachment happens at the expense of the original context, how is it useful? Away from its original context and in a different one, the analysis might provide insights into how the interactants construct a piece of text in a joint manner, taking into consideration what has happened so far in the dialogue and incorporating the other speaker’s contribution to modify or adjust their own. Mann and colleagues (Mann et al. 1992: 41) consider “some dialogues as structured collections of texts”. It is that structural relation that I will try to unveil here: how the dialogue as text is organized and what general characteristics we can find in a corpus of such texts.4 The second point of view from which we could examine these dialogues is that of the interactants. In this view, the conversation is a process to which both speakers contribute in their respective turns. Each turn would be an independently created text, a response to the overall context, but a text in its own right. This type of analysis reveals how and when each turn is autonomous, and how speakers create mini-pieces of discourse with the goal of contributing to a dialogue. Another consideration is that RST is, in a sense, a theory of influence; it is very well suited to describe what the writer/speaker does to influence the reader/hearer. A look at the inventory of relations will suffice, especially the ones grouped under the presentational heading, “those whose intended effect is to increase some inclination in the reader” (Mann and Thompson 1988: 257). Presentational relations are: Motivation, Antithesis, Background, Enablement, Evidence, Justify, and Concession. Text structuring relations are functional; the character that they all share can be stated in terms of the categories of effects that they produce. They can be described in terms of the purposes of the writer, the writer’s assumptions about the reader, and certain propositional patterns in the subject matter of the text. The text structuring relations reflect the writer’s options of organization and presentation; it is in this sense that an RST structure is “rhetorical”. In contrast, one could assume that text structuring relations simply represent relations in the subject matter (e.g. of succession, cause or conditionality). (Mann et al. 1992: 44–45)
I will show in this corpus that the speaker makes decisions and presents his or her material in a particular way to achieve an effect. The choice and the presentation might not be as elaborate as it is in written discourse, due to time
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.12 (116)
Chapter 5
constraints and the impossibility to revise the material, but we still observe options of organization on the part of the speaker. The turn-by-turn analysis looks at each utterance as a text in itself that seeks some reaction – a reaction, unlike in written discourse, immediate and usually clearly visible. This type of analysis ignores the undoubtedly important relations holding between turns, but approximates the original RST analyses in that it looks at text as the product of one mind that projects towards a recipient. Undertaking both types of analysis bridges the gap between purely interactional and purely textual approaches. Redeker (2000), in a review of coherence approaches in dialogue, criticizes the emphasis on exchange patterns in the work of conversation analysts and others (Schegloff and Sacks 1973; Clark and Schaefer 1989), in the form of adjacency pairs. This emphasis disregards the internal structure of turns, often structured in an independently coherent way. Briz and Hidalgo (1988) also point out that coherence in a conversation is established both at the monologic level (each speaker’s contribution) and at the dialogic level (the exchange system). These two types of analysis capture that internal coherence, without disregarding the relations between turns. The rest of this section disscusses some of the issues in carrying out an RST analysis of spoken language. The first problem in applying RST to the dialogues was related to their very nature. In naturally-occurring dialogue there are different elements pertaining to the specific genre, such as greetings and goodbyes, and there are also different degrees of spontaneity in the language, such as the presence of selftalk versus the more planned turns in which a speaker’s utterance evolves in an easily definable pattern. Let us focus our attention, in the first place, on the turn-by-turn analysis of the conversations. There we will look at different phenomena: single-span utterances, self-talk, parallel relations and genre-specific stretches of talk. RST establishes a relation between spans of text, either nucleus-nucleus or nucleus-satellite relations. In the corpus here analyzed, however, there are examples of utterances that consist of only one span and are, therefore, not analyzable in RST terms. Some of them were related to the rest of the dialogue, holding a relation with a previous or following utterance, such as requests for elaboration, and some of them were independent as far as rhetorical relations are concerned, and only coherent in terms of the context of situation and the genre in question. In Example (76), the speaker is formulating a request to the next speaker. This is the Opening of the conversation, and as a consequence, there is only a relation holding between this span and the rest of the conversation, but there is no relation internal to the turn. In Example (77), from the
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.13 (117)
Rhetorical relations in dialogue
Spanish data, there is also a single span, a request for a free time slot. (Refer to Chapter 3 for transcription conventions and for an explanation on how the translations were carried out.) (76) [FBNT_MJFG_1] fbnt_1_01: /oh/, would you like to meet, /uh/, for a two hour appointment, *pause* on, *pause* Monday? at, five P M, (77) [FFCS_FSNM_01] fsnm_01_04: no tienes ninguna hora entre martes y viernes de esa semana? don’t you have any time between Tuesday and Friday in that week?
The next example shows two different groups of spans within the same utterance. The first part is mainly interactional and not related to the task itself, but to the performance of the task (“well, totally just missed what you said, but if it’s a date oh yeah, yeah”). The second part is related to the task, reflects a reaction to what the previous speaker said, and consists of three text spans: the first two in an elaboration relation, and the last one as the nucleus of a concession relation, as illustrated in Figure 5.2. The span numbers displayed in the figure are shown between square brackets in the example. (78) [FCAD_FJAB_12] fjab_12_02: well, /um/ totally just missed /begin_lg/ what you said, /end_lg/ but if it’s a date, oh yeah. yeah. [1] I’m out of town then. [2] I’m out of town from Monday through Wednesday. [3] but /um/, the next week I’m free, for the most of the time. /lg/
Figure 5.2 Rhetorical relations in Example (78)
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.14 (118)
Chapter 5
The previous example is close to what Hayashi (1991), based on studies by Edelsky (1981) and Shultz and colleagues (Shultz et al. 1982), calls nonpropositional floor, that is, floor-holding while not contributing to the conversation in a totally conscious manner. Non-propositional floor is talk that does not directly address the interlocutor(s), but that serves as floor-holder, “moments when a speaker is absorbed by his/her own thoughts and momentarily creates a highly self-preoccupied floor which has nothing to do with the on-going floor, so that nobody else joins in or pays attention to the speaker” (Hayashi 1991: 9). In the following Spanish example, the speaker is holding the floor, presumably while checking her calendar. She has just given a possible time for a meeting, and then she realizes that she is not free then, or that it is not a convenient time. She holds the floor, uttering a sequence of four ‘nos’ until she finds the time when she can meet. (79) [FKND_FKAS_01] fknd_01_02: bueno. | el lunes? | no. | tengo una reunión en la mañana y un almuerzo. | me quedaría muy apretado el tiempo. | no. | qué tal el día viernes? el viernes doce. entre las once y media y las dos de la tarde? | no, | no | no | no. | mejor el viernes entre las once y la una. well. | on Monday? | no. | I have a meeting in the morning and a lunch. | it would be really tight. | no. | how about Friday? Friday the twelfth. between eleven thirty and two in the afternoon? | no, | no | no | no. | it’s better on Friday between eleven and one.
The presence of such instances poses a problem for the RST analysis because they are parentheses in the development of the turn. They break the Coherence Rule as described by Tsui: a sequencing rule governing what can occur if the discourse is to be coherent: an utterance must be related to either the illocutionary intention or the pragmatic presuppositions of the preceding utterance; if neither, it will fail to form a coherent sequence. (We shall refer to this rule as the Coherence Rule, to distinguish it from the rule governing an “adjacency pair”).(Tsui 1991: 123)
Tsui refers to the instances of a breakdown of the Coherence Rule, whether intentional or not, as non-sequiturs (Tsui 1991: 124). Her examples refer to extreme cases of purposeful violation or blatant misunderstanding, which are not usually found in the corpus I analyzed. The examples found in the corpus are nevertheless problematic for a theory that relies on connectedness of neighbouring spans. One could decide to interpret them as parenthetical, and
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.15 (119)
Rhetorical relations in dialogue
continue with the analysis of the preceding and following segments, or to segment the turn at that point, analyzing the previous stretch as a unit, then the self-talk part (if suitable for analysis), and finally analyzing the following stretch of talk as another unit. The reason for this is that sometimes the flow does not continue where it was left before the parenthetical self-talk, but takes a different direction. Such is the approach illustrated in Figure 5.2 above. This violates the constraint of connectedness for the whole text unit, but here I am considering smaller stretches of text to be text units, and the authors of RST do acknowledge that “a very few texts [. . . ] can be analyzed only if the adjacency constraint is relaxed.” (Mann and Thompson 1988: 249). Another problematic case is that of parallel relations between two turns of talk. In Example (80), the two underlined segments hold a relation with each other, but no relation with the rest of the turn itself. In those cases, I decided to divide the turn into two independent groups of relations. The connections will play a role in the analysis of the conversation as a whole, but not in the turn-by-turn analysis. (80) [FBNT_MJFG_1] fjab_12_04: well Mrs Daley, I, happen to /eh/ be in a seminar until six o’clock, on Friday, so I really have a true excuse. | but, I like /um/ the Monday thing, two weeks from now, the twenty sixth, | it’s a good time for me. fcad_12_05: that’s Ms Daley to you, | and, what time would you like to have it on the twenty sixth, | how ’bout, say, that lunch thing again, | you can buy me lunch again at like, twelve to two? | that sounds good.
The next problematic aspect is that of genre-related aspects of the dialogues. Apart from the cases already described, the purpose of some stretches of talk seems to be merely that of complying with the characteristics of dialogue in general, or task-oriented dialogue in particular. Mann and colleagues (Mann et al. 1992: 41) describe three different types of structure a text can have: 1. Holistic Structure, deriving from the genre or variety of text. In a letter, this would account for expressions such as “Sincerely”. 2. Relational Structure, which expresses the organization of coherent contiguous text. 3. Syntactic Structure. RST can interface with different syntactic theories, although a functional theory will simplify the interface considerably. In this analysis the particular stretches of talk discussed above correspond to the holistic structure of the text. Turns in which only greetings or goodbyes
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.16 (120)
Chapter 5
are exchanged do not show any coherent organization in relational terms; they merely serve to establish or maintain the communication, and to ascribe the texts to a specific genre, which both speakers know and exploit (see Chapter 2 on general issues related to genre and Chapter 7 on the staging of these particular dialogues). Example (81) shows the end of a conversation where speaker MRCT confirms the appointment, and speaker FRET agrees, repeats the date again, and signals that the conversation is over. Example (82) illustrates a similar situation in a Spanish dialogue. (81) [FRET_MRCT_1] mrct_1_06: okay, what do you say we meet at, my office at, two? | is that good for you? fret_1_07: that’s fine, | I’ll see you the sixteenth, at two. (82) [FJGC_FKAS_04] fjgc_04_04: /um/ el veintiséis no hay problema. | /um/ nos podemos reunir de once a una de la tarde. | está bien? fkas_04_05: sí | está bien. | perfecto. | hasta luego. um the twenty-sixth there is no problem. | um we can meet from eleven to one in the afternoon. | is that good? yes | that’s good. | perfect. | see you later.
The following example shows a combination of the two types of structure. This is the beginning of the conversation. Speaker FRJP greets the other speaker, identifies herself and then proceeds onto specifying what it is that she wants to achieve with this conversation. In her turn both types of structure coincide. In MPMM’s turn the acknowledgment that the message has been understood is still interactional, or holistic, whereas the rest of the turn displays a relational structure. Holistic elements of the conversations were discarded for the RST analysis. (83) [FRJP_MPMM_1] frjp_1_01: /ah/ Peter. this is Renee. | I need, to schedule a two hour meeting with you, between March, eighth, and March nineteenth? mpmm_1_02: okay, I’m completely free, on the seventeenth of March. | so, let’s try that date, first.
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.17 (121)
Rhetorical relations in dialogue
. Results: Turn-by-turn analysis As described in Section 5.4, two different types of analysis were carried out. This section and the next will describe the results for each of them in both languages, listing the relations found and providing examples. I also mention whether they were signaled or not by a discourse marker. For a complete summary of discourse markers in the corpus, please refer to Section 5.7. The units of analysis are the Semantic Dialogue Units in which the corpus is already segmented (see Chapter 3 for a formal definition of SDUs). The talk included in each of the SDUs can be separated into further RST-like spans. For the RST analysis I divided clausal circumstances, time adjuncts, and conditionals into different spans from their host clauses, even though some of those were included in the same SDU. The units of analysis are, initially, the same as those in the thematic analysis (see Section 4.3 in Chapter 4). However, some of the units were discarded because they were part of the holistic structure of the conversations (see previous section). A number of units were also not included in the turn-by-turn analysis, in cases where a turn contained a single unit, therefore showing no internal rhetorical structure. Those units did become part of the analysis of entire conversations. Table 5.2 shows a list of the relations found in the corpus, for Spanish and English. As can be seen from the table, the results for English and Spanish, in terms of relations and their distributions, are very similar. In both cases, Elaboration shows the highest number of occurrence, followed by Concession and Condition. Lower in the list the relations vary only minimally in percentage distribution. In the next few sections I provide some details and examples of the most frequent relations. .. Background Background is mostly found at the beginning of the conversation, when the first speaker is trying to explain why the meeting is needed. The effect of this relation is that the hearer’s ability to comprehend the nucleus increases. In Example (84), speaker MAKK provides some background (“I think it’s time we made another appointment”) so that the hearer understands his question, “what would be good for you”.
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.18 (122)
Chapter 5
Table 5.2 Number and percentages in the turn-by-turn analysis English Antithesis Background Circumstance Concession Contrast Condition Elaboration Enablement Evaluation Evidence Interpretation Joint Justify Motivation Non-Volitional Cause Non-Volitional Result Otherwise Purpose Restatement Sequence Solutionhood Summary Volitional Cause Volitional Result n
Spanish
T
%
T
%
1 13 5 71 9 66 166 10 8 0 1 29 34 3 37 43 4 10 28 7 7 4 7 17 580
0.17 2.24 0.86 12.24 1.55 11.38 28.62 1.72 1.38 – 0.17 5 5.86 0.52 6.38 7.41 0.69 1.72 4.83 1.21 1.21 0.69 1.21 2.93
1 16 9 70 8 89 140 4 11 4 11 26 22 0 54 29 4 26 47 20 0 6 11 48 656
0.15 2.44 1.37 10.67 1.22 13.57 21.34 0.61 1.68 0.61 1.68 3.96 3.35 – 8.23 4.42 0.61 3.96 7.16 3.05 – 0.91 1.68 7.32
(84) [MAKK_MKGD_10] makk_10_01: hi I, think it’s time we made another appointment? | /um/? what /uh/, would be good for you preferably in the week of the tenth to the fourteenth.
In Spanish, Background also occurs at the beginning of the conversation. However, in Spanish all instances of Background happened at the beginning, whereas in English it is present only in 10 out of 30 conversations at the beginning – the rest of the Background relations were found elsewhere in the dialogue. Spanish speakers introduce a Background before they make a request. As in English, it is never marked. In Example (85), speaker MARC provides some background about a previous meeting, and explains the need for another one, before an actual request is made.
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.19 (123)
Rhetorical relations in dialogue
(85) [MARC_MPBH_02] marc_02_01: hola Pablo. | te recuerdas de la reunión que tuvimos el otro día? | /eh/ no quedé muy claro por los puntos que conversamos | y quisiéramos que fijáramos nuevamente una reunión para aclararlos. | qué te parece? hello Pablo. | do you remember the meeting we had the other day? | uh I wasn’t very clear on the issues we talked about | and we would like that we settled a meeting again to sort them out. | what do you think?
.. Concession Most of the purely negotiating exchanges contain examples of the Concession relation. The formula is equivalent to “although I cannot meet on X (the date you proposed), I’m free on Y”. Of the total number of relations, 12.24% are concessive, showing the centrality of this type of exchanges to the conversation and the repeated use of this resource again as a politeness strategy, in order to minimize the face-threatening act (Brown and Levinson 1978) of rejecting a date proposed by the other speaker. Characteristic of the Concession relation is the frequent use of discourse markers to signal it. More than fifty-three percent of the concessive relations (38 out of 71) used a marker. Within the marked ones, the most common one is but, present in 89.47% of the marked Concession relations, together with though, present in 5.26%, and however, found also in 5.26% of them. The three of them together account for 21.23% of the total number of discourse markers found in the dialogues, making Concession the most heavily marked relation. Example (86) contains two different instances of a Concession, one marked with but embedded within another Concession relation marked with however. Note that a discourse marker is not the only signal for a given relation. In (87) we find another source of marking, the presence of the emphatic do to mark the contrast between nucleus and satellite. (86) [FCRE_FJSL_1] fjsl_1_02: well I’d like to make it as soon as possible, | /glottal/ but I’m not free for two hours, /um/, on, Monday the eighth. | however on the ninth I’m free after twelve o’clock, ... (87) [FSJB_FKDO_1] fkdo_1_02: /oh/ on Monday the eighth I have a meeting from two to four. | /um/ I do think we should get together and it’s probably gonna take us a couple of hours to hash through all that material. ...
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.20 (124)
Chapter 5
Figure 5.3 Two possible analyses for Example (89)
Concession in Spanish features the same contexts of appearance as in English: it is used to express the unavailability to meet on the date proposed and to provide a different date instead. It is also heavily marked: 54.29% of the Concession relations are signaled by a discourse marker, with pero present 50% of the time when a signaled Concession relation occurs. Example (88) shows the Concession relation without a discourse marker. (88) [MJNM_FAMM_06] mjnm_06_09: no. | no puedo porque tengo una reunión. | qué te parece el martes de doce a dos de la tarde? no. | I can’t because I have a meeting. | what do you think about Tuesday from twelve to two in the afternoon?
Such instances of Concession could also be interpreted as Solutionhood: the problem is that the speaker cannot meet on a particular date, and the solution is to propose a new date. For example, in (89), spans 1–2 and 3 could be related through a Concession relation, as some of the examples above (“although I’m away the second to the fourth, I can meet the seventh or the ninth”). But they could also be interpreted as being in a Solutionhood relation (“the second to the fourth poses a problem; the solution is to meet the seventh or the ninth”). Both possible analyses are represented in Figure 5.3). I decided for Concession in cases such as (89), because they seemed similar to instances of Concession clearly marked as such, as some of the examples above suggest. Replication of the analyses is a thorny issue in RST, and other analysts may disagree with this interpretation, but in all difficult cases, I chose the most plausible interpretation, which was also consistent with evidence elsewhere in the corpus. (89) [FJBT_MSRH_11] msrh_11_02: [1] well I’ll be on vacation the, /uh/ second, third, and fourth. | [2] I’m taking the family to Cancun. | [3] /um/ how ‘bout the seventh. or the ninth.
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.21 (125)
Rhetorical relations in dialogue
.. Condition Most instances of Condition are centered around the notion of proposing a date and submitting it to the judgment of the other speaker. Some of them are clearly analyzable as Condition, as in the following example, where speaker FACR proposes a date and signals its character as provisional with a satellite headed by the marker if. (90) [FACR_FJYK_AU] facr_au_03: /uh/ August eighth at nine thirty would be, fine. | if, that’s okay with you as well.
A more borderline case of Condition, which I considered as such because of plausibility judgments, occurs under the same circumstances as the previous one, but is realized by an interrogative clause. In Example (91) below, speaker FFMW, after mentioning a problem with the date proposed by the other speaker, presents other dates as a result of the problem, and expresses the condition through the question “do you have time any of those days?” (91) [FCJA_FFMW_10] ffmw_10_02: I forgot to tell you, /uh/ Christina, that, I have, /um/ a vacation planned, | I’m leaving this Sunday, the second, | and, I’m not coming back until, May the sixteenth. | so, /um/ it’ll have to be some time, during that week, either, Tuesday, Wednesday, Thursday, or Friday. | /uh/ do you have time, any of those days?
In English, Condition is, after Concession, the relation with the most number of markers. Fifty percent (33 out of a total of 66) of the Condition relations were marked. The most frequent marker is if, present in 72.73% of all the marked Conditions. Other markers are: and then if, depending, except for, if . . . then, otherwise, so, unless and whether. Condition in Spanish appears primarily in two different contexts: the presence of a real condition usually marked by the speaker with a discourse marker, and the question at the end of a turn indicating the conditional character of the proposal. In the first case the discourse markers are, as in English, varied: si is the most frequent one, appearing in 22.47% of all the Condition relations, and representing 83.33% of the marked ones. Other markers are a menos que and depende que. The latter case is usually realized through a question and is never signaled by a discourse marker. Example (92) is a clear case of Condition, signaled by a discourse marker, si te parece. Example (93) is a different context, presenting a hypothetical situ-
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.22 (126)
Chapter 5
ation, “if we want to make the meeting two hours long...”. This same example ends with a Condition that affects the whole turn, referring back to the date proposed. It is of the same type described in (90), a Condition expressed through an interrogative clause, qué te parece? (‘what do you think?’). (92) [FSMA_MENF_05] menf_05_04: /eh/ mirá. | el miércoles cuatro, por la tarde, yo puedo. | así que, si te parece, quedamos el miércoles cuatro, de dos a cuatro de la tarde. uh look. | on Wednesday the fourth, in the afternoon, I can. | so, if you want, we’ll meet on Wednesday the fourth, from two to four in the afternoon. (93) [MARC_MPHB_02] mphb_02_04: pero qué tú no almuerzas? hombre, <si lo hacemos de dos horas> pues, por lo menos dos horas tenemos conversar | y, tenemos que [comar (comer)]. | me viene mejor el jueves, | /gl/ por ejemplo empezar a las dos de la tarde. | qué te parece? but don’t you have lunch? man, since, at least two hours we need to talk | and, we have to eat. | it’s better for me on Thursday, | for instance beginning at two in the afternoon. | what do you think?
.. Elaboration Elaboration is, by far, the most frequent relation: 28.62% of all the relations found in English were Elaborations. Unlike other relations, Elaboration has no salient context. It is present in many different circumstances, adding detail about the situation or some element of subject matter presented in the nucleus. It is rarely marked and when it is, and is the only marker found in the corpus. In (94), the Elaboration expands the concept “end of the day” by exactly referring to the time. The addition to the previous thought is signaled by and. (94) [MTJH_MRIL_3] mtjh_3_07: yeah, why don’t we just, /um/ put it at the end of the day, and, go, from three to five. | if that’s okay with you,
In (95), the Elaboration satellite (“that will actually work pretty well”) served as support for the adequacy of the date proposed.
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.23 (127)
Rhetorical relations in dialogue
(95) [FECE_MNFH_10] fece_10_03: well why don’t we try for another lunch meeting on the eighteenth from, twelve to two. | /um/, that will actually work out pretty well. | where do you want to meet. /sniff/
In Spanish Elaboration is also the most frequent relation: it is the relation in 21.34% of the cases, and is found in very different contexts. Unlike the instances of the English Elaboration relation, in Spanish it is more frequently signaled by markers such as además, pero inclusive, por ejemplo, pues, and y (the latter being the most frequent marker). Elaboration is marked 6.43% of the time. Example (96) shows an Elaboration relation signaled by además. The satellite starts at the point where we find the marker, and it affects the preceding stretch. (96) [MEJH_MJJG_01] mjjg_01_05: /eh/ de qué viernes estás hablando. | del doce, o del diecinueve. | porque el diecinueve es mi santo. | ya que soy de San José, | entonces no puedo ir. | además las citas deben ser <de> de ocho a cinco. no después de las cinco. | que las cinco hay que ir directamente a la casa y preparar el diner. o la comida en este caso. | así que aclárame todas esas cosas por favor. which Friday are you talking about. | twelfth, or nineteenth. | because the nineteenth is my patron saint’s day. | since I am from San José, | so I can’t go. | besides the appointments must be from eight to five. not after five. | because at five one has to go straight home and prepare dinner. or lunch in this case. | so clear all those things up for me please.
Example (97) presents an Elaboration or explanation on what the speaker is asserting through a statement (eso está bien) and a Restatement (suena de lo más de bien). She elaborates on what it is that is fine – that they meet at the established time. (97) [MJNM_FAMM_06] famm_06_18: eso está bien. | suena de lo más de bien. | a las dos de la tarde el sábado el dos de octubre? that is fine. | it sounds great. | at two in the afternoon on Saturday the second of October?
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.24 (128)
Chapter 5
.. Joint Joint is a schema, not exactly a relation. The schema is multinuclear, and there is no relation holding between the nuclei. In the corpus, examples of Joint schemata are found in lists, especially lists of possible meeting times. Joint is sometimes signaled by and in English. In Spanish it is marked more frequently: 42.31% of the total number of Joint relations are marked, by y and así. In Example (98), the two proposed dates (Monday and Friday) are in a Joint relationship, signaled by y. The same structure is used to join the times (9–10 and 3–5 for Monday; 11–1 and 3–4 for Friday), but since those were not independent clauses, they were not considered spans for the RST analysis. (98) [FLCM_FKAS_08] flcm_08_01: ... el lunes voy a poder, *pause* de nueve a diez, y de tres a cinco, | y el viernes, voy a poder de once *pause* a una, y de tres a cuatro. ... Monday I can, from nine to ten, and from three to five, | and Friday, can from eleven to one, and from three to four.
.. Non-Volitional Cause Instances of both Non-Volitional Cause and Non-Volitional Result are presented as inevitable situations deriving from conflicts in the speakers’ agendas. In Non-Volitional Cause, the speaker usually apologizes for not being able to accept a certain date through this relation. It is used as a politeness strategy (Brown and Levinson 1978): the speaker does not want his or her interlocutor to lose face. As a result, the reason for the unavailability is presented as something external to the speaker, that he or she would avoid if they could. Although the speakers in the recordings were provided agendas and did not in fact have a choice, they emphasize this fact much more when there is a case of unavailability than when the case is acceptance of a date. Non-Volitional Cause is signaled by different means, most prominently by because or ‘cause, followed in frequency by since, and, and then, in that and so. Example (99), although it contains no overt discourse marker, is signaled by an emphatic do, a signal of contrast already discussed above for the Concession relation. (99) [FJMB_FMLZ_6] fjmb_6_03: okay, actually I forgot to say that what we need is a two hour meeting. /um/ | so, you gave me a couple of, what seemed to be one and
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.25 (129)
Rhetorical relations in dialogue
a half hour times, | how ’bout two hour times. | /uh/ twenty ninth is not good for me in the afternoon, | I do have a seminar all day that day, | and the afternoon of the thirtieth is also, /uh/ booked from two thirty to five. | /um/ do you have two hour block on the first? in the afternoon?
In Spanish, Non-Volitional Cause is typically found in the same contexts as in English, and it is marked in 57.40% of the cases where it is present, most often by porque, followed by es que, y and ya que. In the following examples we can find different instances where this relation occurs. The first example presents a Non-Volitional Cause marked by porque, with a Condition embedded within it. In the second example, (101), there is no marker. (100) [FLCM_FKAS_08] flcm_08_05: a las doce y media me va a parecer muy difícil | porque si tenemos que reunirnos por dos horas, yo tengo una reunión a la una. | y no voy a poder. at twelve thirty is going to look very difficult | because if we have to meet for two hours, I have a meeting at one. | and I won’t be able. (101) [MRAP_FKAS_01] fkas_01_06: no es posible. | yo tengo una reunión a las dos. | /ah/ mejor *pause* martes, <el> *pause* dieciséis, *pause* a mediodía? o a la una? it is not possible. | I have a meeting at two. | uh better Tuesday, the sixteenth, at noon? or at one?
.. Non-Volitional Result In Mann and Thompson’s account, both instances of Result – Volitional and Non-Volitional – seem to be the result of some previous action. In my corpus, Results most frequently involved a certain state of the world. More specifically, they are results of the previously established schedule that each one of the speakers is consulting. As with Non-Volitional Cause, this relation is used as a politeness strategy to protect the other speaker’s face when rejecting his or her suggestions. The conflict in schedules is an unwanted result. This is clearly expressed in Example (102) below, where the speaker hedges her contribution by using well and then saying “fortunately or unfortunately” before she mentions the result of her
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.26 (130)
Chapter 5
busy schedule. Whether she desires that outcome or not, it is a fact established before this conversation. (102) [FSJB_FKDO_1] fsjb_1_03: well fortunately or unfortunately I have to go to a family <member> member’s wedding, /uh/ in the middle of the week. | and I’m not free at all Friday the twelfth. | so it looks like we’re into the week of the fifteenth, | <what about,> we can have Tuesday afternoon, or, *pause* we could have, Thursday, | nope, strike Thursday. | we could do Tuesday afternoon. | I hope you can do Tuesday afternoon.
Non-Volitional Results are marked quite frequently. Of the English relations, 62.79% are signaled by some discourse marker: so in most of the cases, and also by since, then and if so. Example (103) is marked with so, together with unfortunately. (103) [FKCF_MJCB_1] fkcf_1_05: unfortunately I have a lunch meeting, from twelve to two. | so next, week looks bad, | how ’bout the next week,
In Spanish, Non-Volitional Result fulfills exactly the same functions as in English, and it is also very often signaled through the use of a discourse marker: así que, entonces, y, o sea que, por lo tanto, por tanto. Example (104) shows the occurrence of entonces. (104) [MRMV_FFCS_06] ffcs_06_04: Rodolfo. | estaba viendo el calendario de septiembre. | creo que estás en el calendario de octubre. | así que, este, para octubre entonces, | <me queda mejor |> el miércoles estoy fuera. | no sé. | [(có)mo] será el <el> jueves <en> después de las cinco de la tarde será. | a ver si te parece mejor así. Rodolfo. | I was looking at September’s calendar. | I think that you are in October’s calendar. | so, I mean, for October then, | on Wednesday I’m out of town. | <so,> I don’t know. | how it will be on Thursday after five in the afternoon it will be. | let’s see if you like that better.
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.27 (131)
Rhetorical relations in dialogue
.. Restatement This is usually a repetition of the date agreed upon, most frequently occurring at the end of the conversations. It is occasionally marked with so, as in the example below. (105) [FKCF_MJCB_1] mjcb_1_10: yep, the sixteenth. | so, next Tuesday at one then.
Restatement is more frequent in Spanish (7.16% of the time versus 4.83% in English), and much more frequently used as an inter-turn relation (see Section 5.6 below on its function between turns of talk). It is often marked by entonces, o sea que, así que, or recién. Example (106) shows a Restatement preceded by some hesitation: y and then bueno with falling intonation (signaled by the period marker). (106) [FSMA_MENF_05] fsma_05_03: mirá. | imposible porque el /eh/ miércoles yo tengo de dos a cuatro y media una reunión, | y el jueves a la mañana tengo una conferencia de nueve a doce. | así que tendría que ser recién, el miércoles de la semana de [yain (ya_en)] agosto, | el miércoles cuatro. | porque yo voy a estar a fuera el treinta, hasta el tres. | bueno. | recién sería el miércoles cuatro, después de la una del medio día, o a lo mejor antes de las once de la mañana. look. | impossible because on uh Wednesday I have from two to four thirty a meeting, | and on Thursday in the morning I have a conference from nine to twelve. | so it would have to be then, Wednesday the week already in August, | Wednesday the fourth. | because I will be out of town the thirtieth, until the third. | well. | then it would be on Wednesday the fourth, after one in the afternoon, or maybe before eleven in the morning.
.. Volitional Result Unlike Non-Volitional Result, Volitional Result is present when the speaker introduces the situation as a desired consequence of the circumstances. Most of the Volitional Result relations, 82.35% of the English total, are marked through a discourse marker: so the most frequent; and only a few, 11.76%, are marked with and.
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.28 (132)
Chapter 5
(107) [FAMS_FCLD_AU] fcld_au_06: what about, the eleventh? | because I have a meeting, with Mark, your favorite person, from three to four, | so, depending on how long it’ll be I can schedule you in, before that?
Marked 100% of the time in Spanish, Volitional Result serves to express the desired consequence of a situation; in these dialogues, the situation typically involves the speakers’ agendas (where they might coincide), or one speaker’s special preference for an outcome. The most frequent marker for this relation is así que, followed by entonces, y, así, o sea que, porque, por lo tanto, sea que, tal forma que, and ya. Example (108) shows a Volitional Result, that the other speaker sees him on that day, marked by por lo tanto. (108) [MJEG_MFAC_01] mfac_01_05: /ah/ yo tengo clases el viernes de las diez a las once y de las dos a las tres, | por lo tanto si <es> tú me podrías ver a las doce, ese mismo día me parecería muy bien. | *pause* qué te parece? uh I have classes on Friday from ten to eleven and from two to three, | so if you could see me at twelve, that same day I would like that. | what do you think?
The following example is somewhat different. As we saw above, the beginning of the conversation often contains a Background relation, which serves as the basis to establish why the speakers should meet again. In this case the speaker presents a situation, a consequence of which is that he needs to know about his interlocutor’s agenda. (109) [MJBP_MMBU_04] mmbu_04_01: Jaime | <se> se nos olvidó algo. | tenemos que concertar una cita, /gl/ entre el lunes veinticuatro de mayo, y el viernes cuatro de junio. | así que necesito saber algo de tu agenda, para ver cuando podemos reunirnos en ese período de tiempo. Jaime | <we> we forgot something. | we have to make an appointment, between Monday the twenty-fourth of May, and Friday the fourth of June. | so I need to know something about your agenda, to see when we can meet during that period of time.
This section has presented the main findings for the most frequent relations in the turn-by-turn analysis, together with examples in both languages. Now we will turn to the findings of the analysis of whole conversations, and to how those two types of analysis compare with each other.
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.29 (133)
Rhetorical relations in dialogue
. Results: Conversation-as-a-whole analysis As mentioned in Section 5.4, analyzing a whole conversation in rhetorical relations entails a certain detachment from the conversation as interaction, as process, towards conversation as finished product. Mann and Thompson (1988) state that one of the main assumptions underlying RST is that “[t]hese relations, which can be described functionally in terms of the purposes of the writer and the writer’s assumptions about the reader, reflect the writer’s opinions for organizing and presenting the concepts.” In the previous section I already studied the effect one speaker wanted to produce on the hearer at that time, and how the hearer, in his or her turn, proceeded in the same manner. This second type of analysis does not pay so much attention to the effects achieved, but to the cohesiveness of the text as a whole, overlooking the fact that there was a target hearer or reader for these texts. In such a situation the procedure was to scan the conversation for the main purpose of the interaction, and from that macro-relation proceed with the analysis of the smaller segments. The span or spans that represented the main point in the interaction were usually the ones containing the proposal for a meeting. Once the problem was expressed, the rest of the conversation dealt with its solution, that is, with finding an appropriate time for holding a meeting. This main point is the Comprehensive Locus of Effect, the “portion of the text that represents the essence of the text as a whole” (Mann et al. 1992: 61). Obviously, at the lower levels of analysis, the relations holding were the same as those found for the turn-by-turn analysis. At the higher level, however, there were relations holding between the turns uttered by the different speakers. Segmentation was not always based on turns (i.e., a segment of text did not always start and end with a turn), although the turn boundaries were frequently a natural breaking point. The decisions taken in the analysis of turns and the analysis of the conversation as a whole influenced each other, because decisions previously taken (e.g., for the internal structure of a turn) biased the analyst’s decision when the conversation was considered as a whole. In order to avoid or at least balance this influence, 15 conversations in each language were first analyzed as whole conversations, and then broken down in turns. Then the other 15 were analyzed in the reverse order. I present here two analyses of full conversations, in English and in Spanish. Example (110) shows a complete sample conversation in English, and its RST analysis is represented in Figure 5.4. The Spanish example can be found after the discussion of the phenomena present in this conversation. The spans in the figures are represented between square brackets in the conversations.
2
4
5
7
Concession
Figure 5.4 Rhetorical relations in Example (110)
1
4–5 6 Elaboration
3 1–2 Purpose
4–6
Concession
1–3 Enablement
Solutionhood
1–6
Solutionhood
10
1–10 11
8
7–8
9
Non-Volitional Result
7–9
1–9
Solutionhood
1–11 Evaluation 12
13
15 16
15–16
17
18
19
17–19 Elaboration Elaboration
15–19 Solutionhood
12–19
Contrast
14
13–14 Condition
13–19 Elaboration
Solutionhood
Elaboration
TSL[v.20020404] Prn:7/10/2004; 9:16 F: PB12905.tex / p.30 (134)
Chapter 5
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.31 (135)
Rhetorical relations in dialogue
In the figures, I have illustrated a complete RST tree for each conversation. The trees show how the conversations can be analyzed as a full text, created by the two interlocutors. The discussion points at which relations are produced across turns, and which are within one speaker’s turn. (110) [FJBT_MSRH_11] fjbt_11_01: [1] maybe we should get together some time, [2] and talk about this a little longer. [3] do you have anything free in your schedule, /uh/ from, day after tomorrow on? msrh_11_02: [4] well I’ll be on vacation the, /uh/ second, third, and fourth. [5] I’m taking the family to Cancun. [6] /um/ how ’bout the seventh. or the ninth. or maybe even the fourteenth or seventeenth. fjbt_11_03: [7] well it looks like I’m out of town, on the seventh and the ninth, [8] but the fourteenth is open, [9] /uh/ what time suits you then. msrh_11_04: [10] how ’bout two o’clock in the afternoon. fjbt_11_05: [11] two o’clock in the afternoon sounds fine. [12] where would you like to meet. msrh_11_06: [13] we’ll meet at your office [14] if you don’t mind. fjbt_11_07: [15] do you still know where it is from last time? [16] or is this something new. msrh_11_08: [18] nope, I still remember where it is. [19] I’ll see you there. fjbt_11_09: [20] see you there,
The Comprehensive Locus of Effect in this dialogue is a relation of Solutionhood: speaker FJBT asks from speaker MSRH whether he has free time from “day after tomorrow on”, in span 3. The problem-posing span is the satellite in a relation that includes 1–3, ultimately solved in span 10, and approved (through an Evaluation) in span 11. The conversation, then, can be divided into two clear segments: in the first part of the conversation, the speakers agree that they need to meet, and set out to find a suitable date (spans 1–11). Once a suitable date has been accepted, details on that date are brought forward (spans 12–19). The top-level relation is one of Elaboration. Let us concentrate on the first part of the conversation, spans 1 to 11. As mentioned, the most important span is the question asked in span 3. Spans 1 and 2 explain why an appointment is needed, with span 3 enabling that appointment. Between 1 and 2, the relation is one of Purpose: the speakers need to get together so that they can talk further. Notice that the Purpose relation is signaled by a very general and, which could also signal a number of other
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.32 (136)
Chapter 5
relations. 1–3 are the satellite of the Solutionhood relation, partially solved (answered) in 4–6. In his turn, speaker MSRH explains that he will be away, and then proposes a new date. The date proposal is interpreted as a Concession relation between spans 4–5 and 6: “although I’ll be away on the second, third and fourth, I can meet the seventh, ninth, or maybe even the fourteenth or seventeenth”. Since the problem of meeting has not been solved yet, the conversation up to this point (spans 1–6) becomes another satellite in a Solutionhood relation, joining another proposal with the answer, in spans 7–9. In this turn, speaker FJBT encounters an obstacle to the meeting, but finds an open time. As a result of that, she is free on the fourteenth. She presents that as a result of her schedule, and asks further about the date in span 9. Span 9 was considered a Result, although another judge made it the nucleus of an Interpretation relation.5 Once it was considered a Result, the problem was determining whether it was Volitional or Non-Volitional. Although the difference is not crucial to the analysis, Non-Volitional Result seemed the most plausible interpretation, given that the speaker seems to have just found that open slot. We reach now the two most important spans in the conversation: in span 10, speaker MSRH proposes two o’clock in the afternoon, making this the nucleus of the overall Solutionhood relation. In span 11, speaker FJBT accepts that date, by providing a positive evaluation (“sounds good”). The rest of the conversation can move on to sorting out the details of the meeting. The second part of the conversation, spans 12–19, stands in an Elaboration relation to the first part. The two parts can also be interpreted as being a Sequence. Part of the background knowledge and knowledge of the generic structure the speakers have is that after a date has been proposed and agreed upon, other details need to be discussed. Sequence could define that knowledge of task structure. But Elaboration was chosen, because it is more specific to what is happening in the conversation. The structure of the second part is very similar to that of the first: questions are asked, and thus the question-answer pairs are represented in Solutionhood relations. Span 12 is the question that moves the conversation on to details. The rest of this second part can be divided in two spans, related by Elaboration. 13–14 propose speaker’s FJBT’s office (with a Condition), and spans 15–19 elaborate on that proposal. First of all, speaker FJBT asks whether her interlocutor remembers the office location, creating a Solutionhood. Speaker MSRH answers that he remembers (span 17), and further elaborates on that answer (spans 18 and 19). Spans 15 and 16 are in a Contrast relation, that be-
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.33 (137)
Rhetorical relations in dialogue
ing the only plausible relation in that case. As in other situations, there seems to be a need for an Alternative relation, in this case signaled by the conjunction or. Finally, span 20 is not part of the analysis, and considered an element of the holistic structure of the conversation. It could be considered a Restatement of the previous “see you there”, but the turn change, and the fact that it serves to close the conversation, make it more than a mere restatement. A comparison of the above analysis with Spanish is provided in Example (111), a dialogue whose corresponding RST analysis is shown in Figure 5.5, and discussed below. (111) [FSMA_MENF_05] fsma_05_01: [1] hola Edu. [2] mirá, [3] /eh/ qué te parece si arreglamos para la semana que viene, /eh/ bueno /eh/ reunirnos un par de horas. [4] /eh/ yo tengo libre el /mm/ lunes veintiséis, después de las doce del mediodía. [5] qué te parece? menf_05_02: [6] no. [7] mirá, [8] yo el lunes no puedo. [9] /eh/ el martes tampoco, [10] el miércoles puedo a la tarde, [11] y el jueves a la mañana. [12] porqué no me decís qué te parece <en> entonces miércoles a la tarde o jueves a la mañana. fsma_05_03: [13] mirá. [14] imposible porque el /eh/ miércoles yo tengo de dos a cuatro y media una reunión, [15] y el jueves a la mañana tengo una conferencia de nueve a doce. [16] así que tendría que ser recién, el miércoles de la semana de [yain (ya_en)] agosto, el miércoles cuatro. | [17] porque yo voy a estar a fuera el treinta, hasta el tres. [18] bueno. [19] recién sería el miercoles cuatro, después de la una del medio día, o a lo mejor antes de las once de la mañana. menf_05_04: [20] /eh/ mirá. [21] el miércoles cuatro, por la tarde, yo puedo. [22] así que, si te parece, [23] quedamos el miércoles cuatro, de dos a cuatro de la tarde. fsma_05_05: [24] bueno, [25] dále. [26] entonces nos vemos el miércoles, /eh/ /mm/ entre las dos y las cuatro. [27] [tá (está)] bien. [28] chau. TRANSLATION: fsma_05_01: [1] hello Edu. [2] look, [3] uh what do you think if we settle for next week, uh well to <see each other> uh meet for a couple of hours. [4] uh I’m free on Monday the twenty-sixth, after twelve noon. [5] what do you think? menf_05_02: [6] no. [7] look, [8] I can’t on Monday. [9] uh on Tuesday, neither [10] on Wednesday I can in the afternoon, [11] and on Thursday in the morning. [12] why don’t you tell me what you think then Wednesday in the afternoon or Thursday in the morning.
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.34 (138)
Chapter 5
fsma_05_03: [13] look. [14] impossible because on uh Wednesday I have from two to four thirty a meeting, [15] and on Thursday in the morning I have a talk from nine to twelve. [16] so it would have to be then, on Wednesday in the week of already August, Wednesday the fourth. [17] because I’m going to be out of town on the thirtieth, until the third. [18] well. [19] then it would be Wednesday the fourth, after one in the afternoon, or maybe before eleven in the morning. menf_05_04: [20] look. [21] on Wednesday the fourth, in the afternoon, I can. [22] so, if you want, [23] we can meet on Wednesday the fourth, from two to four in the afternoon. fsma_05_05: [24] good, [25] okay. [26] then we meet on Wednesday, uh mm between two and four. [27] that’s good. [28] bye.
The Spanish dialogue shares many of the structural characteristics we already observed in the English example. It is also built around a Solutionhood relation, with spans 1 to 5 as the satellite and the rest of the dialogue for nucleus. The problem statement includes a Condition relation: span number 5 is the satellite which literally means “what do you think?”, and which I interpreted to be a condition on the proposal presented by the current speaker. A paraphrase could be “I’d like to meet on this date, if you want to/are available”. Spans 4 represents an Enablement, the available date to meet, for the proposal of meeting presented in span 3. Spans 1 and 2 were not considered for the RST analysis, because they are interactional. Our next focus of attention is the main body of the Solutionhood relation, the nucleus in spans 6–29. These are again divided into two main parts which stand in a Solutionhood relation to each other. Spans 6 to 12 are the satellite, presenting another question on when to meet, and spans 13 to 29 constitute the answer to that question. I will describe the relations within those two groups of spans separately. The satellite of the Solutionhood relation is one utterance of speaker MENF, where he presents the constraints on his calendar by means of a Concession relation. He is being collaborative and, as a consequence, although he expresses his unavailability on the date speaker FSMA proposed, he contributes some other options. The satellite of the Concession relation – spans 6 to 9 – consists of an Elaboration on the “no” in spans 8 and 9 which is presented as a joint relation, two nuclei expressing the two dates when he is not available. Span 7 is not considered for the RST analysis. The nucleus of this Concession is presented as a Volitional Result at the top level with another embedded Joint relation. This time the two Joint nuclei present the two free slots of the
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.35 (139)
Rhetorical relations in dialogue
Solutionhood 6–29
1–5 Condition
Solutionhood
5
3–4
6–12
Enablement 3
13–29
Concession 6–9
4
10–12
Elaboration
Volitional Result
7–9
6
10–11 Joint
9
8
12 Joint
10
11
Volitional Result 13–21
22–24
Non-Volitional Result
13–15 N-V Cause 13
14
Elaboration 22–25
16–21 N-V Cause 15
Elaboration Elaboration
Evaluation 16–19
16–18
19
Non-Volitional Cause 16
22–23
21
Restatement
17
Figure 5.5 Rhetorical relations in Example (111)
26
Condition 22
23
24
25
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.36 (140)
Chapter 5
speaker. As a result of this availability, the speaker would like to know what his interlocutor thinks of those dates. The Volitional Result is marked by entonces. Now we turn to explore the structure of the nucleus in the second Solutionhood relation, spans 13 to 29. The first relation we encounter is one of Volitional Result, used again to express the idea that the speaker would like to meet, once the calendars have been described. The nucleus in this relation goes from span 13 to 21 and extends across speakers’ turns, starting at FSMA’s turn (the third turn in the conversation) and continuing onto the fourth turn, that of speaker MENF. This stretch starts with the presentation of a problem: speaker FSMA cannot meet on either of the days proposed. There are two Non-Volitional Causes that build the background for the unavailability. The Non-Volitional Result of this is that they will have to start considering a week in August for their meeting. The satellite of this Volitional Result includes an Evaluation (satellite in span 21), a Restatement (satellite in span 19), and a Non-Volitional Cause (satellite in span 17). The satellite in the Volitional Result relation – spans 22 to 24 – is the acceptance and restatement of the proposed date. As we saw in the English conversation, this last part of the interaction displays a completely different structure, where no Concessions or relations of Cause are present, only Elaboration and Evaluation of the date proposed. Table 5.3 shows the number and percentages found in the analysis of the conversations as a whole. As can be seen from the table, the most frequent relation is still Elaboration, but this time two other relations are prominent in their increase: Evaluation and Solutionhood. There is also a slight increase in the Restatement relation. I proceed now to explain the occurrence of the relations that increased with respect to the turn-by-turn analysis and provide some examples. .. Evaluation In the Evaluation relation, the reader/hearer recognizes that the situation presented in the satellite assesses the situation presented in the nucleus and recognizes the value it assigns. In my corpus this is mostly an inter-turn relation because the evaluations refer to what the other speaker said previously, and rarely to what the speaker is currently saying. This relation is never signaled by a discourse marker in the corpus. Example (112) below exemplifies the evaluation relation in the second turn, that of speaker FACR, responding and evaluating the proposal put forth by speaker FACR.
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.37 (141)
Rhetorical relations in dialogue
Table 5.3 Number and percentages of the analysis of whole conversations English Antithesis Background Circumstance Concession Contrast Condition Elaboration Enablement Evaluation Evidence Interpretation Joint Justify Motivation Non-Volitional Cause Non-Volitional Result Otherwise Purpose Restatement Sequence Solutionhood Summary Volitional Cause Volitional Result n
Spanish
T
%
T
%
1 14 4 71 7 67 175 11 82 0 1 29 34 3 40 45 4 10 24 7 121 4 4 17 775
0.13 1.81 0.52 9.16 0.90 8.64 22.58 1.42 10.58 – 0.13 3.74 4.39 0.39 5.16 5.81 0.52 1.29 3.10 0.90 15.61 0.52 0.52 2.19
1 18 7 70 8 89 140 4 67 4 11 26 22 0 54 30 4 26 71 20 126 7 11 49 865
0.12 2.08 0.81 8.09 0.92 10.29 16.18 0.46 7.75 0.46 1.27 3.01 2.54 – 6.24 3.47 0.46 3.01 8.21 2.31 14.57 0.81 1.27 5.66
(112) [FACR_FJYK_AU] fjyk_au_04: okay nine thirty in the morning, until, eleven thirty. | and then perhaps we can /um/ get some lunch afterwards. | how ’bout, you come over to my office. and then, perhaps we can go over to the conference room that’s right next to my office. facr_au_05: that would be fine on August eighth at nine thirty, ...
The Spanish example below, (113), contains a combination of Restatements and Evaluations in reaction to the previous agreement to meet. (113) [FNBA_FCBA_04] fcba_04_05: perfecto. | entonces nos reunimos | qué tal <en> en algún restaurante local de doce a una, | y luego nada más necesito re-
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.38 (142)
Chapter 5
unirme contigo un par de horas. | qué tal de una a tres. | porque yo sé que tu tiempo es muy valuable. fnba_04_06: sí, | para mí eso está muy bien | y así podemos hacer eso, | y ya. | entonces hasta el jueves, | el tres es jueves. | allí nos vemos. perfect. | then we meet | how about at some local restaurant from twelve to one, | and then I only need to meet with you for a couple of hours. | how about from one to three. | because I know that your time is very valuable. yes, | for me that’s very good | and that way we can do that, | and already. | then see you Thursday, | the third is Thursday. | I’ll see you there.
.. Solutionhood Mann and Thompson (1988) define Solutionhood as a situation, expressed in the nucleus, that presents a solution to the problem stated in the satellite. The terms problem and solution are broad; they cover questions, requests (including requests for information), expressions of needs, and conditions that carry negative values. I have considered the question-answer pairs to be in a Solutionhood relation. Question and answer here are understood, in most cases, as more than one clause, not in the strict adjacency pair sense of two turns containing one clause or sentence each (Sacks et al. 1974). A problem might be expressed in a long stretch of talk with embedded relations within it, and the same applies for the solution. As with Evaluation, Solutionhood is not signaled through discourse markers, although the most clear marking is the presence of an interrogative clause in the satellite. (114) [MGCT_MYAW_1] myaw_1_02: oh, yeah that sounds fine. | /um/, according to my schedule I’m, going to be free all day on, Wednesday the seventeenth. | are you, free any time then? mgct_1_03: /um/, seems like I have a seminar? from nine to four thirty? | so, do you wanna like meet [lat (at)], five o’clock then? at that time?
Some of the examples are candidates for a multiple analysis, as in Example (115) below, which could be interpreted as a Solutionhood or Evaluation relation. In general, I decided that those relations expressing an acceptance of a proposal were Solutionhood. Evaluation was assigned to those that elaborated
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.39 (143)
Rhetorical relations in dialogue
on the proposal, with some qualification of the speaker’s attitude towards the proposal. (115) [MAKK_MKGD_10] makk_10_05: /um/, how’s /uh/, nine o’clock then? mkgd_10_06: nine is fine. | /um/, I’ll see you then?
In Spanish, Solutionhood is only present as an inter-turn relation, never within a turn. As in English, it is never signaled by discourse markers, only by the presence of an interrogative clause. (116) is a typical example. (116) [FVGC_FSNM_09] fsnm_09_06: qué dices. | que tienes entre las *pause* diez y las dos libre? fvgc_09_07: sí. | tengo clase de dos a cinco. | así que entre las diez y las dos, o a la una y media, estoy libre. what did you say. | that you are free between ten and two? yes. | I have a class from two to five. | so that between ten and two, or at one thirty, I’m free.
.. Restatement Although there are intra-turn Restatements, some of them refer to something mentioned in the previous turn. In English we find only four of these inter-turn Restatements, which are never signaled by any discourse marker. The following is one example of those. (117) [FRET_MRCT_1] mrct_1_06: okay, what do you say we meet at, /glottal/ my office at, /glottal/ two? | is that good for you? fret_1_07: that’s fine, | I’ll see you the sixteenth, at two.
Unlike in English, the number of inter-turn Restatements in Spanish is very high compared to the intra-turn relations. If in English the number increased only from 24 to 28, in Spanish the intra-turn relations number 47 and the inter-turn ones increase to 71. In Spanish, speakers repeat much more often the previous utterances, sometimes even repeating two or three times the same information at the end of a conversation, in a closure sequence of sorts. These repetitions are also marked, with 21 out of the 71 relations marked with entonces (representing 29.58% of the total number of Restatements). This is indeed one of the most significant differences found in the analysis of the English and Spanish data. In the two examples below, the first one displays a
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.40 (144)
Chapter 5
Restatement of the previous turn. The second example contains three turns, the second a restatement of the first, and the third a restatement of the second. (118) [FYMM_FKAS_02] fymm_02_06: |s|okay. | está bien el martes treinta. | a partir de las tres de la tarde, estoy yo desocupada también. fkas_02_07: entonces <nos> nos reunimos a las tres, hasta las cinco de la tarde, martes. | hasta luego. okay. | it’s fine Tuesday the thirtieth. | from three in the afternoon on, I am free too. then <we> we meet at three, until five in the afternoon, Tuesday. | see you. (119) [MBRP_MFJR_O5] mbrp_05_09: ya | yo creo que el miércoles cuatro de agosto, estaría excelente, luego de la una | porque yo tengo una reunión de once a una. | así que qué tal si almorzamos juntos, y y conversamos. no? | así es que tú puedes venir a mi oficina, | y nos vamos a almorzar. | qué te parece? mfjr_05_10: perfecto. | entonces nos vemos el miércoles cuatro de agosto. | chau |s|Brent. mbrp_05_11: |s|okay |s|Freddy. | entonces en eso quedamos. | nos vemos. | chau chau. I see. | I think that on Wednesday August fourth, it would be great, after one | because I have a meeting from eleven to one. | so how about we have lunch together, and we talk. no? | so you can come to my office, | and we go for lunch. | what do you think? perfect. | then we meet Wednesday August fourth. | bye Brent. okay Freddy. | then that’s settled. | see you. | bye bye.
There are some other instances of inter-turn relations, such as Volitional or Non-Volitional Result, but above I have discussed the most relevant ones. In Chapter 7, we will see how we can relate the presence of certain rhetorical relations to the generic structure of the conversations. The next section discusses the presence of discourse markers.
. Discourse markers The analysis of discourse markers is part of the more general analysis of discourse coherence – how speakers and hearers jointly integrate forms, meaning, and actions to make overall sense out of what is said. (Schiffrin 1987: 49)
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.41 (145)
Rhetorical relations in dialogue
Schiffrin provides some informal definitions of underlying relations between parts of talk, such as support, position, contrasting, interpretation, etc. These relations are formalized in RST, and in this section we will explore how rhetorical relations are signaled by discourse markers. The analysis presented here is a formalization of Schiffrin’s intuitions on markers as exponents of internal coherence. We know that rhetorical relations between stretches of text or talk are present, whether signaled or not, as part of the organization of a coherent text that has a purpose in the culture where it is produced. This analysis of 60 English and Spanish conversations shows that particular genres display structure and meaning, which are preserved even without discourse markers. However, discourse markers provide further information on the type of relation that holds between two pieces of text. Research has shown that discourse markers guide the grounding that is taking place when speakers interact. At the beginning of the dialogue, the speakers assume a set of background assumptions and mutual knowledge. This is the common ground on which speakers start (Clark and Schaefer 1989). As the conversation proceeds, the speakers build up more common knowledge from the information communicated. New information conveyed by one speaker is usually acknowledged as received by the other speaker, thus entering the common ground. This process is what Clark and Schaefer (1989) call grounding. Grounding is a collaborative process: the listener sends signals indicating understanding, and the speaker watches for those signals. If not present he or she usually rephrases, repeats or performs other comprehension checks. Clark and Wilkes-Gibbs (1986) demonstrated that task-oriented dialogues typically require stronger evidence of grounding than casual discourse. Taskoriented conversation is a collaborative process in which progress is made only when the two participants share the same beliefs. Byron and Heeman (1998) explored one way in which that grounding takes place in task-oriented dialogues, and discovered that discourse markers guide the interpretation of utterances to come. Thus, an unexpected second part in an adjacency pair will usually be preceded by a discourse marker, to signal the unexpected character of the utterance. I will discuss in this section how discourse markers signal the presence of rhetorical relations in the conversation, probably helping in the understanding of the relations between parts of the text. Schiffrin (1987: 52) poses another problem in the study of discourse markers as exponents of internal structure. This problem is related to the amount of discourse to be included within the scope of a discourse marker, i.e., how much of the preceding or following speech is affected by the use of the dis-
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.42 (146)
Chapter 5
Figure 5.6 Rhetorical relations in Example (120)
course marker. Again, in RST this is already solved within the theory, since it is hierarchical: the marker will affect some other part of the discourse insofar as it is a part of the relation in hierarchical terms. In other words, a marker will have scope over the spans constituting a same schema when it is found within one of them; it will have indirect scope over the higher and lower levels of that hierarchy; and only marginal scope over the sister relations of that schema. For an exemplification of this, let us observe Example (120) and its representation in Figure 5.6. This example shows the relation Volitional Cause, composed of two spans of text, numbers 2 and 3, and signaled by because, as part of the nucleus for another relation. The other relation is a Volitional Result, signaled by so, whose nucleus is spans 2 and 3 and whose satellite is span 4. The scope of so is everything in those three spans, but the scope of because is only spans 2 and 3, since the Volitional Cause is embedded within the Volitional Result.
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.43 (147)
Rhetorical relations in dialogue
Elaboration
1
1–2
3–4
Sequence (and then)
Sequence (and then)
2
3
4
Figure 5.7 Rhetorical relations in Example (121)
(120) [FCKA_FFMW_10] fcka_10_03: [1] well, /uh/ you’re available, after, /um/ the seventeenth? in the week, <seventeenth,> /eh/ seventeenth, the eighteenth, the nineteenth, the twentieth, the twenty first? | [2] well, the seventeenth and the twentieth would be the most convenient for me | [3] because I have, nothing planned for these two days. | [4] so, what do you think.
A different case is exemplified in (121), the analysis of which is presented in Figure 5.7. This stretch of talk contains two sister relations, and two parallel sequence relations, both marked by and then, where there is no scoping from one to the other. They are both daughters of the elaboration relation, but that is where their coincidence stops, as the two sequence relations and their markers are independent of each other. (121) [FACR_FJYK_AU] fjyk_au_04: [1] okay nine thirty in the morning, until, eleven thirty. | [2] and then perhaps we can /um/ get some lunch afterwards. | [3] how ’bout, you come over to my office. [4] and then, perhaps we can go over to the conference room that’s right next to my office.
I have not attempted to provide a formal definition of discourse markers, or to have a closed inventory. Knott and Dale (1994), for instance, provide an extensive inventory of cue phrases, which they intend to be a motivation for building a definitive set of relations (see Section 5.1 on the number of relations in a theory of rhetorical structure). My approach proceeds in the inverse direction:
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.44 (148)
Chapter 5
given the set of relations, and given the relations I found, I examined whether that particular relation was somehow marked by one of these devices. The decision on which discourse markers to include in the study was a difficult one. Choices can range from the strict classification, including only conjunctions, to the most inclusive one. An example of the latter is what Grimes calls a Pesky Little Particle: “Most languages have particles whose use seems to be related to gluing the parts of discourses together but which are never easy to pin down” (Grimes 1975: 93). Martín Zorraquino (1988) also calls for a classification that does not necessarily follow grammar or grammatical categories (conjunction, adverb, etc.). Portolés (1998) proposes a definition where the main characteristic of discourse markers is that they guide inferences in communication.6 Even the name for the particles or markers is at issue. Jucker and Ziv (1998), in an introduction to a volume on such phenomena, use the term discourse marker because it seems to be able to encompass a number of different elements. Aijmer (2002) uses discourse particles, and Brinton (1996) pragmatic markers. I will use the term discourse marker, although I considered only a subset of all the possible discourse markers. The study has focused on some specific markers that could be described as discourse connectives. I have not considered pause fillers, such as I mean or you know, because those realize many different functions, and they do not always relate two spans of talk as tightly as other markers do. Well is excluded for the same reasons: it is present in a variety of contexts, and its consideration might very well require a study of its own. Among its functions we can include the signaling of a dispreferred second part in an adjacency pair (Byron and Heeman 1998), and the preparation of communication, in the form of a filled pause. I have, then, taken into consideration discourse connectives such as and, but, or. I have also considered discourse markers that are usually labeled conjunctions, or markers of subordination (because, so, if ), and their Spanish counterparts in both cases: y, pero, o; porque, así que, si, etc. Most sentence adverbials (nevertheless, also, on the other hand) are not included. Webber et al. (2003) suggest that they do not contribute to establishing a structural connection between elements of discourse, but an anaphoric one. Table 5.4 displays the occurrences of discourse markers in the corpus. The analysis of discourse markers is based on the turn-by-turn analysis. That is, the association of a discourse marker with a particular relation in the table refers to the presence of that relation in the turn-by-turn analysis. None of these markers was found to mark relations in the higher-level analysis. For instance, a Solutionhood relation across a turn is marked by the presence of the interrog-
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.45 (149)
Rhetorical relations in dialogue
Table 5.4 Occurrence of discourse markers English T Antithesis Background Circumstance Concession Contrast Condition Elaboration Enablement Evaluation Evidence Interpretation Joint Justify Motivation Non-Volitional Cause Non-Volitional Result Otherwise Purpose Restatement Sequence Solutionhood Summary Volitional Cause Volitional Result n
Spanish %
T
%
0 1 0 38 2 33 2 0 0 0 1 16 3 0 23 27 2 9 3 4 0 1 1 14
– 7.69 – 53.52 22.22 50.00 1.20 – – – 100.00 55.17 8.82 – 62.16 62.79 50.00 90.00 10.71 57.14 – 25.00 14.29 82.35
1 0 5 38 6 24 9 0 0 0 11 11 17 0 31 21 4 25 17 17 0 1 11 48
100.00 – 55.56 54.29 75.00 26.97 6.43 – – – 100.00 42.31 77.27 – 57.41 72.41 100.00 96.15 36.17 85.00 – 16.67 100.00 100.00
179
30.86
301
45.88
ative mood, but mood is not considered in this taxonomy of markers. Some other connectives, such as I mean, that is or well, can signal Restatements or Evaluations across turns, but again only conjunctions were included for this study. Therefore, the analysis of discourse markers is to be understood within the analysis of turns in isolation. The percentages represent the number of times the relations are marked in comparison to the number of occurrences of that relation. That is, Condition in English is marked 50% of the time it occurs in the corpus. Roughly the same relations in both languages are marked, and they are signaled to a similar extent. The highest numbers of markers correspond to Concession, Condition, Cause and Result (both Volitional and Non-Volitional). The most significant differences in the way the two languages mark the relations are in Justify, Vo-
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.46 (150)
Chapter 5
litional Cause and Volitional Result. These three relations are marked much more frequently in Spanish. In the last two the overall occurrence of the relation is also higher in Spanish. Justify seems to be marked much more often in Spanish. In general, Spanish uses discourse markers to a higher extent (45.88% of the relations) than English does (30.86%). We can also see that relations that are never, or rarely marked, are the same in both languages. The relations that are never marked are: Enablement, Evaluation, and Solutionhood. Rarely marked are: Antithesis, Background, and Summary. This replicates Knott and Dale’s results, who provide a study of a large corpus where Background, Evaluation, and Elaboration are never marked (Knott and Dale 1994). Moreover, the fact that less than half the relations are marked overall leads us to think that any kind of identification based on lexical discourse markers is bound to fail in identifying all relations in a text. Let us just look at one of the most straightforward cases of signaling, Purpose. This relation is marked around 90% of the times it appears in both languages, slightly more so in Spanish. Even when there is a discourse marker, there are four different markers in each language (and, so that, that and to in English; así, para, para que and y in Spanish). In addition, relations are sometimes underspecified, that is, the connective that is used to link them does not accurately point to the relation. Spooren (1997) discusses underspecification, illustrated with an example similar to (122). In the example, the link after points to a temporal relation, but we could more likely consider this a causal relation (i.e., he was fired because he had stolen paper clips). In the most extreme instances of underspecification, the relation remains totally implicit, as we have seen in some relations in the corpus. Spooren points to a Gricean principle at work in these cases, which could be phrased as “say no more than necessary”. But we could also think the intention here is to create an effect. Example (122) is probably more effective because of what it implies than because of what it actually says. (122) Bill was fired after he had confessed to stealing paper clips from the supply room.
Our last problem in this respect lies in the fact that there are few safe correlations between relation and marker. In English, and is used to mark Elaboration, Joint, Non-Volitional Cause, Purpose and Sequence. So is present in Background, Condition, Justify, Non-Volitional Cause, Non-Volitional Result, Restatement, Summary and Volitional Result. Tables 5.6 and 5.7 show the discourse markers found in the corpus, and the relation in which they were found, broken down according to the language.
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.47 (151)
Rhetorical relations in dialogue
Table 5.5 Equivalences of numbers and relations in Tables 5.8 and 5.9 1 – Antithesis 2 – Background 3 – Circumstantial 4 – Concession 5 – Contrast
6 – Condition 7 – Elaboration 8 – Interpretation 9 – Joint 10 – Justify
11 – Non-Volitional Cause 12 – Non-Volitional Result 13 – Otherwise 14 – Purpose 15 – Restatement
16 – Sequence 17 – Summary 18 – Volitional Cause 19 – Volitional Result
Table 5.6 Discourse markers in the English corpus 2
4 5
6 7 8
9 10 11 12 13 14 15 16 17 18 19
and 2 16 and then and then if 1 because but 34 1 but otherwise ‘cause depending 1 except for 1 however 2 if 24 if so 2 if. . . then in that or 1 otherwise 1 since so 1 1 so that that then though 2 to unless 1 whether 1 which 1 TOTAL 1 38 2 33 2 1 16
1 1
3
2 2
2
1 11 1 1
4
1 1 1
1
4 1 1 24
3
1
1 2 1 3
3 23 27
2
9
3
4
T
1
26 3 1 12 35 1 1 6 1 1 2 24 1 2 1 2 1 5 12 44 1 2 1 2 3 1 1 1 1 14 180
Table 5.5 provides the corresponding name for the number of the relation used at the top of both Tables 5.6 and 5.7. (Table 5.5 was necessary to keep the other two tables at a manageable size.) Tables 5.6 and 5.7 are to be read as follows: if we look at the marker and, for instance, we can see that there are two of those markers for relation number 7. According to Table 5.5, that relation is
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.48 (152)
Chapter 5
Table 5.7 Discourse markers in the Spanish corpus 1 3
4 5
6 7
8
9 10 11 12 13 14 15 16 17 18 19
a menos que 1 además 1 ahora 1 así 1 2 así que 9 como te decía cuando 2 de todas formas 1 depende que 1 digo entonces 1 5 es que 1 luego o 6 1 o sea 1 o sea que 8 2 para para que pero 1 35 1 pero inclusive 1 por ejemplo 1 por lo que 1 por lo tanto 1 por tanto 1 porque 12 28 pues 1 2 que recién sea que 1 si 20 si no so (Eng.) 1 1 tal forma que y 1 5 9 1 2 y después y entonces 1 y si no 2 y ya ya que 1 TOTAL 1 5 38 6 24 9 11 11 17 31 21
1
2 23
2 1
1 8
1
10
1
4
2
21 2
1 10 1 1
2
1 1 1 1
1
13 2 1
2
1 5
2
4 25 17 17
1 1 1 15 48
T 1 1 1 6 34 1 2 1 1 1 25 1 1 7 1 16 21 2 37 1 1 1 2 1 52 4 1 1 2 21 1 2 1 39 2 2 4 1 2 301
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.49 (153)
Rhetorical relations in dialogue
Elaboration. Further on in the same row, there are 16 instances of and in relation number 9, which again according to Table 5.5 is the Joint relation. There is one occurrence of and in Non-Volitional Cause, relation number 11; three instances in Purpose, number 14; two instances in Sequence, number 16; and under 19, which is Volitional Result, we find two instances. The total number of occurrences of and in the corpus is 26. The last line in Tables 5.6 and 5.7 is the sum total of markers for the relation in that column. In Table 5.6, which contains the English markers, we can see that there are no markers for relation number 1, Antithesis, and only one marker in the whole corpus for relation number 2, Background.
. Summary This chapter has taken us one step further in the description of task-oriented dialogue. After we observed the patterns in the surface organization of information in the previous chapter, we turned to look at the realization of underlying information through rhetorical relations. The description of choice was Rhetorical Structure Theory (RST). I first concentrated on a description of both RST and its predecessors. Since RST had been predominantly used for the description of written text, I explained some of the issues that arise in its application to spoken language. The main body of the study is divided into two types of analysis: the analysis of individual turns and the analysis of whole conversations. Results show, first of all, that spoken language can be subjected to a rhetorical structure analysis, and that both English and Spanish conversants tend to resort to the same relations, with some small differences in their frequency of use. The next section (Section 5.7) studied the correlations between relations and discourse markers, presenting the quantitative results of the analysis. The results provide information on which relations are most typically marked. In addition, I listed all the discourse markers that appeared in the corpus, showing how they are present in particular relations. No significant differences were found between the two languages. The results show that a taxonomy of discourse relations based solely on the presence of discourse markers will ignore a good number of relations that are never or rarely signaled by a discourse marker. This study of rhetorical relations has applications in at least two different areas: Second Language Teaching and Computational Linguistics. By having a complete account of discourse markers and how they are used in the organi-
TSL[v.20020404] Prn:7/10/2004; 9:16
F: PB12905.tex / p.50 (154)
Chapter 5
zation of discourse, we can help learners of both languages understand how native speakers structure their discourse, and how they make use of discourse markers. Schleppegrell (1996) points out that ESL learners and native speakers of English use a particular discourse marker (because) for different discourse functions. An account like the one just presented will help define the contexts of use for each rhetorical relation and each discourse marker. Another possible application can be found in Computational Linguistics. Indeed, RST has been used in a number of text processing systems, most notably in Natural Language Generation and Text Summarization. Marcu (2000) created an algorithm for rhetorical parsing of texts. The results of such parsing are applied to text summarization, by constructing a summary with the nuclei of the relations found in the text. In Natural Language Generation, concepts from a Knowledge Base are presented into a coherent text built using RST schemas (Cawsey 1990; Corston-Oliver 1998; Moore and Paris 1993; O’Donnell et al. 2001; Scott and de Souza 1990; Vander Linden et al. 1992). Most importantly, RST can be used as a neutral representation of text to be generated into different languages (Bouayad-Agha 2000; Kosseim and Lapalme 2000), or into different media (Bateman et al. 2001; Rocchi and Zancanaro 2003).
Notes . Longacre (1983) traces back the idea of a “taxonomy of the deep structure of interclausal relations” to his own work, together with colleagues (Ballard et al. 1971a; Ballard et al. 1971b; Longacre 1976), and also to the work of Beekman and Callow (1974), Grimes (1975) and, ultimately, Fuller (1959). . Members of the Discourse Studies group at Tilburg University have used their own theory of coherence relations in the analysis of Dutch (Oversteegen 1997; Sanders 1997; Spooren 1997). . Bernárdez (1995) provides a summary of the theory and translation of the relation definitions into Spanish, but no corpus analysis. . Martin (1992: 262) discusses the RST view of text as product. . In order to check for consistency, some of the English conversations were analyzed by a second analyst. We did separate analyses, and then we checked them against each other. In most cases, we had similar analyses, or we agreed on one plausible interpretation. My thanks to Dennis Storoshenko for his help with this analysis. . Inferences in communication understood within the framework of Relevance Theory (Sperber and Wilson 1995), and also in Argumentation Theory (Anscombre and Ducrot 1983). Portolés (1988) discusses the relation between discourse markers and Argumentation theory.
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.1 (155)
Chapter 6
Cohesion in dialogue
This chapter presents the results of an analysis of cohesion in the corpus. My approach to cohesion is based on Halliday and Hasan’s 1976 work, which is summarized in Section 6.1. Section 6.2 provides the results of the analysis in terms of types of cohesion, distances and types of cohesive chains. This is the last chapter of the three devoted to the analysis of texture, and thus completes our examination of how coherence and cohesion are created interactively.
. A brief introduction to cohesion On an ordinary, day-to-day level, people equate cohesion with the simple and sustaining fact that some sentence sequences make sense and others do not. Cohesion elevates a random collection of sentences to the status of a text, and in the process imparts meaning, insight and purpose to those sentences. Without cohesion, the text can hardly be said to exist at all, for cohesion provides the textual means for initiating comprehension or sense. Since that description attributes to cohesion the survival of civilization, the maintenance of friends, lovers, marriage, and peace, not to mention the successful preparation of a Stouffer’s spinach souffle or a simple paragraph, some clarification from the term, as distinct from the traditional terms, unity and coherence, seems appropriate. (Markels 1981: 3)
This section will provide such clarification. In the definitions of cohesion, the approach taken here is that of Halliday and Hasan, from their 1976 and 1985 work, plus other work by Hasan (1984a). For Spanish, additional accounts such as those of Mederos Martín (1988) and Casado Velarde (1997) were used. According to Halliday and Hasan, the property of being a text is called texture. Texture is what distinguishes a text from a non-text. Texture is derived from the fact that the text functions as a unity with respect to its environment. But if texture is an abstract entity, a property, how can we distinguish a text that has texture from one that does not? Texture is realized in relations existing between parts of a text. Let us look at one example.
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.2 (156)
Chapter 6
(123) [FACR_FJYK_AU] facr_au_01: ... would you like to meet possibly, between the seventh to the tenth? | anytime during those days would be fine.
In this example, those days refers to between the seventh to the tenth. There is a relation between those two phrases that makes the two sentences become a text, because they hang together as one unit. This relation is a cohesive relation, and the pair of related items is a cohesive tie. We must note that no single item is cohesive in itself. Although I will be referring to those days, and categorizing it as a particular type of cohesion, we should always bear in mind that cohesiveness is established through the relation between the two items, not by one item in isolation. The meaning of this relation is that the two items refer to the same thing – they share corefentiality. Identity of reference, as we shall see, is not the only type of cohesive relation. There also exist relations of similarity. But in a more general sense, cohesion occurs when the interpretation of some element in the discourse depends on the interpretation of another one, whether preceding or following. In our example above, those days presupposes between the seventh to the tenth. We need to refer to the latter in order to resolve the presupposition. The fact that it is resolved successfully establishes the cohesive relation between the two clauses. There are, in fact, two different types of tie between those days and its presupposed element. Those establishes a relation of demonstrative reference, whereas days is a general word that subsumes specific instances of days. We will see, in a later section, what types of cohesive relations can be established, and how those enter into chains that make the text hang together. But first, we will examine the relations above and below cohesion. .. Texture and structure, coherence and cohesion A text is not a structural unit in the same sense that a sentence is. The relations that hold the parts of a text together are not similar in nature to those holding a sentence together. In generative terms, a sentence can be generated through a number of phrase structure rules. Even if we had the same type of rules to generate a text – something like van Dijk’s (1980) macrostructures – we could not account for the properties of a text. Halliday and Hasan believe that any unit that is structured is cohesive in itself. In general, any unit which is structured hangs together so as to form text. All grammatical units – sentences, clauses, groups, words – are internally ‘cohe-
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.3 (157)
Cohesion in dialogue
sive’ simply because they are structured. The same applies to the phonological units, the tone group, foot and syllable. Structure is one means of expressing texture. (Halliday and Hasan 1976: 7)
Texts, however, extend beyond the confines of structure for Halliday and Hasan: “a text typically extends beyond the range of structural relations, as these are normally conceived of. But texts cohere; so cohesion within a text – texture – depends on something other than structure” (Halliday and Hasan 1976: 7). There are certainly some structural relations in a text, what Halliday and Hasan term discourse structure. The structure of adjacency pairs (Sacks et al. 1974), the hierarchy of structure found in classroom exchanges (Sinclair and Coulthard 1975), and the modelling I propose in Chapter 7 are all based on the idea that a text is structured in particular ways. Stoddard (1991) also proposes a definition of texture that includes organization of the material in the text and rhetorical structure. On the other hand, there are relations between the parts of a text that are not exclusively structural relations, and that is where cohesion plays a role. Cohesion is a non-structural text-forming relation that works alongside the structural relations to provide texture (Halliday and Hasan 1976, Hoey 1991, Leckie-Tarry 1995). If cohesion is not concerned with structural relations at the sentence/clause level, we might be tempted to affirm that all cohesive relations are between sentences. This is not always true, since there are cohesive relations within the sentence. In Example (124), my office establishes a link with we. The possession relation indicated by my is related to the speaker, one of the components of the pronoun we. (124) [FACR_FJYK_AU] fjyk_au_04: ... and then, perhaps we can go over to the conference room that’s right next to my office....
Although these intra-sentence relations do exist, they are not directly responsible for the hanging together of the sentence – structural relations are. In the description of a text, it is the intersentence cohesion that is significant, because that represents the variable aspect of cohesion, distinguishing one text from another. But this should not obscure the fact that cohesion is not, strictly speaking, a relation ‘above the sentence’. It is a relation to which the sentence, or any other form of grammatical structure, is simply irrelevant. (Halliday and Hasan 1976: 9)
It is for this reason that we pay more attention to the cohesive ties outside the sentence. In my case, I would like to rephrase that to consider cohesive ties only
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.4 (158)
Chapter 6
outside the clause. The reason is that the clause in spoken language has more independence than in the written language. Some instances of subordinate or main clauses can stand on their own. The clause is also the basic unit for the other two types of analysis in this study (thematic and rhetorical relations). I have mentioned that a cohesive tie is established between a presupposed and a presupposing element. When the presupposed element is to be found in the text, the type of reference is endophoric, that is, it happens inside the text. But the presupposed element can also be found somewhere else, in which case we have an exophoric reference, one that takes place outside the text. The reference can be made to the context of situation, the context of culture or to the context we share as humankind. In (125) below, the plant, you, me and last Christmas all refer to the context of situation that the speakers share in that particular instance. (125) The plant you gave me last Christmas is dying.
In (126), this president has a referent in the world that will be instantiated according to the context of culture or of the situation. For example, one referent in the context of Spain, at the time of writing, could be José Luis Rodríguez Zapatero, the president of the Spanish Government. I might also just be referring to a different context of situation, for instance, to Simon Fraser University, in which case the name of the president on that same date is Michael Stevenson. Equally, the previous one is dependent on the particular context in which the utterance is used. (126) This president is taller than the previous one.
Finally, in (127), most speakers will interpret the sun as the one that provides our planet with light and heat. However, in a different context, the sun could be resolved to the 60 Watt lightbulb that I used to represent the sun in my toy model of the world. (127) The sun will eventually burn out.
The hanging together of a text with relation to its context of situation or culture is called coherence. The above utterances will only be coherent if they can be successfully interpreted in particular contexts. Hasan clarifies that her use of the term coherence is fairly close to its everyday meaning. “I used the word to refer to the property of ‘unity,’ of ‘hanging together.’ By this definition, any object is coherent to the extent that its parts hang together” (Hasan 1984a: 181). In general, coherence is reserved for the hanging together of the text with something else outside itself, whereas cohesion is the internal hanging together of
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.5 (159)
Cohesion in dialogue
the text. The relations of texts to the context of situation and culture were first identified by Malinowski (1923) and then extended by Hymes (1972), Halliday and collaborators (1964), and by Gregory and Carroll (1978). See also Chapter 2 for a more detailed description. Although some researchers, for instance Cha (1985), attempt to bring the two together in a unified theory, coherence and cohesion are best studied separately first, to then consider their contribution to the creation of a text. Thus, this chapter is devoted to the study of cohesion. Its interaction with generic structure is pursued in Section 7.6. .. Types of cohesive relations Cohesion is part of the system of a language, and like other semantic relations, it is expressed through the stratal organization of language. Language is organized around three different levels of coding, with each level being realized (coded) in the one immediately below it. The semantic level, or stratum, represents the meanings in the language, and it is realized through the lexicogrammatical stratum, which includes the forms of the language. The lexicogrammar is in turn realized in the expressions of the language – phonological and orthographic systems. This is represented in Figure 6.1, adapted from Halliday and Hasan (1976: 5), where the downward pointing arrow means “is realized in”. The lexicogrammatical system, or wording, comprises both the grammar and the vocabulary. In the Hallidayan tradition the relationship between grammar and lexicon is one of delicacy or depth: the lexicon represents the most delicate choices in the system. The more general meanings are expressed through the grammar, whereas the more specific ones are expressed through the vocabulary of the language. Cohesion follows this pattern as well, with some meanings expressed through the grammatical system, and some through the lexical
Figure 6.1 The three levels of coding in the language
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.6 (160)
Chapter 6
one. Grammatical cohesion – the cohesion expressed through the grammatical system – is divided into the reference, substitution, ellipsis and conjunction types. Lexical cohesion – which is part of the lexical system – is also broken down into relations of repetition, synonymity, collocation, and other semantic relationships.
Grammatical cohesion The types of ties established under grammatical cohesion are all resources found in the grammar of the language. They enter into cohesive relations only when they relate to some other item outside the clause where they occur. Reference is, with exceptions in the demonstrative adverbs, part of the system of determiners in the noun phrase.1 The referring item is one that cannot be interpreted semantically on its own, but needs to find its resolution somewhere else. We have already seen how the pointing can be exophoric, that is, to the context of the text. We will be concerned here only with endophoric reference, the type that is solved with reference to some other element in the text itself. In Example (128), ahí is resolved with reference to mi oficina. Note that the two might be considered as being in the same sentence, but not in the same clause, which I established as my unit of analysis. (128) [FPDV_FSNM_04] fpdv_04_08: ... qué te parece si te presentas en mi oficina alrededor de las once? | y de ahí salimos a almorzar y conversamos todo lo que tenemos que conversar. ... how about you come to my office around eleven? | and from there we go out for lunch and we talk about everything that we need to talk about.
Endophoric reference is of two types: anaphora, the one that refers to the preceding text, and cataphora, which refers to the text that is to follow. The example in (128) is an instance of anaphora. Cataphora, which is much less common in the corpus, is represented in Example (129). The speaker uses here to refer to something she is saying in the following text, namely a list of her available times. (129) [FJMB_FMLZ_6] fmlz_6_08: ... here are the times I actually have. | before one on the twenty seventh, before twelve, on the twenty eighth, *pause* /uh/ after twelve on the twenty ninth. *pause* and after two on the thirtieth. | those, are the only two hour slots I have. ...
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.7 (161)
Cohesion in dialogue
Reference can be further classified into the personal, demonstrative and comparative types. Personal reference includes personal pronouns (I, you, he, she, we...; me, him, her, us...), possessive determiners (my, your, his, her, our...) and possessive pronouns (mine, yours, his, hers, ours...). These will only be considered cohesive when they link to some other element in the text. Thus, (130) contains some exophorically cohesive elements, such as contigo, yo, para tí. Those refer to the speakers, and would be resolved in the context of situation: the referent of contigo (‘with you’) is the listener, whereas the referent of yo (‘I’) is the speaker. These elements were not included in the analysis. (130) [FJGC_FKAS_04] fkas_04_01: hola Jeannette. | me gustaría reunirme contigo /uh/ entre el veinticuatro de mayo y el cuatro. | <de viernes> /eh/ yo tengo libre este lunes el veinticuatro desde el mediodía hasta <el [c(inco)]> <el [c(inco)]> /uh/ las cinco de la tarde. | qué tal para ti reunirnos, el veinticuatro. hello Jeannette. | I’d like to meet with you uh between the twenty-fourth of May and the fourth. | uh I’m free this Monday the twentyfourth from noon until uh five in the afternoon. | how is it for you to meet, the twenty-fourth.
In Example (131), however, the it in what’d it cost you refers to lunch in the previous clause; it is therefore endophoric, and considered for the analysis. (131) [MBMS_MLAB_10] mlab_10_06: yeah, I don’t know, Ben, | the, last time we let you pay, you took me to /uh/ that dump, for lunch, | what’d it cost you, about, /uh/ three fifty, /lg/ for both of us, | I don’t know. | I think, I’m picking the restaurant, this time, downtown,...
Demonstrative reference uses determiners (either as modifiers or as heads) and adverbs, such as this, these, that, those, here, there, then, to point to other items in the text. Example (128) above shows an adverb (ahí) with demonstrative reference. Comparative reference establishes relations of identity of similarity with the use of adjectives and adverbs: same, identical, better, more, less, etc. Ties of comparative type were scarce in the corpus; none were found in English and only a few in Spanish. In Example (132), así (‘like that’, ‘that way’) refers to the whole arrangement that has been previously discussed.
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.8 (162)
Chapter 6
(132) [FMGL_FMCS_01] fmcs_01_11: bueno. | /mm/ entonces quedamos así. ... okay. | mm then it is settled like that ...
The next type of grammatical cohesion is substitution. Substitution refers to a previous element in the text through the use of a substitute term. Reference is distinguished from substitution in that it is a relation between meanings, which implies an identity in the meaning or ultimate referent for the terms entering into the relation. Substitution, on the other hand, is a relation between linguistic items: one linguistic item is used to substitute and point to another linguistic item, not to its referent. Substitution comes in three flavours: nominal, verbal or clausal, depending on the item being substituted. In (133) below, one is a substitute term for meeting, an example of nominal substitution. (133) [MAEM_MJAY_1] maem_1_01: okay. Jules. /um/ thanks for the meeting, | let’s start the next one,
One or ones are the terms most commonly used for nominal substitution in English. Verbal substitution is realized through an auxiliary verb (do, be, have), sometimes together with another substitute term such as so or the same. Example (134) shows the substitution of looks pretty good in the first clause with so does in the second one. The next example, (135) is one of clausal substitution, where so substitutes the previous clause. The terms used in clausal substitution are so and not. (134) [FECE_MNFH_10] fece_10_01: ... /ah/ Thursday the sixth looks pretty good, and, so does Monday the tenth. | how ’bout for you. (135) [FECE_FPAM_3] fpam_3_04: do you think we’ll need an hour? | if so, how ’bout, the twenty sixth, three to four?
The examples above are all from the English corpus, since the Spanish conversations did not contain a single example of substitution. I will discuss this in detail in the next section, but it is my belief that the Spanish speakers do not use substitution because they tend to favour ellipsis. Ellipsis is a special instance of substitution, in that it involves substitution by zero. Instead of one of the lexical items mentioned for substitution, no item is used, and the hearer/listener is left to fill in the gap where the substitute item, or the original item, should have appeared. Since it is a special case of substi-
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.9 (163)
Cohesion in dialogue
tution, the subtypes for ellipsis are the same: nominal, verbal and clausal. In (136), the speaker omits cita when he says me gustaría mucho que [la cita] fuera el miércoles. This is an instance of pro-drop, where the Subject (in Spanish) is dropped if it is known from context. In (136), the translation uses parenthesis around the pronoun representing the pro-drop. (136) [MEJH_MJJG_01] mjjg_01_01: bueno, querido profeta /eh/ tenemos que hacer una cita. | así que este te pongo de tu conocimiento, que <se> a mí me gustaría mucho que fuera el miércoles, diecisiete. ... well, <dear> dear <what’s your name,> prophet uh we have to make an appointment. | so I’ll let you know, that I would like it very much if (it) were on Wednesday, the seventeenth. ...
Example (137) below is an example of verbal ellipsis. The speaker does not repeat the previous Verb (and its complement) when she says we can spend more than two hours if we need to [spend more than two hours]. The full version sounds contrived; the fact that we are describing ellipsis as a phenomenon does not mean that it is extraordinary. In fact, ellipsis is probably the default case, more so in spoken language. (137) [FSJB_FKDO_1] fkdo_1_04: ... and then, if we /um/ need to run over, we might both have enough time that we can spend /uh/, more than two hours if we need to. | how’s that sound.
Finally, (138) represents clausal ellipsis occurring across turns. Speaker FJAB finishes her turn with a mere otherwise yes which omits a full clause of agreement. I have provided examples of cohesion within turns, in order to avoid lengthy examples. However, most instances of cohesion occur across turns, as we will see when we study the average length of distance between ties. (138) [FCAD_FJAB_12] fcad_12_05: ... how ’bout, say, that lunch thing again, | you can buy me lunch again at like, twelve to two? | that sounds good. fjab_12_06: that’s great, | but, I think we might have to cut a little short this time, because I have a class, from two to three thirty. | but, otherwise, yes. /lg/
The last instance of grammatical cohesion is conjunction. Conjunction is represented in a number of conjunctive elements that are not cohesive in and of
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.10 (164)
Chapter 6
themselves. Rather, the meanings of conjunctive elements establish relationships among other parts of the text. Conjunctive elements are words such as and, but, yet, so, then, therefore, however, etc. Conjunction links parts of the text (clauses) which are not linked through any other structural relation. More specifically, conjunctive elements signal the presence of a conjunctive relation, such as succession in time, causal, adversative, etc. Conjunctive relations are, mutatis mutandis, the same type of relations I described in Chapter 5 as rhetorical relations. The discourse markers discussed in that chapter can also be considered conjunctive elements. The main difference is probably that Halliday and Hasan restrict conjunctive elements to those appearing at the beginning of the clause. They also include continuatives in their model (now, well, anyway...) and intonation, both of which I excluded in my study. Smith and Frawley (1983) reported the low use of conjunctive cohesion in four different genres, as compared to other types of cohesion. Important as a complete analysis might be, and since a whole chapter has already been devoted to rhetorical relations and discourse markers, I decided not to include conjunction in this analysis of cohesion.
Lexical cohesion Lexical cohesion contributes to the cohesion of a text through the selection of vocabulary. The repetition (exact reproduction) or reiteration (related word) of a lexical item in the text produces a cohesive effect. Lexical cohesion does not entail identity of referent; two items in related semantic fields can also enter in a cohesive tie. The two turns in (139) are cohesive because speaker FCLD repeats fourteenth and the fifteenth as uttered by the previous speaker. (139) [FAMS_FCLD_AU] fams_au_03: ... so, how does, /um/ the morning of the fourteenth, or the fifteenth sound? fcld_au_04: I have a meeting on the fourteenth and the fifteenth. | but, I’m completely free on Wednesday the twenty third, if that would work out with you,
Lexical cohesion is divided into different types, according to the relationship between the terms. The first type is one where the same item is repeated, as in the example above. For the purposes of the analysis, I established a further subcategorization, a distinction between exact repetitions, and repetitions where there was some minor reformulation.
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.11 (165)
Cohesion in dialogue
A second type of lexical cohesion includes synonyms or near synonyms. Halliday and Hasan (1976: 338) include hyponyms here, but I decided to make two different categories for those. Same item and synonym relations are at the same level of specificity. At different levels are the superordinate and subordinate types (the latter being my addition). On a higher level of specificity, there are relations where the cohesion is done through a ‘general’ item. Such is the case in (140), where opciones is a general term to cover the two dates proposed. (140) [MRBZ_MCRA_03] mcra_03_02: ... y el martes tengo de doce a tres. o de tres en adelante. o el miércoles a cualquier hora, excepto entre diez y doce. | qué te parecen esas dos opciones? ... ... and on Tuesday I have from twelve to three. or from three on. or on Wednesday anytime, except between ten and twelve. | what do you think of those two options? ...
The last category in lexical cohesion is collocation. Collocation is achieved through the association of lexical items that regularly co-occur. In the corpus, many items were cohesive simply because they all had to do with meetings, agendas, calendars and availability. In Example (141), a lunch meeting establishes a cohesive tie with food because the two are related in the same semantic field. (141) [MBMS_MLAB_10] mbms_10_05: /ah/ how ’bout we do another lunch meeting, /uh/ eleven to one, | /uh/ what do you want for lunch this time for food, Italian or Polish. /lg/
Mederos Martín (1988: 127ff.) criticizes Halliday and Hasan’s approach to lexical cohesion as part of a system of anaphora.2 In his view, there is no need to refer back to an antecedent in order to resolve most instances of lexical cohesion. The referent resolution is often postulated as a defining characteristic of anaphora, and of cohesion in general. However, he includes conjunction (‘connectivity’) as a type of cohesion, even though connectives, or conjunctive elements, do not refer to any element in the previous text either. The general role of cohesion is to provide the thread against the warp of the text. Part of that thread consists of relations between elements that are related through the lexical part of the language, rather than through the grammatical part. The term anaphora might be left to describe relations of the grammatical cohesion type, where a given cohesive link is only interpretable as a reference to something
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.12 (166)
Chapter 6 Grammatical R. Reference R1. Personal R2. Demonstrative R3. Comparative
Lexical L. Lexical cohesion Reiteration L1. Same item L1a. Exact L1b. Rephrased
S. Substitution S1. Nominal S2. Verbal S3. Clausal
L2. Synonym L3. Superordinate L4. Superordinate L5. General Word
E. Ellipsis E1. Nominal E2. Verbal E3. Clausal
L6. Collocation
Figure 6.2 Summary of cohesion types
else in the text. However, lexical cohesion is legitimately a part of cohesion, as a different type of relation that makes the text hang together. Before we move on to discuss cohesive chains, I will provide a graphic summary of the types of cohesion discussed above. Figure 6.2 represents the types discussed, with the labels that will be used in the description of the analysis. Figure 6.2 is based on Halliday and Hasan’s (1976) labels, with the exception of conjunctive cohesion, not included for this part of the analysis, and thus not present in the figure. .. Distance of cohesion and cohesive chains Most of the examples introduced in this section display cohesive ties that are very close together in the text. That is not always the case, and we find numerous instances where the two terms are far apart. Halliday and Hasan (1976) devised a system for categorizing the distance between two cohesive ties. The second element might come immediately after the first one, thus making the distance immediate. It might have any number of intervening clauses, and then it is called remote, with a number identifying the number of clauses between the two. Or it might be mediated, that is, the tie is established with a previous element through other elements in the text. Finally, a tie can be cataphoric: the element in one of the cohesive types described above could refer to an element that is mentioned in the discourse to come. These relations are coded as follows.
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.13 (167)
Cohesion in dialogue
– – – –
Immediate – I Remote [number of intervening clauses] – R[n] Mediated [number of intervening clauses – M[n] Cataphoric – K
The concept of mediated ties brings us to the next important concept in cohesion, that of cohesive chains. If two ties are mediated by a third intervening one, then the three of them enter into a cohesive chain. In Example (142), one of the cohesive chains running through the two turns relates to food. The speakers have decided on a lunch meeting, and thus restaurante, comida and estomaguito can all be related to this topic. The chain begins earlier on in the conversation, but here we can see part of it with seven links (underlined). If there was only one preceding item, the chain would then be eight links long. All of the links are instances of lexical cohesion: repetition of the same item and collocation. Most texts contain more than one chain. In the example there is another chain about meeting times. (142) [FMEM_MEOC_02] fmem_02_04: /um/ |s|okay, | <está> acordamos entonces que sea martes, | me parece mejor a las doce y treinta, | así yo termino clases a las doce, | y alcanzo a llegar al restaurante a las doce y treinta. | *pause* /mm/ lo que me parece más bien que definamos es a dónde quieres ir? | tú dime qué tipo de comida quieres para el martes. | o si tu religión te impide /begin_laugh/ algún tipo de comida especial | entonces evitamos ese tipo /end_laugh/ de restaurantes. meoc_02_05: correcto Mónica. | nos vemos entonces ese día a las doce y treinta. | /eh/ como te decía yo pasaré con |s|Len, | y no tengo ningún problema con la comida. | como buen pobre y colombiano que soy *pause* tengo acceso a cualquier /begin_laugh/ tipo de comida /end_laugh/. | mi estomaguito me lo permite. | gracias mija. um okay, | so we agree that it’s Tuesday, | I’d like it better at twelve thirty, | that way I finish my classes at twelve, | and I’ll manage to get to the restaurant at twelve thirty. | mm what I’d like is for us to define is where you want to go? | you tell me <what> what kind of food you want for Tuesday. | or whether your religion forbids any type of special food | then we’ll avoid that type of restaurant. correct Mónica. | we’ll meet then on that day at twelve thirty. | uh as I was saying I’ll come with Len, | and I have no problem with food. | being a good poor and Colombian as I am I have access to any type of food. | my stomach allows it. | thank you dear.
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.14 (168)
Chapter 6
Cohesive chains and chain interaction are the most interesting constructs in describing cohesion and, ultimately, coherence in a text. Many scholars have pointed out that cohesion and coherence are not all-or-nothing categories, but a matter of degree. Parsons (1996) states that, in any given text type, there is a gradation dependent on the extent to which a text relies on cohesion to provide coherence. He takes as starting point the work of Hasan (Hasan 1984a; Halliday and Hasan 1985) in the development of the concept of cohesive harmony. Hasan defined a chain as a set of items, each of which is related to the others by semantic relations of co-reference, co-classification and co-extension. Coreferenciality relates two or more items in a cohesive chain because they have the same referent. Co-classification ties have a different referent, but a relationship of belonging to the same class. Co-extension ties belong to what Hasan terms the same “general field of meaning”. These three types of relationships allow us to classify chains as belonging to one of two types: identity and similarity chains. In an identity chain, the members are related by co-reference: every member of the chain refers to the same thing, event, etc. Identity chains are established in the text. In a similarity chain, the relations are of co-classification or co-extension. These chains contain items referring to non-identical members of the same class of things, events, etc., or to members of related classes of those things and events. Because similarity chains build on our knowledge of vocabulary and of the world, they are established outside the text. Example (142) above contains a similarity chain, where the items relating to food and restaurants are in a relation of co-extension. Chains do not usually occur in isolation, but alongside other chains. However, the mere presence of two or more chains in a text does not guarantee a cohesive effect. To use Hasan’s example (Halliday and Hasan 1985: 91), reproduced below in (143), the fact that 100% of a text’s items enter into cohesive chains (as it happens with this text) does not necessarily make it cohesive. (143) girls bananas two spend shopkeeper apples own girls dollars grapes buy fifty sell cents shopkeeper girls fruit
Indeed, although the chains contribute to cohesion in a text, they need to be related to each other somehow. This relationship is called chain interaction. The relationships are mostly grammatical, in the Transitivity structure. If we reorder the previous example, we see the Actor–Patient–Instrument relationships established among the chains that relate to the girl, the fruit and the money.
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.15 (169)
Cohesion in dialogue
(144) The girl had two dollars to spend on fruit. She decided to buy bananas from the shopkeeper in the corner. She still had fifty cents left, so she bought a couple of apples. She was so happy that she sang all the way back home.
Hasan establishes a minimum requirement for chain interaction: at least two members of one chain should stand in the same relation to two members of another chain. For a better definition of the interactions, she divides the tokens in a text into three categories: –
Relevant tokens: All tokens that enter into identity or similarity chains, further divided into: – Central tokens: relevant tokens that interact (girl, she, fruit, apples, bananas, dollars, cents in (144)). – Non-central tokens: relevant tokens that do not interact (shopkeeper).
–
Peripheral tokens: Tokens that do not enter into any kind of chain (sang).
We are, finally, in a position to define cohesive harmony, which is the function of three phenomena: 1. Low proportion of peripheral tokens to the relevant ones. 2. High proportion of central tokens to non-central ones. 3. Few breaks in the interaction. Hasan affirms that coherence is a function of cohesive harmony. Our perception that a text is coherent, that it somehow makes sense, is dependent on its cohesive harmony. In the next section we will examine the apparently low cohesive harmony of the dialogues, given their low chain interaction.
. Cohesion in scheduling dialogues This section describes the results of the analysis, and a comparison of English and Spanish. I describe the types of cohesion found, the direction and distance of reference and the nature of cohesive chains. .. Cohesion types In the corpus, I identified the types of cohesion described above (see Figure 6.2 on page 166). Each instance in Table 6.1 is to be interpreted as the relationship
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.16 (170)
Chapter 6
of one cohesive item to the item preceding it, that is, as one link. The frequencies are also displayed in Figure 6.3, where the legend for the x axis refers to the types of cohesion in Table 6.1. Note that the figures are the overall numbers in the corpus.
Figure 6.3 Cohesion types, graphical representation Table 6.1 Cohesion types in the corpus English
Spanish
T
%
T
%
5 108 0
1.08 23.28 –
7 49 3
1.13 7.90 0.48
S1 – Substitution, nominal S2 – Substitution, verbal S3 – Substitution, clausal
1 1 2
0.22 0.22 0.43
0 0 0
E1 – Ellipsis, nominal E2 – Ellipsis, verbal E3 – Ellipsis, clausal
8 8 4
1.72 1.72 0.86
19 15 5
3.06 2.42 0.81
L1a – Lexical, same, identical L1b – Lexical, same, rephrased L2 – Lexical, synonym L3 – Lexical, superordinate L4 – Lexical, subordinate L5 – Lexical, general word L6 – Lexical, collocation
141 67 23 9 58 8 21
30.39 14.44 4.96 1.94 12.50 1.72 4.53
274 43 33 25 45 33 69
44.19 6.94 5.32 4.03 7.26 5.32 11.13
n, Total number of links
464
100.00
620
100.00
R1 – Reference, personal R2 – Reference, demonstrative R3 – Reference, comparative
– – –
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.17 (171)
Cohesion in dialogue
I will discuss first the results for each language, then focus on the comparison of the two. In English, by far, the resource used most often is lexical cohesion (70.48% of the links are lexical), and more specifically the repetition of the same item. I divided the repetition of same items into two different categories, one where the item is repeated verbatim, and another one where there is some difference in the tie. For instance, in (145), speaker MPMM repeats the date of the meeting exactly as the other speaker uttered it: the sixteenth of March. In Example (146), on the other hand, the speaker repeats himself, but adds the day of the week as extra information, so that the tie is established between on the eighteenth and on Tuesday the eighteenth. (145) [FRJP_MPMM_1] frjp_1_05: ... I have class from nine to twelve. but after that, I’m free on the sixteenth, of March. mpmm_1_06: /uh/ the sixteenth of March sounds good, ... (146) [FECE_MNFH_10] mnfh_10_02: ... and on the eighteenth /um/, it’s [actual(ly)] it looks pretty good. | I /ah/ have class from two to four. | so would you like to meet with me on Tuesday the eighteenth?
The reason for splitting repetition is that I had the impression that speakers tend to repeat exactly the same terms. We already saw how this affects the thematic progression in the dialogues (see Section 4.5), making most Themes non-derived. In the thematic progression section I mentioned that repetition included many instances of Subject pronouns. This is not the case in the analysis of cohesion, since the pronouns that refer to the speakers have exophoric reference, and are thus left out of the analysis. What the speakers are repeating so often, apart from Subject pronouns, are dates and times. Because the discussion of dates places a high burden on working memory, dates are tossed back and forth and repeated very often. Dates and times are proposed and abandoned constantly. The speakers need to make sure that they are discussing the same date, and thus they tend to repeat the date at the same time as they mention their availability or unavailability for that particular date. The same reason explains the high number in the L4 category, lexical subordinate items. The speakers start out with a two-week window, and then narrow down from that. Overall, lexical cohesion is the most frequently used device to establish cohesion in these conversations. After lexical cohesion, reference is preferred. Within reference, English showed no instances of the comparative, only a few of the personal, and a high percentage (23.28%) of the demonstrative type. Instances of personal reference
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.18 (172)
Chapter 6
are personal pronouns used to refer to a previous date. The demonstrative can be a modifier or head of a nominal group, sometimes accompanied by a general word, such as ‘days’ or ‘times’ (147). Another frequent instance of the demonstrative is the use of an adverb, such as then or there, to refer to the date or place agreed upon, as in Example (148). (147) [MAEM_MJAY_1] mjay_1_04: ... so let’s, shoot for, not this week, next week, fifteenth, sixteenth, seventeenth, or eighteenth. | /um/ any time good for you? on those, four days? (148) [FACR_FJYK_AU] facr_au_05: that would be fine on August eighth at nine thirty, | /um/ I would love to get some lunch as well afterwards, fjyk_au_06: okay, I’m looking forward to see you then. | bye bye.
The last two types of cohesion in order of frequency are substitution and ellipsis. They are both closely related, as we saw in the previous section, since ellipsis is substitution by zero. It is clear, from the table, that speakers do not favor these two types. My hypothesis is that they both place a heavy burden on the speakers’ minds. It takes extra effort to resolve elliptical references, and, to a certain extent, substitution as well. Hoey (1991: 6) points out that substitution is rare outside conversation, and possibly also rare even in conversation. The figures for the two are very small, making it difficult to draw comparisons. However, ellipsis is used more often. That is, speakers prefer to leave something unsaid than to use a substitute term for it. Now let us turn to the study of the Spanish data. Most of the phenomena described in English apply to Spanish: higher numbers in lexical cohesion and, within grammatical cohesion, preference for the reference type. There are a few differences worth noting. At first sight, Spanish seems to employ a higher number of ties (620 versus the 464 of the English corpus). Surprisingly enough, they are in exactly the same ratio per word as in English (Table 6.2).3 Both languages use exactly the same ratio of cohesive devices per word; they only use different proportions of each type of cohesive device, as we shall see. Table 6.2 Ratios of cohesive links to words in the corpus
Cohesive Links Words Ratio
English
Spanish
464 6804 0.068
620 9112 0.068
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.19 (173)
Cohesion in dialogue
The tendencies in Spanish show, again, a high preference for repetition of the same lexical item, whether in identical form or somewhat rephrased. This type of lexical cohesion is used more frequently than in English, with the identical repetition being the preferred type. The Spanish speakers used lexical cohesion 84.19% of the times a cohesive link was used, a slightly higher percentage than in English (70.48% for lexical cohesion). In reference and in ellipsis the percentages are also similar to the English numbers. The most intriguing category for Spanish was substitution. I could not find any instance of substitution in the Spanish corpus. As I noted above, the frequency of substitution is low in English, but in Spanish it was simply nonexistent in my data. Now let us ponder what substitution would look like in Spanish. Consider the following example. (149) Which book do you want? The red one. ¿Qué libro quieres? El rojo.
In English, the answer uses one as a nominal substitute for book. In Spanish, the noun is omitted, leaving us with a case of ellipsis. The category of rojo could be either adjective or nominalized adjective. But it is clear that there is no substitution. Even if we argue that rojo is a substitute for libro rojo, we then have a case of repetition, and thus lexical cohesion, not substitution. Nominal substitution seems to be ruled out for Spanish. In verbal substitution, English utilizes a number of elements to replace the Process or the Process plus Complements, mainly the auxiliary verb do or do so, as in (150), where do replaces know her.4 In the Spanish counterpart to this example, ellipsis is used instead, leaving the Subject pronoun yo to stand on its own. (150) Nobody knows her better than I do. Nadie la conoce mejor que yo.
This is not to be interpreted as a total absence of substitution in Spanish. The verb hacer, a rough equivalent of do, is used in some instances. In (151), hacer replaces invadir. (151) ¿Van a invadir Kosovo? Parece que eso es lo que van a hacer. ‘Will they invade Kosovo? It looks like that’s what they’re going to do.’
Mederos Martín (1988) discusses the status of substitution in Spanish at length. He argues for creating a new category, ‘proform anaphora’, where both reference and substitution would be included. He also discusses the situation of the verb hacer in Spanish as a proform (pp. 100–104). However, later on (p. 109),
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.20 (174)
Chapter 6
he includes it in the category of ‘general words’. The intuition in this case is that there is a very fine line between hacer as substitution and the lexical cohesion instance where a word is referred to with the class of general words. After all, hacer is not as grammaticalized in Spanish as do is in English, and we could very well account for (151) by taking it to be an example of lexical cohesion: the use of a general word to refer to a subordinate instance of the same class. Casado Velarde (1997) also conflates substitution and reference under a ‘substitution’ heading in his description of textual phenomena in Spanish. There he includes lexical proforms, among them the verb hacer, but also reference. Both Mederos Martín and Solé (1966) admit that hacer is not a substitute for all instances of verbs, only a reduced class of them. Bernárdez (1982: 105–108) also points to the character of hacer as a substitute for verbs of action. Bello (1984, paragraph 1093) considers it as a ‘verb of frequent use’ that replaces other verbs. Although do is not a general substitute in English either, its use as a function word seems to be more extended, in the formation of negatives and interrogatives. Finally, let us look at clausal substitution, where an entire clause is presupposed by an item, in English typically so and not. Example (152), adapted from (Halliday and Hasan 1976), substitutes the previous clause by so. A possible Spanish translation would use a demonstrative pronoun, eso, for the same purpose. Unless we interpret the class of demonstrative pronouns as doing double duty, that is, providing reference and substitution, this is an instance of reference. (152) Is there going to be an earthquake? They say so. ¿Va a haber un terremoto? Eso dicen.
In other instances of short answers, ellipsis again seems to be the alternative in Spanish, rather than substitution. Example (153) is again substitution in English but ellipsis in Spanish. One could argue that sí and no are the substitute terms for the previous clause. However, the sentences could be completed in Spanish without alterations (Creo que sí se han marchado; Espero que no se hayan marchado), which is not possible in English. This leads me to believe that the Spanish speakers are simply leaving the information out. (153) Did they leave? I think so. / I hope not. ¿Se han marchado? Creo que sí. / Espero que no.
These observations on the nature of cohesion in Spanish are, of course, preliminary. Since my corpus did not contain any instances, I can only speculate about what substitution could be like in Spanish. A larger corpus across different genres and modes would be needed.
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.21 (175)
Cohesion in dialogue
.. Distances: Types and length An interesting measure of cohesion is the type of distances between the links, and the length of their spread. Table 6.3 displays the four types of direction and distances, with the absolute numbers in the corpus. There is a clear preference for mediated types, because most links enter in long cohesive chains, where all items refer ultimately to the first item, mediated by the second. After mediated, the most frequent type is immediate. In the remote type, the speakers link a term with another one in a remote clause, which occurs less often in the corpus. Finally, the forward-looking link, the cataphoric link, is hardly present. The next table, Table 6.4, displays the averages of these types, and the length of the distance in the mediated and remote types per conversation. The table shows that, on average, a conversation will have mediated ties as the most likely type, followed by immediate ones, and then by remote. Cataphoric links are the least common. The tendencies are as described above: in each conversation, the most likely type of link will be a mediated one. In this type, the average distance between the link under consideration and the first referent – measured in clauses – is 3.41 and 4.25 clauses, for English and Spanish respectively. Thus, even though mediated is preferred over immediate, the length of intervening material in a Table 6.3 Direction and distances of links in the corpus English
Spanish
I – Immediate M – Mediated R – Remote K – Cataphoric
155 244 61 4
183 360 75 2
n, Number of links
464
620
Table 6.4 Average directions and distances per conversation
I – Immediate M – Mediated R – Remote K – Cataphoric M[n] – Average distance R[n] – Average distance
English
Spanish
33.45 51.21 14.61 0.72
30.92 56.71 12.07 0.30
3.41 1.16
4.25 1.79
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.22 (176)
Chapter 6
mediated link is not long: about four clauses. In the case of a remote link, the constraints on distance are more stringent: only one or two clauses on average will separate the two elements in a cohesive tie. Speakers avoid placing an excessive burden on each other’s working memories. Elements that are relevant – supposedly the ones in mediated chains – are repeated frequently. Even when there is a remote link, the distances are very close. This could be a characteristic of spoken language, since it is difficult to keep a large number of items in working memory at the same time – seven, plus or minus two, is the magical number (Miller 1956). In addition, this particular type of spoken language requires an extra effort from the participants, given the number of dates, times and places being passed from one interlocutor to the other one, since chains typically cross over turns. According to Walker (1993), redundancy in conversation is a result of memory limitations. Vigara Tauste (1995) offers a wider range of functions for what she terms recurrence in conversation. Comprehension checks, emphasis, reformulations and readjustements of what has just been said are, according to Vigara Tauste, some of the uses of repetition in conversation. Johnstone (1994) also examines some of the functions of repetition in discourse, among them the use of repetition as a way to deal with an interruption (what was being said before the interruption is repeated), and the use of repetition to maintain the floor while the speaker searches for something to say. .. Chains: Types and length The analysis of cohesive types in the corpus is interesting per se. It comes to have full importance only if we consider those links as they are integrated in cohesive chains running through the text. A cohesive chain is a series of links that are connected to each other by means of some of the relations already described. Most of the cohesive chains in the dialogues included days, times and events. In a few instances, they referred to speakers’ activities. Numerically, the English corpus contained a total number of 133 chains; the Spanish 120. The average numbers per conversation are presented in Table 6.5. Each conversation has, on average, about four chains, with the English chains showing a slightly higher number. However, these chains are longer in Spanish than they are in English. The most interesting aspect in the study of the chains running through the conversations is that they hardly interact with each other. The typical conversation structure is one where a chain is introduced, discussed, and abandoned.
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.23 (177)
Cohesion in dialogue
Table 6.5 Average number and length of chains
Average number of chains per conversation Average length of chains (number of links)
English
Spanish
4.43 4.99
4.00 6.56
After abandoning a chain, a new one is introduced, with little, if any, interaction with the first one. Chain interaction is, according to Hasan (Hasan 1984a; Halliday and Hasan 1985), a necessary component in cohesive harmony. The conversations would, then, show low cohesive harmony.5 If we follow the chains in most conversations, starting a new chain where the previous one finished, the graphical representation of the chains is a right-leaning slope. In order to illustrate this more clearly, I have included a conversation from the English corpus, in (154), which is represented in Figure 6.4. There are a few minor chains running through the text, which interact to a certain extent with the major ones. The interaction is shown through horizontal arrows in the figure. The main four chains, however, do not interact at all with each other. The links in the figure are underlined in the text. (154) [FHKR_FDMR_JA] fhkr_ja_01: Heather, | how’s your schedule for, the fourth. | I’m free all day. fdmr_ja_02: /hm/ the fourth, | that’s today isn’t it? fhkr_ja_03: okay, | never mind . | what about, /hm/ the fifth, like at noon. fdmr_ja_04: the fifth works for me, | I’m free all day. /lg/ {crosstalk} fhkr_ja_05: we have two hours between there, | well, let’s see | I have a lab, at two o’clock. | what about, Friday the sixth, | I’m free all day. fdmr_ja_06: /hm/ that’s kind of rough. | /lg/ /um/ unless we did it like, eight to ten. | then I’d have to rush to a meeting with Dave. | ’cause that’s at ten. fhkr_ja_07: okay. | /um/ let’s try the next week. | on, the ninth and the tenth, I can, meet you | from one thirty, on, I’m free. | how ’bout you. fdmr_ja_08: /lg/ {crosstalk} /oh/ /lg/ fhkr_ja_09: Heather, | I didn’t get that last bit for the ninth and the tenth. | /lg/ I just had blank. | can you repeat that? fdmr_ja_10: I’d love to repeat it | ’cause I messed up, anyways, | /lg/ /um/ the ninth, I can’t, | ’cause I have to meet Michael at the airport at three fifteen. | but the tenth, I just have to prepare a seminar. | so, what
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.24 (178)
Chapter 6
time’s best for you on the tenth. fhkr_ja_11: /um/ after, <we’ll see,> one fifteen to the rest of the day, I’m free, | so, how’s your schedule look for that. fdmr_ja_12: well, I can prepare that seminar all morning, | and, like later on in the evening, if need be so, how ’bout two o’clock. fhkr_ja_13: that’s fine, | I’ll see you January tenth, at two o’clock for, two hours, | and /lg/ I’ll just see you then. /lg/ fdmr_ja_14: okay, | later. /lg/
We could, in the face of these results, argue that these conversations do not exhibit cohesive harmony, given that the conditions for calculating it are not present. However, we are confident that the speakers left the experiment with
Figure 6.4 Cohesive chains in Example (154)
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.25 (179)
Cohesion in dialogue
the impression of having produced cohesive conversations. Indeed, the dialogues seem coherent and cohesive to any outsider. The answer is probably, then, that different types of texts require different measures of cohesive harmony. Fox (1987), in a study of anaphora, concludes that patterns of anaphora are local (specific to a particular genre) rather than global (specifiable for all of the language, English in her case). Stoddard (1991) measures cohesion by the cohesive ties generated by the reader, rather than through anything intrinsic in the text. With respect to the type of chain (identity or similarity), most of the chains found in the corpus are of identity, because the same element is repeated and referred to very often. This holds true even despite the low numbers of reference itself: the speakers achieve identity by mere repetition. The identity of the chains depends on the text and the context. The elements hold together much more closely because of the context. The identity of Tuesday with any date is only true in the context of the speakers’ calendars. Again, this is a characteristic of spoken discourse, since spoken discourse is not meant to be repeated or read again – except for, maybe, by the inquisitive linguist.
. Summary This chapter concludes my study of structural and non-structural components in the conversations. Cohesion represents the non-structural component, in Halliday and Hasan’s view (1976). For the study of the cohesive phenomena in the corpus I followed their formulation of cohesion in English. Some adaptation of that theory was necessary in the case of Spanish. The most notable application revealed the questionable presence of substitution in Spanish. The comparison of the number of cohesive links used for each language discovered exactly the same ratio (0.068) of cohesive elements to words in both English and Spanish. The types of cohesive elements used by the speakers were also very similar in both languages. Lexical cohesion, and specifically the repetition of the same item, is the most widely used type. Lexical cohesion was followed by the use of reference. In terms of direction of reference, both languages preferred anaphoric reference, with cataphoric reference occurring a very small percentage of the time. The anaphoras were most frequently mediated, that is, they entered into a relatively lengthy chain of elements. However, the links are often activated, making the distance between a link and its mediating element only slightly longer than four clauses. After mediated reference, in both languages, comes immediate
TSL[v.20020404] Prn:29/07/2004; 9:35
F: PB12906.tex / p.26 (180)
Chapter 6
reference. All of this shows that the speakers avoid establishing links between elements located far from each other. I argue that this is a characteristic of spoken language in general, and of this genre in particular. Finally, and with regard to chains, there were also striking similarities between the two languages. Each conversation contained about four chains (4.43 in English and 4.00 in Spanish, on average). Those chains rarely interacted with each other, because one was abandoned as soon as a new one was initiated. This holds true for major chains involving times and dates to meet. There were also minor chains throughout the dialogues, which showed slightly higher degrees of interaction. Although the Spanish conversations contained lower numbers of chains than the English, those chains were lengthier (an average of 4.99 links in English versus 6.56 in Spanish).
Notes . I will use the terms phrase and group interchangeably. . Also reporting on criticism by Huddleston (1978). . The similarity in ratio is probably related to the similarity in text length across the two languages. Ratios of cohesive elements per words increase as text length increases (Haswell 1988). . Note that, in English, there is an ellipsis alternative to the substitution, further complicated with the shift of case, from I to me: ‘Nobody knows her better than I/me’. . Cohesive harmony has as prerequisite the interaction of chains. Once interaction is established, a measure of central, non-central and peripheral tokens determines the harmony level. Given the low interaction, I did not carry out the full analysis.
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.1 (181)
Chapter 7
The generic structure of scheduling dialogues
The previous three chapters have examined the textual qualities of conversation from three different points of view: thematic, rhetorical and cohesive structure. In them, I have described how the participants in the conversations built coherent and cohesive conversations that can be analyzed as a text. The texts they produce jointly are instances of a genre (scheduling dialogues), which could be categorized within the class of task-oriented dialogues. As we saw in Chapter 2, once a genre has been defined through its purpose, it can be described in terms of stages. This chapter is devoted to identifying the stages in the conversations, by examining the lexicogrammatical patterns that characterize each stage. The genre analysis follows the steps proposed by Eggins and Slade (1997), described in Chapter 2, Section 2.3. Each type of analysis (thematic relations, rhetorical relations, and cohesion) is correlated to how it surfaces in the different stages of the conversations.
. Stages in scheduling dialogues The conversations can be divided, at a first level of organization, in three clear stages: Opening, Task Performance, and Closing. This tripartite organization has been found in most instances of spoken interaction (Stenström 1994), in telephone conversations (Schegloff and Sacks 1973), task-oriented conversation (Maier 1996a), including service encounters (Ventola 1987), in business meetings (Bargiela-Chiappini and Harris 1997), and in most speech events (Polanyi and Scha 1983) in general. Clark (1996: 37) writes that all joint activities have an entry, a body and an exit. Analysis of both the English and the Spanish dialogues used in this study revealed the organization of stages shown in Figure 7.1, applicable to every single dialogue under consideration. The sequence is represented in a finite state machine (Hopcroft and Ullman 1979), with each stage enclosed in a circle. The arrows on the left indicate the potential starting points (either Opening or Date Proposal), and the double circle around the Closing stage indicates the final
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.2 (182)
Chapter 7
Opening Task Performance Place Proposal Date Proposal
Closing
Figure 7.1 Sequencing of stages in scheduling dialogues
stage in the conversations. If a conversation starts with an Opening stage, then the speakers proceed on to a Date Proposal stage, which can be repeated. After one or more Date Proposal stages have been produced, the option is to go onto a Place Proposal, or to move directly to the Closing stage. All conversations contained a Closing stage, and all of them finished at this stage. Example (155) provides an illustration of a usual conversation flow, with delimited stages. The conversations starts with an Opening stage, in this case a simple greeting. It then proceeds to the Task-Performance stage. The TaskPerformance stage contains three different Date Proposals, plus one Place Proposal. Once the Task-Performance stage has been completed, the conversation ends with a Closing stage. The double vertical bars () represent the boundaries between stages. (155) [FCAE_FPAM_3] Opening stage: fcae_3_01: Patty Task-Performance stage: 1st Date Proposal: /um/ we need to get together to talk about /um/ this secretary project, this anything that floats race. | /um/, what’s a good time for you. fpam_3_02: well next week, I’m out, the twenty seventh through the twenty ninth, | I have some meetings on the twenty fifth and the twenty sixth, | /uh/ the following week I’m out on February first, and, some meetings the rest of the week. | does it have to be during the day?
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.3 (183)
The generic structure of scheduling dialogues
or would you prefer in the evening. fcae_3_03: /um/ probably, some time between nine and five, would be good, | don’t really look forward to, meeting over, business, after five o’clock, fpam_3_04: do you think we’ll need an hour? | if so, how ’bout, the twenty sixth, three to four? fcae_3_05: you know what we probably need /um/, probably maybe the two or three hours, | and actually the twenty sixth, I’m in a seminar from nine to four, 2nd Date Proposal: /um/ are you in the twenty seventh? | the only thing I have on the twenty seventh, is from ten to eleven, in the morning, fpam_3_06: no, the twenty seventh? through the first of February, I’m out of town. 3rd Date Proposal: how ’bout February the third, at two. | we could do, two to four. fcae_3_07: /uh/ that sounds pretty good. | /um/ I have a meeting from nine thirty to noon, and, that’ll give me some time to catch some lunch, Place Proposal: and /uh/ <we’ll > where do you want to meet. | in <my> my office? or your office. fpam_3_08: your office seems to be bigger than mine. | let’s meet in yours. | on the, third of February, two to four. fcae_3_09: sounds good to me, End of Task-Performance stage Closing stage: fpam_3_10: see you then, | bye. fcae_3_11: ten four good buddy,
The first stage varies in length and content. It can range from a simple “hello”, or just a throat-clearing noise to a lengthy exchange involving questions about the other person’s health, work, family, etc. I considered an Initialization, or Opening, to be any segment of talk from the beginning of the conversation until the need to establish a meeting is stated by one of the speakers. Here there is a remarkable language difference: the Openings of Spanish conversations are, on average 1.2 units (Semantic Dialogue Units) long, whereas the Openings of the English conversations are 0.6 units long. For an example, let us look at (156). It is a typical English example, where the first speaker starts with a simple oh. Other words, or sounds, include okay, let’s see, and um.
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.4 (184)
Chapter 7
(156) [FBNT_MJFG_01] fbnt_1_01: /oh/, would you like to meet, /uh/, for a two hour appointment, *pause* on, *pause* Monday? at, five P M,
The next example is a typical Spanish Opening. Although not very long – the second speaker immediately asks about the reason for the “call” – the speakers greet each other, use their names, and ask about each other’s well-being. (157) [MBRP_MFJR_05] mbrp_05_01: |s|okay |s|Freddy. | cómo estás? mfjr_05_02: bien |s|Brent, | cómo estás tú? | para qué me has llamado? okay Freddy. | how are you? good Brent, | how are you? | why did you call me?
The Opening stage includes a couple of clauses uttered by the speaker who initiates the conversation. Only in one case in the Spanish corpus does it span over two turns. The greeting by the first speaker is never corresponded in English, and only in some occasions in Spanish. This violates principles of reciprocity in formulaic greetings (Firth 1972; Ferguson 1981). This is a rare instance of conversations starting directly with the task. In most descriptions, either of telephone conversations or other casual, face-to-face conversations, the first stage is always one of identification, acknowledgement of each other, and maybe questions about the interlocutors’ health, general state of mind, etc. Given that Openings are usually socially enforced, it was strange that some conversations started directly with the task. However, we need to take into account that the conversations started after the speakers had talked to each other already. It will be recalled (Chapter 3) that the speakers were brought into a lab and introduced to the equipment before the conversation started. This could have made the presence of the Opening stage optional. It is therefore an exceptional characteristic of the corpus conversations, and not typical of other instances of this genre. Now we turn to the Closing stage, where the speakers usually confirm the date agreed upon, say goodbye, and close the conversation. Closings were considered to begin at the point where one of the speakers’ proposal has been accepted by the other speaker. In this stage, the English and the Spanish conversations differ in length: whereas the English are usually short (3.4 units on average), the Spanish Closings are longer (5.3 units). Example (158) is an instance of an English Closing, and (159) is an average Spanish one. I provide, in italics, some context that is not part of the Closing stage. The Closing stage is not optional: it helps bring closure to the task itself, serving both a
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.5 (185)
The generic structure of scheduling dialogues
social purpose (taking leave) and a practical one (ensuring that no further arrangements need to be made). The acceptance of the date agreed upon is secured through an exchange of farewells and repetitions of the date. Schegloff and Sacks also point to closings as solutions to problems of conversational organization (Schegloff and Sacks 1973: 290). (158) [FEAS_MTMR_11] mtmr_11_06: /eh/ let’s call it two, | and I’ll take a long lunch after my seminar. feas_11_07: okay, that’s , Friday the eleventh, at two, from two to four? | I’ll see you then. mtmr_11_08: you got it, (159) [FFCS_FSNM_01] ffcs_01_09: perfecto. | porqué no nos reunimos /ah/ de una a las tres. | qué tal? fsnm_01_10: será muy bueno esa hora, | entonces te veré allí. ffcs_01_11: bueno. | está bien. | te veo el martes a la una. | hasta luego. perfect. | why don’t we meet /uh/ from one to three. | how’s that? it’ll be good at that time, | then I’ll see you there. okay. | that’s good. | I’ll see you Tuesday at one. | see you.
The Task-Performance stage itself is composed of different substages. The usual flow of the conversation is represented in Figure 7.2. There are some variations on this general scheme, but it applies to most conversations, equally in English and in Spanish. The Task-Performance stage contains sequences of Proposals and possible Rejections of such Proposals. The speakers negotiate a suitable time to meet, and when a Rejection takes place, a new Proposal is always put forth. The maximum number of Proposals observed in the corpus was seven, but most typically there were two or three. In the figure, other possible elements are in parentheses: reasons for the rejection, and elaborations on either rejection or acceptance. Example (160), from the Spanish corpus, illustrates a relatively short Task-Performance stage. There are five proposals in this example, each one rejected shortly, without much further detail (there is one reason for the first proposal, that speaker FYMM has class that day). Date proposals two, three and four consist of only Proposal-Rejection sequences, and the final one consists of a Proposal followed by the definitive Acceptance.
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.6 (186)
Chapter 7 A- Proposal
B – (Reason) Rejection (Reason/Elaboration) New Proposal
A – (Reason) Rejection (Reason/Elaboration) New Proposal
B – (Reason) Acceptance (Reason/Elaboration/Details) End of Stage
A – (Reason) Acceptance (Reason/Elaboration/Details) End of Stage
B – (Reason) Acceptance (Reason/Elaboration/Details) End of Stage
Figure 7.2 Flow of conversation in the Task-Performance stage
(160) [FYMM_FKAS_02] 1st date proposal Proposal fkas_02_01: ... tenemos que reunirnos para dos horas. /ah/ cómo te parece, el lunes veintinueve en la mañana. ‘we have to meet for two hours. uh what do you think, Monday the 29th in the morning.’ Rejection fymm_02_02: el lunes veintinueve, en la mañana, es imposible para mí. ‘Monday the 29th, in the morning, is impossible for me.’ Reason porque tengo clase. ‘because I have class.’ 2nd date proposal Proposal qué te parece? si, nos reunimos el lunes, cinco de abril. ‘what do you think? if, we meet on Monday, April 5th.’ Rejection fkas_02_03: el lunes cinco de abril no tengo dos horas libres. ‘on Monday April 5th I don’t have two hours free.’ 3rd date proposal
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.7 (187)
The generic structure of scheduling dialogues
Proposal /ah/, pero, el día siguiente? tengo tiempo en la mañana? ‘ah, but, the next day? I have time in the morning?’ Rejection fymm_02_04: *pause* no disculpa. el lunes veintinueve, *pause* este estoy desocupada, a partir de las dos y media. ‘no sorry. on Monday the 29th, uh I’m free, from two thirty on.’ 4th proposal Proposal pause* así que podríamos reunirnos a partir <de> de tres, a cinco de la tarde? te parece? ‘so we could meet from three to five in the afternoon? what do you think?’ Rejection fkas_02_05: estoy ocupada lunes /ah/ en la tarde. ‘I’m busy Monday uh in the afternoon.’ 5th proposal Proposal martes /gl/ /ah/ treinta *pause* tengo libre en la tarde? ‘Tuesday uh 30th I’m free in the afternoon?’ Acceptance fymm_02_06: |s|okay. está bien el martes treinta. a partir de las tres de la tarde, estoy yo desocupada también. ... ‘okay, that’s good, Tuesday the 30th. from three in the afternoon, I’m also free. ...’
The Place-Proposal stage is less common: three of the English conversations and two in the Spanish part of the corpus contained a Place Proposal. The development is usually one of Propose-Accept, with some elaboration on the details of the place where the speakers are to meet. In (161), a proposal for a meeting place, the speaker’s office, is immediately accepted. (161) [FCRE_FJSL_1] fjsl_1_06: ... and where shall we meet. is my office alright? fcre_1_07: your office would be fine. ...
The next step in the generic description of the dialogues (Eggins and Slade 1997) is the specification of optional and obligatory stages, which can be done in schematic form, by providing a structural formula. In (162) I summarize the sequences in Figures 7.1 and 7.2. The caret sign indicates linear ordering. Parentheses represent optional stages, and recursion (more than one of
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.8 (188)
Chapter 7
the same stage) is represented by a superscript n to indicate any number of repetitions (with seven as the observed maximum). The scope of repetition is enclosed in square brackets. In the structural formula, details, reasons, and other elaborations have been left out. (162) (Opening) ^ Proposal ^ [ ( Rejection ^ New Proposal ) ]n ^ Acceptance ^ (Place Proposal) ^ Closing
. Speech acts According to the Birmingham school of discourse analysis – e.g., Sinclair and Coulthard (1975) – an interaction could be described through a rank scale of transaction – exchange – move – act, where each level is more delicate than the previous ones. The transaction corresponds to a whole conversation. The exchange refers to each turn taken by the speakers, and a move is a possible sequence of speech acts. Moves are the smallest free units of discourse and are made up of one or more acts. Acts are similar to speech acts, although the main difference is that in the Birmingham model they are units of discourse, and thus they are always interpreted in context. Since my analysis was based on a generic framework, I altered the levels of representation to a certain extent, so that a conversation is represented first in generic stages. Those stages contain sequences of turns by different speakers. The contents of a stage, however, are best represented in sequences of speech acts. Some of those sequences correspond roughly to the ethnomethodologists’ adjacency pairs (Sacks et al. 1974). And because the speech acts are only interpreted in their context, they are very much like Sinclair and Coulthard’s acts. This representation is robust, yet simple, for the purposes of a computational model. I shall describe, in Section 7.3, the development of each stage through a sequence of speech acts. Before we proceed on that front, this section provides the taxonomy of speech acts found in the corpus. I will not attempt here a thorough review of the definitions of speech act. I take it to be the minimal unit in the analysis of sequencing in conversational texts (Gallardo Paúls 1998). This unit represents the communicative intention of the speaker. A number of different taxonomies exist, from the original ones by Austin (1962) and Searle (1969, 1979), to the developments within the Birmingham school (Sinclair and Coulthard 1975; Coulthard and Montgomery 1979; Brazil 1995). Tsui provides an extensive taxonomy of what she calls discourse acts (Tsui 1994), and Aijmer a corpus-based classification for
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.9 (189)
The generic structure of scheduling dialogues
English (Aijmer 1996). Efforts in computational linguistics have resulted in inventories, most of them designed for task-oriented dialogue (Allen and Core 1997; Jekat et al. 1995; Traum and Hinkelman 1992). Geis (1995) proposes a more dynamic view of speech acts, where context is allowed to play a role. He argues that speech acts are social actions, as opposed to linguistic actions. He uses these concepts to model conversational interaction through Discourse Representation Theory (DRT) (Kamp and Reyle 1993). The speech acts presented here are the result of generalizations extracted from the corpus, and are not meant to be definitive or exhaustive. In addition, these speech acts could be organized at different levels. Thus, request-date could be subsumed under request-info, as an instance of a request for information. As a result, some of the speech acts are very general, whereas some of them are found at a deeper level of specificity. I chose to represent the most common speech acts, or those related to the task, at the most specific level. The speech acts related to the social circumstance of the situation, such as instances of thanking or of good wishes, are represented more generally. The reason is that, in a translation system (the original goal of the data collection effort), it is very important to parse the speech acts pertaining to the task. On the other hand, and although it is not pragmatically optimal, the translation does not need such specificity or variety when it comes to, e.g., thanking the other speaker. Table 7.1 lists the speech acts included in the taxonomy, and their frequencies in the corpus. The percentages shown are with respect to the n for that language. Table 7.2 shows the ratio of speech acts per SDU. It will be recalled (see Section 3.3, in Chapter 3), that SDUs (Semantic Dialogue Units) are portions of a speaker’s turn that can stand alone semantically. These portions often, but not always, coincide with a clause or sentence. In some cases, an SDU will contain more than one speech act. Table 7.2 reflects this fact, in that the ratio is above 1. The difference between the languages lies in the higher ratio of speech act per SDU in English. The Appendix includes a description of each speech act, and some illustrative examples. In summary, the speech acts in the corpus can be classified, in Conversation Analysis terms, as representing constitutive or linking moves – see, for instance, Gallardo Paúls (1998). Speech acts in the constitutive move category are the ones pertaining to the task at hand, such as ask-date, inform-availability or reject-date. Linking moves refer to the discourse and social embedding of the conversations. Filled-pauses and greetings could be included here. Also included are instances of repeat-confirm, inasmuch as they function as turnyielding speech acts.
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.10 (190)
Chapter 7
Table 7.1 Frequencies of speech acts in the corpus English Accept Ask-date Backchannel Channel Filled-pause Goodbye Greeting Inform Inform-availability Other Politey Propose-action Propose-place Reject-date Repeat-confirm Request-action Request-confirmation Request-date Request-information Request-meeting Request-place Self-introduction Vocative n
Spanish
T
%
T
%
73 97 22 9 24 29 3 307 249 4 10 9 8 20 39 8 35 20 8 22 4 2 11
7.21 9.58 2.17 0.89 2.37 2.86 0.30 30.31 24.58 0.39 0.99 0.89 0.79 1.97 3.85 0.79 3.46 1.97 0.97 2.17 0.39 0.20 1.09
130 127 19 2 88 42 16 451 311 2 14 37 8 62 61 11 70 23 21 26 2 0 19
8.43 8.24 1.23 0.13 5.71 2.72 1.04 29.25 20.17 0.13 0.91 2.40 0.52 4.02 3.96 0.71 4.54 1.49 1.36 1.69 0.13 – 1.23
1013
1538
Table 7.2 Ratio of speech acts to SDUs
Speech acts SDUs Ratio
English
Spanish
1013 784 1.29
1542 1294 1.19
. Development of stages through speech acts Each of the stages in Figure 7.1 contains a sequence of speech acts that constitute the local structure of the conversation. In this section I describe how each stage unfolds in a sequence of speech acts.
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.11 (191)
The generic structure of scheduling dialogues (Speaker 1) 1 – filled-pause / vocative / greeting / politey 2 – (greeting) / (politey) / (filled-pause) / (self-intro)1 3 – (filled-pause) (Speaker 2) 1 – politey 2 – request-info
Figure 7.3 Sequence of speech acts in the Opening stage
The Opening stage is, as we saw earlier, optional. Figure 7.3 displays the sequences of speech acts in the Opening stage. In the interest of clarity, I have abandoned the finite state machine representation for a more schematic one. In the figure, and in the ones to follow, a sequence of speakers is identified as ‘Speaker 1’ and ‘Speaker 2’. Such labeling does not refer to the absolute ordering in the conversation, but to the ordering within the stage. ‘Speaker 1’ is always the speaker who initiates this stage, regardless of whether he or she was the first speaker in the conversation. ‘Speaker 2’ refers to the other participant. A parenthesis around a speaker indicates optionality. Thus, in Figure 7.3, the Opening stage could finish after the first speaker’s contribution. A parenthesis around a speech act indicates that it is optional. The numbers ‘1’, ‘2’ and ‘3’ refer to the ordering of speech acts with respect to each other. Speech acts separated by slashes indicate that there is a choice among them. A speech act between parentheses is optional, and one preceded by an asterisk might be repeated. The figure then reads: Speaker 1 initiates the conversation with either a filled-pause, a vocative, a greeting or a politey. After that, the rest of the stage is optional. If the same speaker utters a second speech act, the choice is one of greeting, politey, filled-pause or self-intro. If a third speech act is uttered, then it is always a filled-pause. This is a possible end of the stage. If the interlocutor takes the floor at that point and continues in the Opening stage, his or her sequence would be a politey followed by a request-info. These representation conventions will be used throughout the present section to describe the unfolding of stages. Here, as in the rest of the stage representations, the ordering is based on corpus analysis. I have excluded some borderline or uncommon cases of sequencing. Predictions that a certain utterance will follow another are probabilistic at best, never as absolute as portrayed in these figures (Bernárdez 1995: 92). In summary, not all conversations in this genre will follow this sequencing, but the great majority will.
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.12 (192)
Chapter 7
We now turn to the Date-Proposal stage, the longest and most central part of any conversation. As we saw at the beginning of this section, a Date Proposal might be repeated. A conversation might contain as many as eight different Proposals, and as few as one. Independently of the number, they all follow the schema represented in Figures 7.4 and 7.5 These two figures outline the complicated structure of the Date-Proposal stage to its maximum length, eight turns. T.1 through T.8, in the figure, refer to the turn numbers. The only obligatory turn is the first one, since a proposal might end after that with the second speaker initiating a new proposal. This amounts to an implicit rejection of the first speaker’s proposal. If a new turn is initiated, only the first speech act is obligatory. In fact, in all of the possible eight turns only one speech act is obligatory at each turn. This is reflected in the lack of parentheses around speech acts at turn one and the presence of parentheses in all other speech acts in the turn. The Place-Proposal stage is less common: three of the English conversations and two of the Spanish contained a Place Proposal. The development of this stage is depicted in Figure 7.6. This Proposal always consists of three obligatory turns. The last stage in the dialogues is the Closing stage. Unlike the Opening stage, this one is present in each of the 60 conversations. In both languages, it has a minimum length of one turn. In English, the maximum length is four turns. In Spanish, three turns are the maximum. A graphical representation of this stage can be found in Figure 7.7. The contrast between Opening and Closing stages is noticeable. The former are optional, the latter always present. This might have to do with the need to solve the task as the most pressing issue. Such a need might also determine the presence of a final stage which serves as closure to the task. Speakers do not part until they are assured of the acceptance, on both sides, of the meeting date. The acceptance is secured through an exchange of farewells and repetitions of the date. Schegloff and Sacks also point to closings as solutions to problems of conversational organization (Schegloff and Sacks 1973: 290). In addition to the speech acts included in each stage, there are a number of discourse-based speech acts that can occur, repeatedly, anywhere in the conversation. Those are: backchannel, filled-pause, vocative, politey and channel. An important point here is the fact that adjacency pairs (Sacks et al. 1974) do not succeed each other in immediate fashion. That is, rarely does a turn consist of a single ask-date followed by an isolated reject-date (dispreferred second pair) or accept (preferred second pair). The sequential implicativeness
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.13 (193)
The generic structure of scheduling dialogues
T. 1
( Speaker 1) 1 – ask-date / request-date / inform-av / request-meeting / inform / request-info 2 – (request-meeting) / (propose-action) 3 – (inform) 4 – *(inform-av) / (request-meeting) 5 – (request-conf) / *(ask-date) 6 – (request-date) / (reject-date) 7 – (inform-av) / (accept) / (request-conf) / (request-action) 8 – (reject-date) [possible end of a failed proposal, the other speaker begins a new one, implicitly rejecting this one]
T. 2
(Speaker 2) 1 – inform-av / ask-date / reject-date / repeat-confirm / request-date / accept_inform / request-date / (greeting)2 2 – *(inform) / (politey) / *(inform-av) 3 – (request-action) / (repeat-confirm) / (reject-date) / (inform-av) 4 – (accept) / *(request-date) / *(inform-av) 5 – (request-conf) / (propose-action) / (request-action) / (reject-date) 6 – (request-place) / (inform) / *(request-conf) / *(inform-av) / (accept) [possible end of a failed proposal, the same speaker begins a new one, becoming then Speaker 1]
T. 3
(Speaker 1) 1 – reject-date / accept / inform-av / ask-date / request-action / inform / repeat-confirm / request-info 2 – *(accept) 3 – *(inform-av) / *(inform) / (propose-place) / (propose-action) 4 – (request-date) / (inform-av) / (inform) / (repeat-confirm) 5 – (ask-date) / (request-conf) / (repeat-confirm) / *(inform-av) 6 – (accept) / (propose-action) / (ask-date) / (request-info) / *(inform-av) [possible end, after acceptance or rejection]
T. 4
(Speaker 2) 1 – inform-av / reject-date / repeat-confirm / accept / ask-date / inform / request-action 2 – (accept) / (request-info) / *(inform) / *(ask-date) / (propose-place) (inform-av) 3 – (accept) / (repeat-confirm) / (reject-date) / *(inform-av) 4 – (ask-date) / (accept) / (request-place) / (request-meeting) 5 – (repeat-conf) / (request-conf) [possible end, after acceptance or rejection]
Figure 7.4 Sequence of speech acts in the Date-Proposal stage
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.14 (194)
Chapter 7 T. 5
(Speaker 1) 1 – ask-date / inform-av / accept / propose-place / inform / repeat-confirm / reject-date 2 – (inform) / (accept) / (propose-place) 3 – (inform-av) / (repeat-conf) / (ask-date) 4 – (accept) / (repeat-conf) 5 – (request-conf) / (inform-av) [possible end]
T. 6
(Speaker 2) 1 – accept / inform / request-info / inform-av / 2 – (inform) 3 – (ask-date) / (request-conf) [possible end]
T. 7
(Speaker 1) 1 – accept / inform 2 – inform [possible end]
T. 8
(Speaker 2) 1 – reject-date 2 – politey
Figure 7.5 Sequence of speech acts in the Date-Proposal stage, continued
T. 1
(Speaker 1) 1 – request-action / propose-place 2 – (propose-action) / (propose-place) 3 – (propose-place) / *(request-info) 4 – (inform) / (request-conf)
T. 2
(Speaker 2) 1 – accept / propose-place / inform 2 – (repeat-confirm) / *(inform) 3 – (accept-action) / (repeat-confirm)
T. 3
(Speaker 1) 1 – accept
Figure 7.6 Sequence of speech acts in the Place-Proposal stage
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.15 (195)
The generic structure of scheduling dialogues T. 1
(Speaker 1) 1 – accept / repeat-confirm / inform / goodbye / request-info / inform-av / propose-place 2 – *(accept) 3 – (inform-av) / (request-info) / (request-conf) 4 – (request-action) / (propose-place) / (repeat-confirm) / *(goodbye) 5 – (inform) / (repeat-confirm) / (goodbye) 6 – (request-conf) / *(inform) / (repeat-confirm) / (politey) [possible end]
T. 2
(Speaker 2) 1 – accept / goodbye / repeat-confirm / bch 2 – *(accept) / (goodbye) / (inform) 3 – *(repeat-confirm) 4 – (politey) / (goodbye) [possible end]
T. 3
(Speaker 1) 1 – accept / inform / goodbye 2 – (accept) / (request-action) 3 – *(goodbye) / (inform) / (accept) 4 – (request-info) / (goodbye) / (inform) 5 – (goodbye) / (repeat-confirm) [conversations continued after this point only in English]
T. 4
(Speaker 2) 1 – accept / inform
Figure 7.7 Sequence of speech acts in the Closing stage
of adjacency pairs (Schegloff and Sacks 1973) is, however, still valid. Sequential implicativeness means that an utterance projects a range of occurrences as its follow-ups. That is, ask-date always projects some kind of response. However, both the first pair utterance and the response rarely occur by themselves, in the same dry manner a tennis ball goes back and forth on a court. Researchers have noted the presence of insertion and side sequences (Schegloff 1972, Jefferson 1972). There are, in addition to those, more elements surrounding an adjacency pair. Speakers hesitate, include false starts, filled-pauses, repetitions of availability or unavailability, all surrounding and peppering the two- (or three-) pair exchange.3 In Example (163), for instance, the request for a meeting formulated by speaker FCKA (“I’ve been wondering [...] whether you have
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.16 (196)
Chapter 7
time”) is never answered explicitly. The negotiation for a date that speaker FFMW engages in serves as an implicit acceptance of that request. Similarly, the response to FCKA’s request for a meeting on a workday is answered by FFMW’s whole turn. It would be difficult to point to a single clause that provides the second pair element, although we know that the prompt was appropriately followed up. It is for this reason that the sequencings of stages in the figures above are complex, repeated, and show a great deal of variability. (163) [FCKA_FFMW_10] fcka_10_01: ... request-meeting {crosstalk} I’ve been wondering, this month whether you have time, | request-meeting I have some problems with some files | and I need your advice on some of them. | requestmeeting I really want your opinion. | inform-av /um/ I’m available, any workday, this month, apart from, Thursday the thirteenth where I have my birthday, | inform-av /er/ all the other days I can always arrange something. ffmw_10_02: inform-av I forgot to tell you, /uh/ Christina, that, I have, /um/ a vacation planned, | inform-av I’m leaving this Sunday, the second, | and, I’m not coming back until, May the sixteenth. | inform-av so, /um/ it’ll have to be some time, during that week, either, Tuesday, Wednesday, Thursday, or Friday. | ask-date /uh/ do you have time, any of those days?
. Thematic structure and staging All the stages are defined in functional terms, that is, in terms of their purpose, which can be a mixture of practical (related to the task) and social (related to the maintenance of a relation between the interlocutors, protecting each other’s face and saving their own). The next step in a generic analysis, once the stages and the speech acts have been defined, is to examine the lexicogrammatical features that distinguish each stage. Discourse and lexicogrammatical features, together with choices in the register variables, distinguish a genre from others. Discourse and lexicogrammatical features also help define the stages within that genre. The next three sections examine each of the analyses carried out from this point of view. There are two aspects to the analysis of the thematic structure of the dialogues: theme selection and thematic progression. I will focus on each one of those, and examine how they relate to staging.
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.17 (197)
The generic structure of scheduling dialogues
The types of Themes chosen in each of the stages vary because the subject matter of the stages is different. In the Opening and Closing stages, the participants interact with each other, and very little ideational content is expressed. In fact, most of the clauses in Openings and Closings are minor, thus showing no Theme-Rheme articulation. When a thematic structure is present, it contains textual Themes, such as conjunctions. In Example (164), a Closing from the English corpus, the speaker closes the conversation with a short series of clauses. The ideational Themes (underlined) refer to the meeting (that), or to the speaker herself (I, in two of the clauses). There is also an imperative clause, with a Finite as Theme. The textual Themes (in italics) are the two instances of so. (164)
fsjb_1_05: ... so that’ll be great, | I’m free the whole afternoon, | so, be prepared to take as much time as you need. | I’ll see you then,
The Themes selected in the Task-Performance stage are more characteristic of the genre: they represent, as already discussed in Chapter 4, the subject matter of the conversations, including the interlocutors and the dates when they are available. Ideational Themes are often first and second person pronouns, and also dates and times, either as Adjuncts (marked Themes) or as Subjects (unmarked). What is interesting in the Theme selection is that marked Themes seem to indicate that a new date is being proposed. Clauses were mapped to their function within the Task-Performance stage, into one of the following substages: Proposal, Rejection, Acceptance, New Proposal, Details, Other. Of those, the New Proposal substage is often initiated with a marked Theme, a date in Adjunct position. Example (165) is interesting: the speaker first rejects a previously proposed date (the 17th), and then introduces two new proposals, the 18th and the 19th. The rejection is encoded as a Date-Subject, but the new proposals are Date-Adjuncts, in Theme position, therefore marked. As with other examples of Theme, textual Theme is in italics, interpersonal Theme is in bold, and ideational Theme is underlined. (165) [FKNB_MCJC_10] mcjc_10_04: okay, /uh/, the seventeenth. is, pretty much completely used up with a seminar for me, | gonna find whoever scheduled, me for that, and, shoot them. | the eighteenth, I have the morning free, and also, noon until two, | the nineteenth, I have most of the day free. at least, after lunch.
In Spanish, marked Themes are also present with proposals, and especially new proposals (after a date has been rejected). The tendency is not as clear as in
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.18 (198)
Chapter 7
English, and marked Themes appear in other situations. In the following example, marked Themes are used both for the first proposal (in the first turn, by speaker MJBP), and for a rejection (in the second turn, by speaker MMBU). Thursday is now proposed, as an apparent unmarked Theme (the structure of the clause is difficult to interpret), and speaker MJBP rejects it, making it the unmarked Theme of the clause. In summary, both proposal and rejection used marked Themes, but a new proposal and a new rejection are both encoded as unmarked, showing no clear tendencies one way or another. An interesting aspect of the example is that speaker MJBP, in his first turn, produces a false start, is, which may have led to an unmarked Theme: Monday the twenty-fourth is.... He instead made it marked: Monday the twenty-fourth I am busy. (166) [MJBP_MMBU_04] mjbp_04_02: <el> bueno. | el lunes veinticuatro <es> tengo una reunión desde las una hasta las cuatro de la tarde. | el martes tengo reunión de dos a cuatro y treinta. | el miércoles tengo el día completo ocupado. mmbu_04_03: *pause* el lunes en la mañana *pause* estoy muy ocupado. [..] creo que el jueves veintisiete, que lo tengo totalmente libre podría ser. mjbp_04_04: bueno. el jueves realmente es un día ocupado para mí. mjbp_04_02: <el> well. | Monday the twenty-fourth I have a meeting from one until four in the afternoon. | on Tuesday I have a meeting from two to four thirty. | on Wednesday I am busy the whole day. mmbu_04_03: *pause* Monday morning *pause* I’m very busy. [..] I think that Thursday the twenty-seventh, that I have completely free could be. mjbp_04_04: well. Thursday really is a busy day for me.
As for Thematic Progression, the only patterns were found within the TaskPerformance stage. As is to be expected, Proposal substages are characterized by New progression patterns. In Rejection and Acceptance substages, it is common, in both languages, to have a constant (themic) pattern, with speakers repeating the previous Theme. Example (167) illustrates a gapped constant Theme, with after noon on Tuesday repeated by the speakers. We can also see how thematic progression is built interactively, as one speaker picks up and repeats another speaker’s Themes. (167) [FCRE_FJSL_1] fcre_1_03: ... I am free on Friday the twelfth from eleven to one, [...] any time after noon on Tuesday would be good.
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.19 (199)
The generic structure of scheduling dialogues
fjsl_1_04: oh, unfortunately I have a seminar all day on Friday the twelfth? | but, after noon on Tuesday the sixteenth would be great. ...
. Rhetorical relations and staging The types of rhetorical relations found in each of the stages in the conversations are slightly different, and constant across the two languages. First of all, in the Openings there are not very many relations. Most of the Openings do not even constitute a full span of text, making it difficult to relate it to any other span. When there are actual relations, those are of Solutionhood, as in Example (168). In this example, the Opening moves immediately onto the Task-Performance stage, with speaker FHKR posing a question, how’s your schedule for the fourth, followed by a justification, I’m free all day. Those two units form the satellite of the Solutionhood relation. The relation takes place across turns, with the nucleus being provided by speaker FDMR. (168) [FHKR_FDMR_JA] fhkr_ja_01: Heather, how’s your schedule for, the fourth. I’m free all day. fdmr_ja_02: /hm/ the fourth, ...
Most of the relations, and the most complicated ones, are found in the TaskPerformance stage. The typical realization of the flow in Figure 7.2 is, in rhetorical relations, a sequence of Non-Volitional Causes or Results for the reasons why the proposal is not accepted. In the sequence Reason-Rejection, the typical relation is one of Non-Volitional Result (“I have another meeting at that time, so I can’t meet”). In the sequence Rejection-Reason, the relation is usually one of Non-Volitional Cause (“I can’t meet on that date, because I already have another meeting”). In Example (169), the speaker rejects Monday, providing a reason for the rejection, in the form of a Non-Volitional Cause. The relation is implicit: no marker is present, and the two spans are presented as two independent sentences. (169) [FFCS_FSNM_01] ffcs_01_03: ... el lunes no me queda bien. | tengo un almuerzo de las doce a las dos. ... ... on Monday it’s not good. | I have a lunch from twelve to two. ...
On the other hand, when an Acceptance takes place, the relations are often Volitional Cause and Result, depending on where the optional Reason is placed.
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.20 (200)
Chapter 7
In (170), the relation is one of Volitional Result (“I’m free after noon. Then we can meet at one.”). (170) [FKND_FKAS_01] fkas_01_05: bueno, yo también tengo libre después del mediodía, | entonces /um/ qué le parece si nos reunimos a la una? ... well, I’m also free after noon, | then um what do you think about meeting at one? ...
Most of the relations in the Task-Performance stage are turn internal. The interturn relations are typically Solutionhood and Evaluation, an example of which we can see in (171). (171) [FHKR_FDMR_JA] fdmr_ja_12: ... so, how ’bout two o’clock. fhkr_ja_13: that’s fine, ...
The other two elements of the Task-Performance stage are New Proposal, usually a Result (Volitional or Non-Volitional) of the speaker’s agenda, followed by a Condition (“we can meet on this other date, if that’s okay with you”). In Example (172), spans 1 and 2 are in a Non-Volitional Result Relation. Span 2, at the same time, proposes a new date, conditioned by span 3, in which the speaker submits the new date for his interlocutor’s approval. (172) [FPDV_FSNM_04] fpdv_04_04: [1] realmente el miércoles veintiséis no tengo horas libres más que muy temprano en la mañana. | [2] tendríamos que diferirlo para el día lunes treinta y uno por la mañana. | [3] *pause* espero que eso sea adecuado. [1] really on Wednesday the twenty-sixth I don’t have any free hours except very early in the morning. | [2] we’d have to postpone it for Monday the thirty-first in the morning. | [3] I hope that’s adequate.
The Details of the meeting, the last element of the Task-Performance stage, are usually Elaboration relations. These can be turn-internal, or cross over turns. In (173), the Elaboration relation is between the spans marked 1 and 2. Speaker MAWK elaborates on his acceptance of the sixteenth, by saying, in the second span, that they can meet any time (after twelve). (173) [MAWK_MANV_1] manv_1_6: ... is the sixteenth okay with you | I have a class from nine to twelve | but we could probably meet at one.
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.21 (201)
The generic structure of scheduling dialogues
mawk_1_07: [1] the sixteenth after, twelve o’clock is fine for me. | [2] we can schedule, anytime you want.
The two languages do differ in the Closing stage. The Spanish Closings are longer (5.3 SDUs vs. 3.4 for the English) and contain a higher number of Restatement and Summary relations, often marked by entonces. Example (174) shows a turn-internal Restatement in the Closing section. The speakers have already agreed on the date, and speaker FCBA repeats the arrangements. Span 2 restates span 1, and, to a certain extent, span 3 also restates span 2, although 3 could also be interpreted as a conversational device to close the conversation, and not part of the RST structure. (174) [FCBA_MEBA_08] fcba_08_06: ... [1] entonces nos vemos el dieciséis, en mi oficina. | [2] y allí te veo. | [3] hasta luego. ... [1] then wel’ll meet the sixteenth, in my office. | [2] and I’ll see you there. | [3] see you.
. Cohesion and staging The relationship of cohesion to staging is mediated by the cohesive chains running through the text. We saw, in Chapter 6, that cohesive ties enter into cohesive chains, through the different types of cohesive relations. In this section I discuss the nature of the chains in each of the stages. Most Openings are short in both languages. The Opening is considered to be over as soon as the speakers start negotiating a date. As a consequence, there are no chain-related elements in an Opening. Example (175) contains the first two turns of a Spanish dialogue. The Opening is over after the first SDU, that is, immediately before speaker FSNM mentions the need to meet. The first chain begins with the mention of a date (lunes veinticuatro de mayo), which is picked up by the other speaker. This chain started two clauses into the Task-Performance stage. There is, then, no chain crossing over from Opening to Task-Performance stage. This is true of all the conversations in the corpus, and for both languages. (175) [FPDV_FSNM_04] fsnm_04_01: bueno Patricia. | parece que tenemos que hacer otra reunión para terminar nuestro asunto. | a ver si tiene algunas horas libres el lunes veinticuatro de mayo por la tarde.
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.22 (202)
Chapter 7
fpdv_04_02: *pause* el lunes veinticuatro por la tarde va a ser difícil. porque tengo una reunión. | no sé si sería posible hacerlo otro día como el miércoles veintiséis /eh/ temprano en la mañana o <el> *pause* el martes veinticinco a cualquier hora de la mañana. well Patricia. | it looks like we need to have another meeting to finish our business. | let’s see if you have some free hours on Monday the twentyfourth of May in the afternoon. on Monday the twenty-fourth it is going to be difficult. because I have a meeting. | I don’t know if it would be possible to do it another day such as Wednesday the twenty-sixth uh early in the morning or on Tuesday the twenty-fifth any time in the morning.
As for the Task-Performance stage, all the chains are initiated in it. Some of them are abandoned for new ones, and some minor chains continue alongside the major chains. The Task-Performance stage is the place where the chains are developed for the most part. If we were to break up the Task-Performance stage into smaller substages, we would discover new chains starting at transition points in these stages. Since these substages would involve the discussion of different dates, it is obvious that we could find a switch over to a new chain as soon as a new date is introduced and discussed. The chains that originate in the Task-Performance phase do not always cross over into the Closing stage. If they do, they include links that represent the closing elements in a chain, which include a restatement of the date or a demonstrative reference to it. There are, on average, exactly the same number of links in the Closing section of the Spanish and English dialogues: 2.33 links. These links always belong in the last major chain of the dialogue, and are most often instances of repetition (lexical cohesion) or demonstrative reference. Such is the case in (176), which contains the last two turns of a conversation. The Closing begins immediately after speaker FJAB has accepted the date (couldn’t be better). The last SDU in her turn repeats the date and the time that have been discussed previously in the chain (eleven to one on the twenty-sixth). Speaker FCAD responds by referring to that time (then). (176)
fjab_12_08: couldn’t be better. | and then you know maybe if we’re, over a little bit, you know, it would be okay. | so, eleven to one, on the twenty sixth. fcad_12_09: sounds great. | see you then.
The most interesting phenomenon regarding cohesion and staging is the presence and evolution of cohesive chains within the Task-Performance stage. As I just mentioned, the Task-Performance stage is divided into smaller substages,
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.23 (203)
The generic structure of scheduling dialogues
Table 7.3 Cohesion chains in Example (177) Chain 1 fece_10_01
sometime between the 3rd and 14th Thursday the 6th Monday the 10th
mnfh_10_02
from the 3rd until the 15th
Chain 2
Chain 3
from the 17th till the week of the 24th the 17th that on the 18th on Tuesday the 18th fece_10_03
on the 18th
with new chains starting at transition points from one of those substages to the next. Those substages correspond to proposals to meet, what we can call DateProposal stages. In addition, some conversations contain a Place-Proposal stage, where the speakers choose a convenient place to meet. The different cohesive chains that run through the Task-Performance stage are closely related to the beginnings and ends of the Proposal stages. That is, a new Proposal initiates a chain, and the rejection or completion of the Proposal stage terminates that chain. In the chains in Figure 6.4 in Chapter 6, we observe that the major chains running through the text correspond to different Proposals. Minor chains do not correspond closely with Proposal stages, because they are supporting, or in close relation to, a major chain. Cohesive harmony is again violated: there are whole parts of the text that do not interact at all with each other. Rather, non-interacting chains mark the transition from one stage to the next. Example (177) contains the Task-Performance stage of one of the dialogues, subdivided into two different Date Proposal stages. The chains for this example are illustrated on Table 7.3. (177) [FECE_MNFH_10] 1st proposal fece_10_01: ... now we’re trying to meet? /um/, sometime between the third and the fourteenth of May. | /ah/ Thursday the sixth looks pretty good, and, so does Monday the tenth. | how ’bout for you. mnfh_10_02: /um/, actually I have vacation from the third, until the fifteenth?
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.24 (204)
Chapter 7
2nd proposal we can actually meet from the, [seventeen(th)], ’till the week [dof (of)] twenty fourth. | so /mm/, on the seventeenth, /um/ I have seminar all day, | /um/ /lg/ that’s not really a good time for me. | and on the eighteenth /um/, it’s [actual(ly)] it looks pretty good. | I /ah/ have class from two to four. | so would you like to meet with me on Tuesday the eighteenth? fece_10_03: well why don’t we try for another lunch meeting on the eighteenth from, twelve to two. | /um/, that will actually work out pretty well. ...
The graphical representation shows clearly that a new (identity) chain is initiated every time a new proposal takes place. The first chain contains dates between the third and the fourteenth of May. Included in that range are the two dates mentioned by speaker FECE. Speaker MNFH, in his turn, rejects that entire range because he will be away. He then, within the same turn, makes a new proposal, that should be between the seventeenth and the twenty-fourth. His first proposed date is the seventeenth, to which he refers with the demonstrative pronoun that. He also proposes the eighteenth in that same turn. Speaker FECE, in her turn, accepts that date and repeats it. Another interesting aspect of the chains running through the text is that they transcend turn boundaries. Speaker MNFH continues a chain initiated by his interlocutor the previous turn (Chain 1), and speaker FECE also continues Chain 3 in her turn. The breaks in the chains of cohesion could lead one to think that the conversations are not cohesive, since no chains – at least no identity chains – run through an entire conversation. The conversations are, however, coherent, because cohesiveness is not necessary, in this case, to produce the feel of coherence. Coherence is achieved through genre expectations: if a date has been rejected, the conversation ought to proceed onto a new date, so that the task can be accomplished. The conclusion is that we can predict a new stage given a new chain, and vice versa. We could use this knowledge in different applications, such as automatic recognition of stages. If we know that the participants in a conversation have begun a new cohesive chain, we can predict that they are in a different generic stage of that conversation. This would be useful in, for instance, information retrieval. In order to retrieve a part (a stage) in these conversations, we only need to search for beginnings and endings of cohesive chains. In the opposite direction, knowing that we have started a new stage usually means that
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.25 (205)
The generic structure of scheduling dialogues
Table 7.4 Thematic, rhetorical and cohesive characteristics of stages Stage Opening Task Performance Proposal Rejection Acceptance New Proposal Details Closing
Theme
TP
RST
Cohesion
Solutionhood Marked Marked
New Constant Constant New
New chain Non-Vol. Cause/Result Vol. Cause/Result New chain Elaboration Evaluation, Restatement, Summary
New chain
anaphora resolution will take place within that stage, because references will usually be only to elements in the same cohesive chain.
. Characterization of stages We have seen, in the previous sections, that we can characterize some of the stages by distinct thematic, rhetorical, and cohesion patterns. Table 7.4 summarizes those characteristics.
. Summary This chapter presented a description of the stages in the conversations in the corpus. In the first part of the chapter, I describe the three main stages of the conversations: Opening, Task Performance and Closing, and the substages within the Task-Performance stage. Each one of them is motivated by functional criteria; each one has a specific purpose within the conversation, from the point of view of the participants. The analysis of stages is completed with a taxonomy of the speech acts typically found within each stage. This formal characterization can be useful in computational applications. For instance, a similar formal model for service encounters with travel agents (booking a flight or a hotel) was used to improve the translation quality in a speech-to-speech machine translation system (Taboada 1997, 2003). The second part of the chapter is devoted to correlating stages with the three types of structures studied in Chapters 4 to 6. Thematic patterns were found to be slightly different in the two languages: whereas English tends to
TSL[v.20020404] Prn:7/10/2004; 9:20
F: PB12907.tex / p.26 (206)
Chapter 7
encode New Proposal stages consistently with marked Themes, the tendency was not so clear in Spanish, and marked Themes are used in most of the stages. Section 7.5 considered the realization of rhetorical relations in the different stages of the dialogue, pointing out that specific clusters of relations tend to appear according to the stage of the dialogue. The cross-linguistic analysis showed no difference between the English and the Spanish data as far as staging and type of relations are concerned. The only difference was found in the length of the Opening and Closing stages of the dialogues, which tend to be longer in Spanish. Finally, Section 7.6 discusses the presence of cohesion according to the different stages. The Opening and Closing phases of the conversations did not establish cohesive links with the rest of the conversation, which contained the most number of links and chains. The Task-Performance stage was characterized by starting new chains every time a new date is proposed. This chapter concludes the analysis and closes the circle around the generic study of the conversations. I described genre, and scheduling dialogues as a genre in Chapter 2, to then move onto the in-depth analysis of the conversations. In this chapter, I returned to genre in order to correlate genre staging and characteristic discursive patterns.
Notes . Self-intro occurs only in English. . Greetings in this stage are present only in Spanish. . See Tsui (1989) on three-pair exchanges.
TSL[v.20020404] Prn:29/07/2004; 11:43
F: PB12908.tex / p.1 (207)
Chapter 8
Conclusions and consequences
This book addresses the construction of coherent dialogue by two speakers. Coherence is interpreted within the Systemic Functional tradition, as comprising a number of textual phenomena that were analyzed in a bilingual corpus of task-oriented dialogues. The dialogues are instances of a scheduling genre, where two speakers communicate in order to find a suitable time for a meeting. I established, in Chapter 2, the generic nature of the dialogues. That chapter provided a review of genre theory as the main framework for the study. According to genre theory, a genre is realized in texts which perform a social purpose. As a result of their goal orientation, these texts develop in a series of clearly defined stages. In a circular closure of sorts, the last chapter, Chapter 7, presents the staging structure of the dialogues. Chapter 7 also includes a characterization of the stages through the speech acts that can be found in those stages. These two structural descriptions – at the global level through staging and at the local level through speech acts – provide a computational model that can be used in the processing of conversations of this type. Chapters 4 through 6 provide the analyses of three types of devices that build coherence and cohesion. For all these analyses, the conversations were considered to be texts and, as such, created by both participants. Each of these analyses also took into account the generic nature of the conversations, examining possible differences in realization that could be related to boundaries between stages. Chapter 4 concentrates on thematic realization and thematic progression, showing a few significant differences between the Spanish and the English conversations, discussed below. Chapter 5 examines the rhetorical structures found in the conversations, applying Rhetorical Structure Theory (RST). The application is performed in two different ways. Since RST was developed for monologic discourse, I first examined each turn in the conversation as a text. The second analysis considered the conversations as whole texts, finding important relations holding turns together. Both English and Spanish conversations contained similar distributions of relations.
TSL[v.20020404] Prn:29/07/2004; 11:43
F: PB12908.tex / p.2 (208)
Chapter 8
The study of cohesive devices is taken up in Chapter 6. I examined the presence of different devices and the characteristics of cohesive chains. I also propose a slightly different characterization of cohesive devices in Spanish, especially as regards substitution. Another interesting discovery in the analysis of cohesion was that cohesive chains do not interact very much with each other. This was because a new stage in the text usually involved a new cohesive chain. I would like to concentrate now on what I consider are the consequences of this study. First of all, I will review the research questions posited in Chapter 1, the introductory chapter. I will also point to a number of possible applications of this analysis of cohesion and texture from a generic point of view. The first research question pondered how we can best characterize the texture of spoken language. I have, throughout this book, defended that we can interpret texture by analyzing three elements: relations among propositions (RST), relations between anchors and their progression (thematic analysis) and cohesive devices. My second research question related to the suitability of written text analysis tools as a means of studying spoken language. I hope to have shown that, indeed, we can characterize spoken language through the analysis of thematic progression, rhetorical structure and cohesion. Although some of these methodologies, especially cohesion, have not been especially designed for written language, it is often true that they are rarely applied to exhaustive analyses of spoken language. This is especially the case with Rhetorical Structure Theory. I have, with very few adaptations, employed these tools in a complete study of a spoken language genre. I determined, in Chapter 2, that the dialogues included in the corpus belong to a scheduling genre. My third research question proposed an investigation of the effects of generic structure, described in stages, on the texture and cohesion of the dialogues. Breaking down the effects by type of analysis, we have observed that theme markedness points to a proposal or rejection within the Task-Performance stage. In terms of thematic progression, proposals are usually signaled with a New Theme being introduced. The rhetorical analysis showed that different stages contain different relations, e.g., Solutionhood in the Opening stage, and sequences of Cause and Result in the Task-Performance stage. But the most noticeable genre-related patterns are to be found in the cohesive patterns. Analysis of major cohesive chains running through the conversations showed that a new major chain was initiated every time the change to a new Proposal stage took place. The next question probed whether there were significant differences in the realization of texture between English and Spanish. A related question is
TSL[v.20020404] Prn:29/07/2004; 11:43
F: PB12908.tex / p.3 (209)
Conclusions and consequences
whether English-based analysis tools are adequate for the analysis of Spanish. The results I reported in different chapters of the book point to a few differences in realization. I summarize those results as follows. In the thematic analysis, the most striking difference lies in the types of Transitivity elements that function as Themes. Two tendencies run in opposite directions. The first tendency is for English to favour Participant Subjects as Themes, whereas Spanish prefers Processes. The second preference is for Circumstance Subjects as Themes in English, and Circumstance Adjuncts in Spanish. This I interpret as basic characteristics of the differences between the two languages: in Spanish pro-drop and the presence of Adjuncts in fronted position are more common. The other most salient contrast in the thematic analysis lies with the results of thematic progression. Spanish tends to have many more New Themes than English. This result leads me to question the adequacy of the interpretation of English-based thematic progression analysis for Spanish. New Themes in Spanish were mostly Processes. Because those were verbal stems that had not appeared before in the conversation, I labeled them as New. However, they all contain Participant reference, either to the speakers or to the dates being discussed in the text. That is, they are not completely new. I conclude that we need a new consideration of the status of Finites or Processes in Spanish, where we do not classify as New a Process containing Participant reference. Turning now to the analysis of rhetorical structures, the analyses speak of no difference between English and Spanish in their distribution of rhetorical relations. The analysis tested for types of relations found in the corpus, not their linguistic realization. The speakers in both languages seemed to avail themselves of roughly the same relations in the same situations. The most likely conclusion here is that both the problem-solving strategies and the politeness constraints – which were constant across languages – determined similar choices to express relations among propositions. The last realization of texture brings us to cohesion. The differences in the analysis of cohesion devices revealed some problems in the types of cohesion present in Spanish, with no examples of substitution. I question an applicability of substitution to Spanish in the same terms as it is interpreted in English. The other most significant difference is that cohesive chains in Spanish are longer; Spanish speakers use more cohesive devices per chain than their English counterparts. The fifth and last research question searched for a characterization of sequencing regularities in this genre. In Chapter 7, I provide a formalization of the stages in scheduling dialogues. Within each stage, I also include a sequence of speech acts that constitute the structure of the stage. That chapter
TSL[v.20020404] Prn:29/07/2004; 11:43
F: PB12908.tex / p.4 (210)
Chapter 8
includes too a characterization of each stage based on its thematic, rhetorical and cohesive patterns. A number of applications can be derived from this study. From a general point of view, this is a contribution to the study of how people solve tasks and what linguistic structures they use for that purpose. I also believe that it is a contribution to theoretical discourse analysis, since it provides some insights into the analysis of spoken genres from a contrastive point of view. To further narrow the applications, I refer to two major areas that can benefit from my results: second language teaching and computational linguistics. In second language teaching, the complete analysis of the characteristics of any genre provides useful information for learners who will need to be proficient in that genre. Biber and colleagues (Biber et al. 1998) point to the importance of studying register variation in English for Specific Purposes. The goal is to design materials and activities that will help students understand and produce these registers appropriately. Indeed, there are numerous studies and materials that distinguish different written genres or registers, or that compare written and spoken language. A notable exception is Ventola’s (1987) study of service encounters. Insufficient attention has been paid, in general, to the study of different genres within the spoken medium. A formal presentation of the options in a genre, such as I present in Chapter 7, would help in the comprehension and production of that genre. The field of computational linguistics concerns itself with the handling of natural language through computers, for very varied purposes. One application I have described is ambiguity resolution in a machine translation system. Other applications include natural language generation: a complete specification of the characteristics of a genre will help in generating texts belonging to that genre. In information retrieval we aim at extracting relevant portions of a text. Linguistic markers (such as cohesive chains) of portions or stages in a text help in deciding which segments to retrieve. In more general terms, characterizations of different task-oriented conversations lead to an understanding of the most frequent type of human-machine interaction, that where a task or problem needs to be solved.
TSL[v.20020404] Prn:7/10/2004; 9:24
F: PB129AP.tex / p.1 (211)
Appendix A
Speech act inventory
This appendix presents the speech acts found in the corpus, and provides examples of each of them. The general framework under which speech acts were defined is explained in Chapter 7. The examples in this appendix point to the specific speech act being discussed by providing the speech act label in small capitals right before the corresponding section of speech. The approach is descriptive, i.e., each speech act is identified and described according to the contexts in which it appears. I have not pursued a cross-linguistic or cross-cultural approach to the classification of contexts of occurrence.1 . Accept Accept represents the acceptance of a proposal, either a date or a place. The action or content accepted is not repeated (that speech act is called repeatconfirm). Accept occurs both in the Task-Performance stage (Date Proposal or Place Proposal) and in the Closing stage. It is a response to speech acts such as ask-date, propose-action or request-action. It is most often represented by okay in English and by perfecto, sí, está bien, muy bien, bueno, and dále in Spanish. Sometimes the wording is lengthier: a Spanish speaker uses in one occasion será muy bueno a esa hora (‘it’ll be great at that time’). Okay is quite commonly used in the Spanish dialogues: 15 of the instances of accept were realized through the English expression. Example (178) shows an instance of accept in Spanish. Speaker FCBA replies with a pause and perfecto to the other speaker’s proposal. (178) [FCBA_MEBA_08] meba_08_05: bueno. | el dieciséis está bien. | qué tal si nos reunimos este a almorzar a las doce. | y luego nos juntamos de una a tres. fcba_08_06: *pause* perfecto. okay. | the sixteenth is fine. | how about we meet for lunch at noon. | and then we have the meeting from one to three. perfect
TSL[v.20020404] Prn:7/10/2004; 9:24
F: PB129AP.tex / p.2 (212)
Appendix A
. Ask-date A fairly common direct questioning to the other speaker, in both languages. With this speech act, speakers ask about a particular date that is mentioned in the same clause. Ask-date only occurs in the Date-Proposal stage. Here, and in request-date, the word ‘date’ is a general term for both days and times. (179) [FAMS_FCLD_AU] fcld_au_06: ... what about, the eleventh?
. Backchannel Backchannel is a very general term for any expression of feedback. It may occur anywhere, but the most frequent placement is at the beginning of a turn. Speakers acknowledge the information provided by the other speaker with words such as okay, alright, okay then, yeah in English, and bueno, sí, ya, vale, okay. These are also sometimes accompanied by expressions such as uh-huh, mm, lip smacks, or deep breaths. In (180), speaker MRCT acknowledges the information provided by FRET on her unavailability, and then goes on to look for successful dates for the meeting. These expressions are different from the speech act accept in that they constitute an acceptance of information, but not a commitment to a proposal. (180) [FRET_MRCT_1] fret_1_01: okay, /uh/ Mr Taylor, we’re gonna have to schedule, another meeting, in the next two weeks, /um/ | next week, I’ll be out of town, Tuesday Wednesday and Thursday. so, those won’t be good days. mrct_1_02: okay then, /um/...
There are, in the literature, different interpretations of the term ‘backchannel’ and the function of such a construct. The term, first introduced by Yngve (1970), was defined by Duncan (1972) as having five forms: – – – – –
sentence completions requests for clarification brief phrases, such as ‘uh huh’ and ‘right’ head nods head shakes
TSL[v.20020404] Prn:7/10/2004; 9:24
F: PB129AP.tex / p.3 (213)
Speech act inventory
The backchannel elements found in the corpus are always vocalized, since the speakers cannot see each other. Requests for clarification, when present, are defined under the speech act request-information, or maybe requestconfirmation. That leaves us with sentence completions and brief phrases, which still do not exactly define the backchannel elements in the corpus. Schegloff (1982) points out that these items are placed within turn-constructional units. Such a placement permits them to be heard as continuers, acknowledgments, agreements, etc. Heritage (1989) calls them ‘response tokens’: items that gain significance from their placement in sequences of talk. The definition of backchannel used here is in line with those of Schegloff and Heritage: a variety of different elements present at turn boundaries, which serve as acknowledgment and continuation of the conversation.
. Channel Relates to the channel of communication, especially when there are problems in the understanding. This category comprises a number of different speech acts in terms of their illocutionary force; for instance, requests for repetition. The classification in this case is based more on semantic criteria, since channelrelated issues are only marginal to the task. Examples (181) and (182) are representative. In the first one, the speaker announces the lack of understanding. In the Spanish example, the speaker prompts for a confirmation of what was just said. (181) [FCAD_FJAB_12] fjab_12_02: well, /um/ totally just missed /begin_lg/ what you said, /end_lg/ (182) [FVGC_FSNM_09] fsnm_09_06: qué dices. | que tienes entre las *pause* diez y las dos libre? what did you say. | that you have between ten and two free?
. Filled-pause Filled pauses are, traditionally, instances of pauses in the speech that are filled by some type of vocal articulation. Unfilled pauses, on the other hand, are simply realized by silence. Different authors consider only non-words as filled pauses: uh, uhm, mm (Maclay and Oswood 1959; Cook 1971; Romero Trillo
TSL[v.20020404] Prn:7/10/2004; 9:24
F: PB129AP.tex / p.4 (214)
Appendix A
1994). I included under this heading any word or non-word used by the speakers to hold the floor. This could be an okay, oh, um, etc., but also lengthier expressions, such as let’s see. Cortés Rodríguez and Bañón Hernández (1997) consider some of these as delaying expressions (‘expresiones retardatarias’), which prepare the channel in order to begin the communication. They are most frequent at the beginning of a turn. A special instance of a filled pause is the repetition of the other speaker’s words. It serves as a floor-holder while thought processes are taking place. In (183), speaker FKND repeats the date, el lunes (’Monday’), presumably while she is pondering it. After a brief consideration, she decides to start a new date proposal, by suggesting Tuesday. The filled pause here is el lunes. (183) [FKND_FKAS_01] fkas_01_03: ... entonces /uh/ el lunes quince tengo /ah/ mucho tiempo libre. | no estoy ocupada en la mañana, desde las ocho <de la mañana hasta> de la mañana hasta las dos de la tarde. fknd_01_04: el lunes? | y qué tal el martes?... ... then uh on Monday the fifteenth I have uh a lot of free time. | I’m not busy in the morning from eight in the morning until two in the afternoon. Monday? | and how about Tuesday?...
Filled pauses are much more frequent in Spanish than they are in English. They are realized, among other sounds, by bueno, mira, a ver, pero, y, eh. In both languages, they can occur in any of the stages, but they are always more frequent at the beginning of the turn. Cook (1971) argues against the belief that the occurrence of a filled pause is determined by the word the speaker is about to utter – lexical or function word. He concludes that the thought processes that might lead to the presence of a filled pause are more likely to take place before a grammatical chunk, a sentence or clause. This might explain the presence of filled pauses at the beginning of a turn. The speaker is already holding the floor, because their interlocutor has passed it on, but they are not yet ready to utter a complete thought. A filled pause indicates the willingness, yet unreadiness, to speak. In some situations, a filled pause will occur immediately after an unfilled one, as in Example (184). The *pause* marker indicates a silence of up to 0.5 seconds, which is then followed by a ver (’let’s see’). In this example, the speaker moves from considering and accepting a date to the next step, setting up a time.
TSL[v.20020404] Prn:7/10/2004; 9:24
F: PB129AP.tex / p.5 (215)
Speech act inventory
(184) [MRMV_FFCS_06] mrmv_06_03: el miércoles después del almuerzo, | /eh/ sí. | pudiera ser. | /eh/ tú te estás hablando del miércoles /eh/ seis. | *pause* a ver. | especifiquemos una hora... Wednesday after lunch, | uh yes. | it could be. | uh you are talking about Wednesday uh the sixth. | let’s see. | *pause* let’s see. | let’s set a time...
. Goodbye A signal of leave-taking, found exclusively in the Closing stage. Realized by goodbye, bye bye, see you then, among other expressions, in English. In Spanish speakers use expressions such as chau, hasta luego, nos vemos, adiós.
. Greeting Greetings initiate the conversation in the Opening stage, although they are not always present. Only three of them were found in English, and a slightly higher number in Spanish (sixteen). The difference between the two languages in this respect is that in Spanish they are usually corresponded by the other speaker in his or her other turn, whereas in English they are never repeated. This results in the presence of greetings once the Task-Performance stage has been launched. Example (185) shows the first two turns in a conversation. The Opening stage contains a greeting by speaker FKAS, who then starts the Task-Performance phase with her first proposal. When the next speaker takes her turn, she also produces a greeting (hola), but overall continues within the Task-Performance phase, rejecting the first proposal and providing a new one. (185) [FJGC_FKAS_04] fkas_04_01: hola Jeannette. | me gustaría reunirme contigo /uh/ entre el veinticuatro de mayo y el cuatro. | <de viernes |> /eh/ yo tengo libre este lunes el veinticuatro desde el mediodía hasta <el [c(inco)]> <el [c(inco)]> /uh/ las cinco de la tarde. | qué tal para ti reunirnos, el veinticuatro. fjgc_04_02: hola. | mira, | /uh/ quería decirte que el veinticuatro no puedo por la tarde. | /um/ qué tal el miércoles veintiséis <de> a las diez de la mañana. | *pause* de diez a doce... hello Jeannette. | I’d like to meet with you uh between the twenty-fourth of May and the fourth. | uh I’m free this Monday the twenty-
TSL[v.20020404] Prn:7/10/2004; 9:24
F: PB129AP.tex / p.6 (216)
Appendix A
fourth from noon until uh five in the afternoon. | how’s it for you to meet, on the twenty-fourth. hello. | look, | uh I wanted to tell you that on the twenty-fourth I can’t in the afternoon. | um how about Wednesday the twenty-sixth at ten in the morning. | *pause* from ten to twelve...
. Inform This is a general term for any furnishing of information, with the exception of informing on the meeting possibilities, which is included in the informavailability speech act. Both languages make use of it in similar percentages. Examples (186) and (187) are typical instances. The beginning of the inform speech act is signaled in the examples. Inform is the most frequent speech act in both languages. (186) [FEAS_MTMR_11] mtmr_11_06: /eh/ let’s call it two, | inform and I’ll take a long lunch after my seminar. (187) [FPLA_MLMG_10] mlmg_10_06: ... si quieres dáme tu teléfono, | inform y yo le digo a mi secretaria que te recuerde el miércoles la cita que tenemos el jueves seis de mayo... fpla_10_07: /lg/ bueno. | inform mi teléfono es seis dos cuatro ocho ocho cinco tres... ... if you want give me your phone number, | and I tell my secretary to remind you on Wednesday of the appointment we have on Thursday May sixth... okay. | my telephone (number) is six two four eight eight five three...
. Inform-availability After inform, inform-availability is the next most frequent speech act in the corpus. The speakers use this speech act to initiate a Proposal, indicating when they are available. They also employ it to reject a date, saving the positive face of the other speaker, and exercising negative politeness (Brown and Levinson 1978). The other speaker’s positive face is preserved by not rejecting a proposal directly. Negative politeness is achieved because the rejection is presented neu-
TSL[v.20020404] Prn:7/10/2004; 9:24
F: PB129AP.tex / p.7 (217)
Speech act inventory
trally, as a result of the speaker’s busy agenda, and not as a wilful rejection. Inform-availability is most common in the Task-Performance stage. Example (188) shows an instance as initiator of a proposal. In (189), on the other hand, the information on the speaker’s availability is used to reject the Monday date proposed. (188) [FSEH_FPWJ_1] fpwj_1_02: ... what about Tuesday. | I have, time, /ah/ from one, on.... (189) [FSJB_FKDO_1] fsjb_1_01: ... what about /um/ <sometimes> sometime next week, /uh/ maybe, Monday afternoon. after, two o’clock. fkdo_1_02: inform-av /oh/ on Monday the eighth I have a meeting from two to four...
. Other This is an ad hoc category, where I included speech acts that did not seem to fit anywhere else. There are four such instances in English, and two in Spanish, and I did not consider it necessary to create six new labels, given their marginality. For instance, in Example (190), ‘I’m a slacker’ seems to be more than a simple inform, but I did not deem it appropriate to create a new category for self-deprecation. (190) [FEAS_MTMR_11] mtmr_11_02: inform-av I will be vacationing in Siberia then. | inform-av I probably can’t do it ’till next week, | other I’m a slacker.
. Politey Politey is another general category where a number of different phenomena have been bundled together. I decided for this not-so-elegant solution because of the inessential nature of these speech acts with respect to the task. The category includes expressions of thanks, regret, formulaic expressions of desire to meet (I’m looking forward to seeing you), good wishes, and the like.
TSL[v.20020404] Prn:7/10/2004; 9:24
F: PB129AP.tex / p.8 (218)
Appendix A
. Propose-action Proposes a particular action, such as meeting for lunch or doing any other activity. Proposals for a place to meet are listed under a different heading. It is different from request-action because the speaker involves both participants in the proposal. Example (191) shows two instances of this speech act. (191) [MRMV_FFCS_06] mrmv_06_12: ... propose-action podríamos, incluso, reunirnos más de dos horas. | puede ser el día sábado entero, | propose-action podríamos ir hasta al cine... ... we could, even, meet for more than two hours. | it could be the entire Saturday, | we could even go to the cinema...
. Propose-place This is a proposal for a place to meet, most often found in the Place-Proposal stage. Example (192) shows a proposal for a place combined with a confirmation of the date. (192) [FBNT_MJFG_1] fbnt_1_09: ... why don’t we meet, at, the Cafe Arabica on Craig Street at nine AM, for breakfast, on the fifteenth.
. Reject-date A direct rejection of the date proposed by the other speaker. In a few occasions, the rejection is directed toward the date just mentioned by the same speaker. A speaker would read a date aloud, only to realize that it is not feasible, rejecting it in the same turn. Overt rejections are much more common in the Spanish corpus (4.02% of the speech acts, a raw count of 62), versus the 1.97% in the English data (raw count of 20). The speech act inform-availability, as I mentioned earlier, is used in many cases as an implicit rejection, instead of a more direct reject-date. (193) [MAGH_MTNZ_11] magh_11_01: ... would you have, two hours, on Thursday afternoon, the third, or Friday, the fourth. mtnz_11_02: reject-date no, that’s not good, ...
TSL[v.20020404] Prn:7/10/2004; 9:24
F: PB129AP.tex / p.9 (219)
Speech act inventory
(194) [FFCS_FSNM_01] fsnm_01_02: ... el lunes para mí será bien | como, a la hora de almorzar entre las once y media, por allí, hasta la una y media podremos. ffcs_01_03: reject-date no. | el lunes no me queda bien... ... Monday for me is good | around, lunch time between eleven thirty, more or less, until one thirty we can. no. | Monday is not good for me...
. Repeat-confirm An acceptance of a proposal, it includes a repetition of the item agreed upon. (195) [FCKA_FFMW_10] ffmw_10_06: so I will see you, May twentieth, at /uh/, eleven AM. (196) [FVGC_FSNM_09] fvgc_09_09: sí. | perfecto. | repeat-confirm entonces el miércoles catorce, de once a una. yes. | perfect. | then Wednesday the fourteenth, from eleven to one.
. Request-action Request-action is very similar to propose-action. The latter usually includes both speakers in the proposal. Request-action is also somewhat more forceful in demanding an action or an agreement to comply with the action from the other speaker. In traditional terms, this could be represented under a directive. It is represented in similar percentages in both corpora, with percentages below 1%. Example (197) shows the difference between the two. The speaker first demands an action from her interlocutor, and then proposes a course of action in which she will be included. (197) [FACR_FJYK_AU] fjyk_au_04: ... request-action how ’bout, you come over to my office. propose-action and then, perhaps we can go over to the conference room that’s right next to my office...
TSL[v.20020404] Prn:7/10/2004; 9:24
F: PB129AP.tex / p.10 (220)
Appendix A
. Request-confirmation This speech act demands a verbal agreement to a proposal. It is commonly a short expression of the type how’s that?. In the analysis of Rhetorical Structure Theory (see Chapter 5), clauses of this type were analyzed as the satellite in a Condition relation, since they condition the proposal to the other speaker’s acceptance. It can be realized, in terms of clause type, through an if conditional (198) or an interrogative (199). (198) [FAMS_FCLD_AU] fcld_au_04: ... inform-av but, I’m completely free on Wednesday the twenty third, request-conf if that would work out with you, (199) [MARC_MPHB_02] mphb_02_04: ... me viene mejor el jueves, | por ejemplo empezar a las dos de la tarde. | request-conf qué te parece? ... it’s better for me on Thursday, | for instance starting at two in the afternoon. | what do you think?
. Request-date Request-date occurs, for the most part, during the first Date-Proposal stage, where one of the speakers expresses the need to meet (through a requestmeeting speech act), prompting the other speaker for a suitable date. Example (200) is the very first turn in a conversation. No Opening is present in this dialogue; the speaker immediately prompts his interlocutor for a date. (200) [MJNM_FAMM_06] mjnm_06_01: cuándo te quieres reunir? when would you like to meet?
. Request-information A general term for any request for information, which is different from the previous request for a date (also information). (201) [FECE_MNFH_10] mnfh_10_04: /uh/ /ah/, | /ah/ can you come over to my office at twelve so, | and we’ll decide where to go from, there. fece_10_05: okay. | request-info where’s your office.
TSL[v.20020404] Prn:7/10/2004; 9:24
F: PB129AP.tex / p.11 (221)
Speech act inventory
(202) [MEJH_MJJG_01] mjjg_01_03: de qué de qué lunes estás hablando. | del ocho? o del quince... which <Mon> Monday are you talking about. | the eighth? or the fifteenth...
. Request-meeting This speech act is present in the first few utterances of the Task-Performance stage. The speakers express the need to meet, and then schedule the date for the meeting. (203) [FJMB_FMLZ_6] fjmb_6_01: well Marcy, request-meeting looks like we, need to schedule another meeting, in the next couple of weeks, /um/...
. Request-place A request for a proposal for a place to meet. It usually happens inside a PlaceProposal stage, but sometimes also in the Closing stage, once the date has been agreed, and the speakers are saying goodbye. (204) [FMEM_MEOC_02] fmem_02_04: ... /mm/ lo que me parece más bien que definamos es a dónde quieres ir? ... mm what I think we should define is where you want to go?
. Self-introduction The conversations resemble telephone conversations because the speakers do not face each other. Some speakers adopt a telephone conversation genre, and thus utilize some of the devices present in those. Such is the case with self introductions, where the speaker announces his or her name. There are only two instances of this speech act in English, and none in Spanish. (205) [FRJP_MPMM_1] frjp_1_01: /ah/ vocative Peter. self-intro this is Renee.
TSL[v.20020404] Prn:7/10/2004; 9:24
F: PB129AP.tex / p.12 (222)
Appendix A
. Vocative An address to the other person, usually by name. It occurs anywhere in the conversations, most frequently in the Opening stage. An example can be found in (205), under the self-introduction speech act.
Note . Examples of cross-linguistic and cross-cultural studies are Blum-Kulka et al. (1989) and Márquez Reiter (2000; 2003).
TSL[v.20020404] Prn:7/10/2004; 9:29
F: PB129RE.tex / p.1 (223)
References
Abelen, Eric, Redeker, Gisela, and Thompson, Sandra A. (1993). “The rhetorical structure of US-American and Dutch fund-raising letters”. Text, 13 (3), 323–350. Ahlén, Sondra (1996). “Spontaneous Speech Data Collection: Scheduling Domain, VerbMobil Fourth Report and Shipment”. Language Technologies Institute Technical Report. Carnegie Mellon University. Ahlén, Sondra (1997). “Enthusiast Data Collection”. Language Technologies Institute Technical Report. Carnegie Mellon University and The University of Pittsburgh. Ahlén, Sondra and Vaidya, Anuj (1997). “Extended Spontaneous Speech Data Collection Travel Domain”. Language Technologies Institute Technical Report. Carnegie Mellon University. Aijmer, Karin (1996). Conversational Routines in English: Convention and Creativity. London: Longman. Aijmer, Karin (2002). English Discourse Particles: Evidence from a Corpus. Amsterdam and Philadelphia: John Benjamins. Allen, James and Core, Mark (1997). “Draft of DAMSL: Dialogue Markup in Several Layers”. Draft produced by the Discourse Research Initiative (DRI) in meetings at the University of Pennsylvania and Schloss Dagstuhl. Anderson, Richard C., Reynolds, R. E., Schallert, D. L., and Goetz, E. T. (1977). “Frameworks for comprehending discourse”. American Educational Research Journal, 14, 367–381. Anscombre, Jean-Claude and Ducrot, Oswald (1983). L’argumentation dans la langue. Brussels: Pierre Mardaga. Asher, Nicholas (1993). Reference to Abstract Objects in Discourse. Dordrecht: Kluwer. Asher, Nicholas and Lascarides, Alex (2003). Logics of Conversation. Cambridge: Cambridge University Press. Askehave, Inger and Swales, John M. (2001). “Genre identification and communicative purpose: A problem and a possible solution”. Applied Linguistics, 22 (2), 195–212. Austin, J. L. (1962). How to Do Things with Words. Harvard: Harvard University Press. Bäcklund, Ingegerd (1992). “Theme in English telephone conversation”. Language Sciences, 14 (4), 545–564. Badger, Richard (2003). “Legal and general: Towards a genre analysis of newspaper law reports”. English for Specific Purposes, 22 (3), 249–263. Bakhtin, Mikhail (1981). “Discourse in the novel”. In M. Holquist (Ed.), The Dialogic Imagination: Four Essays by M. M. Bakhtin (pp. 259–422). Austin: University of Texas Press.
TSL[v.20020404] Prn:7/10/2004; 9:29
F: PB129RE.tex / p.2 (224)
References
Bakhtin, Mikhail (1986). Speech Genres and Other Late Essays. Austin: University of Texas Press. Ballard, D. Lee, Conrad, Robert J., and Longacre, Robert E. (1971a). “Interclausal relations”. Foundations of Language, 7 (1), 70–118. Ballard, D. Lee, Conrad, Robert J., and Longacre, Robert E. (1971b). More on the Deep and Surface Grammar of Interclausal Relations. Santa Ana, CA: Summer Institute of Linguistics. Bargiela-Chiappini, Francesca and Harris, Sandra (1997). Managing Language: The Discourse of Corporate Meetings. Amsterdam and Philadelphia: John Benjamins. Bargiela-Chiappini, Francesca and Nickerson, Catherine (Eds.). (1999). Writing Business: Genres, Media and Discourses. London: Longman. Bargiela-Chiappini, Francesca and Nickerson, Catherine (2002). “Business discourse: Old debates, new horizons”. International Review of Applied Linguistics in Language Teaching (IRAL), 40 (4), 273–286. Bartlett, Frederic C. (1932). Remembering. Cambridge: Cambridge University Press. Bateman, John and Rondhuis, Klaas Jan (1997). “Coherence relations: Towards a general specification”. Discourse Processes, 24, 3–49. Bateman, John, Kamps, Thomas, Kleinz, Jörg, and Reichenberger, Klaus (2001). “Towards constructive text, diagram, and layout generation for information presentation”. Computational Linguistics, 27 (3), 409–449. Beekman, John and Callow, John (1974). Translating the Word of God. Grand Rapids, MI: Zondervan Publishing House. Bello, Andrés (1984 [1847]). Gramática de la lengua castellana. Madrid: Edaf. Bentivoglio, Paola (1983). “Topic continuity and discontinuity in discourse: A study of spoken Latin-American Spanish”. In T. Givón (Ed.), Topic Continuity in Discourse: A Quantitative Cross-Language Study (pp. 255–311). Amsterdam and Philadelphia: John Benjamins. Bernárdez, Enrique (1982). Introducción a la lingüística del texto. Madrid: Espasa-Calpe. Bernárdez, Enrique (1995). Teoría y epistemología del texto. Madrid: Cátedra. Berry, Margaret (1981). “Towards layers of exchange structure for directive exchanges”. Network, 2, 23–32. Berry, Margaret (1987). “The functions of place names”. In T. Turville-Peter and M. Gelling (Eds.), Leeds Studies in English, New Series XVIII: Studies in Honour of Kenneth Cameron (pp. 71–88). Leeds: University of Leeds. Berry, Margaret (1996). “What is Theme? A(nother) personal view”. In M. Berry, C. Butler, R. Fawcett, and G. Huang (Eds.), Meaning and Form: Systemic Functional Interpretations (Meaning and Choice in Language: Studies for Michael Halliday) (pp. 1–64). Norwood, NJ: Ablex. Bhatia, Vijay K. (1993). Analysing Genre: Language Use in Professional Settings. London: Longman. Bhatia, Vijay K. (1997). “Introduction: Genre analysis and world Englishes”. World Englishes, 16 (3), 313–319. Biber, Douglas (1988). Variation across Speech and Writing. Cambridge: Cambridge University Press.
TSL[v.20020404] Prn:7/10/2004; 9:29
F: PB129RE.tex / p.3 (225)
References
Biber, Douglas (1995). Dimensions of Register Variation: A Cross-Linguistic Comparison. Cambridge: Cambridge University Press. Biber, Douglas and Conrad, Susan (2001). “Register variation: A corpus approach”. In D. Schiffrin, D. Tannen, and H. E. Hamilton (Eds.), The Handbook of Discourse Analysis (pp. 175–196). Malden, Mass: Blackwell. Biber, Douglas, Conrad, Susan, and Reppen, Randi (1998). Corpus Linguistics: Investigating Language Structure and Use. Cambridge: Cambridge University Press. Biber, Douglas and Finegan, Edward (1986). “An initial typology of English text types”. In J. Aarts and W. Meijs (Eds.), Corpus Linguistics II: New Studies in the Analysis and Explotation of Computer Corpora (pp. 19–46). Amsterdam: Rodopi. Biber, Douglas and Finegan, Edward (1989). “Drift and the evolution of English style: A history of three genres”. Language, 65 (3), 487–517. Biber, Douglas, Johansson, Stig, Leech, Geoffrey, Conrad, Susan, and Finegan, Edward (1999). Longman Grammar of Spoken and Written English. Harlow, Essex: Pearson Education. Blum-Kulka, Shoshana, House, Juliane, and Kasper, Gabriele (Eds.). (1989). Cross-Cultural Pragmatics: Requests and Apologies. Norwood, NJ: Ablex. Bolinger, Dwight L. (1954). “English prosodic stress and Spanish sentence order”. Hispania, 37, 152–156. Bolinger, Dwight L. (1955). “Meaningful word order in Spanish”. Boletín de filología Universidad de Chile, 7, 45–56. Bouayad-Agha, Nadjet (2000). “Using an abstract rhetorical representation to generate a variety of pragmatically congruent texts”. In Companion Volume to the Proceedings of the 38th Meeting of the Association for Computational Linguistics (ACL’00), Student Workshop (pp. 16–22). Hong Kong. Brazil, David (1995). A Grammar of Speech. Oxford: Oxford University Press. Bregman, Alvan and Haythornthwaite, Caroline (2003). “Radicals of presentation: Visibility, relation, and co-presence in persistent conversation”. New Media and Society, 5 (1), 117–140. Brinton, Laurel J. (1996). Pragmatic Markers in English: Grammaticalization and Discourse Functions. Berlin: Mouton de Gruyter. Briz, Antonio and Hidalgo, Antonio (1988). “Conectores pragmáticos y estructura de la conversación”. In M. A. M. Zorraquino and E. M. Durán (Eds.), Los marcadores del discurso: Teoría y análisis (pp. 121–142). Madrid: Arco. Brown, Gillian and Yule, George (1983). Discourse Analysis. Cambridge: Cambridge University Press. Brown, Gretchen P. (1984). “Linguistic and situational context in a model of task-oriented dialogue”. In L. Vaina and J. Hintikka (Eds.), Cognitive Constraints on Communication (pp. 229–240). Dordrecht: Reidel. Brown, Penelope and Levinson, Stephen (1987 [1978]). Politeness: Some Universals in Language Usage. Cambridge: Cambridge University Press. Brown, T. Pascal and Lewis, Marilyn (2002). “An ESP project: Analysis of an authentic workplace conversation”. English for Specific Purposes, 22 (1), 93–98.
TSL[v.20020404] Prn:7/10/2004; 9:29
F: PB129RE.tex / p.4 (226)
References
Bublitz, Wolfram (1999). “Introduction: Views of coherence”. In W. Bublitz, U. Lenk, and E. Ventola (Eds.), Coherence in Spoken and Written Discourse: How to Create It and How to Describe It (pp. 1–7). Amsterdam and Philadelphia: John Benjamins. Büring, Daniel (1997). The Meaning of Topic and Focus: The 59th Street Bridge Accent. London: Routledge. Butt, John and Benjamin, Carmen (2000). A New Reference Grammar of Modern Spanish (3rd edition). Chicago: McGraw-Hill. Byron, Donna K. and Heeman, Peter A. (1998). “Discourse Marker Use in Task-Oriented Spoken Dialog”. Computer Science Technical Report 664. University of Rochester. Caffarel, Alice (2000). “Interpreting French Theme as a bi-layered structure: Discourse implications”. In E. Ventola (Ed.), Discourse and Community (pp. 247–272). Tübingen: Gunter Narr Verlag. Carlson, Matthew (2003). “Preverbal elements in non-agentive Spanish predicates: Theta roles, case, and control”. In P. Kempchinsky and C.-E. Piñeros (Eds.), Theory, Practice, and Acquisition: Papers from the 6th Hispanic Linguistics Symposium and the 5th Conference on the Acquisition of Spanish and Portuguese (pp. 121–139). Somerville, MA: Cascadilla Press. Casado Velarde, Manuel (1997). Introducción a la gramática del texto del español. Madrid: Arco. Cawsey, Alison (1990). “Generating explanatory discourse”. In R. Dale, C. Mellish, and M. Zock (Eds.), Current Research in Natural Language Generation (pp. 75–101). London: Academic Press. Cha, Jin Soon (1985). Linguistic Cohesion in Texts: Theory and Description. Seoul: Daehan Textbook Printing. Chafe, Wallace (1980). “The deployment of consciousness in the production of a narrative”. In W. Chafe (Ed.), The Pear Stories: Cognitive, Cultural and Narrative Aspects of Narrative Production (pp. 9–50). Norwood, NJ: Ablex. Chomsky, Noam (1971). “Deep structure, semantic structure and surface representation”. In D. Steinberg and L. Jakobovits (Eds.), Semantics: An Interdisciplinary Reader in Philosophy, Linguistics and Psychology (pp. 183–216). Cambridge: Cambridge University Press. Christie, Frances (Ed.). (1985). Children Writing: Reader. Geelong, Victoria: Deakin University Press. Christie, Frances (1994). “Developing an educational linguistics for English language teaching: A systemic functional linguistic perspective”. Functions of Language, 1 (1), 95–127. Christie, Frances and Martin, James R. (Eds.). (1997). Genre and Institutions: Social Processes in the Workplace and School. London: Cassell. Clark, Herbert H. (1996). Using Language. Cambridge: Cambridge University Press. Clark, Herbert H. and Schaefer, E. F. (1987). “Collaborating on contributions to conversation”. Language and Cognitive Processes, 2 (1), 19–41. Clark, Herbert H. and Schaefer, E. F. (1989). “Contributing to discourse”. Cognitive Science, 13, 259–294. Clark, Herbert H. and Wilkes-Gibbes, Deanna (1986). “Referring as a collaborative process”. Cognition, 22, 1–39.
TSL[v.20020404] Prn:7/10/2004; 9:29
F: PB129RE.tex / p.5 (227)
References
Cloran, Carmel (1995). “Defining and relating text segments: Subject and Theme in discourse”. In R. Hasan and P. H. Fries (Eds.), On Subject and Theme: A Discourse Functional Perspective (pp. 361–403). Amsterdam and Philadelphia: John Benjamins. Coe, Richard M. and Freedman, Aviva (1998). “Genre Theory: Australian and North American approaches”. In M. L. Kennedy (Ed.), Theorizing Composition: A Critical Sourcebook of Theory and Scholarship in Contemporary Composition Studies (pp. 136– 147). Wesport, CT: Greenwood. Contreras, Heles (1978). El orden de palabras en español. Madrid: Cátedra. Cook, Mark (1971). “The incidence of filled pauses in relation to part of speech”. Language and Speech, 14, 135–139. Cope, Bill and Kalantzis, Mary (Eds.). (1993). The Powers of Literacy: A Genre Approach to Teaching Writing. London: The Falmer Press. Corston-Oliver, Simon (1998). “Beyond string matching and cue phrases: Improving efficiency and coverage in discourse analysis”. In Proceedings of AAAI 1998 Spring Symposium Series, Intelligent Text Summarization (pp. 9–15). Madison, Wisconsin. Cortés Rodríguez, Luis and Bañón Hernández, Antonio M. (1997). Comentario lingüístico de textos orales I. Teoría y práctica (La Tertulia). Madrid: Arco. Coulthard, Malcolm and Montgomery, Martin (Eds.). (1979). Studies in Discourse Analysis. London: Routledge and Kegan Paul. Crystal, David (1969). Prosodic Systems and Intonation in English. London: Cambridge University Press. Cui, Songren (1986). A Comparison of English and Chinese Expository Rhetorical Structures. Master’s thesis. UCLA. Cummings, Michael (1995). “A systemic functional approach to the thematic structure of the Old English clause”. In R. Hasan and P. H. Fries (Eds.), On Subject and Theme: A Discourse Functional Perspective (pp. 275–316). Amsterdam and Philadelphia: John Benjamins. Dahl, Östen (1969). Topic and Comment: A Study in Russian and General Transformational Grammar. Göteborg: Acta Universitatis Gothoburgensis. Dahl, Östen (Ed.). (1974). Topic and Comment, Contextual Boundness and Focus. Hamburg: Helmut Buske. Daneš, František (1974a). “Functional Sentence Perspective and the organization of the text”. In F. Daneš (Ed.), Papers on Functional Sentence Perspective (pp. 106–128). Prague: Academia. Daneš, František (Ed.). (1974b). Papers on Functional Sentence Perspective. The Hague: Mouton. Daradoumis, Thanasis (1996). “Towards a representation of the rhetorical structure of interrupted exchanges”. In G. Adorni and M. Zock (Eds.), Trends in Natural Language Generation: An Artificial Intelligence Perspective (pp. 106–124). Berlin: Springer Verlag. de Geest, Dirk and van Gorp, Hendrik (1999). “Literary genres from a systemic-functionalist perspective”. European Journal of English Studies, 3 (1), 33–50. Dorval, Bruce (Ed.) (1990). Conversational Organization and Its Development. Norwood, NJ: Ablex. Downing, Angela (1990). “Sobre el tema tópico en inglés”. Revista española de lingüística aplicada, Anejo I, 119–128.
TSL[v.20020404] Prn:7/10/2004; 9:29
F: PB129RE.tex / p.6 (228)
References
Downing, Angela (1991). “An alternative approach to Theme: A systemic-functional perspective”. Word, 42 (2), 119–143. Downing, Angela (2000). “Talking topically”. In A. Downing, A. J. Moya Guijarro, and J. I. Albentosa Hernández (Eds.), Talk and Text: Studies on Spoken and Written Discourse (pp. 31–50). Cuenca: Universidad de Castilla-La Mancha. Downing, Angela and Locke, Philip (2002). A University Course in English Grammar (2nd edition). New York: Routledge. Dryer, Matthew S. (1995). “Frequency and pragmatically unmarked word order”. In P. Downing and M. Noonan (Eds.), Word Order in Discourse (pp. 105–135). Amsterdam and Philadelphia: John Benjamins. Du Bois, John W., Schuetze-Coburn, Stephan, Cumming, Susanna, and Paolino, Danae (1993). “Outline of discourse transcription”. In J. A. Edwards and M. D. Lampert (Eds.), Talking Data: Transcription and Coding in Discourse Research (pp. 45–89). Hillsdale, NJ: Lawrence Erlbaum. Dubois, Betty Lou (1987). “A reformulation of thematic progression typology”. Text, 7 (2), 89–116. Dudley-Evans, Tony (1994). “Genre analysis: An approach to text analysis for ESP”. In M. Coulthard (Ed.), Advances in Written Text Analysis (pp. 219–228). London and New York: Routledge. Dudley-Evans, Tony and Henderson, Willie (Eds.). (1990). The Language of Economics: The Analysis of Economics Discourse. London: Modern English Publications. Duncan, Starkey (1972). “Some signals and rules for taking speaking turns in conversations”. Journal of Personality and Social Psychology, 23 283–292. Edelsky, Carole (1981). “Who’s got the floor?”. Language in Society, 10, 383–421. Edwards, Jane A. and Lampert, Martin D. (Eds.). (1993). Talking Data: Transcription and Coding in Discourse Research. Hillsdale, NJ: Lawrence Erlbaum. Eggins, Suzanne (1994). An Introduction to Systemic Functional Linguistics. London: Pinter. Eggins, Suzanne and Martin, James R. (1997). “Genres and registers of discourse”. In T. A. van Dijk (Ed.), Discourse as Structure and Process. Discourse Studies: A Multidisciplinary Introduction (pp. 230–256). London: Sage. Eggins, Suzanne and Slade, Diana (1997). Analysing Casual Conversation. London: Cassell. Erteschik-Shir, Nomi (1997). The Dynamics of Focus Structure. Cambridge: Cambridge University Press. Esgueva, Manuel and Cantarero, Margarita (1981). El habla culta de la ciudad de Madrid: Materiales para su estudio. Madrid: CSIC. Fairclough, Norman (1992). Discourse and Social Change. Cambridge: Polity Press. Fairclough, Norman (1995). Critical Discourse Analysis. London: Longman. Fant, Lars (1996). “Regulación conversacional en la negociación: Una comparación entre pautas mexicanas y peninsulares”. In T. Kotschi, W. Oesterreicher, and K. Zimmermann (Eds.), El español hablado y la cultura oral en España e Hispanoamérica (pp. 147–183). Madrid: Iberoamericana. Fawcett, Robin P. and Davies, Bethan L. (1992). “Monologue as a turn in dialogue: Towards an integration of Exchange Structure and Rhetorical Structure Theory”. In R. Dale, E. Hovy, D. Rösner, and O. Stock (Eds.), Aspects of Automated Language Generation (pp. 151–166). Berlin: Springer Verlag.
TSL[v.20020404] Prn:7/10/2004; 9:29
F: PB129RE.tex / p.7 (229)
References
Félix-Brasdefer, J. César (2003). “Declining an invitation: A cross-cultural study of pragmatic strategies in American English and Latin American Spanish”. Multilingua, 22 (3), 225–255. Ferguson, Charles A. (1981). “The structure and use of politeness formulas”. In F. Coulmas (Ed.), Conversational Routine: Explorations in Standardized Communication Situations and Prepatterned Speech (pp. 21–35). The Hague: Mouton. Fernández Soriano, Olga (1999). “El pronombre personal: Formas y distribuciones. Pronombres átonos y tónicos”. In Gramática descriptiva de la lengua española, I. Bosque and V. Demonte (Eds.), Volume 1: Sintaxis básica de las clases de palabras (pp. 1209– 1273). Madrid: Espasa. Firbas, Jan (1974). “Some aspects of the Czechoslovak approach to problems of Functional Sentence Perspective”. In F. Daneš (Ed.), Papers on Functional Sentence Perspective (pp. 11–37). Prague: Academia. Firbas, Jan (1992). Functional Sentence Perspective in Written and Spoken Communication. Cambridge: Cambridge University Press. Firth, J. R. (1968). Selected Papers of J. R. Firth. Bloomington/London: Indiana University Press. Firth, Raymond (1972). “Verbal and bodily rituals of greeting and parting”. In J. S. L. Fontaine (Ed.), Interpretation of Ritual (pp. 1–38). London: Tavistock. Fox, Barbara A. (1987). Discourse Structure and Anaphora: Written and Conversational English. Cambridge: Cambridge University Press. Foz, M. Carmen (2000). “A contribution to a descriptive functional-semantic framework for Theme analysis”. Word, 53 (2), 173–183. Francis, Gill (1989). “Thematic selection and distribution in written discourse”. Word, 40 (1–2), 201–221. Freedman, Aviva (1999). “Beyond the text: Towards understanding the teaching and learning of genres”. TESOL Quarterly, 33 (4), 764–767. Freedman, Aviva and Medway, Peter (Eds.). (1994a). Genre and the New Rhetoric. London: Taylor and Francis. Freedman, Aviva and Medway, Peter (Eds.). (1994b). Learning and Teaching Genre. Portsmouth, NH: Boynton/Cook. Fries, Peter H. (1983). “On the status of Theme in English: Arguments from discourse”. In J. Petöfi and E. Sözer (Eds.), Micro and Macro Connexity of Texts (pp. 116–152). Hamburg: Helmut Buske. Fries, Peter H. (1994). “On Theme, Rheme and discourse goals”. In M. Coulthard (Ed.), Advances in Written Text Analysis (pp. 229–249). London: Routledge. Fries, Peter H. (1995a). “A personal view of Theme”. In M. Ghadessy (Ed.), Thematic Development in English Texts (pp. 1–19). London: Pinter. Fries, Peter H. (1995b). “Patterns of information in initial position in English”. In P. H. Fries and M. Gregory (Eds.), Discourse in Society: Systemic Functional Perspectives. Meaning and Choice in Language: Studies for Michael Halliday (pp. 47–66). Norwood, NJ: Ablex. Fries, Peter H. (1995c). “Themes, methods of development, and texts”. In R. Hasan and P. H. Fries (Eds.), On Subject and Theme: A Discourse Functional Perspective (pp. 317–359). Amsterdam and Philadelphia: John Benjamins.
TSL[v.20020404] Prn:7/10/2004; 9:29
F: PB129RE.tex / p.8 (230)
References
Fuller, Daniel P. (1959). The Inductive Method of Bible Study. Pasadena: Fuller Theological Seminary. Gallardo Paúls, Beatriz (1998). Comentario de textos conversacionales I. De la teoría al comentario. Madrid: Arco. Garrido Medina, Joaquín (1997). Estilo y texto en la lengua. Madrid: Gredos. Geis, Michael L. (1995). Speech Acts and Conversational Interaction. Cambridge: Cambridge University Press. Ghadessy, Mohsen (Ed.) (1988). Registers of Written English: Situational Factors and Linguistic Features. London: Pinter. Ghadessy, Mohsen (Ed.). (1995). Thematic Development in English Texts. London: Pinter. Ghadessy, Mohsen (1997). “Thematic organisation as a criterion for establishing English text-types”. Estudios ingleses de la Universidad Complutense, 5, 129–145. Giannoni, Davide Simone (2002). “World of gratitude: A contrastive study of acknowledgement texts in English and Italian research articles”. Applied Linguistics, 23 (1), 1–31. Giora, Rachel (1983). “Segmentation and segment cohesion: On the thematic organization of the text”. Text, 3 (2), 155–181. Givón, Talmy (1979). On Understanding Grammar. New York: Academic Press. Givón, Talmy (1983). “Topic continuity in discourse: An introduction”. In T. Givón (Ed.), Topic Continuity in Discourse: A Quantitative Cross-Language Study (pp. 1–41). Amsterdam and Philadelphia: John Benjamins. Gómez-González, María de los Ángeles (1995). “Theme: A heuristic method for discourse analysis”. Estudios ingleses de la Universidad Complutense, 3, 43–54. Gómez-González, María de los Ángeles (2001). The Theme-Topic Interface: Evidence from English. Amsterdam and Philadelphia: John Benjamins. Goodwin, Charles (1981). Conversational Organization: Interaction between Speakers and Hearers. New York: Academic Press. Gregory, Michael (1988). “Generic situation and register: A functional view of communication”. In J. D. Benson, M. J. Cummings, and W. S. Greaves (Eds.), Linguistics in a Systemic Perspective (pp. 301–329). Amsterdam and Philadelphia: John Benjamins. Gregory, Michael and Carroll, Susanne (1978). Language and Situation: Language Varieties and their Social Contexts. London: Routledge and Kegan Paul. Grimes, Joseph E. (1975). The Thread of Discourse. The Hague: Mouton. Grosz, Barbara J. and Sidner, Candace L. (1986). “Attention, intentions, and the structure of discourse”. Computational Linguistics, 12 (3), 175–204. Gundel, Jeanette K. (1977). Role of Topic and Comment in Linguistic Theory. Bloomington: Indiana University Linguistics Club. Gundel, Jeanette K. (1999). “On different kinds of focus”. In P. Bosch and R. v. d. Sandt (Eds.), Focus: Linguistic, Cognitive, and Computational Perspectives (pp. 293–305). Cambridge: Cambridge University Press. Gundel, Jeanette K. and Fretheim, Thorstein (2004). “Topic and focus”. In L. Horn and G. Ward (Eds.), The Handbook of Pragmatics (pp. 175–196). Oxford: Blackwell. Gutiérrez Ordóñez, Salvador (1997). Temas, remas, focos, tópicos y comentarios. Madrid: Arco.
TSL[v.20020404] Prn:7/10/2004; 9:29
F: PB129RE.tex / p.9 (231)
References
Hajiˇcová, Eva (1987). “Focussing: A meeting point of linguistics and artificial intelligence”. In P. Jorrand and V. Sgurev (Eds.), Artificial Intelligence II (pp. 311–321). Amsterdam: Elsevier. Hajiˇcová, Eva (1991). “Topic-focus articulation and coreference in models of discourse production”. Journal of Pragmatics, 16, 157–166. Hajiˇcová, Eva, Korbayová, Ivana, and Kruijff, Geert-Jan (1998a). “Salience in dialogues”. In S. Cmejrková and et al. (Eds.), Proceedings of Dialogue Analysis VI. Proceedings of the 6th IADA Congress in Prague (pp. 381–393). Tübingen: Max Niemeyer. Hajiˇcová, Eva, Partee, Barbara H., and Sgall, Petr (1998b). Topic-Focus Articulation, Tripartite Structures, and Semantic Content. Dordrecht: Kluwer. Hajiˇcová, Eva and Sgall, Petr (1988). “Topic and focus of a sentence and the patterning of a text”. In J. S. Petöfi (Ed.), Text and Discourse Constitution: Empirical Aspects, Theoretical Approaches (pp. 70–96). Berlin and New York: Walter de Gruyter. Halliday, Michael A. K. (1967). “Notes in Transitivity and Theme in English. Part II”. Journal of Linguistics, 3, 199–244. Halliday, Michael A. K. (1978). Language as Social Semiotic: The Social Interpretation of Language and Meaning. London: Edward Arnold. Halliday, Michael A. K. (1989 [1985]). “Context of situation”. In M. A. K. Halliday and R. Hasan (Eds.), Language, Context and Text (pp. 3–14). Oxford: Oxford University Press. Halliday, Michael A. K. (1994). An Introduction to Functional Grammar (2nd edition). London: Edward Arnold. Halliday, Michael A. K. and Hasan, Ruqaiya (1976). Cohesion in English. London: Longman. Halliday, Michael A. K. and Hasan, Ruqaiya (1985). Language, Context, and Text: Aspects of Language in a Social-Semiotic Perspective. Oxford: Oxford University Press. Halliday, Michael A. K., MacIntosh, Angus, and Strevens, Peter (1964). The Linguistic Sciences and Language Teaching. London: Longman. Harold, Bruce B. (1995). “Subject-verb word order and the function of early position”. In P. Downing and M. Noonan (Eds.), Word Order in Discourse (pp. 137–161). Amsterdam and Philadelphia: John Benjamins. Harris, Zellig (1963). Discourse Analysis: Papers on Formal Linguistics. The Hague: Mouton. Hasan, Ruqaiya (1977). “Text in the systemic-functional model”. In W. Dressler (Ed.), Current Trends in Textlinguistics (pp. 228–246). Berlin and New York: Walter de Gruyter. Hasan, Ruqaiya (1984a). “Coherence and cohesive harmony”. In J. Flood (Ed.), Understanding Reading Comprehension (pp. 181–219). Newark, DE: International Reading Association. Hasan, Ruqaiya (1984b). “The nursery tale as genre”. Nottingham Linguistics Circular, 13, 71–102. Hasan, Ruqaiya (1992). “Speech genre, semiotic mediation and the development of higher mental functions”. Language Sciences, 14 (4), 489–528. Hasan, Ruqaiya and Fries, Peter H. (Eds.). (1995a). On Subject and Theme: A Discourse Functional Perspective. Amsterdam and Philadelphia: John Benjamins.
TSL[v.20020404] Prn:7/10/2004; 9:29
F: PB129RE.tex / p.10 (232)
References
Hasan, Ruqaiya and Fries, Peter H. (1995b). “Reflections on subject and Theme”. In R. Hasan and P. H. Fries (Eds.), On Subject and Theme: A Discourse Functional Perspective (pp. xiii–xiv). Amsterdam and Philadelphia: John Benjamins. Haswell, Richard H. (1988). “Critique: Length of text and the measurement of cohesion”. Research in the Teaching of English, 22 (4), 428–433. Hatch, Evelyn and Lazaraton, Anne (1991). Design and Statistics for Applied Linguistics. Boston: Heinle and Heinle. Hatcher, Ann G. (1956a). “Syntax and the sentence”. Word, 12, 234–250. Hatcher, Ann G. (1956b). “Theme and underlying question: Two studies of Spanish word order”. Supplement to Word, 12, 1–52. Hayashi, R. (1991). “Floor structure of English and Japanese conversation”. Journal of Pragmatics, 16, 1–30. Heritage, John (1989). “Current developments in conversation analysis”. In D. Roger and P. Bull (Eds.), Conversation: An Interdisciplinary Perspective (pp. 21–47). Clevedon: Multilingual Matters. Hernández Cabrera, Clara, Samper Padilla, José Antonio, and Troya Déniz, Magnolia (Eds.). (1998). Macrocorpus de la norma lingüística culta de las principales ciudades del mundo. Las Palmas de Gran Canaria: Servicio de Publicaciones de la Universidad de Las Palmas de Gran Canaria. Hetzron, Robert (1971). “Presentative function and presentative movement”. Studies in African Linguistics, Supplement 2, 79–105. Hobbs, Jerry (1979). “Coherence and coreference”. Cognitive Science, 6, 67–90. Hobbs, Jerry (1990a). Literature and Cognition. Menlo Park, CA: CSLI. Hobbs, Jerry (1990b). “Topic drift”. In B. Dorval (Ed.), Conversational Organization and its Development (pp. 3–22). Norwood, NJ: Ablex. Hockett, Charles F. (1958). A Course in Modern Linguistics. New York: Macmillan. Hoey, Michael (1983). On the Surface of Discourse. London: George Allen and Unwin. Hoey, Michael (1991). Patterns of Lexis in Text. Oxford: Oxford University Press. Hopcroft, John E. and Ullman, Jeffrey D. (1979). Introduction to Automata Theory, Languages, and Computation. Reading, Mass: Addison-Wesley. Hopper, Paul (1988). “Emergent grammar and the a priori grammar postulate”. In D. Tannen (Ed.), Linguistics in Context: Connecting Observation and Understanding (pp. 117–134). Norwood, NJ: Ablex. Hovy, Eduard (1993). “Automated discourse generation using discourse structure relations”. Artificial Intelligence, 63, 341–385. Hovy, Eduard and Maier, Elisabeth (1992). “Parsimonious of Profligate: How Many and Which Discourse Structure Relations?”. Technical Report ISI/RR-93-373. Information Sciences Institute. Huddleston, Rodney (1978). “On classifying anaphoric relations: A review of M. A. K. Halliday and R. Hasan, Cohesion in English”. Lingua, 45, 333–354. Hunt, Kellogg W. (1977). “Early blooming and late blooming syntactic structures”. In C. R. Cooper and L. Odell (Eds.), Evaluating Writing: Describing, Measuring and Judging (pp. 91–104). Urbana, Il: NCTE.
TSL[v.20020404] Prn:7/10/2004; 9:29
F: PB129RE.tex / p.11 (233)
References
Hymes, Dell (1972). “Models of interaction of language and social life”. In J. J. Gumperz and D. Hymes (Eds.), Directions in Sociolinguistics: The Ethnography of Communication (pp. 35–71). New York: Holt, Rinehart and Winston. Hymes, Dell (1974). Foundations in Sociolinguistics: An Ethnographic Approach. Philadelphia: University of Pennsylvania Press. Hyon, Sunny (1995). A Genre-Based Approach to ESL Reading: Implications for North America and Australia. Ph.D. dissertation. University of Michigan. Hyon, Sunny (1996). “Genres in three traditions: Implications for ESL”. TESOL Quarterly, 30 (4), 693–722. Hyon, Sunny (2001). “Long-term effects of genre-based instruction: A follow-up study of an EAP reading course”. English for Specific Purposes, 20 (Supplement 1), 417–438. Jackendoff, Ray (1972). Semantic Interpretation in Generative Grammar. Cambridge: MIT Press. Jaffe, Joseph and Feldstein, Stanley (1970). Rhythms of Dialogue. New York: Academic Press. Jefferson, Gail (1972). “Side sequences”. In D. Sudnow (Ed.), Studies in Social Interaction (pp. 294–338). New York: Free Press. Jekat, Susanne, Klein, Alexandra, Maier, Elisabeth, Maleck, Ilona, Mast, Marion, and Quantz, J. Joachim (1995). “Dialogue Acts in VerbMobil.”. VerbMobil Technical Report 65. DFKI. Jiménez Juliá, Tomás (1986). Aproximación al estudio de las funciones informativas. Málaga: Librería Ágora. Johns, Ann (Ed.) (2002). Genre in the Classroom: Multiple Perspectives. Mahwah, NJ: Lawrence Erlbaum. Johnson-Laird, Philip N. (1980). “Mental models in cognitive science”. Cognitive Science, 4, 71–115. Johnstone, Barbara (1994). “Repetition in discourse: A dialogue”. In B. Johnstone (Ed.), Repetition in Discourse: Interdisciplinary Perspectives (pp. 1–20). Norwood, NJ: Ablex. Jones, Linda Kay (1977). Theme in English Expository Discourse. Lake Bluff, Illinois: Jupiter Press. Jucker, Andreas H. and Ziv, Yael (Eds.). (1998). Discourse Markers: Descriptions and Theory. Amsterdam and Philadelphia: John Benjamins. Kamp, Hans and Reyle, Uwe (1993). From Discourse to Logic: Introduction to Model Theoretic Semantics of Natural Language, Formal Logic and Discourse Representation Theory. The Netherlands: Dordrecht. Kehler, Andrew (2002). Coherence, Reference, and the Theory of Grammar. Stanford, CA: CSLI. Kessler, Brett, Nunberg, Geoffrey, and Schütze, Hinrich (1997). “Automatic detection of text genre”. In Proceedings of the 35th Annual Meeting of the Association for Computational Linguistics (pp. 32–38). Madrid, Spain. Kintsch, Walter (1989). “The representation of knowledge and the use of knowledge in discourse comprehension”. In R. Dietrich and C. Graumann (Eds.), Language Processing in Social Context (pp. 185–209). Amsterdam: Elsevier. Knott, Alistair and Dale, Robert (1994). “Using linguistic phenomena to motivate a set of coherence relations”. Discourse Processes, 18 (1), 35–62.
TSL[v.20020404] Prn:7/10/2004; 9:29
F: PB129RE.tex / p.12 (234)
References
Knott, Alistair and Sanders, Ted (1998). “The classification of coherence relations and their linguistic markers: An exploration of two languages”. Journal of Pragmatics, 30, 135– 175. Koike, Dale A. and Biron, Christina M. (1996). “Genre as a basis for the avanced Spanish conversation class”. Hispania, 79 (2), 290–296. Kong, Kenneth C. C. (1998). “Are simple business request letters really simple? A comparison of Chinese and English business request letters”. Text, 18 (1), 103–141. Kosseim, Leila and Lapalme, Guy (2000). “Choosing rhetorical structures to plan instructional texts”. Computational Intelligence, 16 (3), 408–445. Kress, Gunther (1982). Learning to Write. London: Routledge and Kegan Paul. Kress, Gunther and Threadgold, Terry (1988). “Towards a social theory of genre”. Southern Review, 21 (3), 215–243. Kuno, Susumu (1972). “Functional Sentence Perspective: A case study from Japanese and English”. Linguistic Inquiry, 3, 269–320. Labov, William (1972). Language in the Inner City: Studies in the Black English Vernacular. Philadelphia: University of Pennsylvania Press. Lakoff, George (1987). Women, Fire, and Dangerous Things: What Categories Reveal about the Mind. Chicago: University of Chicago Press. Lascarides, Alex and Asher, Nicholas (1993). “Temporal interpretation, discourse relations and commonsense entailment”. Linguistics and Philosophy, 16 (5), 437–493. Lavid, Julia (1999). “Contextual constraints on thematization in written discourse: An empirical study”. In D. Gabbay and M. Cavalcanti (Eds.), Formal Aspects of Context (pp. 119–130). Dordrecht: Kluwer Academic Publishers. Lavid, Julia (2000a). “Text types, chaining strategies and Theme in a multilingual corpus: A cross-linguistic comparison for text generation”. In J. Bregazzi, A. Downing, D. López, and J. Neff (Eds.), Homenaje a Jack C. White: Estudios de filología inglesa, (pp. 107–122). Madrid: Universidad Complutense. Lavid, Julia (2000b). “Theme, focus, given and other dangerous things: Linguistic and computational approaches to information in discourse”. Revista canaria de estudios ingleses, 40, 355–369. Leckie-Tarry, Helen (1995). Language and Context: A Functional Linguistic Theory of Register. London: Pinter. Leong Ping, Alvin (2000). “The inference-boundary model: Reinterpreting Theme and Rheme”. Language Sciences, 22 (1), 1–26. Longacre, Robert E. (1976). An Anatomy of Speech Notions. Lisse: The Peter de Ridder Press. Longacre, Robert E. (1983). The Grammar of Discourse. New York: Plenum. Luzón Marco, María José (2002). “A genre analysis of corporate home pages”. LSP and Professional Communication, 2 (1), 41–56. Macken-Horarik, Mary (2002). “"Something to shoot for": A systemic functional approach to teaching genre in secondary school science”. In A. Johns (Ed.), Genre in the Classroom: Multiple Perspectives (pp. 17–42). Mahwah, NJ: Lawrence Erlbaum. Maclay, Howard and Oswood, Charles E. (1959). “Hesitation phenomena in spontaneous English speech”. Word, 15, 19–44. MacWhinney, Brian (1977). “Starting points”. Language, 53 (1), 152–168.
TSL[v.20020404] Prn:7/10/2004; 9:29
F: PB129RE.tex / p.13 (235)
References
Maier, Elisabeth (1996a). “Context construction as subtask of dialogue processing: The VerbMobil case”. In Proceedings of 11th Twente Workshop on Language Technology, TWLT 11. Maier, Elisabeth (1996b). “Textual relations as part of multiple links between text segments”. In G. Adorni and M. Zock (Eds.), Trends in Natural Language Generation: An Artificial Intelligence Perspective (pp. 68–87). Berlin: Springer Verlag. Malinowski, Bronislaw (1923 [1946]). “The problem of meaning in primitive languages”. In C. K. Ogden and I. A. Richards (Eds.), The Meaning of Meaning (pp. 296–336). London: Kegan Paul. Mann, William C. (1983a). “An Overview of the Nigel Text Generation Grammar”. Technical report ISI/RR-83-113. Information Sciences Institute, University of Southern California. Mann, William C. (1983b). “An Overview of the Penman Text Generation Grammar”. Technical report ISI/RR-83-114. Information Sciences Institute, University of Southern California. Mann, William C. and Matthiessen, Christian M. I. M. (1991). “Functions of language in two frameworks”. Word, 42 (3), 231–249. Mann, William C., Matthiessen, Christian M.I.M., and Thompson, Sandra A. (1992). “Rhetorical Structure Theory and text analysis”. In W. C. Mann and S. A. Thompson (Eds.), Discourse Description: Diverse Linguistic Analyses of a Fund-Raising Text (pp. 39– 78). Amsterdam and Philadelphia: John Benjamins. Mann, William C. and Thompson, Sandra A. (1987). “Rhetorical Structure Theory: A Theory of Text Organization”. Technical report ISI/RS-87-190. Information Sciences Institute, University of Southern California. Mann, William C. and Thompson, Sandra A. (1988). “Rhetorical Structure Theory: Toward a functional theory of text organization”. Text, 8 (3), 243–281. Marcu, Daniel (2000). “The rhetorical parsing of unrestricted texts: A surface based approach”. Computational Linguistics, 26 (3), 395–448. Markels, Robin Bell (1981). Cohesion Patterns in English Expository Paragraphs. Ph.D. dissertation. Ohio State University. Márquez Reiter, Rosina (2000). Linguistic Politeness in Britain and Uruguay: A Contrastive Study of Requests and Apologies. Amsterdam and Philadelphia: John Benjamins. Márquez Reiter, Rosina (2002). “A contrastive study of conventional indirectness in Spanish: Evidence from Peninsular and Uruguayan Spanish”. Pragmatics, 12 (2), 135–151. Márquez Reiter, Rosina (2003). “Pragmatic variation in Spanish: External request modifications in Peninsular and Uruguayan Spanish”. In R. Núñez-Cedeño, L. López, and R. Cameron (Eds.), A Romance Perspective on Language Knowledge and Use (pp. 167– 180). Amsterdam and Philadelphia: John Benjamins. Martin, James R. (1984). “Language, register and genre”. In F. Christie (Ed.), Children Writing: Reader (pp. 21–30). Geelong, Victoria: Deakin University Press. (Revised and published as Martin 2001b). Martin, James R. (1985a). Factual Writing: Exploring and Challenging Social Reality. Victoria: Deakin University Press.
TSL[v.20020404] Prn:7/10/2004; 9:29
F: PB129RE.tex / p.14 (236)
References
Martin, James R. (1985b). “Process and text: Two aspects of human semiosis”. In J. Benson and W. Greaves (Eds.), Systemic Perspectives on Discourse (pp. 248–274). Norwood, NJ: Ablex. Martin, James R. (1992). English Text: System and Structure. Amsterdam and Philadelphia: John Benjamins. Martin, James R. (1993). “Genre and literacy: Modeling context in educational linguistics”. Annual Review of Applied Linguistics, 13, 141–172. Martin, James R. (1995). “More than what the message is about: English Theme”. In M. Ghadessy (Ed.), Thematic Development in English Texts (pp. 223–258). London: Pinter. Martin, James R. (1996). “Evaluating disruption: symbolising Theme in junior secondary narrative”. In R. Hasan and G. Williams (Eds.), Literacy in Society (pp. 124–171). London: Longman. Martin, James R. (2001a). “Cohesion and texture”. In D. Schiffrin, D. Tannen, and H. E. Hamilton (Eds.), The Handbook of Discourse Analysis (pp. 35–53). Malden, Mass: Blackwell. Martin, James R. (2001b). “Language, register and genre”. In A. Burns and C. Coffin (Eds.), Analysing English in a Global Context (pp. 149–166). London/New York: Routledge. (Revised version of Martin 1984). Martin, James R., Christie, Frances, and Rothery, Joan (1987). “Social processes in education: A reply to Sawyer and Watson (and others)”. In I. Reid (Ed.), The Place of Genre in Learning: Current Debates (pp. 58–82). Victoria: Deakin University Press. Martin, James R., Matthiessen, Christian M. I. M., and Painter, Claire (1997). Working with Funtional Grammar. London: Arnold. Martin, James R. and Rose, David (2003). Working with Discourse. London: Continuum. Martín Martín, Pedro (2002). “A genre analysis of English and Spanish research paper abstracts in experimental social sciences”. English for Specific Purposes, 22 (1), 25–43. Martín Zorraquino, M. Antonia (1988). “Los marcadores del discurso desde un punto de vista gramatical”. In M. A. M. Zorraquino and E. M. Durán (Eds.), Los marcadores del discurso: Teoría y análisis (pp. 19–53). Madrid: Arco. Mathesius, Vilém (1939). “O tak zvanem aktualnim cleneni vetnem (On the so-called Functional Sentence Perspective)”. Slovo a Slovestnost, 5, 171–174. Matthiessen, Christian M. I. M. (1992). “Interpreting the textual metafunction”. In M. Davies and L. Ravelli (Eds.), Advances in Systemic Linguistics: Recent Theory and Practice (pp. 37–82). London: Pinter. Matthiessen, Christian M. I. M. (1995). Lexicogrammatical Cartography: English Systems. Tokyo: International Language Sciences Publishers. Matthiessen, Christian M. I. M. and Bateman, John (1991). Text Generation and SystemicFunctional Linguistics: Experiences from English and Japanese. London and New York: Pinter. Matthiessen, Christian M. I. M. and Thompson, Sandra A. (1988). “The structure of discourse and “subordination””. In J. Haiman and S. A. Thompson (Eds.), Clause Combining in Discourse and Grammar (pp. 275–329). Amsterdam and Philadelphia: John Benjamins.
TSL[v.20020404] Prn:7/10/2004; 9:29
F: PB129RE.tex / p.15 (237)
References
Maynard, Senko K. (1986). “Interactional aspects of thematic progression in English casual conversation”. Text, 6 (1), 73–105. McCabe-Hidalgo, Anne (1999). Theme and Thematic Patterns in Spanish and English History Texts. Ph.D. dissertation. Aston University. McKeown, Kathleen R. (1985). Text Generation: Using Discourse Strategies and Focus Constraints to Generate Natural Language Text. Cambridge: Cambridge University Press. Mederos Martín, Humberto (1988). Procedimientos de cohesión en el español actual. Tenerife: Aula de Cultura de Tenerife. Meinhof, Ulrike H. and Smith, Jonathan (2000). Intertextuality and the Media: From Genre to Everyday Life. Manchester: Manchester University Press. Miller, Carolyn (1984). “Genre as social action”. Quarterly Journal of Speech, 70 (2), 151–167. Miller, George A. (1956). “The magical number seven, plus or minus two.”. Psychological Review, 63, 81–97. Minsky, Marvin (1975). “A framework for representing knowledge”. In P. H. Winston (Ed.), The Psychology of Computer Vision (pp. 211–277). New York: McGraw-Hill. Minsky, Marvin (1977). “Frame-system theory”. In P. N. Johnson-Laird and P. C. Wason (Eds.), Thinking: Readings in Cognitive Science (pp. 355–376). Cambridge: Cambridge University Press. Mitchell, T. F. (1957). “The language of buying and selling in Cyrenaica: A situational statement”. Hesperis, 26, 31–71. Moessner, Lilo (2001). “Genre, text type, style, register: A terminological maze?”. European Journal of English Studies, 5 (2), 131–138. Mohan, Bernard (1986). Language and Content. Reading, Mass: Addison-Wesley. Moore, Johanna D. and Paris, Cécile (1993). “Planning text for advisory dialogues: Capturing intentional and rhetorical information”. Computational Linguistics, 19 (4), 651–694. Moore, Johanna D. and Pollack, Martha E. (1992). “A problem for RST: The need for multilevel discourse analysis”. Computational Linguistics, 18 (4), 537–544. Morris, Terry (1998). “Topicity vs. thematicity: Topic prominence in impromptu Spanish discourse”. Journal of Pragmatics, 29, 193–203. Moya Guijarro, A. Jesús (2000). “Towards a definition and hierarchization of topic”. In A. Downing, A. J. Moya Guijarro, and J. I. Albentosa Hernández (Eds.), Talk and Text: Studies on Spoken and Written Discourse (pp. 97–116). Cuenca: Universidad de Castilla-La Mancha. Moya Guijarro, A. Jesús and Albentosa Hernández, José Ignacio (2001). “Points of departure in news items and tourist brochures: Choices of Theme and topic”. Text, 21 (3), 347– 371. Murray, Denise E. (1991). Conversation for Action: The Computer Terminal as Medium of Communication. Amsterdam and Philadelphia: John Benjamins. Nariyama, Shigeko (2004). “Subject ellipsis in English”. Journal of Pragmatics, 36 (2), 237– 264. Nwogu, Kevin and Bloor, Thomas (1991). “Thematic progression in professional and popular medical texts”. In E. Ventola (Ed.), Functional and Systemic Linguistics: Approaches and Uses (pp. 369–384). The Hague: Mouton de Gruyter.
TSL[v.20020404] Prn:7/10/2004; 9:29
F: PB129RE.tex / p.16 (238)
References
Ocampo, Francisco (1990). “The pragmatics of word order in constructions with a verb and a subject”. Hispanic Linguistics, 4 (1), 87–128. Ocampo, Francisco (1995). “The word order to two-constituent constructions in spoken Spanish”. In P. Downing and M. Noonan (Eds.), Word Order in Discourse (pp. 425–447). Amsterdam and Philadelphia: John Benjamins. Ochs, Elinor (1979). “Planned and unplanned discourse”. In T. Givón (Ed.), Discourse and Syntax. Syntax and Semantics Volume 12 (pp. 51–80). New York: Academic Press. O’Donnell, Michael, Mellish, Chris, Oberlander, Jon, and Knott, Alistair (2001). “ILEX: An architecture for a dynamic Hypertext generation system”. Natural Language Engineering, 7, 225–250. Östman, Jan-Ola and Virtanen, Tuija (1999). “Theme, comment and newness as figures in information structuring”. In K. v. Hoek, A. A. Kibrik, and L. Noordman (Eds.), Discourse Studies in Cognitive Linguistics (pp. 91–110). Amsterdam and Philadelphia: John Benjamins. Oversteegen, Leonoor E. (1997). “On the pragmatic nature of causal and contrastive connectives”. Discourse Processes, 24, 51–85. Painter, Claire and Martin, James R. (Eds.). (1986). Writing to Mean: Teaching Genres across the Curriculum. Melbourne: Applied Linguistics Association of Australia. Paltridge, Brian (1993). “Writing up research: A systemic functional perspective”. System, 21 (2), 175–192. Paltridge, Brian (1995a). “Analyzing genre: A relational perspective”. System, 23 (4), 503– 511. Paltridge, Brian (1995b). “Genre and the notion of prototype”. Prospect, 10 (3), 28–34. Paltridge, Brian (1995c). “Working with genre: A pragmatic perspective”. Journal of Pragmatics, 24 (4), 393–406. Paltridge, Brian (1996). “Genre, text type, and the language learning classroom”. English Language Teaching Journal, 50 (3), 237–243. Paltridge, Brian (1997). Genre, Frames and Writing in Research Settings. Amsterdam and Philadelphia: John Benjamins. Paltridge, Brian (2000). “Genre knowledge and the language learning classroom”. English Australia Journal, 18 (2), 52–59. Paltridge, Brian (2001). Genre and the Language Learning Classroom. Ann Arbor: The University of Michigan Press. Parsons, Gerald (1996). “The development of the concept of cohesive harmony”. In M. Berry, C. Butler, R. Fawcett, and G. Huang (Eds.), Meaning and Form: Systemic Functional Interpretations (Meaning and Choice in Language: Studies for Michael Halliday) (pp. 585–599). Norwood, NJ: Ablex. Paulson, Eric J. and Goodman, Kenneth S. (1999). “Influential studies in eye-movement research”. Reading Online (January 1999), http://www.readingonline.org/research/eyemove.html (Accessed August, 2003). Perfetti, Charles A. and Goldman, Susan R. (1974). “Thematization and sentence retrieval”. Journal of Verbal Learning and Verbal Behaviour, 13, 70–79. Péry-Woodley, Marie-Paule (2001). “Modes d’organisation et de signalisation dans des textes procéduraux”. Langages, 141, 28–46.
TSL[v.20020404] Prn:7/10/2004; 9:29
F: PB129RE.tex / p.17 (239)
References
Pike, Kenneth L. (1964). “Discourse analysis and tagmemic matrices”. In R. Brend (Ed.), Advances in Tagmemics (pp. 285–306). Amsterdam: North Holland. Pinto dos Santos, V. B. M. (2002). “Genre analysis of business letters of negotiation”. English for Specific Purposes, 21 (2), 167–199. Polanyi, Livia and Scha, R. J. H. (1983). “The syntax of discourse”. Text, 3 (3), 261–270. Portolés, José (1988). “La teoría de la argumentación en la lengua y los marcadores del discurso”. In M. A. M. Zorraquino and E. M. Durán (Eds.), Los marcadores del discurso: Teoría y análisis (pp. 71–91). Madrid: Arco. Portolés, José (1998). Marcadores del discurso. Madrid: Ariel. Posner, Michael I. and Keele, Steven W. (1968). “On the genesis of abstract ideas”. Journal of Experimental Psychology, 77, 353–363. Potter, John M. (1980). “What was the matter with Dr. Spooner?”. In V. A. Fromkin (Ed.), Errors in Linguistic Performance: Slips of the Tongue, Ear, Pen and Hand (pp. 13–34). New York: Academic Press. Prince, Ellen F. (1981). “Towards a taxonomy of given-new information”. In P. Cole (Ed.), Radical Pragmatics (pp. 223–255). New York: Academic Press. Propp, Vladimir (1977 [1928]). Morfología del cuento. Madrid: Editorial Fundamentos. Ramsay, Guy (2000). “Linearity in rhetorical organisation: A comparative cross-cultural analysis of newstext from the People’s Republic of China and Australia”. International Journal of Applied Linguistics, 10 (2), 241–258. Ravelli, Louise J. (1995). “A dynamic perspective: Implications for metafunctional interaction and an understanding of Theme”. In R. Hasan and P. H. Fries (Eds.), On Subject and Theme: A Discourse Functional Perspective (pp. 187–234). Amsterdam and Philadelphia: John Benjamins. Redeker, Gisela (2000). “Coherence and structure in text and discourse”. In W. Black and H. Bunt (Eds.), Abduction, Belief, and Context in Dialogue. Studies in Computational Pragmatics (pp. 233–263). Amsterdam and Philadelphia: John Benjamins. Reid, Ian (Ed.) (1987). The Place of Genre in Learning: Current Debates. Geelong, Victoria: Deakin University Press. Reitter, David (2003). Rhetorical Analysis with Rich Feature Support Vector Models. Master’s thesis. University of Potsdam. Rich, Charles and Sidner, Candace L. (1997). “COLLAGEN: When agents collaborate with people”. In Proceedings of the First International Conference on Autonomous Agents. Marina del Rey, CA. Reprinted in M. Huhns and M. Singh (Eds.) Readings in Agents. San Francisco, CA: Morgan Kauffmann. 117–124. Riesbeck, Christopher K. and Schank, Roger C. (1978). “Comprehension by computer: Expectation-based analysis of sentences in context”. In W. J. M. Levelt and G. B. Flores d’Arcais (Eds.), Studies in the Perception of Language (pp. 247–293). New York: Wiley. Rivero, María Luisa (1980). “Topicalization and Wh movement in Spanish”. Linguistic Inquiry, 9, 513–517. Rocchi, Cesare and Zancanaro, Massimo (2003). “Generation of video documentaries from discourse structures”. In Proceedings of 9th European Workshop on Natural Language Generarion (EWNLG 9). Budapest, Hungary.
TSL[v.20020404] Prn:7/10/2004; 9:29
F: PB129RE.tex / p.18 (240)
References
Roger, Derek and Bull, Peter (Eds.). (1989). Conversation: An Interdisciplinary Perspective. Clevedon: Multilingual Matters. Romero Trillo, Jesús (1994). “Ahm, ehm, you call it theme?... A thematic approach to spoken English”. Journal of Pragmatics, 22, 495–509. Rosch, Eleanor (1977). “Classification of real-world objects: Origins and representations in cognition”. In P. N. Johnson-Laird and P. C. Wason (Eds.), Thinking: Readings in Cognitive Science (pp. 212–222). Cambridge: Cambridge University Press. Rosch, Eleanor and Mervis, C. B. (1975). “Family resemblances: Studies in the internal structure of categories”. Cognitive Psychology, 7, 573–605. Rose, David (2001). “Some variations in Theme across languages”. Functions of Language, 8 (1), 109–145. Rösner, Dietmar and Stede, Manfred (1992). “Customizing RST for the automatic production of technical manuals”. In R. Dale, E. Hovy, D. Rösner, and O. Stock (Eds.), Aspects of Automated Language Generation (pp. 119–214). Berlin: Springer Verlag. Rothery, Joan (1985). Teaching Writing in the Primary School: A Genre Based Approach to the Development of Writing Abilities. Sydney: Department of Linguistics, University of Sydney. Rumelhart, David E. and Ortony, Andrew (1977). “The representation of knowledge in memory”. In R. C. Anderson, R. J. Spiro, and W. E. Montague (Eds.), Schooling and the Acquisition of Knowledge (pp. 99–135). Hillsdale, NJ: Lawrence Erlbaum. Sacks, Harvey, Schegloff, Emmanuel, and Jefferson, Gail (1974). “A simplest systematics for the organization of turn-taking in conversation”. Language, 50, 696–735. Samraj, Betty (1989). “Exploring current issues in Genre Theory”. Word, 40 (1–2), 189–200. Samraj, Betty (2002). “Introductions in research articles: Variations across disciplines”. English for Specific Purposes, 21 (1), 1–17. Sanders, Ted (1997). “Semantic and pragmatic sources of coherence: On the categorization of coherence relations in context”. Discourse Processes, 24, 119–147. Sanders, Ted and Noordman, Leo (2000). “The role of coherence relations and their linguistic markers in text processing”. Discourse Processes, 29 (1), 37–60. Sanders, Ted, Spooren, Wilbert, and Noordman, Leo (1992). “Toward a taxonomy of coherence relations”. Discourse Processes, 15, 1–35. Sanders, Ted, Spooren, Wilbert, and Noordman, Leo (1993). “Coherence relations in a cognitive theory of discourse representation”. Cognitive Linguistics, 4 (2), 93–133. Sanford, Anthony J. and Garrod, Simon C. (1981). Understanding Written Language: Explorations of Comprehension Beyond the Sentence. Chichester: John Wiley. Schank, Roger C. (1972). “Conceptual dependency: A theory of natural language understanding”. Cognitive Psychology, 3, 552–631. Schank, Roger C. (1973). “Identification of conceptualizations underlying natural language”. In R. C. Schank and K. M. Colby (Eds.), Computer Models of Thought and Language (pp. 187–247). San Francisco: Freeman. Schank, Roger C. and Abelson, R. (1977). Scripts, Plans, Goals and Understanding. Hillsdale, NJ: Lawrence Erlbaum. Schegloff, Emmanuel (1968). “Sequencing in conversational openings”. American Anthropologist, 70, 1075–1095.
TSL[v.20020404] Prn:7/10/2004; 9:29
F: PB129RE.tex / p.19 (241)
References
Schegloff, Emmanuel (1972). “Notes on a conversational practice: Formulating place”. In D. Sudnow (Ed.), Studies in Social Interaction (pp. 75–119). New York: Free Press. Schegloff, Emmanuel (1982). “Discourse as an interactional achievement: some uses of uh huh and other things that come between sentences”. In D. Tannen (Ed.), Analyzing Discourse: Text and Talk. Georgetown University Roundtable on Languages and Linguistics (pp. 71–93). Washington, DC: Georgetown University Press. Schegloff, Emmanuel and Sacks, Harvey (1973). “Opening up closings”. Semiotica, 8, 289– 327. Schiffrin, Deborah (1987). Discourse Markers. Cambridge: Cambridge University Press. Schiffrin, Deborah (1994). Approaches to Discourse. Malden, MA: Blackwell. Schleppegrell, Mary J. (1996). “Strategies for discourse cohesion: Because in ESL writing”. Functions of Language, 3 (2), 235–254. Scholtens, Anneke (1991). “Planning in ordinary conversation”. Journal of Pragmatics, 16 (1), 31–58. Scott, Donia and de Souza, Clarisse Sieckenius (1990). “Getting the message across in RSTbased text generation”. In R. Dale, C. Mellish, and M. Zock (Eds.), Current Research in Natural Language Generation (pp. 47–73). London: Academic Press. Searle, John (1969). Speech Acts. Cambridge: Cambridge University Press. Searle, John (1979 [1995]). “A taxonomy of illocutionary acts”. In A. Martinich (Ed.), The Philosophy of Language (pp. 141–155). New York: Oxford University Press. Shepherd, Michael and Watters, Carolyn (1998). “The evolution of cybergenres”. In Proceedings of the 31st Annual Hawaii International Conference on System Sciences (HICSS ’98) (pp. 97–109). Hawaii: IEEE. Shultz, Jeffrey J., Florio, Susan, and Erikson, Frederick (1982). “Where’s the floor?”. In P. Gilmor and A. A. Glatthorn (Eds.), Children in and out of School (pp. 88–123). Washington, DC: Center for Applied Linguistics. Silva-Corvalán, Carmen (1983). “On the interaction of word order and intonation: Some OV constructions in Spanish”. In F. Klein-Andreu (Ed.), Discourse Perspectives on Syntax (pp. 117–140). New York and London: Academic Press. Sinclair, John and Coulthard, Malcolm (1975). Towards an Analysis of Dicourse: The English Used by Teachers and Pupils. Oxford: Oxford University Press. Smith, Raoul N. and Frawley, William J. (1983). “Conjunctive cohesion in four English genres”. Text, 3 (4), 347–374. Solé, Yolanda P. (1966). ’Hacer’: Verbo funcional y lexical. Washington: Georgetown University Press. Solso, Robert L. (1988). Cognitive Psychology (2nd edition). Boston: Allyn and Bacon. Sperber, Dan and Wilson, Deirdre (1995). Relevance: Communication and Cognition (2nd edition). Oxford: Blackwell. Spooren, Wilbert (1997). “The processing of underspecified coherence relations”. Discourse Processes, 24, 149–168. Stede, Manfred, Wanner, Leo, and Hovy, Eduard (Eds.). (1998). Proceedings of the Workshop on Discourse Relations and Discourse Markers. Montréal, Canada: Association for Computational Linguistics. Steiner, Erich H. and Ramm, Wiebke (1995). “On Theme as a grammatical notion for German”. Functions of Language, 2 (1), 57–93.
TSL[v.20020404] Prn:7/10/2004; 9:29
F: PB129RE.tex / p.20 (242)
References
Stenström, Anna-Brita (1994). An Introduction to Spoken Interaction. London: Longman. Stent, Amanda (2000). “Rhetorical structure in dialog”. In Proceedings of First International Conference on Natural Language Generation (INLG’2000) (pp. 247–252). Mitzpe Ramon, Israel. Stoddard, Sally (1991). Text and Texture: Patterns of Cohesion. Norwood, NJ: Ablex. Suñer, Margarita (1982). Syntax and Semantics of Spanish Presentational Sentence-Types. Washington: Georgetown University Press. Suñer, Margarita (1988). “The role of agreement in clitic-doubled constructions”. Natural Language and Linguistic Theory, 6, 391–434. Swales, John M. (1990). Genre Analysis: English in Academic and Research Settings. Cambridge: Cambridge University Press. Swales, John M., Jakobsen, Heriette, Kejser, Christina, Koch, Lena, Lynch, Joan, and Mølbæk, Lone (2000). “A new link in a chain of genres?”. Hermes, 25, 133–141. Taboada, Maite (1995). Theme Markedness in English and Spanish: A Systemic-Functional Approach. Unpublished manuscript. Universidad Complutense de Madrid. Taboada, Maite (1997). “Improving translation through contextual information”. In Proceedings of the 35th Annual Meeting of the Association for Computational Linguistics, ACL/EACL ’97 (student session) (pp. 510–512). Madrid, Spain. Taboada, Maite (2000). “Cohesion as a measure in generic analysis”. In A. Melby and A. Lommel (Eds.), The 26th LACUS Forum (pp. 35–49). Chapel Hill, NC: The Linguistic Association of Canada and the United States. Taboada, Maite (2003). “Modeling task-oriented dialogue”. Computers and the Humanities, 37 (4), 431–454. Taboada, Maite (2004). “Rhetorical relations in dialogue: A contrastive study”. In C. L. Moder and A. Martinovic-Zic (Eds.), Discourse across Languages and Cultures (pp. 75– 97). Amsterdam and Philadelphia: John Benjamins. Taboada, Maite and Lavid, Julia (2003). “Rhetorical and thematic patterns in scheduling dialogues: A generic characterization”. Functions of Language, 10 (2), 147–179. Taglicht, Josef (1984). Message and Emphasis: On Focus and Scope in English. London: Longman. Tannen, Deborah (1980). “A comparative analysis of oral narrative strategies: Athenian Greek and American English”. In W. Chafe (Ed.), The Pear Stories: Cognitive, Cultural and Linguistic Aspects of Narrative Production (pp. 51–87). Norwood, NJ: Ablex. ten Have, Paul (2002). “Comparing telephone call openings: Theoretical and methodological reflections”. In K. K. Luke and T.-S. Pavlidou (Eds.), Telephone Calls: Unity and Diversity in Conversational Structure across Languages and Cultures (pp. 233–248). Amsterdam and Philadelphia: John Benjamins. Thompson, Geoff (1996). Introducing Functional Grammar. London: Arnold. Threadgold, Terry (1988). “The genre debate”. Southern Review, 21 (3), 315–330. Threadgold, Terry (1989). “Talking about genre: Ideologies and incompatible discourses”. Cultural Studies, 3, 92–118. Todorov, Tzvetan (1990 [1978]). Genres in Discourse. Cambridge: Cambridge University Press. Tomlin, Russell S. (1986). Basic Word Order: Functional Principles. Beckenham: Croom Helm.
TSL[v.20020404] Prn:7/10/2004; 9:29
F: PB129RE.tex / p.21 (243)
References
Tomlin, Russell S. (1995). “Focal attention, voice, and word order: An experimental, crosslinguistic study”. In P. Downing and M. Noonan (Eds.), Word Order in Discourse (pp. 517–554). Amsterdam and Philadelphia: John Benjamins. Torrego, Esther (1984). “On inversion in Spanish and some of its effects”. Linguistic Inquiry, 15, 103–129. Traum, David (1993). “Rhetorical relations, action and intentionality in conversation”. In O. Rambow (Ed.), Proceedings of ACL SIG Workshop on Intentionality and Structure in Discourse Relations (pp. 132–135). Columbus, Ohio. Traum, David and Hinkelman, Elizabeth (1992). “Conversation acts in task-oriented spoken dialogue”. Computational Intelligence, 8 (3), 575–599. Trosborg, Anna (Ed.). (2000). Analysing Professional Genres. Amsterdam and Philadelphia: John Benjamins. Tsui, Amy B. M. (1989). “Beyond the “adjacency pair””. Language in Society, 18 (4), 545–564. Tsui, Amy B. M. (1991). “Sequencing rules and coherence in discourse”. Journal of Pragmatics, 15, 111–129. Tsui, Amy B. M. (1994). English Conversation. Oxford: Oxford University Press. Upton, Thomas A. and Connor, Ulla (2001). “Using computerized corpus analysis to investigate the textlinguistic discourse moves of a genre”. English for Specific Purposes, 20 (4), 313–329. Ure, Jean and Ellis, Jeffrey (1977). “Register in descriptive linguistics and linguistic sociology”. In O. Uribe-Villas (Ed.), Issues in Sociolinguistics (pp. 197–243). The Hague: Mouton. Vallduví, Enric (1990). The Informational Component. Ph.D. dissertation. University of Pennsylvania. Vallduví, Enric and Engdahl, Elisabet (1996). “The linguistic realization of information packaging”. Linguistics, 34, 459–519. van Dijk, Teun A. (1972). Some Aspects of Text Grammars. The Hague: Mouton. van Dijk, Teun A. (1980). Macrostructures: An Interdisciplinary Study of Global Structures in Discourse Interaction and Cognition. Hillsdale, NJ: Lawrence Erlbaum. van Dijk, Teun A. (1998). “Principles of critical discourse analysis”. In J. Cheshire and P. Trudgill (Eds.), The Sociolinguistics Reader. Vol. 2: Gender and Discourse (pp. 367– 393). London: Arnold. van Dijk, Teun A. (2001). “Critical discourse analysis”. In D. Schiffrin, D. Tannen, and H. E. Hamilton (Eds.), The Handbook of Discourse Analysis (pp. 352–371). Malden, Mass: Blackwell. Vander Linden, Keith, Cumming, Susanna, and Martin, James H. (1992). “Using system networks to build rhetorical structures”. In R. Dale, E. Hovy, D. Rösner, and O. Stock (Eds.), Aspects of Automated Natural Language Generation (pp. 183–198). Berlin: Springer Verlag. Vasconcellos, Muriel (1992). “The Theme as message onset: Its structure and characteristics”. Linguistics, 30 (1), 147–163. Ventola, Eija (1987). The Structure of Social Interaction: A Systemic Approach to the Semiotics of Service Encounters. London: Frances Pinter. Ventola, Eija (1989). “Problems of modelling and applied issues within the framework of genre”. Word, 40 (1–2), 129–161.
TSL[v.20020404] Prn:7/10/2004; 9:29
F: PB129RE.tex / p.22 (244)
References
Vigara Tauste, Ana M. (1995). “Comodidad y recurrencia en la organización del discurso coloquial”. In L. Cortés Rodríguez (Ed.), El Español Coloquial: Actas del I Simposio sobre análisis del discurso oral (pp. 175–208). Universidad de Almería. Waibel, Alexander H. (1996). “Interactive translation of conversational speech”. Computer, 29 (7), 41–48. Walker, Marilyn (1993). Informational Redundancy and Resource Bounds in Dialogue. Ph.D. dissertation. University of Pennsylvania. Webber, Bonnie, Joshi, Aravind K., Stone, Matthew, and Knott, Alistair (2003). “Anaphora and discourse structure”. Computational Linguistics, 29 (4), 545–587. Webber, Bonnie, Knott, Alistair, Stone, Matthew, and Joshi, Aravind K. (1999). “Discourse relations: A structural and presuppositional account using lexicalised TAG”. In Proceedings of 37th Annual Meeting of the Association for Computational Linguistics (pp. 41–48). College Park, Maryland. Weigand, Edda (1994). “Discourse, conversation, dialogue”. In E. Weigand (Ed.), Concepts of Dialogue. Considered from the Perspective of Different Disciplines (pp. 49–75). Tübingen: Max Niemeyer. Weiss, Gilbert and Wodak, Ruth (Eds.). (2003). Critical Discourse Analysis: Theory and Disciplinarity. New York: Palgrave. Williams, M. P. (1988). “Functional Sentence Perspective in the context of Systemic Functional Grammar”. In E. H. Steiner and R. Veltman (Eds.), Pragmatics, Discourse and Text: Some Systemically-Inspired Approaches (pp. 76–89). London: Pinter. Wittgenstein, Ludwig (1989 [1953]). Philosophical Investigations. Oxford: Basil Blackwell. Wodak, Ruth and Meyer, Michael (Eds.). (2002). Methods of Critical Discourse Analysis. London: Sage. Yngve, Victor H. (1970). “On getting a word in edgewise”. In Papers from the Sixth Regional Meeting of the Chicago Linguistics Society (pp. 567–577). Chicago: University of Chicago. Zubizarreta, María Luisa (1998). Prosody, Focus and Word Order. Cambridge, Mass: MIT Press. Zwaan, Rolf A. (1994). “Effect of genre expectations on text comprehension”. Journal of Experimental Psychology: Learning, Memory and Cognition, 20 (4), 920–933.
TSL[v.20020404] Prn:23/09/2004; 13:24
F: PB129NI.tex / p.1 (245)
Name index
A Abelen, Eric 110 Abelson, R. 20 Ahlén, Sondra 37, 38 Aijmer, Karin 148, 188, 189 Albentosa Hernández, José Ignacio 61 Allen, James 189 Anderson, Richard C. 21 Anscombre, Jean-Claude 154 Asher, Nicholas 106, 108 Askehave, Inger 36 Austin, J. L. 24, 188 B Bäcklund, Ingegerd 75, 104 Badger, Richard 31 Bakhtin, Mikhail xv, 7–10, 19, 22, 27, 29, 34–36, 54 Ballard, D. Lee 154 Bañón Hernández, Antonio M. 214 Bargiela-Chiappini, Francesca 30, 181 Bartlett, Frederic C. 21 Bateman, John 65, 108, 154 Beekman, John 154 Bello, Andrés 174 Benjamin, Carmen 103 Bentivoglio, Paola 67 Bernárdez, Enrique 67, 154, 174, 191 Berry, Margaret 55, 63, 64, 82, 104, 113 Bhatia, Vijay K. 30, 31 Biber, Douglas 19, 32, 101, 210 Biron, Christina M. 29 Bloor, Thomas 85, 100
Blum-Kulka, Shoshana 222 Bolinger, Dwight L. 67, 69, 103 Bouayad-Agha, Nadjet 154 Brazil, David 188 Bregman, Alvan 31 Brinton, Laurel J. 148 Briz, Antonio 116 Brown, Gillian 20, 21 Brown, Gretchen P. 34 Brown, Penelope 78, 123, 128, 216 Brown, T. Pascal 30 Bublitz, Wolfram 1 Bull, Peter 39 Büring, Daniel 56 Butt, John 103 Byron, Donna K. 145, 148 C Caffarel, Alice 63 Callow, John 154 Cantarero, Margarita 39 Carlson, Matthew 72 Carroll, Susanne 10, 159 Casado Velarde, Manuel 155, 174 Cawsey, Alison 154 Cha, Jin Soon 159 Chafe, Wallace 44 Chomsky, Noam 56, 58 Christie, Frances 7, 25, 29, 30 Clark, Herbert H. 5, 26, 116, 145, 181 Cloran, Carmel 91 Coe, Richard M. 29 Connor, Ulla 30 Conrad, Robert J. 154 Conrad, Susan 19, 32, 210 Contreras, Heles 67
TSL[v.20020404] Prn:23/09/2004; 13:24
F: PB129NI.tex / p.2 (246)
Name index
Cook, Mark 213, 214 Cope, Bill 29 Core, Mark 189 Corston-Oliver, Simon 154 Cortés Rodríguez, Luis 214 Coulthard, Malcolm 16, 157, 188 Crystal, David 44 Cui, Songren 110 Cumming, Susanna 44, 154 Cummings, Michael 79, 83 D Dahl, Östen 56 Dale, Robert 107, 114, 147, 150 Daneš, František 1, 55, 86–90, 102 Daradoumis, Thanasis 113 Davies, Bethan L. 113 de Geest, Dirk 24 de Souza, Clarisse Sieckenius 154 Dorval, Bruce 22–23 Downing, Angela 55–57, 59–63, 65, 69, 80, 103 Dryer, Matthew S. 67 Du Bois, John W. 44 Dubois, Betty Lou 89, 90, 102 Ducrot, Oswald 154 Dudley-Evans, Tony 29, 30 Duncan, Starkey 212 E Edelsky, Carole 118 Edwards, Jane A. 41 Eggins, Suzanne 5, 7, 10, 12, 17, 18, 25, 26, 28, 32, 33, 35, 36, 38, 55, 103, 181, 187 Ellis, Jeffrey 10 Engdahl, Elisabet 55, 56 Erikson, Frederick 118 Erteschik-Shir, Nomi 56 Esgueva, Manuel 39 F Fairclough, Norman 31 Fant, Lars 52
Fawcett, Robin P. 113 Feldstein, Stanley 40 Félix-Brasdefer, J. César 52 Ferguson, Charles A. 184 Fernández Soriano, Olga 71 Finegan, Edward 32, 101 Firbas, Jan 55, 86 Firth, J. R. 10, 16 Firth, Raymond 184 Florio, Susan 118 Fox, Barbara A. 179 Foz, M. Carmen 63 Francis, Gill 85, 86, 100 Frawley, William J. 164 Freedman, Aviva 8, 29 Fretheim, Thorstein 56 Fries, Peter H. 56–59, 62, 63, 66, 82, 84, 100, 101, 104 Fuller, Daniel P. 154 G Gallardo Paúls, Beatriz 5, 188, 189 Garrido Medina, Joaquín 34 Garrod, Simon C. 21 Geis, Michael L. 189 Ghadessy, Mohsen 56, 100, 101 Giannoni, Davide Simone 31 Giora, Rachel 93 Givón, Talmy 67, 68 Goetz, E. T. 21 Goldman, Susan R. 54 Gómez-González, María de los Ángeles xvii, 56, 57, 62, 79 Goodman, Kenneth S. 102 Goodwin, Charles 40 Gregory, Michael 10, 19, 29, 159 Grimes, Joseph E. 106, 148, 154 Grosz, Barbara J. 56, 106, 107, 112 Gundel, Jeanette K. 56, 58 Gutiérrez Ordóñez, Salvador 67, 103 H Hajiˇcová, Eva 55, 56, 58 Halliday, Michael A. K. xv, 2, 5, 7, 10–12, 14, 17, 22, 38, 44, 53,
TSL[v.20020404] Prn:23/09/2004; 13:24
F: PB129NI.tex / p.3 (247)
Name index
55–60, 62, 63, 65, 80, 81, 83, 85, 103, 112, 155–157, 159, 164–166, 168, 174, 177, 179 Harold, Bruce B. 67 Harris, Sandra 30, 181 Harris, Zellig 14 Hasan, Ruqaiya xv, 2, 7, 11–14, 16–19, 22, 24, 33, 36, 38, 56, 63, 66, 155–159, 164–166, 168, 169, 174, 177, 179 Haswell, Richard H. 180 Hatch, Evelyn 46, 52 Hatcher, Ann G. 67, 68 Hayashi, R. 118 Haythornthwaite, Caroline 31 Heeman, Peter A. 145, 148 Henderson, Willie 30 Heritage, John 213 Hernández Cabrera, Clara 39 Hetzron, Robert 68 Hidalgo, Antonio 116 Hinkelman, Elizabeth 189 Hobbs, Jerry 23, 106, 108 Hockett, Charles F. 56 Hoey, Michael 106, 157, 172 Hopcroft, John E. 181 Hopper, Paul 5 House, Juliane 222 Hovy, Eduard 106, 107, 111, 113 Huddleston, Rodney 180 Hunt, Kellogg W. 58 Hymes, Dell 17, 159 Hyon, Sunny 8, 29, 30 J Jackendoff, Ray 56, 58 Jaffe, Joseph 40 Jakobsen, Heriette 31 Jefferson, Gail 43, 44, 142, 157, 188, 192, 195 Jekat, Susanne 189 Jiménez Juliá, Tomás 55, 66 Johansson, Stig 32 Johns, Ann 29, 30 Johnson-Laird, Philip N. 21
Johnstone, Barbara 176 Jones, Linda Kay 57 Joshi, Aravind K. 106, 148 Jucker, Andreas H. xvii, 148 K Kalantzis, Mary 29 Kamp, Hans 189 Kamps, Thomas 154 Kasper, Gabriele 222 Keele, Steven W. 25 Kehler, Andrew 106, 108 Kejser, Christina 31 Kessler, Brett 31 Kintsch, Walter 21 Klein, Alexandra 189 Kleinz, Jörg 154 Knott, Alistair 106–108, 114, 147, 148, 150, 154 Koch, Lena 31 Koike, Dale A. 29 Kong, Kenneth C. C. 110 Korbayová, Ivana 54 Kosseim, Leila 154 Kress, Gunther 12, 22, 29 Kruijff, Geert-Jan 54 Kuno, Susumu 56, 58 L Labov, William 39 Lakoff, George 36 Lampert, Martin D. 41 Lapalme, Guy 154 Lascarides, Alex 106, 108 Lavid, Julia xvi, xvii, 55, 100, Lazaraton, Anne 46, 52 Leckie-Tarry, Helen 17, 19, 33, 100, 157 Leech, Geoffrey 32 Leong Ping, Alvin 55 Levinson, Stephen 78, 123, 128, 216 Lewis, Marilyn 30 Locke, Philip 55, 62, 63, 65, 80 Longacre, Robert E. 154
TSL[v.20020404] Prn:23/09/2004; 13:24
F: PB129NI.tex / p.4 (248)
Name index
Luzón Marco, María José 30 Lynch, Joan 31 M MacIntosh, Angus 10–12, 17 Macken-Horarik, Mary 29 Maclay, Howard 213 MacWhinney, Brian 54 Maier, Elisabeth 106, 107, 111–113, 181, 189 Maleck, Ilona 189 Malinowski, Bronislaw 10, 159 Mann, William C. xv, 1, 2, 65, 106–113, 115, 119, 129, 133, 142 Marcu, Daniel 154 Markels, Robin Bell 155 Márquez Reiter, Rosina 52, 222 Martin, James H. 154 Martin, James R. xv, 2, 7, 10, 12, 17–19, 25–30, 35, 36, 54–56, 63, 65, 105–108, 113, 154 Martín Martín, Pedro 31 Martín Zorraquino, M. Antonia 148 Mast, Marion 189 Mathesius, Vilém 57 Matthiessen, Christian M. I. M. 55, 63, 65, 107, 112, 113, 115, 119, 133 Maynard, Senko K. 93 McCabe-Hidalgo, Anne 103 McKeown, Kathleen R. 107 Mederos Martín, Humberto 155, 165, 173, 174 Medway, Peter 8, 29 Meinhof, Ulrike H. 31 Mellish, Chris 154 Mervis, C. B. 24 Meyer, Michael 31 Miller, Carolyn 8, 29, 31 Miller, George A. 176 Minsky, Marvin 20 Mitchell, T. F. 13, 16 Moessner, Lilo 8 Mohan, Bernard 29 Mølbæk, Lone 31
Montgomery, Martin 188 Moore, Johanna D. 107, 112, 154 Morris, Terry 67 Moya Guijarro, A. Jesús 61 Murray, Denise E. 27 N Nariyama, Shigeko 81 Nickerson, Catherine 30 Noordman, Leo 106–108, 113 Nunberg, Geoffrey 31 Nwogu, Kevin 85, 100 O O’Donnell, Michael 154 Oberlander, Jon 154 Ocampo, Francisco 67 Ochs, Elinor 11, 79 Ortony, Andrew 21 Östman, Jan-Ola 57 Oswood, Charles E. 213 Oversteegen, Leonoor E. 154 P Painter, Claire 29, 55 Paltridge, Brian 19, 24, 29–31 Paolino, Danae 44 Paris, Cécile 107, 154 Parsons, Gerald 168 Partee, Barbara H. 54 Paulson, Eric J. 102 Perfetti, Charles A. 54 Péry-Woodley, Marie-Paule 110 Pike, Kenneth L. 14 Pinto dos Santos, V. B. M. 30 Polanyi, Livia 181 Pollack, Martha E. 112 Portolés, José 148, 154 Posner, Michael I. 25 Potter, John M. 103 Prince, Ellen F. 68, 99 Propp, Vladimir 16 Q Quantz, J. Joachim 189
TSL[v.20020404] Prn:23/09/2004; 13:24
F: PB129NI.tex / p.5 (249)
Name index
R Ramm, Wiebke 64 Ramsay, Guy 110 Ravelli, Louise J. 54 Redeker, Gisela 110, 112, 116 Reichenberger, Klaus 154 Reid, Ian 29 Reitter, David 110 Reppen, Randi 19, 210 Reyle, Uwe 189 Reynolds, R. E. 21 Rich, Charles 5, 34 Riesbeck, Christopher K. 20 Rivero, María Luisa 67 Rocchi, Cesare 154 Roger, Derek 39 Romero Trillo, Jesús 104, 213 Rondhuis, Klaas Jan 108 Rosch, Eleanor 23, 24 Rose, David 56, 63, 99 Rösner, Dietmar 110–112 Rothery, Joan 7, 25, 27, 29 Rumelhart, David E. 21 S Sacks, Harvey 43, 44, 116, 142, 157, 181, 185, 188, 192, 195 Samper Padilla, José Antonio 39 Samraj, Betty 29, 31 Sanders, Ted 106–108, 113, 114, 154 Sanford, Anthony J. 21 Scha, R. J. H. 181 Schaefer, E. F. 26, 116, 145 Schallert, D. L. 21 Schank, Roger C. 20 Schegloff, Emmanuel 43, 44, 77, 116, 142, 157, 181, 185, 188, 192, 195, 213 Schiffrin, Deborah 5, 144, 145 Schleppegrell, Mary J. 154 Scholtens, Anneke 21 Schuetze-Coburn, Stephan 44 Schütze, Hinrich 31 Scott, Donia 154 Searle, John 188
Sgall, Petr 55, 58 Shepherd, Michael 31 Shultz, Jeffrey J. 118 Sidner, Candace L. 5, 34, 56, 106, 107, 112 Silva-Corvalán, Carmen 68, 103 Sinclair, John 16, 157, 188 Slade, Diana 5, 32, 33, 38, 181, 187 Smith, Jonathan 31 Smith, Raoul N. 164 Solé, Yolanda P. 174 Solso, Robert L. 25 Sperber, Dan 154 Spooner, Reverend William A. 53, 103 Spooren, Wilbert 106, 108, 113, 150, 154 Stede, Manfred 106, 110–112, 114 Steiner, Erich H. 64 Stenström, Anna-Brita 181 Stent, Amanda 113 Stoddard, Sally 157, 179 Stone, Matthew 106, 148 Strevens, Peter 10–12, 17 Suñer, Margarita 67, 68, 72 Swales, John M. 22, 23, 27, 29, 31, 34, 36 T Taboada, María Teresa (Maite) xvi, 26, 41, 44, 67, 70, 205 Taglicht, Josef 63 Tannen, Deborah 21 ten Have, Paul 77 Thompson, Geoff 63 Thompson, Sandra A. xv, 1, 2, 63, 106–113, 115, 119, 129, 133, 142 Threadgold, Terry 12, 13, 22, 29 Todorov, Tzvetan 36 Tomlin, Russell S. 66, 67 Torrego, Esther 67 Traum, David 106, 189 Trosborg, Anna 31 Troya Déniz, Magnolia 39 Tsui, Amy B. M. 118, 188, 206
TSL[v.20020404] Prn:23/09/2004; 13:24
F: PB129NI.tex / p.6 (250)
Name index
U Ullman, Jeffrey D. 181 Upton, Thomas A. 30 Ure, Jean 10 V Vaidya, Anuj 37 Vallduví, Enric 54–56, 103 van Dijk, Teun A. 14, 31, 156 van Gorp, Hendrik 24 Vander Linden, Keith 154 Vasconcellos, Muriel 70 Ventola, Eija 16, 17, 19, 24, 29, 181, 210 Vigara Tauste, Ana M. 176 Virtanen, Tuija 57 W Waibel, Alexander H. xvi, xvii, 37, 38 Walker, Marilyn 176
Wanner, Leo 106, 114 Watters, Carolyn 31 Webber, Bonnie 106, 148 Weigand, Edda 49 Weiss, Gilbert 31 Wilkes-Gibbes, Deanna 145 Williams, M. P. 56 Wilson, Deirdre 154 Wittgenstein, Ludwig 24 Wodak, Ruth 31
Y Yngve, Victor H. 212 Yule, George 20, 21
Z Zancanaro, Massimo 154 Ziv, Yael 148 Zubizarreta, María Luisa 56, 67 Zwaan, Rolf A. 30
TSL[v.20020404] Prn:28/09/2004; 9:36
F: PB129SI.tex / p.1 (251)
Subject index
A academic genres see genres Accept (speech act) 190, 192–195, 211, 212 Acceptance 185–192, 196–200, 205 adjacency constraint 119 adjacency pairs 116, 118, 142, 145, 148, 157, 188, 192, 195 Adjunct 59, 60, 63, 65, 66, 70, 71, 72, 74, 77–81, 83, 84, 101, 102, 197, 209 see also conjunctive adjunct; modal adjunct agreement marker 72 All-focus structure 103 Alternative (rhetorical relation) 111, 113, 137 anaphora 160, 165, 173, 179, 205 Antithesis (rhetorical relation) 110, 111, 115, 122, 141, 149–151, 153 Argumentation Theory 154 Ask-date (speech act) 189, 190, 192, 193–196, 211, 212 attitudinal Theme 60, 61 Attribute 80, 85
B Backchannel (speech act) 190, 192, 212, 213 backchannel 212, 213 Background (rhetorical relation) 110, 111, 115, 121, 122, 132, 141, 149–151, 153 background in information structure 55, 56
background knowledge 7, 20, 21, 136, 145 body (of a conversation) 181 Brand-new 68 business communication 30 business genres see genres
C Carrier 80, 85 casual conversation 32, 34, 38, 79, 80, 82, 145, 184 see also spontaneous conversation cataphora 160 central tokens 169, 180 chain interaction 168, 169, 177, 180, 208 Channel (speech act) 190, 192, 213 Chinese 110 chunk 32 thematic chunk 93 Circumstance 60, 66, 67, 70, 74, 78, 80–85, 100–102, 209 Circumstance Adjunct 80, 81, 102, 209 Circumstance Subject 80, 81, 102, 209 Circumstance (rhetorical relation) 109, 110, 122, 141, 149 circumstantial framework 61 classroom discourse 16, 157 classroom genres see genres clausal ellipsis 163, 166, 170 clausal substitution 162, 166, 170, 174
TSL[v.20020404] Prn:28/09/2004; 9:36
F: PB129SI.tex / p.2 (252)
Subject index
clause and cohesive structure 156–160, 164, 166, 167, 175 and filled pauses 214 and RST analysis 110, 111, 128, 142 and SDUs 48, 72, 73, 121, 189 as unit of analysis 44 boundary 44 intonation 43 thematic structure of the clause xv, 2, 53, 55–67, 75, 79–85, 102, 103 topic 57–62 Transitivity structure 56, 75, 85 see also clause relations; declarative clauses; imperative clauses; interrogative clauses; minor clauses clause as exchange 56 clause as message 56 clause as representation 56 clause complex 58, 73 clause relations 106 clitic 70–72, 83, 103 clitic doubling 71, 83 Closing stage xv, 27, 34, 181–185, 188, 192, 195, 197, 201, 202, 205, 206, 213, 215, 221 cluster analysis 101 co-classification 168 co-extension 168 co-reference 168 coherence 1, 5, 23, 86, 106–108, 204, 207 and cohesion xv, 5, 155, 156, 158, 159, 168, 169, 207 and discourse markers 144, 145 in dialogue 116 relations 105–108, 113, 114, 154 Coherence Rule 118 cohesion xv, 1–5, 14, 33, 35, 99, 113, 155–166, 201–209 and staging 201–205
distance of 166, 167 in scheduling dialogues 169–180 see also coherence and cohesion cohesive chains 155, 166–169, 175, 176, 178, 201–205, 208–210 cohesive devices 2, 4, 172, , 176, 179, 208, 209 cohesive harmony 168, 169, 177–180, 203 cohesive relations 156, 157, 159, 160 cohesive ties 156–158, 164–168, 171, 176, 179 cohesiveness 133, 156, 204 collaboration 5, 34, 55 collaborative dialogue xv, 5, 113, 138, 145 collocation 160, 165–167, 170 combiners 57 see also separators comment 56, 58 see also focus; topic common ground 145 Communicative Dynamism (CD) 86 comparative reference 161, 166, 170, 171 Comprehensive Locus of Effect 133, 135 computational linguistics 31, 37, 65, 114, 153, 154, 188, 189, 205, 209, 210 Concession (rhetorical relation) 110, 115, 117, 121–125, 128, 134, 136, 138–141, 149, 151 Condition (rhetorical relation) 110, 121, 122, 125, 126, 129, 134, 136, 138, 139, 141, 149–151, 200, 220 conjunction xv, 2, 14, 33, 160, 163–165 conjunctive adverbs 59, 60, 66, 74–76 conjunctive relations 2, 105, 108, 164, 166 conjunctive Theme 61
TSL[v.20020404] Prn:28/09/2004; 9:36
F: PB129SI.tex / p.3 (253)
Subject index
connectedness 118, 119 connexity 1, 86 constant thematic progression 87, 89, 90–92, 94, 96, 97, 101, 102, 198, 199, 205 Constraints on the Nucleus 109 Constraints on the Satellite 109, 111 context of culture 10, 12, 18, 52, 158 context of situation 1, 10, 12, 18, 90, 93, 116, 158, 159, 161 contextual configuration 14, 15, 24 contextual deduction 36 contiguous thematic progression 86, 87, 89, 90, 92, 94–97, 101 continuative elements 60, 65, 74–77, 164 continuous Theme 87, 88, 90 Contrast (rhetorical relation) 109, 110, 122, 134, 136, 141, 149, 151 contrastive generic analysis xvi, 35 conversation as collaborative effort xv, 5, 21 as opposed to dialogue 49 as product/process 114–116, 133 as text 1, 207 conversational organization xvi, 23, 192 planning in conversation 21 structure 37, 121, 181–188 see also casual conversation; task-oriented dialogue; telephone conversation Conversation Analysis 5, 41, 44, 49, 116, 189 conversation-as-a-whole analysis 105, 114 –116, 119, 133 co-reference 71, 168 Critical Discourse Analysis 31 cross-cultural analysis 52, 211, 222 cross-linguistic analysis 52, 112, 113, 206, 211, 222 cross-talk 40, 44 cue phrases 114, 147 curriculum design 7, 8
curriculum genres see genres cybergenres see genres D Date-Proposal stage 181–183, 185–187, 192–194, 203, 211, 212, 220 declarative clauses 59–61, 65, 71, 72 delaying expressions 214 demonstrative reference 156, 161, 166, 170–172, 202 derived Theme 87–97, 104, 171 Details (in a Proposal) 135, 136, 186–188, 197, 200, 205 dialect 10, 17, 19 of Spanish 52 dialogic xvi, 2, 10, 54, 116 Dialogic RST 113 dialogue see conversation; dialogue; task-oriented dialogue discourse acts 188 discourse markers 2, 105, 107, 114, 121, 123–125, 128, 130, 131, 140, 142–145, 147–154, 164 discourse particles 148 discourse psychology 7 Discourse Representation Theory (DRT) 108, 189 discourse structure 34, 106, 112, 157 discourse (structure) relations 106, 114, 153 distance of cohesion 155, 163, 166, 169, 175, 176, 179 dominant constituent 55 dummy pronoun 93, 99, 103 Dutch 110, 154 dynamic framework 54 dynamic perspective 13, 54 E Effect (in rhetorical relations) 109, 111, 112, 115, 121, 133, 135
TSL[v.20020404] Prn:28/09/2004; 9:36
F: PB129SI.tex / p.4 (254)
Subject index
Elaboration (rhetorical relation) 110, 117, 121, 122, 126, 127, 134–136, 138–141, 147, 149–151, 153, 200, 205 ellipsis 1, 14, 33, 63, 81, 160, 162, 163, 166, 170, 172–174, 180 elliptical Theme 63, 81 embedded clauses 58, 73, 110 Enablement (rhetorical relation) 109–111, 115, 122, 134, 138, 139, 141, 149, 150 end of message signal 40 end-weight 68, 69 endophoric reference 158, 160, 161 English for Academic Purposes 30 English for Specific Purposes 30, 210 entry 181 Evaluation (rhetorical relation) 110, 122, 134–136, 139–142, 149, 150, 200, 205 Evidence (rhetorical relation) 110, 111, 115, 122, 141, 149 exchange (in classroom discourse) 16, 188 exchange model 113 exit 181 exophoric reference 158, 160, 161, 171 Experiencer 71, 72 F face-saving strategies 78, 128, 129, 196, 216 face-threatening act 34, 78, 123 false start 42, 43, 45, 195, 198 family resemblance 24 field 10–12, 15, 17, 18, 27, 28 filled pause 74, 148, 214 Filled-pause (speech act) 189, 190–192, 195, 213 fillers 20 Finite 59, 61, 65, 66, 68, 70–72, 77, 80, 82, 83, 98, 99, 197, 209 finite state machine 181, 191
floor 32, 40, 76, 96, 118, 176, 191, 214 flowchart 16, 24, 29 focus 55–58, 103 folktales 16 foreign language teaching see second language teaching formal language 11, 79 see also informal language formulaic greetings 184, 217 frames 7, 12, 20, 21 framework-setting 59, 69 French 110 functional genre theory 10, 12 Functional Sentence Perspective 55, 93 functional variation 36 G gapped thematic progression 89–92, 94–98, 198 general word 156, 166, 170, 172, 174 generic configuration 1, 28 generic structure 2, 4, 7, 10, 13, 16, 18, 19, 24, 27, 29, 32, 136, 144, 159, 181, 208 Generic Structure Potential 7, 10, 13, 19, 24 generic variation 17 genre 1, 2, 7–28 and cohesion 201–205 and thematic progression 100, 101 and thematic structure 196, 197–199 and rhetorical analysis 116, 119, 145, 199–201 applications 29–31 scheduling dialogues as genre 34–36 genres academic 31 business 30, 31, 181 classroom 29 curriculum 29
TSL[v.20020404] Prn:28/09/2004; 9:36
F: PB129SI.tex / p.5 (255)
Subject index
cybergenres 31 literary 8, 9, 24, 30, 36 on-line 31 genre staging 13, 28, 206 genre theory 5, 8, 10, 12, 17, 29, 32, 207 German 64, 110, 113 Given 2, 55, 57, 58, 86 Goodbye (speech act) 184, 190, 215, 221 goodbye 190, 195, 215 grammatical cohesion 2, 160, 162, 163, 165, 166, 168, 172 grammatical metaphor 80 grammatical mode 67 Greeting (speech act) 189, 190, 191, 193, 206, 215 Gricean principle 150 grounding 145 H holistic structure 119–121, 137 human noises 42, 45, 48 hyperonomy 90 hypertheme 88–90, 93, 102 hyponymy 90, 165 hypotactic relations 58, 75 I ideational metafunction 12, 56, 103, 197 ideational Theme 59, 60, 62–65, 66, 70, 71, 73–83, 88, 90, 93, 94, 101–104 identity chain 168, 169, 179, 204 immediate cohesive distance 166, 167, 175, 179 imperative clauses 59, 61, 72, 197 Indirect Object 66, 70–72 individual framework 61 Inform (speech act) 190, 193–195, 216, 217 informal language 11, 12 see also formal language
Inform-availability (speech act) 189, 190, 193–196, 216–218, 220 information retrieval 31, 204, 210 information structure 33, 56, 57 information unit 44, 53 Initialization 183 see also Opening insertion sequence 195 integration thematic progression pattern 89–92, 94, 97, 98 inter-turn relations 114–116, 131, 140, 143, 144 Interactive Systems Lab (ISL) 37–39, 41, 51 interjections 42, 44 interpersonal metafunction 12, 56, 59 interpersonal Theme 59, 60–63, 65, 66, 70–78, 83, 101–103, 197 Interpretation (rhetorical relation) 110, 122, 136, 141, 149, 151 interrogative clauses 59–61, 72, 74, 83, 125, 126, 142, 143, 174, 220 intonation 5, 43, 44, 56, 72, 79, 131, 164 intonation markers 5, 43, 44 intra-turn relations 113–116, 143 J JANUS project xvi, 37, 41 Justify (rhetorical relation) 110, 111, 115, 122, 141, 149–151 L language metafunctions xv, xvi, 2, 12, 28, 56, 59, 60, 62, 102, 112 see also ideational metafunction; interpersonal metafunction; textual metafunction layers of context 18 lexical cohesion 1, 2, 14, 33, 160, 164–167, 170–174, 179, 202 lexicogrammatical features 8, 14, 19, 33, 196
TSL[v.20020404] Prn:28/09/2004; 9:36
F: PB129SI.tex / p.6 (256)
Subject index
lexicogrammatical patterns 28, 36, 181 liaison 53 linear thematic progression 86–90, 92–97, 101 literary genres see genres logico-semantic relations 105
M machine translation 37, 205, 210 macrostructure 2, 156 major chains 180, 202, 203, 208 marked (ideational) Theme 59–61, 63, 65–67, 70, 72, 78–81, 83, 93, 101, 103, 197, 198, 200, 201, 205, 206, 208 markedness 67, 79, 103, 208 mediated cohesive ties 166, 167, 175, 176, 179 mental models 21 mental processes 71 metafunction see language metafunctions metathesis 42 method of development 58, 84–86, 100, 101 minor chains 177, 180, 202, 203 minor clauses 72, 73 mispronunciations 42, 45 modal adjuncts 59, 60, 66, 77, 78 mode 9–12, 14, 15, 17–19, 25, 27, 28, 40, 102, 112 mode continuum 12 monologic 105, 116, 207 monorhematic sentences 103 Mood 56, 59, 65–67, 70–72, 77, 149 mood-marking elements 60, 66, 77 Motivation (rhetorical relation) 109–111, 115, 122, 141, 147, 149 move (in classroom discourse) 16, 188, 189 multiple thematic progression 89–92, 94, 97, 104
multiple Theme 59–61, 65, 66, 75, 76
N N-Rheme 58 narrative 11, 33, 79 Natural Language Generation 65, 107, 110, 113, 154, 210 naturally-occurring conversation 38, 39, 107, 116 see also spontaneous conversation new datum 69 New Proposal 185, 186, 188, 192, 197, 198, 200, 203–206, 208 New Rhetoric 29, 31 New thematic progression NewInfo 55 Nigel grammar 65 nominal ellipsis 163, 166, 170 nominal substitution 162, 166, 170, 173 non-central tokens 169, 180 non-derived Themes 90–94, 96, 97, 171 non-hierarchical dyads 38 non-human noises 42, 45, 48 non-propositional floor 118 non-sequiturs 118 non-structural cohesion xv, 2, 33, 157, 179 non-text 24, 155 Non-Volitional Cause (rhetorical relation) 110, 122, 128, 129, 139, 140, 141, 149–151, 153, 199 Non-Volitional Result (rhetorical relation) 110, 111, 122, 128–131, 136, 139–141, 144, 149–151, 199, 200 North American genre theory 8, 29 nucleus 109, 111, 112, 116, 117, 121, 123, 126, 128, 136, 138, 140, 142, 146, 154, 199 nursery tales 7, 13, 33
TSL[v.20020404] Prn:28/09/2004; 9:36
F: PB129SI.tex / p.7 (257)
Subject index
O Object 31, 59, 66, 70–72 obligatory stages 14, 15, 24, 33, 187 observer’s paradox 39 Old English 79, 83 on-line genres see genres Opening stage xv, 13, 27, 34, 101, 116, 181–184, 188, 191, 192, 197, 199, 201, 205, 206, 208, 215, 220, 222 open-proposition 55 optional stages 15, 16, 20, 24, 33, 184, 187, 191, 192 Other (speech act) 190, 217 Otherwise (rhetorical relation) 110, 122, 141, 149, 151 P paratactic relations 75, 107 parsing 31, 154 Participant 60, 61, 66, 71, 80, 81, 83–85, 99, 101–103, 209 Participant Adjunct 80, 81, 83 Participant Subject 71, 80, 81 pause 26, 40, 42, 44, 69, 74, 104, 211, 213 see also filled pause Penman system 65 peripheral tokens 169, 180 personal pronouns 98, 161, 172 personal reference 161, 166, 170, 171 Pesky Little Particle 148 Place-Proposal stage 182, 183, 187, 188, 192, 194, 203, 211, 218, 221 planned discourse 1, 11, 79, 80 plausibility judgments 109, 112, 124, 125, 136, 137, 154 point of departure 44, 57–59, 61–63, 84, 99 politeness 123, 128, 129, 209, 216 Politey (speech act) 190–192, 217 possessive determiners 161 possessive pronouns 161
pragmatic markers 148 pragmatic mode 67, 68 Prague School 55, 58, 86 Precondition (rhetorical relation) 111 presentational construction 68, 69 presentational rhetorical relations 111, 112, 115 presentative construction see presentational construction presupposed element 156, 158, 174 presupposition 55, 156 primary genres 9, 19 pro-drop 68, 163, 209 Process 60, 61, 63, 65, 66, 71, 74, 80–85, 99–102, 104, 173, 209 proform anaphora 173, 174 Proposal stage see Place-Proposal stage; Date-Proposal stage Propose-action (speech act) 190, 193, 194, 211, 218, 219 Propose-place (speech act) 190, 193–195, 218 prototype theory 20, 21, 23–25 Purpose (rhetorical relation) 110, 122, 134, 135, 141, 149–151, 153 push-to-talk 40, 41 R reciprocity 184 recursion 110, 187 reference 1, 14, 33, 63, 83, 99, 104, 156, 158, 160–162, 165, 166, 168–174, 179, 180, 202, 205, 209 register 7–13, 16–19, 27, 28, 32, 36, 53, 100, 105, 196, 210 Register and Genre Theory 17 register variation 17, 210 reiteration 164, 166 Reject-date (speech act) 189, 190, 192–194, 218, 219 relation definition 109, 110, 154 relational structure 119, 120 Relevance Theory 154
TSL[v.20020404] Prn:28/09/2004; 9:36
F: PB129SI.tex / p.8 (258)
Subject index
relevant tokens 169 remote 166, 167, 175, 176 Repeat-confirm (speech act) 189, 190, 193–195, 219 repetition 160, 164, 167, 171, 173, 176, 179, 202 Request-action (speech act) 190, 193–195, 211, 218, 219 Request-confirmation (speech act) 190, 220 Request-date (speech act) 189, 190, 193, 212, 220 Request-information (speech act) 190, 213, 220 Request-meeting (speech act) 190, 193, 196, 221 Request-place (speech act) 190, 193, 221 response tokens 213 Restatement (rhetorical relation) 110, 122, 127, 131, 137, 139, 140, 141, 143, 144, 149–151, 201, 205 rhemic 90–92, 96, 97 see also linear thematic progression Rhetorical Genre Studies 8 rhetorical predicates 106 Rhetorical Structure Theory (RST) xvi, 1, 2, 4, 5, 105, 108–114, 153, 207, 208, 220 S saccades 53 satellite 109, 111, 112, 116, 123, 125–127, 135, 136, 138, 140, 142, 146, 199, 220 Sayer 85 scenarios 21, 24, 39 scheduling dialogues 7, 34, 40, 51, 181, 182, 206, 209 scheduling genre 34–36, 207, 208 schema types in RST 109, 110, 128, 146, 154 schemata 7, 12, 20–23, 55
scope (of discourse markers) 145, 146 scripts 7, 20, 21 SDU see Semantic Dialogue Unit second language teaching 7, 8, 29, 30, 153, 210 secondary genres 9, 19 Segmented Discourse Representation Theory 108 Self-introduction (speech act) 190, 221, 222 self-talk 116, 119 Semantic Dialogue Unit (SDU) 44, 48–51, 72, 73, 121, 183, 189, 190, 201, 202 Semantic End of Segment (SEOS) 44, 45, 51, 72 sentence adverbials 148 SEOS see Semantic End of Segment separation thematic progression pattern 90–92, 94, 97, 98 separators 57 see also combiners Sequence (rhetorical relation) 109–111, 118, 122, 136, 141, 147, 149–151, 153 sequential implicativeness 192, 195 service encounters 7, 13, 16, 19, 27, 33, 181, 205, 210 SFL see Systemic Functional Linguistics side sequences 195 silence 42, 213, 214 similarity chain 156, 161, 168, 169, 179 simple contiguous constant thematic progression 87, 89, 92, 94 simple contiguous linear thematic progression 87, 89, 92, 94 simple gapped constant thematic progression 89, 92, 94
TSL[v.20020404] Prn:28/09/2004; 9:36
F: PB129SI.tex / p.9 (259)
Subject index
simple gapped linear thematic progression 89, 92, 94, 95 simple gapped thematic progression derived from hypertheme 89 simple linear thematic progression 86, 87 simple thematic progression derived from hypertheme 88–90, 93, 102 single Theme 60, 61, 63, 64, 75 slots (in scripts) 20 social semiotics 19, 29 sociorhetorical discourse community 34 Solutionhood (rhetorical relation) 110, 122, 124, 134–136, 138–143, 148–150, 199, 200, 205, 208 span 109, 110, 116–118, 121, 124, 128, 133, 135–138, 140, 146, 148, 184, 199–201 speech acts 2, 16, 188–196, 205, 207, 209, 211–222 speech events 181 speech genres 2, 4, 7–10, 34 split Rheme 88–90 spoken interaction 181 spoken language 4, 5, 9, 11, 15, 23, 25, 30, 32, 53, 79, 101, 102, 208, 210 and cohesion 158, 163, 176, 179, 180 and rhetorical analysis 112–120, 153 spontaneous conversation 39, 70, 79 spoonerism 53, 103 stages 2, 16, 26, 27, 32, 33, 36 and cohesion 201–205 and rhetorical relations 199–201 and speech acts 190–196, 211–222 and thematic structure 196–199
in scheduling dialogues xvi, 4, 34, 35, 101, 181–188, 208–210 staging structure xvi, 4, 12, 13, 18, 25, 28, 105, 207 starting point see point of departure Step-sequence (rhetorical relation) 111 structural cohesion 2, 3, 156, 157, 164, 179 structural formula 13–15, 18, 33, 187, 188 structure 13, 14, 156–159 see also texture style of discourse 11 Subject 59, 61–63, 65–72, 74, 80–85, 93–99, 101–104, 110, 163, 171, 173, 197, 209 Subject ellipsis 68, 70, 81, 83, 102. 163 subject matter 11, 74, 93, 100–102, 115, 126, 197 subject matter rhetorical relations 111, 112 Subject omission see Subject ellipsis Subject-Verb inversion 68, 70, 72 subordinate clauses 73, 75, 107, 111, 158 subordinate lexical cohesion 165, 170, 171, 174 substages 185, 197, 198, 202, 203, 205 substitution (cohesive relation) 1, 14, 33, 160, 162, 166, 170, 172–174, 179, 180, 208, 209 Summary (rhetorical relation) 110, 122, 141, 149–151, 201, 205 superordinate lexical cohesion 165, 166, 170 SV 67–69 SVO 67, 68, 70 synonymy 160, 165, 166, 170 synoptic framework 13, 54
TSL[v.20020404] Prn:28/09/2004; 9:36
F: PB129SI.tex / p.10 (260)
Subject index
Syntactic thematic progression 93, 94, 99 system networks 60, 61, 65 Systemic Functional Linguistics (SFL) 5, 7, 8, 13, 16, 18, 19, 28, 29, 36, 55, 56, 58, 60, 65, 103, 207 T T-unit 58, 59 tag questions 42 task-oriented dialogue 1, 5, 34, 51, 52, 79, 114, 119, 145, 153, 181, 189, 207, 210 Task Performance stage xv, 27, 181–183, 185, 186, 197–208, 211, 215, 217, 221 taxonomy of rhetorical relations 107, 111, 112, 114, 153, 154 taxonomy of speech acts 188, 189, 205 taxonomy of speech genres 34, 101 telephone conversation 15, 75, 77, 181, 184, 221 tenor 10–12, 15, 17, 18, 27, 28 text generation see Natural Language Generation Text Summarization 154 text types 8, 12, 17, 19, 31, 32, 79, 100, 101, 108, 168 textual metafunction xiii, xvi, 2, 4, 12, 56, 59, 102, 181 textual prediction 36 textual Theme xiii, 59–63, 65, 66, 70, 71, 73–76, 101–103, 197 texture xvi, 2, 4, 5, 14, 18, 155–157, 208, 209 thematic chains 91 thematic chunks 93 thematic progression xv, 1, 4, 33, 53, 56, 83–88, 91–96, 99–102, 104, 171, 196, 198, 207–209 thematic realization 4, 53, 64, 72, 81, 207 thematic selection 65, 71, 72, 75, 80, 83, 102, 196, 197
thematic structure xv, 2, 35, 53, 57, 196, 197 thematization 54, 100 Theme and Subject 65, 66, 71 Theme in English 55–67 Theme in Spanish 67–72 Theme markedness 79, 208 Theme-Rheme 2, 33, 58, 67, 72, 86, 197 ThemeF 64 ThemeM 64 themic progression 90–92, 96, 97, 101, 102, 198 see also constant thematic progression thetic judgments 103 three-pair exchanges 195, 206 tone group 44, 157 topic 1, 19, 53, 55, 56, 58–62, 68, 82, 88, 93, 100, 103, 104 topic continuity 68 Topic-Focus Articulation 55 topical Theme 59–63, 65, 66, 72, 83, 103, 104 TP see thematic progression transactions in classroom discourse 16, 188 transcriber comments 42–45 transcription conventions xiv, 41–45, 117 Transitivity 56, 72, 75, 78, 80–85, 100, 102, 168, 209 turn-by-turn analysis 116, 119, 121, 122, 132, 133, 140, 148 turn-constructional units 44, 213 turn-holding device 77 U underspecification 150 unit markers 44, 45, 51 unmarked (ideational) Theme 59–61, 63, 65–67, 70–72, 78–81, 83, 84, 93, 102, 197, 198 see also marked Theme
TSL[v.20020404] Prn:28/09/2004; 9:36
F: PB129SI.tex / p.11 (261)
Subject index
unplanned discourse 11 Until (rhetorical relation) 111 Unused 68 utterance 8, 9, 36, 37, 41, 44, 49 vs. turn 49
Volitional Result 110, 122, 129, 131, 132, 136, 138–141, 144, 146, 149–151, 153, 200 VS 67, 68, 70 VSO 67, 70
V Verb 31, 68–72, 83, 98, 99, 162, 163, 173, 174 verbal ellipsis 163, 166, 170 verbal substitution 162, 166, 170, 173 Vocative (speech act) 190–192, 221, 222 vocative as Theme 60, 66, 72, 77 Volitional Cause 110, 122, 141, 146, 149, 151, 199
W wh-question 59, 61, 65, 66, 72, 74, 77, 81 word order 51, 59, 64, 67, 68, 70, 72, 78, 103 written language 75, 76, 79, 100, 112, 113, 115, 116, 158, 208 Y yes-no question 65, 71, 72, 77
In the Pragmatics & Beyond New Series the following titles have been published thus far or are scheduled for publication: 79 ANDERSEN, Gisle and Thorstein FRETHEIM (eds.): Pragmatic Markers and Propositional Attitude. 2000. viii, 273 pp. 80 UNGERER, Friedrich (ed.): English Media Texts – Past and Present. Language and textual structure. 2000. xiv, 286 pp. 81 DI LUZIO, Aldo, Susanne GÜNTHNER and Franca ORLETTI (eds.): Culture in Communication. Analyses of intercultural situations. 2001. xvi, 341 pp. 82 KHALIL, Esam N.: Grounding in English and Arabic News Discourse. 2000. x, 274 pp. 83 MÁRQUEZ REITER, Rosina: Linguistic Politeness in Britain and Uruguay. A contrastive study of requests and apologies. 2000. xviii, 225 pp. 84 ANDERSEN, Gisle: Pragmatic Markers and Sociolinguistic Variation. A relevance-theoretic approach to the language of adolescents. 2001. ix, 352 pp. 85 COLLINS, Daniel E.: Reanimated Voices. Speech reporting in a historical-pragmatic perspective. 2001. xx, 384 pp. 86 IFANTIDOU, Elly: Evidentials and Relevance. 2001. xii, 225 pp. 87 MUSHIN, Ilana: Evidentiality and Epistemological Stance. Narrative Retelling. 2001. xviii, 244 pp. 88 BAYRAKTAROĞLU, Arın and Maria SIFIANOU (eds.): Linguistic Politeness Across Boundaries. The case of Greek and Turkish. 2001. xiv, 439 pp. 89 ITAKURA, Hiroko: Conversational Dominance and Gender. A study of Japanese speakers in first and second language contexts. 2001. xviii, 231 pp. 90 KENESEI, István and Robert M. HARNISH (eds.): Perspectives on Semantics, Pragmatics, and Discourse. A Festschrift for Ferenc Kiefer. 2001. xxii, 352 pp. 91 GROSS, Joan: Speaking in Other Voices. An ethnography of Walloon puppet theaters. 2001. xxviii, 341 pp. 92 GARDNER, Rod: When Listeners Talk. Response tokens and listener stance. 2001. xxii, 281 pp. 93 BARON, Bettina and Helga KOTTHOFF (eds.): Gender in Interaction. Perspectives on femininity and masculinity in ethnography and discourse. 2002. xxiv, 357 pp. 94 McILVENNY, Paul (ed.): Talking Gender and Sexuality. 2002. x, 332 pp. 95 FITZMAURICE, Susan M.: The Familiar Letter in Early Modern English. A pragmatic approach. 2002. viii, 263 pp. 96 HAVERKATE, Henk: The Syntax, Semantics and Pragmatics of Spanish Mood. 2002. vi, 241 pp. 97 MAYNARD, Senko K.: Linguistic Emotivity. Centrality of place, the topic-comment dynamic, and an ideology of pathos in Japanese discourse. 2002. xiv, 481 pp. 98 DUSZAK, Anna (ed.): Us and Others. Social identities across languages, discourses and cultures. 2002. viii, 522 pp. 99 JASZCZOLT, Katarzyna M. and Ken TURNER (eds.): Meaning Through Language Contrast. Volume 1. 2003. xii, 388 pp. 100 JASZCZOLT, Katarzyna M. and Ken TURNER (eds.): Meaning Through Language Contrast. Volume 2. 2003. viii, 496 pp. 101 LUKE, Kang Kwong and Theodossia-Soula PAVLIDOU (eds.): Telephone Calls. Unity and diversity in conversational structure across languages and cultures. 2002. x, 295 pp. 102 LEAFGREN, John: Degrees of Explicitness. Information structure and the packaging of Bulgarian subjects and objects. 2002. xii, 252 pp. 103 FETZER, Anita and Christiane MEIERKORD (eds.): Rethinking Sequentiality. Linguistics meets conversational interaction. 2002. vi, 300 pp. 104 BEECHING, Kate: Gender, Politeness and Pragmatic Particles in French. 2002. x, 251 pp. 105 BLACKWELL, Sarah E.: Implicatures in Discourse. The case of Spanish NP anaphora. 2003. xvi, 303 pp. 106 BUSSE, Ulrich: Linguistic Variation in the Shakespeare Corpus. Morpho-syntactic variability of second person pronouns. 2002. xiv, 344 pp.
107 TAAVITSAINEN, Irma and Andreas H. JUCKER (eds.): Diachronic Perspectives on Address Term Systems. 2003. viii, 446 pp. 108 BARRON, Anne: Acquisition in Interlanguage Pragmatics. Learning how to do things with words in a study abroad context. 2003. xviii, 403 pp. 109 MAYES, Patricia: Language, Social Structure, and Culture. A genre analysis of cooking classes in Japan and America. 2003. xiv, 228 pp. 110 ANDROUTSOPOULOS, Jannis K. and Alexandra GEORGAKOPOULOU (eds.): Discourse Constructions of Youth Identities. 2003. viii, 343 pp. 111 ENSINK, Titus and Christoph SAUER (eds.): Framing and Perspectivising in Discourse. 2003. viii, 227 pp. 112 LENZ, Friedrich (ed.): Deictic Conceptualisation of Space, Time and Person. 2003. xiv, 279 pp. 113 PANTHER, Klaus-Uwe and Linda L. THORNBURG (eds.): Metonymy and Pragmatic Inferencing. 2003. xii, 285 pp. 114 KÜHNLEIN, Peter, Hannes RIESER and Henk ZEEVAT (eds.): Perspectives on Dialogue in the New Millennium. 2003. xii, 400 pp. 115 KÄRKKÄINEN, Elise: Epistemic Stance in English Conversation. A description of its interactional functions, with a focus on I think. 2003. xii, 213 pp. 116 GRANT, Colin B. (ed.): Rethinking Communicative Interaction. New interdisciplinary horizons. 2003. viii, 330 pp. 117 WU, Ruey-Jiuan Regina: Stance in Talk. A conversation analysis of Mandarin final particles. 2004. xvi, 260 pp. 118 CHENG, Winnie: Intercultural Conversation. 2003. xii, 279 pp. 119 HILTUNEN, Risto and Janne SKAFFARI (eds.): Discourse Perspectives on English. Medieval to modern. 2003. viii, 243 pp. 120 AIJMER, Karin and Anna-Brita STENSTRÖM (eds.): Discourse Patterns in Spoken and Written Corpora. 2004. viii, 279 pp. 121 FETZER, Anita: Recontextualizing Context. Grammaticality meets appropriateness. 2004. x, 272 pp. 122 GONZÁLEZ, Montserrat: Pragmatic Markers in Oral Narrative. The case of English and Catalan. 2004. xvi, 410 pp. 123 MÁRQUEZ REITER, Rosina and María Elena PLACENCIA (eds.): Current Trends in the Pragmatics of Spanish. 2004. xvi, 383 pp. 124 VINE, Bernadette: Getting Things Done at Work. The discourse of power in workplace interaction. 2004. x, 278 pp. 125 LERNER, Gene H. (ed.): Conversation Analysis. Studies from the first generation. 2004. x, 302 pp. 126 WU, Yi’an: Spatial Demonstratives in English and Chinese. Text and Cognition. 2004. xviii, 236 pp. 127 BRISARD, Frank, Michael MEEUWIS and Bart VANDENABEELE (eds.): Seduction, Community, Speech. A Festschrift for Herman Parret. vi, 196 pp. + index. Expected Fall 2004 128 CORDELLA, Marisa: The Dynamic Consultation. A discourse analytical study of doctor–patient communication. xvi, 254 pp. Expected Fall 2004 129 TABOADA, María Teresa: Building Coherence and Cohesion. Task-oriented dialogue in English and Spanish. 2004. xvii, 261 pp. 130 HALMARI, Helena and Tuija VIRTANEN (eds.): Persuasion Across Genres. A linguistic approach. viii, 244 pp. + index. Expected Winter 04-05 131 JANOSCHKA, Anja: Web Advertising. New forms of communication on the Internet. xiv, 221 pp. + index. Expected Winter 04-05 132 ONODERA, Noriko: Japanese Discourse Markers. Synchronic and diachronic discourse analysis. xiv, 240 pp. + index. Expected Winter 04-05 133 MARNETTE, Sophie: Speech and Thought Presentation in French. Concepts and strategies. Expected Spring 2005
A complete list of titles in this series can be found on the publishers website, www.benjamins.com
