This page intentionally left blank
Conversation Analysis “Conversation analysis” is an approach to the study of social interaction that focuses on practices of speaking that recur across a range of contexts and settings. The early studies in this tradition were based on the analysis of English conversation. More recently, however, conversation analysts have begun to study talk in a broader range of communities around the world. Through detailed analyses of recorded conversations, this book examines differences and similarities across a wide range of languages including Finnish, Japanese, Tzeltal Mayan, Russian and Mandarin. Bringing together interrelated methodological and analytic contributions, it explores topics such as the role of gaze in question-and-answer sequences, the organization of repair, and the design of responses to assessments. The emerging comparative perspective demonstrates how the structure of talk is inflected by the local circumstances within which it operates. jac k si dn e l l is an Associate Professor in the Department of Anthropology at the University of Toronto.
Studies in Interactional Sociolinguistics editors Paul Drew, Majorie Harness Goodwin, John J. Gumperz, Deborah Schifrin 1. Discourse Strategies John J. Gumperz 2. Language and Social Identity edited by John J. Gumperz 3. The Social Construction of Literacy Jenny Cook-Gumperz 4. Politeness: Some Universals in Language Usage Penelope Brown and Stephen C. Levinson 5. Discourse Markers Deborah Schiffrin 6. Talking Voices: Repetition, Dialogue, and Imagery in Conversational Discourse Deborah Tannen 7. Conducting Interraction: Patterns of Behaviour in Focused Encounters Adam Kendon 8. Talk at Work: Interaction in Institutional Settings edited by Paul Drew and John Heritage 9. Grammar in Interaction: Adverbial Clauses in American English Conversations Cecilia E. Ford 10. Crosstalk and Culture in Sino-American Communication Linda W.L. Young (with a foreword by John J. Gumperz) 11. AIDS Counselling: Institutional Interaction and Clinical Practice Anssi Perakyla 12. Prosody in Conversation; Interactional Studies edited by Elizabeth Couper-Kuhlen and Margret Selting 13. Interaction and Grammar edited by Elinor Ochs, Emanual A. Schegloff and Sandra A. Thompson 14. Credibility in Court: Communicative Practices in the Camorra Triales Marco Jacquement 15. Interaction and the Development of Mind Anthony J. Wootton 16. The News Interview: Journalists and Public Figures on the Air Steven Clayman and John Heritage 17. Gender and Politeness Sara Mills 18. Laughter in Interaction Philip Glenn 19. Matters of Opinion: Talking about Public Issues Greg Myers 20. Communication in Medical Care: Interaction between Primary Care Physicians and Patients edited by John Heritage and Douglas Maynard 21. In Other Words: Variation in reference and narrative Deborah Schiffrin 22. Language in Interaction in an urban school Ben Rampton 23. Discourse and Identity edited by Anna De Fina, Deborah Schiffrin and Michale Bamberg 24. Reporting Talk: Reported Speech in Interaction edited by Elizabeth Holt and Rebecca Clift 25. The Social Construction of Literacy, 2nd edition by Jenny Cook-Gumperz 26. Talking Voices, 2nd edition by Deborah Tannen 27. Conversation Analysis: Comparative Perspectives edited by Jack Sidnell
Conversation Analysis Comparative Perspectives Edited by
Jack S i dn ell University of Toronto
CAMBRIDGE UNIVERSITY PRESS
Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo, Delhi, Dubai, Tokyo Cambridge University Press The Edinburgh Building, Cambridge CB2 8RU, UK Published in the United States of America by Cambridge University Press, New York www.cambridge.org Information on this title: www.cambridge.org/9780521883719 © Cambridge University Press 2009
This publication is in copyright. Subject to statutory exception and to the provision of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published in print format 2009 ISBN-13
978-0-511-63432-1
eBook (EBL)
ISBN-13
978-0-521-88371-9
Hardback
Cambridge University Press has no responsibility for the persistence or accuracy of urls for external or third-party internet websites referred to in this publication, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.
Contents
List of figurespage vii List of tablesix List of contributorsxiii Transcription conventionsxv Part I Introduction 1
Comparative perspectives in conversation analysis Jack Sidnell
Part II Repair and beyond 2
Repetition in the initiation of repair Ruey-Jiuan Regina Wu
3
A cross-linguistic investigation of the site of initiation in same-turn self-repair Barbara Fox, Fay Wouk, Makoto Hayashi, Steven Fincke, Liang Tao, Marja-Leena Sorjonen, Minna Laakso, and Wilfrido Flores Hernandez
4
Repairing reference Maria Egbert, Andrea Golato, and Jeffrey D. Robinson
Part III Aspects of response
1 3
29 31
60
104
133
5
Projecting nonalignment in conversation Anna Lindström
135
6
Two answers to inapposite inquiries Trine Heinemann
159
vi
Contents 7
Gaze, questioning, and culture Federico Rossano, Penelope Brown, and Stephen C. Levinson
8 Negotiating boundaries in talk Makoto Hayashi and Kyung-eun Yoon Part IV Action formation and sequencing 9
Alternative responses to assessments Marja-Leena Sorjonen and Auli Hakulinen
10 Language-specific resources in repair and assessments Jack Sidnell 11 Implementing delayed actions Galina B. Bolden Part V Conclusion 12 One perspective on Conversation Analysis: Comparative Perspectives Emanuel A. Schegloff
187
250
279 281
304 326
355 357
Bibliography 407 Index436
Figures
4.1 4.2 4.3
Open-class repair German “Wa?” page 108 Open-class English “What?” 109 Underspecified referent repair initiation (URRI) German “Was.” 110 4.4 URRI German “Was denn.” 110 4.5 URRI English “What.” 111 7.1 Location of three cultures 204 7.2 Gaze toward other participant, by language 208 7.3 Gaze behavior across the three cultures clustered by participation role 209 7.4 Distribution of actions by languages 214 7.5 Distribution of percentages of each action in which the Q-speaker looks at the addressee, 215 by language 7.6 Distribution of percentages of each action in which the Q-recipient looks at the Q-speaker, 217 by language 7.7 Italian dyads 227 7.8 Tzeltal dyads 228 7.9 Yélî Dnye dyads 229 9.1 Responses to assessments with a verb repeat282
Tables
3.1 Post-beginning and pre-completion repair initiation 3.2 Recognizable completion and repair initiation in English 3.3 Recognizable completion and repair initiation in Indonesian 3.4 Recognizable completion and repair initiation in Finnish 3.5 Recognizable completion and repair initiation in Bikol 3.6 Recognizable completion and repair initiation in Japanese 3.7 Recognizable completion and repair initiation in Sochiapam Chinantec 3.8 Recognizable completion and repair initiation in Mandarin 3.9 Site of initiation of simple recycling repair initiated in monosyllabic words 3.10 Site of initiation of simple replacement repair initiated in monosyllablic words 3.11 Site of initiation in recycling repairs in words of three or more syllables 3.12 Site of initiation in replacement repairs in words of three or more syllables 3.13 Frequency of monosyllabic words for each repair type 3.14 Frequency of multisyllabic words for each repair type
page 68 75 77 77 78 78 79 79 81 82 84 84 86 86
x
Tables
3.15 Site of initiation of recycling repairs in bisyllabic words 89 3.16 Percentage of recycling repairs initiated in bisyllabic words 90 3.17 Recycling repairs: Bisyllabic words by syntactic class 91 3.18 Recycling repairs: Multisyllabic words by syntactic class 92 3.19 English replacement repairs by word type for monosyllabic words 93 3.20 Replacement repairs by word type for monosyllabic words 94 3.21 Word length in recycling repairs 95 7.1 Instances of Q-speaker and Q-recipient gaze toward other participant, by language 207 7.2 Logistic regression analysis predicting gaze toward other participant in relation to role (Q-recipient as reference group), by language 209 7.3 Logistic regression analysis predicting Q-speaker gaze toward Q-recipient in relation to language (Italian and Tzeltal as reference group) 210 7.4 Logistic regression analysis predicting Q-recipient gaze toward Q-speaker in relation to language (Italian and Tzeltal as reference group) 211 7.5 Instances of mutual gaze by language 213 7.6 Logistic regression analysis predicting mutual gaze in relation to language (Italian and Tzeltal as reference group) 213 7.7 Questions in which speakers look at recipients and recipients do not look back, by language 214 7.8 Distribution of actions performed by questions, by language 214 7.9 Q-speaker looking at Q-recipient, by action and by language 215 7.10 Q-recipient looking at Q-speaker, by question type and by language 217 7.11 Distribution of response types by language 218 7.12 Distribution of responses to questions, by language 218
Tables
xi
7.13 Distribution of questions that do not get responded to in which the recipient does not look at the speaker 220 7.14 Logistic regression analysis predicting no response 221 7.15 Logistic regression analysis predicting no response if Q-recipient is not gazing, by language 222 7.16 Q-speaker gaze, by language 222 7.17 Q-recipient gaze, by language 223 7.18 Questions in which lack of Q-speaker gaze toward Q-recipient motivated by question content or competing activity 224 7.19 Questions in which lack of Q-recipient gaze toward Q-speaker motivated by question content or competing activity 224 11.1 Transcription symbols 352 12.1 Distribution of “so”- and “oh”-prefaced sequence initiators388
Contributors
Galina B. Bolden, Department of Communication, Rutgers University, New Brunswick, N.J., USA Penelope Brown, Language Acquisition Group, Max Planck Institute for Psycholinguistics, Nijmegen, The Netherlands Maria Egbert, Department of Business Communication and Information Science, University of Southern Denmark, Denmark Steven Fincke, Nuance Communications, Cambridge, Mass., USA Barbara Fox, Department of Linguistics, University of Colorado, Boulder, Col., USA Andrea Golato, Department of Germanic Languages and Literatures, University of Illinois at Urbana-Champaign, Ill., USA Auli Hakulinen, Department of Finnish Language and Literature, University of Helsinki, Finland Makoto Hayashi, Department of East Asian Languages and Cultures, University of Illinois at Urbana-Champagne, Ill., USA Trine Heinemann, Department of Business Communication and Information Science, University of Southern Denmark, Denmark Wilfrido Flores Hernandez, Department of Applied Language Studies and Linguistics, University of Auckland, New Zealand Minna Laakso, Department of Speech Sciences, University of Helsinki, Finland Stephen C. Levinson, Language and Cognition Group, Max Planck Institute for Psycholinguistics, Nijmegen, The Netherlands
xiv
Contributors
Anna Lindström, School of Humanities, Education and Social Sciences, Örebro University, Sweden Jeffrey D. Robinson, Department of Communication, Portland State University, Oregon, USA Federico Rossano, Department of Development and Comparative Psychology, Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany Emanuel A. Schegloff, Department of Sociology, University of California, Los Angeles, Calif., USA Jack Sidnell, Department of Anthropology, University of Toronto, Canada Marja-Leena Sorjonen, Research Institute for the Languages of Finland, Finland Liang Tao, Department of Linguistics, Ohio University, Athens, Ohio, USA Fay Wouk, Department of Applied Language Studies and Linguistics, University of Auckland, New Zealand Ruey-Jiuan Regina Wu, Department of Linguistics and Asian/ Middle Eastern Languages, San Diego State University, Calif., USA Kyung-eun Yoon, Department of African and Asian Languages and Literatures, University of Florida, Gainesville, Fla., USA
Transcription conventions (from Schegloff 2000)
I. Temporal and sequential relationships Overlapping or simultaneous talk is indicated in a variety of ways. [
Separate left square brackets, one above the other on two successive lines with utterances by different speakers, indicates a point of overlap onset, whether at the start of an utterance or later.
] ]
Separate right square brackets, one above the other on two successive lines with utterances by different speakers indicates a point at which two overlapping utterances both end, where one ends while the other continues, or simultaneous moments in overlaps which continue.
=
Equal signs ordinarily come in pairs – one at the end of a line, and another at the start of the next line or one shortly thereafter. They are used to indicate two things:
(1) If the two lines connected by the equal signs are by the same speaker, then there was a single, continuous utterance with no break or pause, which was broken up in order to accommodate the placement of overlapping talk.
(2) If the lines connected by two equal signs are by different speakers, then the second followed the first with no discernable silence between them, or was “latched” to it.
A single equal sign indicates no break in an ongoing piece of talk, where one might otherwise expect it, e.g., after a completed sentence.
(0.5) Numbers in parentheses indicate silence, represented in tenths of a second; what is given here in the left margin indicates 0.5 seconds of silence. Silences may be marked either within an utterance or between utterances.
xvi (.)
Transcription conventions A dot in parentheses indicates a “micropause,” hearable but not readily measurable without instrumentation; ordinarily less than 0.2 of a second.
II. Aspects of speech delivery, including aspects of intonation The punctuation marks are not used grammatically, but to indicate intonation. The period indicates a falling, or final, intonation . contour, not necessarily the end of a sentence. Similarly, a ? question mark indicates rising intonation, not necessarily a , question, and a comma indicates “continuing” intonation, not ¿ necessarily a clause boundary. The inverted question mark is used to indicate a rise stronger than a comma but weaker than a question mark. ::
Colons are used to indicate the prolongation or stretching of the sound just preceding them. The more colons, the longer the stretching. On the other hand, graphically stretching a word on the page by inserting blank spaces between the letters does not necessarily indicate how it was pronounced; it is used to allow alignment with overlapping talk.
-
A hyphen after a word or part of a word indicates a cut-off or self-interruption, often done with a glottal or dental stop.
word
Underlining is used to indicate some form of stress or emphasis, either by increased loudness or higher pitch. The more underlining, the greater the emphasis.
word
Therefore, underlining sometimes is placed under the first letter or two of a word, rather than under the letters which are actually raised in pitch or volume.
WOrd
Especially loud talk may be indicated by upper case; again, the louder, the more letters in upper case. And in extreme cases, upper case may be underlined.
o
The degree sign indicates that the talk following it was markedly quiet or soft.
o
wordo
When there are two degree signs, the talk between them is markedly softer than the talk around it.
Combinations of underlining and colons are used to indicate intonation contours: _: If the letter(s) preceding a colon is/are underlined, then there is an “inflected” falling intonation contour on the vowel (you can hear the pitch turn downward). : If a colon is itself underlined, then there is an inflected rising intonation contour on the vowel (i.e., you can hear the pitch turn upward)
Transcription conventions
xvii
↑ ↓
The up and down arrows mark sharper rises or falls in pitch than would be indicated by combinations of colons and underlining, or they may mark a whole shift, or resetting, of the pitch register at which the talk is being produced.
> < < > <
The combination of “more than” and “less than” symbols indicates that the talk between them is compressed or rushed. Used in the reverse order, they can indicate that a stretch of talk is markedly slowed or drawn out. The “less than” symbol by itself indicates that the immediately following talk is “jump-started,”, i.e. sounds like it starts with a rush.
hhh (hh)
Hearable aspiration is shown where it occurs in the talk by the letter h – the more h’s, the more aspiration. The aspiration may represent breathing, laughter, etc. if it occurs inside the boundaries of a word, it may be enclosed in parentheses in order to set it apart from the sounds of the word. If the aspiration is an inhalation, it is shown with a dot before it.
.hh
III. Other markings (( ))
Double parentheses are used to mark transcriber’s descriptions of events, rather than representations of them: ((cough)), ((sniff)), ((telephone rings)), ((footsteps)), ((whispered)), ((pause)), and the like.
(word)
When all or part of an utterance is in parentheses, or the speaker identification is, this indicates uncertainty on the transcriber’s part, but represents a likely possibility.
(lit/bit)
Where alternate hearings are possible these are enclosed in parentheses and separated by a back slash.
( )
Empty parentheses indicate that something is being said, but no hearing (or, in some cases, speaker identification) can be achieved.
IV. Multi-linear transcription conventions Many of the transcriptions in this book are of talk in languages other than English and include one or two lines of glossing. In three-line transcripts the first line is a broad phonetic representation of the talk in the original language. The second line is a morpheme-by-morpheme gloss using a combination of word-for-word translation and abbreviations such as ASP to indicate particles and other functional items that do not admit of a direct translation into English. Authors who use these abbreviations in the transcripts they present include a key explaining them at the conclusion of the chapter. The third line presents an idiomatic English gloss – an attempt to get as close as possible to the contextual sense or meaning of the utterance. The following example from the chapter by Wu illustrates:
xviii
Transcription conventions 01 L: wo dou hai mei jian guo ta.
I
all still N see ASP he
‘I haven’t met him yet.”
In two line transcriptions the morpheme-by-morpheme gloss is not included. The following example is from the chapter by Sidnell:
01 C: him mada biilongz tuu Hamilton.
his mother is from Hamilton.
Part I
Introduction
1 Comparative perspectives in conversation analysis Jack Sidnell
Introduction Comparison of the various ways in which people talk in different sociocultural and linguistic communities can easily lead to two, apparently contradictory, conclusions. On the one hand, the diversity of conduct is striking. People speak different languages, they are oriented to markedly different sociocultural norms of posture and tone, they inhabit very different social (as well as economic, political, etc.) worlds and thus find occasion to talk in vastly different “contexts.” One need only compare, for instance, the Mayan villagers of Tenejapa with Yélî Dnye speaking Rossel Islanders to see this diversity (as Rossano, Brown and Levinson do in this volume). On the other hand, the commonalities are what are remarkable. Everywhere turns-at-talk are constructed and opportunities to speak distributed, courses of action are launched and co-ordinatively managed, troubles of speaking, hearing and understanding are located and their repair attempted. These commonalities suggest that, for all the diversity we see, people everywhere encounter the same sorts of organizational problems and make use of the same basic abilities in their solutions to them – a capacity for reading other’s intentions, anticipating and projecting actions, calculating inferences and processing information available to them (see Levinson 2006, Schegloff 2006). If these abilities and problems appear universal and generic, the particular ways in which they are implemented or solved is anything but. After all, whatever happens in interaction happens through the medium of some specific set of locally available semiotic resources. Over the past forty years, conversation analysts
4
Jack Sidnell
have shown that actions in talk-in-interaction are formed through the use of distinctive prosodic patterns, lexical collocations, wordorder patterns as well as language-specific objects such as English “oh” (Heritage 1984), Mandarin “ a” (Wu 2004), Finnish “ nii” (Sorjonen 2001) and so on (see Stivers and Sidnell 2005 for an overview). Of course, these semiotic resources vary significantly and systematically across different languages and communities. Some languages such as Vietnamese and Mandarin use a set of distinct tonal contours to signal differences of lexical meaning; others do not. Some languages are, like English, left-headed, whereas others, such as Korean and Japanese , are right-headed. In some languages, such as Vietnamese again, words are typically made up of one morpheme having a single meaning whereas in others words often consist of several morphemes some of which may embody multiple meanings (e.g., -ó in Spanish “habló” expresses indicative mode, third person, singular, past tense, and perfective aspect). Because every turn-at-talk is fashioned out of the linguistic resources of some particular language, the rich and enduring semiotic structures of language must be consequential in a basic way for social interaction. So, although the problems are generic and the abilities apparently universal , the actual forms that interaction takes are shaped by and adapted to the particular resources that are locally available for their expression. The chapters in this book contain detailed analyses of talk in English, Finnish , German , Italian , Japanese , Mandarin , Tzeltal (Mayan ), Russian , Swedish and Yélî Dnye and provide the beginnings of an answer to the question of how the local resources of these particular languages and other semiotic systems shape, constrain, torque or inflect the otherwise generic and universal underlying organization of talkin-interaction. In recent work, both Schegloff (1996c) and Fox (2007) have argued that over time the organizational contingencies of interaction significantly shape (though do not determine) grammar. This book offers a complementary view in which, at some given point in time, the semiotic resources of any particular language – especially grammar – essentially define the possibilities for social action accomplished through talk. Notice, that this way of putting things largely obviates the theoretical syntactician’s appeal to a “deep” level of linguistic structure or “logical form” at which such “surface” differences between
Introduction
5
languages more or less completely evaporate. For social interaction does not work with such “underlying” mental representations; it works with the actual, observable surfaces of grammar and prosody as well as the particular words and phrases that constitute possibly complete turn units and recognizable actions. As we will see in this chapter, it is this surface patterning of language that has consequences for turn-taking , repair and other aspects of the organization of talk-in-interaction. The linguistics that forms a natural companion to conversation analytic studies is thus the descriptive and empirical tradition that investigates not mental models but rather actual living languages in all their peculiarity and nuance often from a comparative perspective.1 In fact, recent work in interactional linguistics has emphasized and shown the value of just this connection (see Ford et al. 2002 for an overview). Some of the complexity of the relationship between (1) the specific semiotic resources of a particular language, (2) language and culture-independent principles, and (3) the contexted courses of action within which language is actually embedded can be seen through a consideration of repetition . Turns designed to show that they are repeats of something that someone has just said are apparently found in all communities and in all languages. But now consider that, strictly speaking, “unmarked” repetition of what someone has just said is interactionally impossible. That is to say, one cannot simply and solely repeat what someone has just said without adding something to it. While it’s possible for a second speaker to do a repeat in such as way as to make it hearable as saying “exactly” what another just said (for example, by preserving not only the words another speaker has used but also the original deictic forms as well the intonation and other aspects of prosody ), to do so is to do something in addition to repetition. Repetition of this kind is hearable as a verbatim repeat and thus possibly as “mimicry.” This points to the fact then that even in repeating what another has said, the speaker necessarily draws on the semiotic resources of some particular language to mark, and thus to frame, the repeat in some way. 2 Typically then, a repeat involves the use of some intonational or periphrastic marking which shows to what end this bit of talk is being repeated. One very common thing is for the repeat speaker to produce his or her talk with what is typically known as “question
6
Jack Sidnell
intonation” as in the following case from an English telephone call in which Ben has called for someone he refers to first as “Mary” and subsequently, after Ann indicates that she does not recognize anyone by that name, as “the tax lady” (Jefferson 1972, Schegloff et al. 1977). (1) XTR1 01 ((click)) 02 Ann: Hello:¿ 03 Ben: hHello, Ma:ry? 04 (0.2) 05 Ann: No: 06 (0.3) 07 Ben: No, not Ma:ry?= 08 Ann: =No, it’s not Ma:ry. There’s no Mary he:re 09 (.) 10 I don’ think:hh 11 (3.0) 12 Ben: The tax lady: 13 (0.2) 14 Ann: → The tax lad(h)y::? 15 Ben: Ya hhh= 16 Ann: =Nah. (.) Wha number were you callin’.
Another common intonation-marking for repeats in English involves stressing the first syllable of a multi-syllabic word. In (2) below Ann and Bev are discussing the diet and regime of a woman who is pregnant. (2) YYZ_T1A_A&D 2.03 01 Bev: anyway [.hhhh [I know. 02 Ann: 03 (0.2) 04 Ann: an’ i- (.) gra:vol an all these (s[ ) 05 Bev: → [gr↑a:vo:l 06 Ann: She takes gravol al- everyda:y.
Notice that whereas in Extract 1 the repeat elicits from Ben a confirmation (“Ya” at line 15) in Extract 2 it does not: instead, Bev continues with the telling. This suggests that in Extract 2 Bev uses the repeat to do something other than check her hearing of “gravol.” While these forms of marking appear to be quite common across a wide range of languages, more obviously language-specific
Introduction
7
resources are used in repeats too. Stivers (2005) for instance has described one kind of repetition in English that involves, among other things, expanding the pronoun + verb contraction of the original utterance (It’s → It is). Consider the following case from a conversation between co-workers. (3) YYZ 27 Clare: Just we’ve got- Michael and I did the resourcing for next 28 → week and it’s just- it’s=just ughu(h)h(hh) .hh and it’s a 29 short wee:k so 30 (.) 31 Alice: → Yeah:- Oh:yeah it is a short week= 32 Clare: Ye[ah 33 Alice: [.hh[hh 34 Clare: [so (.) I’m rilly rilly sorry,
Indeed, for English, next-turn repeats take a great variety of forms and are deployed in a wide range of actions such as confirmation (Schegloff 1996b), other-initiation of repair (Schegloff et al. 1977), correction (Schegloff et al. 1977; Goodwin 1984), displays of appreciation (Jefferson 1972) and agreement (Pomerantz 1984). Coupled with the particular semiotic resources of English grammar, prosody and lexis, the apparently universal practice of repetition is employed to effect a range of interactional outcomes. A number of the chapters in this book consider practices that involve repetition including Hayashi and Yoon on Japanese and Korean , Sorjonen and Hakulinen on Finnish , Wu on Mandarin , Heinemann on Danish , Bolden on Russian . These studies provide further evidence of the ways in which an apparently universal aspect of interactional organization is shaped and, in some ways, constrained by the semiotic resources of a given language thus providing unique possibilities for social action. 3 A comparative perspective in conversation analysis The chapters collected in this book examine the organization of social interaction and, more specifically, talk-in-interaction, from a comparative, cross-linguistic perspective. Although some of them draw at points on a number of other traditions (particularly functional/interactional linguistics ), the analyses they present are largely
8
Jack Sidnell
developed within the framework of conversation analysis (hereafter CA) – an approach to talk and social interaction which emerged in the writings and lectures of Harvey Sacks , Emanuel Schegloff and Gail Jefferson .4 CA takes not language per se as its focus but rather the practical activities in which language (along with gesture , gaze and other aspects of bodily comportment) is deployed, that is, talk-in-interaction. However, comparison of such practical activities across different languages brings the relevance of grammatical (and other) structures into view as the chapters in this book demonstrate. In this chapter, I sketch some of the conceptual background of comparative studies in CA. I begin with a consideration of previous studies distinguishing the approach taken within CA from that known as cross-cultural pragmatics. I then move to consider the CA methodology. Here I note a similarity between structuralist or distributionalist methods in linguistics and the CA approach of using collections. I argue that the former, but not the latter, essentially undermine the very possibility of comparison. Structuralism, taken to its logical conclusion, reveals the particularity of any system (e.g., of pronouns, or verbal tenses) and the elements of which it is composed. Since each element is defined by its relation to all the others, each element is a unique outcome of the particular system in which it is embedded. The result, as is well known, are accounts which are wholly hermetic and incapable of being compared to one another. In contrast, CA focuses on generic interactional problems which find solutions in the local resources of particular languages and social systems. For example, turn-taking and repair address interactional problems which must be solved in any community if its members are to coordinate their actions through talk-in-interaction. This approach actually encourages comparison: by looking across different languages/communities we can see the way in which the same interactional problem is solved through the mobilization of different resources (Schegloff 2006). Conversation analysts are keenly aware of the method ological and theoretical issues raised by a comparative perspective. In an early co-authored paper, Schegloff and Sacks (1974: 234–235) wrote: The materials with which we have worked are audiotapes and transcripts of naturally occurring interactions […] with differing numbers of participants and different combinations of participant attributes. There is a
Introduction
9
danger attending this way of characterizing our materials, namely, that we be heard as proposing the assured relevance of numbers, attributes of participants, etc., to the way the data are produced, interpreted, or analyzed by investigators or by the participants themselves. Such a view carries considerable plausibility, but for precisely that reason it should be treated with extreme caution, and be introduced only where warrant can be offered for the relevance of such characterizations of the data from the data themselves. […] The considerations just adduced […] restrain us from further characterizing it here.
And the authors go on: For example, they restrain us from characterizing our findings as relating to “some general features of conversation rules in American English” – a suggestion offered by Dell Hymes (personal communication), for that suggests implicitly an ethnic or national or language identification as a relevant putative boundary for both our materials and findings. We cannot offer a warrant for asserting such a boundary and we suspect others cannot either.
The passage sums up quite succinctly the distinctive position which conversation analysts have adopted toward what are referred to in this passage as “participant attributes.” Whereas other approaches such as linguistic anthropology and sociolinguistics typically treat participant attributes as self-evident features of “context” or, alternatively, as “external variables” to be correlated with aspects of speech behavior, CA shifts the burden of evidence by requiring researchers to show that putative “characteristics” of the participants (such as race, ethnicity, gender) have some “procedural consequentiality” or demonstrable relevance for the participants themselves in terms of the specific ways in which the interaction is organized.5 In the passage quoted above, Schegloff and Sacks characterized this as a reluctance to invoke a boundary where none seemed to be warranted by the data being examined. Importantly, Schegloff and Sacks were not denying the existence of differences either between groups within a society or between societies. Instead, what they were pointing to were the method ological and theoretical issues involved in making claims about the relevance of such differences for the organization of interaction. Within CA then, researchers have emphasized the various ways in which “participant attributes” are picked out and made relevant to the talk of the moment (see for instance Goodwin 1987; Schegloff 1991, 1992b; Sacks 1995).
10
Jack Sidnell
There were deep underlying issues at stake here – ones that Sacks had initially addressed in his work on membership categorization devices and that Moerman (1974, for instance) had applied to problems of ethnic categorization in anthropology . As Goodwin puts it, “showing that a category can be accurately applied to a participant does not demonstrate either that the participants themselves are dealing with each other in terms of such a category, or that it is in fact a relevant feature of the activity being studied” (1987: 120). A speaker (or recipient) may be Catholic, six feet tall, male, a dentist, a vegetarian or whatever else without this being in any way relevant to the way he talks on some particular occasion. Sacks noted in his lectures that all societies appear to divide people into groups based on at least two axes of difference: gender and age (Schegloff 2002c). It was not an interest in universals of social organization that prompted Sacks ’ noticing of this. Rather, the existence of at least two axes of difference meant that any given individual could be described or categorized in more than one way. As such, any description of a participant (as a male, a child, etc.) required some “warrant” or motivation, some evidence that this and not some alternative was the relevant descriptive category. It is not difficult to see how these considerations bear on the problems of comparison to which the chapters in this book are addressed. How can we, for instance, warrant a description of the data as “requests in Polish,” “French compliments” or whatever where, typically, the participants do not display any overt orientation to the relevance of the language being used or their ethnicity?6 This is a particularly difficult problem and this book does not offer a single, conclusive solution to it. Rather, different authors take somewhat different approaches. One possible solution is discussed in this introduction: this involves linking the practices of talk-in-interaction to particular, distinctive aspects of the language being spoken thus warranting a characterization of those practices in terms of the language employed. The idea then is that particular languages provide specific resources and also establish unique constraints for the organization of interaction. Grammar, culture, and the organization of repair A few years after the Schegloff and Sacks paper cited above was published, Moerman (1977) addressed the issue of comparison
Introduction
11
in a study of repair in Thai -Lue materials. If differences could be located between the Thai-Lue and English materials, these might be related to differences in language structure, social arrangement or, perhaps, culture. However, in the main, the study did not reveal significant differences in the organization of repair and Moerman (1977: 875) was led to conclude: “the detailed, systemic, and massive parallels between” Thai and American English corpora “support the claim that the domain described by Sacks , Schegloff and Jefferson is conversation – without respect to the language, nation, class, or culture in which it occurs.” In contrast to work in both cross-cultural pragmatics and linguistic anthropology, early comparative research in CA provided evidence of massive, underlying structural similarities in interaction from very different communities. Whereas both crosscultural pragmatics and linguistic anthropology tended to work from the assumption of significant differences between communities/cultures, conversation analysts began with a focus on generic interactional problems which had to be, in some way, solved in any community. For instance, the phenomena to which a conversational repair mechanism is addressed – misunderstandings, misspeakings and the like – are universal and ubiquitous. If participants are to engage in coordinated action they will necessarily require some mechanism for dealing with these issues. At a finer level of detail, misunderstandings and misspeakings seem to take similar forms in every community. Misunderstandings arise because the recipient does not know a word the speaker has used or did not hear it adequately. Misspeakings arise when a speaker uses the wrong word or cannot find the one they need and so on. Given that everywhere the issues are more or less the same, it is hardly surprising to find that the solutions developed for dealing with them are also strikingly similar. This is not, however, to say that there are no differences. Here, I will briefly note a few of these staying with the problem of repair; below (see “Mobilization of Local Resources to Solve Generic Problems of Interaction” ) I discuss some claims for profound differences in turn-organization and turn-taking between Japanese and English. In a comparison of repair in English and German, Egbert (1996: 587) argued that “repair is sensitive to the linguistic inventory of a given language” in several different ways. First, Egbert showed that grammatical markings present in German but not English (such
12
Jack Sidnell
as articles marked for gender, i.e. “das” and “der”) were subject to repair. Second, she observed that case-marking in German can serve as a resource for framing, and thus locating, the repairable item. Third, she showed that, in other-initiated repair, grammatical congruities across turns, the marking of an item as masculine singular accusative for instance, may serve as a resource for linking repairable , repair initiator and the repair itself. In a study which I discuss in greater detail below, Fox et al. (1996) noted three differences between self-repair in English and Japanese and link these to the very different “syntactic practices” of the two languages. First, similar to Egbert ’s finding for German, Fox et al. find that certain kinds of morphological repair are possible in Japanese but not English. They illustrate with the following example: (4) K: ja nanji goro ni kurida[shi-*]soo? Then what.time about OBL go.out “Then about what time (shall we) go out?”
The authors (Fox et al. 1996: 202) observe that in this example the “speaker K replaces only the inflection al ending of the verb with another. […] The first form that K produces in this example (kurida-shi) has the adverbial ending -shi. […] However K cuts off the verb and replaces -shi with the ‘cohortative’ ending -soo. In other words, K has replaced one bound morpheme with another.” In English, repair does not operate on inflection al endings. This difference can likely be accounted for in terms of verbal morphology in the two languages. The Japanese verb endings -shi and -soo are full syllables consisting of a consonant and a vowel. In contrast, English verb endings (with the exception of -ing and -s in certain contexts) are, typically, pronounced as a single consonant sound. They are thus unpronounceable independently of the word to which they are attached. This prevents their production as independent units in self-repair as in the following invented example (from Fox et al. 1996: 202). (5) She look[ed]-*s at the table.
Moreover, as the authors point out, Japanese is an agglutinating language in which each morpheme has roughly a single grammatical meaning. In contrast, English is characterized by a relatively
Introduction
13
greater degree of fusion in inflection al morphology (Comrie 1981). So, for instance, verbal -s carries the meaning of present, third person and singular. Bound verbal morphemes are thus semantically complex in English but not in Japanese . English verbal morphemes (or at least verbal -s) may also indicate agreement with the subject (third person) of the clause and thus link back to something earlier in the utterance where as Japanese verbal markers do not. The authors (Fox et al. 1996: 203) conclude that “these […] differences between English and Japanese verb endings suggest to us that at a variety of levels verb endings in English are more tightly ‘bonded’ to the verb than are verb endings in Japanese and hence less available for individual replacement than are verb endings in Japanese.” Another difference noted by these authors involves practices for delaying the production “of next noun” due in self-repair . The following illustrates one common practice in English: (6) on the back of his pickup truck [with a,*] (0.4) with a jack.
Here the speaker begins a prepositional phrase, initiates repair and then recycles the preposition and article before then producing the rest of the phrase. Recycling the preposition (and article) thus “constitutes a procedure for delaying the production of the next item due. This procedure could, for example, be part of a word search , a request for recipient gaze, management of overlapping talk, and/or production of a dispreferred” (Fox et al. 1996: 204). Since Japanese has postpositions rather than prepositions (and no articles), this procedure for delay is not available. Fox et al. show that Japanese speakers often use a quite different procedure to effect the same result: employing a demonstrative pronoun (translated as “that”) as a “place holder while the speaker looks for a lexically specific noun” (Fox et al. 1996: 206). The studies of Egbert and Fox et al. thus link language-specific practices of repair to particular grammatical features of the language in question. Some of the variations in repair observed in the literature are accounted for not by reference to the grammar of the language but rather to the beliefs of the speakers. Ochs (1984: 328–329) writes of the type of other-initiated repair she terms “clarification ’: While clarification is a universal activity, the manner in which clarification is accomplished varies cross-cultural ly. Prefer ences for
14
Jack Sidnell
accomplishing clarification are embedded in local principles of social order and local epistemologies. More specifically, I would like to suggest that both the conditions under which clarification takes place (what gets clarified, who participates in the activity of clarification, in which roles), and the discourse procedures speakers prefer to use, index members’ views of knowledge, particularly members’ views on the limits of knowledge (what can be known) and the paths to knowledge (how knowledge is acquired).
Ochs goes on to contrast two basic practices in clarification – a “minimal grasp strategy” in which the one initiating clarification employs quizzical expressions, statements of non-understanding, WH-questions and other directives “to elicit from the child a reformulation of all or part of the unclear utterance or gesture”, and an “expressed guess strategy” in which it is the one initiating clarification “who attempts a reformulation of the unclear act” (1984: 331). Ochs argues that the more or less total absence of explicit guessing in the clarification practices of Western Samoans in favor of a minimal grasp strategy is linked to, among other things, local conceptions of knowledge: “ Samoans generally display a strong disprefer ence for guessing at what is going on in another person’s mind. This dispreference has reflexes in a range of verbal activities,” including those of clarification (Ochs 1984: 337, see also Ochs 1991). CA and Cross-Cultural Pragmatics I want to briefly discuss here the way in which the approach taken in these chapters differs from that commonly known as crosscultural pragmatics. One major set of differences is method ological. Work done under the rubric of cross-cultural pragmatics typically employs some kind of questionnaire method in which respondents are asked to select the most appropriate or least appropriate response to some described situation. Alternatively respondents may be required to engage in a “discourse completion task.” In contrast, CA employs a method ology of observation and collection of naturalistic interactional data (see below). At an even more fundamental level, these two approaches exhibit quite different conceptualization of the object of study. Cross-cultural pragmatics tends to focus on the different ways in which actions such as
Introduction
15
“promises,” “requests,” “invitations,” “refusals” are managed and understood by recipients, often in terms of their implications for “face” and notions of politeness (see for instance Blum-Kulka et al. 1989). In other words, actions such as “requests,” “invitations” and the like are treated as though they were objective categories of social action providing a template through which one can view interactional data. CA adopts a different perspective. As Schegloff observed in an early paper (1984: 30): “it is misleading to start to account for such categories of action as questions, promises, and so on as the analytic objects of interest. They are commonsense, not technical, categories and should be treated accordingly.” Rather than begin with such “common sense categories of action” as promises and such, CA begins with relatively stable organizations of practice in talk-in-interaction, e.g., answering the phone with “hello?,” referring to persons by name or some alternative recognitional form (see Sacks and Schegloff 1979; Schegloff 1996a; Stivers 2007), multiple sayings such as “no no no” (Stivers 2004), etc.7 The practices of repair discussed in the previous section provide another example. Notice that with a format such as (partial repeat + WH word), a speaker may accomplish a wide range of actions such as expressing surprise in response to a news report (“They’re going where?”), challenging a previous speakers version of events (“You did what?”), requesting clarification (“ Move what”), etc. (see Schegloff 1997b). The contrast here between CA and cross-cultural pragmatics is a decisive one. A focus on practices of speaking, rather than “action-types,” suggests that whatever action a given utterance (or part of an utterance) is understood by the participants to be doing is a contingent outcome. Participants understand some practice of speaking as doing a particular action by considering it in relation to the local sequential context in which it is produced (i.e. after a recognizable news report, an excuse, a suggestion, etc.). In that sense, categories of action are themselves the product of local interpretive work. This perhaps accounts for a basic problem within cross-cultural pragmatics of identifying comparable “actions” in the first place. That is, identifying invitations in Russian , Japanese, Samoan, and Greek so that they can then be compared necessarily involves a reification – a treatment of
16
Jack Sidnell
utterances as if each transparently conveyed a unique “illocutionary force.” Practices, structures, distributions Bloomfield (e.g., 1946, 1970) is often chided today for his distributional method in which one accounts for the function of a given item by looking, one might say, dispassionately, – that is, without regard to its apparent “meaning” – at the positions in which it can be found to occur and the various alternatives for which it may substitute. This basic conceptual framework behind structuralism , in which relations in presentia and in absentia are more important than links between the sign and the object (designatum), has broad theoretical and method ological consequences. Perhaps most importantly, it offers an arguably more objective view than an impressionistic sense of what any given instance of a phenomenon “means” or is “doing.” I do not think it does CA any injustice to say that it employs a, broadly speaking, distributional method – though it is one tempered by detailed attention to each particular case. Analysis is based on a collection of instances. Such a collection not only allows for the identification of an unmarked or default pattern (for discussion of this usage see Stivers , Enfield and Levinson 2007; Stivers 2007; Enfield 2007), it also reveals the operation of underlying prefer ences or principles. Collections allow also for the identification of “deviant” instances in which the practice does not (for whatever reason) follow its normal course – a “deviance” to which participants themselves orient. Such deviant cases frequently reveal participants’ own orientations to the various contingencies involved in the execution of that practice – orientations which are typically unobservable in cases which conform to normative expectations and which are therefore literally unremarkable. In the context of an examination of practices for referring to persons, for instance, a collection of instances allows the analyst to distinguish cases in which just reference is being done from those in which something in addition to this is being accomplished. Consider, for instance, a case in which a spouse returns home from work and is met immediately with, “What did she say?”. Use of a locally subsequent form (the anaphor “she”) in a
Introduction
17
locally initial position (i.e. without any previous use of a name) can bring off that this has “been on my mind throughout the interim,” that this is, in effect, “a continuation of the earlier conversation” (Schegloff 1996a: 451). Referring to oneself by name (“Jack doesn’t like that!”) is another non-default usage that can be used to convey footing shifts of various kinds (see Schegloff 1996a). It is worth noting the clear resonances with structuralism here. Consider Schegloff ’s analysis of continuers such as “uh huh” presented in his paper “Discourse As Interactional Achievement: Some Uses of ‘Uh Huh’ and Other Things That Come Between Sentences” (Schegloff 1982). In characterizing “discourse as an interactional achievement,” Schegloff draws attention to the fact that a spate of talk composed of more than one “sentence” (or turn-constructional unit, etc.) is the product of methodic work by speaker and recipient(s) since it necessarily involves the negotiation of places at which transfer to a next speaker might have otherwise been effected. One key resource in such an accomplishment are vocalizations such as “uh huh,” “mm hmm,” “yeah” and the like typically labeled “backchannels” in the literature. Schegloff (1982: 78) notes that in the literature, two related characterizations have been offered to deal with these bits of behavior. According to one, they are evidence of attention, interest, and/or understanding on the listener’s part. (Thus Fries 1952: 49, “signals of this continued attention,” or Kendon 1967: 44, “appears to do no more than signal […] that he is attending and following what is being said”). A second use of such behavior proposed in this literature is that it “keeps the conversation going smoothly” (Dittman and Llewellyn 1967: 342), or “appears to provide the auditor with a means for participating actively in the conversation, thus facilitating the general coordination of action by both participants” (Duncan and Fiske 1977: 202–203). After noting that while a coherent set of cases can be assembled in which participants use “uh huh” and other such objects to deal with issues of attention (cases in which, e.g., “attention was indeed problematic for the parties” (Schegloff 1982: 79) or in which the “uh huh” responds to the use of a try-marked recognitional person reference), Schegloff goes on to point out a number of problems with such characterizations. They are, in the end, of course largely
18
Jack Sidnell
impressionistic. Schegloff (1982: 81) offers an alternate account, arguing that: perhaps the most common usage of “uh huh,” etc. is to exhibit on the part of its producer an understanding that an extended unit of talk is underway by another, and that it is not yet, or may not yet be (even ought not yet be), complete. It takes the stance that the speaker of that extended unit should continue talking, and in that continued talking should continue that extended unit. “Uh huh,” etc. exhibit this understanding, and take this stance, precisely by passing an opportunity to produce a full turn at talk. When so used, utterances such as “uh huh” may properly be termed “continuers.”
Within such an account, the notion of signaling attention takes on a different cast. Points of possible completion in a continuing speaker’s talk are structurally relevant positions in which the recipient may properly display their understanding of the current state of the talk and “show their intention to pass the opportunity to take a turn-at-talk that they might otherwise initiate at that point” (Schegloff 1982: 81). So, with “uh huh” etc., a recipient passes the opportunity to do any sort of fuller turn. The analysis begins then with a specification of the particular sequential environment in which “uh huh” and other such tokens are deployed. Schegloff then goes on to consider what else can occur in that same position noting that were a fuller turn done it would necessarily be a turn of some particular type and would thus be doing some specific action or actions. Thus, the opportunity being passed up by use of “uh huh” is an opportunity “to do whatever might have relevantly been done at that point” (1982: 87). Schegloff (1982: 87) asks “are there any kinds of actions which have some kind of ‘general relevance’ in conversation,” actions “not made relevant by the particulars of someone’s immediately preceding talk or behavior?” The obvious candidate is other-initiated repair which Sacks described in his first transcribed lecture as an “occasionally usable device” meaning that “there doesn’t have to be a particular sort of thing preceding it; it can come at any place in a conversation” ( 1995: I, 6). As Schegloff (1982: 87) puts it, “it appears that there are no systematic exclusion rules on the possible relevance of nextturn repair initiation in any possible turn position.” “Uh huh,” etc., in passing the opportunity to do a full turn at talk, “can be seen to be passing an opportunity to initiate repair on the immediately preceding talk” thus providing an obvious basis for “the
Introduction
19
ordinary inference that the talk into which they are interpolated is being understood” (Schegloff 1982: 88). Schegloff (1982: 88) concludes: It is not that there is a direct semantic convention in which “uh huh” equals a claim or signal of understanding. It is rather that devices are available for the repair of problems of understanding the prior talk, and the passing up of those opportunities, which “uh huh” can do, is taken as betokening the absence of such problems.8
Schegloff ’s alternative to the impressionistic account in which “uh huh,” etc. are treated as “displays of attention”9 involves identifying a position, the transition relevance place constituted by a turn’s possible completion, and examining the particular object in relation to the other objects which might have occurred there – and which participants themselves can see could have been done in that position. The result is an interactionally generated, positionally sensitive “paradigm”. 1. Continuer (e.g., “uh huh”). 2. Full-turn (whatever might have otherwise been done). 3. Other-initiation of repair. The sense of “uh huh” (as, variously, conveying attention, understanding, agreement ) is a product of its place within a positionally sensitive yet highly restricted set of alternatives – relations in absentia. Though CA draws, as I hope to have shown, in a basic way on a structural or distributional method , it does not fall into the trap of structuralist analysis which, taken to its logical conclusion, reveals the particularity of any system (e.g., of pronouns or verbal tenses) and the elements of which it is composed. Since each element is defined by its relation to all the others, each element is a unique outcome of the particular system in which it is embedded. The result are accounts that are wholly hermetic and incapable of being compared to one another. The item in the phonemic inventory designated as /p/ for a given language will necessarily be unique to that system of distinctions. The consequences of a strict-structuralism have been noted with respect to diachronic comparision – that is, comparisons of the “same language” at different points in time. As Harris and Taylor
20
Jack Sidnell
(1997) note, if the consequences of Saussure’s structuralism are accepted, the comparative assumptions of historical linguists , who happily trace the evolution of consonants and vowels from IndoEuropean to the present, are undermined. Accepting structuralist premises, it makes no sense to suppose that the earlier b was the “same consonant” as its later manifestation p, or that “the same word” appeared in Latin as causa but later in French as chose. For if linguistic units did not exist except as structural units defined within a single linguistic system, it was impossible for any given unit to “survive” from one system A into a different system B at a later point in time. (Harris and Taylor 1997: 215)
Even where a consonant (e.g., m) appears to have survived and maintained its phonetic character from Latin to French because at each stage there was a slot that could be filled by a consonant having those phonetic qualities, we should not suppose that it is the same: “For in spite of this apparent identity of phonetic substance, the consonant in question. […] enters into a different set of contrastive relations with other consonants in Latin from those which characterize its position in the French system” (Harris and Taylor 1997: 216–217). CA avoids this problem of particularism through its focus on generic problems of interaction. For example, Schegloff ’s focus in the argument summarized above is the generic problem of coordination and turn-distribution – a problem that has to be solved within any community. We can look across communities and languages to see how local resources are mobilized to solve the recurrent, generic problems of interaction within human groups. Mobilization of local resources to solve generic problems of interaction Consider the following example (Sacks et al. 1974: 721): (7) [T. Labov: Battersea:B:1] 01 Tourist: Has the park cha:nged much, Parky: Oh:: ye:s, 02 03 (1.0) 04 Old man: Th’Funfair changed it’n [ahful lot [didn’it. Parky: [Th- [That05 06 Parky: That changed it,
Introduction
21
At line 05 of this example Parky begins to speak and finds himself in overlap with Old Man. Where does Parky begin to speak? Not after “The,” not after “Fu” or “Fun,” not after “funfair.” Rather, the beginning of Parky’s turn is designed to coincide with the end of “it,” so, after the “The funfair changed it.” Parky has then targeted the first point of possible completion in Old Man’s turn as a place to begin speaking. Notice that Parky starts up here and again at the next two points of possible completion (after “lot” and “it”), not by virtue of any silence (by the time he starts there is no hearable silence) but by virtue of the possible completion of the turn which he has anticipated. The coordination of speakership – i.e. turn-taking – with minimal gap and overlap thus seems to involve, in some kind of essential way, the participants’ ability to project the possible course of the current turn and to thus anticipate when it will reach a point of possible completion. In their 1974 paper, Sacks et al. (1974: 702) write: There are various unit-types with which a speaker may set out to construct a turn. Unit-types for English include sentential, clausal, phrasal, and lexical constructions. Instances of the unit-types so usable allow a projection of the unit-type under way, and what, roughly, it will take for an instance of that unit-type to be completed. Unit-types lacking the feature of projectability may not be usable in the same way. [Emphasis added]
In this way, they link the notion of turn constructional units – the building blocks out of which turns-at-talk are composed – to a feature of “projectability.” In a paper originally delivered in 1973, Schegloff (1987c: 71) expands on the notion of projectability, writing: One important feature of turn construction […] and the units that turn construction employs (e.g., lexical, phrasal, clausal, sentential constructions) is that they project, from their beginnings, aspects of their planned shape and type. […] There are other sorts of projection that are, or can be involved from the very beginning of a turn. For example, question projection. […] Or: beginnings can project “quotation formats.” Or: a beginning like “I don’t think” can project, in certain sequential environments, “disagreement” as a turn type for its turn. […] Again: turn beginnings are important because they are an important place for turn projection, and, given the importance of turn projection for turn taking, they are important structural places in conversation.
Recent work on Japanese (beginning with Fox et al. 1996) suggests that the grammatical properties that make such “early
22
Jack Sidnell
syntactic projections” possible in English may not be shared by all languages. Fox et al. (1996) and later Hayashi (2003) argue that various “syntactic practices” of Japanese result in a “limited projectability.” These syntactic practices include: 1. a flexible word order particularly at the beginning of the turn; 2. the use of post- rather than prepositions; 3. a prevalence of unexpressed constituents (including subject and object); 4. the use of final particles capable of retroactively transforming the constituents to which they attach. The authors (Fox et al. 1996: 208) argue that: the kind of clausal TCU structure these facts commonly lead to typically […] starts with some kind of discourse marker (e.g., ano, nanka), followed by adverbials, or nouns indicating setting of some kind, followed by the verb, and possibly followed by so-called final particles. So what occurs early in the TCU is often only loosely associated structurally with what is to follow. Conversational utterances in Japanese thus seem not to show tight syntactic organization.
Whereas in English “the subject begins a tightly knit clause structure and hence syntactically is the ‘beginning’ of the clause,” in Japanese “ these is no constituent element that serves as the beginning of a tightly knit syntactic unit – in fact, there is no such tightly knit unit” (Fox et al. 1996: 209). These syntactic practices then mean that turn-taking must be managed in a different way in Japanese than it is in English. The authors (Fox et al. 1996: 213) summarize: English recipients are able to use the beginning of a TCU to project a possible course for that utterance, while Japanese recipients “wait and see” how the utterance develops. It is thus possible that English recipients are able to predict with such accuracy how the utterance-under-construction will come to an end that they are able to plan their own utterance to start up exactly at the moment the current utterance comes to a possible completion point (this is, in fact, the claim made by Sacks et al. (1974) with regard to turn-taking in English conversation), with no pause between the end of the current turn and the start-up of their own turn. In Japanese , on the other hand, since recipients are not able to make such detailed predictions about the course of the current utterance, they wait until they hear the last few syllables of the turn (which often contain such “ending
Introduction
23
signals” as final particles, or special completion-relevant verb forms) before starting their own utterance.
The general interactional consequences of these various syntactic practices then include the adoption of a “wait and see”-type approach to turn-taking (and specifically self-selection), limited mid-turn projectability of turns-in-progress and an “incremental” or “bit-by-bit” approach to turn construction (Tanaka 2000; Hayashi 1999, 2003). The consequences are manifested in an absence of clause-level recycling in self-repair (Fox et al. 1996), various kinds of delay in co-participant completions including a prefer ence for terminal item completions (Hayashi 2003), an orientation to the end of the predicate component as transition relevant and a vulnerability of post-predicate components to overlap (Tanaka 2000: 7–8). Japanese grammar may then help to account for the practices of aizuchi (Miller 1983; LoCastro 1987). As Schegloff et al. (1996: 32) remark: The interpolations which in English conversation ordinarily come at the boundaries of larger chunks of extended turns understood to be not yet complete (Schegloff 1982) are produced – and solicited – for smaller chunks of utterance in Japanese , in part to offset otherwise potentially problematic indeterminacies built into Japanese grammar, indeterminacies which are problematic precisely because of the exigencies of recipient parsing in real time.
As Schegloff (2000c) notes, “the orderly distribution of opportunities to participate in social interaction is one of the most fundamental preconditions for viable social organization.” In other words, for human groups, turn-taking is a generic problem of interaction. Like any other such problem, this one must be solved through the mobilization of available local resources . With respect to turn-taking, one particularly important resource appears to be grammar (or “syntactic practices”). In both English and Japanese, the turn-taking problem gets solved but differently in each case due to the availability of different grammatical resources. And notice then that differences in the organization of turn-taking appear to have consequences for other practices of talk-in-interaction – specifically the deployment of so called “response tokens” (See Hayashi and Yoon, this volume).
24
Jack Sidnell
Overview of the volume The chapters in this volume are divided into three thematically organized sections. In conclusion here I want to briefly point to some of the issues involved in creating such groupings in the first place. The chapters grouped together in the first section of the book share a focus on repair. Each involves a detailed consideration of the various ways in which particular kinds of interactional trouble are addressed in conversation. At the same time, the reader will find that these chapters deal with a wide range of quite varied phenomena. For instance, Fox et al.’s analysis of self-repair initiation in seven typological and genetically distinct languages leads to a consideration of the influence of word length and syllable structure. The examination by Egbert et al. of a species of other-initiation leads to a consideration of prosody and pre-announcements . And though Wu ’s chapter begins with other-initiated repair, it comes eventually to an examination of epistemic issues implicated in talk-in-interaction. The point then is that within a given single domain a very wide range of practices are implicated. Indeed, it is quite impossible to deal solely with the phenomenon of repair, for instance, without at the same time considering the organization of turns and the sequencing of actions. So, while the chapters in each section address particular domains of talk-in-interaction, they do not attempt to reduce the phenomena they examine to that domain. The second section of the book considers responses, and, again, although there is clear thematic continuity across the chapters, each raises distinct issues and locates a particular dimension for comparative analysis. Here the prefer ences for contiguity and agreement described initially by Sacks (1987) are particularly important as are issues of epistemic access. The chapters show that although there is much in common across languages in terms of the resources available for constructing responses, the particular grammatical and lexical characteristics of these languages also result in considerable variability. Hayashi and Yoon compare not only different languages (English, Japanese, Korean) but also different approaches to social interaction. The authors conclude their chapter with a discussion of the ways in which the phenomenon they
Introduction
25
consider in Japanese and Korean has been analyzed in previous studies as an expression of cultural themes (e.g., “mutual dependency”), before offering an alternate account that emphasizes the particular syntactic practices and grammatical characteristics of these languages. Taken together the chapters in this section show that while responding clearly presents similar problems everywhere, interactional solutions are adapted and shaped by the available resources of particular languages. The chapters of the final section maintain a focus on languagespecific resources in the construction and sequencing of action investigating the use of language-specific resources to respond to assessments (Sorjonen and Hakulinen, Sidnell) and to show relations between utterances that are not adjacent (Bolden). Sorjonen and Hakulinen show that, in Finnish, word order does not code basic grammatical relations (as it does in English and many other languages) and can thus be deployed to index subtle distinctions of epistemic access, speaker stance and interactional alignment . While there are roughly comparable practices in languages that use word-order to code grammatical relations, these Finnish practices of responding are intimately tied to the grammar of this language and thus ultimately unique to it. The broad theme relating all the chapters, then, is how the generic problems of interaction among humans (coordination of opportunities to participate, fixing problems, responding in such a way as to convey an understanding of what the previous speaker has done, constructing a course of action in a sequence, etc.) are solved through the mobilization of specific local resources . Those resources include not only the grammar and lexicon of a particular language but also prosody and the body. It is the mutual shaping of these resources, the structures of interaction and the production of coordinated courses of action that these chapters investigate. Acknowledgments I thank Nick Enfield , Chuck Goodwin , Makoto Hayashi , Manny Schegloff, and Tanya Stivers for helpful written comments on an earlier version of this chapter.
26
Jack Sidnell
Notes 1 Variously termed descriptive, functional or cognitive linguistics . For a useful overview of language universals that contrasts these two approaches see Comrie 1981. Some major landmarks in the descriptive/ functional tradition are, for instance, Li and Thompson (1981), Shopen (1985). 2 So-called “non-configurational” languages which exhibit free word order provide a striking illustration of the way grammar shapes the practice of repetition. Thus Hale writes of Warlpiri: “sentences containing the same content words in different linear arrangements count as repetitions of one another” (Hale 1983:5). 3 While each language offers unique possibilities for social action, there are also clear points of convergence as well. Indeed in some cases it is possible to identify the very same practice in typologically and genetically unrelated languages. For instance, Stivers (2004) shows that a practice involving the use of multiple sayings such as “no no no” to halt an ongoing course of action is found not only in English but also in Japanese and Russian . 4 The chapters by Fox et al. and Rossano et al. depart from the collection-based, qualitative CA approach described here and develop their analyses through the use of quantitative methods. For discussion of such methods applied to the materials of social interaction see these chapters as well Schegloff ’s commentary. 5 See Goffman (1964) for an early juxtaposition of two research traditions: one in which features of context are correlated with the features of speech and another in which the focus is on the structures of social interaction themselves, that is, “the neglected situation.” 6 See, however, Egbert (2004) for an examination of some occasions in which just these issues come to the interactional surface. 7 As Makoto Hayashi points out, CA studies do sometimes begin with action-types: Pomerantz (1978) on compliments, Goodwin and Goodwin (1987) on assessments, and Drew (1984) on invitations are examples. These studies differ from those in cross-cultural pragmatics by taking the action-implementing turn as part of a larger organization of the turn at talk (i.e. the Goodwins’ study of assessments) or sequence of action (Pomerantz and Drew). Emanuel Schegloff notes that although the titles of the papers by Pomerantz and Drew refer to compliments and invitations, the actual texts are about responses to compliments and invitations. 8 Schegloff goes on to note “Further, the use of other-initiated repair as one way of pre-indicating the imminent occurrence of disagreement
Introduction
27
(Schegloff et al. 1977: 380) suggests why uh huh’s and the like can be taken as indications of agreement with the speaker of an ongoing extended unit. For if disagreement were brewing, then opportunities to initiate repair would supply a ready vehicle for the display and potential deflection of that disagreement. Passing the opportunity to raise problems of understanding may be taken as indicating the absence of such problems. It may also be taken as indicating the absence of that which such problems might have portended – disagreement – and thus be taken as indications of agreement.” 9 As Schegloff notes, “attention signals” is a gross under-specification. First because it is not clear why attention should be signaled on the occasion in which “uh huh” is used and not others in which it is not and second because it is not clear why “uh huh” should be used rather than some other signal such as gaze.
Part II
Repair and beyond
2 Repetition in the initiation of repair Ruey-Jiuan Regina Wu
Introduction The organization of repair has received sustained interest in the study of language and social interaction over the past decades. Although, until recently, research in this area has based its claims primarily on English materials, there has been a growing interest, especially in the past ten years or so, in exploring how repair operates in languages other than English. This expanding body of research includes studies of German (e.g., Egbert 1996, 1997b, 2004; Selting 1988, 1992, 1996; Uhmann 2001), Japanese (e.g., Fox et al. 1996), Korean (e.g., Kim 1999a, 2001), Thai (e.g., Moerman 1977), and Mandarin Chinese (e.g., Chui 1996; Tao et al. 1999; Wu 2006; Zhang 1998), among others. Some of these studies have focused on the mechanisms of self- and other-initiation of repair in the languages examined (e.g., Chui 1996; Kim 1999a, 2001; Moerman 1977 ; Zhang 1998), while others have uncovered the linkages between repair and other aspects of interactional practice, such as prosody (e.g., Selting 1996; Tao et al. 1999) and bodily conduct (e.g., Egbert 1996), and still others have explored the relation between repair and syntax from a cross-linguistic perspective (e.g., Fox et al. 1996). As part of this growing effort to understand the organization of repair across languages, this chapter investigates two repeatformatted other-initiated repair practices in Mandarin conversation. Using the methodology of conversation analysis (CA), this study will show that the two Mandarin repair initiations under examination, like other-initiation of repair in English, serve not only to initiate repair but also as vehicles for accomplishing additional
32
Ruey-Jiuan Regina Wu
negatively valenced actions, such as displaying a stance of disbelief or nonalignment . However, in further explicating the common sequential and activity contexts of these two practices, I will also describe a previously undescribed division of labor between repair initiations and demonstrate how the use of these two Mandarin practices is sensitive to two intertwining axes: the epistemic stance of the speaker who initiates the repair and the sequential context and positioning of the initiation of repair. The analysis draws upon a corpus of approximately two hours of telephone conversations and seventeen hours of videotaped face-to‑face conversations among family members, friends, and acquaintances. The data were collected in mainland China , Taiwan, and the USA. In what follows, I will first briefly review previous work on other-initiation of repair and lay out a few initial observations regarding the use of the two Mandarin practices under examination in clear-cut contexts of repair. I will then examine occasions in which they are used to implement actions beyond repair initiations and discuss how these uses can be related to the basic properties they exhibit as repair initiators.1 The candidate phenomenon: repeat-formatted other-initiation of repair In general, research on other-initiation of repair has been pursued along two different lines. The first line of inquiry (e.g., Curl 2005; Egbert 1996; Kim 1999a, 2001; Schegloff 2000b; Schegloff et al. 1977; Zhang 1998) is concerned mainly with the internal organization of repair sequences, exploring issues such as the locus of repair initiation relative to the trouble source turn as well as the repertoire of practices for the initiation or resolution of repair. The second line of inquiry (e.g., Couper-Kuhlen 1992; Drew 1997; Egbert 1997b, 2004; Kim 1999a; Robinson 2006; Schegloff 1997b; Selting 1988, 1992, 1996) focuses not so much on repair sequences per se, but rather on the larger sequential contexts in which such repair initiations and sequences figure, as well as their interactional uses in these contexts. One prominent theme along this second line of inquiry is the double-barreled nature of other-initiated repair – that is, its
Repetition in the initiation of repair
33
capacity for, and frequent use in, implementing additional interactional projects while simultaneously serving to initiate repair. Egbert (1997b), for example, demonstrates how initiating repair in multiparty German conversation can be used strategically not only as an entry or exit device to such conversations but also as a display of affiliation among conversational co-participants. Selting (1988, 1992, 1996) shows that prosodically marked other-initiation of repair in German can carry an emotive overtone of “astonishment” and indicate a problem of expectation on the part of the speaker. Drew (1997) focuses on what he terms “open class” repair initiators in English (e.g., “Pardon?,” “Sorry?,” “What?”) and shows how their use can embody the speaker’s treatment of the turn being targeted as either topically or sequentially disjunctive or in some way inappropriate. And the strong connection between the use of other-initiated repair and the display of incipient nonalignment on the part of its speaker has also been documented in several studies (e.g., Heritage 1984: 339–344; Kim 1999a; Pomerantz 1984; Schegloff 1997b, 2007a; Schegloff et al. 1977 ). The following example, from Schegloff (2007a: 102–103) , is a case in point: (1) Schegloff (2007a:102– 103) 1 Bee: a→ =[Why] whhat’sa mattuh with y- Yih sou[nd HA:PPY,] hh 2 Ava: [Nothing. ] 3 Ava: b→ u- I sound ha:p[py?] 4 Bee: c→ [Yee]uh. 5 (0.3) 6 Ava: d→ No:,
Here, Ava initiates repair (Arrow B) of Bee’s assessment (“Yih sound HA:PPY,” Arrow A) of Ava’s emotional state. To this repair initiation, Bee responds in the next turn (Arrow C) with a confirmation . Although this response by Bee suggests that she is treating Ava’s repair initiation, prima facie at least, as involving a hearing or understanding problem, Ava’s subsequent disconfirmation (Arrow D) nonetheless retroactively indicates that this prior repair initiation serves not merely to initiate a repair but to portend disagreement as well. The present study pursues the second line of research and considers a similar phenomenon in Mandarin conversation as that illustrated in Extract 1. An initial sense of the presence of a similar
34
Ruey-Jiuan Regina Wu
operation – and a further complication – involved in the use of repeat-formatted other-initiated repair in Mandarin is evident in the following two excerpts, both paralleling Extract 1 closely: (2) (CMC01–403A) 1 C:
ta yi xiaoshi:: fu ni duoshao, xianzai. she one hour pay you how:much now “How much does she pay you- for an hour:, now?”
2 W: a→ pianyi de le. sanshi. cheap NOM CRS thirty “Cheap. Thirty dollars.” 3 C: b→ yi ge xiaoshi sanshi? one C hour thirty “Thirty dollars for an hour?” 4 W: c→ (uh [huh) PRT PRT “(uh [huh.]” 5 C: d→ [Bu pianyi. N cheap [ “Not cheap.” (3) (CMC 05–06, 00:20, audio 315a) 1 L: a→ hai, hebei jiu mei shenme hao difang. (sigh) (province) just N what good place “((sighs)) There are not many good places in Hebei.” 2 M: b→ hebei a?= (province) PRT “Hebei A?”= 3 L: c→ =uh. PRT = “Yeah.” 4
(0.5)
5 M: d→ langfang hai xing. (place) still okay “Langfang is okay.”
As with Extract 1, in each of these excerpts, a speaker initiates repair (Arrow B) by repeating part of the prior assessment by another (Arrow A); and in each excerpt, following a recipient confirmation (Arrow C ), the repeat speaker moves to launch a
Repetition in the initiation of repair
35
disagreeing response (Arrow D): In Extract 2, the speaker delivers a straightforward negation of the prior assessment; and in Extract 3 the speaker introduces an exception to the recipient’s just-proffered negative evaluation of the province by claiming that one of its cities – i.e., Langfang – is “okay.” Despite the similarity, it can be noted, however, that the repair initiations in these extracts are not in entirely the same format: Whereas the repeat in Extract 3 is suffixed with the particle a, 2 the one in Extract 2 is not. What Extracts 2 and 3 suggest, then, is both a commonality and variation between English and Mandarin . On the one hand, as in English conversation, repeat-formatted other-initiation of repair in Mandarin apparently can serve not only to initiate repair but also as a vehicle for accomplishing negatively valenced actions such as projecting disagreement . On the other hand, however, in initiating a repair through the use of repeat to accomplish such an action, a Mandarin speaker will have to select between at least two alternative practices: a question-intoned repeat3 and a repeat suffixed with the final particle a. The second of these practices appears to have no direct analogue in English.4 As will be shown below, the selection between these two Mandarin practices in implementing negatively valenced projects is not random; rather, the division of labor appears to be consistent with, and may in fact be carried over from, the basic meanings they index when serving as straightforward repair initiators. To demonstrate such a connection, let us turn first to a brief overview of the use of these practices in clear-cut contexts of repair initiation. Question-intoned repeats and a-suffixed repeats in clear-cut contexts of repair As straightforward repair initiations, question-intoned repeats and repeats suffixed with a commonly adumbrate different problems their speakers are having with respect to the trouble source turns targeted and frequently engender different types of responses from their recipients. Repeats suffixed with a are commonly heard as confirmation questions, i.e. the use of this format makes a recipient’s confirmation or disconfirmation relevant in next position. This is illustrated respectively in Extracts 4 and 5, both from a conversation recorded in
36
Ruey-Jiuan Regina Wu
Beijing between two acquaintances. Immediately prior to Extract 4, M has just asserted that life in Beijing is boring. In Extract 5, the topic has turned to a neighborhood where M used to live. (4) (CMC 05–05, 00:40, audio 243a) 1 L:
suibian ni. as:wish you “Whatever.”
2
(0.3)
3 L: a→ chabuduo zhao ren jia le de le. almost find person marry ASP get CRS “(I’ll) just get married to some guy.” 4
(0.5)
5 M: b→ zhao ren jia le de le a?= find person marry ASP get CRS PRT “Just get married to some guy A?”= 6 L: c→ =uh.= PRT = “Yeah.”= 7 M: d→ =wo zher hen duo ren dasuan qu ne. I here very many person plan marry PRT = “I know lots of guys who are ready to marry.” (5) (CMC 05–03, 01:45, audio 130a) 1 M:
bu, yuanlai women jia zhu nar? no originally we home live where “No, where did we live before?”
2 M:
3 M: a→ =[dongcheng liuyin jie na kuair. (place) (place) street that area = “Near Liuyin Street in Dongcheng.” 4 L:
=[((clears throat))
5
(0.5)
6 L: b→ [[guyun jie a?= (place) street PRT [[ “Guyun Street A?”=
Repetition in the initiation of repair 7 M:
37
[[shi bbe [[ “(It) is-”
8 M: c→ =aiyahh, liuyin jie. (exclamation) (place) street = “aarrgh! Liuyin Street.” 9
(.)
10 L: d→ ao:. PRT “Oh:”
In each of these cases, a speaker displays a hearing or understanding problem with respect to a prior turn (Arrow A) by initiating a repair through a partial repeat suffixed with the final particle a (Arrow B); and, in each case, this a-suffixed repeat is responded to with either a confirmation or disconfirmation (Arrow C). That such a confirmatory type of response is adequate for repair initiations in the form of a-suffixed repeats can be evidenced by the repeat speakers’ subsequent moves (Arrow D): They either proceed to produce a sequentially relevant pending response, as in Extract 4, and, by implication, treat the recipient’s confirmatory response as sufficient for their prior inquiries. Or, as in Extract 5, they receive the confirmation with a free-standing ao “oh,” which, by proposing a change in the speaker’s local state of information as a result of the response, reaffirms the status of the a-suffixed repeat as one which is meant to request a confirmation or disconfirmation . By contrast, a confirmatory type of response is not always adequate for question-intoned repeats. In the database, although some such repeats appear to be treated as candidate hearings offered for confirmation of a prior turn’s talk, the majority of such repeats turn out to involve recognition or understanding problems. Consider, for example, Extract 6 below, in which the participants are talking about a variety of foods in the northeast area of Mainland China . (6) (CMC 01–01, 03:22, audio 034a) 1 Z:→ ranhou ta nar ( . . . ) yi yang, dongbei de ming cai, then it there one C northeast ASSC famous dish “And there’s ( . . . ) another famous northeastern dish,”
38
Ruey-Jiuan Regina Wu
2 Z:→ jiao zuo guobaorou. call CP (dish) “called ‘guobaorou.’” 3
(0.5)
4 N:→ guobaorou? (dish) “‘Guobaorou?’” 5 Z:→ ((nod))= 6 N:→ =she- na shenme [guo? that what (word) = “Wh- which [‘guo’ is that?” 7 Z:
[guo. (word) [ “‘Guo.’”
8 W:
[ ↑nang ba. (word) PRT [ “↑Isn’t it ‘nang’?”
9 Z:
[guo: jiu shi guo: de guo:. (word) just be (word) ASSC (word) [ “‘Guo:’ is just the ‘guo:’ in ‘guo:.’”
In line 4, N initiates repair on the name of the food just mentioned through a question-intoned repeat. Although Z’s response – a simple affirmative nodding (line 5) – displays her treatment of N’s repair initiation as possibly checking a candidate hearing, N’s immediate move to follow Z’s response with a further clarification question (line 6) retrospectively marks that the trouble she has been proposing involves a recognition problem, rather than a hearing problem. Extract 7 provides another instance of question-intoned repeats. Here, despite the non-presence of a follow-up clarification question, the question-intoned repeat is nonetheless treated as embodying an understanding problem. In this excerpt, the participants are exchanging their views of the city of Chengdu, which both of them have been to. (7) (CMC 05–02, 04:40, audio 105a) 1 M:
ni lai nabianr::(.) you come there “(When) you were there::(.)”
Repetition in the initiation of repair 2 M:
ni zui xi(huan) chengdu shenme? you most like (city) what “what did you like most about Chengdu?”
3
(0.5)
4 L:
chi de. qihou.= ou, wo zui xihuan qihou. qici shi eat NOM weather PRT I most like weather next be
39
chi de. eat NOM “Food. Weather. = Oh, I like its weather the best, and then the food.” 5 M:
you tong gan. zhende. have same feeling really “(I) feel the same. Really.”
6 M: → nar qihou jianzhi tai lan le. =zhende. there weather truly too bad CRS really “The weather there is just too bad. =Really.” 7
(.)
< na zhong- [na zhong 8 M: that kind that kind “The kind of- the kind of-” 9 L: (0.3) 10
[↑qihou tai lan le? weather too bad CRS “The weather is too bad?”
11 M: → ni shuo- ni (tian tian jian zhao taiyang mei zai nar)? you say you day day see CP sun N at there “Tell me- did you (see the sun every day there)?” 12
(0.7)
<↑a:, wo jiu xihuan mei taiyang a. 13 L: PRT I just like N sun PRT < “↑Ah::, I just like (the fact) that the sun does not come out.”
Here, consistent with how he behaves in the remainder of this conversation, M interacts with L in an ironic and joking manner. In line 5, M first emphatically aligns with L’s displayed favorable stance toward the weather of the city but then abruptly ends his responses with a sharp contrary view about this topic (line 6). To this unexpected twist, L, after a slight delay, reacts by repeating part of what M has just said with a raised pitch (↑qihou tai lan le,
40
Ruey-Jiuan Regina Wu
“↑the weather is too bad?”; line 9). Note here that M’s subsequent response (line 11), albeit with a shade of challenge , is clarificatory in nature and apparently treats L’s prior repeat as displaying an understanding problem. Contrasting Extracts 6–7 with 4–5, we find that whereas in all of these excerpts a party undertakes to claim some kind of trouble with a prior turn’s talk through a repeat, these differently formatted repeats have different sequential implications. When suffixed by the particle a, the repeat is intended and understood as a candidate hearing or understanding of the trouble source turn, ordinarily making a recipient confirmation relevant in next position. Without the a-suffixing and with question intonation, however, the repeat is commonly offered not as what the speaker proposes to have understood but as the speaker’s momentary breakdown in comprehension of the element being repeated in its current sequential context. Of particular relevance to our discussion below is that, as Extract 7 has also demonstrated, the speaker’s displayed noncomprehension of the trouble source turn embodied in the use of a question-intoned repeat may arise not only from the repeat speaker’s lack of familiarity with a reference previously mentioned (e.g., Extract 6); it may also arise from an apparent sense of abruptness and unexpectedness associated with how the trouble source turn figures in its sequential context (cf. Drew 1997). As noted, M begins his response to L’s favorable view of the weather in Chengdu with an aligning response (line 5), which appears to be leading up to a subsequent reinforcing positive comment. However, just as this response by M appears to be on its way to completion, it takes an abrupt twist and ends with a contradictory comment about this subject (line 6). It is notable that L’s subsequent repeat (line 9) targets just this seemingly incoherent part of the response by M, giving evidence that it is the unexpected twist manifested by this response that has caused her puzzlement. As we will see in the following sections, this use of questionintoned repeats to mark what is being repeated as unexpected or having come out of the blue, as well as the use of a-suffixed repeats to mark a relatively higher degree of speaker certainty, are generic features of these two practices. These features and the different epistemic stances so instantiated are present not only in contexts in
Repetition in the initiation of repair
41
which these two practices serve straightforwardly to initiate repair but also in contexts in which they accomplish actions beyond repair initiation.
The use of question-intoned repeats and a-suffixed repeats in the service of negatively valenced work To explicate the division of labor between the use of question-intoned repeats and a-suffixed repeats in implementing additional negatively valenced work, I will focus on sequences sharing the following features: • The repeat is responded to with a repair-relevant response – which commonly (though not invariably) takes the form of a confirmation or clarification of the matter addressed by the repeat. • While the repair-relevant response displays the repeat recipient’s treatment of the repeat, prima facie at least, as initiating a repair, namely, as indicating a hearing or understanding problem, the repeat speaker’s subsequent move suggests that his or her prior repeat is not merely repair-implicated. • Commonly, the repeat speaker follows the recipient’s response with a negatively valenced response regarding the matter being addressed – a move which suggests that the prior repeat-formatted repair initiation was employed (and perhaps exploited) in the service of projecting the follow-up negative response. In the following sections, I will demonstrate that while a-suffixed repeats and question-intoned repeats both figure in such sequences, they differ along two dimensions: the epistemic stance of the speaker who initiates the repair, and the sequential context and positioning of the repair initiation. I will consider these issues in that order. The effect of the epistemic stance of the repeat speaker We have seen that compared to question-intoned repeats, repair initiations in the form of an a-suffixed repeat commonly convey a stronger epistemic stance on the part of the speaker. To illustrate
42
Ruey-Jiuan Regina Wu
that such a difference is also visible when the use of these two practices implicates a problem of interactional alignment , let us consider first Extract 8 below, which involves a question-intoned repeat that serves to project a negatively valenced response. In this excerpt, the participants are talking about the expense of visiting Jiuzhaigou – a well-known scenic spot in China . (8) (CMC 05–06, 03:30, audio 368a) 1 M: jiuzhaigou, qu na bian, dei hua duoshao qian a? (place) go that side must spend how:much money PRT “How much does it cost to go (on a trip) to Jiuzhaigou?” 2
(.)
3 L:a→ tch! wo juede liang qian yinggai gou le. yi ge ren. I feel two thousand should enough CRS one C person “tch! I think two thousand should be enough, for one person.” 4 L:
[((clears throat))
5 M: b→ [yi (ge) ren liang qian?= one C person two thousand [“Two thousand for one person?”= 6 L: c→ =uh. PRT = “Yeah.” 7 M: d→ na jiu xia bei le. then just next life CRS “Then (wait) until the next life.” 8
(1.0)
9 L:
ganma? what “What?”
10 M:
tai gui le. too expensive CRS “(It’s) too expensive!”
11 L:
↑gui ma? expensive Q “↑Is (it) expensive?”
12 M:
tai gui le. tai gui le. too expensive CRS too expensive CRS “(It’s) too expensive! Too expensive!”
Repetition in the initiation of repair
43
Here, it can be noted that the repeat-formatted repair initiation (Arrow B) is followed by a negatively valenced response produced by the repeat speaker (Arrow D) after the information being targeted is confirmed (Arrow C). It can also be noted, however, that whereas the repeat speaker M subsequently displays a less-thanembracing stance toward the information being targeted, he does not challenge or contest the information itself. Rather, the production of the question-intoned repeat appears to have more to do with the speaker’s display of a sense of unexpectedness toward the information. This is embodied in M’s production of a cry of dismay (na jiu xia bei le, “then (wait) until the next life”; line 7), which, like his subsequent comments (tai gui le, “too expensive”) in lines 10 and 12, apparently treats the estimated travel expense as excessively high rather than untrue. The embodiment of a weaker epistemic stance in the selection of a question-intoned repeat is also apparent in cases in which the repeat speakers subsequently move to convey a sense of doubt about the truth-value of the target information: In these latter cases, it is observed that despite the expression of doubt, the repeat speaker nonetheless commonly designs and/or delivers his or her subsequent query in an epistemic ally downgraded fashion. A case in point is Extract 9, where the two participants, L and M, are talking about a friend of M’s, whose girlfriend was L’s classmate for several years at a university in Chengdu. (9) (CMC 05–09, 00:55, audio 096b) 1 L: wo dou hai mei jian guo ta. I all still N see ASP he “I haven’t met him yet.” 2
(.)
3 M: mei jian guo ta ma?= N see ASP he Q “Haven’t met him?”= 4 L: → =<wo zai chengdu de:shihou jiu lao tingshuo.= I at (city) when then always hear =< “I heard about him a lot when I was in Chengdu.”= 5 L: =ranhou (yao) dao beijing ye chang tingshuo. than will come (city) also often hear = “And then I also heard about him a lot in Beijing.”
44
Ruey-Jiuan Regina Wu
6
(.)
7 M:
[niyou [ “You-”
8 L:
[jiu wen da ming. long hear big name [ “Have heard his name for a long time,”
9 L:
ranhou mei jian guo ren. then N see ASP person “but have never met him in person.”
10 M: → ↑ni zai chengdu lao tingshuo guo? you at (city) always hear ASP “↑You heard about him a lot in Chengdu?” 11 M: → tamen liang [bu shi zai, chengdu bu shi renshi,= they two N be at (city) N be know “Didn’t they [in Chengdu- didn’t (they) know (each other)-”= 12 L: [wo tingshuo guo. I hear ASP [“I did.” 13 M: → =bu shi hen chang shijian (ma)? N be very long time Q = “not for a long time?’” 14 L:
zen::me bu shi hen chang?= how N be very long “How: could it not be a long time?”=
15 L: =dasan jiu renshi le ba. junior then know CRS PRT = “(I think) they knew each other since (they were) juniors.” 16 L:
yi nian duo me. one year more PRT “(They’ve known each other) for more than a year.”
17 M:
shi ma? be Q “Have they?”
Here, L’s claim of having frequently heard about this man in Chengdu (line 4) suggests that the romantic relationship between this man and L’s former classmate had already begun and was possibly even well established at that time. In response to this information and
Repetition in the initiation of repair
45
implication, M produces a question-intoned repeat (line 10), to which L responds with a confirmation (line 12). That this question-intoned repeat by M serves to do more than initiate repair can be seen in the fact that immediately following it, and without waiting for a responsive affirmation, M produces a question (tamen liang bu shi zai, chengdu bu shi renshi, bu shi hen chang shijian [ma], “didn’t they in Chengdu . . . didn’t they know each other . . . not for a long time?”; lines 11 and 13) which can be, and is in fact, heard as a challenge to the validity of L’s prior claim (notice L’s rebuttal in line 14). Note, however, that although this follow-up question by M is challenge -implicating, it is delivered and designed in an epistemic ally weakened fashion: Not only is the question delivered with many hesitations and disfluencies, but the final shape it takes – a double negative construction (bu shi renshi bu shi hen chang shijian (ma), “didn’t (they) know each other not for a long time?”) – is epistemically much weaker than its affirmative version would be (bu shi (zhi) renshi hen duan de shijian (ma), “didn’t (they) know each other for a short period of time (only)?”). In Extract 9, then, there is evidence that even though the question-intoned repeat embodies a certain degree of doubt on the part of the speaker toward the information just received, such a stance , as in many other similar instances in the corpus, is substantially weakened and mitigated. In fact, if we consider the questionintoned repeats examined earlier (e.g., Extracts 7–8), and the fact that across these instances a predominant usage of question-intoned repeats is to embody a sense of unexpectedness on the part of the speaker, there is a sense that the use of the question-intoned repeat in Extract 9 is strategic in nature. That is, by using a question‑ intoned repeat, and thereby invoking a sense of the unexpectedness commonly associated with its use, the repeat speaker in effect marks the newly received information as less than expected and in need of clarification , rather than as less than truthful and in need of correction . This difference, between marking a piece of information as less than expected versus marking it as less than truthful, can be more clearly demonstrated if we compare Extracts 8–9 with those involving an a-suffixed repeat in similar negatively valenced contexts. In
46
Ruey-Jiuan Regina Wu
these latter instances, the repeat speakers commonly follow their a-suffixed repeats with a statement or query which in some way contradicts or compromises the information just received and commonly displays a stronger epistemic stance toward the matter being addressed. Extracts 10 and 11 illustrate. Extract 10 provides an expanded version of Extract 3 and comes from the same conversation as Extract 9. Of particular interest here is that whereas in both Extracts 9 and 10 M initiates repair through a partial repeat of what his interlocutor has said in a prior turn, in Extract 10, below, M suffixes his repeat with the final particle a (line 2), in contrast to what he does (in line 10) in Extract 9. (10) (CMC 05–06, 00:20, audio 315a) 1 L: hai, hebei jiu mei shenme hao difang. (sigh) (province) just N what good place “((sighs)) There are not many good places in Hebei.” 2 M: → hebei a?= (province) PRT “Hebei A?”= 3 L:
=uh. PRT = “Yeah.”
4
(0.5)
5 M: → langfang hai xing. (place) still okay “Langfang is okay.” 6 L:
uh::? (.) bu hui ba.= (PRT) N likely PRT “Mm::? (.) Are (you) serious?”=
7 M: =langfang yinwei xianzai yao jian yi daxue cheng. (place) because now will build one college town = “‘cause now (they) plan to build a college town in Langfang.” 8 L:
diang. (na) xia zhao le.= (exclamation) that scare CP CRS “Right! How impressive!”=
9 M:
=jiu shi li beijing jin.= just be away:from (city) close = “(It’s) just very close to Beijing.”=
Repetition in the initiation of repair
47
10 L: =wo dangnian kao- daxue de:shihou, I that:year test college when = “The year when I applied to- colleges,” 11 L: wo bao le yi ge- wujing kexue yuan. I apply ASP one C police science college “I applied to- the Police Academy.” 12
(0.5)
13 L: qu nar mianshi,<wo jiu faxian,.hhh= go there interview I then find “(I) went there for an interview.< Then I found out .hhh”= 14 L:
=langfang hai meiyou (chengdu) xiancheng- (.) (place) yet N (place) county = “Langfang wasn’t as- (.)”
15 L: ganjing (.) zhengjie- (.) da. clean tidy big “clean- (.) tidy- (.) or as big as (Chengdu) county.” 16 M:
langfang nar mei ren guan.= (place) there N person take:charge “Nobody is in charge in Langfang.”=
17 M: → =xiao shihou wo ye zai langfang dai guo.= little when I also at (place) stay ASP = “When (I) was little, I also stayed in Langfang.”= 18
((The storytelling continues.))
Here, as noted earlier, in response to L’s assessment of Hebei (line 1), M not only produces an a-suffixed repeat (line 2) but, following the recipient confirmation (line 3), also moves to produce a disagreeing response (line 5). Notice that compared with the negatively valenced response which M produces subsequent to his question-intoned repeat in Extract 9, this disagreeing response by M (langfang hai xing “Langfang is okay”; line 5) is designed and delivered with a much stronger epistemic stance : Not only is this disagreeing response delivered in a straightforward manner without significant hesitation or disfluency, it is also cast as a straightforward assertion which presents information contradictory to that which is being targeted, rather than as a query displaying weaker epistemic strength. Notice also that in Extract 10 not only does M have prior firsthand knowledge concerning the city of Langfang, but it is precisely
48
Ruey-Jiuan Regina Wu
by reference to this knowledge that he finds L’s prior comment in line 1 problematic (lines 7, 9, and 16–18). In this respect, therefore, M’s subsequent delivery of firmly asserted information that is contrastive with what L has just said, as well as his move to invoke his prior independent experience of this matter, are compatible with his selection of an a-suffixed repeat as the repair initiator; all of these actions display a stronger epistemic stance on the part of the speaker toward the matter being addressed. Extract 11, from a conversation between two college roommates, provides a similar instance. Where this excerpt begins, H is in the midst of reporting an incident she saw on her way to school one day. In line 6, D interrupts H’s storytelling by first launching a repair initiation clarifying the place where this incident took place. After the place reference is clarified (lines 7–8), D then retrieves a piece of information that H had reported earlier (line 1), by repeating part of that information and suffixing it with a (line 9). (11) (CMC 09 audio 150a) 1 H:
=jiu kan yi liang huang se chuzuche guo:qu le. then see one C yellow color taxi pass:by CRS = “and then (I) saw a yellow cab passing by,”
2 H: .h ranhou faxian libian de n de then find inside ASSC female NOM “.hhh then (I) found the lady inside-” 3 H:
↑zhang de hao yanshou [ou. look CSC very familiar PRT “↑looked really familiar.”
4 D: 5 H:
[↑hhheng:: (exclamation) xiang xiang xiang xiang, think think think think “(I) kept thinking and thinking and thinking,
↑ a:: [( . . . ) (exclamation) oh:.” 6 D: [( . . . ) zai nar a? at where PRT “(. .) where?” 7 H:
jiu- jiu zai na- zanmen xuexiao na ge- (0.3) lu shang. just just at that we school that C road on “(It was) just- just on- the road- (0.3) in the school”
Repetition in the initiation of repair 8 H:
49
xuexiao de lu shang. school ASSC road on “On the road in the school.”
9 D: → ↑huang se chuzuche a? yellow color taxi PRT “↑A yellow cab A?” 10 H:
um.= PRT “Yeah.”=
11 D: → =xiali a? (cab company) PRT = “A Xiali (cab)?” bu shi. huang se de. 12 H: N be yellow color NOM “No. (It was) a yellow one.” 13
(0.5)
chuzuche [a? 14 D: → ↑zher na: you huang se de here where have yellow color ASSC taxi PRT “↑How can there be yellow cabs here?” 15 H:
[you.= have “There are.”
=shang bian yi banr shi huang de, yi banr (bie . . . ) 16 H: top side one half be yellow NOM one half other = “The top of the cab was half yellow, (and) half (another . . . )” 17 D:
ou::. PRT “Oh::.”
In response to H’s affirmation (line 10) of the a-suffixed repeat, D initiates another repair by producing an understanding check (xiali a, “a Xiali (cab)?”; line 11), which takes the form of a best guess about the specific cab company that owns the yellow cab in question.5 This understanding check is disconfirmed by H (line 12), who goes on to reaffirm the color of the cab, and, by implication, the validity of her prior report as well. In answer to this disconfirmation , D produces a question (line 14), which is designed with the question word nar, “where,”6 with slight stretching, as well as stress on the word huang, “yellow,” which represents the issue in question. In and through this design, the question can be heard
50
Ruey-Jiuan Regina Wu
as “pro-forma” in nature. That is, it is not meant to request new information but to question a prior utterance by introducing the relevance of a corresponding negative assertion, namely, the assertion that there are no yellow cabs at all here (cf. Koshik 2003). And this is just the way H treats it: In line 15, she moves to confirm the accuracy of her report, though backing down with a modification of her prior claim (line 16). Extract 11, then, offers us another instance in which the speaker of an a-suffixed repeat displays an orientation to independent knowledge concerning the matter being addressed – and an instance in which the repeat speaker subsequently produces a doubt-implicated disaligning response that displays a stronger epistemic stance . If an epistemically weakened follow-up query, as argued earlier, retrospectively marks the prior question-intoned repeat as the speaker’s display of a sense of unexpectedness toward the information just received (e.g., Extract 9), then the subsequent challenge -implicated query produced by D here, as well as her invocation of prior knowledge on the matter, can be seen as indicating that her previous a-suffixed repeat is not only unexpectedness-implicated but is, in fact, challenge -relevant. This section has shown that the choice between question-intoned repeats and a-suffixed repeats in accomplishing negatively valenced work is largely determined by, and itself embodies, the epistemic stance of the speaker toward the information in question: Whereas the use of a question-intoned repeat is commonly accompanied by a follow-up negatively valenced response that is epistemically weakened or unexpectedness-implicated, an a-suffixed repeat is ordinarily followed by a negative response which displays the speaker’s orientation to prior knowledge of the matter at issue, and which is often built with a stronger epistemic stance . Such a difference, I have also suggested, is consistent with the features these two practices exhibit in the context of straightforward repair initiation. The effect of the sequential context and positioning of the initiation of repair The fact that the speaker’s epistemic stance, albeit relevant, is not the only axis pertinent to the selection between these two
Repetition in the initiation of repair
51
repair initiations is made clear by a subset of cases containing question-intoned repeats. These cases are similar to the previously examined excerpts involving a-suffixed repeats (e.g., Excerpts 10–11) in that the repeat speakers move subsequently to invoke prior knowledge regarding the target information. A unique feature of the majority of these cases is that such questionintoned repeats almost always occur in a sequentially disjunctive position, produced to address a prior turn by another which appears to have been abruptly launched or ill-fitted in its sequential context from the repeat speaker’s perspective.7 This usage of a question-i ntoned repeat, I would suggest, is consistent with its use to mark that which is being repeated as unexpected or having come out of the blue, as we have observed in earlier examples (e.g., Extracts 7–8). Consider Extract 12, from a telephone conversation between two friends who practice a Chinese martial art together at the same association. Prior to this excerpt, B, who had been absent from classes for several days, has just told F that she does not know how to get to the sessions due to a recent change in bus schedules. This excerpt begins with B’s reiteration of her problem with transportation, which appears to be produced here as an implicit request for information, but which apparently is treated by F as an account for B’s absence and is merely responded to with an information receipt (line 4). (12) (T-mom & bo 165) 1 B: dao na ge (wulun) chang a, arrive that C (name) field PRT “(As for) how to get to the (Wulun) Association” 2 F:
um. PRT “Uh huh.”
3 B:
zuo shenme lu hai bu zhidao lei. sit what road still N know PRT “(I) don’t know which bus to take yet.”
4 F:
ou, zhe yangzi ou. PRT this manner PRT “Oh, I see.”
52 5 B: 6
Ruey-Jiuan Regina Wu ↑wu:lao kongpa y- .hhh ta- ta ye bu zhidao ba? (person) afraid he he also N know PRT “↑Mr. Wu probably also- .hh He- he doesn’t know either, right?” (1.0)
7 F: → w ulun:: chang: wo- (0.5) (name) field I “The Wulun::Association:I- (0.5)” 8 B: → ↑bu shi,= N be “↑No,”= 9 F:
=ta xianzai shi zuo- (.) haoxiang zuo yibai lu. he now be sit seem sit one:hundred road = “Now he takes- (.) seems to take Bus 100.”
10
(.)
11 B: →
one:hundred road < “Bus 100?/!”> 12 F:
dui. wu:lao shi zuo yibai. right (person) be sit one:hundred “Yes. Mr. Wu takes 100.”
13
(1.2)
14 B: → ou:? PRT “Yeah?” 15 F:
en ehnhh ((smile voice)) PRT PRT “Uh huh.”
16 B: → oh, ta xianzai gai le yibai lu.= PRT he now change ASP one:hundred road “Oh, now he’s changed to Bus 100.”= 17 F:
=dui::. right = “Right::.”
18 B: → bu shi shiyi lu. N be eleven road “Not Bus 11?” 19 F:
dui, bu shi. right N be “That’s right. No.”
Repetition in the initiation of repair 20 B:
53
haha (laugh) “haha”
Here, not having been provided with any information on the proposed transportation problem, and possibly also taking F’s reaction (line 4) as a sign of her not having a solution, B brings up the name of another fellow practitioner (Wu Lao “Mr. Wu”), asking whether he might know about how to get to the association with the new bus schedules (line 5). Whereas this question by B in line 5 is designed in such a way that requires (minimally) only a confirmation or disconfirmation from the recipient, F appears to exhibit some difficulty in giving a response: Following B’s question, there is first a one-second pause (line 6); thereafter, F attempts to provide a response, which nonetheless ends with yet another pause (line 7). Whereas it is clear from F’s response in line 7 that she is still searching for an answer or searching for a way to deliver a solution, it is opaque as to where her answer-in-progress is headed. In fact, the way in which F has begun her turn in line 7 (wulun:: chang: wo-, “the Wulun:: Association:: I-”) can promote a hearing that what is projected is something about F’s own knowledge in relation to some aspects of this association, rather than about Mr. Wu’s, as B’s prior question (line 5) has asked. On this hearing, then, F can be taken by B as having misunderstood her (B’s) prior question. And, indeed, this seems to be what happens here. In line 8, apparently attempting to remedy such a perceived misunderstanding on F’s part, B initiates what Schegloff (1992a) terms a “third position repair,” which regularly takes the form, “No, I don’t mean X. I mean Y.” However, just as B finishes producing the first component, bu shi, “no,” and is apparently on her way to completing this third-position repair, she is interrupted by F, who moves at this point to resume her earlier-suspended response (line 9). And it is to this interruptive response by F that B launches a question-intoned repeat (yibai lu, “Bus 100?/!”; line 11). It should be noted that F’s response in line 9 is not only interruptive of B’s prior turn in progress (line 8) but also shifts away from the line of response introduced (yet not completed) in F’s own prior turn (line 7). In this regard, F’s response can be seen, from B’s perspective, as “doubly disjunctive” and hence may strike B as having come out of the blue. I would suggest that it is just this (dual) sense
54
Ruey-Jiuan Regina Wu
of the unexpectedness of the occurrence of F’s response at this point in interaction that has prompted B – who otherwise appears to have some prior knowledge of the matter being inquired about (cf. her reactions in lines 14, 16, and 18) – to choose a questionintoned repeat, rather than an a-suffixed repeat, in initiating the repair. As Extract 12 demonstrates, sequential context and positioning appear to play a role in the selection of the two repair-initiation formats under examination. In addition, such an effect can, and often does, prevail in cases where the repeat speakers turn out to have independent prior access to the matter being addressed – cases in which an a-suffixed repeat would otherwise be commonplace, as demonstrated in the previous section. This suggests that although the selection of a question-intoned repeat versus an a-suffixed repeat is closely related to the distribution of knowledge between the repeat speaker and his or her interlocutor, what the selection conveys is not always how the state of knowledge is objectively distributed nor the speaker’s subjective evaluation of it but can rather be a display by the speaker of how he or she perceives the trouble source turn to figure in its sequential context. Question-intoned repeats in the context of intensified nonalignment In the data so far examined, we have observed two major contexts in which question-intoned repeats are commonly chosen over a-suffixed repeats: (i) contexts in which the repeat speaker subsequently delivers a negatively valenced response that is epistemic ally weakened or unexpectedness-implicated, and (ii) contexts in which the repair initiation figures in a sequentially disrupted environment. In both contexts, I have argued, the use of question-intoned repeats frequently serves to embody a sense of unexpectedness on the part of the speaker toward the matter being addressed. However, question-intoned repeats are sometimes observed to occur in contexts which are neither sequentially disruptive nor in which the negatively valenced responses they project appear in any way mitigated. While instances like this initially seem to conflict with the preceding account, upon examination, the use
Repetition in the initiation of repair
55
of question‑intoned repeats in these cases appears again to mark “unexpectedness”. A return to Extract 2, which is reproduced here with a few ensuing turns as Extract 13, provides a case in point. (13) (CMC01–405A). 1 C: ta yi xiaoshi:: fu ni duoshao, xianzai. she one hour pay you how:much now “How much does she pay you- for an hour:, now?” 2 W:
pianyi de le. sanshi. cheap NOM CRS thirty “Cheap. Thirty dollars.”
3 C: → yi ge xiaoshi sanshi? one C hour thirty “Thirty dollars for an hour?” 4 W:
(uh [huh) PRT PRT “(uh [huh.)”
5 C: → [Bu pianyi. N cheap [ “Not cheap.” 6 W: SHENme BU pianyi a. ta yanjiusheng. what N cheap PRT she graduate:student “WHat do you mean by NOT cheap? She’s a grad student.” 7
(.)
8 W: wo jiao ta gudai hanyu. I teach her ancient:time Chinese “I teach her Classical Chinese.” 9
(1.0)
10 C: u-oh.=ta shi yanjiusheng, =zheme da yi ge ren le a. PRT she be graduate:student such big one C person ASP PRT “e- Oh.=She is a grad student?/! Such an old person.” 11
(0.5)
12 W:
ni: yiwei [ne ((infiltrated with laughter)) you think PRT “What did you suppose?”
13 C: [buguo sanshi kuaiqian queshi- suan hen hao de.= but thirty dollar actually count very good NOM “But thirty dollars is truly a very good pay.”
56 14 C:
Ruey-Jiuan Regina Wu =ni zhiyao- gen: na ge- (0.4)= you only with that C = “If you- with the uh:(0.4)”=
=na shei gei ta zhao le yi ge= 15 C: that who for she find ASP one C = “Who’s the person that has found one for her?”= 16 C:
=luo gei- dan zhao le [yi ge= (name) for (name) find ASP one C “Luo has- found a (tutoring job) for Dan.”
17 W: 18 C:
=gei su:dan zhao [le yi ge, yi ge xiaoshi cai shi wu kuaiqian. for (name) find ASP one C one C hour only fifteen dollar “Found one for Dan Su. Only fifteen dollars an hour.”
19 W: 20
[uh:: PRT “Yeah,”
[uh. PRT “Yeah,”
(.)
21 W: .hhhh jiao shei a. nei ge jiao shenme ia. (in-breath) call who PRT that C teach what PRT “.hhhhh Who’s the person? Er- what does she teach?”
As noted earlier, C follows W’s confirmation (line 4) of her prior repair initiation (line 3) with a straightforward disagreement (line 5). Note here that whereas C chooses a question-intoned repeat in initiating the repair, this excerpt is different from those examined earlier (e.g., Extracts 8–9, 12) in that not only is C’s follow-up negative response not epistemic ally weakened but the question-intoned repeat is also not produced in a sequentially disjunctive context. A remarkable feature of instances like Extract 13 is the striking discrepancy between the information targeted by the questionintoned repeat and what the repeat speaker understands the issue to be. In this excerpt, C subsequently (lines 14–16, 18) reports a case she knows of in which a mutual friend has only been paid half of what W has reportedly received. In light of this background information, and perhaps also in light of the way in which W has initially oriented her recipients to assess the amount of pay to be reported (pianyi de le, “cheap”; line 2), the exact figure
Repetition in the initiation of repair
57
which W ends up reporting can be seen by C as so unlikely that this information may appear initially to have been arrived at out of the blue. Here, I would suggest, C’s question-intoned repeat appears to embody just this kind of initial reaction – that is, marking a sense of unexpectedness on the part of the speaker. And the unexpectedness being so marked here concerns not so much the abruptness with which another’s prior talk is brought up in the local context but the significant gap between the newly received information and the repeat speaker’s prior knowledge of the matter being addressed.8 Before leaving this section, it may be worth registering that given the capacity of question-intoned repeats to mark that which is being repeated as utterly beyond expectation, the possibility can be entertained that this feature may make question-intoned repeats an apt and well-suited resource for projecting intensified or outright disagreement . That is, by (strategically) marking what another has said in a prior turn as having come out of the blue, the speaker can in effect problematize the target information while at the same time reinforcing the adequacy of the position he or she is holding. However, while such a usage seems plausible, and, on occasion, even familiar to the ear, there are no clear exemplars of this kind in the data at hand. This proposal thus awaits further investigation and should be treated as impressionistic at present. Conclusion Over the years, the other-initiation of repair has been a recurrent theme in CA and related fields and is perhaps one of the most well studied interactional phenomena cross-linguistically. Using a CA framework, this study has shown that, like other-initiation of repair in English, the two Mandarin repair-initiation formats under examination can similarly serve not only to initiate repair but also as vehicles for accomplishing additional negatively valenced projects. In further explicating the common sequential contexts of their occurrences, this study has also shown that these two repair initiation formats, while occurring in seemingly overlapping contexts, exhibit a consistent fine-tuned division of labor: Whereas a-suffixed repeats generally embody a stronger epistemic stance on the part of the speaker, question-intoned repeats, across
58
Ruey-Jiuan Regina Wu
various contexts, consistently embody a speaker’s display of a sense of unexpectedness toward the information being targeted. Before closing this chapter, one final note regarding the implications for cross-linguistic and cross-cultural work on repair should be registered. As the foregoing discussion has demonstrated, in both Mandarin and English, repeat-formatted other-initiation of repair can serve not only to locate a hearing or understanding problem but to implicate alignment issues as well. And yet, the linguistic resources in Mandarin provide two alternative repeat formats through which speakers can calibrate a fine-grained epistemic stance toward the matter being addressed vis-à-vis their interlocutors while initiating the repair. These findings raise interesting questions as to whether other languages make a similar distinction in embodying negatively valenced stances in the context of other-initiation of repair, and if so, how this is achieved without the availability of final particles. Clearly, more studies are needed to further our understanding of the organization of repair in Mandarin as well as across languages, and more studies are needed if we are to understand how the organization of repair is, as suggested (e.g., Egbert 1996; Fox et al. 1996), an interactional phenomenon that is universal across languages and cultures while at the same time sensitive to the linguistic and cultural repertoire of each given language. Appendix: Abbreviations ASSC Associative (-de) ASP Aspectual marker CRS Currently relevant state (le) CSC Complex stative construction C Classifier N Negator NOM Nominalizer (de) PRT Particle Q Question marker Notes 1 Part of the analysis discussed in this chapter has appeared in Wu (2006).
Repetition in the initiation of repair
59
2 Over the years, various scholars have proposed different meanings for the particle a. (For a brief review, see Wu 2004.) In a more recent work, Wu (2004: 128) demonstrates that in TCU (turn constructional unit)final position, the particle a generally exhibits a “contrast-invoking” property, used to “mark a discrepancy in knowledge, expectation or perspective regarding some state of affairs between the current speaker (i.e., the a user) and the prior speaker.” 3 The matter of what constitutes “question intonation” in a tonal language such as Mandarin is complicated and remains unsettled. Over the years, different proposals have been made as to how question intonation is realized in Mandarin. These proposals, albeit not in complete agreement with each other, generally point to some exaggerated gestures with which question intonation is realized (e.g., Chang 1998; Gårding 1987; Ho 1977; Schack 2000; Shen 1989; Tseng 2003; Yuan and Shih 2004; Yuan et al. 2002; Zeng et al. 2004). In this study, I have taken into consideration these findings while basing my identification of question-intoned repeats on my auditory perception after repeated listenings. The exaggerated gestures claimed in the literature to be associated with question intonation are reflected in the transcripts in the forms of underlinings, upward-pointing arrows, and/or left-facing arrows. 4 There are other repeat formats in Mandarin , e.g., repeats suffixed with the question particle ma (e.g., line 11 in Extract 8) or the particle ou (see Wu 2004). The exact functional differences among these formats await further investigation. 5 “Xiali” is the name of a cab company whose cabs, however, are red. 6 In this question, D chooses to design the question with nar, which literally means “where” in English but which can also serve to mark a question as being rhetorical. 7 Question-intoned repeats are observed to also occur in contexts in which the repeats are produced in a sequentially abrupt manner. For this usage, see Wu (2006). 8 That the difference between W’s pay and the pay of this friend can be perceived as huge can also be supported by W’s subsequent reaction after she learns of this friend’s pay – a big in-breath (line 21), apparently adumbrating a big surprise.
3 A cross-linguistic investigation of the site of initiation in same-turn self-repair Barbara Fox, Fay Wouk , Makoto Hayashi , Steven Fincke , Liang Tao , Marja-Leena Sorjonen , Minna Laakso, and Wilfrido Flores Hernandez Introduction Same-turn self-repair is the process by which speakers stop an utterance in progress and then abort, recast or redo that utterance. While same-turn self-repair has become a topic of great interest in the past decade, very little has been written on the question of where within a word speakers tend to initiate repair (the main exceptions being Schegloff 1979 and Jasperson 1998). And, to our knowledge, no work has been done on this question from a crosslinguistic perspective. The goal of this paper is twofold: First, we explore existing proposals regarding where in words speakers initiate repair (what we will call the “site” of initiation) using data from our seven languages; and, second, we present and explain site of initiation data from those seven languages. Our findings suggest that there is a great deal of cross-linguistic variation with respect to favored sites of initiation but that most of the variation can be accounted for by a few simple interactional factors. This paper is the first study we are aware of which considers word length in explanations of selfrepair data. The current study is part of a larger project that has as its goal an understanding of the universal principles of self-repair and their language-specific manifestations. Prior studies have shown that the linguistic resources available to speakers of different languages shape the specific methods by which repair is accomplished (Fox et al. 1996; Wouk 2005; Fincke 1999; Egbert 2002; Sidnell 2007c; Karkainnen et al., 2007 ). We hope to learn more about the details of recycling and replacement repairs and how such
Initiation in same-turn self-repair
61
linguistic factors as major constituent order, frequency of ellipsis, use of prepositions or postpositions , and phonological structure of the word impact the organization of repair. The seven languages examined in this study are: Bikol (spoken in the Philippines), Sochiapam Chinantec (spoken in Mexico), English, Finnish , Indonesian, Japanese , and Mandarin . The language sample is one of convenience, but it represents languages from six language families, three continents and three archipelagos. The languages differ in major typological features, such as basic word order , phonotactic structure, and morphological structure. In other words, the sample is genetically, typologically, and areally diverse. Below we give very brief descriptions of each language: Bikol , a language of the Philippine branch of the Austronesian family, is verb-initial and prepositional. As in other Philippine languages, content words are multisyllabic and multimorphemic. Much of the bound morphology is prefixal, although there are infixes, suffixes, and circumfixes as well. Sochiapam Chinantec, of the Chinantecan branch of the OtoManguean language family, is spoken in Oaxaca, Mexico. It is a verb-initial, prepositional language. Words tend to be monosyllabic and monomorphemic, but there is a small amount of derivational and inflection al morphology. Sochiapam Chinantec is a tonal language. Finnish , in the Finno-Ugric language family, is verb-medial, with both prepositions and postpositions . It also has an extremely rich system of inflection al and derivational morphology, much of which is suffixing. Indonesian , an Austronesian language spoken in Indonesia, is verbmedial and prepositional. Words tend to be bisyllabic, although there is a considerable amount of derivational morphology which produces longer words. The genetic classification of Japanese remains in dispute, some treating it as essentially a language isolate, while others believe it belongs to the Altaic language family. Japanese is verb-final and postpositional. Japanese words are not divided into syllables, but rather into units called mora, which determine syllable weight (which, in its turn, determines stress and timing). Words tend
62
Barbara Fox et al.
to consist of a large number of mora, and thus to be quite long in terms of their phonological structure, even though they may look no longer than a two syllable word in English. For example, “Nippon,” which looks like it is only two syllables, is actually four mora, with the double consonant and the final n each constituting a separate mora. Japanese has a rich system of bound morphology, all of which is suffixing. Mandarin , in the Sino-Tibetan language family, is verb-medial and prepositional. It is thought of as a monosyllabic language, but modern Mandarin has such a large number of homophones that the practice of compounding has become quite prevalent, and many content words are compounds. Mandarin is also a tonal language. Data collection As part of a larger project on the syntax of self-repair (cf. Fox and Wouk 2003), we collected 250 instances of same-turn selfrepair for each language (except English, for which we have 500 instances). The instances were taken from recordings of naturally occurring speech. The recordings were not made specifically for this project; rather, they had been collected by the individual researchers at different times and in different places, for a variety of purposes. Each instance of repair was coded for a variety of features. The coding scheme was developed for the larger project, and it included information relating to type of repair and site of initiation. Self-repair can involve a variety of different operations, but the larger study was restricted to quite specific types. Only instances in which there was a clear syntactic relationship between the trouble source (the original utterance) and the repair were included in the collection. Thus, instances in which the speaker started an utterance, cut it off, and started a new utterance were excluded from the larger collection. In fact, the larger study is limited to cases which involve either replacement of one or more elements of the trouble source or repetition of the trouble source (or some part thereof), although either replacement or repetition (which we term “recycling”) may be accompanied by other repair operations.
Initiation in same-turn self-repair
63
In the present study, we further limit our focus to simple repairs, both simple recycling and simple replacement . By “simple” we mean that no other repair operation is involved in the repair. Thus, simple recycling repairs are just repetitions of words, with no add itions, deletion s or replacements; simple replacement repairs are instances in which a morpheme, word or phrase is replaced, without recycling, addition or deletion .1 Since the overall study includes both simple and non-simple repetition and replacement repairs, the total number of simple repairs is less than 250 per language (or 500 for English). In this paper we examine • 146 Indonesian repairs (117 recycling and twenty-nine replacement); • 201 Sochiapam Chinantec repairs (185 recycling and sixteen replacement); • 339 English repairs (285 recycling and fifty-four replacement); • 200 Japanese repairs (147 recycling and fifty-three replacement); • 150 Mandarin repairs (115 recycling and thirty-five replacement); • 185 Bikol repairs (162 recycling and twenty-three replacement); • 162 Finnish repairs (116 recycling and forty-six replacement). As can be seen by these figures, languages differ in the proportion of simple to non-simple repairs and also in the proportion of replacement to recycling repairs, but overall simple recycling is far more frequent than simple replacement. Method ology There is a large body of work on self-repair in the conversation analysis (CA) framework (e.g., Jefferson 1974; Schegloff et al. 1977; Goodwin 1979, 1981; Schegloff 1979, 1987c, 2000c, 2002b). This work is based on qualitative analysis of individual cases in their interactional environment, or of collections of such cases. The studies are grounded in the understanding that orderliness is achieved first in single cases and that a perceived orderliness in the aggregate arises from orderliness on a single case basis. “Quantitative analysis is, in this sense, not an alternative to single case analysis, but
64
Barbara Fox et al.
rather is built on its back” (Schegloff 1993: 102). In other words, we see similar patterns in multiple instances because speakers manage interactional contingencies with recurrent practices. It is entirely possible that orderliness at the single case basis will not lead to orderliness at the aggregate level. In fact, Schegloff (1993) expresses serious concerns about undertaking quantitative analysis of conversational phenomena. In that paper, he raises questions about what counts as numerator and what counts as denominator. As Schegloff (1993) points out, these are not trivial questions, and in many cases the phenomenon may not be well enough understood to identify either all of the environments of relevant possible occurrence (i.e. the denominator) or all of the possible items that could count as instances of the phenomenon under investigation (i.e. the numerator). As Schegloff has suggested “We may need to defer full dress quantitative analysis in many areas until its appropriate constraints can be met” (1993: 119). For these and other reasons, CA researchers have typically avoided quantitative analysis, preferring instead to use “informal quantification.” As Schegloff says: Terminology such as occasionally or massively reports an experience or grasp of frequency, not a count; an account of an investigator’s sense of frequency over the range of research experience, not in a specifically bounded body of data; a characterization of distribution fully though tacitly informed by the analytic import of what is being characterized. (Schegloff 1993: 118–119)
In this paper, however, we take findings from relevant CA work and explore their cross-linguistic implications utilizing quantitative method s. Although we are cognizant of the concerns mentioned above, we have chosen to use a quantitative approach because we are working with phenomena which are difficult for an analyst to notice at a conscious level, and to have a general grasp of, such as how often repair is initiated early in the word in a noun rather than in an article. We thus make use of quantification as a tool to help us arrive at a sense of frequency. Although this paper is full of numbers (and even statistics), we do not believe that the fact that there were, for example, X instances of repair initiated early in a noun and Y instances of repair initiated early in an article in data set P of Language Q is significant in and of itself. It only serves to
Initiation in same-turn self-repair
65
give us a sense of whether both of these syntactic types can be considered frequent sites of early repair initiation in that language. The questions we have asked here are not just linguists ’ questions; they can also be participants’ questions in that they involve the resources for interactional tasks such as delaying next item due or replacing a current item. For example, languages with a large number of verbal prefixes provide speakers with an easy way of delaying the verb root by recycling the prefix, while in languages which lack verbal prefixes speakers will need to find other means to delay production of the verb. We begin our study with an exploration of the first proposal in the literature regarding site of initiation. Site of initiation The first proposal we will explore concerning site of initiation appears in Schegloff (1979). There, Schegloff writes, “Just postinitiation and just pre-completion of various unit types seem to be specially common loci of repair initiation. Thus, just after the start of a turn-constructional unit (e.g., a sentence) or just before its completion; after the first sound of a word or just before its last sound” ( 1979: 275). This is just one point in a long paper covering many topics, and Schegloff does not elaborate on it nor give any examples. Nor is it clear from the paper exactly what is meant by “after the first sound of a word or just before its last sound”; after the first sound could mean after the first sound is recognizable or after the first sound is complete; just before its last sound could likewise mean before the last sound is articulated or before the last sound is complete. However, Schegloff (personal communication ) has suggested that the relevant domain for just post-beginning starts after the first sound is recognizable, thus has begun to be articulated, and continues until the first sound is complete, while the relevant domain for just pre-completion begins just before the final sound is articulated, in the penultimate sound, and continues until just before the final sound is complete. He has further suggested that site of initiation of repair may depend on the type of repair involved. The first goal of our study is to explore Schegloff ’s characterization of English site of initiation in the simple repairs found in the
66
Barbara Fox et al.
seven languages of our sample. In order to avoid confusion of terminology, we will use the term “post-beginning” for just post-initiation of the word and “initiation” only to refer to repair initiation. Post-beginning and pre-completion in English We began our exploration with English. In our collection of 339 simple repairs in English, we found twenty-four simple recycling repairs (8 percent of all simple recycling repairs) and seventeen simple replacement repairs (31 percent of all simple replacement repairs) in which repair was initiated during or right after the first sound (post-beginning). Extract 1 below illustrates a simple recycling repair initiated post-beginning, in this instance in the initial vowel: (1) Chinese dinner Ann: the biggest debate i- in our department. in:, at Trenton was that when we had these faculty meetings.
Extract 2 illustrates a simple replacement repair initiated postbeginning, in this instance just after the bilabial nasal closure (which could be heard to begin the word “my,” replaced by a): (2) Sports Allan: I wish I’d had m- a camera earlier today because
We further found 128 simple recycling repairs (45 percent) and twenty-two simple replacement repairs (41 percent) in which repair was initiated just before the last sound (pre-completion). Extract 3 illustrates initiation in the penultimate sound of the demonstrative determiner “that”: (3) SBC1 [715] Lynne: .hh Not this time of the year, because, you know they’ve had a dry year, bu-< .hhhh usually they’re too we:t, and tha:- (0.7) that shoe can just pull ri:ght o:ff:, because it’s just, .hhh (.) the hoof wa:ll is so so:ft.
Extract 4 shows initiation in the final sound of the pronoun she; the pronoun is replaced with a full noun phrase (this girl):
Initiation in same-turn self-repair
67
(4) HG Hyla: and she- this girl’s fixed up on a da- a blind da:te.
It is worth noting that the vast majority of these pre-completion initiations occur in the final sound of the word, rather than prior to it. Combining post-beginning and pre-completion frequencies for each type, we find that 152 (53 percent) of simple recycling repairs and thirty-nine (72 percent) of simple replacement repairs are in accord with Schegloff ’s proposal. It thus seems that for simple replacement repairs in English, Schegloff ’s proposed description of the common sites of repair initiation provides a rough characterization of our data. However, while post-beginning and precompletion are both common sites of initiation of simple recycling repairs, they comprise only approximately half of these repairs. We thus wondered if other explanations might be available that would account for a greater proportion of the data. We explore other possible explanations for our findings beginning in the section ‘Recognizable Completion’ below. Among simple replacement repairs, on the other hand, the positions mentioned by Schegloff comprise 72 percent of all repair initiations, so it appears that Schegloff ’s generalization holds quite strongly for this category of repair. As Schegloff (personal communication ) noted, it is to be expected that different repair types behave differently. 2
Cross-linguistic comparison Table 3.1 provides the figures for post-beginning and pre-completion sites of repair initiation for both types of repairs in all of the languages in our study. For each language, we give the percentage of all recycling repairs and replacement repairs that are initiated postbeginning and pre-completion. We also combine repairs initiated post-beginning and pre-completion, thus giving the percentage of all repairs initiated in the positions suggested by Schegloff (1979). In the discussion below, we explore the pattern of distribution of post-beginning and pre-completion repair initiation for each language, starting with Indonesian. As can be seen from Table 3.1, in Indonesian , post-beginning is not a common site of repair initiation. Out of 146 repairs, only
Japanese
Finnish
Bikol
Mandarin
Sochiapam Chinantec
Indonesian
English
24 (8%) 17 (31%) 6 (5%) (0%) 2 (1%) 1 (6%) 5 (4%) 2 (6%) 8 (5%) 1 (4%) 21 (18%) 13 (28%) 30 (20%) 2 (4%)
Recycling Replacement Recycling Replacement Recycling Replacement Recycling Replacement Recycling Replacement Recycling Replacement Recycling Replacement
Post-beginning
7 (44%) 33 (29%) 6 (17%) 37 (23%) 11 (48%) 19 (16%) 7 (15%) 39 (27%) 17 (32%)
128 (45%) 22 (41%) 51 (35%) 8 (28%) 22 (12%)
Pre-completion
8 (50%) 38 (33%) 8 (23%) 45 (28%) 12 (52%) 40 (34%) 20 (43%) 69 (47%) 19 (36%)
152 (53%) 39 (72%) 57 (39%) 8 (28%) 24 (13%)
Combined
Table 3.1 Post-beginning and pre-completion repair initiation
16 115 35 162 23 116 46 147 53
285 54 117 29 185
Total repairs
Initiation in same-turn self-repair
69
six simple recycling repairs (5 percent) and no simple replacement repairs were initiated immediately post-beginning. Pre-completion, on the other hand, does seem to be a moderately frequent site, as fifty-one simple recycling repairs (35 percent) and eight simple replacement repairs (28 percent) are initiated pre-completion. Combining post-beginning and pre-completion frequencies for each type, we find that fifty-seven (39 percent) simple recycling repairs and eight (28 percent) simple replacement repairs are in accord with Schegloff ’s proposal. In Sochiapam Chinantec and Mandarin , as in Indonesian , postbeginning is not a common site of repair initiation; in Sochiapam Chinantec only two simple recycling repairs and one simple replacement repair are initiated post-beginning, while in Mandarin two (6 percent) simple replacement repairs and five (4 percent) simple recycling repairs were initiated post-beginning. The two languages differ from each other in the degree to which speakers seem to orient to pre-completion. In Sochiapam Chinantec, pre-completion is a proportionally more frequent site for replacement repair initiation – being found in seven cases (44 percent) – than for recycling repair initiation, where it is found in twenty-two cases (12 percent). In Mandarin, it is a more frequent site for recycling repair initiation – being found in thirty-three cases (29 percent) – than for replacement repairs, where it is found in six cases (17 percent). Combining post-beginning and pre-completion frequencies for each type, we find that in Sochiapam Chinantec twenty-four (13 percent ) simple recycling repairs and eight (50 percent) simple replacement repairs are in accord with Schegloff ’s proposal, while in Mandarin thirtyeight (33 percent) simple recycling repairs and eight (23 percent) simple replacement repairs are. In Bikol as well, post-beginning is not a frequent site of repair initiation; only nine (5 percent) simple repairs were initiated immediately post-beginning, eight simple recycling repairs and one simple replacement repair. Among simple recycling repairs, thirty‑seven (23 percent) were initiated just pre-completion, while among simple replacement repairs, eleven (48 percent) were initiated pre-completion. Combining post-beginning and pre-completion frequencies for each type, we find that forty-five (28 percent)
70
Barbara Fox et al.
s imple recycling repairs and twelve (52 percent) simple replacement repairs are in accord with Schegloff ’s proposal. In Finnish , as in English, post-beginning is a moderately frequent site of repair initiation, particularly for replacement repairs. We found that thirty-four (21 percent) Finnish simple repairs, twenty-one (18 percent) recycling repairs and thirteen (28 percent) replacement repairs, were initiated immediately post-beginning. Pre-completion is also used with some frequency, although less frequently than post-beginning, as a site of repair initiation in Finnish. Among simple recycling repairs, nineteen (16 percent) repairs were initiated just pre-completion. Among simple replacement repairs, seven (15 percent) were initiated just pre-completion. Combining post-beginning and pre-completion frequencies for each type, we find that forty (34 percent) simple recycling repairs and twentysix (43 percent) simple replacement repairs are in accord with Schegloff ’s proposal. In Japanese , repair is initiated in post-beginning position with moderate frequency for recycling, but not for replacement, as thirty-two (16 percent ) repairs – thirty (20 percent) simple recycling repairs and two (4 percent) simple replacement repairs – were initiated immediately post-beginning. Pre-completion initiation is moderately frequent for both repair types. Among simple recycling repairs, thirty-nine (27 percent) were initiated just pre-completion. Among simple replacement repairs, seventeen (32 percent) were initiated just pre-completion. Combining post-beginning and pre-completion frequencies for each type, we find that sixty-nine (47 percent) simple recycling repairs and nineteen (36 percent) simple replacement repairs are in accord with Schegloff ’s proposal. Summary of cross-linguistic comparison What we have found for simple repairs in our seven languages, then, is that Schegloff ’s proposed description matches English the best and is a particularly useful generalization about site of initiation in simple replacement repairs. However, the pattern is clearly not universal , and most of the matching in our other languages is accounted for by pre-completion, not by post-beginning. In fact, in several languages (Indonesian , Bikol , Sochiapam Chinantec, and
Initiation in same-turn self-repair
71
Mandarin ) speakers do not orient toward post-beginning at all, while in others speakers show an orientation toward it only for one type of repair (simple replacement repairs in English and simple recycling repairs in Japanese), and only in Finnish do speakers seem to orient toward it for both types. In all languages, speakers showed some orientation toward pre-completion, but varied greatly in how much. Furthermore, while speakers of all of the languages in our study show an orientation toward at least one of these two sites of repair initiation, only in English are these sites particularly common. Recognizable completion We have seen that post-beginning is far less relevant crosslinguistically than pre-completion, as illustrated in Table 3.1. Additionally, the majority of the pre-completion repair initiations in all the languages of our study occurred in the final segment, rather than just before it. Furthermore, for several of the languages in our study (Bikol , Finnish, Indonesian, and English) between one quarter and one third of recycling repairs were initiated after the word had been completely articulated, and for Mandarin and Sochiapam Chinantec the majority of recycling repairs were initiated in this location. Only in Japanese was this site rarely used. It would seem, then, that for most languages recycling repairs tend to be initiated either in the final sound of the word or after it. It might be useful to consider why this is the case. We will refer to initiation in or after the last sound of the word as initiation at recognizable completion. In four of the languages in our study (English, Mandarin , Sochiapam Chinantec, and Indonesian), speakers favor initiating repair at recognizable completion in recycling repairs. However, not all languages in the study showed this orientation; we demonstrate in ‘CrossLinguistic Comparison’ below (as illustrated in Tables 3.4, 3.5, and 3.6), that speakers of Finnish, Bikol, and Japanese seem to have no special prefer ence for sites of initiation after the word is recognizably complete. We will consider a number of factors below in explaining the difference between these two groups of languages.
72
Barbara Fox et al.
Given the work that the concept “recognizable completion” seems to do for speakers of at least four of our languages, it seems promising to broaden Schegloff ’s categories of post-beginning and pre-completion to “before recognizable completion” and “recognizable completion” as possible sites of repair initiation oriented to by speakers. Recognizable completion includes most cases where repair is initiated pre-completion and also those cases where repair is initiated after the word is complete. Initiations in the penultimate sound (part of Schegloff ’s original characterization of pre-completion) would be excluded, but there turned out to be very few of them in our data, in any language. All other sites of repair initiation would be considered a single category, “before recognizable completion,” which would contrast with “recognizable completion.”3 It has long been known (Sacks et al. 1974) that participants orient to recognizable completion in their monitoring of the turn-so-far for transition relevance; however, it could be argued that speakers do not orient to the recognizable completion of a word but rather only to the recognizable completion of a turn constructional unit (TCU), and therefore that the relevant characterization of withinword repair sites would be something else. Another possible characterization, for example, could be that speakers tend to initiate repair late in a word, without regard to projection of completion. Although little research has been done on speakers’ and recipients’ monitoring of word production (but see Jasperson 1998; Jefferson 1974), our findings suggest that “recognizable completion” better captures the phenomenon. If lateness were the appropriate characterization, we would expect to find speakers frequently initiating repair in the position just prior to the final sound, as well as in that sound and after it, since just before the final sound is also late. However, we do not find this. What we find is that speakers overwhelmingly orient to recognizable completion in initiating repair. This is a fact that needs to be captured. So, how is it that speakers and recipients orient to the completion-relevant trajectory of words, in addition to TCUs? We suggest here that while recipients monitor the course of an utterance for where it might come to possible completion, they are also monitoring the unfolding of an utterance for its relevance to other forms of co-participation (e.g., head nod, gaze organization,
Initiation in same-turn self-repair
73
a ssessment‑relevant activity) (cf. Ford et al. 1996), and for the moment-by-moment, word-by-word, shifting course of the construction of the TCU (cf. Ford et al. 1996). Each word shapes the trajectory of the TCU, either renewing the current trajectory or changing it. It thus makes sense that speakers and recipients should orient to the word as a unit of production, and to its completion. Furthermore, although it may not be the case for English, in many languages a word’s identity is not certain until it is complete. For example, in Japanese and Finnish, which make use of a large number of inflection al and derivational suffixes, a word’s syntactic class, and therefore the trajectory of the turn so far, can remain indeterminate until the end. In the remainder of the study, then, we will make use of the concepts “recognizable completion” and “before recognizable completion” as within-word sites of repair initiation. And how do replacement repairs pattern with regard to recognizable completion? It appears that they are treated quite differently from recycling repairs, at least in some languages in our survey. For example, recognizably complete initiations comprise only 35 percent of all simple replacement repairs in English. This interesting difference between recycling and replacement in the use of the category recognizably complete will be discussed further below. We got our inspiration about how to think about this difference between the two repair types from the only major work on site of repair initiation, Jasperson (1998). We turn now to a summary of that work. Site of initiation and repair type Jasperson first introduces a distinction between two different types of cut-offs, word-disruptive cut-off and word-preserving cut-off, a distinction he clarifies as follows: “Word-preserving cut-off is cut-off which occurs upon recognizable completion of the word, so that by the time cut-off occurs, a word has been produced intact. In contrast, word-disruptive cut-off is cut-off which occurs at any point before recognizable completion of the word-in-progress, resulting in a ‘word fragment’” (1998: 140). Repair that is initiated before a word is recognizably complete is thus word-disruptive. During the final sound of a word, speakers
74
Barbara Fox et al.
can produce either word-disruptive or word-preserving cut-off: In word-preserving cut-off, the cut-off is done softly, and the final segment is not shortened or perturbed in any way, while in worddisruptive cut-off the cut-off is louder, the final segment is typically shortened, and glottalization may begin during a vowel preceding the final segment in a consonant-final word. Jasperson then goes on to suggest a correlation between these cut-off������������������������������������������������������������� types and the type of repair they construct: “The orientation of word-disruptive cut-off is typically retrospective, insofar as it routinely initiates repair which operates to change some part of the turn in progress (cf. Schegloff 1979). [. . .] Word-preserving cut-off on the other hand does not seem to have this retrospective orientation” (1998: 77). Jasperson ’s findings thus suggest that repair initiated before a word is recognizably complete tends to be engaged with retrospective work, while repair at recognizable completion is not engaged with retrospective repair. In our study we included a wide variety of initiation types, not just cut-off , and did not make use of phonetic distinctions between different types of cut-offs. Rather, we focused on where in the word initiation took place. This decision was largely practical; many of the data sets being used were not of sufficiently high sound quality to make detailed phonetic distinctions, nor did the researchers all possess the auditory acuity for accurate coding . Therefore, we could not test Jasperson ’s proposal against other languages. However, Jasperson’s findings encouraged us to look for other correlations between repair type and repair initiation. In particular, we looked at whether retrospective vs. non-retrospective repair might correlate with initiation when the word is recognizably complete, or earlier in the word, before it is recognizably complete. We therefore looked to see whether the difference between recycling and replacement noted above could be explained in terms of sensitivity to recognizable completion. Based on Jasperson’s study, we hypothesized that initiation before the word is recognizably complete is associated with repairs that change an element in the preceding talk (for example by replacing it), while initiation when the word is recognizably complete is associated with repairs that operate on upcoming talk, for example by delaying next item due (by recycling current word). The following section tests this hypothesis for English.
Initiation in same-turn self-repair
75
Table 3.2 Recognizable completion and repair initiation in English
Recycle Replace Total
Not recognizably complete
Recognizably complete
Total
58 (20%) 35 (65%) 93 (27%)
227 (80%) 19 (35%) 246 (73%)
285 54 339
Chi-square = 45.09, d.f. = 1, p = .0000
Repair type and site of initiation in English Our English data support this claim strongly; 65 percent of all simple replacement repairs are initiated before the word is recognizably complete, while only 20 percent of simple recycling repairs are (see Table 3.2 below). This difference is significant (p = .0000), and both repair types contribute to the significance. It thus appears that in English there is a strong association between site of initiation and repair type, even when a wide range of types of repair initiation are included. Before examining the patterns in our other languages, it is worth exploring possible explanations for this association in English. One obvious explanation is the function of each repair type. Recycling can be used to elicit gaze from recipients or to treat the start of an utterance not accompanied by recipient gaze as defective (Goodwin 1981); it can also be used to delay the next item due, either in a word search or as an indication that the speaker is choosing between alternatives (Jefferson 1974); further, recycling is found in overlaps, as a device to produce talk in the clear (Schegloff 1987c). While all of these functions are found in our English data, we have found that recycling in English tends to be used to repeat function words, which we hypothesize is most often done in order to delay the production of the next content word due (80 percent of all simple recycling in our English data repeat function words). Because their purpose in most cases is exactly to delay production of next item due, there is no hurry in initiating the repair. Replacement, on the other hand, tends to be used to replace content words (61 percent of simple replacement repairs in
76
Barbara Fox et al.
English replace content words) and may occur in cases where an inappropriate word or pronunciation has been produced. Jefferson (1974) provides some insight into this phenomenon; she suggests that speakers make use of the format word1 + hesitation + word 2 to replace one word with another, and that speakers can employ this device to produce just enough of a word for the word to be recognizable, at least retrospectively, thereby allowing the speaker to produce an inappropriate word without being interactionally accountable for it. Or, for example, one can propose ‘I am not like this but am talking by reference to the fact that you are’ by finding ways to show that the terms one produces are not the terms which first come to mind. (Jefferson 1974: 193) One means of achieving that display may be the production of just enough error to convey one’s habitual terminology without inheriting complaints from its recipient (i.e. not having “officially” produced the word in question) and then correcting it with a term which can be seen as selected by reference to one’s situation or recipient. (Jefferson 1974: 193)
As Jefferson indicates, in such cases there may be pressure to initiate repair as early as possible, to produce a word without being accountable for its production. Of course, not all replacements are so interactionally significant, but speakers may nevertheless prefer to minimize the production of an errorful or inappropriate word (cf. Jasperson 1998).4 However, the pressure may be less in these cases than in the type of instances that Jefferson examines, thus allowing for initiation after recognizable completion. 5 Our explanation for differences in preferred site of repair initiation for recycling and replacement repairs thus appears to rest on what might be universal motivations. But do other languages actually manifest the same patterns as English? The following section examines how the other languages pattern. Cross-linguistic comparison If we look at Indonesian , the results appear promising. Indonesian patterns much like English, as shown in Table 3.3 below, with the majority (74 percent) of recycling repairs being initiated when the word is recognizably complete, while only a minority (28 percent) of the replacements are initiated in that site. As in English, the
Initiation in same-turn self-repair
77
Table 3.3 Recognizable completion and repair initiation in Indonesian
Recycle Replace Total
Not recognizably complete
Recognizably complete
Total
30 (26%) 21 (72%) 51 (35%)
87 (74%) 8 (28%) 95 (65%)
117 29 146
Chi-square = 22.37, d.f. = 1, p = .0000
Table 3.4 Recognizable completion repair initiation in Finnish
Recycle Replace Total
and
Not recognizably complete
Recognizably complete
Total
52 (45%) 35 (76%) 87 (54%)
64 (55%) 11 (24%) 75 (46%)
116 46 162
Chi-square = 12.95, d.f. = 1, p = .0003
difference is highly significant (p = .0000), and both repair types contribute to the significance. Finnish is somewhat similar, with speakers strongly favoring initiations before the word is recognizably complete for replacement repairs; recycling repairs, however, are more evenly distributed between initiations before the word is recognizably complete and when the word is recognizably complete, as shown in Table 3.4. Note that while the overall distribution in Table 3.4 is significant (p= .0003), the significance comes entirely from the Replace row. Recycling repairs are actually distributed evenly with respect to repair initiation type (chi-square = 1.24; p = 0.27). However, when we examine Bikol (Table 3.5) and Japanese (Table 3.6), we can see immediately that the pattern as manifested by English and Indonesian is not universal . For Japanese (p = .5649) and Bikol (p = .8428) initiation type does not pattern significantly according to repair type. From Table 3.5 we can see that in Bikol speakers have no prefer ence for either site of initiation, with both recycling and
78
Barbara Fox et al.
Table 3.5 Recognizable repair initiation in Bikol
Recycle Replace Total
completion
and
Not recognizably complete
Recognizably complete
Total
88 (54%) 13 (56%) 101 (55%)
74 (48%) 10 (44%) 84 (45%)
162 23 185
Chi-square = 0.04, d.f. = 1, p = .8428
Table 3.6 Recognizable completion repair initiation in Japanese
Recycle Replace Total
and
Not recognizably complete
Recognizably complete
Total
106 (72%) 36 (68%) 142 (71%)
41 (28%) 17 (32%) 58 (29%)
147 53 200
Chi-square = 0.33, d.f. = 1, p = .5649
replacement repairs being initiated before recognizable completion roughly 55 percent of the time; Table 3.6 suggests that speakers of Japanese favor repair before recognizable completion, with replacement repairs initiated before recognizable completion 68 percent of the time and recycling repairs initiated before recognizable completion 72 percent of the time. The difference between the two languages is highly significant (chi-square = 11.11, d.f. = 1, p = .0009, chi-square calculated on the totals for not recognizably complete and recognizably complete for each language). In fact, Japanese proved to be the only language in our sample where speakers actually favor initiation before recognizable completion for recycling repairs. Turning now to Sochiapam Chinantec (Table 3.7) and Mandarin (Table 3.8), we see yet another pattern, namely that speakers strongly favor initiation when the word is recognizably complete for both recycling and replacement repairs, with no significant difference between the two types of repair (Sochiapam Chinantec p = .2385, Mandarin p = .4006).
Initiation in same-turn self-repair
79
Table 3.7 Recognizable completion and repair initiation in Sochiapam Chinantec
Recycle Replace Total
Not recognizably complete
Recognizably complete
Total
26 (14%) 4 (25%) 30 (15%)
159 (86%) 12 (75%) 171 (85%)
185 16 201
Chi-square = 1.39, d.f. = 1, p = .2385
Table 3.8 Recognizable completion repair initiation in Mandarin
Recycle Replace Total
and
Not recognizably complete
Recognizably complete
Total
12 (10%) 2 (6%) 14 (9%)
103 (90%) 33 (94%) 136 (91%)
115 35 150
Chi-square = 0.71, d.f. = 1, p = .4006
There is no significant difference between Mandarin and Sochiapam Chinantec with respect to initiation type of simple recycling repairs (chi-square = .840, d.f. = 1, p = 3595, chi-square calculated on the values for recognizably complete and not recognizably complete for recycling repairs in the two languages). There are too many cells with values below 5 to attempt a statistical comparison of the two languages with respect to simple replacement repairs, but the pattern is clear: In both languages speakers strongly favor initiation after recognizable completion for replacement repairs. Summary With a sample of seven languages we have seen five of the logically possible patterns for site of initiation and repair type: 1. no difference, with speakers favoring initiation before the word is recognizably complete for both types of repairs (Japanese);
80
Barbara Fox et al.
2. no difference, with speakers favoring initiation when the word is recognizably complete for both types of repairs (Mandarin , Sochiapam Chinantec); 3. no difference, with speakers favoring neither type of initiation (Bikol); 4. speakers favoring initiation when the word is recognizably complete for recycling repairs and initiation before the word is recognizably complete for replacement repairs (English, Indonesian); 5. speakers favoring initiation before the word is recognizably complete for replacement repairs, and showing no prefer ence for recycling repairs (Finnish). It is difficult to argue for a universal pattern when so many of the logical possibilities are found in such a small sample of languages. Given our discussion of the interactional functions of recycling and replacement repairs in “Monosyllabic Words and Recognizable Completion,” this is a surprising finding. We would have expected that data from other languages would give evidence of speakers orienting to the interactional pressures we discussed, and at least on the surface they appear not to. Nonetheless, we believe that there are underlying universal patterns at work here. Recycling tends to be initiated after recognizable completion, as it is frequently employed to delay the next content word due. The only exception to this is Japanese , where recycling repairs are generally initiated prior to recognizable completion and where function words generally follow content words. Replacement tends to be initiated prior to recognizable completion to reduce accountability for inappropriate words. However, word length, that is the number of syllables in a word,6 and syntactic class of the word in which repair is initiated, are also important factors, and their influence can at least partially explain our surprising finding. In the next section, we explore in detail the underlying universal patterns relating to site of initiation, word length and syntactic class. Word length organization and site of initiation The languages in our sample vary in their morphological structure and especially in the length of words involved in repair: Mandarin ,
Initiation in same-turn self-repair
81
Table 3.9 Site of initiation of simple recycling repair initiated in monosyllabic words Not recognizably complete Sochiapam Chinantec Indonesian Mandarin Bikol English Finnish
Recognizably complete
Total
4 (3%)
137 (97%)
141
1 (4%) 5 (6%) 7 (12%) 39 (16%) 8 (21%)
24 (96%) 86 (94%) 43 (88%) 209 (84%) 30 (79%)
25 91 50 248 38
Chinantec, and English speakers tend to initiate repair in monosyllabic words;7 Indonesian speakers most often initiate repair in bisyllabic words; and Bikol , Japanese, and Finnish speakers tend to initiate repair in multisyllabic words of three or more syllables.8 So, one obvious question to ask of the data is whether word length might play a part in site of initiation for the two types of repair. We explored this question by looking at site of initiation for words of varying length in each language. Our findings suggest that word length is, indeed, a factor in site of initiation and that cross-linguistic patterns underlie that outcome. Monosyllabic words and recognizable completion The first positive result that came from our study was that in all of the languages speakers favor “recognizably complete” initiation for recycling repair initiated in monosyllabic words. Tables 3.9 and 3.10 below provide the data for all seven languages.9 From Table 3.9 it is clear that in all the languages in the sample speakers favor initiation when the word is recognizably complete for recycling repairs in monosyllabic words, although with some variation in the strength of that preference. In four of the languages (Japanese , Sochiapam Chinantec , Indonesian, and Mandarin), speakers show a very strong preference, with well under 10 percent initiation before the word is recognizably complete. In the remaining three (Bikol , Finnish, and English) speakers show a strong preference, with percentages of initiation before the word is recognizably complete above 10 percent but below 20 percent.
82
Barbara Fox et al.
Table 3.10 Site of initiation of simple replacement repair initiated in monosyllabic words Not recognizably complete Indonesian Bikol Finnish Japanese Mandarin Sochiapam Chinantec English
0 (0%) 0 (0%) 0 (0%) 0 (0%) 2 (6%) 1 (8%) 17 (50%)
Recognizably complete
Total
0 (0%) 3 (100%) 4 (100%) 5 (100%) 28 (94%) 11 (92%)
0 3 4 5 30 12
17 (50%)
34
If we consider site of initiation in replacement repairs in monosyllabic words, we find a similar pattern for all languages except English (and Indonesian , which had no instances of replacement repair initiated in monosyllabic words). Note that in some cases it is not possible to determine the number of syllables of a word that gets replaced, since its lexical identity is not always clear. This proved to be a problem particularly for the English and Finnish data sets. Therefore, in this section of the paper, the data reported for English and Finnish includes only cases where word length could be determined (forty-six cases for English and twenty-six cases for Finnish). As shown in Table 3.10, in several of the languages in the study, those where speakers regularly use a large number of longer words in ordinary conversation, speakers rarely initiate simple replacement repairs in monosyllabic words. However, when they do, they follow a consistent pattern. With the exception of English (and Indonesian ), the preference for initiation when the word is recognizably complete in monosyllabic word replacement ranges between 90 percent and 100 percent. Frequencies are too low for reliable statistical analysis, but the difference between English and all the other languages in the study is noticeable, and we expect that larger data sets would bear it out. If we sum the results from the languages besides English and Indonesian , there are three instances of initiation prior to recognizable completion and fifty instances after recognizable completion. It is possible to compare these totals with English, and the result is highly significant (chi‑square = 23.47,
Initiation in same-turn self-repair
83
d.f. = 1, p = .0000). The difference between English and all the other languages in the study in this regard will be taken up in “Explaining English” below. We have thus uncovered our first universal pattern: Recycling repair is initiated when the word is recognizably complete in monosyllabic words. Replacement repairs also tend to be initiated when the word is recognizably complete in monosyllabic words, except in English. This is surprising, as our understanding of the interactional functions of replacements would lead us to expect that other languages would be more like English in initiating repair more often before recognizable completion. However, we will argue later (“Explaining English”) that there are reasons why we see this difference, reasons that are compatible with out understanding of interactional functions of repair. We will suggest that monosyllabic words are typically function words in many languages, and therefore pattern differently than one might expect (for content words), as they are less likely to be interactionally delicate. Multisyllabic words and recognizable completion If we compare the pattern of repair initiation in monosyllabic words to the patterns of repair initiation in multisyllabic words, we come to another important finding. Consider Table 3.11, which provides the frequencies for site of initiation in recycling repair in words of three syllables or more: We can notice first that the numbers for Sochiapam Chinantec , English, and Mandarin are very small, because in those languages speakers tend to use mostly words of one and two syllables in ordinary conversation. In fact, the totals for Mandarin and Sochiapam Chinantec are so low that one really cannot talk about tendencies in any meaningful way. Therefore, the discussion here will be restricted to the five languages that have ten or more multisyllabic words in which repair is initiated. In all languages except Indonesian (46 percent), speakers show a preference for initiation before the word is recognizably complete in these longer words, with Bikol , Finnish, and Japanese speakers showing the greatest preference for such initiation (between 80 percent and 90 percent). If we compare Table 3.11 to Table 3.10, we see that in each of the languages for which we have sufficient numbers to make a
84
Barbara Fox et al.
comparison, speakers show a dramatic difference in preferred site of repair initiation between multisyllabic words and monosyllabic words, with a higher frequency of initiations before the word is recognizably complete in multisyllabic words. Table 3.12 provides a similar set of figures for replacement repairs in multisyllabic words. For each language, we give the numbers for site of initiation in replacement repair in words of three syllables or more. Table 3.11 Site of initiation in recycling repairs in words of three or more syllables
Mandarin Indonesian English Finnish Japanese Bikol Sochiapam Chinantec
Not recognizably complete
Recognizably complete
Total
1 (33%) 12 (46%) 7 (70%) 24 (77%) 80 (87%) 61 (91%) 1 (100%)
2 (67%) 14 (54%) 3 (30%) 7 (23%) 12 (13%) 6 (9%) 0 (0%)
3 26 10 31 92 67 1
Japanese/Finnish: chi-square = 1.60, d.f. =1, p= 0.2 Finnish/English: chi-square = 0.23, d.f. = 1, p = .63 English/Japanese: chi-square = 2.07, d.f. = 1, p = .1505 Japanese/Bikol: chi-square = 0.65, d.f. = 1, p = .4218 Indonesian/Bikol: chi-square = 22.36, d.f. = 1, p = .0000
Table 3.12 Site of initiation in replacement repairs in words of three or more syllables
Bikol Finnish Indonesian Finnish Japanese English Sochiapam Chinantec Mandarin
Not recognizably complete
Recognizably complete
Total
8 (62%) 7 (64%) 10 (83%) 27 (87%) 31 (89%) 4 (100%) 0 (0%)
5 (38%) 4 (36%) 2 (17%) 4 (13%) 4 (11%) 0 (0%) 0 (0%)
13 11 12 31 35 4 0
0 (0%)
1 (100%)
1
Initiation in same-turn self-repair
85
When we consider site of replacement repairs, we find that in most of the languages in our study speakers show a preference for initiation before the word is recognizably complete in multisyllabic words, as illustrated in Table 3.12. Sochiapam Chinantec and Mandarin lack significant numbers of multisyllabic words and are thus not relevant to this question. In all the remaining languages, speakers most frequently initiate repair in multisyllabic words prior to recognizable completion. It is likely that this finding relates to the fact that multisyllabic words are generally content words and thus more likely to be interactionally delicate or potentially inappropriate.10 We now have two important findings: repair is initiated when the word is recognizably complete in monosyllabic words and is initiated before the word is recognizably complete more frequently in multisyllabic words than in monosyllabic ones. We are thus in a position to suggest that if speakers of a language favor monosyllabic words as the site of repair initiation, they will tend to initiate repair when the word is recognizably complete. If speakers of a language favor multisyllabic words, they will tend to initiate repair before the word is recognizably complete. If we look at the languages through this perspective, we will see that the tendencies do hold up, although there are some interesting exceptions. Word length, repair type and recognizable completion In this section we further explore the correlation between repair type and word length in the languages of our study. Tables 3.12 and 3.13 give the percentage of monosyllabic and multisyllabic words in both repairs types for each language. Table 3.12 provides the results for monosyllabic words. From Table 3.13, we can see that in Sochiapam Chinantec, English, and Mandarin , speakers typically initiate repair in onesyllable words, in both kinds of repair. In Bikol , Finnish, and Indonesian, about one-third to one-fifth of recycling repairs are initiated in one-syllable words, and in Bikol and Finnish replacement repairs, an even lower proportion is found (Indonesian has no monosyllabic words in which replacement repairs are initiated). Japanese is unique in having a very low rate of one-mora words in which recycling repairs are initiated; however, this may
86
Barbara Fox et al.
Table 3.13 Frequency of monosyllabic words for each repair type Recycle
Japanese Indonesian Bikol Finnish Sochiapam Chinantec Mandarin English
Replace
N
%
N
%
8 25 50 37 141
5 21 30 32 76
5 0 3 4 12
9 0 13 9 75
91 284
79 87
30 34
86 74
Table 3.14 Frequency of multisyllabic words for each repair type Recycle
Sochiapam Chinantec Mandarin English Indonesian Finnish Bikol Japanese
Replace
N
%
N
%
1
1
0
0
3 10 26 31 67 92
3 4 22 27 41 63
1 4 12 31 13 35
3 9 41 67 56 66
relate to the fact that one-mora words are somewhat rare in Japanese. Moreover, one-mora words are mostly postpositions (e.g., ni, ga, de, no), and because postpositions follow rather than precede their nouns, they are generally not used to delay next content word due. They thus have a limited functionality in recycling repairs. Turning now to multisyllabic words, we see a picture that is complementary to the one in Table 3.13. From Table 3.14 we can see that in Sochiapam Chinantec , Mandarin, and English speakers do not often initiate either recycling or replacement repair in words of three or more syllables. Indonesian and Finnish have a somewhat higher percentage of recycling repairs initiated in long words and a
Initiation in same-turn self-repair
87
substantially higher percentage of replacement repairs; the increase is particularly striking for Finnish, which actually has the highest percentage of multisyllabic words in which repair is initiated in the languages of our sample. Japanese and Bikol have a fairly high percentage of repair initiation in longer words for both recycle and replacement repairs. If we take these findings back to our original question of site of initiation, it seems that we can now explain some of the basic patterns. Sochiapam Chinantec and Mandarin speakers initiate repair mainly in monosyllabic words, and, since there appears to be a universal tendency toward initiation when the word is recognizably complete when repair is initiated in monosyllabic words, these two languages show an overwhelming prefer ence for initiation at recognizable completion for both kinds of repair. Finnish and Bikol have a moderate number of monosyllabic words in which recycling repairs are initiated and a low number of monosyllabic words in which replacement repairs are initiated. Finnish has a moderate rate of multisyllabic words in which recycling repairs are initiated and a quite high rate of multisyllabic words in which replacement repairs are initiated. Bikol has a moderately high frequency of multisyllabic words in which repair is initiated for both types of repair. We would thus expect an increased frequency of initiation before the word is recognizably complete, when compared to Sochiapam Chinantec and Mandarin , because repairs in multisyllabic words tend to be initiated before the word is recognizably complete, and both Bikol and Finnish have a greater proportion of multisyllabic words than do Sochiapam Chinantec and Mandarin. Both languages match this prediction in having an increase in initiation before recognizable completion for both repair types. Japanese is unusual in having an extremely low rate of monomoraic words as sites of initiation for both kinds of repair, and a fairly high rate of multimoraic words as sites of initiation for both kinds of repair. We would thus expect a strong preference for early initiation for both kinds of repair, since repairs in longer words tend to be initiated before recognizable completion, and this is what we find. In Indonesian , speakers initiate recycling repairs in monosyllabic words with only moderate frequency, and no replacement repairs at all are initiated in monosyllabic words. Recycling repairs are
88
Barbara Fox et al.
initiated in multisyllabic words with a moderately high frequency. Indonesian thus clearly does not pattern with English, Mandarin, and Sochiapam Chinantec in terms of the length of word in which repair is most often initiated. Nor does it pattern with Japanese , as the frequency of repair initiation in multisyllabic words is considerably lower in Indonesian than in Japanese. Indonesian is more like Finnish or Bikol , as all three languages have only moderate numbers of monosyllabic words in which repair is initiated and the frequency of multisyllabic words in which repair is initiated in Indonesian is closest to their frequency in Finnish for recycling repairs and in Bikol for replacement repairs. Indonesian might thus be expected to pattern with Finnish and Bikol and to have an increase in initiation before recognizable completion for both repair types compared with Mandarin and Sochiapan Chinantec. However, the Indonesian data does not show the same prefer ence for site of initiation for the two repair types. Instead, Indonesian speakers prefer initiation after recognizable completion for recycling repairs and initiation prior to recognizable completion for replacement repairs. Indonesian speakers’ preference for initiation prior to recognizable completion in replacement repairs fits our prediction, since longer words show a preference for initiation before recognizable completion, but their preference for initiation after recognizable completion for recycling repairs does not (they should show a greater preference for early initiation even in recycling repairs because repairs are initiated in multisyllabic words with moderate frequency). Indonesian thus appears to be an exception to our pattern; below we suggest that there is a competing cross-linguistic factor at work in Indonesian which produces this result. English is another exception to the findings regarding word length. A preponderance of English recycling repairs are initiated in monosyllabic words, so we would expect English speakers to favor initiation after recognizable completion for recycling repairs, which it does. However, as noted above, English speakers favor initiation prior to recognizable completion for replacement repairs, even in monosyllabic words. We have already suggested that this is a response to interactional pressures, and we will further explore the difference between English and most of the other languages in our corpus below.
Initiation in same-turn self-repair
89
Table 3.15 Site of initiation of recycling repairs in bisyllabic words
Indonesian Mandarin Finnish Bikol English Sochiapam Chinantec Japanese
Not recognizably complete
Recognizably complete
Total
17 (26%) 6 (29%) 20 (43%) 20 (44%) 12 (44%) 21 (49%)
49 (74%) 15 (69%) 27 (57%) 25 (56%) 15 (56%) 22 (51%)
66 21 47 45 27 43
26 (55%)
21 (45%)
47
It appears that only five languages (Sochiapam Chinantec, Mandarin , Finnish, Bikol, and Japanese ) fit the predictions from our general findings about word length and site of repair initiation, while two (Indonesian and English) do not. We argue below that there are principled reasons for these two exceptions. Explaining Indonesian: Longer words and recognizable completion We have shown that Indonesian is exceptional in that, according to the word-length criterion, speakers should show a greater preference for initiation prior to recognizable completion in recycling repairs than they do. We propose to explain this exception in terms of word length and syntactic class. The careful reader will no doubt have noticed by now that we have discussed monosyllabic words and multisyllabic words but have ignored bisyllabic ones. The time has come to rectify this matter. Table 3.15 presents the results for bisyllabic words in recycling repairs. Due to the small number of replacement repairs initiated in bisyllabic words, we will not discuss them in this paper. We find two different patterns in our data. In Indonesian and Mandarin , speakers prefer initiation after recognizable completion in bisyllabic words. To put it another way, they seem to treat bisyllabic words much as they treat monosyllabic ones. In the remaining languages, speakers do not show a preference for site of initiation in bisyllabic words, suggesting that for speakers of those languages
90
Barbara Fox et al.
bisyllabic words are treated as intermediate between monosyllables, in which repair is typically initiated after recognizable completion and multisyllabic words, in which repair is typically initiated prior to recognizable completion. The preponderance of initiations after recognizable completion in bisyllabic words in Mandarin is entirely in line with the general pattern of initiation after recognizable completion in Mandarin.11 Indonesian is unusual in our sample in that repair is most often initiated in bisyllabic words, rather than monosyllabic or multisyllabic words (see Table 3.16). More than half (sixty-six or 56 percent) of simple recycling repairs in Indonesian are initiated in bisyllabic words, whereas the proportion for most of the other languages ranges from one fifth to one third, and for English is far lower (twenty-seven or 9 percent). Only Finnish approaches Indonesian in the frequency with which repair is initiated in bisyllabic words, with forty-seven (41 percent), but even in this case the difference between the two languages is statistically significant (chi-square 5.89, d.f. = 1, p = .0152). The fact that Indonesian does not fit our prediction, and that speakers of Indonesian should prefer repair initiation prior to recognizable completion more frequently than they do in recycling repairs appears to be traceable to the high frequency with which repair is initiated in bisyllabic words. Why should Indonesian speakers so frequently initiate recycling repairs in bisyllabic words? We believe the key lies in the nature of function words in Indonesian: Most function words in Indonesian are bisyllabic (including pronouns, prepositions , and determiners). Indonesian is like English in that function words precede content Table 3.16 Percentage of recycling repairs initiated in bisyllabic words
Indonesian Finnish Japanese Bikol Sochiapam Chinantec Mandarin English
Bisyllabic
All recycling
66 (56%) 47 (41%) 47 (32%) 45 (28%) 43 (23%) 21 (18%) 27 (9%)
117 116 147 162 185 115 285
Initiation in same-turn self-repair
91
words, so function words can be used in recycling repairs as they are in English, to delay next item due. This leads to a high rate of recycling of bisyllabic function words, which, like function words in English, generally show initiation after recognizable completion. Consider Table 3.17, which shows how the relative frequency with which recycling repairs are initiated in bisyllabic function words and in bisyllabic content words in each language: If we compare the number of bisyllabic function words and content words in which recycling repairs are initiated in the languages of our sample, we see that it is indeed the case that Indonesian speakers initiate repair far more often in bisyllabic function words than speakers of most of the other languages. Only Finnish has a higher percentage, and Finnish was the other language with an unusually high percentage of bisyllabic words in which recycling repair was initiated (see Table 3.16). This aspect of Finnish repair will be discussed further below. We now consider the relative frequency of repair initiation in multisyllabic function and content words in the languages in our sample. If we look at Table 3.18 below, we see, first of all, that Mandarin and Sochiapan Chinantec speakers initiate repair in multisyllabic words far too infrequently for data from these languages to be relevant to this question. Of the other languages, Japanese , Bikol, and Finnish speakers are much more likely to initiate repair in multisyllabic content words than in multisyllabic function words. In Indonesian and English, on the other hand, speakers initiate repair in multisyllabic content words and function words with almost equal frequency. Table 3.17 Recycling repairs: Bisyllabic words by syntactic class
Finnish Indonesian Bikol English Sochiapam Chinantec Japanese Mandarin
Function
Content
Total
37 (79%) 42 (64%) 18 (40%) 10 (37%) 14 (33%)
10 (21%) 24 (36%) 27 (60%) 17 (63%) 29 (67%)
47 66 45 27 43
15 (32%) 5 (24%)
32 (68%) 16 (76%)
47 21
92
Barbara Fox et al.
Just as with bisyllabic words so too with multisyllabic; Indonesian has an unusually high number of multisyllabic function words compared with most of the other languages in our sample. In fact, Indonesian by itself accounts for 31 percent of the multisyllabic function words in our data, although it only constitutes 11 percent of the multisyllabic words in which recycling repair is initiated. We can now see that the unusual finding that Indonesian speakers initiate recycling repairs after recognizable completion in longer words can be explained by the tendency to recycle function words to delay the next content word due. Explaining English: Syntactic class and recognizable completion So we are left with English as the language which cannot be fully predicted by our cross-linguistic findings about the correlation between word length and repair initiation. English speakers are unusually likely to initiate replacement repairs prior to recognizable completion, given that 62 percent of replacement repairs in English are of monosyllabic words, which would predict a tendency toward initiation after recognizable completion. We believe that English speakers are less attentive to length in replacements than they are to syntactic class; that is, in replacement repairs, English speakers far more frequently engage in initiation prior to recognizable completion whenever the word is a content word (adjective, noun or verb) and make use of initiation after recognizable completion whenever the word is a function word.
Table 3.18 Recycling repairs: Multisyllabic words by syntactic class
Japanese Bikol Finnish Indonesian English Mandarin Sochiapan Chinantec
Function
Content
Total
6 (7%) 5 (7%) 9 (29%) 11 (42%) 4 (40%) 0 (0%) 0 (0%)
86 (93%) 62 (93%) 22 (71%) 15 (58%) 6 (60%) 3 (100%) 1 (100%)
92 67 31 26 10 3 1
Initiation in same-turn self-repair
93
Table 3.19 gives the relevant data for replacement repairs initiated in monosyllabic words. The majority of replacement repairs initiated in monosyllabic function words are initiated after recognizable completion, and the majority of replacement repairs initiated in monosyllabic content words are initiated prior to recognizable completion. It seems likely that this difference relates to different interactional implications of function words and content words, where content words are more likely to be interactionally sensitive or inappropriate. It would be desirable to compare English with other languages; however, for many of our languages the number of monosyllabic words in which replacement repairs are initiated is so small that the data cannot be considered representative, and any patterns they might seem to show must be treated with caution. Only Chinantec and Mandarin have more than five exemplars. Nonetheless, we have provided data for all five languages below in Table 3.20 (Indonesian has no replacements initiated in monosyllabic words).12 It appears that in Japanese, Finnish and Bikol, speakers show a tendency that is familiar to us by now: Monosyllabic function words are initiated after recognizable completion. Although the numbers are too small to be anything but suggestive, it looks as though speakers of these languages behave like speakers of English. Of course, unlike these three languages, the English corpus shows a large number of replacement repairs initiated in monosyllabic content words, a fact which makes English look, on the surface, quite different. However, in both Sochiapam Chinantec and Mandarin , repairs in both function words and content words are generally initiated after recognizable completion. In other words, in neither language do speakers seem to be influenced by word type, which is Table 3.19 English replacement repairs by word type for monosyllabic words
Function words Content words Total
Not recognizably complete
Recognizably complete
Total
3 (18%) 14 (82%) 17 (50%)
14 (82%) 3 (18%) 17 (50%)
17 17 34
94
Barbara Fox et al.
not surprising, given that we have seen that speakers of these two languages consistently favor initiation after recognizable completion. It is unclear to us at this point why English speakers (and probably speakers of other languages such as Japanese , Finnish and Bikol ) seem to orient to syntactic class, and, thus, to interactional pressures, while Mandarin and Sochiapam Chinantec speakers orient mainly to word length. Further data from other languages may provide a clue as to why speakers of these two languages are so consistent in their preference for initiation at recognizable completion. Both are tone languages; one possible avenue to explore is whether tonality may be relevant. A closer look at Bikol and Finnish: Word length and site of initiation Bikol and Finnish both require further discussion because, although they match our predictions about site of repair initiation and word length, in Finnish recycling and both types of repair in Bikol speakers show no statistically significant preference for one or another site of initiation when all the data is grouped together. Because both of these languages have many multisyllabic words in their lexicons, according to the word-length criterion, we would have expected speakers to show a clear preference for initiation prior to recognizable completion. The answer to this puzzle lies in the distribution of word lengths of those words in which repair is actually initiated. Table 3.20 Replacement repairs by word type for monosyllabic words
Sochiapam Chinantec Mandarin Japanese Finnish Bikol
Not recognizably complete
Recognizably complete
Function
Content
Function
1
0
5
6
12
0 0 0 0
0 0 0 0
9 5 2 3
18 0 2 0
27 5 4 3
Content
Total
Initiation in same-turn self-repair
95
Table 3.21 Word length in recycling repairs
Monosyllabic Bisyllabic Multisyllabic Total
Finnish
Bikol
38 (33%) 47 (41%) 31 (27%) 116
50 (30%) 45 (28%) 67 (41%) 162
In both Finnish and Bikol, as shown in Table 3.21, the distribution of site of repair initiation according to word length is fairly even. Repairs initiated in monosyllabic words in these languages tend to be initiated after recognizable completion, those initiated in bisyllabic roots show fairly even division between the two types of initiation, and those in multisyllabic roots are mainly initiated prior to recognizable completion. As a result, although each syllable type follows our predictions, the data for the languages as a whole shows no preference for site of initiation of repair. Other languages in the sample did not show this type of balance between word lengths. In Japanese , for example, almost no repairs are initiated in monosyllabic words (5 percent), while in English (87 percent), Sochiapam Chinantec (76 percent) and Mandarin (79 percent) most repairs are initiated in monosyllables, and in Indonesian , as noted above, repair is most frequently initiated in bisyllabic words (56 percent). Explaining monosyllabic repair initiation One of the central findings of the current study is that speakers of languages of many different genetic, typological, and areal affili ations initiate repair after recognizable completion in monosyllabic words. This striking finding calls out for an explanation. In this section we proffer several hypotheses and examine evidence that might help us choose among them. In our discussion we focus primarily on recycling repairs. Shortness vs beat of delay As Schegloff (1979) has suggested, recycling is done, basically, to delay next item due. For whatever specific interactional or cognitive
96
Barbara Fox et al.
reason a speaker recycles on a particular occasion, recycling always has the function of stopping progressivity of the turn-so-far and delaying next item due. One obvious explanation, then, for our finding regarding site of initiation in monosyllabic words is that speakers shape their repair to give themselves a beat of delay. If we can say that one syllable gives roughly one conversational beat of delay, then initiating repair at the end of a syllable and repeating that syllable provides the speaker with one conversational beat of delay. Of course, speakers may also lengthen the syllable in which repair is initiated and may also allow silence to develop after the repair initiation, thereby increasing the temporal delay beyond a single beat; nonetheless, repeating a single syllable gives at least a minimum of one beat of delay. This proposal finds support in research regarding the management of overlap. The majority of overlaps are resolved in one to two beats (Schegloff 2000c, 2002b),13 and one common device used to provide those one or two beats is turn-beginning recycling (Schegloff 1987c), in which the speaker recycles the first one or two syllables of the turn to secure a turn in the clear. This fact provides some support for our claim that many interactional difficulties can be resolved within one to two conversational beats and thus that recycling monosyllabic words may be used to provide the speaker with one beat of delay. Moreover, in languages where function words precede content words, speakers have highly entrenched recurrent practices for delaying next content word due, practices which involve the recycling of high-frequency monosyllabic function words. For example, English speakers have as highly entrenched recycling devices “I- I,” “the- the,” “what- what,” “that’s- that’s,” “it’s- it’s” (but they do not have a comparable recurrent practice of, for example, “receivereceive”). These devices provide the speaker with one beat of delay just before a content word is due. According to this explanation for our finding, it is beneficial for a speaker to initiate repair in such a way as to create one (or possibly two) extra conversational beats. In other words, a speaker may choose to initiate recycling repair after recognizable completion in a monosyllabic word, in order to achieve just this amount of delay. With longer words, recycling of only the first one or two
Initiation in same-turn self-repair
97
syllables of the word will produce the same kind of delay, thus allowing speakers to initiate repair earlier in multisyllabic words. But a second possible explanation is also plausible. Monosyllabic words are short, often 200 milliseconds or less, and short words pose challenges to speakers who would like to initiate repair: Will the speaker decide to initiate repair in time to do so before the word reaches recognizable completion? If the speaker decides to initiate repair a bit late, it might not be possible to actually initiate the repair before the word is recognizably complete. The recycling still accomplishes delay, but the repair is initiated after recognizable completion because the speaker didn’t recognize a need for delay until the last moment. We thus have two clear possible explanations for our finding. We recognize that in any given case it is impossible to know when or why the speaker decided to initiate repair. However, when we look at the aggregate, we may be able to discover patterns which will provide clues as to what types of pressures speakers respond to most frequently. In the next section we explore three pieces of evidence that address this issue. Three sources of evidence The first piece of evidence comes from comparing sites of initiation for content words and function words in our languages. We have suggested that function words are most likely to be used to delay the next content word due and, therefore, to provide speakers with an extremely useful device. We would expect that content words would not be recycled with the same frequency as function words, because, while they may do important work in providing occasional delay for gaze organization, overlap resolution, etc., such functions are less frequent than the need to delay next content word due. In fact, in our sample of seven languages, in five of them speakers range from moderately to highly more likely to initiate repair in a function word than a content word. The two exceptions are Japanese , where function words follow content words and thus are less likely to be used to delay the next content word due, and Bikol , where one third of simple recycling repairs are initiated in verbs, and which has a rich system of verb prefixes which can be used to delay verbs.
98
Barbara Fox et al.
A related piece of evidence comes from Indonesian . In Indonesian, most function words are bisyllabic. In spite of a cross-linguistic tendency for lack of preference in site of initiation in bisyllabic words, Indonesian speakers prefer initiation after recognizable completion for bisyllabic words, possibly because these tend to be function words that are used to delay the next content word due. This fact provides evidence that speakers indeed use function words to create a beat (or two) of delay. English offers additional support for the idea that speakers use recycling to create a beat (or two) of delay. First, the vast majority of English recycling repairs (80 percent) are initiated in function words, which would be a surprising finding if there weren’t some natural connection between function words and recycling, such as the one we have suggested of delaying next content word due. Second, even when content words are recycled, there is often evidence that they are being used to provide a beat of delay. Of the thirty-four monosyllabic content words in our collection, twenty-one were initiated after recognizable completion. Of those, the majority appear to be designed to provide delay in response to interactional pressures. Twelve are in overlap and attempt to bring the speaker into the clear. Another three occur in word searches within compound or complex noun phrases, where they are delaying the next noun due, and two are recyclings of the phrase “he goes,” which delays an upcoming clause of reported speech . It therefore appears that English speakers often recycle complete words to provide a beat of delay, rather than because they have decided too late to initiate repair. In addition, we have seen evidence from replacement repairs that English speakers are perfectly capable of initiating repair before recognizable completion in monosyllabic content words. One could argue that monosyllabic function words are even shorter than monosyllabic content words, due to frequency-induced phonological reduction, thus allowing for even less time to initiate repair; while this is no doubt true, the evidence from English still suggests that shortness alone is not the crucial factor in site of initiation in monosyllabic words. We have thus found several facts to support the hypothesis that speakers initiate repair after recognizable completion in monosyllabic words in order to provide a beat of delay. While it is impossible to prove, or disprove, a claim that in some cases speakers
Initiation in same-turn self-repair
99
don’t decide to initiate repair until they have reached the end of a word, and it is likely that this type of situation occurs, both when speakers initiate repair in a monosyllabic word and in a longer word, it is unlikely to be the explanation for the frequency with which we find repair initiation after recognizable completion in monosyllables. However, Mandarin and Sochiapam Chinantec provide evidence that for some languages, neither of our explanations is appropriate. In these languages, repair is consistently initiated after recognizable completion in monosyllabic words, both content and function, and most repair is initiated in monosyllabic words. It seems that for these two languages speakers have a recurrent practice of initiating repair after recognizable completion, a practice that is only rarely overridden. We are not sure why this should be the case. Given that the two languages in question are both tonal, it may relate to tonality. Or it could be a kind of grammaticalization of a highly useful practice. Further research is needed to understand this issue more deeply. Summary and conclusion This is the first study of site of repair initiation from a crosslinguistic perspective that we know of. We began this study by noticing that there is a wide range of variation in relative frequency of site of initiation across the languages in our sample. We then asked the question: Is this variation just random variation, or do universal principles manifesting through language-specific means create the variation? What we have proposed in this study is that there are indeed universal principles at work in shaping site of initiation patterns cross-linguistically. First, we noticed that different repair types can have very different patterns of initiation, probably relating to differences in interactional function. Second, we noticed that length of word is one factor that correlates with site of initiation. We found that in both recycling and replacement repairs monosyllabic words tend to be repaired after recognizable completion in all the languages of our study; furthermore, multisyllabic words tend to be repaired prior to recognizable completion in all but one of the languages that have enough
100
Barbara Fox et al.
multisyllabic words to show a pattern. Indonesian is the only exception; in this language, speakers show no preference for initiation before recognizable completion in multisyllabic words (see Table 3.11). We also observed that in most languages in our sample speakers show no preference for site of initiation in bisyllabic words. In Indonesian , however, which has an unusually high frequency of repair initiated in bisyllabic words, speakers strongly prefer initiation after recognizable completion. This means that in languages in which speakers initiate repair mainly in monosyllabic words (such as Sochiapam Chinantec , Mandarin, and English), they tend toward initiation after recognizable completion, while in languages in which speakers prefer initiation in multisyllabic words (such as Japanese ), they tend toward initiation prior to recognizable completion. On the other hand, if speakers of a language shows no preference as to the length of the words in which they initiate repair (such as in Bikol and Finnish), they will also show no overall preference for a single site of repair initiation, even though for each of the three categories of word length (monosyllabic, bisyllabic, and multisyllabic) they follow the patterns we have observed. Third, we noticed that replacement repair tends to be initiated prior to recognizable completion in four of our seven languages (English, Indonesian , Japanese, and Finnish ). Sochiapam Chinantec and Mandarin , in which speakers almost always initiate repair after recognizable completion, do not participate in this pattern, nor does Bikol, where speakers show no preference. We offered several possible explanations for these cross-linguistic patterns. First, we suggested that speakers tend to initiate repair after recognizable completion in monosyllabic words because of late decisions to initiate repair. Since speakers may not decide to initiate repair until they have gotten partway through a word, if those words are monosyllabic, then by the time the speaker chooses to initiate repair, the word is recognizably complete. It is worth noting that high-frequency function words are often somewhat and sometimes radically phonologically reduced, at least in English (Jurafsky et al. 1998). Phonological reduction might make it more difficult to initiate repair prior to recognizable completion. Second, we explored the possibility that speakers initiate repair after recognizable completion in monosyllabic words because they want to create a beat of delay. We suggested that there may be a tendency for monosyllabic words to be function words, and that
Initiation in same-turn self-repair
101
function words tend to be recycled to delay next content word due; in such cases – in languages for which function words precede content words – speakers may want a beat (or two) of delay, since one of the main purposes of recycling in these instances is to provide a temporal delay. This same reasoning could explain why in Indonesian repair tends to be initiated in bisyllabic words after recognizable completion, since in Indonesian function words are frequently bisyllabic. Clearly such reasoning does not apply to Japanese , since in Japanese function words follow content words. Japanese has developed other techniques for accomplishing delay (Fox et al. 1996). We proposed that while either of these explanations could be the relevant one in any given case, the overall patterns of repair initiation that we have observed suggest that speakers frequently recycle monosyllabic words to achieve a beat of delay, and that repair initiation after recognizable completion is a recurrent practice. However, Mandarin and Sochiapam Chinantec , where speakers have a consistent practice of initiating repair after recognizable completion in all word-lengths, under almost all circumstances, seemed to suggest the possibility that in some languages speakers may have generalized the pattern to be a fixed practice. We also suggested that multisyllabic words provide many opportunities to initiate repair before coming to the recognizable completion of the word. It may be that in most cases of repair initiation, the speaker needs only one or two extra conversational beats, and with multisyllabic words a speaker doesn’t need to produce the whole word to get those beats. We further suggested that speakers favor early initiation in replacement repairs in several languages just because replacements are generally of content words, which are generally longer in most languages in our study. Additionally, speakers may choose to initiate repair sooner so as not to be held accountable for alternate formulations or so as not to produce inappropriate terms. In Mandarin and Sochiapam Chinantec this tendency is not expressed. We hope to have demonstrated that site of repair initiation can be a window onto differences across languages, as well as onto universal principles of repair organization. We also hope to have provided evidence that sometimes differences across languages mask the universal interactional patterns that organize them, and that research on cross-linguistic differences also benefits from a search for those universal patterns (cf. Sidnell 2007c).
102
Barbara Fox et al.
Acknowledgments We would like to thank Vincent Sarich for statistical and editorial assistance, Emanuel A. Schegloff for clarifying comments and for providing the original inspiration for this study, and Jack Sidnell for inviting us to participate in this volume and for insightful comments on an earlier draft. Funding was provided by NSF grant number BCS0406512 . Authorship Barbara Fox analyzed the English data. Fay Wouk analyzed the Indonesian data. Makoto Hayashi analyzed the Japanese data. Steven Fincke analyzed the Bikol data. Liang Tao analyzed the Mandarin data. Marja-Leena Sorjonen and Minna Laakso analyzed the Finnish data. Wilfrido Flores Hernandez analyzed the Sochiapam Chinantec data. Barbara Fox and Fay Wouk wrote the majority of the paper with data and input provided by the other co-authors. Notes 1 Replacements and deletions differ in that in replacements the word type remains although the token has been replaced, while in deletions the word type is eliminated. 2 This has been amply demonstrated for English by Jasperson (1998). 3 This might seem to be a substantial change from Schegloff ’s proposal, since the category post-beginning has been expanded from including just the first sound of the word to including all but the very last sound. However, in actual fact, at least for English the difference is not as great as it sounds. It turns out that for English (the language on which Schegloff based his original proposal), initiation after the first syllable but prior to recognizable completion is an extremely rare occurrence; in our corpus of 339 simple repairs, only eight were initiated in these in-between locations. The same is true of Sochiapam Chinantec and Mandarin . It’s only when we turn to languages with large numbers of multisyllabic words that the middle category becomes robust. 4 By inappropriate we do not mean offensive to the recipient; we simply mean that it is not the word that the speaker finds best suited to the context. 5 And in some cases the speaker may wish to be accountable for an inappropriate or offensive term. 6 For Japanese, the relevant category is actually the mora, not the syllable. All Japanese counts in the following discussion represent mora
Initiation in same-turn self-repair
7
8 9
10
11
12
13
103
rather than syllable counts. However, for simplicity of exposition, we will use the term “syllable” throughout when referring to all languages in the study. The term mora will only be used in discussions restricted to Japanese. All three of these languages have a range of word lengths; 80 percent of words in the modern Mandarin lexicon are compound words, thus usually bisyllabic; Sochiapam Chinantec has mainly monosyllabic roots but makes use of a wide range of prefixes to form bisyllabic words; the English lexicon clearly has many words of two or more syllables. However, according to our findings, in these languages repair is initiated primarily in monosyllabic words. Studies of the frequency of different word lengths in everyday conversation in these languages might shed light on this discrepancy. These differences are no doubt in part a reflection of the average length of words in general the different languages. Word length in Mandarin was determined as follows. If the syllable (where repair is initiated) and its following syllable(s) form a tight unit in that they cannot be further divided into sub-components (Adj-N), then it is a bisyllabic or multisyllabic word. If the syllable and its following syllable(s) form a relatively loose unit that allows insertion of grammatical elements into it (i.e., the site of initiation is a syllable that has a meaning of its own to contribute to the whole utterance), then this syllable is a monosyllabic word. Because content words are open class, there are a larger number of potential candidates in any given context than there are for function words. In addition, content words are generally of lower frequency than are function words. Thus speakers face a greater challenge in selecting the appropriate term. In fact, it is more puzzling that Sochiapam Chinantec speakers do not show this preference than that Mandarin speakers do, given the many similarities we have seen between these two languages where repair is usually initiated in monosyllabic words. It would appear that the difference relates to the presence of verbal prefixes in Sochiapam Chinantec and the frequency with which verbs are broken off after the prefix but before the root as a way of delaying the next verb due. However, a complete discussion of this matter is beyond the scope of this paper. Although Mandarin has thirty simple replacement repairs, three had to be excluded from this analysis, as the syntactic class of the word that was replaced was ambiguous. Resolution in the first two beats is done without recourse to techniques of competitive production, while those that go over two beats typically involve competitive production (Schegloff 2000c, 2002b).
4 Repairing reference Maria Egbert, Andrea Golato, and Jeffrey D. Robinson
Introduction Languages provide speakers with resources for both referencing and repairing.1 While referencing is crucial for sustaining intersubjectivity, repairing is crucial for re-establishing intersubjectivity (Auer 1984; Schegloff 1992a; Schegloff et al. 1977). This chapter describes a repair-initiation action that claims that a particular aspect of the prior turn presents “trouble” for its speaker (e.g., understanding the prior turn; for a review of “repair” and its associated terminology, see Schegloff 1992a). The trouble source is treated as being a pro-term/indexical expression (e.g., “it”, “this”, “that”). The repair operation – that is, the interactional move that is performed by the speaker of the trouble source in response to the repair-initiation action being investigated, and that is designed to deal with/resolve the trouble – involves producing a full-reference form/full noun phrase (NP) (e.g., “the folder”, “the token”, “the ticket”) that is relevantly associated with the trouble-source indexical expression. Given that the trouble-source speaker orients to the trouble source as being an indexical expression, and given that the trouble-source speaker resolves the trouble by producing a recipient-designed full-reference form, the repair-related trouble can be characterized as a referent that is (claimed to be) underspecified for the person initiating repair. Languages provide for different types of referents, such as people, places, times, and things. This chapter focuses on a comparative analysis of one repair-initiation action in English and German that always (in our data) is taken by troublesource speakers as targeting underspecified “thing”-referents. The repair-initiation action being examined is implemented by “Was denn” or “Was.” in German and “What.” in English. (For initial
Repairing reference
105
observations, see Schegloff 1997b; Schegloff et al. 1977; Selting 1987a, 1987b, 1987c, 1988, 1992). This repair-initiation action can be characterized as seeking specification of a claimedly underspecified referent. This chapter is organized as follows: After a brief description of the data and transcriptions, the analysis is presented in five steps. First, we show how final intonation plays an essential role in differentiating “Was denn”/“Was.”/“What.” from open-class repair initiators “Was?”/“What?” delivered with rising intonation (Drew 1997; Schegloff et al. 1977). Second, we analyze cases in both German and English to demonstrate that “Was denn”/“Was.”/“What.” implement a repair-initiation action that is produced to seek the specification of a claimedly underspecified referent. The third section provides additional evidence that “Was denn”/“Was.”/“What.” claim trouble with the specificity of an indexical expression by examining repair-initiation turns that contain two succeeding repair-initiation attempts (i.e. doubles ; Schegloff et al. 1977). The fourth section contrasts “Was denn”/“Was.”/“What.” with a different but allied repair-initiation action implemented by “Für was.”/“For what.”, which receives, as a repair operation, an increment (i.e. an extension) rather than a specification of a referring expression in the trouble-source unit. Finally, the fifth section discusses the commonalities among, and differences between, “Was denn”/“Was.”/“What.” in terms of range of usage and function. In this section, we also discuss one non-repair-initiation action that can be implemented by “Was.”/“What.”, which is a “go-ahead” to a pre-announcement action. This chapter closes with a discussion of method ological issues related to a comparative analysis.
Data The repair-initiation implemented by “Was denn”/“Was.”/“What.” appears to be a relatively rare action in conversation. The German data consist of thirty hours of mundane conversation (eighteen hours of videotaped face-to-face interaction and twelve hours of audio-taped telephone conversation), including two-party and multi-party interaction with speakers from different dialectal regions within Germany. These data produced fifteen instances of “Was denn” and six instances of “Was.” used by others to initiate repair.
106
Maria Egbert et al.
The American English data consist of sixty-five hours of mundane conversation (sixty-two hours of audio-taped telephone conversation and three hours of videotaped face-to-face interaction), which yielded a total of seventeen instances of “What.” used by others to initiate repair. The aforementioned numbers do not include instances of “Was denn”/“What.” (“Was.” is not found) used as “go‑ahead” responses to pre-sequence-initiating actions, i.e. instances that do not initiate repair (Schegloff 2007b ). All data were transcribed using the transcription system developed by Gail Jefferson (Atkinson and Heritage 1984). As described below, intonation is a constitutive feature of “Was denn/“Was.”/“What.”. Unit-final intonation is represented with a period for falling intonation and a comma for slightly rising intonation. In the text, German “Was.” and English “What.” are referred to with periods to reiterate their final-intonation contours. For the German transcripts, each utterance is represented by three lines of transcription: The first line presents the original utterance; the second consists of a word-for‑word gloss; and the third provides an idiomatic English translation. When the idiomatic translation corresponds to the gloss, only the idiomatic translation is given. (For reasons to be discussed, “Was denn” is idiomatically translated as “What.”). Sound files were analyzed with the acoustic software Praat (Boersma and Weenink 2006). Final-intonation as a constitutive feature of “Was denn”/“Was.”/“What.” Final-intonation contour is part of action formatting (CouperKuhlen and Ford 2004; Couper-Kuhlen and Selting 1996; Golato and Fagyal 2008; Schegloff 1998a; Selting 1995) and plays an important role in conveying the precise function of practices of other-initiated repair (Schegloff 1997b; Selting 1987a, 1987b, 1987c, 1988, 1992). The lexical items “Was denn”, “Was”, and “What”, which are used to implement the practice under investigation, can also be used to implement a different repair-initiation practice: open-class (e.g., “What”). However, in contrast to the practice described in this paper, open-class repair initiators are regularly produced with final-rising intonation, such as “Was?”/“What?” (Drew 1997; Schegloff 1997b; Selting 1987a, 1987b, 1987c, 1988,
Repairing reference
107
1992). Furthermore, open-class -repair initiators do not specify what part or aspect of a trouble-source unit is problematic, i.e. they leave “open” the type of trouble in play (Drew 1997; Schegloff et al. 1977). Although repair operations in response to open-class repair initiators can vary (e.g., Schegloff 2004), they frequently take the form of repeats of the trouble source. The following two examples show such a repair operation in response to German “Was?” and English “What?” (1) Konferenz [Kirsten:1:16:24] 01 K: .hhh Du [ich wollt dich was-] .hhh You [I wanted you something] .hhh You [I wanted to (ask) you ] [ ] 02 → R:
[ [
Wie lang wars]ten da? How long were yo]u there?
03 K: Wa? Wha’? 04 (0.2) 05 K: .h Ich war nur [ä:hm .h I was only [u:hm [ 06 → R:
] ] ]
[Wie lang ] warstn>da< [How long ] were you >there<
07 K: Nur- bisschen weniger als ne woche. Only- little less than a week. (2) IMPOSSIBLE [Marcia and Joe] 01 M: 02→ J: 03 04 M: 05→ J: 06 07 M:
He’s just impo:ssible. Did juh tell ’im that? (0.4) Wha:t? D’=juh tell ’im that? (1.4) No:
In response to “Wa?” and “What?”, Rita and Joe respectively produce verbatim repeats of the entirety of their trouble-source units (compare lines 6 and 5, respectively, with line 2 in both cases). There are occasions when open-class repair initiators are treated as
108
Maria Egbert et al.
dealing with reference trouble, as evidenced by responses that are “replacements of ‘specialized’ reference terms – whether technical ones, recognitional ones, uncommon ones, or others – by ones designed to be more accessible, better recipient designed, and so forth” (Schegloff 2004: 96–98). For example, see Extract 3, supplied by Schegloff (2004: 97). (3) SPC, NYE, 6 (NTRI #130b)] 01→ 02 03→ 04
A: Do you have some church affiliation, now? B: What? A: Do you belong to a church now? B: No I went to church I haven’t been …
In contrast, “Was denn”/“Was.”/“What.” are not delivered with strong final-rising intonation. Furthermore, as shown below, “Was denn”/“Was.”/“What.” always (at least in our data) target reference trouble of a particular type, i.e. underspecification.
0.254349886
Pitch (Hz)
500
75
0
Time (s)
wa?
Figure 4.1 Open-class repair German “Wa?”
0.5108
Repairing reference
109
The intonation contour of “Was denn” ranges from final-falling to fi nal-slight‑rising, where the slight rise neither exceeds that of the unit’s initial pitch nor reaches the pitch ranges seen in “Was?”/“What?”. Figures 4.1–4.5 exemplify the waveforms, spectrograms, and pitch contours of different practice classes of other-initiated repair presented in this chapter, beginning with open-class repair initiation (Figures 4.1–4.2) and ending with “Was denn”/“Was.”/“What.” repairs (Figures 4.3–4.5). “Was denn”/“Was.”/“What.”: A practice of other-initiated repair As a preview, our argument is that, when co-participants use “Was denn”/“Was.”/“What.” as a repair initiator, trouble-source speakers understand it to claim that trouble-source units themselves contain a “thing”-referent that is underspecified for the co-participant and, thus, to request its specification. The trouble
0.31
Pitch (Hz)
500
75 0
Time (s)
what?
Figure 4.2 Open-class English “What?”
0.62
110
Maria Egbert et al. 0.238004535
Pitch (Hz)
500
75 0
Time (s)
0.476
was
Figure 4.3 Underspecified referent repair initiation (URRI) German “Was.”
0.383129252
Pitch (Hz)
500
75 0
Time (s)
was
denn
Figure 4.4 URRI German “Was denn.”
0.7663
Repairing reference
111 0.308333333
Pitch (Hz)
500
75
0
Time (s)
0.6167
what
Figure 4.5 URRI English “What.”
sources are almost always, but not necessarily exclusively, proterms/indexical expressions. 2 Before beginning the analysis, two conceptual/definitional issues need to be addressed. First, although “thing”-referents are commonly real-world physical objects (e.g., photo album, parking token, airplane ticket), they can also be more abstract objects, such as sources of humor and characterizations of events. In our collection of “Was denn”/“Was.”/“What.” that implement repair initiation, the trouble source is never a person, place or time referent. Second, we chose the term “underspecified” for the following reasons. Repair operations consist of producing full-form “thing”-referring expressions that correspond to indexical “thing” -referring expressions in trouble-source turns. Thus, trouble-source speakers orient to the trouble, as made relevant by “Was denn”/“Was.”/“What.” as being one of underspecification for the people who initiate repair. In addition, we provide evidence (in the form of “doubles”; see below) that interactants who initiate repair orient to “Was denn”/“Was.”/“What.” as being produced to
112
Maria Egbert et al.
target reference-underspecification trouble. Thus, the characterization “underspecified” is drawn from participants’ orientations, and “Was denn”/“Was.”/“What.” involve a claim to be experiencing reference-underspecification trouble. Importantly, we do not use the term “underspecified” to suggest that, during the production of trouble-source units, their speakers necessarily design them to contain underspecified referents, as in the case of puzzles (although this can be the case). During the following analysis, readers’ attention is drawn to: 1. the distinctive unit-final intonation contours of “Was denn”/ “Was.”/“What.”, which either fall (in the case of “Was.”/ “What.”) or only slightly rise (in the case of “Was denn”); 2. the trouble-source units, whose grammatical structure embodies “thing”-referents in form of pro-terms or indexical s; 3. the repair operation, which involves trouble-source speakers specifying trouble-source referring expressions and thereby orienting to them as having been somehow underspecified. In each extract, trouble-source units are labeled with Arrow A, “Was denn”/“Was.”/“What.” with Arrow B and repair operations with Arrow C. This section begins by examining German data, including two cases of “Was denn” and one of “Was.”, and then proceeds to examine three English cases of “What.” Extract 4 is drawn from a family get-together involving Martha, Fina, Elfi, Laura, and Bärbel, who have been looking at pictures and singing songs from a book with memorabilia. Prior to this segment, in line with the activity of reminiscing, Laura is looking for a folder with photos she brought. Laura has first asked Martha where the folder is, to which Martha has referred her to Elfi, who refers Laura to Fina, whose response is displayed in line 1. In line 4, Bärbel is the last one being addressed in Laura’s search for the photo folder. Bärbel has been sitting at the table, displaying listenership. (4) MAPPE [CAE2 5:20] 01 F: Ick häff hier nix liggen. I have here nothing lie. I have nothing lying here.
Repairing reference
113
02 (1.0) ((M, F, and E looking for photo-folder)) F: Nee nee nee nee. 03 No no no no. 04→a L: Bärbel du hast se zuletzt gehabt. Bärbel you had it last have. Bärbel you had it last. 05→b B: Was denn, what PRT, What. 06→c L: Die mappe m- oder hab ich die The folder w- or have i it The folder w- or did i have it 07 ((Bärbel leans forward and reaches out for the folder on the table between herself and Laura))
The trouble-source unit, “Bärbel du hast se zuletzt gehabt.”/ “Bärbel you had it last.” (line 4), contains the direct object “se”/“it.” It is marked for grammatical gender , case, and number to correspond to the locally initial form, “die mappe”/“the folder” (feminine, accusative, singular). In response, Bärbel initiates repair with “Was denn,”/“What.” (line 5). The unit-final intonation slightly rises, but this rise does not reach the same height as the unit’s pitch onset. Laura responds by specifying the trouble-source referring expression in that she produces its full-reference form, “se”/“it,” as “die mappe”/“the folder” (line 6), thereby treating “Was denn,”/“What.” as seeking such specification. This repair operation successfully resolves the trouble in that Bärbel reaches for the folder and hands it to Laura. The larger activity of searching for the photo folder is thus also completed. Here, like almost all other repair-initiators, “Was denn,”/“What.” is treated as a post-first insertion sequence (Schegloff 2007b ), where it serves to resolve trouble in order to continue a larger course of action. Extract 5 is drawn from a videotape of a dinner conversation. Prior to this extract, Sybille and Annette have been talking about parking tokens that local businesses distribute when customers exceed a certain spending limit. Three other speakers are present, including Bernhard, who has up until now been engaged in a side conversation (not displayed in transcript).
114
Maria Egbert et al.
(5) PAULA [Fischessen:1:08:06] 01→a S: Da hättst es kriegn müssen.= There should=have it get must.= There you should have gotten it.= 02 [((A shaking her head)) [ 03→b B: [=Was denn, [=What PRT, [= What. 04→c S: Die stadette. The token.
At line 1, Sybille asserts that Annette should have received a token for purchasing a corkscrew, “da hättst es kriegn müssen.”/“there you should have gotten it.”. Grammatically , Sybille’s “es”/“it” is a direct object of this trouble-source unit. The real-world referent of the pro-form is clear to the addressee Annette, who is responding in line 2 with a headshake and who is not initiating repair. However, Bernhard initiates repair with “Was denn,”/“what.”, which has slightly rising final intonation (similar to Extract 3, above). At line 4, Sybille responds by specifying the trouble-source referring expression, “es”/“it’’, as “die stadette.”/“the token.”. Bernhard’s repair initiation is perhaps not surprising as the turn at line 1 was not addressed to and thus not designed for him. As Lerner (2004: 163–4, 178) noted, turns designed for an addressed party can be problematic for non-addressed parties and hence result in repair. Repair initiation (like Bernhard’s in this case) can be a practice for entering already ongoing conversations or moving out of schisming (Egbert 1993, 1997a). For a final German example, this time of “Was.”, see Extract 6, which is drawn from a coffee-table conversation, at which two women named “Susan” are present and a third is referred to (line 4). Prior to line 1, one of the two co-present Susans (and the only Susan featured in Extract 6) has been the center of attention. When she asked for the sugar pot, other co-participants commented that she is the only one taking sugar with her coffee. Susan responded by disclaiming that she typically takes artificial sweetener for weight reasons; therefore, her taking of sugar is a potentially sensitive topic . The request sequence of asking for sugar and
Repairing reference
115
the topic of taking sugar are complete when Georg launches his turn at line 1, which is addressed to the weight-conscious Susan. Note that Georg’s turn contains the indexical “ d as”/“that”.3 (6) Zucker [CAE1 3_20] (simplified) 01→a G: Du Sue das ham die susans You Sue that have the susans Sue! That seems to be something 02 wohl so an sich.= PRT like that with them.= the Susans have in common.= 03→b S: =Was.= =What.= 04→c G: =Zucker. Sue in Ahaus die nimmt auch zucker= =Sugar. Sue in PLACE she takes also sugar= =Sugar. Sue in Ahaus she always takes sugar too= 05 S: =Nee ich nehm eigentlich süßstoff =No i take PRT sweetener =No i usually take sweetener
At lines 1–2, Georg points out that “die Susans”/“the Susans” share “das”/“that.”. Grammatical ly, Georg’s “das”/“that” is the direct object that refers to a shared habit (i.e. not a real-world object as in the previous examples). At line 3, Sue initiates repair with “Was.”, which is produced with final-falling intonation. Georg responds by specifying his referring expression: “Zucker. Sue in Ahaus die nimmt auch zucker”/ “Sugar. Sue in Ahaus she always takes sugar too”. This operation is successful in that the trouble in understanding is resolved, as can be gleaned from Susan’s rejecting response at line 5. Georg’s turn is interesting in the followings ways. It reopens an already closed topic which is sensitive to Susan; it implies that “all” Susans take sugar although the second co-present Susan does not; and it ignores the first Susan’s explanation that she usually takes sweetener. Note also that Georg turn-initially singles out the first Susan as addressee (“du susan”). Given these features, we may speculate whether his use of the indexical is designedly underspecified and that he sets Susan up to initiate repair and thus taking a bait. The bait would consist in insisting that she has the habit of taking sugar although she says she usually takes sweetener. Georg
116
Maria Egbert et al.
would thus imply Susan has lied, and, in addition, he covertly draws attention to Susan being overweight. In response, Susan insists that she usually takes sweetener, thus rejecting Georg’s assumption, but by initiating repair she has already taken the bait. Such cases where the trouble source may potentially be designed as underspecified are rare. Our collection contains two such cases.4 While these two cases could entail a bait or puzzle, they are nevertheless treated as repair initiations. The repair-initiation action accomplished by the German “Was denn”/“Was.” is paralleled by that of the English “What.”. For the first of three English examples, see Extract 7, which is drawn from a three-way telephone call between a mother, father, and their young-adult son who no longer lives at home. At lines 1–3, the mother reminds her son that he and his sister are supposed to provide their parents with a list of desired Christmas presents. (7) Ticket [CF:5051] 01 M: Uh:m I need you gu:ys to:uh: (0.2) start (.) 02 workin’ on your (0.2) Chris’mas lists:,=small 03 duh big, please, for each of you:,= 04 S: =M[m:that’s ri]ght. 05 M: [Uh:( )- ] 06 M: We’re gunna go:to[: ] 07→a S: [Wel]l you got thuh big one. 08 (0.5) 09 S: .hhhh (0.2) ( )-= 10→b D: =Wha[:t. ] 11 M: [Wel]l12 (.) 13→c S: Thuh- (.) thee:=uh:ticket. 14 (.) 15 M: Mm.= 16 D: =O[h ] 17 M: [W]e:ll °(th)-° some- (.) D: °#Ye[ah.#°] 18 19 M: [That ]’s somewhat. .hhhh Uh:m () 20 >you=know< >we’re gunna< go: to: 21 Spokane ’n another week or 22 so so (.) I’d like tuh start lookin.’ 23 (0.7) 24 S: °(Oh)-° Okay. I:’ll=uh I’ll think about it.
After the son merely acknowledges (but arguably does not agree with) his mother’s reminder/request (Lindström 1997), “Mm: that’s
Repairing reference
117
right.” (line 4), the mother pursues his agreement by beginning to account for needing the list (line 6; see the mother’s later pursuit at lines 20–22, especially her repeat of “we’re gunna go to . . .”). The mother’s request contains an implicit potential offer to buy “big” gifts (lines 2–3, “small duh big”), and, at line 7, the son interrupts her by saying that a large gift has already been agreed upon, “Well you got thuh big one.”. Here, the son refers to an as-of-yet-unnamed “big” present with “the big one” (which turns out to be an airplane ticket). Grammatical ly, “the big one” is the direct object in this trouble-source unit and is a placeholder that is specified in the repair operation with “thee: =uh: ticket.” (line 13). In other words, the reference in the trouble-source unit is a more “granular” reference form (Schegloff 2000a) than the reference form in the repair operation. In this case, “the big one” is treated as being underspecified by the dad, who initiates repair, but not by the mom, who displays that she understands “the big one” by beginning to respond at line 11, “Well-,” which is comparable to how she re-begins at line 17, “We: ll”. At line 10, the dad initiates repair with “Wha:t.” (which is produced with final-falling intonation). The son responds by specifying “the big one” with “thee:=uh: ticket” (line 13). There is evidence that the dad’s “Wha:t.” constitutes repair initiation. At line 16, his “Oh,” which sequentially is a third-positioned action relative to the repair sequence (Schegloff 2007b ), claims a change of state from unknowing to knowing (Heritage 1984). And, at line 18, the dad responds to the son’s “Well you got thuh big one” by partially discounting it with a (delayed and) dismissively produced “°#Yeah.#°”, thereby treating the “Wha:t.” as having delayed the progression of the son’s claim in order to better understand it. (The mom similarly dismisses the son’s claim at lines 11, 17, and 19.) For a second English example, see Extract 8, which is drawn twenty-seven minutes into a telephone call between Sue and her boyfriend Max. Sue has initiated all of the major topics and has done a majority of the talking. (8) Phone (CF:6507) 01 S: 02 03 M: 04 05→a S:
.hhh Anyway there’s nothing on you(r)- more on your mind is there. (0.4) Uh:not really. h=Oh:.=hh (.) .hhh (and) a:re- (.) are all our
118
Maria Egbert et al.
06→a our conversations gunna end up like this, 07 (.) 08 S: On thu[h phone,-] 09→b M: [Wha:t. ] 10→c S: .hhhhhh (0.2) I mea[n] 11 M: [(O]h./No.) h= 12→c S: =.hhh (I’ll=say) something:you know pretty:=h (.) 13→c >somewhat-< (.) you know pretty good that hap14→c you know happened (to you then you) say something 15 .hhhhhhhhhh[h h-=[or ] 16 M: [.pt=We:l[l >you kno]w< you gotta 17 understand th’t I’m:kind=of a little ti:red,
Although Sue’s action at lines 1–2 solicits additional conversational topics (Button and Casey 1984), it is grammatical ly designed to prefer (and thus presume) a “No”-response. Given that Max has not previously initiated any major topic , Sue’s turn can be heard as an interpersonal complaint regarding Max’s lack of enthusiasm or interest in terms of talking to her (for in-depth research on complaints, cf. Heinemann and Traverso, forthcoming; Yoon 2006). In the wake of Max’s (delayed) agreement, “ U h: not really” (line 4), Sue asks/accuses: “are all our conversations gunna end up like this” (lines 5–6; note the extreme-case formulation “all our conversations”; Pomerantz 1986). Grammatically, Sue’s “like this” is an adjunct (in the form of an adverbial expression of manner) and refers to the immediately prior sequence of action (at lines 1–4). In response, Max initiates repair with “Wha:t.” with final-falling intonation (line 9), to which Sue responds by specifying what she “meant” (see line 10, “I mean”) by “this.”. That is, she specifies her complaint by providing a characterization of how the call has “ended up” (lines 12–14). Max orients to his “Wha:t.” as having initiated repair on Sue’s turn and thus as having delayed its progression to better understand it, by subsequently responding to it by defending himself: “We:ll >you know< you gotta understand th’t I’m: kind=of a little ti:red” (lines 16–17). For a final English example, see Extract 9, which is drawn from an American telephone call between two college students, Carla and Rich. (9) Humorous [UTCL:D08] 01 02
C: Peet’s a:lways there. ((i.e. at the apartment)) R: °Y°:eah he is.=
Repairing reference
119
03 C: =B’t your never there. ((i.e. at the apartment)) 04 (1.6) C: (No:pe) 05 06 (1.0) R: Hh hh [hh °hh° ] 07 08 C: [Where are] you. on thuh pho:ne, R: .hhhh ((laugh-relevant inbreath)) 09 C: ↑What’re you↓ sa:ying? 10 R: .huh ((laugh)) 11 12 (0.4) C: Te:ll me:. 13 R: .hhh hh=heh (0.5) .huh (0.2) No:thing I’m 14 15→a j’st- I find it humorous. 16→b C: ↑Wha:t.↓ 17→c R: Just (tw)- listening duh you. C: What I sa:id? 18 R: Yes. 19 C: Peet’s always the:re, 20 R: Yes. an’ I’m n:ever ‘ere. 21 C: So where are you. 22
When Rich fails to respond (at line 4) to Carla’s assertion that Rich is never at his apartment, “B’t your never there.” (line 3), Carla produces a candidate response (for Rich) that aligns with her position: “No:pe” (line 5). After a long gap of silence at line 6, Rich laughs, “Hh hh hh °hh°” (line 7), and continues laughing at lines 9, 11, and 14. Perhaps because of Rich’s displayed lack of attention (i.e. his failure to respond at line 4), or because Rich’s laughter (at line 7) is disjointed from prior talk (i.e. lines 1–5, which neither participant oriented to as being funny), Carla begins to suspect – evidenced by her chastising question “↑What’re you↓ sa:ying?” (line 10) and its pursuit “Te:ll me:.” (line 13) – that Rich might be talking to, and laughing about something with, his roommate Peet. Rich ultimately answers Carla at lines 14–15: “No:thing I’m j’st- I find it humorous.”. Grammatically, Rich’s “it” is the direct object of this trouble-source unit. As a pro-form, “it” is analyzable as being underspecified relative to its full-reference form, which is the source of what Rich finds humorous. Indeed, as the repair operation to Carla’s “↑Wha:t.↓” (line 16), Rich responds by specifying the humor source: “Just (tw)- listening duh you.” (line 17). To review Extracts 4–9, “Was denn”/“Was.”/“What.” are produced with either clearly falling intonation (in the case of “Was.” and “What.”) or with non-strong-rising intonation (in the case of
120
Maria Egbert et al.
“Was denn”), which differentiates them from open-class forms (Drew 1997). The trouble sources are pro-terms or indexical expressions (e.g., “it”, “that”, “this”; in the data), adverbial expressions of manner (e.g., “like this”), and placeholder nouns (e.g., “the big one”) that refer to “things”, which include not only realworld physical objects (e.g., photo album, parking token, airplane ticket), but also more abstract objects (e.g., a source of humor, a characterization of a telephone call, and a habitual physical action). “Was denn”/“Was.”/“What.” implement repair initiation by halting the forward progression of a prior course of action (regarding progressivity , see Lerner 1996) in order to deal with a particular type of referential trouble, i.e. its claimed underspecification. “Was denn”/“Was.”/“What.” are responded to with specifications of referring expressions. The referring expression in the trouble sources may or may not have been used in sequentially proximate prior environments, including the prior turn. Further evidence: The case of “doubles ” So far, we have examined formal features of the repair initiators “Was denn”/“Was.”/“What.” and their uptake by co-participants. These results are strengthened by observations of a special kind of context in which “Was denn”/“Was.”/“What.” can occur, and that is when they are produced in the same turn with another repairinitiation action. In such multi-unit turns (called “doubles” by Schegloff et al. 1977), the second repair-initiation attempt is typically “stronger” (more specific) than the first in terms of locating the trouble source . In doubles, “ Was denn”/“Was.”/“What.” are placed first, followed by stronger confirmation requests of candidate specifications of trouble-source referring expressions. For a German example, see Extract 10, which is drawn from a dinnertime interaction. (10) WEIN [Fischessen_1.31.10] (simplified) 01→a A: Ptz ptzptzptz ºderº is lecker. der g- der Pts pts pts ºitDEMº is tasty. that- p- that Pts pts pts ºthat oneº’s tasty. that- p- I 02→a schmeckt mir am besten von allen. tastes to me the best of all. like that one best of all of them.
Repairing reference
121
03 B: Ja? Yeah? A: [Mh ºhmmº] 04 [Uh ºhumº] [ ] 05→b D: [Was denn. ] [What PRT. ] [ “What.” ]
[Dieser wein hier? ] [This wine here? ] [This wine here? ] [ ]
06→c A: [((Points at wine)) ] 07→c A: mh ºhmmº uh ºhumº 08 (0.5) ((D starts turning bottle to read label))
As context, two minutes prior to Extract 10, the interactants discussed what they would drink next. Annette had suggested continuing to drink another bottle of the same white wine, whereas Sybille had suggested a bottle of red wine. In Extract 10, at lines 1–2, Annette produces an extreme-positive assessment (Auer and Uhmann 1982; Pomerantz 1984) of the same bottle of white wine that she had suggested earlier: “Der schmeckt mir am besten von allen.”/“I like that one best of all of them.”. Annette produces a slight head nod toward the bottle of wine during the production of “der”/“that one”. Grammatical ly (in German), Annette’s “der”/“that one” is the subject. Note that the German “der” has the same form as the article “der” in “der Wein”. Annette’s “der”/ “that one” relies on a demonstrative determiner to refer to a referent indexical ly (Hanks 1992). In spoken German, this form is often used as a pronoun (i.e. instead of “er”) or as a demonstrative (i.e. instead of “dieser”). German grammar thus provides for an indexical form which is not easily translated directly into English. At line 5, David treats the referring expression as being underspecified by initiating repair with “Was denn” (line 5), but then immediately following it up with a request for confirmation of a candidate referent: “Dieser wein hier?”/“this wine here?”. David’s candidate solution, “Dieser wein hier?” is grammatically congruent with the pronoun “der” in the trouble-source turn in that both are morphologically marked for the grammatical gender masculine and case-marked for the nominative. In moving from “Was denn” to a
122
Maria Egbert et al.
candidate solution, David displays an orientation to his own “Was denn” as doing the job of targeting a “thing” referent in need of specification to him. Note that Annette similarly orients to the function of “Was denn”; that is, immediately after its production, and before David produces his candidate solution, she specifies her “der”/“that one” by pointing at a specific bottle of wine. Thus, in doubles with “Was denn”, the second repair-initiator offers a candidate specification of the troublesome indexical . For an English example, see Extract 11, which is drawn from a phone conversation between Joe and his mother. Prior to this fragment, Joe and Marsha have discussed a variety of topics dealing with Marsha. Deep into the call, Joe asserts that “he can’t talk very long” (line 1). With a delayed “by thuh way” (line 3), Joe orients to his assertion as being “misplaced” relative to prior talk (Schegloff 1984, 1987c), and, at lines 6–7, Joe returns to prior talk by producing a summary assessment of it. (11) Stanford (Marcia and Joe) 01 J: 02 03 04 J: 05 06→a J: 07 08 09→b M: 10→c J: 11 12 J: 13 14 M: 15
Oh we can’t talk very long ‘cause I=‘as jus’ talkin’ tuh gra’ma. (0.4) by thuh way. (0.4) But this sounds like=ay- very interesting situation you’re in you’re in here. (0.2) What. Mi::ne? Ye:ah. (0.5) With Stanford research an’ everything. (1.4) Oh I myself am auditioning for Ca:melot this June.
In the trouble-source unit (lines 6–7), Joe’s “this”, in combination with “a very interesting situation”, alludes to, and thus is analyzable as being underspecified relative to, a particular “situation”. After a slight pause (line 8), Marcia initiates repair with “What.” (line 9), targeting “this” in the trouble-source turn, and then immediately follows it up with a request for confirmation of a candidate domain of “situation”, “Mi::ne?”, which attempts to clarify the nature of the “interesting situation” being referred to. Here, Marcia displays
Repairing reference
123
an orientation to her own “What.” as doing the job of targeting an underspecified referring expression. Joe treats Marcia’s request for confirmation as such by first confirming, “Ye:ah.”, and then by supplying a particular situation: “With Stanford research an’ everything.” (line 12). A contrasting practice with a similar form: “Für was.”/“For w hat.” The function of “Was denn”/“Was.”/“What.” can be put into relief by examining a different, but allied, repair-initiation action: “Für was.”/“For what.” The comparison of these two different repair‑initiation actions is presented to shed further light on the unique function of “Was denn”/“Was.”/“What.”. In Extracts 4–9, repair-related trouble was associated with referring expressions contained in trouble-source units themselves. Alternatively, we will now discuss one other way in which repair-related trouble can be associated with referents. In these cases, none of the referring expressions in the trouble-source action are treated as presenting “trouble”, but, rather, the trouble-source action presupposes a referent (e.g., an elaboration of a direct object), and the knowledge / understanding of this referent is necessary to adequately “grasp” the trouble-source action. This alternative type of reference trouble can be dealt with by an alternative repair-initiation practice, “Für was.”/“For what.”, which prompts (Lerner 2004: 169) the prior speaker to elaborate a prior utterance. The trouble-source utterance is syntactically complete, and the repair operation in the form of an increment provides pragmatic completion. For a German example, see Extract 12. The trouble-source unit is Oma’s announcement: “ Ich hab’ jetz’ wieder ‘nen neues rezept gekriecht”/“I have now gotten another prescription again” (line 4). (12) Für was (Oregon 1A_1:38) 01 M: =.hh Wie oft musst’n zur krankengymnastik. =.hh How often must+you to+the physical therapy. =.hh How often do you have to go to physical therapy. 02 O: Äh dreimal die woche °ne.° Uh three times the week °PRT.° Uh three times per week you know.
124
Maria Egbert et al.
03 (.) 04a→ O: Ich hab’ jetz’ wieder ‘nen neues rezept gekriecht, I have now again a new prescription got, I have now gotten another prescription again, 05b→ M: Für was. For what. (0.5) M: Ach so f[ür die gymnastik. 06 Oh i see f[or the therapy. [ 07c→ O:
[Für die krankengymnastik. [For the physical therapy.
Grammatically, Oma’s “rezept”/“prescription”, is the direct object. Based on the analysis of Extracts 4–9 (above), to initiate repair with “Was denn”/“Was.” would be to claim that “rezept”/“prescription” is itself underspecified. However, Markus initiates repair with “Für was.”/“For what.”, which alternatively claims that the troublesource unit elided a referring expression necessary for his understanding and solicits an increment (Schegloff 1996b; 2001) to the trouble-source unit. In this case, Markus’s “Für was.”/“For what.” is responded to with an elaboration of the direct object (i.e. the source mandating the prescription). Both participants orient to this as the repairable . Oma responds by producing a grammatical increment to her trouble-source question that supplies a new referent: “die krankengymnastik”/“the physical therapy” (line 7). Prior to Oma’s response, Markus himself claims that he has resolved the trouble with “Ach so”/“Oh I see” (Heritage 1984; Golato and Betz 2008), and then repairs the trouble with “für die gymnastik”/“for the therapy”, which displays his understanding of his “Für was.”/“For what.” as seeking an elided referring expression. For an English example, see Extract 13, in which the troublesource unit is Ada’s question at line 6. (13) Loan (CH:4365) 01 A: .mtch So:u- (.) I=s- I:was thinking 02 o[f sen]ding mom an’ da:d (.) uh:m about a 03 M: [.hhh] 04 A: hundred bucks. 05 (.) 06a→ A: .hhh Does that sound reasonable to you,=
Repairing reference
125
07b→ M: =For what. 08 (0.9) 09c→ A: Off of what I owe them.
Grammatically, Ada’s “that” (line 6) is the subject. Based on the analysis of Extracts 4–9 (above), to initiate repair with “What.” would be to claim that “that” is itself underspecified and seek its specification, such as a recharacterization of Ada’s proposal at lines 1–4. However, Max initiates repair with “For what.”, which alternatively claims that the trouble-source unit elided a referring expression necessary for his understanding and solicits an increment to the trouble-source unit. Similar to Extract 12 (above), Max’s “For what.” seeks an elaboration of the subject (i.e. Ada’s reason for sending the money). This is oriented to by Ada, who responds by supplying her reason, “Off of what I owe them.” (line 9). (The) Difference(s) between “Was.” and “Was denn” As mentioned in the introduction, there is an easily describable difference in form between “Was denn” and “Was.”. “Was denn” contains the particle “denn” and is delivered with an intonation contour that ranges from falling to slightly rising. “Was.” does not contain a particle and is always delivered with clear final-falling intonation. Given these form features, German “Was.” is akin to English “What.”. However, distributionally, when used to initiate repair, “Was denn” (75 percent of corpus) occurs much more frequently than “Was.” (25 percent of corpus). Furthermore, and perhaps more importantly, similar to English “What.”, in German only “Was denn” is used to accomplish a specific nonrepair function, namely a “go-ahead” action to pre-sequence-initiating actions (Schegloff 2007), such as a summons, preannouncement , and story/ joke preface (see examples below). In our data, we found no cases of “Was.” used as “go-ahead” action, suggesting that “Was.” is reserved for repair-initiation actions only. Such distributional and action-based differences suggest that “Was denn” (1) is the more generic, common, or default form relative to “Was.”; and (2) may be the more appropriate equivalent of English “What.”. What follows are two examples of pre-announcements (Terasaki 1976) that are relevantly responded to, and progressed, by “Was denn”/“What.”, respectively. As Schegloff (1997b) argues, there is
126
Maria Egbert et al.
not a one-to-one correspondence between practices and actions, and, in the following cases, “Was denn”/“What.” are not used to implement repair initiation. For a German example, see Extract 14, from an interaction between two adults and two children, with one of the children beginning to tell a story. (14) Weisstduwas (SandL_T1_0.22.48) 01→a S: Un weißt du (.) wa:as, An you know (.) wha:t, 02→b A: Was denn, What PRT, What. 03→c S: Und die .hh (ni.hhna) .H hat(te) auch And the .hh ((name .hh)) .H has/had also And .hh ni.hhna has also once 04 mal ihre ((2 syll.)).H[HH PRT her ((2 syll.)).H[HH (had) her ((2 syll.)).H[HH [ 05 A:
[Ja? [Yes? [Really?
Extract 15 is drawn from an English telephone conversation between two sisters, Kathy and Janet, the former of whom is planning a vacation to Cabo San Lucas and has been searching for reasonably priced airline tickets. (15) Cabo (Sister Calls.2) 01→a K: An’ then uhm guess what. 02→b J: What. 03→c K: ’Cause I called ‘er because 04 .hh Kaytlen fo:und uh::m (0.2) 05 things for a hundred an’ fifty. 06 to go duh Cab[o.] 07 J: [.H]HHH She di:d?
In both Extracts 14–15 (above), line 1 involves the first part of a preannouncement sequence, which initiates a course of action and solicits a response. Similar to many preannouncement sequences, these (see “Arrow A”) project the presence of announcables without
Repairing reference
127
explicating them. Here, they are alluded to by “Was denn,” and “What.”, respectively, which refer to actions or events. These preactions function to frame their projected announcements as “news” for recipients (Terasaki 1976) and make use of underspecified referents (actions or events in the lives of the interactants) to which the prefer red response “Was denn,”/“What.” then engenders the actual announcements. At line 2 in both examples, “Was denn,”/“What.” functions as a second-part “go-ahead”, which is oriented to as such by pre-announcers when they proceed to deliver their announcements (lines 3–4 and 3–6, respectively), thereby explicating the referring expression. Rather than halting the progressivity of the pre-announcement sequence, “Was denn,”/“What.” propels the course of action. Extracts 14–15 were presented for two reasons. First, the interactional achievement of “Was denn”/“What.” as second pair parts of a pre-announcement sequence is different from initiating repair, while they share the feature that they request more specification. In response to both functions, a reference is “unpacked”. Second, our preceding comparison of pre-announcement go‑aheads and repair initiators sheds light on a hypothesis of how “Was denn”/“Was.”/“What.” may differ in their repair initiation function. In German, relative to “Was.”, “Was denn” is the more default repair-initiation form, and it may be the case that “Was.” is doing something unique. This aligns with Thurmair’s (1991: 382) observation that German w-Fragen (wh-questions) typically contain the modal particle “denn” and that such questions without “denn” are marked. All of this, though, begs the question of what “denn”’s uniqueness might be in the context of repair-initiation. Because we have too little data to pursue this hypothesis further, we offer some speculative possibilities. An initial way of getting at differences between “Was denn” and “Was.” is to examine aspects in which they do not differ. Our limited data suggest no differences in terms of idiolectal and dialectal factors, nor medium (i.e. telephone vs. co-present interaction), nor number of participants (i.e. two or more), nor distance between the trouble-source referring expression and the reference form to which it (sometimes) connects back. There also seems to be no difference in prefer ence organization, i.e. “Was denn” and “Was.” can target trouble-source units which are preferred or dispreferred .
128
Maria Egbert et al.
Regarding this last observation, however, relative to “Was denn” both native-speaker authors of this chapter (Maria Egbert and Andrea Golato) perceive “Was.” as sounding more “terse”. Their vernacular understanding is also expressed by research in linguistics , which reports that “denn” makes utterances friendlier and possibly less blunt (Durrell 1992; Helbig 1988: 107; Hentschel and Weydt 1983: 266; Schwitalla 2003: 153). One subtle difference might exist at the sequential level. In each of our six cases of “Was”., the trouble source is always the unit immediately prior to “Was.”. In our fifteen cases of “Was denn”, this is so for thirteen cases, but there are two cases in which the trouble-source unit is farther removed. Thus, it is possible that the effectiveness of “Was.” relies on a more proximate relationship between itself and the trouble-source unit, whereas “Was denn” may remain effective given a more distant sequential relationship. We speculate that the particle “denn” serves to claim that repairinitiators have a general sense of intersubjectivity amid the need for reference specification. Such a sign of general intersubjectivity would not be necessary with “Was.” because the trouble source is immediately prior. Although more research on actual conversation is necessary, the aforementioned speculation is in line with linguistic descriptions of the function of the particle “denn”. When “denn” is used in questions, linguists have argued that it indicates that such questions refer to immediately prior talk (Dittmann 1980; Franck 1980: 222; Helbig 1988: 107; Thurmair 1989: 164, 166), and thus are not asked “aus heiterem Himmel” (“out of the blue”) (Zifonun et al. 1997: 1219). The particle “denn” is also said to target things that are known, or assumed to be known, by hearers (Helbig 1988: 107). Along these lines, Franck (1980: 225) argued that “denn” is frequently used in utterances that request further explanation before its speakers are able to respond to the content of the utterance (i.e. in repairs). Discussion Given the important relevance of referencing to intersubjectivity, it makes sense that the organization of repair provides conversationalists with specific practices for repairing reference-related
Repairing reference
129
troubles. Lerner (2004: 180–181, footnote 25) noted that there are multiple practices with which speakers can deal with “inadequate or insufficiently complete or a not yet response-ready turn that has reached possible completion.” Our small collection of thirty-eight instances from a corpus of ninety-five hours suggests that there is a specific category of repair initiation designed to request specification of “thing”-referring expressions which the repair-initiating speaker claims to be underspecified and thus inhibiting full understanding. In German and English, this can be realized with “Was denn”/“Was.”/“What.” produced with nonrising intonation. This chapter extends prior research on German (Selting 1987a, 1987b, 1987c, 1988, 1992) and American English (Schegloff et al. 1977; Schegloff 1997b) by comparatively analyzing one particular type of repair initiation that requests the specification of “thing”referents that are somehow embodied in the components of trouble-source units themselves. This practice was implemented by “Was denn”/“Was.” in German and “What.” in English, produced with either final-falling or slight (but not strong)-rising final intonation. Future research needs to examine other types of troublesource specific repair initiators, including those that target person, place, and time-referring expressions (such as “Wo.”/“Where.”, and “Wann.”/“When.”) (See Schegloff et al. 1977, for comments on English “Where”, see also Lerner 2004). For now, we consider the term “thing” (as in “thing”-referent) to be an analytic one. In the data, “things” include not only realworld physical objects (e.g., photo album, parking token, airplane ticket), but also more abstract objects (e.g., a source of humor, a characterization of a telephone call, a habitual physical action, and an action/assertion). Future research needs to explore in more detail how members orient to persons, places, times, and “things”. As argued most forcefully by Selting (1987a, 1987b, 1987c, 1988, 1992), our analysis reiterates that intonation is an important, constitutive feature of practices of other-initiated repair. As suggested by our analysis of “Was denn”, when languages provide particles, they too can be constitutive features of repair initiation. Relative to the distribution of previously analyzed classes of other-initiated repair, such as open-class initiators (Drew 1997), rising‑intoned interrogative words (e.g., “Who?”, “When?”, “Where?”), and various types of repeats and reformulation of trouble
130
Maria Egbert et al.
sources (Sacks and Schegloff 1979), “Was denn”/“Was.”/“What.” repair initiations seem to be rare. This rarity might be explained by the presence of a set of first-order conversational practices and prefer ences regarding reference that obviate the relevance of “Was denn”/“Was.”/“What.”. For example, the preference for selfcorrection (Schegloff et al. 1977) and the associated practice of “try-marking” (Sacks and Schegloff 1979) tend to preclude the need for “others” to initiate repair, for example, through iterations of trouble-source unit completion, recipient silence, and same-speaker continuance (Heritage 2007). Furthermore, there is a preference for using recognitional reference forms (at least when referring to persons; Sacks and Schegloff 1979), which, in our data, are not targeted by “Was denn”/“Was.”/“What.”. Finally, there are practices regarding the use of locally initial and subsequent reference forms that promote intersubjectivity (Schegloff 1996a). Thus, “Was denn”/“Was.”/“What.” are only “needed” when other conversational safeguards have “failed”. Our findings add to the growing evidence that at least the context-free, structural organization of repair outlined first for English (Schegloff et al. 1977; see also Schegloff 1992a), and replicated for Thai (Moerman 1977) and for German (Egbert 1996), seems to be universal . Our findings further suggest that different languages (in our case, German and English) may contain similar linguistic and intonational resources (in our case, the lexical item “Was.”/“What.” produced with falling intonation) to accomplish similar repair-initiation actions. Method ological issues in conducting a cross-linguistic analysis One conversation analytic goal is to analyze how linguistic resources are used in talk-in-interaction to achieve social actions, thus contributing to a theory of social order. Since languages provide different resources, cross-linguistic research contributes to the theoretical question of what dimensions of human sociality are universal and what are specific to particular linguistic communities (see the introduction to this volume, see also the collection in Enfield and Stivers 2007, and here in particular, Hanks 2007 and Levinson 2007). Given this theoretical relevance, we would like to make transparent how method ological decisions and procedures in conducting a comparative analysis may influence the analysis
Repairing reference
131
and thus potentially impact the theoretical development. We are addressing how the role of English as data and as a lingua franca in the conversation-analysis community may shape our findings. Similar to many other comparative studies, the present chapter analyzes the commonalities and differences of one phenomenon across two languages. All cross-linguistic studies on repair have taken English as the basis for comparison, for example, Moerman ’s (1977) analysis of repair in Thai, Egbert ’s (1996, 2002) work on repair in German, Kim’s (1993) study of generalized other-initiated repair in Korean , or Sidnell ’s (2007a) analysis of repairing person reference in a small Caribbean community. When doing so, it may be tempting to look for those formats that have been found for English, yet it is vital that the researcher remain open for all kinds of formats, including ones not attested in English. 5 Similarly, it may be tempting to compare items across languages based on form congruities (e.g., downward intoned German “Was.” and downward intoned English “What.”). However, the present chapter has shown that more can be gained by comparing functions of utterances, i.e. actions across languages. Similarly, our analysis cautions against translating utterances based on (perceived) phonological or lexical correspondences. For example, based on having similar lexical forms, linguistic meanings, and prosodic shapes, one might be biased toward translating the German “Was.” as the English “What.”. In contrast, our analysis suggests that the German “Was denn” is a more appropriate translation because it and the English “What.” are more similar in terms of the actions they accomplish. That is, unlike the German “Was.”, the German “Was denn” and the English “What.” accomplish both repair initiation and goahead responses to pre-sequences. In sum, we suggest that translations be based on an analysis of action, sequential placement, and function rather than on perceived form congruities (for a similar argument, see Betz and Golato 2008). Many comparative studies show that it is difficult or impossible to accurately translate interactional data into English. Even when a morpheme-by-morpheme gloss is provided with the transcript, the original language is not fully represented. One way that a reader of English can be made aware of this dilemma is by leaving words such as particles in the original language within the English rendering (Sorjonen 2001). This works well for highlighting the focal phenomenon, yet it would render a transcript illegible if used for all
132
Maria Egbert et al.
items which lack an appropriate translation. Descriptions of what the English translation does not capture can be cumbersome for the reader, and the researcher is forced to find a balance between focusing on the relevant features of the phenomenon and to what level the nature of the data can be sustained. Acknowledgments All three authors contributed equally to this research. We thank Emma Betz for helping with transcription and bibliographic work and Christopher Stewart for creating the PRAAT files. Additionally, we thank Emma Betz, Trine Heinemann, Anna Lindström, MarjaLena Sorjonen, and Jakob Steensig for their feedback on individual data segments. We are grateful to Emanuel A. Schegloff for feedback on an earlier draft and to Jack Sidnell, the Editor, for both feedback and support. All remaining errors are of course our own. Notes 1 “Repair” is defined as the mechanism by which trouble in speaking, hearing, and understanding is claimed and resolved (Schegloff et al. 1977). 2 We intentionally avoid the terms “locally subsequent” (vs. locally initial) and “non-recognitional” (vs. recognitional ) because they have only been empirically verified with regard to person-reference, and this chapter deals with thing-reference (but see Golato 2005 for a discussion of locally initial and locally subsequent forms for thing-references in German). 3 Here and in other data samples we chose to summarize stretches of the interaction which would have been too long to display. 4 We would like to thank Emanuel A. Schegloff for drawing our attention to designedly underspecified references. 5 The idea that our knowledge of talk-in-interaction in most languages may be shaped by what is already described for English is due to the history of the field, whose first-generation researchers have reported profound results mainly on English. Therefore, taking English as a basis of comparison is more a decision due to the history of science than to scientific reasoning.
Part III
Aspects of response
5 Projecting nonalignment in conversation Anna Lindström
Projecting nonalignment Early conversation-analysis (CA) research established that nonaligning actions are typically delayed and that turn-initial objects can project nonalignment. These features provide a resource for recipients, allowing them to anticipate nonaligning responses and to revise their prior action before such a response is realized in this way avoiding the occurrence of rejection or disagreement . Sacks (1987) showed that in conversation there is an association between agreement and contiguity on the one hand, and disagreement and noncontiguity on the other. Agreeing utterances tend to be contiguous while disagreeing utterances are noncontiguous. Consider Extract 1. (1) (Sacks 1987: 58) 01 A: 02 B: 03 04
Yuh comin down early? Well, I got a lot of things to do before gettin cleared up tomorrow. I don’t know. I w- probably won’t be too early.
As Sacks noted, B’s disagreement – “ probably won’t be too early” – is not only weak, additionally, and more important for this study, it is not produced until the end of B’s turn. The disagreement is prefaced by “Well,” an account “ I got a lot of things to do before gettin cleared up tomorrow,” a hedge “I don’t know,” and a selfinterruption “I w-” all of which serve to delay its production. The distinctive features of agreeing and disagreeing turns were explored
136
Anna Lindström
further by Pomerantz (1975, 1978, 1984). Focusing on how predisagreements shape sequence trajectories she noted that: a gap after a first pair part to which agreement /disagreement is relevant may function as an “unstated” or “as yet unstated” disagreement . That is, if a recipient delays in initiating his talk post the completion of a turn to which agreement /disagreement is relevant, that delay-gap may be disagreement implicative for his [sic] co-participant. As such, that co-participant may resume talk post a gap with talk which displays a sensitivity toward a discrepancy between the parties. (Pomerantz 1975: 79)
This is illustrated in the following Extract 2. (2) (SBL:3.1.-8). (Pomerantz 1975:79, line numbers added) 01 B: 02 03 B: 04 A:
. . an’ that’s not an awful lotta fruitcake. (1.0) Course it is. A little piece goes a long way. Well that’s right.
B makes an assertion in line 1 “an’ that’s not an awful lotta fruitcake.” Faced with no uptake from A (line 2), she revises her assertion in line 3 “Course it is. A little piece goes a long way.” B’s revision potentiates A’s agreement in the ensuing turn “Well that’s right.” In Pomerantz’s terms, B’s revision shows her sensitivity toward the discrepancy between A and B’s displayed points of view. In this example, a silence was treated as disagreement implicative. Other devices for projecting nonaligning responses include prefaces, partial repeats, anticipatory accounts , and “pro-forma” agreements (Heritage 1984b, 1988; Levinson 1983; Pomerantz 1975; Schegloff 1995b, 2007b). Pomerantz (1975) argued that modifications and revisions such as B’s “Course it is. A little piece goes a long way.” coupled with the tendency to delay or withhold actual disagreement has “a minimizing effect on the occurrences of disagreements and a maximizing effect for agreements . In such instances, it is not only what would be a disagreement may not get said, it is also that what comes to be said may be said as an agreement” (Pomerantz 1975: 81). In line with this work, alignment can be understood as the product of finely tuned negotiation between speaker and recipient. The present study focuses on how this negotiation is sequentially realized over the course of each emerging TCU by demonstrating how a prosodic variant of the affirmative response
Projecting nonalignment in conversation
137
token “ja” can be used to project a nonaligning responsive action in Swedish conversation. The dictionary-assigned meaning for “ja” is “yes.” There are several prosodic variants of the Swedish “ja”. I use the term “curled” to draw attention to two prosodic features: a lengthened vowel and a slight rise in pitch toward the end of the syllable (represented by colons and underlining in the transcript). The collection of curled “ja” turns include tokens that begin with an open /a/ as well as ones that begin with a soft /j/. The arrowed lines in Extract 3 show a curled “ja” response to a proposal. (3) Visit (MOL:1:A). In line 10, A is referring to the fact that she has to help her children get accustomed to institutional day care by spending time with them at the day care facility. 01 A: → Hörru ja tänkte höra ifall ni ville ha Listenyou I thought hear in case you wanted have Listen I was going to ask if you wanted to be 02
en liten påhälsning¿< a little visit paid a visit
03 V:
↑Ja: de vill vi gärna?↑ Yes that want we gladly Yes we would love to
04 A:
Utav mej å ba:rnen¿ By me and the children
05 V:
Ja::: jättegä:rna? Yes very gladly Oh yes of course
06 A: I veckan¿ In the week This week 07 V: → .hh Ja:: (.) de e lite kärvt ((smilevoice)) men .hh Ja:: (.) ‘t’s little difficult but .hh Ja:: (.) it’s a little tricky but 08
de- vi ska se vicken da hade ‘ru tänkt då, ‘t- we will see which day had “you thought then ‘t- let’s see which day had you thought of then
09 A: Nej ja har inte tänkt nånting för nästa vecka No I have not thought anything cause next week No I hadn’t thought anything it’s just that I begin
138
Anna Lindström
10
börjar “ja” sko:la in dom, begin I school in them schooling them next week
[a discussion about Vera’s new job omitted]
50 A: >Är det kärvt den här veckan sa du< Is it difficult this here week said you You said it’s tricky this week
Anita asks Vera whether she would like a visit in lines 1–2, “Hörru ja tänkte höra ifall ni ville ha en liten påhälsning”/“Listen I was going to ask if you wanted to be paid a visit.” Anita is addressing her proposal to Vera and family with the second person plural form of the pronoun “ni” in line 1. Vera enthusiastically accepts the proposal on her family’s behalf with a two-unit turn in line 3, “↑Ja: de vill vi gärna?↑”/“Yes we would love to.” Upon hearing Vera’s acceptance, Anita specifies the proposal by stating who the prospective visitors will be. This is done with an increment that is syntactically parasitic on Anita’s prior turn in lines 1–2 (Lerner 1987, 1989). Vera responds with heightened enthusiasm in line 5, “Ja::: jättegä:rna”/“Oh yes of course.” Anita then adds another increment that proposes a time for the visit, “I veckan”/“this week.” After an inbreath, Vera responds with a curled “ja” turn. The curled “ja” projects but does not implement a nonaligning stance . This analysis is supported by the fact that turn-transfer does not occur during the micropause after the curled “ja”. In refraining from claiming the floor during this pause, Anita shows an orientation to the relevance of an extended turn. The type of nonalignment that is being instantiated is specified within the ensuing parts of Vera’s turn. Instead of accepting Anita’s proposal, Vera states “de e lite kärvt”/“it’s a little tricky.” Vera begins the next turn constructional unit (TCU) with a contrastive conjunction “men”/“but.” This syntactic linkage to the prior TCU indicates that the emerging TCU will contrast with its predecessor. Thus it projects that although the time Anita has proposed is problematic for Vera she might still be willing to work something out. However, Vera does not bring this TCU to completion (note the self-interruption of de- in line 8). Instead she initiates an insertion sequence, “ vi ska se vicken da hade ‘ru tänkt då”/“so which day had you thought of” that indicates that her acceptance is contingent on the time frame for the visit, and the turn is thereby
Projecting nonalignment in conversation
139
hearable as a possible pre-rejection of the proposal. The past tense formulation “had you thought of” coupled with the inference marker “då” conveys that Vera has inferred that Anita already had a particular day in mind when she made the proposal. Anita rejects Vera’s assumption that she had called with a pre-planned agenda in her next turn, “Nej ja har inte tänkt nånting”/“No I have not thought anything.” That she is addressing this assumption is partially conveyed through the reuse of the word “thought” from line 8. Anita then gives an account “ för nästa vecka börjar ja sko:la in dom”/“it’s just that I begin schooling them next week” that emphasizes that from her perspective it is urgent that the visit takes place in the near future. This is an ability account (Labov and Fanshel 1977) that has a no‑fault quality (Heritage 1984a, 1988) in that it “does not implicate a lack of willingness” (Heritage 1988: 18). Rather, it references a circumstance that is beyond Anita’s control (i.e. the requirements posed by her children’s day-care provider) as the reason why the visit needs to be scheduled sooner rather than later. Vera then pursues a discussion about Anita’s new job (transcript not shown). Anita returns to the issue of scheduling the visit in line 50 by formulating the upshot of Vera’s curled “ja” turn, “Är det kärvt den här veckan sa du”/“You said it’s tricky this week.” This formulation does not merely show that Anita has understood Vera to convey that the proposed time frame for the visit is problematic but also that Anita may be willing to back down from her original proposal. The analysis of this example illustrates that the curled “ja” projects nonalignment and that the recipient, in this case Anita, can use this as a resource for repositioning her actions in a way that promotes alignment. The results of this study are in line with CA work that has shown that “close and systematic attention to phonetic detail can open the way for a better understanding of the organization of conversational interaction” (Local and Kelly 1986: 204, cf. Couper-Kuhlen and S elting 1996; Egbert et al.). A case in point is Sorjonen ’s analysis of the Finnish response particle “nii” (Sorjonen 1997, 2001). Sorjonen suggested that intonational variants of this response particle embody different epistemic claims. When “nii” is spoken with a falling intonation contour in an environment where a display of recognition of a referent is relevant it claims
140
Anna Lindström
recognition. The “nii” in Extract 4 occurs after Asta has referred to a third (nonpresent) party by first and last name. (4) (Tuire/Danish Visitor) (Sorjonen 1997:431–432). Simplified transcript, syntactic glossing line omitted. 01 T:
Haloo?, Hello?,
02 A:
.mth ö Tuire Haimakainen?,=
03 T:
=.mt >Puhelime-ssa,< =.tch >Speaking,<
04 A:
.hh Asta Kiiski täällä hei .hh This is Asta Kiiski hi:.
05 T:
>↑Hei,< >Hi:,<
06 A:
.hh Kuule tota: .hh Listen uh:
07 T:
<.nhh[h>
08 A: 09
[e toi: (0.2) Marja Niemelä, er tha:t (0.2) Marja Niemelä,
(0.5)
10 T: → >Nii.< 11 A:
on otta-nu yhteyt-tä (.) semmose-s asia-ssa --has contacted me (.) by reference to ---
Sorjonen argued that in withholding talk during the silence in line 9, Asta is showing an orientation to the relevance of a display of recognition by Tuire. Tuire’s “Nii” in the next line (line 10) is produced with falling intonation. It claims recognition of Marja Niemelä (Sacks and Schegloff 1979). This analysis is supported by Asta’s next turn where she continues without further describing Marja Niemelä. When “nii” is used as a response to utterances that contain a referent that is treated as unfamiliar to the recipient by contrast it is delivered with a nonfalling terminal contour. In these cases, “nii” receipts the entire prior utterance as part of a larger incomplete unit of talk and invites the recipient to continue with that unit. This “nii” thus embodies a weaker epistemic claim than the “nii” that is produced with falling intonation. Extract 5 shows a “nii” that is spoken with a nonfalling terminal contour.
Projecting nonalignment in conversation
141
(5) (Sorjonen 1997: 438). Simplified transcript, syntactic glossing line omitted. 01 T: – ja sit so-i puhelin ja .hhh jah tota (.) siel and then the phone rang and .hhh and well (.) there 02 ol-i semmonen (.) tyttö joka me ol-la-an tava-ttu sillo was this (.) girl whom we met 03 neljä vuo-tta sitte Rooma-ssa. four years ago in Rome 04 A: Nii?, 05 T: .mh Ja tota uusseelanti-lainen tyttö. .hh ni se (0.2) .mh And um a New Zealand girl. .hh she (0.2) 06 ol-i tulo-ssa Suome-en pyöräl-ile-mä-än, was coming to Finland to bike, 07 A: Äö:?, Ea:?
Tiina is telling Arto about her midsummer celebration. In her turn in lines 1–3 she introduces a character who will feature centrally in her telling with a nonrecognitional reference (Sacks and Schegloff 1979): “this (.) girl whom we met four years ago in Rome.” Arto responds with “Nii” that is produced with a slightly rising terminal contour. Sorjonen argued that this response token treats Tiina’s talk as incomplete and invites her to continue. This also entails an agreement with the epistemic assumption of the prior utterance, i.e. that Arto does not know the person who Tiina introduced in her prior turn. This claim is supported by Tiina’s next turn where she continues her story by characterizing the person she introduced in lines 1–3. A comparison of recipient uptake after “nii” in these two examples demonstrate that variations in prosody alter how otherwise identical lexical items are understood. Beginning with Sacks (1987), I reviewed early CA research that has established a link between contiguity and agreement in conversation. Sacks (1987) and Pomerantz (1975, 1978, 1984) showed that agreement can be a negotiated outcome by describing how speakers revise the action of a prior turn to accommodate the displayed stance of their recipient. Another set of CA studies have explored the link between prosody and interaction by showing that the prosodic features of response tokens are sequentially and interactionally significant. Sorjonen (1997, 2001) argued that intonational variants of the
142
Anna Lindström
Finnish response particle “nii” embody different epistemic claims. When produced with a falling intonation, “nii” can be heard as a claim to recognition of a referent. When delivered with nonfalling terminal intonation by contrast, it invites the recipient of the “nii” turn to continue with the talk that was begun in her prior turn. The following section is devoted to establishing the central claim of this chapter, namely that the curled “ja” projects nonalignment. I do this by first exploring a range of cases in which the curled “ja” initiates a turn where a nonaligning action is implemented. Second, I examine deviant cases where the recipient claims the floor interruptively immediately after the curled “ja” has been produced. By using this position to revise the stance enacted in the first pair part of the curled “ja” sequence, recipients display that the curled “ja” was understood as projecting a nonaligning responsive stance . Data The data consists of audio recordings of naturally occurring telephone conversation which were made in three Swedish families during 1991. The conversations that figure in this corpus were made between landline telephones, and none of the families that participated in the study used telephones with automatic caller identification. The corpus that forms the basis for this chapter consists of twenty-five sequences where the curled “ja” occurs in turn-initial position of a turn in which a responsive action is due. Initiative actions in the curled “ja” sequence include information requests , action requests, advice , assessments, offers , and proposals. These actions are accomplished through a variety of linguistic devices such as interrogatives, statements, and follow-up questions. Unlike syntactic interrogatives, follow-up questions do not include a subject. They are constructed with a clausal or phrasal TCU and are syntactically and semantically parasitic of the sequential context (cf. Extract 3, lines 4 and 6). All first pair parts (FPPs) in the curled “ja” collection are built to take a “yes”/“no”-type response. Analysis Prima facie evidence for the argument that the curled “ja” projects nonalignment is that the curled “ja” can frequently be found in
Projecting nonalignment in conversation
143
the initial position of turns which implement a nonaligning action. The curled “ja” prefaces three types of nonaligning actions: hedges, pre-rejection s, and rejections. In the following I will explicate how nonalignment is done with the curled “ja” turn in each of these categories . I have used the term “nonaligning” rather than “disaligning” in this chapter to capture the most prevalent action initiated with curled “ja” turns namely the hedge. Twelve of the curled “ja” turns in the corpus either preface an actual hedge or are oriented to as projecting a hedging response. A hedging response does not fully align with the prior turn nor does it reject it. This is shown in Extract 6. (6) Wallpaper [GRU:7:B]. Tore and Cajsa are building a house together with Tore doing most of the construction. He is calling Cajsa at work to give her an update on the work. Parts of the interior walls of the house are being covered with panelling. 39 T:
=.hhhh Få- få hålla på å sätta upp den här .hhhh Have- have hold on and set up this here .hhhh Will- will have to put up this
40
panelen i da ‘rå paneling in day then paneling today then
41 C:
Ja:? Yes
42 T:
Mm:,
43
(.)
44 C:
Då: blir de panelat färdit i helgen då, Then becomes it panelled ready in weekend then Then it’ll be all panelled this weekend then
45 T: → Aa: de vete tusan, ((creaky voice)) Aa: that know thousand Aa: lord knows 46 C:
.hhh För ja börja fundera me Leif också de .hhh For I began wonder with Leif also this here .hhh Cause I started to wonder with Leif and the
47
här me: tapetserarbordet å hhh here with the wallpapering table and hhh stuff about the wallpapering table and hhh
48
allting, everything
144
Anna Lindström
Here Tore describes the construction work that he will do in his turn at lines 39–40, “Få- få hålla på å sätta upp den här panelen i da rå”/“Will have to put up this paneling today then.” Cajsa receipts this with an acknowledgment (line 41). Tore’s subsequent third-position acknowledgment may propose topic closure (see Hayashi and Yoon, this volume). After a micropause, Cajsa reopens the previous topic by drawing an inference from the prior talk “Då: blir de panelat färdit i helgen då”/“Then it’ll be all panelled this weekend then.” Tore responds with a curled “ja” followed by the idiomatic expression “de vete tusan”/“Lord knows.” This hedging response casts doubt on Cajsa’s inference without actually disagreeing with it. Cajsa orients to Tore’s curled “ja” response as nonaligning by accounting for why it is relevant for her to propose that the panelling should be finished by the weekend (line 46–48). Extract 7 is from a conversation between six-year-old Henrik and his father, Torkel. The curled “ja” turn is produced as a response to the father’s question whether Henrik will return home (lines 2 and 4). (7) Come home some time [VAT:11: B]. Henrik is calling home from a friend’s house. 01 H:
H[EJ’RÅ, Hi’then Bye then
02 T: [Kom- kommer’ru hem sen då¿ Com- come’you home later then Are- are you coming home later then 03
(0.3)
04 T:
Nån gång, Some time At some point
05 H: → Ja:hh ja vet inte, Ja:hh I know not Ja:hh I don’t know 06 T:
Ja:¿ Yes
Henrik proposes a conversational closing in line 1 by initiating a good-bye sequence. Torkel claims the floor interruptively to ask
Projecting nonalignment in conversation
145
Henrik, “Kom- kommer’ru hem sen då”/“Are- are you coming home later then.” Faced with no uptake from Henrik (line 3), Torkel recompletes his prior turn with the adverbial phrase “Nån gång”/ “At some point.” While the question in line 2 made the specification of a time possible but not necessarily relevant, this question merely requests a “yes”/“no”- type response. The recompletion can thus be hearable as adjusting to Henrik’s lack of uptake in line 3. Henrik responds in line 5 with a curled “ja” turn. The curled “ja” is separated from the second TCU by an audible outbreath. Torkel treats the curled “ja” as incomplete by refraining from claiming the floor at this point. Henrik continues with a disclaimer, “ja vet inte”/ “I don’t know” that neither affirms nor denies Torkel’s question. Another hedging curled “ja” response is shown in Extract 8. Torkel is calling his wife Liv at work to ask her what clothes their son should wear for soccer practice. (8) Pants and trousers (VAT:11:B). 01 T:
.h Va heter’e va skulle Henrik ha på sej What call’t what should Henrik have on himself What was I gonna say what should Henrik wear
02
de grö:na¿ the green the green stuff
03 L:
Ja:: fö[r han lär väl ha långa byxor, Yes for he must väl have long pants Yes because he should probably wear pants
04 T:
[VicketWhichWhat-
Short spate of talk omitted where Torkel and Liv dispute the colors of the different items of clothing.
15 T:
Bå:de byxorna å tröjan? Both the pants and the sweater
16 L: → Ja: han kan väl ta tröjan runt mi:djan Ja: he can väl take the sweater around the waist Ja: he can at least put the sweater around the waist 17 → åtminstone om de bir kallt °får han väl at least if it gets cold gets he väl if it gets cold he can
146
Anna Lindström
18 → sätta på° sen har han väl nån svarta put on then has he väl some black put it on then he has some black 19 → te:shirt eller fotbollste:shirtar t-shirt or soccer t-shirts 20 → ligger i garderoben, lying in the closet 21 T:
°Ja°¿ Yes
Torkel requests confirmation in lines 1–2 by asking whether Henrik should wear “the green stuff”: “Va heter’e va skulle Henrik ha på sej de grö:na”/“What was I gonna say what should Henrik wear the green stuff.” Liv confirms in the next turn noting the need to wear pants. In Swedish, “långbyxor”/“long pants” are contrastive with “kortbyxor”/“short pants or shorts” (see the literal translation of line 3). A side sequence ensues with Torkel disputing Liv’s characterization of the color of the clothing (lines 4–14). Torkel then asks with a follow-up question in line 15 whether Henrik should wear both the pants and the sweater: “Bå:de byxorna å tröjan”/“Both the pants and the sweater.” Liv gives a curled “ja” response that neither completely aligns with nor disconfirms Torkel’s question by stating that Henrik should bring the sweater but not put it on: “han kan väl ta tröjan runt mi:djan”/“he can at least put the sweater around the waist.” In Extracts 6–8, the curled “ja” is followed by a hedge that neither aligns with the FPP nor completely rejects it. The next category of curled “ja” responses that I will discuss, pre-rejection s, are more strongly disaffiliative. There are eight pre-rejection relevant instances of curled “ja” in the corpus. In Extract 9, Jan is calling his sister-in-law Vera in Sweden from a London airport. (9) Landvetter [MOL:3:B]. Vera and her family live within two hours driving distance of Landvetter airport. Ulla, mentioned in line 24, is Vera’s sister and Jan’s fiancee. 16 J:
Ja ja e: i London allså, Yes I ‘m in London see/in other words Yes I’m in London you see
Projecting nonalignment in conversation 17 V:
↑E:: ’RU:¿ Are you You are
18 J:
Så ja tänkte fråga (.) pt om eh- (0.4) om ni So I thought ask (.) pt if eh- (0.4) if you So I was going to ask (.) pt if eh (0.4) if you
19
kunde- (.) ja skulle behöva liksom ha: could- (.) I would need kind of have could- (.) I would sort of need
20
h skjuss från Landvetter h ride from Landvetter h a ride from Landvetter
147
21 V: → .hh Aa: när då ’rå? .hh Aa: when then then .hh Aa: so what time would that be then Ja kommer klockan- vid midnatt kommer ja22 J: I come clock- at midnight come I I arrive at- around midnight I come- at 23
klockan tolv, clock twelve twelve o’clock
24 V: → pt heh heh heh ˙hh Aa:: ˙hh eh kanske U:lla kommer? pt heh heh heh ˙hh Aa:: ˙hh eh maybe Ulla comes pt heh heh heh ˙hh Aa::˙hh eh Ulla might come 25 J:
Ja just de, Yes right that Yes right
Vera treats Jan’s announcement that he is in London as news in line 17. Instead of responding to the newsmark, Jan begins to make a request for a ride from the airport beginning with a preliminary “Så ja tänkte fråga”/“So I was going to ask,” which marks the upcoming action as delicate (Schegloff 1980). Jan’s request for a ride from the airport imposes on Vera in at least two ways. First, it is made late (i.e. on the day of Jan’s arrival). Second, compliance with the request would require Vera to make a two-hour drive to the airport in the middle of the night. Jan’s orientation to the first of these issues is evident in that later in the conversation he indeed states that he had been trying to reach Vera earlier. The request
148
Anna Lindström
includes two self-interruptions “pt om eh-”/“pt if eh-” and “om ni kunde-”/“if you could-.” The second of the self-interruptions could be heard to be going to a direct request “if you could pick me up at the airport.” This formulation is dropped for an indirect one that states a need, “ja skulle behöva liksom ha:h skjuss från Landvetter”/“I would sort of need a ride from Landvetter.” Vera responds to the request with a curled “ja” turn, “Aa: när då rå,” literally translated as “.hh Aa: when then then.” The duplication of the inference marker emphasizes that Vera’s willingness to grant the request is contingent. Like the request in lines 18–20, Jan’s answer to Vera’s question is hesitant and replete with selfrepairs, “Ja kommer klockan- vid midnatt kommer ja- klockan tolv”/“I arrive at‑ around midnight I come- at twelve o’clock.” This may indicate an orientation to the time as problematic as does Vera’s laughter in line 24. Rather than granting the request, Vera states equivocally with another curled “ja” turn that “kanske U:lla kommer”/“Ulla might come” in line 24. Jan’s neutral uptake to this turn “Yes right” is compatible with the analysis that the request has not been granted. The curled “ja” is followed by a rejection in the next two examples (10 and 11). (10) Dagens Nyheter (MOL:2:A). From a telephone call to a street vendor who mans a newsstand where papers and fast food is sold. The caller (C) has just asked the street vendor (S) whether a copy of last Sunday’s newspaper is still available (he is calling after Sunday). 08 S: [De kan no hända men iså:fall finns de It can probably happen but in that case is there It may be possible but in that case there won’t be 09
ingen e:xtrabila:ga till den utan då e’re no extraenclosure to it but then ‘t’s any magazine section then it’s
10
bara Dagens Ny:heter¿ only Dagens Nyheter just Dagens Nyheter
11 C: Ja bara Dagens Ny:heter ja¿= Yes just Dagens Ny:heter yes 12 S:
=Jo men de [tror ja att vi ha:r,] =Jo but it think I that we have =Jo but I think we have that
Projecting nonalignment in conversation 13 C: 14
[Ka- kan Ca- can Ca- can
du you you
149 ] titt(h)a:, look go check
(.)
15 S: → Ja:: de fi:nns Ja:: it exists Ja:: it’s there 16 C:
KH[HH
17 S: [ja ve:t de. I know that I know it 18 C:
.hh Eh nu e’re så här att ja: sitter i- i Orvenbo .hh Eh now “s’t so here that I sit in- in Orvenbo .hh Eh now I am in- in Orvenbo now
19
här å kommer väl inte iväg förrn om en timma here and come probably not away until about one hour and I probably won’t leave for about an hour
20
när stänger’u? when close you when do you close
The street vendor’s response in lines 8–10 suggests that parts of the Sunday paper are still available. When the caller replies that he is only interested in the paper itself (without the magazine section), the street vendor states that he believes he has it, Jo men de tror ja att vi ha:r”/“Jo but I think we have that.” The caller interrupts the street vendor’s turn in progress to request that he go look for the paper, “Ka- kan du] titt(h)a:”/“Ca- can you go check.” The street vendor’s response to the request is slightly delayed. Instead of granting the request, the street vendor asserts with a curled “ja” prefaced turn that he has the paper, “Ja:: de fi:nns”/“Ja:: it’s there.” Upon hearing this, the caller lets out a big sigh (line 16). This is treated as a display of irritation by the street vendor who retorts with an upgraded epistemic assertion, “ja ve:t de”/“I know that.” A different kind of rejection is shown in Extract 11. (11) Tax money (MOL:4:A). Runar has just asked his nephew Allan whether Allan’s father (Runar’s brother), Orvar has returned home. Orvar is a medical doctor employed by the Swedish government who apparently has been out on business travels. Allan responds that he does
150
Anna Lindström
not know whether Orvar has returned. This response is acknowledged and the sequence continues as follows. 34 R:
Han e bara ute å (lurvar)¿ He is just out and putzes He is out (all the time) putzing around
35 A:
pt Aa: han e bara ute å förlustar se:j, pt Yes he is just out and covorts himself pt yes he is out covorting (all the time)
36 R:
Ja*. ((*=glottal stop)) Yes
37 A:
Gör av me våra skattepengar. Makes off with our taxmoney Spends our tax money
38 R: → Ja::= 39 A:
=˙hA,= =Yes=
˙h ja har (precis en) känsla 40 R: → =Hörrudu- =Listenyouyou- ˙h I have (exactly a) feeling =Listen up ˙h I have a certain feeling 41 R:
av att [du också gör de, of that you also do that that you do that as well
42 A: 43 A:
[˙hh
Att ja: också hh g(h)ö- [heh heh heh˙heh heh heh That I also hh d(h)o- heh heh heh˙heh heh heh That I also spend-
44 R:
[HAH HAH HAH HAH
45 A:
J(h)a:(h) de kanske e nåt släktdra:g J(h)a:(h) it perhaps is some familytrait (Yes) maybe it is a family trait
46
hörru[du¿ listenyouyou (don’t you think)
In line 34, Runar jokingly evaluates Orvar’s travels, “Han e bara ute å (lurvar)”/“He is out (all the time) putzing around.” The Swedish word “lurvar” denotes walking or moving about without a sense of purpose. Allan responds to Runar’s turn with an upgraded assessment that treats Orvar’s travels as more illicit, “han e bara
Projecting nonalignment in conversation
151
ute å förlustar se:j”/“he is out covorting (all the time).” After an acknowledgment by Runar in line 36, Allan reclaims the floor to make another assessment that is stronger and more negative than the other two as it underscores that Orvar’s travels have a negative impact on others: “Gör av me våra skattepengar”/“spends our tax money.” The assessment makes relevant a second assessment (Pomerantz 1975, 1984). To merely discuss this spate of talk in terms of the relevance of a second assessment, however, does not fully capture the social intricacies involved. In a study of American family dinner conversations, Ochs et al. (1989, 1992) argued that one resource for building solidarity and cohesiveness within the family was to align the members of the family against a third nonpresent party. The possessive pronoun in line 37, “våra”/“our” could be heard to refer to taxpayers in general as well as Runar and Allan in particular. Allan thereby builds an alliance with Runar against Orvar (who is Runar’s brother and Allan’s father). With his assessment, Allan invites Runar to express solidarity with him (via a prefer ence for agreement ). However, in response to this, Runar produces a curled “ja” in line 38. On the immediate completion of the curled “ja,” Allan reclaims the floor with “.hA.” Runar orients to Allan’s turn as an interruption. The turn-initial summons in line 40, “Hörrudu-“/“listen up” is doubly contextual (Heritage 1984b). It links back to the prior turn by admonishing Allan for claiming floor prematurely. It also points forward by marking the upcoming action as problematic (Schegloff 1970, 1995b, 2007b). Runar agrees with Allan’s assessment but rejects the alliance by suggesting that Allan is just as bad as Orvar in his next TCU, “ja har (precis en) känsla av att du också gör de,” “I have a certain feeling that you do that as well.” Although Runar produces this without laughter, Allan treats it as humorous by laughing in his next turn and then attempts to turn the tables on Runar by suggesting that this is a family trait (lines 45–46). When the curled “ja” turn is brought to completion, it typically implements a nonaligning action. Most commonly it does this with a hedge that teeters between alignment and nonalignment. My collection also includes curled “ja” turns that are prerejections. Only two curled “ja”s in my database are followed by rejection. This distribution suggests that the curled “ja” is more closely associated with minor nonalignment than with rejection
152
Anna Lindström
or disagreement per se. That the curled “ja” typically precedes a nonaligning action does not prove that the curled “ja” response token projects nonalignment. To establish the latter point, we must isolate the recipient’s orientation toward the curled “ja” in and of itself. In the next set of examples, the recipient claims the floor after the curled “ja” is produced but before the remainder of the turn it prefaces has been articulated. The actions they engage in show that they orient to the curled “ja” as foreshadowing a nonaligning response. The analysis presented in the following builds on and contributes to CA studies that have established that interactants orient to the emergent and dynamic quality of each turn at talk, parsing it for the action(s) that are being accomplished (cf. Goodwin and Goodwin 1992; Heritage 1988; Jefferson 1973; Lerner 1987; Sacks et al. 1974; Schegloff et al. 1996). Schegloff (1996c) observed that any utterance in conversation may be understood to go through three phases: as (incipient) next, as current, and as prior. That is, as a currentrecipient-of some-talk/potential-next-speaker parses it in the course of its progressive articulation, potential response types and lines are engendered, subject to revision and replacement as the current talk is further produced bit by bit. (Schegloff 1996c: 97)
Recipients’ orientation to the bit-by-bit construction of talk becomes particularly clear when one examines talk produced in overlap. As Jefferson (1973, 1983) has shown, overlap onset is precisely placed with respect to the turn that is already in progress, and its resolution is delicately managed. Discussing overlap, Lerner (1996) argued: Just as each next turn in a conversation is ordinarily designed to follow the just prior turn, so a recipient’s utterance begun internal to a turn’s talk (i.e. in the midst of the projected turn space of the current speaker) is shaped by, and understood by reference to, its placement. (Lerner 1996: 251)
In Extract 12, the recipient of the curled “ja” turn enters the turn space of her interlocutor. The placement of her turn relative to the turn-in-progress coupled with the kind of action she performs lends support to my argument that curled “ja” is understood to project nonalignment. The example is from a conversation between two young mothers, Nina and Liv. Nina’s young children are in
Projecting nonalignment in conversation
153
Liv’s home-based day care. Nina is a nursing aide who works on a shift schedule that involves night work. Her husband (Janne) is a teacher. In the excerpt below they are discussing Nina and Janne’s prospective work schedules. (12) Work schedule [VAT:4:B]. Nina is in the midst of describing days when her and her husband’s working hours are so spread out that they do not need to leave their young children in day care. 01 N: [>Å likadant som tillexempel< >And same as for example< >And it’s the same if say< 02 om ja börjar- när ja börjar fy:ra när ja har if I begin- when I begin four when I have if I begin- when I begin at four when I have 03
natten¿ the night the night shift
04 L:
Ja:? Yes
05
(0.4)
06 N: (Ja) behöver ju inte åka före (0.2) tjugi i fyra,= (I) need ju not go before twenty to four I would not need to go earlier then twenty to four right 07 i fyra,= to four to four right 08 L:
=Å [Janne kommer ha:lv fem, =And Janne comes half five =And Janne comes home four thirty
09 N:
[(De e ju-) (It’s ju-)
10 N:
[Ja seYes laYes at the la-
10 L:
[Före halv fem, Before half five Before four thirty
154 11 N:
Anna Lindström Senast [alltså, Latest that is At the latest that is
12 L: 13 L:
[Ja:, Yes .hh Ja han [slutar fyra då, .hh Yes he quits four then .hh Yes he gets off at four then
14 N:
[(Då hinner ja) (Then manage I) (Then I’ll have the time)
Ne::j när slutar’om slutar före 15 L: No when quit they quit before four what No when do they get off they get off before four don’t they 16
fyra va? four what four don’t they
olika )] 17 N: → Ja::: d[e (e lite Ja::: i t (‘s little different) Ja::: it (depends they have-) 18 L: → 19 N:
[De bero:r lite på- a::¿] It depends little on yes It depends yes
Dom har inte riktit (.) samma ti:der, They have not really (.) same times They don’t have exactly (.) the same times
Nä:e, 20 L: No
In line 13, Liv states that Nina’s husband’s workday ends at four o’clock. This is offered as a concessive inference across a series of turns beginning in line 8. Although her turn in line 13 is formatted as a statement, it could be heard as a question because it focuses on a B-event (Labov and Fanshel 1977). Instead of answering Liv, Nina reclaims the floor interruptively to continue her own previous turn by beginning to formulate its upshot, “(Då hinner ja)”/“Then I’ll have the time.” She yields the floor to Liv before her turn has been brought to completion. In lines 15–16, Liv self-corrects the time reference she gave in her previous turn, “Ne::j när slutar’om
Projecting nonalignment in conversation
155
slutar före fyra va”/“No when do they get off they get off before four don’t they.” Here, Liv uses a locally subsequent reference form “they,” in locally initial position (Schegloff 1996a). Presumably she is referring to Janne and his colleagues at the school. The turn is done as a self-correction through the turn-initial “ Ne::j”/“No,” and the stress on före/“before.” In contrast with her turn in line 13, which also focused on Janne’s work schedule, this turn is formatted as a syntactic interrogative and includes a turn-final tag, “va,” and is thus hearable as in pursuit of a response which Nina indeed begins to provide in line 17 with a turned prefaced by a curled “ja”. Liv reclaims the floor just after the articulation of the first consonant of Nina’s second TCU, i.e. after “de e lite olika” (literally translated as “it’s little different”). The “early” placement of Liv’s turn is not haphazard but sequentially motivated. Specifically, this is an instance of a recognition point entry. Lerner (1996) proposed that one feature of many recognition point entries is that they are done not merely as responses to a forecasted TCU, but are used in circumstances where it would be felicitous to pre-empt the current speaker’s utterance before completion – i.e. to forestall the action that is currently underway and recognizable as part of the TCU’s projectability. (Lerner 1996: 252)
The positioning of Liv’s turn in line 18 is thus not early but precisely placed to pre-empt the nonaligning stance that was projected with the curled “ja.” While the position of Liv’s turn claims an understanding of the action projected by the curled “ja” it does not demonstrate that understanding. The demonstration is achieved through the utterance itself, “De bero:r lite på- a::”/“it depends a little on the circumstances yes,” which is spoken in overlap with and aligns with the second TCU of Nina’s turn, “de e lite olika”/“it’s little different.” By capitalizing on the nonalignment projected by the curled “ja,” Liv has thus been able to maneuver herself from a position of projected nonalignment to actual alignment with her co-participant. Extract 13 is from a conversation between Malena and Pihlander. Malena has just bought a used car from Pihlander. Pihlander is calling Malena to remind her to register the car in her name as he
156
Anna Lindström
otherwise would be responsible for any traffic violations that she may incur. (13) Gasoline (GRU:4:B). After having underscored the importance of changing the registration, Pihlander informs Malena that he has an extra set of keys for the car that she can pick up at her own convenience. Once these two issues have been settled, Malena raises a new topic in line 42 by asking Pihlander what grade of gasoline he has been using in the car. “Ninety-six” and “ninety-eight” refer to two grades of gasoline with ninety-eight being the more expensive grade. 39 P:
[>Får hämta dom nån gång då sen< >Can fetch them some time then later< (You) can pick them up at some later time then
40 M:
Ja? Yes
41 P:
A[a:,
42 M: [˙hh Du “ja” tänkte på bensi:n (eh*) visst ˙hh You I thought on gas (eh) no doubt ˙hh Hey I was thinking about gas you usually do 43
tankar du s- nittisex? fill you s- ninetysix fill up with s- ninetysix I suspect
44
(0.6)
45 P: → Ja::[(h) 46 M: → 47 P:
[Eller tanka’ru nittiåtta?= Or fill you ninetyeight Or do you fill up with ninetyeight
=Ne:j nittisex, =No ninetysix
48 M: Nittisex (bör ja väl [kunna tanka’rå,) Ninetysix ought I probably able to gas’then I ought to be able to fill up with ninetysix then
Malena’s question in lines 42–43 is built to take an affirmative response through the adverb “visst.” Pihlander begins to respond in the arrowed turn with a curled “ja.” Like the recipient of the curled “ja” turn in Extract 12), Malena revises her prior turn upon hearing the curled “ja.” She does this by giving a contrastive alternative to the one she had offered in line 43, “Eller tanka’ru nittiåtta”/“Or
Projecting nonalignment in conversation
157
do you fill up with ninetyeight.” Her turn is built as a revised query through its placement (immediately after Pihlander’s curled “ja”) and through the use of the turn-initial “ Eller”/“Or.” In revising her query, Malena is trying to head off the nonaligning response that she heard to be underway at line 45. Pihlander then goes on to assert with an other-correction in line 47 that he indeed uses 96-octane gasoline. In this example, then, the curled “ja” was not produced (or “intended”) to project nonalignment. That Malena nonetheless heard it that way underscores that for the negotiation of agreement and disagreement , and indeed for interactional practices generally, it is not “the intention” but the interactional effect that matters. Discussion When used in turn-initial position of a responsive turn where a display of alignment is relevant, curled “ja” projects nonalignment with the action of the prior turn. Since it projects rather than accomplishes nonalignment, the curled “ja” furnishes a resource for recipients as they can recast their prior action upon hearing the curled “ja” to change the sequence trajectory from projected nonalignment to actual alignment. This was illustrated by Extract 12, where the recipient reclaimed the floor upon hearing the curled “ja” to revise the position she had assumed in her prior turn. As my analysis showed, the placement of her turn was crucial in that it allowed her to maneuver herself from a position of projected nonalignment to actual alignment with her co-participant. The curled “ja” is therefore an important tool for the achievement of alignment and affiliation as a negotiated outcome. This is the first empirical study of curled ja. Future studies could examine the syntactic form of the FPP to see whether it shapes the way alignment and nonalignment are done within the responsive turn. For the purposes of the analysis presented in this chapter I have not found any reason for treating “ja:” and “aa:” as interactionally distinct. However, this should be examined in future work. I suggested that the curled “ja” may be more closely associated with slight nonalignment than outright disagreement or rejection. This should also be explored in future studies.
158
Anna Lindström
Acknowledgments The study draws on analyses originally presented in my doctoral dissertation (Lindström 1997, 1999). I would like to thank Emanuel Schegloff, John Heritage, Steve Clayman, Trine Heinemann, and Jack Sidnell for valuable comments on earlier versions of this manuscript. This study is one of several investigations in the project Language and Social Action: A Comparative Study of Affiliation and Disaffiliation Across National Communities and Institutional Contexts (financed by the Swedish Council for Working Life and Social Research and the Swedish Research Council).
6 Two answers to inapposite inquiries Trine Heinemann
Introduction In recent years, one major focus of conversation analysis (CA) has been to investigate how knowledge is distributed and negotiated among participants in interaction. Research has demonstrated that for conversationalists what each of them knows and how they know it are accountable matters (Heritage 1984a, 1998; Drew 1991; Schegloff 1996b; Roth 2002; Heritage and Raymond 2005; Raymond and Heritage 2006; Stivers 2005; Clift 2006a, 2006b; Golato and Fagyal 2008 ). As is the case for most CA research, the majority of these studies have been based on the analysis of English data. However, as Schegloff (2006) argues, human beings as a species face the same problems and issues in everyday life independently of language or culture. As “the primary, fundamental embodiment of sociality” (Schegloff 2006: 70), interaction is designed to address these problems and issues, though the way in which this gets done varies across languages or cultures (cf. Schegloff 2002a). It thus stands to reason that practices parallel or analogous to those described for English exist in other languages. Once identified, such practices may provide a basis for cross-linguistic comparison – allowing us to see the way in which the local resources of a particular language are mobilized to solve generic problems of interaction. In this chapter, I explore how Danish speakers orient to knowledge in interaction. Specifically, I identify two Danish practices, through which a recipient can index that a question is inapposite, because it inquires into something that should already be known to the questioner, and compare these with one previously described for English.
160
Trine Heinemann
In what follows, I first discuss the use of the English changeof‑state token “oh” (Heritage 1984a, 1998). I then identify the Danish practices used for accomplishing equivalent actions. Having established that Danish, like English, has a practice for conveying that a question targets something that should already be known to the questioner, I proceed to identify another Danish practice used for the same purpose. This second practice constitutes a stronger implementation of inappositeness than the English “oh” because speakers here explicitly reveal why a question ought not to have been asked. The chapter concludes with a discussion of how the two practices reflect the different ways in which a question can be understood to be inapposite and provides some suggestions as to how this difference may be reflected in other languages. Answers that index a question as inapposite The English change-of-state token “oh” and its use in question–answer sequences Question–answer sequences are concerned in a basic way with the exchange of information and, as such, invoke the relevance of who knows what. In the following extract, both questioner and answerer display themselves as being more or less knowledgeable in relation to each other. First, J asks a question about the future activities of a person called Susan. J thus portrays herself as less knowledgeable about this matter, than the recipient, M. In turn, M aligns with the expectation that she is more knowledgeable than J by providing an answer to the question. Finally, by producing the change-of-state token “oh” as a receipt of the answer, J claims to have been informed on a matter where she was previously uninformed (Heritage 1984a). (1) (WPC:1:MJ(1):1) (Heritage 1984a, ex. 20, pp. 308) 01 02 03 04 05 06
J: When d’z Sus’n g[o back.= M: [.hhhh J: =[( ) M: =[u-She:goes back on Satida:y= J: =O[h:. M: [A:n’ Stev’n w’z here (.) all las’week…
In Extract 1, the distribution of knowledge is clear-cut, and the participants’ roles as questioner and answerer are in alignment
Answers to inapposite inquiries
161
with their display of who is the least and the most knowledgeable. This is not always the case. Interaction is abundant, with instances in which a speaker asks a question to which she already knows the answer. In ordinary, everyday conversations, such questions are commonly used as vehicles for actions such as challenging the recipient (Heritage 2002; Heinemann, 2008 ) or making a direct complaint (Monzoni, in press ).1 Even when they do not serve directly as vehicles for disprefer red actions, answerers often treat such questions as inapposite and problematic. One way in which recipients mark this, in English, is by prefacing the answer with “oh”. In the following example, Steven’s description of Robert Maxwell’s parents as having been killed by the Nazis (→1) causes Lesley to make the inference that Maxwell is of Jewish origin (→2). Steven confirms this with an “oh”-prefaced response (→3), which marks the answer as self-evident and the question as one that need not have been asked (Heritage 1998). (2) (Field:2:3:9) (Drew and Holt 1998, ex. 6, pp. 500)2 01 Ste: →1 Well he didn’t either ‘ee had a bad start I mean ‘ee had 02 ‘iz (0.3) .t.k.hh father shot by the Nazis ‘nd ‘is uh .hh 03 mother died in:Auschwitz yih kno:w [so 04 Les: [Oh really:?= 05 Ste: =So ‘eez [had the:( ] )06 Les: →2 [Oh ‘z a Je:w] is he Je:w? 07 (.) 08 Ste: →3 Oh yeah. 09 (.) 10 Ste: He’s had k- eez a Czechoslovakian Jew so so [eez had 11 Les: [Yes 12 Ste: k- eez had quite a- checkered career already=
“Oh” is used in different sequential positions in Extracts 1 and 2 to implement different actions. In Extract 1, it occurs in third position, after a response to a question, and is used to claim that the speaker has been informed about something she was previously uninformed about. In Extract 2, by contrast, “oh” occurs in second position, as a preface to the answer to a question, and it indicates that the question, by inquiring into something that was already available to the questioner in the preceding sequence of talk, is inapposite. Heritage (1998) argues that “oh” has an overarching function as a change-of-state token, where variations such as those between
162
Trine Heinemann
Extracts 1 and 2, “are managed through sequential and contextual particularizations of the particle’s placement” (Heritage 1998: 327). Thus, the “oh” in Extract 1 marks a shift in the state of the speaker’s knowledge , whereas the “oh” in Extract 2 marks a shift in the speaker’s state of attention (Heritage 1984a, 1998). “Oh” provides a solution to two different interactional problems in these extracts. In Extract 1, the problem that “oh”-prefacing addresses is that of how a participant can show that an answer to her question has been informative. In Extract 2, by contrast, “oh”-prefacing displays to a questioner that her question was inapposite. In the following section, I explore how these interactional problems are solved in Danish. The Danish change-of-state token “nå” The two functions that “oh”-prefacing is used for in English are accomplished through two different expressions in Danish. This language, like English, has what may be described as a changeof‑state token (see Nielsen 2002), i.e. one that can be used to mark that an answer to a question has informed the speaker where she was previously uninformed, as in the following extract. Here a question is asked (→1), an answer given (→2), and, in third position, the answer is receipted (→3). The extract comes from a face-to-face conversation between two female students, Birthe and Carina. In lines 1–3, Birthe inquires whether Carina has found any flats advertised in the newspaper. Initially, Carina, who is chewing some food, disconfirms this question with a headshake and a negative minimal response token “Mm mm” (line 5). She then readdresses the question with a confirming “Jo”/“Yes”, followed by a specification of the flat’s location. In line 11, Birthe receipts the information with the particle “nå”, registering that she has now been informed. (3) (DLAU:KC “Flat hunting”)3 01 Bir: →1 Angående den lejlighed d[er. ] Var der andre About that flat. [ ] Were there any others 02 Car: 03 Bir: →1 i avisen, in the newspaper, 04
(0.4)
05 Car:
Mm mm.
[Mm, ]
Answers to inapposite inquiries 06
163
(0.4)
07 Bir: Ne[j. ] N[o. ] 08 Car: →2 09
[Jo ]. Ude ve” Bilka, [Yes]. Out by Bilka, (0.4)
10 Car: →2 eller i Store Husby. or in Store Husby. 11 Bir: →3 Nå. Oh.
As this extract attests, “nå” works as a change-of-state token similar to “oh” in English (see Extract 1). In contrast to English “oh,” however, “nå” is never used as a response to an inquiry, to mark that the answer to the inquiry was already known.4 Instead, this is accomplished through the production of multiple response particles produced within the same prosodic unit, as demonstrated in the following section. Answers to inapposite questions in German and Danish It seems safe to assume that it is possible to ask a question to which the answer is already known in all languages, and prior research has established that this is certainly the case for Danish (Heinemann 2008 ). Consequently, it can be expected that Danish speakers have ways to index the inappositeness of such questions, just as English speakers have, even if this cannot be accomplished by using the change-of-state token “nå”. Heritage (1998: 294) describes the function of “oh”, when used in response to inquiry, as proposing “that the question questioned something which could (or should) be taken for granted.” A similar description of a Danish practice is found in Heinemann (2003). Here I note that multiple productions of the negative response particle “nej”/“no” within the same prosodic contour marks a previous turn “as in some way conveying something so obvious that it need not have been stated in the first place” (Heinemann 2003: 380). Likewise, Golato and Fagyal (2008 ) suggest that in German, certain multiple response particles are used to index a question as “requesting (obvious) information that was already presented or implied [ … ] in prior turns.”5
164
Trine Heinemann
Extracts 4 and 5 are instances of multiple response particles used to answer questions in German and Danish. When compared with the English Extract 2, we can see that in all three cases a question (→2) questions something that is implied in prior turns (→1), and that this inappositeness is indexed in the recipient’s answer (→3). As earlier noted for the English example (Extract 2), the fact that Maxwell had a father who was shot by the Nazis and a mother who died in Auschwitz strongly implies that he is of Jewish origin, and the question “Oh ‘z a Je:w is he Je:w?” questions something which is self-evident.6 Similarly, in the German example below, M, in lines 14–15, asserts that his dentist subjected him to unnecessary dental work. For M to know and assert this, he has to have been informed of this after the dental work was done. Thus, with the inquiry in line 18, “hat sich das nachher rausgestellt?”/“did that come to light later?”, G questions information that is already conveyed through M’s assertion. Here, M indexes the inappositeness of the question by producing two instances of the positive response particle “ja”/“yes” within the same prosodic contour.7 (4) (OregonT1A_514_offenerBiss) (Golato and Fagyal 2008 , ex. 8, pp. 258–259)8 05 M:
[.hh denn ähm denn nen zahnarzt, ich (g-) der [.hh cause uhm cause a dentist, i (g-) who
06
damals als ich bei mei- bei de- bei unserm back then when I was at my- at the- at our
07
zahnarzt war, .hh dentist, .hh
08 G:
mhm. uhum.
09 M:
zu dem meine ganze familie immer hingegangen is. to whom my entire family always went.
10
>und auch der kieferorthopäde=die ham mir ja da >and also the orthodontist=they quacked
11
im mu:nd< rumgedoktert, .hh around< in my mouth, .hh
13 G:
mhm. uhum.
14 M: →1 und ähm:ha:m sachen gemacht die ich gar nich and uhm:did stuff that I actually did not
Answers to inapposite inquiries
165
15 →1 ge↑braucht hab. .hh ↑need at all. 16 G:
ach so:? oh rea:lly?
17 M: ich hab äh:[m, I have uh:[m, 18 G: →2
[hat sich das nachher rausgestellt?= [did that come to light later?=
19 M: →3 =ja^ja. =yeah^yeah. =that’s ^right. 20
(.)
21 M:
.h der zahnarzt hat mich zum kieferorthopäden .h the dentist sent me to the orthodontist,
22
geschickt, und hat ä:hm hat gesacht ich bräucht ne a:nd said that I needed
23 M:
feste spange, braces,
In the following Danish example, Kirsten is describing a trip she took several years back. During this description, she mentions a church she visited and expresses her belief that this is the church where a friend’s husband is buried. This leads Gunnar to make the inference – somewhat delayed – that the friend’s husband is dead. Being buried clearly entails being dead. Hence, Gunnar’s inquiry “A’ han død”/“Is he dead” questions something that is conveyed in Kirsten’s prior turns at talk. As in the German example, this inappositeness of the question is indexed through the production of two positive response particles within the same prosodic contour. (5) (MITHVDP31Aug0606 “Dead and buried”) 01 Kir:
Å’ der lå nemlig en kirk’ som lå °så’n° And you see there was a church situated °like°
02
>meget meget meget< højt oppe.→Å’ der tror jeg< >very very very< high up.→And there I think<
03 →1 .hhh Der tror jeg faktisk øh Tine’s tidligere mand .hhh There I actually think eh Tine’s former husband 04 →1 °der han- e-° >al’så børnenes far< ligger °there he e-° >you know the children’s father< is
166
Trine Heinemann
05 →1 begravet, buried, 06
(1.1)
07 Gun: →2 A’ han død, Is he dead, 08 Kir: →3 Jah=jah. Jahm’ han døde .hh (2.0) Han var Yes=yes. Yes but he died .hh (2.0) He bloody 09
sgu ikk’ død da vi var herovre.= well wasn’t dead when we were over here.=
Extracts 2, 4, and 5 show how three languages – English, German, and Danish – allow speakers to index that they are answering a question the answer to which should already be known to the questioner. Whereas this is accomplished in English through prefacing the answer turn with the change-of-state token “oh”, in Danish (and German) this is done through the production of multiple response particles within the same prosodic unit. Generic features of inapposite questions and their answers Heritage (1998) demonstrates that the question–answer sequences in which “oh”-prefaced answers occur have a number of features in common beyond the fact that the question being asked is one to which the answer is already known. Here, I will briefly discuss how some of these features are also present in the Danish (and German) question–answer sequences . First, questions that receive an “oh”-prefaced answer are typically “mispositioned”, or “off the sequential track” (Heritage 1998: 307). This means that such questions are oblique, in that they target ancillary matters and have the potential to disrupt the progressivity of an ongoing activity (Koenig 2007). In Extract 2 above, for instance, Steve is describing the “bad start” to life that Robert Maxwell had, exemplifying this by providing the information that his parents were killed during World War II. Though the reason for their being killed was that they were Jewish, this is not central to the description of Maxwell’s “bad start”. When Leslie inquires into this matter, she suspends the progressivity of Steve’s extended description. This is emphasized by the position of the inquiry as being in overlap with – and interruptive of – Steve’s turn in line 5.
Answers to inapposite inquiries
167
Inapposite questions in German and Danish are similarly mispositioned. The inquiry in the German example is produced in overlap and suspends M’s description of the excessive dental work he has endured. In the Danish example, there is no overlap, and the inquiry is produced after a pause of 1.1 seconds. However, the matter of “Tine’s former husband” is specifically introduced as ancillary, through Kirsten’s use of the adverb “faktisk”/“actually” (see Clift 1999). When Gunnar, through his inquiry, focuses on this issue, Kirsten’s telling about her trip is suspended. That the questions in Extracts 2, 4, and 5 are oblique or mispositioned is evidenced by the fact that recipients return to the activity they were engaged in after providing an answer. In the English example, Steve resumes the description of Maxwell’s life with a repeat of the conjunction “so” then closes the description with the figurative expression “checkered career” (Drew and Holt 1998). Likewise, in the German example, M continues his description of the unnecessary dental work he endured, after answering G’s oblique question. In the Danish case, this type of evidence is not available, as Kirsten self-interrupts her answer when she realizes that though her friend’s husband is dead he may not have been buried (or dead) at the time at which she went to the church. However, in the following case, the answerer’s treatment of the question as mispositioned is evident. Extract 6 comes from a telephone call between the local mayor, JO and a council member, Jens. JO has called Jens to let him know that a political meeting, taking place the day after, has been moved forward so that Jens will be able to participate before going on to a committee meeting. As will become evident, however, Jens was under the impression that the political meeting was scheduled on the day of the phone call and is, in fact, ready to go to that meeting when JO calls. (6) (TH/S2/78 “Changing clothes”) 01 JO:
Ø::h Hvis vi så ku’ holde mødet der E::h If we could then have the meeting there
02
lidt i seks::, just before six,
03
(0.5)
04 Jen:
Jerh,= Yeah,=
168
Trine Heinemann
05 JO:
=ne’e ve’ Kurt, =at Kurt’s place,
06 Jen:
Jerh, Yeah,
07 JO:
Så ka’ du nå å’ komme te:eh:te’ bå’e det ene Then you can manage to get to both one
08
å’ d[et andet ikke¿] and t[he other TAG¿] and [the other right¿ ]
09 Jen:
[ . h h h h h [ . h h h h h [.hhhhh
J]Ahm:så si’r vi da det, Y]Es-but then say we then that, O]kay then that’s what we’ll do,
10 JO: Ska’ vi ikk’ si’:det¿= Shall we not say that¿= Isn’t that what we’ll do¿= 11 Jen: →1 =Johm’ jeg var ellers klædt om¿ =Yes-but I was otherwise dressed about¿ =Yes but I had actually changed my clothes¿ 12
(0.4)
13 JO: →2 >.hh< Var du klædt om? >.hh< Were you dressed about? >.hh< Had you changed your clothes? 14 Jen: →3 Jah=jah=ja:h=å’:: f [orb]eredt.=.hhh [Det var Yes=yes=yes=and p[rep] ared.=.hhh [Those were 15 JO: 16 Jen:
[( )]
[(
fan]d’me hunderd’tredive hårde sider. blo] ody well a hundred and thirty tough pages.
17 JO: )] 18 JO:
Ja det a’ hårdthh. Det a’ hårdt ahah= Yes it’s tough. It’s tough ahah=
In lines 1–5, JO proposes when and where the political meeting will be held. His turn in lines 7–8, “Så ka’ du nå å’ komme te: eh: te’ bå’e det ene å’ det andet ikke¿”/“Then you can manage to get to both one and the other right” implies that the meeting has been coordinated to accommodate Jens. Jens’ response does not align with this implication as, for instance, an expression of gratification would have done. His acceptance is delayed by a loud and long inbreath, potentially projecting a disprefer red answer
Answers to inapposite inquiries
169
(Heritage 1984b; Pomerantz 1984). The inbreath in combination with the way in which Jens’ acceptance is formulated, “JAhm: så si’r vi da det,”/“OK then that’s what we’ll do,” comes across as resigned concession. This lack of enthusiasm makes JO pursue a more adequate response, with the inquiry “Ska’ vi ikk’ si’ det”/“Isn’t that what we’ll do”. As a negative interrogative, this inquiry is strongly biased toward a positive, confirming answer (Heritage 2002; Heinemann 2006). Jens, however, produces only a minimally accepting “Jo”/“Yes”, then goes on to explain that he had already changed his clothes, i.e. gotten ready for the meeting (line 11).9 This turn provides a possible explanation for Jens’ unenthusiastic acceptance of the rescheduling of the meeting: He believed the political meeting to be taking place on the day of the telephone call and had thus already incorporated the meeting into his schedule. With the rescheduling, he now finds himself in a position where on one day (the day of the telephone call) he is ready to go to a meeting that will not take place, whereas the next day he will be forced to go to two meetings in close succession. In line 13, JO reformulates Jens’ statement as an inquiry, requesting Jens to reconfirm that he had “changed his clothes”. This question entirely fails to take the implications of Jens’ statement into consideration and is, in that sense, off the sequential track. Jens’ response to the inquiry orients to this misplacement. First, he treats the question as inapposite through the production of multiple response particles (“jah=jah=jah”). Second, he continues on a track that focuses on his expectation that the meeting is taking place on the day of the phone call: Not only had he changed his clothes, he had also prepared by reading “130 tough pages” (presumably a report to be discussed at the meeting). He does so, through latching an “and”-prefaced turn (Heritage and Sorjonen 1994) to the last of the response particles, thus signaling that this turn is part of the overall sequence of which the question–answer sequence was a suspension. A second observation made by Heritage (1998) about “oh”prefaced answers to inquiry is, that though these answers indicate “some inappositeness or difficulty about a prior inquiry”, they only do so implicitly and leave “the source of the difficulty as something to be searched for and resolved by the questioner” (Heritage 1998: 295). In Extract 2, for instance, Steve does not overtly state
170
Trine Heinemann
that it follows from the information provided by him (that Nazis shot Maxwell’s father and that his mother died in Auschwitz), that Maxwell is of Jewish origin. The Danish and German extracts above suggest that multiple response particles are similarly inexplicit as answers to inapposite inquiries. In Extract 4, Kirsten does not state that the fact that a person is buried entails that that person is dead. Nor does Jens, in Extract 6, remark on the fact that he has already explicitly stated that he had changed his clothes at the point at which JO inquires about this. Rather, as Golato and Fagyal (2008 ) note, the answerer, through the production of multiple response particles merely indicates the misaligned placement of the turn to which she is responding. The inexplicit nature of “oh”-prefaced responses and their functional equivalents in Danish and German seems to be a direct consequence of the fact that the questions to which these practices are used as answers are oblique. Should answerers deal with the inappositeness of such questions in a more explicit manner, this would only serve to further suspend the progressivity of the activity that the participants were engaged in before the question was asked. Moreover, by explicating exactly why a question should not have been asked, the answerer would be challenging the questioner’s inapposite behavior in an overt manner, making a reaction relevant from the challenged party (see, for instance, Dersley and Wootton 2000; Koshik, forthcoming ; Monzoni 2008 ). Such a reaction would, in turn, lead the participants further away from the suspended activity and perhaps even result in a complete abandonment of this activity. As practices used for answering questions to which the answer is already known, “oh”-prefaced responses and multiple-response particles are thus perfectly designed to handle the answerer’s conflicting interests so that she can index the inappositeness of the question without further diverting attention from her overall interactional project. Explicit targeting of the inappositeness of a question: The use of the modal adverb “da” The previous examples have shown how recipients can index that the question to which they are responding is inapposite without explicating exactly what was problematic about the question. On
Answers to inapposite inquiries
171
occasion, however, recipients do explicate why a question need not – indeed, ought not – have been asked. In the following, I identify one practice for doing so in Danish. Extract 7 is a first instance of this. Here, a multiple “jaja” indexes the prior question as inapposite and is then followed by additional material which explicates what knowledge the recipient possessed that should have prevented her from asking the question. Kirsten and Gunnar have been married for about twenty years. Prior to this extract, they have been talking about Carl Peter, who is married to Gunnar’s sister. Throughout their discussion, Kirsten has treated Gunnar as being the most knowledgeable about all things concerning Carl Peter. She does so again in lines 4–5, where she inquires whether Gunnar has ever met Carl Peter’s father. (7) (MITHVDP31Aug0606 “Carl Peter’s father”) 02 Gun:
KR HH
03
(2.5)
04 Kir:
>M’ har du< no’ensinde mødt:e:h hans far >But have you< some-time met:e:h his father >But did you< ever meet eh his father
05
egentli’ >al’så fordi det var< no’et me’ actually >ADV because it was< something with actually >you know because there was< something
06
(1.1)
07 Gun: → Jah=jah ham kender du da ås.’ Yes=Yes him know you adv also. Yes=Yes you da know him as well. 08
(0.5)
09 Kir:
Carl Peters far? Carl Peters father?
Kirsten’s inquiry is positively framed. With the inclusion of the negative polarity item “nogensinde”/“ever” and the adverb “egentlig”/“actually”, she displays her belief that Gunnar has never met Carl Peter’s father (Quirk et al. 1985). Because of the distribution of knowledge , the inquiry implies that Kirsten herself has never met Carl Peter’s father. If Kirsten had met him, then presumably Gunnar would have as well – and, more importantly, Kirsten would know this.
172
Trine Heinemann
The kind of knowledge that is at stake here is not available in prior turns. Prior to this extract, Gunnar and Kirsten have discussed only Carl Peter, not his father. Any knowledge that Kirsten has about Carl Peter’s father is thus dependent on who she is in relation to Gunnar, rather than on what he has said in prior turns. In addition to this, Kirsten’s inquiry is produced in a sequential environment different from the ones discussed in prior sections. It follows a lapse in the conversation (see lines 1–3) and does not suspend an ongoing activity. Nevertheless, Gunnar treats the inquiry as inapposite by producing multiple instances of the response particle “ja”/“yes”. Furthermore, and in contrast to the extracts discussed previously, he goes on to explicate exactly what it was that Kirsten knew and should have taken into consideration before asking the question. Thus, in line 7, Gunnar states that Kirsten has herself met Carl Peter’s father. The statement concerns a “B-event” (Labov and Fanshel 1977) in that it makes a claim about the recipient, Kirsten’s experiences. In addition, the use of the verb “kender”/“to know” is an upgrade of the verb used by Kirsten, “mødt”/“met”, and implies that Kirsten has had more than one encounter with Carl Peter’s father. Finally, the addition of the term “ås”/“as well” links Kirsten’s experience directly to Gunnar’s. This statement is thus exquisitely designed to convey exactly what was problematic about Kirsten’s question, namely the presupposition that she herself had not met Carl Peter’s father. It is, however, only through the presence of the modal adverb “da”, that Gunnar explicitly rebukes Kirsten for failing to consider already known information in designing her question. Though this adverb can be translated into English as “surely” or “really”, translation fails to capture its interactional import. To the best of my knowledge, “da” has not been described in any systematic fashion. However, the similar adverbs, “doch” in German and “toch” in Dutch , have.10 Though these descriptions are based on invented examples, some initial observations about them can serve as a point of departure for the analysis of “da”. For the German adverb “doch”, Blass (2000) argues that this is a marker of mutual manifestness (Sperber and Wilson 1986), which serves to strengthen what the speaker is saying, in order to weaken or strengthen a previous assumption (stated or implicit). Blass (2000) provides the following case as an example of how “doch” is used and how it should be interpreted.
Answers to inapposite inquiries
173
(8) (Blass 2000, ex. 14, pp. 48) Komm doch rein! Come as-it-is-manifest-to-do in contrary-to-what-you may-be-thinking “Why don’t you come in?”
For this example, Blass (2000: 48) notes, that doch indicates that the speaker and hearer have manifest a cultural norm, which is “to just come in.” Doch also indicates that the speaker is negating any assumption that made the hearer act against the manifest norm. To paraphrase, “It is the norm to enter without being told and any other assumption has to be eliminated.” In English the question form has the same implications: “It is normal to come in, so why don’t you?”
In a similar fashion, Foolen (2006) argues that the Dutch “toch” is used in contexts where the speaker had assumed that the recipient shared the same knowledge and attitude toward a particular matter but where there is now reason to be uncertain about this. He provides the following example, from a novel by the Dutch writer Maarten ‘t Hart. (9) (Foolen 2006:67, ex. 12)
Maar ik kan toch thuis koffie zetten But I can make coffee at home, can’t I?
For both “doch” and “toch”, there are indications, then, that speakers use these adverbs to imply that the recipient has failed to behave in a manner that is congruent with some knowledge, attitude or normative expectation that is shared by both parties. “Doch” and “toch” serve to emphasize the existence of this shared knowledge , attitude or normative expectation by forcefully rejecting any suggestions to the contrary. Gunnar’s use of “da” in the extract above suggests that this Danish modal adverb functions in a similar manner. The presence of “da” emphasizes that the fact that Kirsten has met Carl Peter’s father is shared knowledge and that any assumption to the contrary is incorrect. Gunnar thus rebukes Kirsten for asking a question that presupposes the opposite to what she already knows, and the question is in this way treated as inapposite. In the following, I will discuss in more detail how “da”, when used in turns that are responsive to a question, serves to explicate
174
Trine Heinemann
why that question was inapposite and thus to imply that the questioner has behaved in a morally problematic manner. First, I demonstrate how the presence of “da” in answer turns serves to target and reject a position taken (implicitly) by the questioner in asking a particular question. Second, I demonstrate how speakers use “da” to indicate that the recipient has committed a moral breach by failing to follow normative expectations. Finally, I show that recipients of “da” turns orient in their responses to these aspects of “da” by conceding that they did have access to the knowledge that should have prevented them from asking the question and that they have consequently behaved in a morally problematic manner. “I don’t know” answers One way in which to illuminate how “da” serves to target and reject a position taken (implicitly) by the questioner is to contrast cases where this adverb is present in the answer, with similar cases where it is absent. The following two extracts allow for such a comparison. These are instances of “I don’t know” answers, i.e. answers where a speaker claims to lack the knowledge required to answer a question. These answers are particularly interesting because they themselves reject an implication – or expectation – of the question: that the person to whom the question is directed will be able to answer. As the contrast between the two extracts attests, the presence of “da” in such answers upgrades the rejection and marks the answerer’s lack of knowledge as self-evident and the question, consequently, as highly inappropriate. In the first extract, an “I don’t know” answer without “da” is delivered. Gunnar and Kirsten are having a break during their bicycle vacation on an island. Both of them have been on the island (separately) before, but Kirsten has repeatedly treated Gunnar as the one most knowledgeable about the island and their whereabouts. In line 2, she does so by inquiring about the distance to a small village, Søby. (10) (MITHVDP31Aug0606 “Distance to Søby”) 02 Kir: .mlhh Hvor langt a’ der egentli’ te’ Søby .mlhh How far is it actually to Søby
Answers to inapposite inquiries
175
03 herfra, from here, 04 (0.8) 05 Gun: Det ve’ jeg ikk.’ ↑ti kilometer, That know I not. ↑ten kilometers, I don’t know. ↑ten kilometers,
In response to Kirsten’s inquiry, Gunnar first claims a lack of knowledge with “Det ve’ jeg ikk’”/“I don’t know”. He then proceeds to attempt to answer the question nevertheless, by providing a possible answer, “ti kilometer”/“ten kilometers”. With this “best guess”, he displays his willingness to answer the question – and does not in any way indicate that he finds it problematic that Kirsten has asked him this question. By comparison, in Extract 11, the answerer adds “da” to her “I don’t know” answer, thus treating as incorrect and inappropriate the expectation that she is more knowledgeable on the matter questioned than is the person asking the question. KM has called to talk to Jens. Jens’ wife, Fie, answers the telephone. After a greeting and a self-identification (in line 2), KM inquires about Jens’ availability by asking “Har du Jens hjemme”/“Do you have Jens home”.11 Fie disconfirms this (in line 5), and the sequence is completed through KM’s third position receipt “Nåh”/“Oh” in line 6. Fie then asks whether she was supposed to have Jens home, suggesting that KM expected this to be the case and that she thus knows more about Jens’ whereabouts than his own wife (in line 8). (11) (TH/M2/9 “Missing husband”) 02 KM:
Jah. Go’dag An’Sophie KristaMerete, Yes. Hello An’Sophie KristaMerete,
03 Fie:
Jah hej, Yes hi,
04 KM:
Har du Jens hjemme¿ Have you Jens home¿ Do you have Jens home¿
05 Fie: Nejh. No. 06 KM:
Nåh. Oh.
176
Trine Heinemann
07
(0.3)
08 Fie:
Sku’ jeg ha’ det? Should I have that? Should I?/Was I supposed to?
09
(0.1)
10 KM: → De::tehhh V(h)e’ jeg da i(h)kk’(hh) Tha::thehh K(h)now I adv n(h)ot(hh) I do da n(h)ot(hh) kn(h)ow Nejh, de:t det tror jeg ikk’. 11 Fie: No, tha:t that think I not. No I don’t think so. 12 KM:
>Nåh.< >Oh.<
As in Extract 7, Fie’s question in line 8 is neither sequentially mispositioned, nor is it inquiring into something that is stated or implied through prior turns at talk. Her question is, however, problematic and inappropriate in various other ways. First, it portrays her as a wife who does not necessarily know her husband’s whereabouts. Second, it potentially initiates a complaint about Jens as the kind of husband who is never home or who never lets his wife know where he is. More importantly for our purposes is that Fie, through the question, treats KM as the one who knows the most – or has the most rights to know – about Jens, though she does so jokingly. As Heritage and Raymond (2005) and Raymond and Heritage (2006) have demonstrated, participants can challenge such claimed rights to know through the way in which the first and second parts of an adjacency pair are formatted. In Extract 11, KM orients to the inappropriateness of Fie’s question through the addition of “da” to her “I don’t know” answer. Whereas the “I don’t know” answer, as in Extract 10, in itself rejects the implication that KM would have access to knowledge about Jens’ whereabouts, the addition of “da” emphasizes this rejection thus indicating the self-evident nature of her lack of knowledge and implying that Fie should have known better than to ask KM this question. The inappropriateness of Fie’s question is further marked through KM’s production of “embarrassed laughter” throughout her answer. The answer thus invokes aspects of morality in that KM chides Fie for asking a question that Fie herself had more rights to know the answer to. In the following section, I will discuss
Answers to inapposite inquiries
177
in more detail how the morality aspect is a generic feature of “da” turns and how recipients of such turns orient to this. Implying immoral behavior with “da” The invocation of a moral dimension appears to be a generic feature of “da” turns and is present outside the context of question– answer sequences, for instance in assertions. Here, the fact that the speaker is telling the recipient something that should already be known to her, rather than informing her of something new, is particularly evident, as is the effect that this has on marking the recipient’s actions as morally problematic. In the following extract, for instance, a speaker uses a “da” turn to treat the recipient’s nonverbal actions as problematic because they are at variance with a norm that the recipient should be aware of. The extract comes from a caregiver’s visit to a care recipient and takes place in the bathroom where the caregiver is helping the care recipient to get ready for the day. Prior to this extract, the caregiver has announced that she will get the care recipient’s shirt from the living room. The care recipient has indicated that the shirt might not be clean as she wore it the day before. When the extract starts, the caregiver (CG) has just returned to the bathroom with a shirt and has begun helping the care recipient (CR) into the shirt. (12) (TH/F4/HH/3–1 “Curry shirt”) 01 CR:
Var den ren? Was it clean?
02 CG: Nejh. No. 03
(.)
Der var kar ry på.= 04 CG: There was cur ry on. It had curry on it.= 05 CR: → =Jahm’ så ska’ jeg da ikk’ ha’ den på. =Yes-but then shall I adv not have it on. =Yesbut then I’m da not supposed to have it put on. 06 CG: Ahm’ det’ (h)heller ikk’ den(h), D(h)et’ en Nyeah-but it’s (h)neither not it(h), I(h)t’s a Nyeah but it(h) isn(h)’t that(h) one(h) (h)either,
178
Trine Heinemann
07 ren den [ her ] Maren(h), .hh hehheh .hhhh clean this [here ] Maren(h), .hh hehheh .hhhh It’s a clean[ one ] this one Maren(h), .hh hehheh .hhhh 08 CR:
[Nåh, ] [Oh. ]
In line 1, the care recipient inquires whether the shirt was clean. The caregiver answers with a disconfirming “no”, followed by an explication of what was wrong with the shirt: it has curry sauce on it. Nevertheless, she continues to help the care recipient into the shirt. In reaction to this, the care recipient in line 5 asserts that she is not supposed to wear a dirty shirt. The presence of “da” means that this is more than a simple informing of how the care recipient believes things ought to be. In addition, she targets the caregiver’s behavior as being in violation of a manifest cultural norm of which the care giver should be aware: One should not be wearing dirty clothes. This casting of a prior action as problematic is evident also in the way in which the caregiver responds to the care recipient’s assertion. She has, in fact, unbeknownst to the care recipient, disposed of the dirty shirt and taken a clean one from a cupboard in the hallway. The caregiver explicates this in her turn in lines 6–7. Thus, her actions are not in violation of the normatively expected behavior to which the care recipient orients in line 5. By laughing throughout the production of her turn, the caregiver displays that she does indeed share the same moral position as the care recipient and would not expect her to wear a dirty shirt – this is in fact so far from her mind that she finds the suggestion that she has done so laughable. In this case then we see how a “da” turn is used to treat an immediately prior action as problematic when that action involves opposing what should be taken for granted. The “da” turn serves to rebuke the recipient for having behaved in opposition to a manifest cultural norm and thus invokes a sense of what is appropriate and what is not. Recipient’s responses to “da” turns In the previous sections, two features of “da” have been emphasized. First, the addition of “da” to a turn upgrades the speaker’s statement from informing the recipient of something new to telling the recipient something she already knew. Second, the presence of “da” casts a prior action as morally problematic and serves to rebuke
Answers to inapposite inquiries
179
the recipient for her failure to take shared knowledge or values into account. Here, I will focus on how recipients deal with these aspects of “da” turns. Recipients’ responses show that they interpret “da” turns as morally loaded judgments that center on their failure to take shared norms or knowledge into account before asking a question. Recipients do this by admitting that they were at fault in asking a question to which they already possessed – or had access to – knowledge that should have prevented them from asking the question. Extracts 13 and 14 are examples of this. As was the case in Extracts 7 and 11, in both of these cases, the knowledge that the questioner should have taken into consideration is not knowledge that was provided in prior turns. Likewise, in neither of these extracts does the question suspend an otherwise ongoing activity. Extract 13 comes from a telephone call made by Mathias, to his aunt, Ester, who lives less than one kilometer away. Mathias has just explained to Ester that he and his father wanted to go sailing earlier that day but had to return home because there was no wind but lots of waves. In lines 1–2 Ester concedes to the weather being “weird”, then inquires whether Mathias thinks it will start to rain (in line 4). (13) (TH/S2/73 “Wet window”) 01 Est: Nej det” rigtigt underligt vejr for det’ ås’ No it’s really weird weather because it’s also 02
) li’ plu’sligt blevet ( all of a sudden gotten (
03 Mat:
Jerh,= Yeah,=
04 Est:
=Tror’d’ det bli’r regnvejr? =Think-you it become rain-weather? =Do you think it’ll rain?
)
05 Mat: → Jahm’ det regner da, Yes-but it rains adv, Yes but it is raining da, 06 Est: Det gør det sgu’ da ås,’ mit vindue a’ That does it adv adv also, my window is It bloody well da is too, my window is 07
jo vådt adv wet jo wet
180
Trine Heinemann
From the perspective of the ongoing interaction, nothing about Ester’s question is inapposite. Whereas Mathias has stated that there is no wind but lots of waves, he has said nothing about rain. Neither does Ester’s question suspend an already ongoing activity. The question is, however, contingent on the presupposition that it is not already raining. If it was raining, the answer to Ester’s question would already be known and the question would be irrelevant and hence inapposite. In his confirming answer “Jahm’ det regner da”/“Yes but it is raining da”, Mathias treats the question as inapposite. First, he rejects the presupposition that Ester’s question is contingent upon. With the addition of “da”, he emphasizes that Ester had access to the knowledge that it is already raining, thus rebuking her for asking an inapposite, irrelevant question to which she supposedly already knew the answer. In this case, however, it is not just the answerer, Mathias, who treats the question as inapposite. Ester accepts the rebuke by displaying in her response that she did have access to the fact that it is raining and hence need not have asked the question. In her response, she uses three emphatic markers, “sgu”/“bloody well”, “da” and “ås”/“as well”, thus marking the evident nature of the fact that it is raining. She then goes on to explicate why this was self-evident and what type of access she had to this information: Her window is wet. In this continuation, she uses another marker of self-evidence and shared knowledge, the modal adverb “jo”/“you know”. Furthermore, the whole of Ester’s turn is delivered with a smile voice, as if she is laughing at herself for having failed to use this accessible information in framing her question. Extract 14 is another example of how the recipient of a “da” answer can show his understanding that he is being rebuked for failing to consider already known information. The extract comes from the beginning of a phone call, made by LC to Jens. After the identification and greeting sequences (lines 1–4), LC produces the statement “Nå nu er du hjemme igen”/“So now you’re home again”. As a statement about a “B-event” (Labov and Fanshel 1977), this minimally invites the recipient, Jens, to confirm or disconfirm and thus to treat it as a question. Jens orients to it as such and confirms the truthfulness of LC’s statement with the positive response token “Jerh”/“Yeah” in line 06. At this point, however, he appears to fail to grasp the implication of LC’s question: that he has been away and has returned home (conveyed by “now” and “again”). Realizing this somewhat later (see line 7), Jens adds further material to his
Answers to inapposite inquiries
181
confirmation , stating, in lines 8–9, that he has been home all the time, thus rejecting the implication of LC’s question. (14) (TH/S2/64 “Home again”) 01 Jen:
Ja det’ Jens, Yes this is Jens,
02 LC:
Det a’ Lars Christian. Go’dag Jen[s, ] This is Lars Christian. Hello Jen[s, ]
03 Jen: 04 Jen:
[Go]’dag [He]llo
Lars Ch[ristian,]
05 LC:
[Nå n]u a’ du hjemme [ igen,] [So n]ow you’re home [again,]
06 Jen: → 07 LC:
=Jahm’ det var dejligt å’ hør[e. ] =Yes but that’s great to he[ar.]
08 Jens: → 09
[ .h h ]h h- Jerh.= [ .h h ]h h- Yeah.=
[ De]t har jeg da [Tha]t have I adv [Tha]t I’ve da
været hele tiden,= been all time-the,= been all the time,=
10 LC: →Jahm’ det ve’ jeg godt men jeg har- .hh Jeg →Yes-but that know I well but I have- .hh I →Yes but I do know but I have- .hh I probably 11 har måske ikk’ la’ den ringe længe nok sidste gang have maybe not let it ring long enough last time didn’t let it ring long enough the last time 12 for jeg var kom så i tvivl om du var på arbej’e.← because I was came then in doubt about you were at work.← because I was unsure whether you were at work.←
As in the other extracts discussed so far, Jens adds “da” to his turn to emphasize the self-evident nature of his answer and to rebuke LC for failing to take known information into consideration in framing his question. In this case, the known information is that Jens has been home “all the time”. As in previous examples where “da” is part of the answer, there is nothing in prior turns that could have caused LC to know this. Rather, the fact that LC states that Jens is now back home suggests that he has indications that Jens had not been home
182
Trine Heinemann
all the time, for instance, because he has previously tried to reach Jens without success. By adding “da” to his turn in lines 8–9, however, Jens implies that LC ought to have known that Jens has indeed been home all day and that LC, through his question, has failed to consider this knowledge . As in Extract 13, the recipient, LC, admits to possessing this knowledge , this time by explicitly stating that he did know (line 10). LC thus not only admits to having possessed the knowledge that renders his question irrelevant and inapposite but also accepts being rebuked for having asked the question. As further support for his “in the know” status, he then goes on to account for how this newly claimed knowledge can be compatible with the contrary implication of his question: He has called previously but failed to let the phone ring long enough for Jens to pick it up. Summary In this section, I have identified a practice with which recipients can convey that a question is inapposite because it questions something already known. Recipients do so, by rejecting a presupposition , expectation or implication of the question, thus challenging its relevance. By adding the modal adverb “da”, to such rejections, recipients indicate that the questioner possessed knowledge that should have prevented her from asking the question. Answers in which “da” is used thus serve to rebuke the questioner for failing to consider this knowledge in asking the question. When answerers use this practice to deal with the inappositeness of a question, they do so in a specific sequential position, i.e. when the question is not mispositioned and does not initiate an oblique sequence. Furthermore, the kind of knowledge that the questioner is rebuked for not considering is knowledge that cannot be inferred from prior turns-at-talk but instead stems from who the participants are in relation to each other or from normative expectations about the world. The circumstances of the questions that receive “da” answers thus provide the answerer with both motive and opportunity for explicating why the question ought not to have been asked. Failing to take shared knowledge assessed through a relationship into account might evidence a breach of greater consequence than one resulting from being “disattentive” or “slow on the uptake”. Consequently, the inappositeness of this type of question is particularly strong, pushing the answerer to deal with it explicitly. In
Answers to inapposite inquiries
183
turn, because the inquiry is not mispositioned and does not interrupt and suspend an ongoing activity, the answerer can deal with the inappositeness explicitly without furthering a digression from the main line of activity. Thus, the practice of rejecting a question’s presuppositions, expectations or implications by adding the modal adverb “da” to the answer, is perfectly suited to deal with the relevancies that are invoked through the sequential positioning of the question and the action that the question is implementing. Discussion This paper has identified two practices in Danish with which recipients can treat a question as inapposite because the questioner failed to consider pre-existing knowledge in framing the question. The existence of two different practices in one language does not necessarily exclude that other languages can accomplish both jobs through a single practice. However, the two practices described here, though serving the same general function, are deployed in distinct sequential positions, to deal with distinct types of knowledge neglect. This suggests that the two practices exist in Danish because participants in interaction more generally need to be able to deal with inapposite questions in different ways depending on the type and strength of inappositeness that has been committed. The fact that both Dutch and German appear to have similar adverbs dedicated to the same job as the modal adverb “da” in Danish supports this suggestion though further study of how these adverbs are deployed in question– answer sequences in these languages is required before we can determine whether this is actually the case. Furthermore, Danish, Dutch, and German are so closely related that it could be argued that the existence of a specific practice in these three languages does not entail that all (or any other) languages would have to be able to do the same. However, findings from a completely unrelated, non-IndoEuropean language suggest that a similar practice is found there as well. Thus, Wu (2004) argues that in Mandarin , the particle “a”, when produced with flat or slightly high pitch, serves to mark that a participant has failed to take already existing knowledge into consideration, when, for instance, asking a question. Whether this particle is actually used to accomplish the same practice as the Danish modal adverb “da” (and the Dutch and German “toch” and “doch”)
184
Trine Heinemann
remains unclear at present though the following extract suggests that these are indeed very similar practices. Here, L has just stated that he has played baseball in the national Taiwan league. In line 1, R, who is L’s long-term girlfriend, inquires when this was. When the answer to this inquiry is provided, R questions the fact that L played in the league. Both of R’s questions are answered with turns suffixed with final “a”, indexing her displayed state of knowledge as less than expected. Furthermore, the third participant, T, overtly orients to this in line 5, with the question “li m chai ou”/“You didn’t know?”. (15) (CS Party 216A) (Wu 2004, ex. 30, pp. 184)12 01 R: shenme shihou When? 02 L: xiao shihou a. When (I) was a kid a. 03 R: ni da guo ou. You played (ball in the league)?/! 04 L: dui a. dui a. (That’s) right a. (That’s) right a. 05 T: %li m chai ou% %You didn’t know?%
In the first section of this paper, I took, as a point of departure, a practice well described in English (“oh”-prefaced responses to inquiry) and identified a similar practice in Danish (and German ). In principle, the same process, but in reverse order, should be possible; that is, we should be able to identify the English practice with which recipients convey what knowledge a questioner should have considered before framing a question, as is done in Danish through the use of the modal adverb “da”. Though this is not the place to do so, there are indications as to where to look when trying to establish what this practice may look like in English. As noted when the modal adverb “da” was introduced, this adverb is a marker of emphasis and contrast that can be translated as “really”. Though this translation does not capture the actual function of “da”, there are other ways of marking emphasis and contrast, in English most notably stress. Indeed, Schubiger (1972) argues that many of the functions accomplished in German through the use of modal adverbs are accomplished through the application of stress
Answers to inapposite inquiries
185
in English. The tie between German and Danish modal adverbs on the one hand, and the tie between German modal adverbs and English stress on the other, thus suggests that stress patterns might be a good place to look for how English speakers accomplish the actions that in Danish are accomplished through the production of the modal adverb “da”. Appendix: Transcription symbols The Danish data is transcribed according to the convention developed by Gail Jefferson . In addition, the following abbreviations are used to mark certain grammatical categories: tag:
Lexical tag used for making a statement into a question. In Danish this is typically done with adverbs that are fitted to the polarity of the statement, so that positively framed statements are followed by the negative adverb “ikke”/“not”, whereas negatively framed statements are followed by “vel”/“probably”. adv: Adverb. Danish has a range of adverbs that all seem to deal with epistemic or evidential aspects of the statement in which they are used. When possible, I have translated the adverbs into similar, English ones. Acknowledgments I am immensely indebted to Anna Lindström, Tanya Stivers, Birte Asmuss, and, in particular, Jakob Steensig, for their helpful comments at the very early stages of this paper. I am also grateful to Andrea Golato, Jack Sidnell, and Sofie Emmertsen for comments on later versions. All remaining faults are my responsibility alone. Notes 1 In various types of institutional interaction, questions to which the answer is already known serve different purposes than they do in ordinary conversation. In educational settings, for instance, teachers use “known answer questions” to check whether their students have the required knowledge (Koshik 2002, 2005; Macbeth 2004), and when lawyers cross-examine witnesses (Drew 1992), radio journalists interview subjects (Heritage 1998) and police officers interrogate suspects
186
Trine Heinemann
(Stokoe and Edwards, 2008), they do so to get known information into the public domain. 2 In order to provide information about what happens in this sequence after Steve has answered the inquiry in line 10, I have used the example from Drew and Holt 1998, rather than Heritage 1998. 3 I am grateful to Jakob Steensig for giving me access to the data from which this excerpt is taken. 4 “Never” in this contexts means that in a data corpus of approximately sixty hours of conversation, I have found no such instances. 5 The production of multiple linguistic items within the same prosodic unit has also been described by Stivers (2004). Her analysis of what this phenomenon accomplishes in English is different to – though not entirely incompatible with – the analysis by Heinemann (2003) and Golato and Fagyal (2008) for Danish and German respectively. The difference can be due to language differences, but it may also be a reflection of the fact that the “multiple sayings” described by Stivers (2004) include other linguistic objects than response particles and occur in different sequential contexts, i.e. not as direct answers to questions. See also Golato and Fagyal (2008). 6 Leslie’s prefacing of the question with “oh” is partially oriented to this, as it displays her as having “just now” come to this conclusion herself (Heritage 1984a). 7 Golato and Fagyal (2008 ) describe this contour as one where the pitch peak lies on the second syllable. They argue that the function of this prosodic version is distinctive from that of multiple ja’s where the pitch peak is placed on the first syllable. Whether this is the case for Danish has not been investigated for this paper. 8 Simplified version. 9 In Denmark, political meetings usually involve a more formal dress than what most Danes wear during the workday. 10 The etymological equivalent of “doch” and “toch” in Danish is “dog”, whereas “da” stems from the same source as the German “denn”/“then” (Fretheim 2000). The functions of these adverbs seem to be somewhat intertwined, so that “dog” is used for some of the functions that are accomplished through “doch” in German, whereas “da” covers other aspects of the same adverb. “Dog” is used only infrequently in Danish and will not be discussed here. 11 In Danish telephone calls, both “Do you have x home” and “Is x home” can be used for this type of inquiry. To my native ear, both sound equal, but to leave open the possibility that one is more idiomatic than the other, I have chosen a translation that sound slightly awkward in English. 12 Simplified version.
7 Gaze, questioning, and culture Federico Rossano, Penelope Brown, and Stephen C. Levinson
“There is nothing so brutally shocking, nor so little forgiven, as seeming inattention to the person who is speaking to you [. . .] I have seen many people who, when you are speaking to them, instead of looking at, and attending you, fix their eyes upon the ceiling or some other part of the room [. . .] Nothing discovers a little, futile, frivolous mind more than this, and nothing is so offensively ill-bred.” Lord Chesterfield 1752 (Letters to his son, Letter CCLXXVIII)1
Introduction This chapter is about gaze behavior in conversation. Using data from three unrelated cultures (or more properly, speech communities), we ask whether there are, on the basis of these three cultures, plausible candidates for universal patterns of gazing during interaction, and where, on the contrary, we may expect to find culturally specific patterns. This focus is already rather different than the tenor of most of the other papers in this volume. Moreover, because of this focus, a central part is played by basic descriptive statistics, rather than the qualitative analysis of specific episodes. Despite these foci and the method s adopted, we see this paper as addressing some central issues posed by conversation analysis (CA). We are interested in a pervasive conversational practice, namely when participants gaze at one another during conversation. There has been relatively little published work in CA on the systematics of gaze, and what there has been (e.g., by Goodwin 1981) will be shown by our data not to provide a possible general analysis of the phenomena. Extensive work by one of us (Rossano 2009) shows that gaze (in Italian conversation at least) can indeed be understood in CA terms, that is, in terms of a sequential analysis
188
Federico Rossano et al.
of actions, but not directly in terms of the turn-taking apparatus or displays of engagement and attentiveness as had been supposed (e.g., Goodwin 1981). 2 Instead, what this work shows is that gaze is used, for example, to coordinate the development and closure of sequences and courses of action, to pressure for responses and pursue them, to indicate special states of recipiency . This means that gaze to speaker can be used to signal that the recipient recognizes that the speaker is launching on an extended turn at talk (as in a story), or gaze away from the other participant can be used to signal the closing of larger units, for example, that the recipient recognizes that a multi-turn sequence is completed. On the foundation of this prior qualitative work, there is a good set of hypotheses that can be addressed by a statistical approach, where the various units and actions over which gaze is coded are fully informed by CA findings. This approach is not alien to recent CA research (cf. Heritage and Maynard 2006), 3 where several works, usually in institutional settings, have looked at an interactional phenomenon and used quantitative method s to assess the association between that phenomenon and some exogenous variable whether that be time (e.g., in Clayman et al. 2007), prescribing outcomes (e.g., in Mangione-Smith et al. 2003) or race and class (e.g., in Stivers and Majid 2007). One reason for adopting a statistical approach is our general goal of trying to understand whether gaze as an interactional practice has universal properties across cultures or not. We cannot make easy judgments about sameness and difference in conversational practices unless we can be sure that the extracts we analyze are reasonably representative of interactions in the culture in question – we need plenty of instances, across different speakers and recipients, where the actions and context are in some sense comparable. To achieve this, we have here adopted the following expedient. First we have established the relevance of looking at gaze in question–answer sequences in three different cultures because of the differences they could reveal with claims made in prior literature. We have, then, piggy-backed off a larger project concerned with the cross-linguistic comparison of question–answer pairs (Stivers et al. 2009), and focused in on gaze behavior in a sample of 300 question–answer pairs from each of the three cultures, as used by roughly ten distinct dyads (or c. twenty individuals) from
Gaze, questioning, and culture
189
each culture who are engaged in naturally occurring conversation in informal settings. The three languages and cultures are completely unrelated: (1) Italians of northern Italy; (2) speakers of Yélî Dnye , a language isolate spoken on Rossel Island, a remote island off Papua New Guinea; and (3) speakers of Tenejapan Tzeltal , a Mayan language spoken in an indigenous community in the highlands of southern Mexico. As we will show, this restricted sample is quite enough to show that earlier analyses of the occurrence of gaze in interaction are not general across functions and cultures. The results are also suggestive of more positive general hypotheses, which would need to be followed up by qualitative analysis of a wider range of actions and sequence types. Given the work earlier referred to (Rossano 2009), it is clear that we cannot expect a full understanding of gaze in interaction without such a wider analysis – question–answer pairs, for example, constitute a very different sequential environment if compared with storytellings. But we hope that these preliminary results will already serve as a useful orientation for this future work. Earlier work on gaze in interaction Human gaze behavior is a highly evolved system: Unique amongst primate species, the human orbit has evolved to display the sclera or white of the eye, the function of which can only have been to make gaze direction discernable to others at a distance ( Kobayashi 1997, 2001; Morris 1985). Moreover, it has long been established that humans can judge the direction of other humans’ gaze to within a few degrees of arc (Gibson and Pick 1963), and this capacity has been claimed to be crucial for the development of joint attention and human social cognition ( Tomasello 1999). The social functions of gaze have been much commented upon, not least by Darwin (1872), who was especially interested in the role of gaze aversion in the display of shame and shyness, and of gaze engagement in the display of mastery, lust, and aggression, and who noted that infants do not display any sensitivity to these roles.4 Simmel also wrote about the “uniquely sociological function” of the eye and the “union and interaction of individuals [being] based upon mutual glances” (1969: 358). In particular he claimed that “the
190
Federico Rossano et al.
totality of social relations of human beings, their self-assertions and self-abnegation, their intimacies and estrangements, would be changed in unpredictable ways if there occurred no glance of eye to eye” (1969: 358). Although there is no doubt that the social functions of gaze are important, one should not forget that gaze is also straightforwardly an indication of visual attention. Foveal vision directly indexes the centre of attention, despite the fact that movement in particular is easily picked up in peripheral vision. If one puts together these observations, there cannot be the slightest doubt that gaze engagement in interaction is a potentially potent tool, carefully deployed. Most of the detailed work on gaze in interaction has been conducted by social psychologists or kinesic researchers working on interaction in English or other European languages. Much of this has inevitably been in laboratory settings. General conclusions include the fact that participants spend a considerable proportion of interaction time looking at each others’ face, and that these moments have some “central importance” in interaction (Argyle and Graham 1976: 6; see also Argyle and Dean 1965; Exline 1963; Goodwin 1981; Gullberg 2003; Gullberg and Holmqvist 2006; Kendon 1967; Nielsen 1962). Gaze in other directions seems likely to be attracted by objects relevant to the tasks at hand (cf. Argyle and Graham 1976), but gaze beyond that might be held accountable (e.g., as lack of interest, cf. Goodwin 1981, 1984). A few researchers have explored the notion that gaze is more than just an ancillary index of interest or attention, namely that it plays a crucial role in organizing and regulating interaction. Besides claims about the importance of gaze for monitoring each other’s behavior and facial expressions (e.g., M. H. Goodwin 1980; Kendon 1967), Kendon (1967) and Duncan and colleagues (Duncan 1975; Duncan and Fiske 1977; Duncan and Niederehe 1974) have argued that speaker gaze has a “floor apportionment” function in conversation and can function as a turn yielding cue, 5 but B eattie (1978, 1979) has countered that the speaker’s gaze away during early utterance production and reengagement during final production are occasioned purely by the need to reduce cognitive load and do not have any regulating function in terms of turn-taking.6 Nevertheless, the idea that gaze is closely related to participant role (speaking, or being addressed, in particular) has
Gaze, questioning, and culture
191
been persistent. Specifically, it has long been claimed that in dyadic interaction people tend to look at the other participant more when they are listening than when they are speaking (Argyle and Cook 1976; Argyle and Dean 1965; Argyle and Graham 1976; Bavelas et al. 2002; Duncan and Fiske 1977; Exline 1963; Goodwin 1981; Kendon 1967, 1973, 1990; Kleinke 1986; Nielsen 1962; Rutter 1984).7 Kendon (1967, 1990; see also Sebeok 1981) provided a more precise description of the different patterns of speaker and hearer gaze: Hearers give speakers long looks interrupted by brief glances away, while speakers alternate looks toward and looks away from the recipient of approximately equal length. Goodwin , relying on a case-by-case analysis of the data, further supported by quantification in a small corpus, proposed two rules (1980: 275, 287; 1981: 57) that should account for participants’ gaze behavior: 1. “A speaker should obtain the gaze of his recipient during the course of a turn at talk.” 2. “A recipient should be gazing at the speaker when the speaker is gazing at the hearer.” If the recipient looks most of the time, then the speaker will find him or her gazing back any time the speaker looks toward the recipient. If the recipient is not looking at the speaker, the latter has resources (phrasal breaks, pauses, restarting the turn) to solicit and obtain the recipient’s gaze. By proposing these as rules, Goodwin claims that participants’ gaze behavior is interrelated rather than independent and suggests a system of norms to which participants are oriented during any turn-at-talk.8 This focus on turns at talk (“utterances” in Kendon ’s terms) as the level at which gaze behavior is organized is shared by those researchers (e.g., C. Goodwin 1980, 1981; Kendon 1967) who were particularly interested in its regulatory functions, which is to say how gaze affects and coordinates the exchange of turns between participants and how it displays engagement and disengagement in the conversation. In contrast, based on sequential analysis of Italian conversations, Rossano (2005, 2006a, b, 2009) suggests that gaze in interaction is not organized primarily by reference to turns at talk. He argues that gaze behavior is mainly organized in relation to sequences of
192
Federico Rossano et al.
talk and the development of courses of action, so that most of the variation in gaze direction should be observed at the beginning or at possible completion of courses of action accomplished through one or more sequences of talk.9 In addition, speaker gaze to recipient seems to serve as a way of putting pressure on the recipient to provide feedback (Bavelas et al., 2002; Kendon 1967) or to produce a response and thus to pursue a response when one is missing ( Rossano 2006b, 2009; Stivers and Rossano, in press). All these claims have been based largely on research done on Indo-European languages or at least in Western societies. They are often challenged by ethnographic reports of anthropologists who have studied the languages and cultures of non-Western societies. Walsh (1991) and Evans and Wilkins (2000), for example, have suggested that the conversational style of Australian Aboriginal communities is fundamentally different from that of white middle-class Australians. While white Australian interaction is usually dyadic and face-to-face, Aboriginal interaction is held to be non-dyadic, talk is broadcast, and eye contact is not important. According to these authors, a preferred seating pattern among friends is side by side (or even back to back) and people will only be “face to face” if there is a significant distance between them or they are separated by something like a fire, and even then the gaze will typically not be directed toward an interlocutor for any significant length of time. (Evans and Wilkins 2000: 582)
Hansen and Hansen (1992) make a related argument suggesting that the gaze patterns typical of European conversation are perceived as offensive in many parts of Aboriginal Australia . Two of the present authors have also claimed that they discern strong cultural differences in the use of gaze in connection with “response systems,” that is, in ways in which active recipiency is demonstrated (Brown and L evinson 2005; Levinson and Brown 2004). Tzeltal recipients appear to avoid gaze and tend to produce verbal responses, repeating part of the prior speaker’s utterance, while Yélî Dnye recipients assume speaker gaze and thus are enabled to deploy a range of facial signals instead of, or in conjunction with, verbal responses (e.g., an eyebrow flash signals “yes,” a nosewrinkling signals “wow!”). One of our goals in this paper is to test whether this kind of cross-cultural difference can be demonstrated on sound numerical
Gaze, questioning, and culture
193
grounds. The following section provides some qualitative ground that accounts for the development of our coding scheme for the comparison and displays initial evidence that prior claims about gaze behavior might not be appropriate for the cultures here examined. A different landscape: Some qualitative counter-evidence to prior claims In this section we provide a few examples that motivate the development of the coding scheme and show a different organization of gaze in interaction. Extract 1 is taken from a dyadic conversation in Italian , and it shows that: 1. In a situation where participants tend to look at objects (pictures) relevant for the task at hand, the speaker nonetheless tends to look at the addressee while asking a question. 2. The speaker looks at the addressee during the questions but the addressee does not look back. 3. No remedial action is taken (cut-off, pauses, restarts, sound stretches) to get the recipient’s gaze. 4. There are questions during which none of the participants looks at the other. In Extract 1, two male friends in their twenties are sitting side by side and are looking at some pictures on the table. The pictures were taken by A during his holidays, and B starts the sequence at line 2 by asking how much it had cost to enter St. Peter’s church in Rome (the picture they are looking at was taken inside this church). We focus on the questions at lines 2, 15, 19, and 21. (See Appendix B for gaze-marking conventions.) (1) Italian 2PCOMP 9:33 A
B
01 (1.0) A
B
((both participants looking down at pictures)) A
B
A
B
02 →B: Soccia quanto hai pagato per entrar qua. Wow how much have paid for enter here Wow how much did you pay to enter here.
194
Federico Rossano et al. A
B
03 (0.2) A
04 B:
B
A
B
Die [ci ten Te[n A
B
05 A: [Poco. °Un euro. D ue euro neanche Little. One euro two euro neither [Little.° One euro. Not even two euros 06 07 B: 08
A
B
(.) A
B
A
Ah Oh Oh A
B
beh well well B
(1.9) A
B
09 A: Per visitar la cupola eh poi. For visit the dome eh then And it was for visiting the dome eh. 10
A
B
(0.2) A
B
> 11 B: Si’. °Vabbe’° Yes Alright Yes. °Alright° 12
A
(0.8)
B
Gaze, questioning, and culture A
195
B
13 A: ↑Anzi niente. Actually nothing ↑Actually nothing. 14
A
B
(0.2) A
B
15 →B: Ni[ente. Nothing No[thing. A
B
16 A: [Perche’ sono andato su’ a piedi. Because be.1s gone up by feet [Because I went up on foot. 17
A
B
(0.2) A
B
A
B
18 A: Se pr[endi] l’ ascensore paghi. If take.2s the elevator pay.2s If you t[ake] the elevator you pay. 19 →B: 20 21 →B:
A
B
[eh?] eh [eh?] A
B
(.) A
B
Si’. Yes Yes.
196
22 A:
Federico Rossano et al. A
B
Si’ Yes Yes A
B
A
B
A
23 B: Allora ci vado anch’ io. Then cl. go.1s also I Then I will go too.
B
At line 2, B asks a question about the cost of entering the building represented in the picture. The speaker (B) turns toward the recipient (A) at the beginning of the question while the latter keeps looking at the picture. After a delay in responding, B starts a candidate answer quantifying the possible cost (ten), but at this point A looks toward B and provides an answer. At line 13, A modifies the answer provided at line 5 by claiming that to enter that place he did not pay anything. Even though A starts the turn with a pitch reset and emphasizes the word “nothing,” A does not look at B. At line 15, B produces an other-initiation of repair by asking for confirmation (and displaying disbelief at the claim that A did not pay anything). B starts turning toward A during the silence at line 14 and therefore starts the initiation of repair already looking at the addressee. B holds his gaze toward A until the end of line 19, where he is not initiating repair but rather pushing for a “yes”/“no” answer to his question at line 15 (“eh” can be used as a tag in Italian questions). At line 21, B asks for confirmation of the last claim about the free entrance to the building, but this time he is looking at the pictures and looks toward B only at line 23, when he produces a closing remark that accounts for the occurrence of the question at line 2 (his possible interest in visiting the place represented in the picture). With Extract 1, we have illustrated, among other things, that even in situations where objects relevant for activities at hand are present, speakers tend to look at recipients while asking questions. With the following extract we want to show what happens when two participants are just talking about future plans and are not dealing with relevant objects in the surrounding environment. Extract 2 shows:
Gaze, questioning, and culture
197
1. Speaker and recipient tend to look up mid-question during the question that initiates a new course of action (see also line 2 in Extract 1). 2. During the course of action, speakers tend to look from beginning to end of each question. 3. Recipients sometimes look away before completion of the question (presumably to plan the answer to the question), though not during every question. In Extract 2, the participants are two female friends sitting at a table (at a 90-degree angle), and the conversation has just started. We focus on the gaze behavior during lines 1, 3, 6–7, 12, 16, and 20. (2) Italian 2GGOSS-stasera 00:15 A
A
B
B
A
B
01 →B: .hh A(h)l l o r a stas(h)era cos’ e’ che fate S(h)o tonight what is that do.2s .hh S(h)o t(h)onight what are you doing A
02 A: Eh Eh Eh
B
A
B
andiamo a Villa Chiara = go.1p to Villa Chiara we go to Villa Chiara =
A
B
A
B
03 →B: = Ma a che ora vi incontrate But at which hour you meet.2p = But at what time do you meet 04 A: 05
A
B
A
B
Vado alle nove e mezza dalla Gloria. Go.1s at nine and half to Gloria I go to Gloria’s (house) at nine thirty. ((long side sequence about Gloria’s recent guests)) A
B
A
B
A
B
06 →B: Nove e mezza ma andate subito a Villa Chiara Nine and half but go.2p immediately to Villa Chiara Nine thirty but do you go immediately to Villa Chiara
198
Federico Rossano et al. A
B
07 →B: alle nove e mezza? = At nine and half at nine thirty? = A
B
A
B
08 A: = Con- le dieci con gli altri. With the ten with the others = With- at ten with the others. 09
A
B
(0.4) A
B
10 A: Ci incontriamo. Cl. meet.1p We meet. 11
A
B
(0.6) A
A
B
B
12 →A: Ci [v i e n i? Cl. come.1p Do you come there? ((to Villa Chiara)) A
B
A
B
A
B
A
B
13 B: [Io esco alle nove_ (0.5) ah io pero’ se arrivo I get.1s out at nine oh I but if arrive.1s [I get out at nine_ (0.5) oh but if I come 14 15
A
B
arrivo a mezzanotte eh arrive.1s at midnight eh I arrive at midnight eh A
B
(0.6)
((A displays shocked facial expression))
Gaze, questioning, and culture A
199
B
16 →A: Alle nove esci? At nine get.1s out At nine you get out? A
B
A
B
17 B: E[h Eh Y[es ((confirming something already said)) 18 A: 19
[Cazzo Dick [Shit ((and then A nods, looking down)) A
B
(0.6)
A
B
(1.2) (1.8)
A
B
A
B
20 →A: Domani sera che fai? Tomorrow evening what do.2s Tomorrow evening what do you do? 21
A
B
(0.5)
At line 1, the recipient starts looking at the speaker before the latter turns toward her; by the time the question at line 3 starts, both participants are looking at each other, and only the recipient withdraws from mutual gaze before completion of the question. The same happens during the question at lines 6–7. All three questions can be taken to be requests for information by B about A’s plans for the night, probably meant to get an invitation by A. At line 12, A does indeed invite B to go out with her and her friends, and, again, we see that the speaker is looking at the recipient from the beginning to the end of the question, while the recipient (B) looks toward the speaker mid-turn when she has already started producing the turn at lines 13–14. The question at line 16 is not only a repair initiation but also a way of displaying surprise and disbelief (as can be seen from the
200
Federico Rossano et al.
co-occurring facial expression) and in this case both participants are looking at each other from beginning to end of the turn. The course of action reaches its completion at line 18 and both participants look down toward the table. At line 20, however, A starts a new sequence and asks B about her plans for the following day. Here we see that the speaker starts looking at the addressee only after the beginning of the question and the recipient does not look at the questioner at all. Also in this case (see questions in Extract 1), no remedial action is taken by the speaker to obtain the recipient’s gaze back and the recipient is not looking at any other relevant objects on the table. These two extracts, which represent patterns observed throughout ten hours of dyadic interactions in Italian , suggest that many of the claims about the organization of gaze in interaction previously proposed do not apply to Italian data. In particular, if we focus on questions, it appears that there is much less gaze mobility throughout the question than compared to previous general claims about gaze in any utterance,10 and there is no particular sensitivity to the lack of recipient gaze (differently from C. Goodwin 1980, 1981). Moreover, speaker gaze appears to be much more predictable and relevant during questions than recipient gaze . If we then look at two typical patterns of gaze during questions in Tzeltal and Yélî Dnye , we can see that while in the former case participants can produce entire sequences without ever looking at each other (see Extract 3), in the latter case participants can be seen to sustain gaze toward each other during questions without any withdrawal (see Extract 4). Extract 3 is taken from an interaction between two middleaged Tzeltal sisters. Both participants are sitting, side by side on a bench, with their backs against a wall, approximately two meters away from each other. They are looking straight ahead, away from each other’s faces. B has been telling A about a relative of theirs, who had come supposedly to the village for a meeting but had gone straight back to town without waiting for the meeting. Lines 2 and 4 are our focus. (3) Tzeltal 2005_v5A_Q19 11:53 01
A
(1.3)
B
Gaze, questioning, and culture A
201
B
02 →A: ma’yuk li’ y-ak’ junta tz’i bi not.at.all here 3E-give meeting PT TAG He didn’t hold a meeting at all here then eh? A
B
03 B: ma’yuk laj not.at.all QUOT Not at all, reportedly 04 →A:
A
B
ju’uk no No? A
B
05 B: ma’yuk laj jalaj ix not.at.all QUOT last.long ACS He did not stay long (enough) at all, reportedly.
During both questions, both participants keep looking away from each other and do not turn to each other even during the production of the answers. This contrasts quite clearly with what can be observed in interactions in our dataset for Yélî Dnye, where, in dyadic interactions, participants seem to look at each other most of the time. Extract 4 is taken from an interaction between two Rossel men who are sitting face-to-face, less than one meter apart. A is complaining about the fact that some individuals who owe him money have not repaid him yet. The gaze behavior at lines 4, 6, and 7 is our focus. (4) Yélî r03_v19_s2 00:13:53 01 A: awêde nga today here 02 A:
anî tóó, 1sFUT.PUNCT sitting
u pyinê 3POSS quest
d:a ngmêê, ngmepe 1sIMMPAST.PUNCT exchange.PROX paying.back I am here today, to search for it ((the shell money)), to pay it back law nkwodo até nî kmungo. law on.top just 1sIMMPAST.PUNCT put.inside I took it up to the law (told the Peace Officer)
202 03
Federico Rossano et al. (0.4) A
B
A
B
04 →B: n:uu ye ngmepe Who that.near.Addr paying.back Who is paying it back (to you) 05 06 →A:
A
B
(0.4) A
B
:êê uh uh?
07 →B: n:uu ye ngmepe Who that.near.Addr paying.back Who is paying it back (to you) 08
A
B
(0.4) A
B
09 A: kî pini dy:eemi knî that.distal man.spec man.and.brother-in-law AUGM That man and his brothers-in-law
In this extract, we can see that participants sustain mutual gaze throughout the three questions: a request for information, its repetition, and a repair initiation. Altogether, the participants sustain mutual gaze across silences and into the answer at line 9, showing the opposite pattern of what was shown in Extract 3. It is noticeable that not only in Italian but also in Yélî Dnye a speaker can sustain gaze toward the addressee from the beginning to the end of the question, without looking away before uttering it or looking toward the addressee only while approaching the completion of the turn. These extracts raise multiple questions. How representative are they of the common gaze patterns during question–answer sequences in each culture? Are the clear differences from prior
Gaze, questioning, and culture
203
claims in the literature indicative of a different type of organization of gaze in interaction, at least in these three cultures? Are there cultures where participants tend not to look at each other at all or tend to look all the time? And if we believe Simmel’s claim (1969) about the importance of mutual gaze for human sociality, what would be changed in the interactions and social relationships of human beings living in a culture that appears to minimize the occurrence of mutual gaze (e.g., Tzeltal )? All these issues invite the development of a proper coding scheme and the collection of a larger sample to assess how many of the patterns here observed are still systematic in a larger dataset. For this purpose, we used extensive video recordings of natural conversation from these three cultures and developed a comparative coding scheme, as described below. Data and method The data we used for our comparative analysis of gaze behavior in interaction comes from a total of twenty-nine dyadic interactions11 in three different languages: Italian , Tzeltal, and Yélî Dnye .12 Italian is a Romance language spoken primarily in Italy, Tzeltal is a Mayan language spoken in Chiapas, Mexico, and Yélî Dnye is an isolate, presumably of Papuan origin, spoken on Rossel Island, Papua New Guinea. More precisely, each dataset has been collected in a specific part of the area where each language is spoken, and we believe the data are representative of the three different cultures, which are relatively homogeneous within each dataset.13 Figure 7.1 shows the location of the three cultures – as the geographical dispersion suggests, there has been no cultural contact between any of them, and thus any similarities in gaze pattern can be presumed to derive from universal tendencies or from systematic functions within an underlying shared interaction system. The ten dyadic interactions in Italian were recorded with two cameras, one on each participant, while the ten interactions in Tzeltal and the nine in Yélî Dnye were recorded with one camera, but with an angle that allowed for a good view of both participants’ faces. Twenty different individuals are included in the Italian dataset, seventeen in the Tzeltal one and eighteen in the Yélî Dnye one. The conversations recorded were ordinary and casual and not
Figure 7.1 Location of three cultures
Italian (Rossano)
Yélî Dnye (Levinson)
Tzeltal (Brown)
Gaze, questioning, and culture
205
solicited by the researchers. The participants of each interaction knew each other, being either relatives, friends or acquaintances. The interactions analyzed were dyadic because the presence of just two participants simplifies issues such as next-speaker selection (see e.g., L erner 2003) and who has been addressed by the speaker ( Goodwin 1979). Specifically, in a dyadic interaction if one person is talking, the other person is the one addressed. This means that if the speaker looks at the addressee during a turn, this is not done in the service of disambiguation to show who is being addressed and who is being selected to speak next. This allows gaze, freed from an address function, to serve a range of other interactional functions. This extensive database of interactions in the three cultures was sampled in a systematic way. Within each cultural sample, we selected roughly ten dyads and searched for question–response sequences until we had 300 such sequences for each cultural sample. By “question” we understood any utterance that functioned as an information soliciting action, regardless of whether it was in interrogative form or otherwise marked morphosyntactically, lexically or prosodically. For example, many “yes”/“no” questions in Tzeltal and Yélî Dnye are delivered in declarative format with falling intonation – they are recognizable as questions just because they appear to be statements about facts that are privileged information of the recipient’s, of the kind “You have a stomach ache” (cf. Labov and Fanshel’s 1977 “B-event statements” ).14 Note also that we excluded requests for things other than information: We were interested only in utterances that required a verbal or visible communicative response, which could thus be coded as an “answer,” “nonanswer response” or missing response. The question–response sequences collected in this way are not homogenous in function. Schegloff ’s (1984) point is well taken: An utterance can be a question and, at the same time, much more besides. Still, it is by virtue of being the first pair part of an adjacency pair (cf. e.g. Sacks 1987; Schegloff 2007b; Schegloff and Sacks 1973), calling for a specific type of second pair part, that questions do the multifarious things they do. For this study we have focused on just three gross functions: requests for new information, requests for repair (e.g., for repetition or clarification ), requests for confirmation (of previously established information or
206
Federico Rossano et al.
presumed facts). The reason for distinguishing these is the different degree to which they push the ongoing business forward (i.e., different degrees of “progressivity”; ( see Schegloff 2007b; Stivers and Robinson 2006), as opposed to invoking prior utterances or sequences. An other-initiated repair, for example, is a momentary hitch, occasioning the insertion of a sequence into whatever business is at hand – it is hoped and expected that completion of the repair sequence is immediate upon response, and work on the interrupted sequence can resume. Question–response sequences have a number of advantages for this study focusing on gaze behavior. First, the category is “emic,” a category for participants – an unanswered question is a potential matter for complaint. Second , like in all adjacency pairs, there is turn-transition relevance (cf. Sacks et al. 1974) immediately after the first pair part; moreover, normally a question is delivered as a simplex unit, a single turn-constructional unit in the great majority of cases (see Extracts 1–4). Third, the two parts, question and response, potentially form a complete sequence, and again in the majority of cases actually do so. Thus over the course of a question–response sequence, two turns have been traversed, the former speaker has become a ratified recipient, and the former recipient a ratified speaker, and, in addition to speaker change occurring, a whole sequence has been jointly engendered. These are all important loci where gaze has been thought, or might be thought, to play a crucial role in organizing interaction. We should add that most if not all the prior claims and rules about gaze in interaction did not specify whether they would apply only within a specific culture or to specific actions. They are usually presented as applying generally to every single turn at talk whenever two people are talking to each other. If we find that those claims do not hold for at least one type of action (questions) and for one or more cultures, the generality and universality of such claims can be considered disproved. This would further confirm the importance of restricting claims in terms of the domain under investigation and, even more importantly, the necessity of moving beyond the “turn level” and including the details of the actions performed and the sequential organization of participants’ talk as important variables that affect gaze in interaction, as proposed by Rossano (2005; 2006a; 2009).
Gaze, questioning, and culture
207
We coded for the participant role (e.g., speaker vs. recipient), the primary action being implemented with the question, the question type, and the occurrence or not of a visible and/or verbal response. We then developed a refined coding for gaze for both speaker and recipient. We wanted to capture, inter alia, absence of gaze at interlocutor and its possible cause, full gaze, unilateral vs. mutual gaze , and gaze at crucial loci within the first pair part. Details about the coding can be found in the appendix. The inter-reliability coding performed on 40 percent of the gaze coding for the Tzeltal data showed a Cohen Kappa statistic k = .875, S.E. = 0.03, p < 0.0001 (Cohen 1960) which is considered an almost perfect agreement (Landis and Koch 1977).15 Results In this section, we report the main outcomes of the statistical analyses, commenting briefly on the import of each finding but leaving aside extended discussion until the next section. First, let us compare speaker gaze and recipient gaze during the question utterances, asking whether speakers or recipients gaze at their interlocutor at least once during the question (from here on we refer to these as Q-speakers and Q-recipients, to remind the reader that we are referring to speakers and recipients of questions only). A first surprise is that, as made clear in Table 7.1 and Figure 7.2, in each language Q-speakers look at Q-recipients more often than vice versa. In particular, speakers gaze at addressees in 65.7 percent of the questions in Tzeltal , in 73 percent in Italian, and in 79.7 percent in Yélî Dnye . Recipients gaze at speakers in 42.3 percent of questions in Tzeltal, in 63.3 percent in Italian, and in 67.3 percent in Yélî Dnye . This is unexpected because it contrasts with all the above-mentioned works that found that participants Table 7.1 Instances of Q-speaker and Q-recipient gaze toward other participant, by language
Speaker Recipient
Tzeltal
Italian
Yélî Dnye
197 (65.7%) 127 (42.3%)
219 (73%) 190 (63.3%)
239 (79.7%) 202 (67.3%)
208
Federico Rossano et al. 90% 80%
Percentage
70% 60% 50%
Speaker gaze up
40%
Recipient gaze up
30% 20% 10% 0% Tzeltal
Italian
Yélî Dnye
Languages
Figure 7.2 Gaze toward other participant, by language
generally look more while listening than while speaking (see section “Earlier Work on Gaze in Interaction”), while it confirms what was shown in Extracts 1 and 2 in Italian. We therefore carefully checked the statistical reliability of more speaker than recipient gaze in each language through logistic regression analysis.16 Table 7.2 shows the results of the logistic regressions run for each language in which the standard error has been corrected for the clustering of the data by interaction.17 The difference between speaker and recipient gaze behavior is statistically significant for each language. Table 7.2 shows that when a Tzeltal participant acts as a Q-speaker, he or she is 2.61 times more likely to gaze at the addressee than when he or she acts as a Q-recipient. It shows that when an Italian participant acts as a Q-speaker, he or she is 1.57 times more likely to gaze at the addressee than when he or she acts as a Q-recipient. It finally shows that when a Yélî Dnye participant acts as a Q-speaker, he or she is 1.90 times more likely to gaze at the addressee than when he or she acts as a Q-recipient. We turn now to consider possible cross-cultural differences in gaze behavior. Figure 7.3 compares the gaze behavior of Q-speakers and recipients across the three languages. It shows that, in both, Q-speaker and recipient role participants look at each other more
Gaze, questioning, and culture
209
Table 7.2. Logistic regression analysis predicting gaze toward other participant in relation to role (Q-recipient as reference group), by language Predictor
β
S.E. β
p
Odds ratio
95% C.I.
Role Tzeltal (speaker) Constant
0.96 −0.31
0.17 0.28
0.000 0.023
2.61
1.86, 3.66
Note: Model Evaluation: Wald χ2(1) = 30.53, p < 0.001
Predictor
β
S.E. β
p
Odds ratio
95% C.I.
Role Italian (speaker) Constant
0.44 0.55
0.15 0.30
0.002 0.073
1.57
1.17, 2.09
Note: Model evaluation: Wald χ2(1) = 9.37, p < 0.01
Predictor
β
S.E. β
p
Odds ratio
95% C.I.
Role Yélî (speaker) Constant
0.64 0.72
0.23 0.32
0.006 0.276
1.90
1.21, 2.99
Note: Model Evaluation: Wald χ2(1) = 7.68, p < 0.01
90% 80%
Percentage
70% 60% Tzeltal Italian Yélî Dnye
50% 40% 30% 20% 10% 0%
Speaker gaze up Recipient gaze up Participation role
Figure 7.3 Gaze behavior across the three cultures clustered by participation role
210
Federico Rossano et al.
Table 7.3. Logistic regression analysis predicting Q-speaker gaze toward Q-recipient in relation to language (Italian and Tzeltal as reference group) Predictor
β
S.E. β
p
Odds ratio
95% C.I.
Tzeltal (speaker) Yélî (speaker) Constant
−0.35 0.37 0.99
0.40 0.29 0.25
0.385 0.212 0.000
0.71 1.45
0.32, 1.54 0.81, 2.59
Note: Model evaluation: Wald χ2(2) = 4.85, p = 0.09
Predictor
β
S.E. β
p
Odds ratio
95% C.I.
Italian (speaker) Yélî (speaker) Constant
0.35 0.72 0.65
0.40 0.35 0.30
0.385 0.039 0.036
1.41 2.05
0.65, 3.09 1.04, 4.05
Note: Model evaluation: Wald χ2(2) = 4.85, p = 0.09
in Yélî Dnye than in Italian and more in Italian than in Tzeltal , although the significance of these differences in frequencies will be shown to vary depending on the participation role . Let us first take the question of speaker gaze: Are Q -speakers more likely to gaze at their recipients during questioning in one of the languages compared to the others? If we compare Italian and Yélî Dnye , the answer is no: Taking Italian as the comparison language, there is no significant difference between Q-speakers’ gaze behavior in Italian and Yélî Dnye (see Table 7.3, top half) and equally no significant difference between Q-speakers’ gaze behavior in Italian and Tzeltal . However, when we take Tzeltal as the comparison language, there is a significant difference in Q-speaker gaze between Tzeltal and Yélî Dnye (see Table 7.3, bottom half). Tzeltal Q-speakers are significantly more economical with their gaze to recipients than Yélî Dnye Q-speakers. These latter results, however, should be taken with extreme caution, as the model evaluation shows that the Wald statistic is not significant. This means that the predictor (language) should be rejected as a good predictor of the outcome (Q-speaker gazing at Q-recipient). Language in general does not significantly predict speaker gaze , at least for these three languages.
Gaze, questioning, and culture
211
Table 7.4 . Logistic regression analysis predicting Q-recipient gaze toward Q-speaker in relation to language (Italian and Tzeltal as reference group) Predictor
β
S.E. β
p
Odds ratio
95% C.I.
Tzeltal (recipient) Yélî (recipient) Constant
−0.85 0.18 0.55
0.40 0.42 0.29
0.035 0.677 0.062
0.43 1.19
0.19, 0.94 0.52, 2.74
Note: Model evaluation: Wald χ2(2) = 7.36, p < 0.05
Predictor
β
S.E. β
p
Odds ratio
95% C.I.
Italian (recipient) Yélî (recipient) Constant
0.85 1.03 −0.30
0.40 0.41 0.27
0.035 0.013 0.273
2.33 2.79
1.06, 5.16 1.24, 6.27
Note: Model evaluation: Wald χ2(2) = 7.36, p < 0.05
In short, Table 7.3 shows that Italian Q-speakers’ gaze behavior does not differ significantly from Tzeltal and Yélî Q-speakers’ gaze behavior, while Tzeltal and Yélî differ significantly but in a model that suggests language does not properly predict differences in Q-speaker gaze behavior . Let us now turn to the gaze behavior of the recipient during questioning. Once again, there is no significant difference between Italian and Yélî gaze behavior, now considering Q-recipients (Table 7.4, top half). But both Italian and Yélî are strikingly different from Tzeltal in this regard: The difference between Q-recipient gaze behavior in Tzeltal and the other two languages is significant (see Table 7.4, bottom half) and so is the model used to test these differences (see model evaluation of Table 7.4). Table 7.4 shows that Yélî Dnye Q-recipients’ gaze behavior is not significantly different from Italian Q-recipients’ gaze, whereas Tzeltal Q-recipients’ gaze behavior is significantly different from Italian and Yélî Q-recipients’ gaze. In particular, Italian and Yélî Q-recipients are respectively 2.33 times and 2.79 times more likely to look at the speaker during a question than a Tzeltal Q-recipient is. In sum, Q-speakers’ gaze behavior is fundamentally similar across the three languages.18 Although there is a significant
212
Federico Rossano et al.
difference between the two extremes (Tzeltal and Yélî Dnye), Italian stands in the middle and is not significantly distinct from either. Moreover, the evaluation of the model shows that language is not a good predictor of differences in speaker gaze. All three languages show a significantly greater proportion of speaker gaze compared to recipient gaze during questioning. On the other hand, gaze behavior by the Q-recipient shows more fundamental differences in the case of Tzeltal: Now Italian and Yélî Dnye show similar, relatively high patterns of gaze by the Q-recipient to the speaker, while Tzeltal shows significantly less. These results suggest a strong tendency to uniformity in speaker gaze behavior, despite the literature (reviewed above) that suggests that speaker gaze in general varies radically across cultures. In contrast, there is a clearer cultural difference in the way in which gaze is deployed by recipients, and the implication is of course that gaze plays a differential role across cultures as a signal of active recipiency . If we now look at the occurrence of mutual gaze between Q-speaker and Q-recipient, that is, whether mutual gaze is engaged sometime during the course of asking a question, we find that mutual gaze occurs in just over 50 percent of questions in Italian and Yélî Dnye, but in only 33.7 percent of questions in Tzeltal (see Table 7.5). Once again, this aligns Italian and Yélî Dnye as highgaze cultures, with Tzeltal as a low-gaze culture, along the lines shown in Brown and Levinson ’s previous comparative work on Tzeltal and Yélî Dnye (2005; Levinson and Brown 2004). Table 7.6 shows that Tzeltal is significantly different from Italian and Yélî Dnye in terms of the occurrence of mutual gaze during a question, while Italian and Yélî Dnye do not differ significantly. We noted earlier that some of the gaze literature (e.g., C. Goodwin 1980; Goodwin 1981) suggests that as a rule speakers check that the recipient is attending, expecting to find a recipient gazing back at that point. This predicts of course that mutual gaze should occur in nearly all cases in which the speaker looks at the recipient, and where not, remedial action (e.g., getting attention by restarting an utterance) might be undertaken. The numbers for our data do not support any such rule, at least for question–answer sequences. Consider that in approximately 20 percent of questions in Italian and Yélî Dnye , and in more than 30 percent of questions in Tzeltal,19 the
Gaze, questioning, and culture
213
Table 7.5. Instances of mutual gaze by language
Mutual gaze
Tzeltal
Italian
Yélî Dnye
101 (33.7%)
159 (53%)
173 (57.7%)
Table 7.6. Logistic regression analysis predicting mutual gaze in relation to language (Italian and Tzeltal as reference group) Predictor
β
S.E. β
p
Odds ratio
95% C.I.
Tzeltal (MG) Yélî (MG)
−0.83 0.18
0.41 0.42
0.042 0.672
0.44 1.19
0.20, 0.97 0.53, 2.69
Note: Model evaluation: Wald χ2(2) = 6.37, p < 0.05
Predictor
β
S.E. β
p
Odds ratio
95% C.I.
Italian (MG) Yélî (MG)
0.83 1.00
0.41 0.43
0.042 0.020
2.29 2.72
1.03, 5.06 1.17, 6.35
Note: Model Evaluation: Wald χ2(2) = 6.37, p < 0.05
speaker is gazing at the recipient and the recipient does not gaze back before completion of the turn (see Table 7.7), and no remedial action is taken (see Extracts 1 and 2 for further qualitative evidence). We turn now to consider the distribution of gaze across finer coding categories. As mentioned, questions can implement rather different actions, and our coding scheme characterized these accor ding to their “progressivity,” distinguishing information-seeking questions from repair initiation questions and from confirmation seeking questions. Table 7.8 and Figure 7.4 show the distribution by language of these main classes of action. The residual category “Other” includes different actions such as offers , invitations , requests for further clarification , topic profferings, etc., that were implemented through the questions – they did not amount to a large number of instances for any given action for each language. There are slight differences in the occurrence of these different types. For example, Yélî Dnye has more requests for information
214
Federico Rossano et al.
Table 7.7. Questions in which speakers look at recipients and recipients do not look back, by language
Speaker gaze, no recipient gaze
Tzeltal
Italian
Yélî Dnye
96 (32%)
59 (19.7%)
66 (22%)
Table 7.8. Distribution of actions performed by questions, by language
Request for information Other-initiated repair Request for confirmation Other
Tzeltal
Italian
Yélî Dnye
107 (35.7%) 34 (11.3%) 115 (38.3%) 44 (14.7%)
126 (42%) 43 (14.3%) 68 (22.7%) 63 (21%)
143 (47.7%) 53 (17.7%) 84 (28%) 20 (6.6%)
60.0% 50.0% 40.0% Tzeltal Italian Yélî Dnye
30.0% 20.0% 10.0% 0.0%
Req. for info
O.I. repair
Req. for conf.
Other
Figure 7.4 Distribution of actions by languages
and other initiation of repair than the other two languages, while Tzeltal has more requests for confirmation . Again, the possibility arises that these differences in the frequencies of action types could account for some of the cross-cultural differences in gaze. If, for example, gaze is especially associated with requests for information and repair-initiating questions, given
Gaze, questioning, and culture
215
Table 7.9. Q -speaker looking at Q-recipient, by action and by language
Request for information Other-initiated repair Request for confirmation
90.0% 80.0% 70.0% 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% 0.0%
Tzeltal
Italian
Yélî Dnye
71 (66.4%) 26 (76.5%) 81 (70.4%)
87 (69%) 33 (76.7%) 50 (73.5%)
110 (76.9%) 43 (81.1%) 72 (85.7%)
Tzeltal Italian Yélî Dnye
Req. for info.
O.I. repair Actions
Req. for conf.
Figure 7.5 Distribution of percentages of each action in which the Q-speaker looks at the addressee, by language
that there are more of these in Yélî Dnye than in Tzeltal , this could explain the distribution of gaze shown in Figure 7.3. However, Table 7.9 and Figure 7.5 show that this is not the case. For each action, Q-speakers look more in Yélî Dnye than in Italian and more in Italian than in Tzeltal, although the difference between Q-speaker’s gaze behavior in Italian and Tzeltal is minimal, exactly as shown in the general distribution. It is noticeable here that, of the three actions, in each language requests for information are produced less often with speaker gaze than are other-initiations of repair and requests for confirmation . Considering Rossano ’s (2005, 2009) claim that gaze behavior is mainly organized around larger structures than turns (sequences and courses of action) and that most of the shifts in gaze direction should occur at the beginning and end of these structures and around self-repair and speech disfluencies (on this point cf. B eattie, 1979), the differences in the amount of gaze per action
216
Federico Rossano et al.
could be accounted for by their sequential position. The fact that other-initiations of repair do not occur at the beginning of a sequence or course of action, and the same usually holds for requests for confirmation, whereas requests for information can occur in first position, may account for this difference. Another general fact emerging from all three cultures is that questions initiating repair are especially likely to involve both speaker and recipient gaze. One possibility is that gaze here reinforces that the Q-speaker, though speaking, is committing to a more attentive engagement as a listener in the conversation. The speaker of the repair initiation is asking the recipient to repair his or her prior talk because of some problem, and the speaker of the repair question can ask to delay the progressivity of the talk by projecting a full recipiency once the repair is produced. Moreover, if speaker gaze can exert pressure for a response, this pressure can be particularly valuable given the social cost of asking a prior speaker to repair his or her own prior talk and the additional social cost of not being able to provide an appropriate response in case of troubles in hearing or understanding. Alternatively, one may take other-initiation of repair to convey disaffiliation, non-alignment or some other problematic stance toward the prior utterance; it may therefore be a kind of interactionally “loaded” moment, comparable to challenges , which are usually delivered while looking at the addressee. Independently of which is the favored account, it is clear that delaying the progressivity of the conversation and asking for repair is an interactionally problematic action, and the occurrence of more gaze by both speakers and recipients might be related to this potential problematicity. We turn now to consider Q-recipient gaze within each action type. Table 7.10 and Figure 7.6 show that, for each action, the distribution of Q-recipient gaze follows the general distribution of recipient gaze as shown in Figure 7.3. This means that Yélî Dnye and Italian Q-recipients look more during each action than Tzeltal speakers do, and the difference between Italian and Yélî Dnye is less marked than the difference between these two languages and Tzeltal. It is noticeable that in each language other-initiation of repair tends to get gaze more often from the Q-recipient than the other two actions. Moreover, in Italian and Yélî Dnye, requests for confirmation are the actions that are least likely to elicit gaze from
Gaze, questioning, and culture
217
Table 7.10. Q -recipient looking at Q-speaker, by question type and by language
Request for information Other-initiated repair Request for confirmation
Tzeltal
Italian
Yélî Dnye
40 (37.4%) 18 (52.9%) 55 (47.8%)
80 (63.5%) 28 (65.1%) 38 (55.9%)
100 (69.9%) 40 (75.5%) 51 (60.7%)
80.0% 70.0% 60.0% 50.0%
Tzeltal
40.0%
Italian Yélî Dnye
30.0% 20.0% 10.0% 0.0% Req. for info
O.I. repair
Req. for conf.
Action
Figure 7.6 Distribution of percentages of each action in which the Q-recipient looks at the Q-speaker, by language
recipients, while in Tzeltal it is requests for information that are the actions least likely to get gaze from the recipient. As we noted at the outset, questions are perceived as first pair parts of adjacency pairs, making a response by the recipient the relevant next action. This is the case independently of whether they are initiating a sequence, they are inserted in it, or they are working as post-expansion . Table 7.11 shows how many of our questions got proper answers, responses that would not count as proper answers (e.g., “I don’t know, maybe”) or no response at all. The results show that in each language proper answers are produced in response to questions two-thirds of the time or more. Moreover, the number of questions that receive no response at all is minimal (on average, in 12 percent of cases).20 It is clear that overwhelmingly questions get responded to (on average, in 88 percent of cases; see Table 7.12).
218
Federico Rossano et al.
Table 7.11. Distribution of response types by language
No response Non-answer responses Answers
Tzeltal
Italian
Yélî Dnye
15 (5%) 41 (13.7%) 245 (81.7%)
45 (15%) 71 (23.7%) 184 (61.3%)
50 (16.7%) 52 (17.3%) 198 (66%)
Table 7.12. Distribution of responses to questions, by language
No response Response
Tzeltal
Italian
Yélî Dnye
14 (4.7%) 286 (95.3%)
45 (15%) 255 (85%)
50 (16.7%) 250 (83.3%)
Gaze per se does not, nor did we expect it to, account for the occurrence of an answer or a nonanswer response, as action types, question format, and many other factors can come into play here. There is, however, an interesting correlation between lack of Q-recipient gaze (and of mutual gaze as a consequence) and the fact that an answer does not get a response. If we take Q-recipient gaze to be a means for displaying recipiency , we should expect that when Q-recipient gaze is lacking there should be some interactional consequence. Given that here we are looking at questions, the most notable consequence would be problems in responding to a question. We are here considering only the most macro display of “trouble” in responding: i.e., not responding at all. It is clear that other micro-consequences can occur and would be worth examining in future work. Notice that we are not expecting that every time the recipient is not looking there will be no response but rather that absence of Q-recipient gaze can be a good predictor of upcoming trouble in responding, or, more specifically, of no response. Extract 5 (as well as Extract 1) provides a qualitative example of cases when the Q-recipient is not looking at all during the question and there seems to be a problem and a response is lacking. In Extract 5, two male Italian undergraduate students are preparing for an exam. B is asking A questions and is holding the notes they have used to prepare for the exam. B has already passed the exam while A has to take it the following day. The topic being discussed
Gaze, questioning, and culture
219
is the kinds of diseases that one should report to the officials while working as a veterinarian. We focus on the questions at line 1, 3, and 5. (5) ITALIAN 2PEXAM-sapere (42:49) A
B
A
B
01 →A: Ma lei vuol sapere ste cose qui? but she want.3s to know these things here Does she want (us) to know these things here? A
02
B
(1.6) A
03 →A:
B
Eh? eh Eh? A
B
A
04 (1.0)
B
A
(0.2) (1.6)
A
B
(0.4)
B
A
B
A
B
05 →A: Lei son ste cose qua [che vuole sapere?] she are these things here that want.3s to know Are these the things [that she wants (us) to know?] A
B
A
B
A
B
06 B: [Lei vuole sape]re quelle she want.3s to know those [She wants (us) to know those A
B
07 ‘segnalazioni immediate’ = si’ questo e’ importante signaling immediate yes this is important ‘inform immediately’ = yes this is important
In all three questions the recipient is not looking at the speaker but at the notes (although the answer to the questions is not in the notes). The question at line 1 is not answered, and A pursues it at lines 3 and 5, until he finally gets an answer at lines 6–7 (and
220
Federico Rossano et al.
Table 7.13. Distribution of questions that do not get responded to in which the recipient does not look at the speaker
No recipient gaze
Tzeltal
Italian
Yélî Dnye
9/14 (64.3%)
23/45 (51.1%)
32/50 (64%)
B looks toward A while producing it). Notice that even though B answers, the formulation of the answer is not straightforward (the “yes” comes only mid line 7). B was not looking at A during the questions and it is clear that he had problems dealing with the question(s). See also Extract 1 for similar evidence. If we focus purely on the numbers from our dataset, we see that in Italian and Yélî Dnye, the proportion of gaze absence by the recipients of the questions that are not responded to is higher than in the total of the questions (see Table 7.13). 21 We can test whether lack of recipient gaze is a good predictor of lack of response by dividing the response possibilities into response vs. no response, as in Table 7.12. We tested this hypothesis by creating a model that takes into account other variables such as speaker gaze, recipient gaze, and language spoken (Table 7.14). Table 7.14 shows that, in general, lack of Q-recipient gaze is a good predictor of lack of response after a question, while speaker gaze and language spoken are not (at least for this dataset). Table 7.15 shows that lack of recipient gaze works differently in each language. It is a significant predictor of lack of response in Italian and Yélî Dnye , but not for Tzeltal . This means that while the proportion of no responses preceded by lack of Q-recipient gaze is significantly higher than for all questions in Yélî Dnye and Italian, this is not the case for Tzeltal. These results are particularly relevant because they show that, in Italian and Yélî Dnye , Q-recipient gaze appears to be related to the doing of recipiency, and, in this respect, the lack of it can predict possible upcoming trouble in terms of the occurrence of a response to the question. However, in Tzeltal, recipient gaze does not seem to have this function, and this fits with the general tendency for Tzeltal recipients to gaze less at the speaker than in the other two
Gaze, questioning, and culture
221
cultures. This raises important issues about the general function of gaze as a display of recipiency: It suggests that this role for gaze is not universal , and there must be other possible ways in which good recipiency can be signaled. All the results just outlined distinguish questions in which the participants looked at any time toward the other participant from those in which the participant never looks at the other. The coding scheme we had developed, however, was more specific, as it distinguished thirteen possible scenarios (six distinct locations for gaze for each of speaker vs. recipient, and the possibility of mutual gaze ). Once these data are taken into account, the results are again surprising, as they contrast with claims by Kendon (1967) for a wider range of actions about when speakers typically look away and when they typically look at the recipient within the turn. Tables 7.16 and 7.17 show that, in each language, only in a small minority of questions do participants look up during the question if they were not already looking at its beginning. In particular, for Tzeltal Q-speakers, this happens in 10.7 percent of the questions, in Italian in 16.4 percent, and in Yélî Dnye in 12.3 percent (these percents are the sums of rows 4, 5, and 6 in Table 7.16). More specifically, if looking toward the recipient near the completion of an utterance invites speaker transition (see Kendon 1967), row 5 in Table 7.16 shows that in our question–answer data this happens very infrequently. For Tzeltal this happens in 7.7 percent of the questions, for Italian in 5 percent, and for Yélî Dnye in 5 percent. Table 7.17 shows percentages for Q-recipients of gaze “immobility” within a question similar to those in Table 7.16 for
Table 7.14. Logistic regression analysis predicting no response Predictor
β
S.E. β
p
Odds ratio
95% C.I.
No recipient gaze Speaker gaze Tzeltal Yélî Constant
0.61 −0.31 −1.00 −0.29 −1.78
0.25 0.23 0.54 0.41 0.35
0.015 0.191 0.063 0.481 0.000
1.84 0.73 0.37 0.75
1.12, 3.02 0.46, 1.17 0.13, 1.06 0.33, 1.68
Note: Model evaluation: Wald χ2(6) = 84.25, p < 0.00001
222
Federico Rossano et al.
Table 7.15. Logistic regression analysis predicting no response if Q-recipient is not gazing, by language Predictor
β
S.E. β
p
Odds ratio
95% C.I.
No recipient gaze (Tzeltal) Constant
0.11
0.54
0.842
1.11
0.39, 3.21
−3.00
0.42
0.000
Note: Model evaluation: LR χ2(2) = 0.40, p = 0.84
Predictor
β
S.E. β
p
Odds ratio
95% C.I.
No recipient gaze (Italian) Constant
0.70
0.33
0.031
2.02
1.07, 3.83
−2.03
0.23
0.000
Note: Model evaluation: LR χ2(2) = 4.61, p < 0.05
Predictor
β
S.E. β
p
Odds Ratio
95% C.I.
No recipient gaze (Yélî) Constant
1.59
0.33
0.000
4.92
2.58, 9.37
−2.29
0.39
0.000
Note: Model evaluation: LR χ2(2) = 25.22, p < 0.001
Table 7.16. Q -speaker gaze, by language
1. No gaze 2. Gaze up all the time 3. Gaze up at beginning, away before completion 4. Gaze up on new referent 5. Gaze up approaching completion 6. Gaze up any other time
Tzeltal
Italian
Yélî Dnye
103 (34.3%) 146 (48.7%) 19 (6.3%)
81 (27%) 162 (54%) 8 (2.7%)
61 (20.3%) 187 (62.3%) 15 (5%)
2 (0.7%) 23 (7.7%)
8 (2.7%) 15 (5%)
7 (2.3%) 15 (5%)
7 (2.3%)
26 (8.7%)
15 (5%)
Q-speakers. In particular, a Q-recipient looks toward the speaker’s face after the beginning of a question in 15.4 percent of questions in Tzeltal , in 16 percent in Italian, and in 7.7 percent in Yélî Dnye (the sums of rows 4, 5, and 6 in Table 7.16).
Gaze, questioning, and culture
223
Table 7.17. Q -recipient gaze, by language
1. No gaze 2. Gaze up all the time 3. G aze up at beginning, away before completion 4. Gaze up on new referent 5. G aze up approaching completion 6. Gaze up any other time
Tzeltal
Italian
Yélî Dnye
173 (57.7%) 63 (21%) 18 (6%)
110 (36.7%) 120 (40%) 22 (7.3%)
98 (32.7%) 162 (53.3%) 17 (5.7%)
0 (0%) 29 (9.7%)
4 (1.3%) 14 (4.7%)
1 (0.3%) 11 (3.7%)
17 (5.7%)
30 (10%)
11 (3.7%)
These results contrast with the patterns claimed by Kendon (1967) to occur in English and suggest, as argued by Rossano for Italian (2005; 2009), that the utterance per se is not an adequate level for describing gaze behavior in interaction, as very little seems to happen to gaze direction in dyadic interactions during a single utterance. 22 Finally, we noticed that looking away, or better our coding “no gaze toward addressee” could actually refer to three different situations: 1. The participant is just looking down or away but at nothing in particular. 2. The participant is looking away because of the content or the delivery of the question (e.g., someone is pointing and looks in the direction of the point, or is talking about an object present in the surrounding environment and looks at it). 3. The participant is involved in a competing activity (e.g., eating, weaving, drawing on the sand). We therefore monitored whether the lack of gaze toward the addressee could be considered “motivated” or not and whether this alternative focus was related to the talk or to a competing activity. Tables 7.18 and 7.19 show the great cross-cultural difference in terms of what else is going on while talking and how often people’s eyes get redirected toward different semiotic entities. Tables 7.18 and 7.19 show that, while in our Italian interactions Q-speakers and Q-recipients can be seen to be looking away for a possible reason around 80 percent of the time, this happens around 50 percent of the time in Yélî Dnye interactions and
224
Federico Rossano et al.
Table 7.18. Questions in which lack of Q-speaker gaze toward Q-recipient motivated by question content or competing activity.
Total motivated Q. delivery/content Other activity
Tzeltal
Italian
Yélî Dnye
6/103 (5.8%) 5/103 4/103
70/81 (86.4%) 16/81 62/81
33/61 (54.1%) 20/61 21/61
Note: The sum of “gaze orientation motivated by question delivery/content” and “gaze orientation motivated by competing activity” does not correspond to the “Total motivated gaze orientation” because in some instances participants are conflating the two possible motivations, e.g., if someone points to an object in a picture while the competing activity is looking at pictures or if someone is drawing on the sand and uses deictics in his or her speech to refer to the object just drawn.
Table 7.19. Questions in which lack of Q-recipient gaze toward Q-speaker motivated by question content or competing activity
Total motivated Q. delivery/content Other activity
Tzeltal
Italian
Yélî Dnye
8/173 (4.6%) 4/173 4/173
85/110 (77.3%) 20/110 77/110
42/98 (42.9%) 10/98 39/98
only around 5 percent of the time in Tzeltal interactions. This means that in the culture where most of the time participants do not look at each other (Tzeltal) this is not because they are looking in the direction of a pointing finger or because they are usually involved in competing activities. Rather, they tend to look down, toward their hands or legs or in the middle distance but not at a specific object or in a recognizable direction. As will be explained in the following section, speakers of Tzeltal and speakers of Italian inhabit very different material worlds, and the interactions were recorded mainly at home in Italian, and mainly outside the home in Tzeltal and outside in Yélî. It is remarkable that the “nonlooking” at the other participant is mainly a clear looking at something else in the Italian interactions, and half of the time in the Yélî interactions, but almost always looking down or mid-distance in Tzeltal. This might be another cultural difference in terms of displaying attentive recipiency. If, in an Italian conversation, one is expected to be looking at the speaker most of the time, then not looking
Gaze, questioning, and culture
225
at the speaker is accountable and should be “motivated”. In contrast, if the default of recipiency is not looking at the speaker, then ostensive focus toward other objects or activities might mean that the recipient is not listening to the speaker at all. So by not looking at anything else, Tzeltal recipients might simply be displaying full commitment to attending to the conversation, while the same behavior enacted by an Italian participant would appear to display complete uninterest toward the conversation. In other words, speakers in these two cultures are dealing with the same problems (full recipiency and commitment to listening) using opposite visible displays. These behaviors follow directly from the nature of preferred gaze behavior in conversation within their respective cultures. To sum up this section, the following general patterns of gaze in question–answer sequences across the three cultures emerged: 1. Q-speakers gaze at recipients in about two-thirds of the questions, but Q-recipients’ gaze is more variable and less frequent. 2. Q-speakers gaze least on the most progressive actions. 3. Q-recipients gaze most on the least progressive actions. 4. Mutual gaze is far from assured during such sequences and never occurs in much more than half of the cases. There are, however, some striking cross-cultural differences, most clear in the Tzeltal vs. Yélî Dnye contrast: Here, both Q-speakers and Q-recipients gaze significantly more in Yelî Dnye than in Tzeltal. In the Tzeltal case, gaze is not a good indicator of active recipiency, and it plays no role as a predictor of whether a response will be forthcoming. Moreover, nonlooking at the other participant usually means very different types of orientation in the three cultures. In particular, participants do very different things with their eyes if they do not orient them toward the other person: They are mainly oriented toward competing activities in Italian interactions, they are oriented toward competing activities or to locations motivated by question delivery only half of the time in Yelî Dnye interactions, and they are mainly oriented down or middistance in Tzeltal interactions. These differences cannot be accounted for in terms of different patterns in the uses of questions within the three cultures.
226
Federico Rossano et al.
Reasons for, and consequences of, cultural variability in gaze behavior in interaction Brown and Levinson (2005; Levinson and Brown 2004) have noted, based on many years of observation, that differences in gaze behavior between Yélî Dnye and Tzeltal speakers correlate with differences in observed seating patterns. Broadly speaking, Yélî Dnye speakers of Rossel Island prefer to sit face-to-face within reach of each other, while Tzeltal speakers of Tenejapa prefer to sit side by side or at an angle.23 In our samples of interaction, these patterns are indeed replicated. There are, of course, additional cultural constraints here – Rossel Islanders do not invite people of other families into their homes, and they have no chairs, consequently all interactions were filmed standing, squatting or seated on the ground in the open air. The Tenejapan home is also relatively private, and visitors would normally be seated on benches or small chairs on a patio or enclosed public area: The recordings are therefore filmed either outside or in the less private parts of the home. The Italian interactions were filmed inside the home, where seating is normally arranged around a table for face-to-face interaction. Our recordings show that participants speaking Italian and Yélî Dnye prefer sitting face-to-face during conversation, while participants speaking Tzeltal prefer sitting side by side. 24 However, as can be seen in Figures 7.7–7.9, in all three cultures participants can be seen sitting side by side or face-to-face and engaging mutual gaze independently of their body position. The fact that in each culture we can observe both side by side and face-to-face configurations, and in both cases participants can be seen to engage mutual gaze , suggests a certain caution in distinguishing these cultures in terms of typical or normative seating configurations. Nevertheless, in our sample in Italian and Yélî Dnye, in most of the dyadic interactions participants sit face-toface, whereas in Tzeltal they sit side by side. 25 These different seating configurations would suggest that in Tzeltal looking at each other is a more marked act (see, e.g., Schegloff 1998c on body torque), given that it is often necessary to turn one’s head to look at the other participant (Brown and L evinson, 2005; Levinson and Brown, 2004), and this, of course, is consonant with the less frequent gaze in interaction. On the other hand, we should bear in mind
Gaze, questioning, and culture
227
Figure 7.7 Italian dyads
that in our dataset Tzeltal speaker gaze is not significantly different from Italian speaker gaze, and this shows that there are distinct limitations on the ecological determinism of gaze behavior. The early literature on human kinesics also throws considerable doubt on any theory of ecological determinism of gaze behavior .
228
Figure 7.8 Tzeltal dyads
Federico Rossano et al.
Gaze, questioning, and culture
229
Figure 7.9 Yélî Dnye dyads
These studies showed that participants tend to arrange themselves according to the interactions they plan to have (Ekman and Friesen 1974; Kendon 1977; Sommer 1959, 1962). While it is true that the physical structure of the environment constrains interactional positioning ( Goffman 1971; Goodwin 1981), these
230
Federico Rossano et al.
physical arrangements tend to directly reflect the preferred cultural patterns. In short, we would argue that the Tzeltal seating patterns, for example, directly reflect the preference for more limited, more controlled gaze behavior, in just the same way that the unfurnished and thus more unconstrained sitting arrangements of Yélî Dnye speakers favor the eyeball-to-eyeball interaction that is the dominant mode. So what does account for the cultural differences? Brown and L evinson (2005; Levinson and Brown 2004) note that the contrast between Tzeltal and Yélî Dnye is partly based on different practices for the display of recipiency . For example, during extended turns at talk such as a telling, Tzeltal recipients are expected to respond at regular intervals with significant verbal material, repeating parts of the immediately prior utterance (Brown 1979: Chapter 4, 1998, 2007). In contrast, Yélî Dnye speakers have an extended inventory of visual feedback signals, including facial gesture s specialized for assent, surprise, continuer function, and so forth ( L evinson 2007). The Yélî system presupposes the likelihood of gaze (or at least close peripheral vision), while the Tzeltal system seems built to assume its absence. The low-key role that gaze plays in Tzeltal interaction implies, as mentioned, other effective displays of recipiency. Extract 6 shows how repetitions can signal recipiency and understanding of what has been asked in a much more precise way than is observable in Italian and Yélî Dnye. 26 It shows a “yes”/“no” (polar) question– answer sequence in Tzeltal. This kind of question particularly allows an answer to be constructed by repetition with minimal modification – a repeat rather than “yes” is in fact the default for an affirmative response to a “yes”/“no” question (unlike what was shown by Raymond [2003] for conversations in American English ); the Tzeltal words for “yes” have much more restricted functions ( Brown 1998, in press). (6) Tzeltal T002020, 13:13 01 AMT: kuxul-0 to wan s-tukel i (.) muk’ul jme’tik i? living-3A still perhaps 3E-self DEIC big Mrs. DEIC? Is “big Mrs.” ((honorific for “your mother”)) still alive? 02 CH: kuxul-0 living-3A She’s alive.
Gaze, questioning, and culture
231
Extract 6, together with Extract 3, illustrate how repeating part of the polar question in the answer is an appropriate and typical way of responding in Tzeltal . 27 If one probes further for why Tenejapan Tzeltal speakers are relatively gaze aversive, there is little doubt that it has to do with politeness and decorum: Tzeltal interactors will even turn their backs, or hide behind a structure, if the conversational matter is in any way face-threatening (e.g., in the case of substantial requests – see Brown 1979 for details). Darwin (1872) was particularly interested in gaze aversion as a symptom of self-conscious shyness and shame, associated especially with the appropriate demeanor of Victorian women. From this, it seems to acquire the semiotics of self-denigration, which figures in many honorific systems (see Brown and L evinson 1987). Such an account predicts gaze aversion, especially in disprefer red responses, which needs further checking in the data. If the semiotics of gaze aversion has a “natural” source, we may expect this to hold in all three cultures. Implications for general patterns of gaze in interaction As already mentioned, one of our main cross-cultural ly stable findings is that Q-speaker gaze is significantly more expectable than Q-recipient gaze in each culture. This was not predicted by the prior literature, where, on the contrary, the assumption has been that generally recipients are expected to gaze, and speakers may well only glance at recipients somewhere in the course of the utterance, e.g., toward its end (e.g., Duncan and Niederehe 1974; Goodwin 1981; Kendon 1967). Two further points are in order. Some of these earlier claims are phrased in terms of rates of gaze, i.e., gaze over time, while our data has been coded as gaze over turns (question or response). Nevertheless, since we have found that if there is gaze it tends to be more or less constant over these units (see Tables 7.16 and 7.17), our findings too imply a temporal preponderance of speaker gaze. This constancy of gaze during the asking of a question also contrasts with another general claim, made most forcefully by Kendon (1967), according to which speakers tend to give quick glances at recipients while recipients tend to hold their gaze toward the speaker. Our results show that only rarely will a participant start
232
Federico Rossano et al.
glancing at the other and then look away and maybe glance back again during a question. 28 Where speaker gaze is momentarily averted this seems to be occasioned by gaze pointing for the other participant, or, in the case of recipient gaze, by the need to redirect the eyes toward the object the other participant is talking about. Gaze redirection in these instances – looking away from the recipient – is not related to the management of turn-taking . The general lack of mobility in gaze direction within a question has implications for the proper level of organization for gaze in interaction. A question typically consists of a single turn constructional unit (TCU). This suggests a general equivalence between our unit “question” and a turn at talk, and, indeed, 90 percent of the questions of the Italian corpus, and 77 percent of those in the Tzeltal and in the Yélî Dnye corpora, constitute a turn by themselves. It follows straightforwardly that gaze cannot be playing an internal function within the turn, e.g., by warning of turn transition (as Duncan and Fiske 1977; Duncan and Niederehe 1974; Kendon, 1967 had supposed). For the same reasons, C. Goodwin’s (1980; 1981) “recipiency” rules for American English interactions do not seem to hold for question turns in these three languages. Recollect that, according to these rules, the recipient’s gaze should be toward the speaker by the time the speaker looks at the recipient, failing which the speaker can solicit such attention through, for example, self-repair . Table 7.7 has already shown that in 20–30 percent of questions in our samples, the speaker looks but the recipient fails to look back, and clearly the speaker has not extended the turn to successfully achieve mutual gaze . In Italian, for example, in only four out of fifty-nine (or 7 percent) questions of this kind do sound stretches, cut-offs or pauses occur.29 If we look at the thirty-eight Italian questions in which the speaker is gazing from the beginning but the recipient looks up only later, we find a cut-off or a sound stretch in only seven (and only three seem to have any timing correlation to the return of recipient’s gaze toward the speaker). The difference between Goodwin’s findings and ours may be due to a number of factors, in particular related to the sampling of the data and the details of the coding . For example, we looked only at question sequences, while Goodwin’s data was coded intact and no particular attention was paid to the specific actions that the participants were implementing with their turns at talk
Gaze, questioning, and culture
233
(C. Goodwin 1980). Still, it is possible that Goodwin’s account in terms of speakers’ restarts eliciting recipient gaze has captured a practice that is especially prominent in American English .30 In any case, what is clear is that neither the importance of obtaining the recipient’s gaze before the completion of the turn nor the general normativity of this behavior can be observed in our samples. What role could the restriction to question sequences play in our data? Clearly, by being first-pair parts of adjacency pairs, questions imply turn transition, and in a dyadic situation they also imply to whom the next turn belongs. If gaze here plays no addressee-selection role, it may especially play the role of exerting pressure for a response ( Rossano 2006b, 2009; Stivers and Rossano in press). It is possible too, that gaze in questions is especially devoid of turn-transition functions, since recipients will know, as soon as they detect a question, that on completion an answer is immediately relevant. Despite finding little evidence for a role for gaze in regulating turn-taking in our three cultures, there does seem to be a correlation between the variation in gaze behavior within a question and the sequential position of the question. This trend can be seen reflected in Table 7.9, where requests for information, which are mostly sequence initial, are accompanied by the least amount of speaker gaze. Requests for repair and confirmation , which deal with prior business, are associated with more speaker gaze. This same pattern is reflected in recipient gaze, at least as regards repair (Table 7.10): There is more recipient gaze during repair initiation, a class of action that is by definition not in first position. These trends have been further explored in the Italian data, which has been coded also in terms of the sequential position of the question, by which we mean, for example, whether it initiates a new course of action or not (see Schegloff 2007b). Fifty-nine of the Italian questions were starting a course of action, which is to say less than 20 percent.31 The question here is whether in this sequential position the gaze behavior is comparable to the one observed for the total of the questions, when their sequential position was not taken into account. We want to focus in particular on whether participants tend to start courses of action already looking at each other or not, which is to say whether they tend to look up only during the sequence-initiating question. The Italian Q-speakers look at Q-recipients before the beginning of turns that start new courses
234
Federico Rossano et al.
of action only in 27.1 percent of the cases, while Q-recipients do so only in 25.4 percent of the cases. In 42 percent of these fifty-nine questions the Q-speaker directs his or her gaze toward the recipient only during the course of the question and not before the beginning. In 35.6 percent of these fifty-nine questions initiating a course of action, the recipient directs his or her gaze toward the speaker only during the course of the question, not before the beginning. If these latter percentages are compared with the general proportion of gaze shifts toward addressee during the question shown in Tables 7.16 and 7.17 (16.3 percent for speakers and 16 percent for recipients), it is clear that there is more “mobility” in terms of gaze when initiating a course of action than when questioning in the middle of a course of action (Rossano 2009). These numbers therefore confirm what was already shown in Extracts 1 and 2. Moreover, these numbers suggest that Italian speakers do not regularly start new courses of action only after they have secured gaze from the recipient. Actually, they look at the addressee before starting speaking only a fourth of the time in this specific sequential position. We have here claimed that participants will tend to look up during the question rather than before the beginning of it mainly during questions that initiate new courses of action. However, if we focus, for example, on Q-recipients in Italian conversations, we can see that out of the forty-eight questions in which Q-recipients look up during the question and not before, twenty-seven do not occur at the beginning of a course of action. Why in these cases is the recipient not looking at the speaker from before the beginning of the question (as appears to be the case in most of the non-initial questions) but instead looks up only during the question? Two clear interactional patterns can be observed in the Italian data. The first one unfolds as follows. Some of these questions are initiating a post-expansion of the sequence at a point when the sequence has been treated as complete by the other participant. In other words, the speaker who initiates the post-expansion re‑opens a finished sequence. The recipient of the post-expansion is the person who had brought the sequence to possible completion. As mentioned earlier, sequence completion typically involves withdrawal of gaze ( Rossano 2005), and the recipient will therefore have withdrawn his or her gaze before the occurrence of the question. When the other participant launches the post-expansion, the recipient
Gaze, questioning, and culture
235
abandons any other activity and reorients gaze toward the current speaker before completion of the question. Extract 7 shows this pattern. A and B are a couple and they are having lunch. A has announced that he is going to borrow a motorbike from his grandfather but he is worried about all the additional expenses. B reassures him about the expenses and promises to lend him her helmet for few months so that he can delay the expense for it. The target is the gaze behavior during the question at line 8. (7) Italian:2PLUNCH1-casco 14:14 01 B: Io il casco te lo presto (0.6) e poi:: I the helmet you it lend.1s and then I will lend you the helmet (0.6) and the::n 02 prima o poi te lo compri before or after yourself it buy.2s sooner or later you will buy one 03 (0.5) 04 B: Perche’ per questa estate ti presto quello integrale, Because for this summer you lend.1s that full face Because for this summer I will lend you the full face one, 05 quando arriva l’inverno che l’integrale when arrive.3s the winter that the full face when winter comes the full face one A
B
A
A
B
B
A
B
06 (0.9) serve a me te lo c o m p r i. Serve.3s to me yourself it buy.2s (0.9) is needed by me you buy one for yourself. A
B
07 (0.9) A
A
B
(0.7) (2.0)
B
A
A
B
(0.4) B
08 →A: Quanto costa un casco? POST-EXPANSION How much cost.3s a helmet How much does a helmet cost? 09 B: Non lo so Not it know.1s I don’t know
236
Federico Rossano et al.
B’s turn between lines 1 and 6 is closing the topic of general expenses for the motorbike and the issue of the cost of the helmet by promising A to lend him her helmet until the following winter (the interaction occurs at the beginning of April). At the end of line 6, both participants are not looking at each other (see Rossano 2005), and for two seconds there is silence. Both participants are oriented toward the dishes in front of them on the table. At line 8, A asks a question about the cost of the helmet, and, by doing so, he reopens the sequence, starting a post expansion, to which B responds with a non-answer response. The sequence is then further expanded. The second pattern of shift in recipient gaze during a question unfolds as follows. The person that has been coded as a recipient for the question was actually speaking in overlap or displaying an intention to initiate a turn at talk by producing a long in-breath when the speaker of the question starts his or her turn. The beginning of the question initiates in overlap and the recipient of the question orients toward the speaker only at the moment in which he or she abandons his or her overlapping turn and acts as a recipient. Extract 8 shows this pattern (see also line 12 of Extract 2 for additional evidence). A and B have met to prepare for an exam, and B is supposed to ask A questions to check whether he is ready for the exam. Our target is the gaze behavior at line 6. (8) Italian:2PEXAM-zonose 40:35 01 B: Cioe’ cosa hai letto della Marelli I mean what have.2s read by Marelli I mean what did you read by Marelli 02 A: Ho letto gli appunti della:dell’ Ilaria che erano have.1s read the notes of of Ilaria that were.3p I read the notes by:by Ilaria that were 03 fra fotocopie e:.hh e articoli di legge among Xeroxes an:d and articles of law among Xeroxes an:d .hh and law articles 04
(1.8) A
B
A
B
05 A: Cioe’ so:: [so le classificazioni,] I mean know.1s know.1s the classes, I mean I kno::w [I know the classes,]
Gaze, questioning, and culture A
B
A
237 B
A
B
A
06 → B: [Cioe’ le zoonosi in Italia principali quali sono I mean the zoonoses in Italy main which are [I mean which are the main zoonoses32 in Italy 07
B
(0.5)
A: Le zoonosi in Italia? 08 The zoonoses in Italy The zoonoses in Italy? 09
(0.4)
B: Eh 10 Eh B: Eh
B’s request for information at line 1 is responded to by A at lines 2 and 3. After a silence of almost two seconds, A starts talking again, providing further specification of what he knows and therefore what B can ask him questions about. However, immediately after the beginning of line 5, B, who is looking at some notes he is holding in his hands, starts talking as well and asks A a possible exam question. This is followed by a repetition of part of line 6 that initiates repair at line 8, which is confirmed by B, and the sequence continues with A’s tentative answer to B’s question. If we look at the gaze of the participants, we can see that while B keeps looking at his notes throughout the entire question at line 6, A, who was looking away toward his right side, turns toward B as soon as the first two words in B’s turn are produced. A, therefore, turns toward B as soon as he hears B talking in overlap and abandons the turn that he was producing soon after. This shows that once the sequential environment of the question is taken into account, some further order is observable. In particular, it shows that the beginning and ending of sequences and courses of action are the most vulnerable environments – in Italian at least – in terms of sudden shifts in gaze direction, and it also accounts for those instances in which the shift in gaze is not due to the sequential position of the question but rather to a parallel shift from speakership to recipiency that had to occur because of overlapping speech.
238
Federico Rossano et al.
Conclusions It is clear that gaze plays a delicate and complex role in social interaction. In this paper we have explored this conversational role mainly through the crude instrument of descriptive statistics. We have done this primarily because we have here been pursuing similarities and differences between conversational practices in unrelated cultures. A statistical approach can firmly establish the existence of distinct practices in the utilization of gaze, practices that have only partly been observed in qualitative analysis. If, for example, we focus just on Extracts 3 and 4, we can immediately notice the importance of having grounded their representativeness in a larger corpus. While being representative of a large number of questions in Tzeltal and Yélî Dnye, Extract s 3 and 4 only represent a minority of cases in each language. Indeed, mutual gaze (as shown in Extract 4) occurs in only half of the questions in Yélî Dnye and is sustained from beginning to end of the question in even fewer cases. In more than 65.7 percent of questions in Tzeltal, the speaker looks at the addressee, and the recipient looks in 42.3 percent of the cases. This means that the number of instances in which neither participant looks during a question (as shown in Extract 3) would necessarily be a minority. Exposing the extremes, even if recurrent and not isolated cases, could risk distracting us from a proper reconstruction of what empirically appears to be the most general pattern in a culture. The use of large corpora, in this sense, is necessary if the goal is to develop a comparative analysis of the systematicity of specific practices across different cultures. By proceeding in this way, we have indeed found distinct cultural practices, especially in the relative lack of use of gaze as an indicator of engaged recipiency in Tzeltal . We have also noticed alternative “home positions” for the eyes when a participant is not looking at the other’s face: down, toward hands or legs or middistance for Tzeltal participants, mainly toward objects or places motivated by question delivery for Italian participants, and a bit of both for Yélî Dnye participants. We know from observation that there are also other distinct practices that could be culturally specific, for example the use of gaze to “point” to unseen, even distant, entities in Yélî Dnye ( L evinson 2007). All these practices
Gaze, questioning, and culture
239
clearly affect the “accountability” and semiotic relevance of specific gaze orientations in conversation and merit further examination to establish their interactional consequences. But of at least equal importance are the general findings that hold across all three cultures and thus are candidates for general interactional properties. First, during questions, it is the Q-speaker rather than the Q-recipient who is more likely to be gazing. As pointed out repeatedly, this is at odds with prior generalizations about the role of gaze in conversational practice in general, where its functions have been presumed to be largely a display of recipiency and attentiveness. Second, the dominant pattern is with gaze being fixed throughout the turn, and thus it cannot play a systematic role (in the question context at least) as a turn-transition cue, or as a cue that participant role is switching. Third, the findings in all three languages are consistent with a detailed qualitative analysis of gaze in Italian (Rossano 2009), which shows that gaze is tied into sequence initiation and sequence completion rather than into the turn-taking system directly. We therefore think it likely that this observation too has general application across cultures. While our study has distinct limitations, especially because of its restriction to question sequences and the relative lack of detailed sequential analysis (at least that we can report on here), it is nevertheless clear that the standard accounts for the role of gaze need modifying either in general or at least with regard to question sequences in particular. Whichever account is preferred, it is clear that our understanding of the systematicity of gaze behavior in interaction has been affected. The most notable piece of news is probably the fact that Q-speaker gaze behavior seems to be much more similar across languages than Q-recipient gaze behavior. In this sense, ethnographic descriptions that describe lack of eye contact and very brief glances could well be partially correct but perhaps only as descriptions of typical recipient behavior. The difference in gaze behavior observed between Tzeltal , Italian, and Yélî Dnye strongly challenges the view that looking at the speaker’s face would always be the default gaze behavior for a recipient. Nonetheless, the nonuse of gaze as a reliable indicator of recipiency in Tzeltal can be shown to require some substitute system, which in Tzeltal seems to be
240
Federico Rossano et al.
provided by the repetition response system (Brown and L evinson 2005; Levinson and Brown 2004). This study therefore confirms that generic conversational functions reliably invoke machinery that will handle them ( Schegloff 2006). There are, however, a series of open questions that emerge from our first empirical pass through the data. What are the functions of speaker gaze in dyadic interaction? How is recipiency displayed and are recipiency cues somehow systematically organized and ranked? How do we account for a system in which Q-recipients look less than Q-speakers and yet in more than 40 percent of the cases? Is there something special about “questions” that affects gaze deployment in a very specific way? It is clear that much more fine-grained analysis of gaze behavior in interaction is needed. Action types, at least at the level of the description that we chose, do not really account for the differences in recipient gaze behavior. It seems reasonable to invoke as an account the material world the participants are inhabiting, but we feel obliged to emphasize that Tzeltal participants are similar to the Italian when they speak but not when they listen to questions. The Tzeltal participants are inhabiting the same environment throughout the interaction and yet behave differently according to their participation role . We are aware of the risks involved in isolating a specific practice (gaze in interaction) and trying to understand how this changes the interactional environment. However, the fact that gaze toward addressee is deployed very often during questions but not “all the time” or “never” in all three cultures here examined (and in all cultures examined in previous studies) suggests that its occurrence is systematic and interactionally managed and probably affects the interaction in significant and specific ways. As the quotation by Lord Chesterfield at the beginning of this chapter suggests, gaze in interaction is a potent thing. And, as Darwin insisted, it has deep ethological roots and so is inevitably deployed in interaction in all cultures as a valenced indicator of visual attention – although, as we have shown, it is not always deployed in the same ways. Yet gaze in interaction has received relatively little attention in the past twenty years. We hope this chapter will encourage further investigations of how gaze is deployed
Gaze, questioning, and culture
241
in different settings, different cultures, different participant roles, and in performing different actions.
Appendix A: Coding For numerical purposes, we developed the following coding scheme. For the verbal component, we used the following five categories (bear in mind that all the interactions were dyadic, so the effective recipient is by default the nonspeaker). A. B. C.
D. E.
SPEAKER/RECIPIENT: who is the speaker of each question. QUESTION TYPE: “yes”/“no” (polar) question, wh-question or alternative question. ACTION: e.g., request for information, request for confirmation , other initiation of repair, offer, invitation, seeking agreement , etc. RESPONSE: answer (interactionally appropriate), nonanswer response (e.g., “I don’t know, maybe”), no response.33 SEQUENTIAL POSITION: initial position or further in the sequence (only in Italian ).
We coded for thirteen different possible gaze configurations: 1. (a) The speaker does not look at the recipient at all during the question. (b) The recipient does not look at the speaker at all during the question. 2 (a) Speaker looks at recipient from beginning to end and was already looking before the beginning of the question. (b) Recipient looks at speaker from beginning to end, and was already looking before the beginning of the question. 3. (a) Speaker is looking at recipient before the beginning of the question but looks away during the question. (b) Recipient is looking at speaker before the beginning of the question but looks away during the question. 4. (a) Speaker looks toward recipient during the introduction of a new referent in the question. (b) Recipient looks toward speaker during the introduction of a new referent in the question.
242
Federico Rossano et al.
5. (a) Speaker looks toward recipient on nearing completion of the question (last two words or last two syllables if question made of two words). (b) Recipient looks toward speaker on nearing completion of the question (last two words or last two syllables if question made of two words). 6. (a) Speaker looks toward recipient in any other position during the question. (b) Recipient looks toward speaker in any other position during the question. 7. The occurrence of mutual gaze at any time during the question turn. (We define mutual gaze as both participants looking at each other’s face or eyes simultaneously.) Finally, we coded for how often the lack of speaker or recipient gaze was motivated by: 1. the content of the question and its delivery (e.g., presence of deictic s, pointing, talking about an object present in the surrounding environment); 2. a competing activity such as eating, weaving, drawing on the sand, etc. Appendix B: Symbols for gaze orientation A A
A
A
B
: Mutual gaze.
B
B
: A looks away and B looks away. : A looks down oriented toward B. B looks away.
B
A
B
A
B
A
B
A
B
: A looks away. B looks down oriented toward A. : A and B are looking down in front of them. : A looks at B. B looks down. : A looks at B. B looks away.
: B looks at A. A looks down.
Gaze, questioning, and culture A
B
A
B
A
B
A
B
A
B
A
B
A
: B looks at A. A looks away.
: A looks down. B eyes closed.
: A eyes closed. B looks down.
: A eyes closed. B eyes closed.
: A looks at B. B eyes closed.
: A eyes closed. B looks at A.
B
A
B
A
B
A
B
A
B
: A away. B eyes closed. : A starts turning toward B who is looking down. : A starts turning toward B who is already looking at A.
B
A
B B
A
B
A
B
A
B
A
B
A
B
A
: A starts raising the gaze toward B who is looking down.
: A starts raising the gaze toward B who is already looking at A.
A
A
243
: A looks away B looks mid-distance up left. : A looks away B looks mid-distance up right. : A looks down B looks mid-distance up left.
: A looks down B looks mid-distance up right.
: B starts turning toward A who is looking down. : A starts raising the gaze toward B who is looking down.
: A looks mid-distance up left. B is looking away.
: A looks mid-distance up left. B is looking down.
B
: A looks mid-distance up right. B is looking away.
244
Federico Rossano et al.
A
B
A
B
A
B
: A starts turning toward B who is looking away.
A
B
A
B
A
B
A
B
A
B
: B starts turning toward A who is looking away. : B starts raising the gaze toward A who is already looking at B.
: B starts raising the gaze toward A who is looking down.
: A looking at B. B looks mid-distance up right.
B
A
B
B
A
B
A
B
A
B
A
B
: B starts turning toward A who is already looking at B. : B starts raising the gaze toward A who is looking away.
A
A
: A looks mid-distance up right. B is looking down.
: A mid-distance right. B eyes closed. : A mid-distance right. B looking at A.
: A mid-distance left. B looking at A.
: A looking at B. B mid-distance right. : A looking at B. B mid-distance left.
: A eyes closed. B looks away. : A looking mid-distance up left. B looking up.
Notes 1 Cited in Goodwin (1981: 62), after Goffman (1953). 2 “Participants utilize both their bodies and a variety of vocal phenomena to show each other the type of attention they are giving to the events of the moment, and, reciprocally, the type of orientation they expect from others. [ . . . ][Engagement displays] permit those present to display
Gaze, questioning, and culture
245
to each other not just speakership and hearership but differentiated attention to, and participation in, the talk of the moment” (Goodwin 1981: 124–125). 3 “To extract robust outcome-based conclusions about how physicians (or patients) should conduct themselves in specific moments in the flow of the medical encounter, it is important to find a meeting point between the two methodologies of coding and microanalysis [ . . . ]. In other words, beyond the intrinsic worth of an analytical framework responsive to very granular, individual moments in the physician– patient encounter, we need one that simultaneously supports coding at a broader level of granularity sufficient to reach beyond individual cases to generate findings at a statistical evidential standard” (Heritage and Maynard 2006: 8). 4 “Children at a very early age do not blush; nor do they show those other signs of self-consciousness which generally accompany blushing; and it is one of their chief charms that they think nothing about what others think of them. At this early age they will stare at a stranger with a fixed gaze and unblinking eyes, as on an inanimate object, in a manner which we elders cannot imitate” (Darwin 1872: 328). 5 Duncan and colleagues actually refer to shift in head direction as a turn-yielding cue but specify that this should be taken as a proxy for “eye direction” (Duncan and Fiske 1977: 211). 6 In a direct response to Beattie’s paper, Kendon (1978) argues that the data used by Beattie (dyadic conversations between students and their supervisors) was not comparable to the one used in his study (ordinary dyadic conversation between Oxford undergraduates), because of the asymmetric status of the participants and the formality of the situation, further indicating that the kind of interactional situation participants are dealing with may well affect the deployment of gaze. 7 There are, however, works (e.g., Erickson 1979; LaFrance 1974; LaFrance and Mayo 1976) that claim racial differences in this respect: Black Americans look more while speaking than while listening, while white Americans follow the opposite pattern. 8 In a footnote, (Goodwin 1981: 57) admits that even though he proposes this as a rule applying to turns in general, this pattern is not found in every turn at talk. 9 With the terms “courses of action accomplished through one or more sequences of talk,” Rossano refers to the fact that, to be considered completed, most actions require at least the occurrence of some sort of response or reaction by the other participant, and, therefore, an initiating action usually starts the development of a course of action produced by more than one participant. For example, the gist of a request
246
10
11 12
13
14
Federico Rossano et al.
for information can be considered accomplished only if the other participant provides the information, and an offer is completed only if it is accepted or rejected and the thing offered is provided to the person to whom it had been offered. This means that the occurrence of an utterance and the action(s) that it implements opens the possibility (and sometimes it normatively expects) the occurrence of another set of utterances or actions that would allow the gist of the initiating action to reach a socially appropriate completion. This can be achieved in two turns or may require larger structures, though the participants’ orientation toward completing the gist of the initial action remains the same. Kendon is one of the few who distinguish between speaker gaze during “short questions” and speaker gaze during “long utterances”; he suggests that during short questions “[the speaker] will look steadily at [the recipient] while he asks his question, and where [the speaker] is asking a series of questions, unless he has to pause in thinking of the question, or unless he has to pause in formulating it.” (Kendon 1967: 47). Ten interactions in Italian, ten in Tzeltal, and nine in Yélî Dnye. In some interactions in Tzeltal and Yélî Dnye, some children and/or other adults can be seen in the videos that capture the interactions. The interaction remains dyadic for current purposes as these bystanders are not addressed, they never intervene in the conversation, nor they are looked at. As we are not interested in the content of their conversation but rather in the gaze behavior of the participants, the presence of neglected bystanders does not have any significant impact on the gaze behavior of the dyad. The Italian speakers were all natives of the region Emilia-Romagna, in the north of Italy, and all the Italian data were recorded in Bologna, Italy. The Tzeltal speakers all belong to the ethnic group of Tenejapans (c. 25,000 strong), and all were recorded in situ within their territory. The Yélî Dnye speakers are the exclusive inhabitants of Rossel Island, and form a small population of 4,000 souls, nevertheless divided into two dialect groups – the recordings are wholly of eastern dialect speakers. From each culture we had participants of both sexes; ages ranged from twenty to fifty years old for Italian and Yélî Dnye speakers and from thirty to sixty years old for Tzeltal speakers. The three authors are either native speakers (Rossano ) or long-term fieldworkers (Brown , Levinson ) in their respective communities reported on here. Some of these could be understood as noticings, but often they address already-raised issues so are functioning as requests for confirmation or prods for expansion.
Gaze, questioning, and culture
247
15 Rossano coded the linguistic/interactional aspects of questions for Italian , Brown for Tzeltal and Levinson for Yélî Dnye . Rossano coded for gaze the Italian and Yélî Dnye data (as well as 40 percent of the Tzeltal data, checking consistency of coding), while Brown coded for gaze the Tzeltal data. 16 Logistic regression is a statistical model used to predict outcome variables that are categorical (e.g., yes or no). In this specific case, if a specific participation role (e.g., being a speaker) predicts significantly better than the other (being a recipient) whether an individual will be looking at the other participant, then we can say that being a speaker rather than a recipient significantly affects the likelihood of the occurrence of gaze toward the other participant. In other words, participants look at each other differently depending on which participation role they have during a question. We are grateful to Tanya Stivers for help with the statistical analyses. 17 From now on, all logistic regression analyses reported here have to be considered corrected for the clustering of the data by interaction. This is meant to take into account the fact that different interactions (and in particular what the participants are doing within a specific interaction) could affect the general pattern more than others and therefore the general distribution could be the product of one or two biased interactions rather than a systematic pattern observable across multiple interactions. 18 Interestingly, in a small corpus of questions in not experimentally elicited dyadic English interactions, Beattie (1978: 13) reports that the speaker gazes at the addressee at least at some point during the question in thirty out of thirty-nine questions, which corresponds to 76.9 percent of the questions. Similarly, Kendon (1967: 45) reports that speakers look toward addressees in 31/41 (75.6 percent) of the questions of his corpus of British English dyadic interactions. While both are very small corpora, probably not fully representative of dyadic interactions in British English, it is remarkable that the percentage of speaker gaze during questions would be in between the ones found for Italian and Yélî Dnye, which are not significantly different from each other. 19 These percentages can be obtained by subtracting the instances of mutual gaze from the instances of speaker gaze. 20 Most of the questions that receive no response are directly followed by the same speaker elaborating or clarifying the question and therefore still pursuing a response. The relevance of a response is still present, and interactional contingencies can account for the lack of responses (Rossano in press).
248
Federico Rossano et al.
21 For Tzeltal the recipient is not gazing in 57.7 percent of the 300 questions here considered, in Italian in 36.7 percent, and in Yélî Dnye in 32.7 percent. 22 There is no doubt that in dyadic conversations there are instances in which gaze direction can vary within one utterance, and this is usually related to the activities at hand and what is going on in the interaction (moreover, specific actions can require a very specific gaze deployment). What we are saying is that during questions in these three cultures we do not see much gaze mobility, and, therefore, there is at least one domain and three cultures in which prior general claims about gaze behavior in any turn at talk do not hold. 23 We should emphasize here once again that there are a number of distinct social groups who speak Tzeltal, with distinct customs; our generalizations here are about Tenejapans. 24 This matches previous claims by Brown and L evinson (2005; Levinson and Brown 2004). However, the fact that alternative sitting configurations are observable within the same samples as well as the rather small number of dyadic interactions for each culture analyzed here suggest caution in terms of generalizations. 25 It is not always clear whether participants have chosen one configuration because of a general preference for that seating configuration or because of external factors such as presence of benches, preference for sitting in the shade, surrounding noise, etc. 26 See for a comparison the following example in American English, recorded in a medical interaction (Boyd and Heritage 2006: 158): 01 Doc: Is your mother alive, 02 Pat: No: .
27 This repetition -as-affirmation default has been shown to apply in a number of languages, not only within the Mayan family but also in completely unrelated languages (e.g., Irish, Welsh, Finnish, Estonian). 28 See row 3 of Tables 7.16 and 7.17 for instances in which a participant is looking at the addressee before the beginning of the question and looks away before completion of it. It never happens more than 6–7 percent of the time in any of the three languages, and in some configurations (e.g., speaker gaze in Italian) it is even rarer. 29 Moreover, the fact that there is a sound stretch, a cut-off or a pause in the question does not mean per se that they have been produced to elicit gaze back, as issues such as self-initiation of repair could be in play here. 30 There are a few cases in our data where a self-repair may play a similar role to that in American English.
Gaze, questioning, and culture
249
31 As expected, most of them were requests for information (44/59, 74.6 percent) with the remaining ones performing different actions such as offering or doing a topic proffer, but clearly not initiating repair. 32 A zoonosis is a disease that can be transmitted to humans from other vertebrate animals. 33 We coded as responses both visible and verbal responses. A head nod or a head shake in response to a polar question was coded as an answer, exactly like pointing to an object on the table in response to a question such as “where did I put my notes?” Shoulder shrugs or similar gestures whose meaning could not be clearly established were instead coded as non-answer responses. “No response” means that no verbal or visible response was provided.
8 Negotiating boundaries in talk Makoto Hayashi and Kyung-eun Yoon
Introduction Across languages, we find minimal vocalizations, such as “mhm” in English and “un” in Japanese , that are used by participants in interaction to respond to co-participants’ prior or ongoing talk. These “response tokens” have drawn a sustained interest in the study of language and social interaction over the past several decades (see Sorjonen 2001 for a useful review). One major line of research in this area has been concerned with distributional analysis, exploring statistical correlations between types of response tokens and turn-taking (e.g., Duncan 1972, 1974; Duncan and Fiske 1977) or correlations between the frequency of response tokens and such parameters as culture and gender (e.g., Maynard 1986, 1990; White 1989; Fishman 1983). Because their primary interest is in discerning patterns for the aggregate occurrences of response tokens across situations, these quantitative studies tend to isolate response tokens from their local contexts of use, glossing the interactional contingencies to which participants orient when producing response tokens in any given context. By contrast, a line of research on response tokens that emerged within conversation analysis (CA) pays more serious attention to the sequential/interactional contexts for the deployment of response tokens (e.g., Jefferson 1984, 1993; Schegloff 1982; Heritage 1984, 1998, 2002; Goodwin 1986; Mazeland 1990; Beach 1993; Sorjonen 1996, 2001, 2002; Gardner 2001; Mori 2006; among others). These studies investigate the kinds of interactional activities the participants engage in that give rise to the deployment of particular response tokens, how the co-participants orient to them, and what consequences their deployment has for the organization of
Negotiating boundaries in talk
251
subsequent talk. This research has shown that the “meanings” of response tokens depend crucially on their placement within the ongoing sequence of actions and activity, and that it is therefore necessary to examine the details of the sequential/activity contexts as part of the “semantics” of response tokens. In the present study, we explore the situated meanings of minimal response tokens in Japanese and Korean (“ un” and its variants and “um” and its variants, respectively; both roughly equivalent to “yeah” or “mhm” in English) produced in a specific sequential context – i.e. in a slot after one’s talk has been responded to by a recipient with a minimal “yeah”/“mhm”-like acknowledgment token.1 To describe this sequential context schematically: →
A: B: A:
turn’s talk minimal yeah/mhm-like acknowledgment token minimal yeah/mhm-like acknowledgment token
Position 1 Position 2 Position 3
The practice to be examined is thus characterized as the “thirdposition deployment” of a minimal response token. 2 It should be noted that the minimal response tokens in Position 3 are produced by the primary speaker-so-far (i.e. the speaker of substantive talk in Position 1) rather than the recipient-so-far. Previous studies have paid almost exclusive attention to response tokens as “recipient activities,” and little has been investigated about primary speakers’ deployment of response tokens. Our study thus aims to shed light on the roles of response tokens as primary speakers’ activities in organizing the unfolding courses of talk in interaction. This study is also intended as a contribution to cross-linguistic analysis of response tokens (cf. Beach and Lindström 1992). Most of the past studies focused either on analyzing multiple functions of a single response form in a given language (e.g., Heritage 1984; Beach 1993; Sorjonen 2002; Mori 2006) or on comparing functions of different response forms within a single language (Jefferson 1984, 1993; Goodwin 1986; Gardner 2001; Sorjonen 1996, 2001). Our study, by contrast, examines comparable response tokens used in comparable sequential/activity contexts in two different languages. Through this examination, we demonstrate how the practice of deploying third-position minimal response tokens provides a resource for speakers of both Japanese and Korean to accomplish similar interactional ends under similar interactional contingencies.
252
Makoto Hayashi and Kyung-eun Yoon
In what follows, we analyze two related yet distinct types of usage of third-position minimal response observed in Japanese and Korean conversations. We will show that in both types of usage, the deployment of third-position minimal response tokens is “closure implicative”; that is, it proposes and projects termination of what has been going on prior to it. In the first type of usage we examine, it is the topic /activity-so-far that is projected to be terminated. In other words, a speaker-so-far deploys third-position minimal response to indicate that the topic talk or activity in which the participants have been engaged so far is now put into a “state of attrition” (Jefferson 1993) and that a new topic or activity can be relevantly initiated. A third-position minimal response is thus used as a resource to negotiate topic/activity boundaries and achieve topic/activity transition. In the second type of usage, it is the turn-so-far that is projected to be terminated. That is, third-position minimal response is used as a “turn-exit device” (Jefferson 1981). As will be discussed below, speakers of Japanese and Korean sometimes produce syntactically incomplete turns that may nonetheless be regarded as “pragmatically complete” (i.e. complete from the perspective of executing an action). Such turns can at times cause a problem for turn transfer because recipients may not treat them as complete turns due to their syntactic incompleteness, while the speaker-sofar may be ready to give up the floor. Thus, when a syntactically incomplete, yet pragmatically complete turn is produced and is only met with minimal response from the recipient, the speaker may deploy third-position minimal response to indicate that his or her turn is over now and that it is the recipient’s turn to do some extended talk. Negotiating topic /activity boundaries The first type of usage of third-position minimal response examined here is its deployment in the context of negotiating topic / activity boundaries.3 As previous studies have shown, bringing an ongoing sequence of topic talk or activity to closure is an interactive achievement – an achievement that commonly involves turn-by-turn negotiation among the parties, rather than unilateral execution by a single speaker (Schegloff and Sacks 1973; Maynard
Negotiating boundaries in talk
253
1980; Jefferson 1984, 1993; Drew and Holt 1998). Deploying third-position minimal response, then, is one method that is available for the participants to propose the relevance of imminent topic/activity closure. By responding to a co-participant’s minimal acknowledgment of the prior talk with another minimal acknowledgment token, the speaker-so-far conveys that there is nothing more to add to the prior topic/activity and therefore that it is an apt moment for topic/activity transition to take place. The following fragment presents an example from a Japanese conversation. This fragment is taken from a phone conversation between two adult friends who used to be on the same track and field team in high school. Prior to this segment, one of the participants, Mika, announced her recent visit to a stadium (called Hattori Stadium) where they often had district track meets when they were in high school. This announcement touches off a reminiscing activity (lines 1–5), which then leads to Eiji’s “my-side telling” (from line 7 onward) about his visits to stadiums they used to compete in. In this telling, he mentions that, though he sometimes visits another stadium (called Nagai Stadium), he has not visited Hattori Stadium for quite a while. Examine in the following fragment how Eiji brings his “my-side telling” to closure with a thirdposition minimal token in line 36. (1) [Japanese: K MI 9 (telephone:Eiji and Mika)] 01 E: sabu torakku mitaina n mo::natsukashii na:::. sub track like N also nostalgic FP I feel nostalgic about the practice track, too. 02 °.hhh° (.) demo anmari::hh anm(h)ar(h)i but not.much not.much °.hhh° (.) But I don’t::hh I don’t 03
ob(h)oet(h)enai na:: [.hhhhhhh] u:::[::n. hhh] don’t.remember FP yeah remember m(h)uch::. .hhhhhhh Yeah::::. hhh
04 M: 05
[huhuhuhu] huhuhuhu
[.hhh moo] already .hhh It’s
haruka mukashi no ohanashi desu ka[ra ne:::] by.much long.ago LK story CP because FP already such a long time ago, so:::
254
Makoto Hayashi and Kyung-eun Yoon
06 E: 07
[ soo desu] so CP That’s
ne:::. nanka:NAGAI::::no hoo WA:: FP like Nagai LK side TP right. As for Nagai::::,
08 M: u:n Mhm 09 E: ano:: yoko ni amehuto joo ga dekite::: uhm next.to PT American. field SP is.built football uhm, they built an American football staduim right next to it, so … : :((14 lines omitted in which Eiji tells Mika :that he visits Nagai Stadium every so often :because he goes to American football games :played in the new stadium next to it.)) : 24 E: . . . nagai no kyoogijoo o [ne::] Nagai LK stadium O FP 25 M:
[u:::]n. Mhm
26
(.)
27 E:
nagameteta n desu ke[do.] was.looking N CP but ((Lines 24 and 27)) So I was looking at Nagai Stadium.
28 M: 29
no hun ni mamireta [( LK dropping by stained stained by pigeons- pigeons’ droppings
30 E: 31
[ano] hato ni:- (.) hato that pigeon by pigeon That ((filthy stadium)) )]
[s(h)oo HHsoo soo soo soo] R(h)ight HHright right
soo. right right right.
32 E: .hhhhhh demo ne:nanka ano::::acchi wa:: but FP like uhm there TP .hhhhhh But uh::::m, the other one,
Negotiating boundaries in talk
255
33
hattori no hoo wa moo::: zen:::zen:: Hattori LK side TP EMPH not.at.all Hattori Stadium, I don’t go there at all
34
ja nai ka na:::.= CP NEG Q FP anymore, I guess.
– Position 1
35 M:
=u::n. Mhm
– Position 2
36 E: →
u::n. Mhm
– Position 3
37
(0.2)5
.hhhh soo ka:::::iya ii desu ne::::: 38 E: so Q well good CP FP .hhhh I see. Well, it’s good ((that you went to Hattori))
After explaining why he still sometimes goes to Nagai (lines 7–27, most of which is not included here), Eiji brings the focus of his talk back to Hattori Stadium and states that he does not go there at all (lines 32–4). Eiji’s utterance here connects back beyond the adjacently prior talk/details (i.e. talk about his visits to Nagai) to an earlier part of the ongoing talk (cf. lines 1–3). It also presents an upshot of his “my-side telling” and its connection to Mika’s original announcement about her recent visit to Hattori; that is, that in contrast to his occasional visit to Nagai, he has not visited Hattori for quite a while. Utterances of this kind, connecting back to an earlier part of an ongoing telling and presenting an upshot of the telling-so-far, often serve to indicate that the teller is becoming disengaged from an item-by-item development of topic talk, and they are therefore commonly associated with topic closure (Drew and Holt 1998). In response to this closure-relevant statement, Mika produces a minimal acknowledgment token, “u::n”/“mhm” (line 35), and continues to display her alignment as a recipient to Eiji’s telling. Eiji then responds to Mika’s display of “passive recipiency” (Jefferson 1984) with a minimal response token of his own (line 36), and, by doing so, indicates that he is ready to disengage from the current line of “my-side telling” about his visits to old stadiums. Indeed, after a brief silence (line 37), Eiji makes a transition from “my-side” talk to “your-side” talk by producing a response to (and a positive assessment about) Mika’s original announcement of her recent visit to Hattori Stadium.
256
Makoto Hayashi and Kyung-eun Yoon
The next fragment from a Korean conversation shows a case in which a parenthetical activity embedded within another activity is brought to closure with third-position minimal response. Prior to the following segment, Gina mentioned a tasty noodle product, and Sumi asked for its brand name. Just as Sumi brings out a paper to write down the name of the product, Gina begins to voice a concern about its availability in stores because it is a seasonal product. Examine how this parenthetical activity of “(potential) problem announcement” is brought to closure with a third-position minimal response in line 13 before the participants resume the activity of giving/writing down the product name. (2) (Korean : Lunch Talk [face-to-face: Sumi and Gina]) 01 S:
[((brings out a paper))
02 G: [yocumey kacyeo-nunci moll -a.= these.days bring -if not.know-INT I am not sure if they bring the noodles ((from the warehouse)) these days. 03 S:
=[ ney:.] Mhm.
04 G:
=[waynya]myen yelum -ey ha:ncham nemwu manhi because summer-in a.long.time very a.lot because they bring a lot of those noodles
05
kacyeo-nuntey, bring -CIRCUM all summer long, but
06 S:
u u:m. m Mhm.
07 G:
kyewul-i[-nikka.] winter-be-CONN because it’s winter now.
08 S:
[ u :m ] Mhm.
09 G:
han pen tto- (0.3) one time again Once again- (0.3)
10 S:
nacwungey (ye- pom -ina yelum later sum- spring-or summer Later (when sum- spring or summer
Negotiating boundaries in talk
257
11
toy -myen) become-when comes,)
– Position 1
12 G:
°um° °Mhm°
– Position 2
13 S: → ey. Mhm 14
– Position 3
(1.2) ((Sumi gets ready to take a memo.))
tongchimi mwul-nayngmyen-i –ntey -yo, pwukkyeng 15 G: Tongchimi mwul-nayngmyen-be-CIRCUM-POL Peking It’s Tongchimi Mwul Nayngmyen, but I guess it’s 16
tongchimi mwul-nayngmyen tongchimi mwul-nayngmyen Peking Tongchimi Mwul Nayngmyen.”
In lines 2, 4–5, and 7, Gina develops a line of talk in which she suggests a potential problem with the product’s availability in stores in the season in which this conversation takes place. This “problem announcement” is responded to by Sumi in lines 10–11 with a suggestion of a “problem solution,” i.e. that one only needs to wait a few months to see the product available again in stores.4 In the context of discussing a problem, presenting a solution can be implicative of the closure of that discussion. Gina responds to this statement of problem solution with “um”/“mhm” produced in soft voice (line 12) and displays minimal acknowledgment/agreement . To this, Sumi produces a third-position minimal response token of her own (line 13) and displays her stance that she is not going to pursue the current line of talk about the problem of the noodle product’s availability any further.5 This proposal of topic / activity closure is further reinforced by Sumi’s nonvocal conduct in the ensuing silence (line 14), in which she moves to prepare to write down the name of the product and proposes the relevance of resuming the activity that the participants were about to engage in before the parenthetical sequence began. Gina concurs with Sumi’s proposal by moving on to mention the brand-name of the noodle product (lines 15–16). Thus, just as in the previous instance, thirdposition minimal response is used in the juncture where topic/ activity closure and transition are jointly accomplished. The two fragments we examined demonstrated that thirdposition minimal response can be used by speakers of Japanese
258
Makoto Hayashi and Kyung-eun Yoon
and Korean as a means to effect imminent closure of the preceding line of talk and transition to a new topic /activity. In both cases, we observed the following sequence of actions around the deployment of third-position minimal tokens. First, the utterance produced in what is labeled as Position 1 above presents some kind of “closurerelevant” statement, such as a reference to what was said earlier, an upshot of what has been discussed, etc. This is receipted by a co‑participant with a minimal acknowledgment in Position 2, which registers no more than minimal receipt of (and/or minimal agreement with) the prior talk. This, then, is responded to by a reciprocal minimal response produced by the speaker of the utterance in Position 1, which serves to decline to take the opportunity in that slot to say anything more about the previous topic/activity. This exchange of reciprocal minimal response tokens displays both speakers’ orientation to withholding any further topical development or elaboration and thereby paves the way for a transition to a next topic/activity. Schematically: Position 1 Position 2
A: Closure-relevant statement B: Minimal response (registers minimal receipt/ agreement) Position 3 A: Minimal response (declines further development of topic/activity) Position 4 B/A: New topic/activity initiation
We should note here that this sequence of actions at the juncture of topic /activity transition is not exclusively observed in Japanese and Korean conversations. For instance, Drew and Holt (1998), who examine the use of figurative expressions in English conversation, describe a similar course of actions observed at topical junctures (which they term the “topic transition sequence”) and explore how figurative expressions regularly occur as topical summaries in the slot equivalent to Position 1 in the schema above (see Extract 3). Jefferson (1993) also describes how a summative commentary is followed by reciprocal exchanges of minimal responses that put the topic talk-so-far into the “state of attrition” and lead to topic shift (see Extract 4); see also Schegloff and Sacks 1973: 318). (3) (English: Drew and Holt 1998: 501–2; slightly modified) 01 R: B’t take this with a dollop’v salt you kno::w I’m02 I’m ba:sic’ly quite happy b’t quite relieved it’s
Negotiating boundaries in talk
259
03 the sheer organization ‘n getting all, everything – Position 1 04 done in th’ da:y. – Position 2 05 L: Yes:that’s ri:ght,= – Position 3 06 R: =Ye[s. 07 L: [Yes. Ye[h .tch .hhhhhh Wuh[O k a y, W’ddiyou wanna talk t’ me 08 R: – Position 4 09 abou(h)t (4) (English: Jefferson 1993: 20; slightly modified) 01 M: .hh Well you never kno:w do you someti:mes 02 yo[u feel as if you don’t want to stay in the= 03 J: [No:. 04 M: =sa:me pla[:ce, .hh tha[t where you’ve been= [(pla:ce.) [Ye:s. 05 J: – Position 1 06 M: =with your pa:ren[ts:.hh [Ye:s. – Position 2 07 J: 08 (.) – Position 3 09 M: Mm[:. .hh [But uh::anyway, 10 J: 11 (0.3) 12 M: .mptlk By the wa:y Janet did you get . . . – Position 4 13
Thus, the practice described in this section is by no means unique to Japanese and Korean . There is, however, a potentially distinctive feature in the way third-position minimal response can operate in Japanese and Korean that has not been observed in English. That is, at least in some instances of third-position minimal response in Japanese and Korean, a speaker uses the minimal token to close down the topic talk/activity-so-far even though his or her immediately prior turn is still grammatical ly incomplete. An example of this is found in Extract 2, examined above, which is partially reproduced in Extract 5 below. (5) (Korean : Lunch Talk [face-to-face: Sumi and Gina]) 10 S:
nacwungey (ye- pom -ina yelum later sum- spring-or summer Later (when sum- spring or summer
11
toy -myen) become-when comes,)
– Position 1
12 G:
°um° °Mhm°
– Position 2
260
Makoto Hayashi and Kyung-eun Yoon
13 S: → ey. Mhm 14
– Position 3
(1.2) ((Sumi gets ready to take a memo.))
15 G: tongchimi mwul-nayngmyen-i –ntey -yo, … Tongchimi mwul-nayngmyen-be-CIRCUM-POL It’s Tongchimi Mwul Nayngmyen, but …
In her “problem solution” statement in lines 10–11, Sumi produces a subordinate clause that ends with the subordinate-clause marker “-myen”/“when.” In more interactional terms, Sumi’s utterance here is the “preliminary component” of a “compound turn constructional unit (TCU)” (Lerner 1991) in the form of [X-myen + Y] ([When/If X + then Y]).6 At the component boundary within this compound TCU, Gina provides a weak acknowledgment/agreement (line 12). Rather than continuing to produce the main clause, or the “final component” of the compound TCU, Sumi then produces a third-position minimal token (line 13) and indicates that she is not going to go on with the turn-so-far even though it has not been brought to syntactic completion.7 Thus, third-position minimal response can be used not only as a device to exit from the topic /activity-so-far, but also from the turn-so-far. While this feature is not observed in all cases of thirdposition minimal response in Japanese and Korean , as evidenced by Extract (1) above, it is observed regularly enough to merit analytic attention in its own right. We thus turn now to the usage of third-position minimal response as a means to exit from a notyet‑complete turn. Negotiating turn closure This section examines how a third-position minimal response token is used as a “turn-exit device,” i.e. a device to propose the closure-relevance of the turn-so-far by conveying that the speaker is not going to add anything more to the current turn. The deployment of this practice is often observed when speakers of Japanese and Korean produce utterances that are incomplete from a syntactic point of view, yet which may nonetheless be regarded as complete from the perspective of executing an action. Such “syntactically incomplete yet pragmatically possibly complete turns”
Negotiating boundaries in talk
261
can sometimes engender a problem in coordinating speakership transfer because recipients may not treat them as complete turns due to their syntactic incompleteness, while the speaker-so-far may regard his or her turns to be complete within the particulars of the given contexts in which they are produced. When this type of incongruence of interpretation arises, the speaker-so-far may use a third-position minimal response token to mark the closure of his or her turn and propose the relevance of speakership transfer. Now, what motivates speakers to leave their utterances syntactically incomplete in the first place? In our database, we observe several specifiable environments in which “syntactically incomplete yet pragmatically possibly complete turns” are recurrently observed. One is where a speaker regards the rest of his or her utterance to be easily inferable from the context and thus avoids stating the obvious. An example is seen in Extracts 2 (and 5) above, in which the main clause or the “final component” that would follow Sumi’s utterance in lines 10–11, “nacwungey ye- pom-ina yelum toy-myen”/“Later when sum- spring or summer comes,” is easily inferable by Gina, given that it alludes to what Gina herself stated in the immediately prior talk, i.e. that people at grocery stores lay in a stock of the noodle product in the summer (lines 4–5). Thus, in terms of executing the action of “stating a solution to a possible problem with obtaining the noodle product,” Sumi’s utterance in lines 10–11 can be regarded as complete for all practical purposes. Sumi then deploys a third-position minimal response token to exit from the turn and move on to resume the activity that has been put on hold. One common environment in which stating the obvious is avoided is where the speaker is voicing something interactionally delicate, such as a disagreement or bad news. In this type of environment, third-position minimal response can be used as a way to exit from a turn while leaving the “core” of the delicate matter unsaid (and inviting the co-participants to infer it). The following fragment from a Japanese conversation provides a case in point. In this fragment, Seiji, who often takes a sauna at a public bathhouse, is telling the recipients, Akira and Harumi, about the kind of conversation he would have with other people while taking a sauna. He claims that, because people want to make sure they can get out when it gets too hot, they would deliberately choose “safe” topics
262
Makoto Hayashi and Kyung-eun Yoon
such as weather, etc., on which they would not disagree with one another or get into an argument (lines 1–2, 4, 7–8). Examine how Akira responds to Seiji’s telling (lines 10–12) and how Seiji in turn responds to Akira’s response (lines 14–15). (6) (Japanese : RKK 24 [face-to-face: Seiji, Akira, and Harumi]) 01 S:
(yoo suru ni) iya sonna koto nai desu yo::: in.short no such thing NEG CP FP
02
toka yuu yoona [koto yuu] to::giron ni= QT say like thing say if argument PT
03 A: 04
[a: : : : : : : .] Oh: : : : : : .
=natteshimatte::= become
((Lines 01–02, 04)) In short, if you say something like, “I don’t think so,” you get into an argument and, 05 H:
=u:(h):(h):(h):(h):n.= Y(h)e(h)a(h)h:::.
06 A:
=[demo::] but But
07 S:
=[nanka:]:moo atsui detai to omotta like already hot want.to.get.out QT think Like, when you are hot and want to get out,
08
toki ni [derenai [heh heh heh heh .hhhhhhhh when PT can’t.get.out you can’t. heh heh heh heh .hhhhhhhhh
09 H:
[heh heh heh heh heh
10 A: [shotaimen no 1st.encounter LK I guess it’s more or 11
kaiwa tte daitai sonnan janai conversation QT more.or.less like.that CP:NEG less like that whenever you talk to people who
12
ka na::. FP FP you meet for the first time.
13
(0.7)
Negotiating boundaries in talk
263
14 S: u::::n .hh demo hora:: warito:s- (.) sauna nakama but y’know sort.of sauna friend Hmm:::.hh But y’know, since people become, like, 15
mitaini nacchau kara sa::, like become because FP “sauna buddies,”
16 H:
hhehh=
17 A:
=u:::n. ((with a slight head nod)) Mhm
18 S: → u:::n. Mhm
((with a slight head nod))
19
(0.4)
20 H:
(nani) sauna suki na no what sauna like CP FP (What), do you like sauna?
– Position 1
– Position 2 – Position 3
In response to Seiji’s characterization of “conversations in a sauna,” Akira comments that the same holds more or less true whenever people meet for the first time and have a conversation, whether in a sauna or not (line 10–12). As it points out the potential generalizability of Seiji’s characterization to just about any conversation among people meeting for the first time, Akira’s comment can undermine the very point that Seiji is trying to make, i.e. that the choice of conversational topics in a sauna is intimately related to the external constraints unique to the setting of sauna-taking. In lines 14–15, then, Seiji responds by stating that the people he converses with in a sauna are his “sauna nakama”/“sauna buddies” that he sees repeatedly, thereby suggesting that what Akira said does not apply to the conversations that he is describing. Note that Seiji’s utterance in lines 14–15 is constructed as a clause marked with “kara”/“because,” a preliminary component of a potential compound TCU in the form of (X-kara + Y) ([Because X + Y]).8 Seiji could go on to produce the final component, such as shotaimen demo nai n da yo ne “it’s not really that we meet for the first time,” and thereby make explicit his point of disagreement with Akira’s prior comment. However, he opts to leave it unsaid; in this context, what would be said in the final component is easily inferable, yet saying it overtly may cause some interpersonal tension. Thus, by proposing to exit from the turn with a third-position minimal
264
Makoto Hayashi and Kyung-eun Yoon
response in line 18, Seiji avoids stating the obvious and potentially delicate. This proposal of exiting from the turn is accepted and confirmed by Harumi, who initiates a new sequence of talk in line 20. In the previous Extract and also in Extract (2), the syntactically incomplete turn produced in Position 1 embodied a responsive action to what was said prior to it, and the third-position minimal response served not only the work of exiting from the turn-so‑far while leaving the obvious and/or delicate unsaid, but also the work of exiting from the topic talk/activity-so-far. Now, when a syntactically incomplete turn in Position 1 embodies a sequence-initiating action that makes a recipient’s response relevant next, third-position minimal response can be used to accomplish the work of pursuing a more-than-minimal response from the recipient. That is, when the speaker considers the recipient’s production of a mere acknowledgment token in Position 2 to be inappropriate as a response to his or her prior talk in Position 1, the speaker may deploy a minimal response token in Position 3 in order to prompt the recipient to revise his or her response and produce a more sequentially/interactionally appropriate response. A case in point is seen in the following Extract. In Extract (7), Young, the host of a dinner gathering at her apartment, mentions her idea of purchasing some kind of rug to put on the living room floor, and she solicits others’ opinions about it (lines 1–3, 5–6). (7) (Korean : Dinner talk among five friends [face-to-face: Young, Joo, and Hoon]) 01 Y: wuli-twu::, yeki-taka mwe -lul ilehkey we -also here-to something-AC like.this Like others, we have to- we are thinking about 02 com ↑kkal↑ -aya- ↑kkal↑ -kka-pwa:, i a.little put.down-must- put.down-what.if this putting something here like this, on here, ((what 03 wi-ey:, on-to do you think?)) 04 H: i wi-ey-ta[ka? this on-to-to On here?
Negotiating boundaries in talk
265
05 Y: [nemwu com ssellenghan ke-kath too a.little chilly thing-seem It looks too chilly, you know? (.) 06 -ay:, (.) ikhey cip-i:, -INT like.this home-NM Like, ((I mean)) this place, you know? 07 (0.5) 08 J: mwe sikthak? what dining.table “What, ((you mean)) the dining table?” : :((5 lines omitted where Young clarifies that :she is referring to the floor.)) : 14 J: a pata:k¿ oh floor Oh ((you mean)) the floor! 15 Y: u:ng. Ye:s. 16 (1.0) 17 Y: nemwu com ssellenghan ke-kath-ay:, too a.little chilly seem.like It looks too chilly, you know? 18 (2.2) 19 Y: °com° ttattushan ke-llwu:, a.little warm thing-with With °a little° warm thi:ng,
– Position 1
20 J: °um° °Mhm°
– Position 2
21 Y: u:m.
– Position 3
22 (.) 23 H: °patak-ey iss -nun ke° [kuntey] floor-on exist-RL thing DM °But the floor material° 24 J: 25 H: (
[kuntey]:(.) DM Bu:t (.) )[(
)]
266 26 J:
Makoto Hayashi and Kyung-eun Yoon [nam -tul]-i sayngkakha-n[u:n]= others-PL -NM think -RL
27 Y: [ e?]= Huh? 28 J: =kule:n↓ khapheythu-kathun ke-malkwu:, such carpet -like thing-not ((Lines 26 and 28)) Not the kind of carpet that others would think,”
After clearing up Joo’s misunderstanding about the referent of yeki (“here”; line 1) and ikhey cip (“this place”; line 6) in an other-initiated repair sequence (lines 8–15; mostly omitted), Young resumes the original activity and pursues Joo’s response to her idea about putting a rug on the living room floor. In line 17, she first repeats a part of what she said in lines 5–6 (“nemwu com ssellenghan ke-kath-ay:, “It looks too chilly, you know?”) and thereby re-proffers an opportunity for Joo to respond to her idea about carpeting. Though this utterance is produced while Young looks down wiping the dining table, she makes it clear that she is pursuing Joo’s response when she subsequently looks up and brings her gaze to Joo while producing her utterance in line 19. In this utterance, she provides a further detail about her idea of carpeting, thereby providing more material for the recipient to respond to. It finally succeeds in eliciting a response from Joo, but only in the form of a minimal acknowledgment token (line 20). It is at this moment that Young deploys a third-position minimal response (line 21). By indicating that she is not going to continue her syntactically incomplete utterance in line 19, Young proposes to exit from the turn and hands the floor back to Joo, thereby offering her another opportunity to respond to her prior statement. This move then succeeds in eliciting more elaborate responses from the recipients: Hoon, another participant who has been attending to Young’s talk, starts to present his opinion about Young’s idea (line 23), joined slightly later by Joo (line 24), who makes a suggestion that Young should not buy the kind of carpet that other people often buy (lines 26 and 28). In this section, we examined the use of third-position minimal response as a turn-exit device . We observed that speakers of Japanese and Korean sometimes leave their utterances syntactically incomplete in certain interactional environments, such as those
Negotiating boundaries in talk
267
where the rest of the utterance has already been clearly adumbrated in the prior context, and/or where saying the rest of the utterance would put the speaker (and the recipient) in an interactionally delicate situation. In those contexts, a speaker may use third-position minimal response to propose the closure of, and an exit from, his or her prior (syntactically not yet complete) turn.9 When the syntactically incomplete utterance in Position 1 embodies an action that makes a recipient’s response relevant next, third-position minimal response can be used to accomplish the work of pursuing a more-than-minimal response from the recipient. That is, when a speaker expects the recipient to produce a more-than-minimal response, yet only receives a minimal acknowledgment token, the speaker can deploy third-position minimal response to reproffer an opportunity for the recipient to revise his or her earlier response and provide a more relevant and appropriate response. Summary and discussion In this study we have explored the practice we termed the “thirdposition deployment of minimal response tokens” observed in Japanese and Korean conversations. The guiding perspective for our exploration is that, in order to understand the meanings of response tokens, it is crucial to examine the details of the sequential/activity contexts in which response tokens are embedded, such as what activities the participants engage in that give rise to the deployment of given response tokens, how the co-participants orient to them, and what consequences their deployment has for the organization of subsequent talk. We identified two broad types of actions that are accomplished through the deployment of thirdposition minimal response. One is to bring closure to the topic / activity-so-far by proposing the speaker’s imminent disengagement from it. The other is to bring closure to a syntactically incomplete turn by indicating that the speaker is not going to continue with the turn-so-far. Thus, we observed that third-position minimal response is used as a resource to regulate boundaries both at the level of turn organization and at the level of topic/activity organization. By way of concluding the present study, we will discuss some implications of our findings.
268
Makoto Hayashi and Kyung-eun Yoon
Response tokens and cultural style In the literature on Japanese communication style, response tokens have been one of the most common objects of study (e.g., Mizutani 1982, 1988; Maynard 1986, 1989; LoCastro 1987; White 1989; to name a few).10 The focus of many, if not all, of these studies has been to claim a direct linkage between response tokens (especially their frequency in Japanese conversation) and cultural traits and values of the Japanese. Thus, it has often been asserted that the behavior of sending frequent “back-channel cues” (or “aizuchi,” the Japanese term for listener response often used in these studies) in Japanese communication is a quintessential manifestation of the Japanese cultural values, such as “omoiyari”/“consideration/ empathy,” “ wa”/“harmony,” “sasshi”/“empathetic understanding,” etc., as seen in the following excerpts: [T]he emphasis on maintaining group harmony and on having smooth relations would probably cause speaker-listeners to use more aizuchi to show willingness to co-operate in the conversation and to show support of and attach value to the speaker on the part of the listener. This could be the underlying factor behind the notion that it is impolite in Japanese not to use back-channel cues. (LoCastro 1987: 110) The cultural value most relevant to the use of backchannels concerns the Japanese concept omoiyari, which, according to Lebra (1976), is a key concept for understanding Japanese people. [. . .] [T]he concept generally refers to the creation and maintenance of smooth and pleasant human interactions. [. . .] To maintain harmony, unanimity, or mutual understanding, people must be most sensitive to the recipient’s point of view and feelings. Being empathetic with others’ ideas and wishes may require going beyond indirectness and politeness [. . .] and involve compliance with the other’s ideas, even if they are opposed to one’s own (Lebra 1976). [. . .] In conversation, the fear of deviating from the speaker’s viewpoint and the eagerness to anticipate, understand, and accommodate the other’s idea may, in part, be demonstrated by the frequency with which the Japanese listener interjects with a backchannel. (White 1989: 67)
Along similar lines, Iwasaki ’s (1997) study of what he calls the “loop sequence,” which is very similar to the practice examined in the present study, presents the following conclusion.11 I propose that the frequent use of the loop sequence is a consequence of the Japanese conversationalist’s preference toward “ mutual dependency,” a concept held to be important in Japanese interaction. [. . .] The producer of the loop-tail [roughly equivalent to third-position minimal response in
Negotiating boundaries in talk
269
our study] reveals his helpless and submissive position (“I cannot contribute to the floor now. Help me out.”) and at the same time his trust in, or dependency (or amae) on, the other who he believes can provide a help. [. . .] The recipient of the loop-tail may either take up his responsibility and become the floor holder, or he may show his own helpless stance and counters the other’s loop-tail with his own loop-tail. That is, taking a floor in conversation is sometimes more of an obligation that a more powerful party must fulfill at a given moment in interaction rather than a right that he is entitled to. (Iwasaki 1997: 688–689)
From the point of view of the present study, however, there are some serious drawbacks in this type of claim for a direct linkage between response tokens and cultural values. That is, these accounts are invoked on behalf of the aggregated occurrence of response tokens (or the loop sequence), yet they do not show in any way whether the participants themselves in fact orient to such notions as empathy , mutual dependency, etc., when they produce response tokens or the loop sequence in the particular interactional contexts in which they are situated. For example, if we look again at the Extract examined above, we do not find any evidence that the producers of third-position minimal token (or what Iwasaki 1997 calls the loop-tail ) display their “helpless and submissive position” or their “trust in, or dependency [. . .] on” their interlocutors. The problem lies in the fact that, because these accounts are formulated in a manner far removed from the details of the actual contexts in which response tokens occur, it is unclear whether they provide any useful explanation of the meanings to which the participants themselves orient (Schegloff 1982, 1993; Zimmerman 1993). Of course, this is not to say that response tokens (or the loop sequence) are never used to display cooperation, empathy , dependency, etc. If one can show that displaying cooperation, empathy , dependency, etc., with a response token is a relevant action for the participants in a particular moment within the trajectory of the particular interactional activity that they engage in, accounts based on such notions may be viable for those particular cases. However, unless one can establish that displaying cooperation, empathy , etc., is invariably relevant for the participants when producing response tokens across different situations, such accounts as those cited above, however intuitively appealing they may be, remain equivocal as general explanations for the “meanings” of response tokens (Schegloff 1982, 1993; Sorjonen 2001).
270
Makoto Hayashi and Kyung-eun Yoon
As our analysis of third-position minimal response has shown, the meanings of response tokens depend crucially on the local sequential/interactional context of their usage and the kinds of tasks the participants orient to accomplishing by using response tokens in that context. Thus, rather than assuming a direct linkage between response tokens and cultural values and proposing premature generalizations about the meanings of response tokens for a particular cultural group, we should pay close attention to the immediate interactional environment within which a response token is precisely placed, and with respect to which it gets its meaning within the moment-by-moment contingencies of interaction. By doing so, we can gain an understanding of the detailed workings of response tokens in situ, not as an undifferentiated class of expressions taken out of context, but rather as deployable “devices” with which to accomplish different interactional ends in different contexts (Jefferson 1984, 1993; Sorjonen 2001). We hope that our study serves as a reminder of the importance of grounding our claims about response tokens (or any bit of human conduct, for that matter) in the careful observation of empirical details of their actual context of production. Cross-cultural /cross-linguistic comparisons While the predominant focus of cross-cultural/cross-linguistic studies is on substantiating differences rather than evidencing similarities (Schegloff 1987a; Beach and Lindström 1992), the present study was not motivated by a concern with demonstrating crosscultural/cross-linguistic differences. Rather, our interest has been to explore the operation of systematic solutions to certain general organizational problems of interaction and to evidence that those systematic solutions are utilized by speakers of different languages under similar interactional circumstances. Regulating the completion (or expansion) of turns, topics, activities, encounters, etc., is an omnipresent problem that requires continual resolution by the participants in interaction in any language community. Our study focused its inquiry on one specific form of conduct (minimal response) produced in one specific sequential environment (after a recipient’s production of a minimal response to one’s prior talk) and demonstrated how it is mobilized in both Japanese and Korean
Negotiating boundaries in talk
271
language communities as a systematic solution to the problem of regulating turn completion and topic /activity closure in that specific sequential environment. As discussed above, Japanese and Korean are not the only languages in which third-position minimal response can be deployed to regulate topic /activity boundaries; it has been reported that English speakers utilize a similar practice to accomplish topic/ activity transition (Jefferson 1993; Drew and Holt 1998). As for the usage of third-position minimal response as a device to exit from a syntactically incomplete turn, we are not aware of any study that has reported on this practice in other languages. Given the limited amount of empirical work done on response tokens in different languages, we cannot claim anything conclusive about whether the usage of third-position minimal response as a turnexit device is unique to Japanese and Korean. We certainly need further research to verify this point. That said, we have some tentative basis for speculating about the possibility that third-position minimal response as a turn-exit device may indeed be more likely to occur in Japanese and Korean because of certain distinctive features of turn constructional practices observed in these two languages, which may not be shared by some other languages, such as English. Our speculation is based on recent work on turnconstructional practices in Japanese and Korean and their relation to turn-taking (Maynard 1989; Fox et al. 1996; Kim 1999b, 2001; Tanaka 1999, 2000; Hayashi 2003, 2004; Morita 2005; Young and Lee 2004; Iwasaki 2007). It has been pointed out that, in conversational Japanese and Korean , speakers often adopt an “incremental” or “bit-by-bit” approach to turn construction. That is, rather than producing a whole turn at once, Japanese and Korean speakers tend to produce smaller units of talk (often phrasal units) during the course of a turn and leave a space at the end of each such unit for the recipients to produce a minimal response to acknowledge the just-produced unit (Maynard 1989; Kim 1999b, 2001; Tanaka 2000; Iwasaki 2007). As they receive such acknowledgments from the recipients at intra-turn phrasal unit boundaries, the speakers build up the whole turn by piecing together these smaller units. This type of incremental turn construction, it has been argued, relates to several characteristics of the “syntactic practices” (Fox et al. 1996)
272
Makoto Hayashi and Kyung-eun Yoon
employed by both Japanese and Korean speakers in conversational interaction. These syntactic practices include the following: In conversational Japanese and Korean, (1) word order within a clause is quite flexible, particularly toward the beginning of the clause; (2) many syntactic constituents (including core arguments) can be, and often are, left unexpressed (i.e. the prevalence of so-called “ellipsis”); and (3) elements placed toward the end of a clause (e.g., clause-conjunctive markers, predicate negation, quotative markers, etc.) can retroactively transform the grammatical status of the clause-so-far (Tanaka 1999; Hayashi 2003). If word order is flexible and many syntactic constituents can be left unexpressed, and if clause-final elements can retroactively transform the grammatical status of what has been produced so far, then the presence of some constituent at the beginning of a given TCU provides relatively little information as to how the rest of the TCU will eventually develop. This observation has led some researchers to claim that Japanese (and Korean) recipients tend to adopt a “wait and see” approach to turn-taking , which manifests itself in the frequent interpolation of acknowledgment tokens at TCU-internal phrasal unit boundaries discussed above. Thus, Fox et al. (1996) argue that, if turn-beginnings do not provide the recipient with much information about how the utterance is going to proceed, then it makes sense for speakers to produce relatively short [units] whose interactional implications the recipient can acknowledge or question as the speaker works on a larger turn (exactly what happens with multi-TCU turns in English). This allows the recipient to acknowledge small pieces without having to know exactly where the speaker is going with the full turn. (Fox et al. 1996: 212)
The discussion presented so far seems to provide some basis for arguing that the usage of third-position minimal response as a turn‑exit device is particularly fitted to the type of turn-constructional practices described for Japanese and Korean . That is, as discussed above, due to the limited mid-turn projectability of how the turn will eventually develop, recipients in Japanese and Korean conversations are systematically motivated to provide acknowledgment tokens at turn-internal junctures where the speaker’s turn is still not yet complete. When, as it happens, the speaker decides that the syntactically incomplete bits of talk that the recipient has just acknowledged are already sufficient from the perspective of executing an action, he or
Negotiating boundaries in talk
273
she can take advantage of the intra-turn acknowledgment token produced by the recipient as an opportunity to counter it with another minimal token and thereby indicate that the speaker is not going to continue with the turn any more. In other words, the very operation of third-position minimal response as a turn-exit device appears to rest on the systematic provision of intra-turn opportunities for recipients to insert acknowledgment tokens during the course of the speaker’s turn in Japanese and Korean. Thus, we would like to suggest that turn constructional practices in Japanese and Korean offer the kind of interactional environment that seems to be more amenable to the deployment of third-position minimal response as a turn-exit device than turn-constructional practices in some other languages, where there is no such systematic provision of intra-turn opportunities for inserting acknowledgment tokens by recipients. Of course, what has been discussed above remains speculative until further research is done on response tokens in languages other than those discussed here. We hope that our study serves as encouragement for others to explore the detailed workings of response tokens in a much wider range of different languages. Acknowledgments Earlier versions of this article were presented at the International Conference on Conversation Analysis (Helsinki, May 10–14, 2006) and the 92nd Annual Convention of National Communication Association (San Antonio, November 15–19, 2006). We are grateful for invaluable feedback provided by audience members at both meetings. We are also indebted to Hiromi Aoki, Kaoru Hayano, Kiyomi Kawakami, Mi-Suk Seo, and Jack Sidnell for their helpful comments on earlier drafts of this paper. Appendix AC CIRCUM CP DM EMPH FP
accusative particle circumstantial copula discourse marker emphasis marker final particle
274 INT LK N NEG NM O PL POL PT Q QT RL SP TP
Makoto Hayashi and Kyung-eun Yoon intimate speech level nominal linking particle nominalizer negative nominative particle object particle plural marker polite speech level particle question particle/suffix quotative particle relativizer suffix subject particle topic particle
Notes 1 We realize that “yeah” and “mhm” in English can work quite differently, especially with regard to whether they function to hand the floor back to the immediately prior speaker (i.e. the continuer usage; Schegloff ) or whether they serve to project their producer’s incipient shift to speakership (Jefferson 1984; Drummond and Hopper 1993a, 1993b). It is not clear to us where “un” in Japanese and “um” in Korean stand in this regard, and it is for this reason that we use the term “minimal acknowledgment” in a somewhat loose manner to capture this equivocality. 2 We exclude from our analysis those cases in which the deployment of a response in Position 3 is made relevant by the type of minimal response produced by a recipient in Position 2. For instance, such minimal responses as “Really?” produced by a recipient in Position 2 make relevant a subsequent response by the speaker-so-far in Position 3, as seen in the following fragment: (a) [Heritage 1984: 302] R: And I got athletic award C: REALLY? R: Uh huh.
Position 1 Position 2 Position 3
The sequential import of a minimal response in Position 3 in cases like this is very different from that of a minimal response in Position 3 when it follows a “yeah”/“mhm”-like acknowledgment token in Position 2, which does not make a subsequent response relevant. It is therefore important to make an analytic distinction between cases like (a) and those examined in this study. The practice we examine in our study,
Negotiating boundaries in talk
275
then, is not simply a third-position deployment of minimal response but is more appropriately characterized as the “post-minimal-response deployment” of a minimal response token. For convenience, however, we will use the label “third-position minimal response” in the rest of this article to refer to the practice of deploying a minimal response after another minimal response in Position 2. 3 We juxtapose the terms “topic/activity” here in order to capture the fact that the practice being examined is used to negotiate not only the boundaries of “what is being talked about” (i.e. topics) but also a shift in “the course of action that participants engage in” (for example, a shift from discussing the availability of a product to writing down its name; see Extract 2). 4 Gina’s self-interrupted utterance in line 9 could be an initiation of a similar statement, though it is not entirely clear. 5 “Ey” is a polite version of “um.” 6 Lerner (1991) uses the term the “preliminary component” to refer to a non-final sub-unit within a compound turn-constructional unit (e.g., [If X] of [If X + then Y]). The term “final component” is used to designate the final sub-unit that completes the whole compound unit (e.g., [then Y] of [If X + then Y]). 7 Compare this with more common, “unmarked” cases in which the speaker continues after the recipient’s acknowledgment token and brings the turn to syntactic completion: [Korean : Two Friends] 01 K: hankwuk ay–tul wena: k chongmyenghay–kacikwu:, Korea kid–PL very smart -because Because Korean kids are very smart, 02 J:
u: m= Mhm
03 K: → =kumpang kumpang ttalaha–nta. quick quick catch.up–DC they catch up so fast.
Below is a similar case from Japanese: [OBS 9] ((Middle-aged mothers are discussing how their daughters now wear their old clothes.)) 01 C:
tenkin no tanbi ni sutechatta kara sa::,= transfer LK every.time threw.away because FP Because I threw away ((my old clothes)) every time we moved,
02 B:
=[u:: n. Mhm
03 ?:
=[u:: n. Mhm
276
Makoto Hayashi and Kyung-eun Yoon
04 C: → nai kedo, attara kirareta not.exist but exist.if could.wear we don’t have any left, but if we did, 05 →
(kamo ne, °soo yuu no.°) maybe FP so say FP maybe ((my daughter)) could wear such clothes.
8 It has been pointed out in the literature that so-called conjunctive particles such as kara (“because”) and kedo (“although”) have come to be used somewhat like final particles in conversational Japanese and that they do not always indicate that the preceding clause is a “subordinate” clause to be linked with a forthcoming (or already-produced) “main” clause (Kokuritsu Kokugo Kenkyuujo 1960; Nakayama and IchihashiNakayama 1997; Iwasaki and Ono 2002). While this is true, examination of our data concurs with Tanaka ’s (1999: 195–199) observation that, when conjunctive particles are used turn-finally (as opposed to turn-internally), they are typically delivered in a turn-final, falling intonation. Furthermore, even when a conjunctive particle is produced in a falling intonation, it can be retroactively recontextualized as a “subordinate” clause-marking by producing a “main” clause after it. Thus, a third-position minimal token such as that observed in Extract 6 can be seen as a way to “seal off” the possibility of structural expansion of the turn-so-far. 9 In a study of some uses of “yeah” by Korean learners of English in spontaneous interaction, Park (2004) reports on a usage of “yeah” by those learners that is strikingly similar to the usage of third-position minimal response as a turn-exit device described in the present study. According to Park (2004: 91): “[N]onnative speakers often end their grammatically complete unit with continuing intonation, rising intonation, or simply non-ending intonation, and thereby, not properly ending their turn intonation-wise. As a result, they get native speakers’ continuer and miss an opportunity to leave their turn properly. They then need to use some extra marker or device to show their interlocutor that they are actually finished and need to leave their turn explicitly.” The “extra marker or device” that Park refers to is the minimal token yeah, as illustrated by the following fragment. [Park, 2004: 91] ((The instructor (T) has asked the student (D) which chapter of the textbook was the most difficult to study.)) 01 D: A lot and in case of:: 11, 02 T: mhm, 03 D: Word is very (0.2) difficult=hhh[h 04 T: [Right. Right. 05 There were a lot of animal name[s and so:: okay 06 D: [Yeah
Negotiating boundaries in talk 07 D: 08 09 T: 10 D: 11 T:
277
um: and:: (2.8) But I like to have a chance with(0.4) with other people and listen to other opinion::, mhm, °Yeah Okay. That’s good. All right. That’s basically that.
D’s utterances in lines 1, 3, and 7–8 constitute her answer to the instructor’s question about which chapter of the textbook was the most difficult for her to study. Notice that D produces the sentential unit in lines 7–8 with continuing intonation at its end (represented by a comma). This appears to prompt T to produce a continuer (“mhm” in line 9), which displays her understanding that D’s answer is not yet complete. Faced with this continuer, D produces “yeah” to indicate that she was in fact finished with her answer before the instructor’s continuer and that she is not going to continue her turn any more. Park (2004) does not address the native-speaker practice of deploying third-position minimal response in Korean, and she appears to regard the use of “yeah” shown above as a practice creatively devised by nonnative speakers in order to cope with their less-than-fully-developed linguistic and interactional competence. However, it seems clear to us that there is a strong connection between what Korean learners of English do with “yeah” to exit from a turn and what native Korean speakers do with “um” (and its variants) in similar interactional contexts. While, of course, nonnative speakers’ conduct cannot always be accounted for in terms of L1 transfer, careful examination of littleknown L1 practices may sometimes help understand what appears to be a peculiarly nonnative practice in a second language. 10 Though response tokens in Korean have not been studied as extensively as the Japanese counterpart, there are a growing number of recent studies that explore their functions in conversation; see K. Kim 1999b; H. Kim 2004; Young and Lee 2004. 11 What Iwasaki (1997) refers to by the “loop sequence” subsumes the type of sequence of utterances examined in this study, but it is not limited to it. Though the schematic representation of the loop sequence Iwasaki provides (see below; slightly modified from Iwasaki 1997: 662) corresponds fairly closely with what we examine in our study, he includes those cases where the second-position utterance by speaker B is an utterance that makes a response relevant next, such as a soo nan desu ka “Oh, is that right?” SPEAKER A (Some utterance) ee (“Yeah”)
SPEAKER B ee (“Yeah”) (Some utterance)
278
Makoto Hayashi and Kyung-eun Yoon
As stated in Note 2, we believe that the sequential import of a minimal response in third position following an utterance like “a soo nan desu ka”/“Oh, is that right?” is very different from that of a minimal response following another minimal acknowledgment token. For that reason, we made an analytic distinction between these two types of sequences and focused only on the cases where a minimal response follows another minimal acknowledgment token.
Part I V
Action formation and sequencing
9 Alternative responses to assessments Marja-Leena Sorjonen and Auli Hakulinen
Introduction In this chapter we examine part of the paradigm of utterance types used to agree with a prior assessment in Finnish. We use the term “assessment” in the same sense as Pomerantz (1984) and Goodwin and Goodwin (1992), for example, to refer to an evaluative act, typically performed by an utterance that contains a negative or positive predication of a referent or a state of affairs expressed by the subject or the object of the sentence. Alternatively, the assessable is something that can be inferred from the context. We focus on cases in which agreement is accomplished by presenting “the same evaluation ” as that of a previous assessment (cf. Pomerantz 1984: 66–68). In each of them, the recipient repeats part of the preceding assessment by her co-participant, leaving out the assessment term.1 Figure 9.1 shows the range of alternatives; those we will discuss are marked in boldface. The schema is a simplification in which we have, for the purposes of illustration, used as the first assessment a prototypical clause-type for assessing something: a predicate nominal clause with an evaluating adjective complement. In practice, there is variation in the clause type, and the evaluating element, “X” in our schema, is often a more complex phrase. Equally there are other verbs besides the copula “on” used in this context, but “on” is by far the most frequent one. What is striking in this schema is that of the six alternative response types, only two begin with an element other than the verb. This reflects the facts that in Finnish the subject is not an obligatory clausal element, nor is the order between the subject and the verb
282
Marja-Leena Sorjonen and Auli Hakulinen S
1st assessment :
V
se mekko on that dress
is
X
| V
S
X
hieno | on(pa) se mekko hieno great | is(CLI)i that dress great ‘that dress is great’
response :
ADV + V niin on
V+S
on se
V (+ V) on on on
S+V
V + PRT on joo
se on
i -pA
is a clitic particle which here adds to the emphatic function of the sentence (see Hakulinen et al 2004: 800).
Figure 9.1 Responses to assessments with a verb repeat
grammatical ly constrained (see, e.g., Heinämäki 1976; Vilkuna 1989; Helasvuo 2001). In contrast with a language like English, for example, in Finnish, the word order verb + subject does not, by itself, convey interrogativity. The crucial element in a “yes”/“no” interrogative is the question particle -ko (in colloquial Finnish also -ks), which is added to the initially positioned verb: “on-ko se totta?”/“is-q it true?” Accordingly, verb-initial word order can be deployed for indicating that a declarative sentence is either emphatic, as in the schema above: “on-pa se mekko hieno”/“that dress is really great” (literally, “is-clitic that dress great”), or responsive – for example, that it is an answer to a question, an assent or an agreement to an assessment (Hakulinen et al. 2004: 1316, 1325). So, one possible answer to the question “on-ko se totta”/“is it true” is a verb-initial sentence “on se”/“it is” (lit. “is it”). In our previous work (Hakulinen and Sorjonen, forthcoming), we examined four alternative assessment responses, “on,” “on on,” “on se,” and “on joo.” We found that a mere verb repeat (V) asserts agreement and is usable in a wide range of sequential contexts. The reduplicated verb repeat (VV), in comparison, occurs in a more narrow range of sequential contexts. The design of the prior assessment is especially relevant for the understanding of its use. Specifically, we found reduplicated verbs used where prior assessments expressed incontestable truths such as culturally conventionalized proverbs and the like. Finally, the (V + joo) format was found to be implicative of topic closure and routinely occurring in sequential environments where closure is a relevant possibility. In this chapter we will focus on three of the alternatives: “on se,” “se on” and “niin on.” In these alternative assessment response
Alternative responses to assessments
283
formats there is, in addition to the verb, a deictic element: either the third-person anaphoric pronoun “se”/“it” (in plural “ne”/“they”) or the adverb “niin.” The adverb “niin” is a derivative of the pronoun “se”/“it”; it is its plural form in the instrumental (instructive) case, meaning approximately “so; in that way.” It is even used as a response particle; in this usage, as well, the anaphoric meaning is still discernible (see Sorjonen 2001: 281–282, 287). Our initial aim was to explore the minimal pair “se on” vs. “on se.” However, it turned out that there were very few instances of “se on” in our data. Instead, the response type “niin on” proved to be extremely frequent. For that reason we examine all three alternatives. We show that the detailed structure of the response, including its word order and the types of elements it contains, specifies the way in which a “same evaluation ” should be understood in a given sequential context. We argue that a key parameter these alternatives index is the extent to which the participants approach the object of evaluation from the same point of view. Our database consists of approximately thirty interactions including both face-to-face conversations and telephone calls. Most of our data come from conversations among friends and family members, but we have also included a number of videotaped institutional interactions, for example from a social-insurance office and from a hair salon. VS response implying difference in perspective We first consider cases where, in her response, the recipient repeats the finite verb, followed by a pronominal subject (or by a pronominal object as in Extract 2 below). We refer to this as a VS response. By using the VS response, which sustains through ellipsis the evaluative term in the prior assessment, the recipient agrees with her co‑participant’s assessment. We argue further that, at the same time, she implies that their experience of the object of evaluation and its relevance for them is of a different kind. The nature of the difference may but need not be specified in the subsequent talk. Especially in conversations between friends, the VS response may be found in sequences where the topic is equally accessible to the participants. These cases thus appear to run against the meaning of the VS response as one that implies a difference
284
Marja-Leena Sorjonen and Auli Hakulinen
in the perspectives. For instance, in the following example the participants are discussing the weather. (1) (SKK/Sg01 B03, 8–9) 01 L:
Minkäslainen sää teillä nyt on.= what.kind weather you.ade now is What kind of weather do you have now.=
02 A:
=.hh Tänään on kaunista. Ihan kirkas taivas näkyy olevan =.hh today is beautiful quite bright sky seems be.prpc =.hh Today it is beautiful. There is quite a clear sky
03
tuol[la.hh there over[ there
04 L: 05
[Niin- e niin täälläki ja niin komee ruska että [so er so here.too and such great autumn.coloring that [So it is here as well and such a great autumn coloring
.hh[h
06 A:
[JOO:.hh [YEA:.hh
07
(.)
08 L:
että nyt:hän siellä [o( )( )( ). that now.cli there [w( )( )( ). that now it (would) ]
09 A:
] ]
[Nyt kelpais olla]k[i. Mut kyllä [now worth.would be.cli but sure [Now it would be something. But it is really
10 L:
[Ai( ) ]
11 A:
sie]l on ihanaa heti kun ei sada. there is lovely immediately when neg rain lovely there as soon as it’s not raining.
12 L:
↑Nii, [Joo, ↑Right, Yeah,
13 A: → 14 →
[.hh Kyl se on: syksy on niin mahdottoman [ prt it is autumn is so impossibly [.hh It really is: the autumn is so extremely
kaunis.[h beautiful.h
Alternative responses to assessments 15 L: →
285
[On se.= [is it 2 .
16 A: → =.Jo[o .Y[ea 17 L:
[>Kyllä mä vi- ei viikonloppuna menen< .hhh mä meen [> sure I we- neg weekend go.1 I go.1 [>I’m surely wee- at the weekend I’m going< .hhh I’m also
18 kans t- kääntää ↑maat ja .hh laittamaan kuntoon varmuuden also turn lands and put shape.ill safety’s going to dig the land over and .hh put everything in shape to be 19 vuoks kaikki jos (.) jos sitte ei tuu enää °mennyks° sake everything if if then Ø neg comes anymore going on the safe side if (.) if one doesn’t get to go there °anymore°.
Anna’s (A) assessment in lines 13–14 is located in a larger segment of weather talk. It was preceded by talk about Leila’s (L) summer cottage which Anna had checked out while visiting hers during the weekend. The assessment in lines 13–14 moves the talk into autumn weather in general. This generalization can be heard as closing implicative. However, it receives a VS response, which, we argue, foreshadows further talk by suggesting that the participants have different perspectives on the topic being discussed. The ensuing talk explicates the difference in the participants’ perspectives . In overlap with a further closing-implicative turn by Anna (an inhaled response token “.joo,” cf. Hakulinen 1993) in line 16, Leila, continuing at the pitch level of her preceding response, tells of her plan to visit her cottage the following weekend. This leads into a troubles-telling; it turns out that Leila is considering selling her cottage. The VS formatted response is thus deployed by the recipient, at a potential place for moving into the closing of the call, as a way of constructing a bridge from the joint talk about the weather to her own, related talk about giving up her summer place. In the next example from a social-insurance office, the official (O) responds with a VS-formatted turn to an assessment by the client (C) in line 10. The client has come to ask whether he should pay back some of his study grant, as he has earned too much in his part-time work to be eligible for the full amount that he received.
286
Marja-Leena Sorjonen and Auli Hakulinen
(2) (Kotus, T1000: 2–3, Social insurance office, Helsinki) 01 O:
Paljokos sulla on ollu niitä tuloja. how.q.cli you.ade is be.ppc those.par income.pl.par How much income have you had.
02 C:
No mull on olluk kaheksantoisttuhatta et well I.ade is had eighteen.thousand prt Well I have had eighteen thousand so
03
seittemäntoisttuhatta seittemänsataa. seventeen.thousand seven.hundred.
04
(.)
05 O:
Joo katotaa (hetki). prt look.pass moment Yeah let’s have a look (a moment).
06
(2.0) Official taking a printout from the printer
Mä otan tom päätöksen tosta samalla mä 07 O: I take.1 that decision there.ela same.ade I I’ll take the decision from there while I’m 08
[Takes a paper from the printer [haen sen. ] [fetch.1 it ] [fetching it ]
09 C: → [Se ov vä]hä hankala arvioida etukätee. [it is some difficult estimate Ø beforehand [It is a bi]t difficult to estimate ((them)) in advance. 10 O: → O:n ne. joo; se o ihan totta. is they prt it is quite true O: n ne. yes; it is quite true. Tuntilaisena teen. 11 C: hour.ag.ess do.1 I do ((the work)) as a timeworker ((“paid by the hour”)) 12 O:
clears throath krhrr
13
(2.0) Official reading the text on the computer screen
14 C:
°Joo° °Yeah°
15
(0.2) Official reading the text on the computer screen
16 O:
.mt Eli kakstuhatta↑kaks: on tulluv valvonta, prt two.thousand.two has come control .tch So two thousand and ↑two: there’s been a control,
Alternative responses to assessments
287
The assessing utterance in line 9 is an account: The client provides his reason for having to pay back some of the allowance. The utterance does not contain any explicit person reference forms referring to the client himself but is formed with a zero-person format (Hakulinen 1987; Laitinen 2006). Through it, the account is constructed as generic (i.e. applicable to anyone), and this invites the official to display affiliation with the client and in that way her acceptance of the account (cf. Sorjonen 2001: 131–140). The official treats the turn as one that sought agreement by responding with a repeat of the copula verb “o:n,” followed by the subject pronoun “ne”/“they” (line 10). The copula “o:n” is here in its singular form since in colloquial Finnish there is no number agreement in the third-person plural (see also Extract 4, line 6). The antecedent to the plural pronoun “ne” is found in line 1, “niitä tuloja”/“the income,” which is plural in Finnish. In line 9, the reference is achieved with a zero anaphora object, marked with a Ø-sign in the English gloss. In Finnish, the conditions for zero anaphora are more flexible than in English; anaphoric subject and object pronouns can alternate with a zero both across sentence boundaries and across speaker change (see Laitinen and Hakulinen, 2008). By using this response format the official implies that the participants’ perspectives on the issue are different. While the client presented his evaluation as someone who is expected to have correctly estimated his income, the official’s response is understood to be based on the experience she has gained in her job, dealing with other clients in a similar position. This difference in perspective is treated as readily understood – it need not be spelled out. From the assertion of agreement , the official proceeds to upgrade her agreement: “It is quite true.” With her turn then the official accepts the client’s account for having earned too much money. The repeated verb is accented, and also lengthened. Our intuition is that the prosody contributes to the reassuring flavor of the response. What we have seen so far is that the VS response, while asserting agreement , conveys a difference in the participants’ perspectives on the assessable item (autumn weather, difficulty in anticipating earnings). Even though the prosody may work to strengthen the agreement (as in Extract 2), the verbal construction opens up a possibility for further talk from a different perspective. The difference hinted at by the design may but need not be specified in the ensuing talk.
288
Marja-Leena Sorjonen and Auli Hakulinen
SV response implying independence of speaker’s stance Most of the alternative responses in the schema presented above are verb-initial – word-order being a means of indicating that the turn is a response. As there are also two alternatives that do not begin with a verb, the question we face is what speakers are doing when using these formats. We shall first examine the S + V format where the recipient starts the response with an anaphoric subject pronoun, followed by a repeat of the verb in the prior assessment. Our intuition is that by starting the response with the subject, the recipient, while agreeing with the evaluation presented by the first assessment speaker, implies having already held the same stance (see also Tainio 1993: 189–193). In the few cases we have, the following is the only one that features an exact repetition of both the subject and the verb in the co-participant’s prior assessment. The example comes from our field notes, which is why some details of the sequential environment are not available. We participated in a conversation in which a colleague was praised for having given a successful job talk. After a characterization by B of the kind of talk the candidate had just given, speaker A, who did not participate in the job talk, presents an assessment which consists of an evaluation of the person as a teacher in general. It is formatted as a predicate nominal clause, with an SV word order , and it evaluates the person in extreme terms. (3) (Field note, spring 2007) 01 A:
Se on ihan hirveen hyvä opettaja. it is quite terribly good teacher He is just an extremely good teacher.
02 B: → Se on. it is He3 is
By responding with the format “se on,” B both asserts agreement with A and implies the independence of her stance . In the prior talk, B was the one who was telling the others about the specific event, portraying the success of the colleague’s teaching. Now that A presents her assessment, B’s turn can be heard as agreeing but simultaneously confirming A’s stance . The function of the SV response comes close to that of the partial modified repeats in
Alternative responses to assessments
289
English investigated by Stivers (2005: 138–143). Stivers examines response turns consisting of a repetition of the subject and the verb of the prior assessment (e.g., “it is”), as in our Extract 3 above. She shows that by producing the repeat verb with marked prosodic emphasis (stress) the second speaker claims primary rights to the claim. In Extract 3 above, the copula verb in line 2 carries an accent too. However this is not an obligatory feature of SV responses as illustrated by Extract 4 below. Thus it seems that an action that in Finnish is performed with the word order of the response is expressed through prosodic means in English. In Extract 3, the fact that the response begins with the subject, instead of the finite verb, is essential for its function. However, unlike in Extract 3, in our other cases of SV-formatted responses, the subject pronoun differs grammatical ly from that of the previous assessment. The following example from a hair salon contains a modification – specifically, a change from singular to a plural in the subject position (lines 3 and 6). (4) (Kotus, T1208: 61, Eastern Finland, hairdresser’s) 01 C:
Joo. Yeah.
02
(33.7) H cutting Client’s hair
03 H: → .mt Voi mahoto mite o itsepäine hius. oh impossible how is obstinate hair .tch Oh my god what an obstinate hair. 04 C:
Mm-m.
05
(0.6)
06 C: → .nf Ne on. they is. 07
(2.0)
08 H:
Mite sie sitä aina kotona ite laitat. how you it.par always home.ess yourself make.2 How do you set it at home by yourself.
09 C: No geelillähän sitä pittää °mh°, (0.7) muottoilla. well gel.ade.cli it.par Ø must shape Well you have to use gel °mh°, (0.7) to shape it. 10
(5.0)
290
Marja-Leena Sorjonen and Auli Hakulinen
11 C: °Kyl°lä se aina asettuu ku s’tä aikasa, surely it always settles when Ø it.par time.poss It does settle alright when one just ((treats)) it 12
aikasa vaa °et[tä° (men]ee). time.poss prt so.that (goes) for awhile so that it (goes ((down)))
13 H:
[Mm-m. ]
At line 3, the hairdresser evaluates the client’s hair. By using the nominative singular “hius”/“hair” instead of the plural form “hiukset,” the variant used in everyday speech, she treats the client’s hair as material for professional designing. The client agrees with the evaluation of his hair, but the subject initial format (line 6) indexes his independent knowledge of the matter at hand. By choosing the plural form of the anaphoric pronoun “ne”/“they,” he brings forth his own experience, that of doing his hair every day. This is then elaborated in the subsequent talk (lines 9–12). This difference in the speakers’ perspectives resembles what we discussed above, in connection with Examples 1–2, with “on se” (VS) format. However, in our present example, the difference in the perspective is conveyed by the change in the grammatical number of the subject (from singular to plural), whereas the SV order is doing another job – that of indicating independent access to, and prior stance toward , the object of the assessment (the speaker’s own problematic hair). Thus the two formats differ in the following way: The VS format conveys a difference in the perspectives of first assessor and recipient, and the SV format conveys the prior existence of the assessment recipient’s stance . In the SV format, expressing agreement may be intertwined with a function of������� confirming the co-participant’s stance . “Niin on” response displaying unmodified agreement In the introduction to this chapter we mentioned that while the subject initial “se on”/“it is” response is infrequent in our current database, there is another, much more frequent response type in which the verb occupies the second position within the turn. This response format begins with the adverb “niin”/“so; in that way” which is then followed by a repetition of the finite verb of the first assessment
Alternative responses to assessments
291
(“niin on”/“so is”). With this response type, the recipient asserts unmodified agreement with the prior speaker, without the further implications characteristic of SV and VS responses. In the following example the speakers discuss the reservation they need to make for a badminton field where they have just started to play. In lines 8–10 they compare this field with the one where they used to play. (5) (SKK/Sg06 A06) 01 S: .mt Et tota niin .hhhhh hh mä [tilaan viel toisen ja sit prt prt prt I [order.1 yet another and then .mt So uhm then .hhhhh hh I’ll [reserve yet another ((time)) and then 02 V: [hhhhhhhhh 03 S: katotaa vähän millanen m:millane se kenttä et onks se look.pass a.bit what.kind what.kind the field prt is.q it we’ll see a bit what kind m:what kind of a field it is and if it is 04
sama järjestely ku viimeks. same arrangement as last the same arrangement as last time.
05 V:
°Jo[o°. °Yeah°.
06 S: 07
[>Sehän ei ollu mikää jättihyvä mut kylhä siin [ it.cli neg was any super.good but sure. cli it.ess Ø [It wasn’t y’know particularly great but one can surely
pelaa.= plays play on it.=
08 V: → =Kyl siin pelaa joo ei se tos >et ne oli sure it.ess Ø plays yeah neg it there prt they was =Sure one can play on it yeah it won’t there >so those were 09 →
kyl kivat kentät oikeestaan tuolla (1.2) .mt Olarissa. sure nice fields actually there Olari.ine real nice fields there in fact (1.2) .tch in Olari.
10 S: → Niin on. Nehän on tosi hyvät. so is they.cli is true good.pl Niin on. They are real good. 11
(1.0)
12 V:
Et kylhän tos tietyst voi sitte (.) harkita. prt sure.cli there of.course Ø can then consider. So one can indeed think about it then of course.
292
Marja-Leena Sorjonen and Auli Hakulinen
13 S:
Välillähän sitä voi harkita sometimes it-par can consider Every now and then one can think about it
14
[(tietyst sitte) [of.course then [(of course)
15 V:
[Täytyy sit vaa kattoo tosiaanki [must then just look indeed [One just must see to it indeed
16
et e(h)i tost r(h)ytmist tipahda ku heh se o that neg that.ela rhythm.ela drop.out when it is n(h)ot to drop out from the rh(h)ythm as it heh seems to be
17
vitunmoinen työ [näköjää päästä toho ryt(h)mii näi fucking job [apparently get that.ill rhythm.ill thus a fucking hard job to get into the rh(h)ythm like this
18 S: 19 V: 20 S: 21 V:
[h(h)h he he he
p(h)uhelim(h)itse ainaki. [.hhhh et mä en sit phone.through at.least prt I neg.1 then over the pho(h)ne at least. [.hhhh so I don’t [hihihihi .mh(h)h £mielellään nyt niinku tipu siitä£. preferably now prt drop.out it.ela £like to drop out it sort of£.
After agreeing with Sami (S) that the present field is manageable, Ville (V) moves to give a positive evaluation of the field where they used to play (lines 8–9). Sami agrees with this using the “niin on” format response in line 10. As was mentioned in the introduction, the adverb “niin” is derived from the anaphoric pronoun “se”; its meaning is not quite equivalent to the English adverb “so,” but could be approximately glossed as “in that way.” “Niin” refers to the evaluative term in the co-participant’s preceding assessment (to “kivat”/“nice” in line 9 in this example). “Niin on” is the only one of the response formats that has an evaluative element as one of its components: After “niin on,” Sami explicates his agreement by upgrading the assessment, further demonstrating his agreement with the evaluation. In the “niin on” response above, there is a change in tense: While Ville presented his evaluation in the past tense (“oli”/“were,” line 8), Sami uses the present tense and thereby highlights the current relevance of the assessment. In this sequential context, this choice of
Alternative responses to assessments
293
tense implies that there is still a possibility of returning to the old field. In the subsequent talk, this option is acknowledged by Ville (line 12). The use of the unmodified, strong agreement presented by “niin on” is not restricted to affiliative contexts. Our next example comes from a situation in which there is disagreement between two speakers. A group of men are discussing what a boss ought to do with an employee who is found drunk at the worksite. What is at issue here is whether drunkenness is a sufficient reason for being fired or whether some additional troubles are required, such as a fight, to warrant firing. In lines 15–16, VM expresses his opinion that fighting is worse than using alcohol. To this, MI responds by saying that fighting should be seen as an aggravating circumstance. In this way he implies that drunkenness should be seen as the primary reason for being fired. The previous speaker VM receives this with a “niin on” response (line 19). (6) (SKK/Sg003 A, 12) 01 V:
=Mut oliks (.) oliks se tää Kivine ku aiheutti sen but was.Q was.Q it this name who caused the =But was it (.) was it this Kivine who caused the
02
luunmurtuman tai [siis (.) tappe[lun.= bone.break or [prt fight fracture or [I mean (.) the[ fight.=
03 M: 04
[No ei: [well neg [Well no:
[tässä sitä ole:kyllä muuta [here it is in.fact other [it doesn’t say so here4 other
ku että. than that than
05 (1.0) 06 V: 07 M: 08 ?:
Sen yhteydes[sä vaan, its connect [ion.ine only Just in connection with it, [Nii nii että (.) [ol ryyppäämine ja [ prt prt that [was drinking and [Yes yes that (.) [((there)) was drinking and [O[W-
294
Marja-Leena Sorjonen and Auli Hakulinen
09 M:
[sitte, [then,
10 V:
[°Joo joo (.) mm°,= [° Yeah yeah (.) mm°,=
11 M:
=Mutta tuota noin ni joka tappauksessa ni, but prt prt prt any case.ine prt =But um er so in any case that,
12
(1.0)
13 M:
↑joko hänen ois kylymästi< (.) potkassa either he be.cond coldly should sack ↑either he should coldly (.) should sack this (.)
14
tämä (.) (vi[hollinen). this (.) (en[emy). 5
15 ?:
[(Tää vih[ollinen.) [(This enemy.)
16 V:
[>Minä katson ainaki siis< että ku [I look.1 at.least prt that when [>I think at least that is< that when
17
tulee tappelu ni se on .hh siis pahe- se on melkeen comes fight so it is therefore worse it is almost there is a fight then it is .hh in fact wor- it is almost
18
pahempi ku: (.) .= worse than alcohol’s use worse than (.) .=
19 R:
[Nii;= [Yeah;=
20 M: → =Se on ras:kauttava asian[. ] it is aggravating circumstance ] =It is an agg:ravating circum[<stance>.] 21 V: →
[Niin on. ] [so is ]
22 →
°se on:° (.) sillon se on (.) pahoinpitely. it is then it is assault °it is:° (.) then it is (.) an assault.
23 M:
Mm-m,
24 V:
Ja (.) vaikka kats- ajattelee asiaa siis ihan tollaviisii and though Ø look thinks the.matter prt quite that.way And (.) even if one look- thinks about the matter just like
Alternative responses to assessments
295
25 noin vaan ö:.mt ei mitenkää hh (.) juridisesti niinku sanotaa like.that neg any.way legally as say.pass that er .tch not in any way hh (.) legally as they say 26
[hianost. .hh [refinely. hh
27 M:
[Mm-m,
28 V:
vaan >tota noin< ihan noin kaverien kesken et jos tulee (.) but prt prt quite prt mates.gen among that if comes but >sort of< just among mates that if there is (.)
29
ryypätessä tappelu nin minusta sillon ni (.) saa kyl drinking.ine fight so I.ela then prt Ø may surely a fight amidst drinking so in my opinion then (.) one can
30
antaa lem:put. give sack give the sack.
31 R:
Niin kyllä °se on°, so sure it is So sure °it is°,
Ei siis pelkkä alkoholinkäyttö nii sitä mä en sano 32 V: neg therefore mere alcohol.use prt that I neg.1 say I mean not just the use of alcohol so I am not saying 33
et saa suoranaisesti [sitä voi varottaa ensin. ] that Ø may simply [prt Ø can warn first ] that one may downright [one can give a warning first.]
From the strong agreement with “niin on” (line 21), V moves to present his own interpretation of the situation. Note that at the beginning of line 22 what looks like an instance of the format “se on” is in fact an instance of self-repair: Speaker V abandons his utterance beginning and restarts it with an adverb “sillon”/“then.” He defines a fight in legal terms, as his co-participant had also done, but he uses the technical term “pahoinpitely”/“assault.” In this way he ignores the implication that the primary reason for getting fired is being drunk, conveyed in the first assessment, thereby holding to his own stance expressed in lines 16–18. In the subsequent talk, he further explicates his stance (see lines 24–33). In this example the recipient thus uses “niin on” strategically: He asserts agreement as a way to get back to enforce his own stance .
296
Marja-Leena Sorjonen and Auli Hakulinen
In the next case, the “niin on” response is also followed by a disaffiliative utterance by the same speaker (see Sorjonen 2001: 174–176 for an analysis of the same response). Iia has called to ask whether Mia would join her to see a play in a summer theatre. Instead of responding to the inquiry, Mia has delayed the response by asking about the price of the tickets. Iia answers by saying that she does not know the price and then goes on to tell Mia what she knows about the availability of tickets. Her turn in line 1 is part of that telling: There are more tickets available for weekdays than for weekends. (7) (SKK/Sg08 A01, 2) 01 I:
J’ >mut sit arki-iltoina on ihan (.) tost and but then weekday-evenings.ess is right that.ela An’ >but then for weekday evenings there are still tickets
02
heinäkuun:viidennest päiväst lähtien ni °lippuja.° july.gen fifth.ela day.ela starting prt tickets right (.) from July:fifth onwards.
03
(1.1)
04 M: Onkohan se hyvä.=Eihän sitä tiedä ennen kun näkee. is.Q.cli it good neg.cli it Ø knows before Ø sees I wonder if it’s good.=One doesn’t know before seeing. 05 I:
Mm:.
06 M: → M’t kyl kuus tuntii on ai:ka kau°hee°. but indeed six hours is quite awful B’t six hours is really qui:te aw°ful°. 07 I: → Niin on; Siin on kaks väl°iaikaa°. so is it.ess is two intermissions Niin on; It has two inter°missions°. 08
(1.2)
09 M:
°Jeoo°. °Yeah°
10
(0.8)
Mut kai se on semmone et se pitäs 11 M: but maybe it is that.kind that it should Ø But I guess it’s the kind of thing that one should 12 13 I:
nähä [°kuitenki°. see [°anyway° [°Mm:.°
Alternative responses to assessments
297
14 (1.8) 15 I: °nJoo.° .hh et jos mä nyt ostan johonki< (0.6) päivään prt so if I prt buy.1 some.ill day.ill °Yeah.° .hh so if I buy tickets for some< (0.6) day, 16 liput, ni ostanks mä sulle °kanssa°. tickets so buy.1.Q I you.all also shall I buy for you °too°. 17 M: ↑Nii sä tarttet sillo varmaan mun opiske- aijaa mut ei prt you need then certainly my student prt but neg ↑Right you probably need then my stu- oh but there 18 ei siell_oo, onks siellä mitää opiskelija neg there is is.Q there any student isn’t, is there any student ((discount))
In line 6, Mia produces a negative assessment of the length of the play. In this context, the assessment implies a possibly untoward stance toward going to the theatre. The recipient responds with “niin on” by which she asserts agreement with the co-participant (line 7). She thus focuses on the general stance -taking character of her co-participant’s utterance. After that, she addresses the implication that the length of the play might discourage them from seeing it. She does this by indicating that the play will include two “intermissions.” This continuation thus disaffiliates with the co-participant in terms of the possible consequences of her assessment. In lines 11–12, following trouble-foreshadowing silences (lines 8 and 10) and her minimal acknowledgment, the first assessment speaker produces a turn with which she implies the possibility of an acceptance of the invitation . In this case, a participant asserts agreement with a “niin on” response in the midst of a pending response by her co-participant who had produced evading and rejection-implicative responses to her invitation . With the “niin on” response she affiliates with her co-participant’s assessment as a general one of the state of affairs. She then, however, moves to express disaffiliation with the kinds of consequences the assessment could be heard to have for the ongoing larger activity. The “niin on” response can also be used as a means to treat an utterance as an assessment although it was not primarily designed as one. In the following extract the recipient responds with “niin on” to an utterance that is hearable as a piece of advice . Tiina
298
Marja-Leena Sorjonen and Auli Hakulinen
is talking over the phone with Ripa and Anni. Both Tiina with her husband and Ripa with Anni are renovating their homes. As part of promising Ripa and Anni that she will visit them, Tiina complain s about the hard work she and her husband have been doing and mentions their intention to take some time off (lines 3–5, 8). Ripa responds with an utterance that is constructed as a positive evaluation of taking a vacation but whose primary function, due to its sequential position and the attention getting particle “kuule” (“listen,” see Hakulinen et al. 2003), is to give advice . The utterance contains the zero person that leaves it unspecified for whom it is good to take a vacation. This utterance gets a “niin on” response by Tiina (line 11). (8) (SKK/Sg06 A01 20) 01 T: .hh Kyl me sit tullaan kuha me täst vähä sure we then come when we here.ela a.bit .hh We’ll definitely come when we have ((sorted out)) a bit here 02 A:
[Joo::. [Yea::h.
03 T:
[Me ollaan nyt ajate]ltu et sit ku Kalle lopettaa [we have now thought that then when 1nameM finishes [We have been thinking that after Kalle finishes
04
hommat ni sen jälkeen me otetaan ii:sisti et ei chores prt it.gen after we take easy prt neg his job then we are going to take it easy as we don’t have
05
[me nyt jakseta tällai [ku me ollaa nyt kak]s [we now have.energy this.way [as we have now two [the energy ((to work)) in this way “cos we have
06 A:
[Joo:. [Yea:.
07 R: 08 T:
] ]
[Joo:. [Yea:.
viik[koo oltu tuolla .hhh wee [ks been there have been there now for two weeks .hhh
09 R: →
] ] ] ] ]
[Ja sit on k- sit on kuule he:le]vetin hyvä [and then is listen then Ø is listen hell.gen good [And then it’s l- then it’s listen da:mn good
Alternative responses to assessments
299
10 → ottaa välillä lomaa?, take now.and.then vacation to take a break every now and then 11 T: → Niin o[n. so is 12 R:
[(-) siit [ju()(-) [(-) from it (-)
] ]
13 T: →
[Se on erittäin tärkeetä. ] [it is extremely important ] [It is extremely important. ]
14 A:
[(Nii. Sit siin on) [prt then there is [(Yea. Then there is)
] se et ] it that ] the fact that
15
vaik me ollaan tehty vaa viikko tätä although we have done just week this although we’ve been doing this for a week only
16
mut siis me ollaan nukahdettu joka ilta kello but prt we have fallen.asleep every night clock but I mean we have fallen asleep every night at
17
kymmenen ([)(-) ten ten o’clock ([) (-)
18 T:
] ] ]
[.mhh Joo ] sama meilläki et ku – – [ prt ] same we.ade.cli that when [.mhh Yeah] the same with us too that when – – -
By responding with “niin on,” Tiina disattends the advice giving function of Ripa’s turn by making use of the alternate interpretation that the zero person reference allows: She treats Ripa’s utterance as one that presented a generic assessment on what would be a good way of conducting one’s life and asserts agreement with it. In this way she transforms the ongoing activity from advice - giving to one in which equal partners assess a state of affairs. Subsequently (line 13), she explicates her stance with a strong positive assessment. By using here the assessment term “tärkeetä”/“important,” she echoes the voice of lifestyle advice , enforcing her personal competence in the matter. In this case, with the “niin on” response and its continuation Tiina emphasizes the joint character of the activity and in so doing resists the advice offered by Ripa.
300
Marja-Leena Sorjonen and Auli Hakulinen
Conclusion In this chapter we have discussed three alternative ways of constructing an agreeing response to an assessment, and the interactional functions of these responses. All the responses we have analyzed are clausal, that is, they contain a finite verb. Two of these – the VS and SV formats – include the anaphoric subject but leave out the evaluative term of the first assessment altogether. By contrast, the subjectless variant “niin on” with the anaphoric adverb “niin”/“so; in that way,” is composed of a predication only: The object of evaluation expressed with the subject-NP in the first assessment is in this format treated as fully shared. The “niin on” response, being a predicating expression, foregrounds the evaluation by having the adverb in the initial position. These two aspects make “niin on” a construction for expressing strong agreement. The three response types we have analyzed show that participants in interaction do not merely express agreement in terms of its strength. We have seen that the detailed structure of the response, including its word order and constituents (subject or object pronoun or the adverb “niin”), conveys the way in which a “same evaluation ” should be understood in some particular sequential context. What seems to be at issue is the extent to which the participants approach the object of evaluation from the same point of view. Thus the “on se”/“is it” response carries an implication that the perspective of the response speaker differs from that of her co‑participant. This difference in perspective may be explicated in the subsequent talk. The subject-initial response “se on”/“it is,” by contrast, implies that the response speaker already held the stance earlier and in so doing underscores the independence of the agreement . This is perhaps most clearly seen in cases where the “se on” speaker’s primary or independent access to the object of evaluation is evident in one way or another from the context. In these cases, the “se on” utterance may perform a double task: While claiming agreement with the prior speaker it also functions as a confirmation of the validity of the co-participant’s assessment due to its conveying the priority of access to the matter at hand. The response types that contain the subject pronoun (“on se”/“is it” and “se on”/“it is”) carry an implication that the response
Alternative responses to assessments
301
speaker qualifies the agreement: Either she implies a difference in the perspective (VS) or conveys that she already held the same stance before the current interaction. In comparison, the subjectless response “niin on”/“so is” focuses on the sharedness of the current assessing activity: It merely asserts agreement. However, we have seen that this type of agreement can be put to various uses. It strikes us as remarkable that in our data, it often functions as a means of asserting agreement that is then deployed for moving to one’s own, opposing stance . Because the three response types are constructed as clauses they make available to the speaker grammatical resources other than word order and the presence or absence of the subject to further specify their point of view on the evaluation. For instance, we have shown cases in which the response speaker changes the tense of the verb or the number of the subject NP so as to modulate or qualify agreement . As we have shown above, a function similar to the one expressed in Finnish by word order may have to be expressed through prosodic means in English (Stivers 2005; see p. 286 above). In addition, the meanings of the Finnish VS and SV responses resemble what Heritage (2002) has described as “oh”-prefaced responses to assessments in English. According to Heritage, “ oh”-prefacing is “a method persons use to index the independence of their access and/or judgment in relation to the state of affairs under evaluation” (2002: 204). This function of “oh” rests on its more general change of state semantics. Even though the issue of epistemic access can be and often is associated with the VS and SV responses, that is not the key issue with them – the difference in perspective (VS) or the independence of the assessment (SV) can get specified in the ensuing talk in several different ways. Furthermore, as compared to “oh,” the VS and SV responses, by being formatted as clauses, have a different range of sequential contexts where they can occur: as responses to assessments and as answers to questions but – unlike the English “oh” – not as responses to informings. The range of morpho-syntactic means, on the other hand, is not available in English given the role that word-order plays in syntactic processes; verb-initial formats are used for interrogatives and imperatives only. We have shown that a number of basic morpho-syntactic characteristics of Finnish are reflected in the organization of assessment
302
Marja-Leena Sorjonen and Auli Hakulinen
sequences. These include, most notably, the nonconfigurationality of Finnish word order , in other words the possibility of using word order to make discoursal and not grammatical distinctions. In addition, there is the possibility of forming a sentence without a subject: with an anaphoric zero, a response is tied to the utterance of the prior speaker. Finnish is not unique in terms of its morphosyntactic characteristics. However, languages that share some of the same typological features may not deploy them for the same interactional purposes. For example, Japanese (Hayano 2007) appears to have an assessment response format coming close to “niin on” but functioning differently both with respect to its structural relation to the prior assessment and with respect to its interactional functions. On the other hand, Estonian, a language closely related to Finnish, does not deploy the same types of response formats: The only verb initial response is (verb + particle) (Renate Pajusalu, personal communication). Acknowledgments We would like to thank Jack Sidnell for most helpful comments on our article. An earlier version of this article was delivered at the Annual Meeting of Finnish Linguists , at the Annual Meeting of Finnish Conversation Analysts and at ICCA-06 in Helsinki, and we are grateful to participants in these events for comments. Appendix: Glossing symbols used ADE AG ALL CLI COND ELA ESS GEN ILL INE NEG PAR
adessive case (“on”) agentive morpheme allative case (“(on)to”) clitic conditional mood elative case (“from”) essive case (“at”; “as something”) genitive case illative case (“(in)to”) inessive case (“in”) negative auxiliary or particle partitive case
Alternative responses to assessments PASS PL POSS PPC PRPC PRT Q 1 2 1nameM S V X
303
passive plural possessive suffix past participle present participle particle question particle 1st person singular 2nd person singular male first name subject verb any complement
Notes 1 It is of course also possible to present a same evaluation by repeating the entire prior assessment, for example, “Se on hieno”/“It is great.” These cases fall out of the scope of our present focus. 2 We will not provide an idiomatic version of the target lines as we do not know what the exact functional equivalents of English in these cases would be. 3 The pronoun se is used for both human and non-human referents in most of the colloquial varieties. 4 The problem that the men were to solve was presented to them in written form. 5 The speaker indeed refers to the problematic worker as an “enemy”.
10 Language-specific resources in repair and assessments Jack Sidnell
Introduction For humans, social action is, by its very nature, embedded in the rich semiotic structures of the life-world. Such structure not only provides the background against which particular actions are recognizable but also the raw materials out of which they are constructed in the first place. In talk-in-interaction, social action is built up out of the particular prosodic, lexical, and grammatical resources of a given language and is thus necessarily endowed with a partially language-specific character. In this chapter, I show that the use of “if” as a preface to next-turn repeats provides speakers of Caribbean English Creoles with apparently unique possibilities for social action with no clear analogue in other languages. In their basic, canonical use, “if”-prefaced repeats are used to initiate repair of a prior turn formatted as a “yes”/“no” question (YNQ). In this sequential context, the “if”-preface acknowledges the ongoing activity of questioning even while it suspends it in order to initiate repair. This use of “if”-prefacing appears to be a reflex of grammar in language varieties which, unlike many other varieties of English, do not use inversion to form YNQs. We can see this use of “if”-prefaced repeats then as the mobilization of local grammatical resources to solve a generic interactional problem: the problem of conveying that one heard a prior trouble-source turn to be a “yes”/“no” formatted question. This basic usage makes possible a number of other distinct practices quite independent of initiating repair. In the second section of the chapter I turn to consider two of these. First, I discuss the use of “if”-prefaced repeats as second assessments. I argue that, as second assessments, “if”-prefaced repeats treat a prior assessment as if it were a question and thus as
Language-specific resources
305
something said with some uncertainty. Second assessors use such “if”-prefaced repeats to not only agree with a prior assessment but also and simultaneously to claim epistemic priority from second position (Heritage and Raymond 2005; Stivers 2005). Second, I discuss the use of “if” as a stand-alone item. In the particular case I examine, stand-alone “if” apparently treats a prior turn as stating the obvious and thereby casts the emerging agreement and alignment of the moment as an artifact of a general consensus. An examination of “if”-prefacing in these language varieties opens up a number of issues for comparative conversation analysis. First, it shows that differences in the way a distinction is coded can have significant, domino-like, effects which link together apparently unrelated practices in talk-in-interaction. Second, it illustrates the way in which practices of speaking that solve generic interactional problems are shaped in part by the specific character of locally available resources such as, for instance, language-specific grammatical resources. Other-intiated repair Other-initiated repair ( OIR) is a subset of practices which participants in conversation employ to address and potentially resolve problems of speaking, hearing or understanding (Schegloff et al. 1977). The following example presents an instance from American English . Here Jon and Guy are discussing where they might be able to play golf later in the afternoon. (1) NB 1.1. 01 J: Well I’m s:↑ure we c’get on et San Juan ↑Hi:lls ↑that’s 02 ni: ce course ah only played it ↑o:nce. 03 G: °Uh huh?° 04 (0.6) 05 G: a→ .hhh °↑It’s not↑ too bad,° 06 (0.4) 07 J: b→ Hu:h? 08 G: c→ ‘S not too ba:d, 09 (.) 10 J: No:.
In extract 1, repair is initiated (arrow B) in the turn directly following the trouble source (arrow A ) and the repair itself follows in the turn after that (arrow C). The repair as a whole constitutes
306
Jack Sidnell
a sequence inserted between the second assessment “It’s not too bad” at line 5 and the confirmation of it at line 10. It is important to see that although repair is used here to fix an apparent problem of hearing – an orientation to which is revealed, in part, through the use of repetition in line 8 – it is implicated, at the same time, in the larger course of action being pursued. That is, at lines 1–2 Jon has offered up San Juan Hills as a place he and Guy might play golf and has recommended it on the basis that it is a “nice course” and that he has “only played it once.” Guy’s “It’s not too bad” responds specifically to the assessment “nice course,” refusing to endorse it with an equivalent or upgraded evaluation (e.g., “yeah, real nice,” etc.). It is within this context of emerging disalignment that Jon initiates repair. In doing so he provides Guy with an opportunity to revise, modify, or withdraw his nonaligning assessment. And notice that when Guy does not do this, Jon treats it as if it were a positive assessment agreeing with “no:.” Jon might have agreed with “yeah” but this would have been to accept the negative assessment and the rejection of Jon’s proposal this embodies. The point is, then, that while repair is often used to fix problems of hearing, speaking, and understanding, it is not simply or solely a piece of the “systems engineering” as Goffman (1976) might have suggested. Rather, when we look at particular instances we find that whatever else repair may be doing it is also a vehicle for social action. A number of comparative studies suggest that other-initiated repair is, in terms of its organization, essentially generic and, in terms of its distribution, universal (see, e.g., Egbert 1996; Moerman 1977; Ochs 1984; Schegloff 1987a; Sidnell 2006; Wu 2006). At the same time, these studies indicate that there are also some important and striking differences across communities. This variation within a clearly identifiable domain makes repair an attractive topic for comparative study as many of the chapters in this book demonstrate (for instance, Fox et al., Egbert et al., Wu). Other-initiated repair in two Caribbean communities The data for this study comes from long-term fieldwork in two Caribbean communities. Between 1994 and 1996, I spent twelve months in a rural Indo-Guyanese village of approximately 700 people that I refer to as Callander (see Sidnell 2005a for an ethnographic
Language-specific resources
307
analysis of talk and knowledge in this village). I recorded about eighty-five hours of talk in the village, the bulk of which consists of ordinary conversation in people’s homes, at gathering places along the road and in the rumshop. Due to the character of the phenomenon examined in this chapter I also draw from a number of interviews I conducted with villagers. Between 1999 and 2004, I spent approximately seven months conducting fieldwork in Bequia – the largest of the Grenadine islands with a population of about 5,000. Although this corpus includes a large number of audio recordings, for the present study I have drawn exclusively on a smaller corpus of approximately ten hours of video-recorded data. The video-recordings capture ordinary conversation in the yard of an extended family and in the homes of island residents. An examination of the recordings collected in these two communities reveals that the organization of other initiated-repair in the Caribbean data is remarkably similar to that in American English as described in Schegloff et al. ’s well-known study (1977).1 Several different strands of evidence may be cited in support of this claim. First, the various formats used to initiate repair are almost identical. Thus, in the Caribbean data the major formats for initiation include open-class “ what?” and “huh?” as well as questionintoned repeats and class-specific question words such as “who?,” “where?,” and combinations thereof (e.g., “They well have who?”). Second, a robust prefer ence for self-correction is observable in that others typically await the first possible completion of a turn before initiating repair; indeed, as Schegloff et al. (1977) observed for American English, others typically delay initiation even further allowing a slight gap to develop between the completion of the turn containing the trouble source and the other-initiation of repair. Moreover, others typically only initiate repair, leaving it up to the speaker who produced the trouble source to provide a correction. Third, other initiated repair is used in a remarkably similar way not only to fix troubles of speaking, hearing, and understanding but also to convey disagreement , disbelief, and other problems of interactional alignment .2 This suggests a base of generic principles organizing repair across differences of language, culture, and social arrangement (see Moerman 1977, Schegloff 1987a, 2006; Sidnell 2001 and 2007a). There are, however, some differences. Elsewhere I’ve discussed one
308
Jack Sidnell
such difference – a repair initiation format apparently specific to Bequia that is particularly fitted to deal with problems encountered in the making of recognitional reference to persons (Sidnell 2007a). In what follows, I discuss another difference having to do with the way in which a participant can, in the course of otherinitiating repair, convey that she heard that the trouble-source turn was a YNQ. “If”-prefaced repeats in the initiation of repair The collection of other-initiated repairs includes a small number of instances in which a partial repeat or wh + partial repeat formatted initiation is prefaced by “if”. Pat’s turn at line 2 of Extract 2 is an instance. Here Pat and Benson are sitting side by side in the yard that adjoins Benson’s small house. It is a week after Carnival and Pat has stopped by for a visit with Benson’s neighbor. (2) #187_Q2 qt 51:50 01 B: yu biin hii fu kanival (.) Pat? were you here for Carnival Pat? 02
(.)
03 P: → if mi bin wa? if I was what? 04 B:
Bekwe fu kanival? Bequia Carnival?
05 P:
yeah: yeah
When Benson asks Pat in line 1 if she was in Bequia for Carnival, Pat responds by initiating repair with “if mi bin wa?”/“If I was what?’ – an “if”-prefaced repeat. The preframing “mi bin”/“I was” combined with the question word “wa”/“what” isolates “hii fu kanival”/“here for Carnival” as the trouble source to be repaired. By prefacing the turn with “if,” Pat also shows that she heard the turn addressed to her as a YNQ. At line 4, when Benson repairs the reference, he preserves the status of his turn as a question by producing it with rising intonation. At line 5, Pat answers the question with “yeah.” The function of this “if”-preface then is to show that the one initiating repair heard that the turn containing the trouble source was produced as a YNQ.
Language-specific resources
309
In most varieties of English, one standard way of forming YNQs is to invert the ordering of subject and auxiliary verb in simple declarative constructions (Quirk et al. 1985). So “You’re going for a nap” becomes “Are you going for a nap?”. Such inverted syntax can be preserved in repeats which initiate repair so that, for instance, repair is initiated with forms such as “Am I GOing?” or “Am I going for a what?”. The fragment given as Extract 3 provides an illustration from American English conversation: (3) TC II(b): #28 6 P: D’yknow where Mister Bowdwin is. 7 (0.2) 8 B: Wha:t? 9 (·) 10 P: hhuh-hhuh-°hu-° [˙hhh 11 B: → [Do I know where who? 12 P: Leo is. 13 B: No. 14 P: Oh. Okay.
Here, by preserving the inverted syntax of the trouble source turn (“Do you know . . . ?” becomes “Do I know . . . ?”), Bush shows that he heard the turn to be a YNQ. Notice then that Pyatt’s repair in line 12 not only fixes a problem of person reference by substituting “Leo” for “Mr Bowdwin,” it also completes the question that Bush began in his turn at line 11. Notice moreover that Bush displays an orientation to the ongoing activity of questioning by providing an answer at line 13. In the creoles of the Caribbean there is no auxiliary-subject inversion in YNQs; indeed, there is no syntactic category of auxiliary verb for such an inversion to operate upon (see Winford 1993). Instead, in these varieties, a turn’s status as a question is constituted through a range of features of design and context. In the following case, Naomi asks her nephew Ozan, who is looking at a newspaper, whether he read a story about “the Brewster boy.” (4) Q1.12.55–14.02 02 N: → Oozan yu riid ting bou i Bruusta bo::y? Ozan did you read the thing about the Brewster boy? 03
(.)
04
hii kripl an ii ina jeel, he’s crippled and in jail
310 05 O:
Jack Sidnell oo gad, mi keer abou hii?= Oh God, do I care about him?
A number of factors intersect to help the turn at line 2 come off as a question. First, it is produced with slightly rising intonation and a sound stretch on “boy.” Second, this is a clear B-event in the sense that whether Ozan has read the story is clearly something Ozan knows better than Naomi. And, third, at the completion of this unit, Naomi stops speaking momentarily thereby providing Ozan with an opportunity to respond. When he doesn’t, Naomi builds a next component on to the turn. In the following case (Extract 5), Emmanuel and Viv are explaining to Baga that the anticipated time of his arrival prevented Emmanuel from going to town earlier in the day. At line 7, Baga formulates the upshot of what has been said with a “so”-prefaced turn. In this case then the B-event character of the turn is established by the immediately subsequent talk. Although there is no discernable rise in the intonation of the turn the recipients immediately recognize this as a question and respond accordingly. (5) Garden Scene qt. 2: 15 03 E: ai woz goin tong tu bo- wii [gain ai gaing I was going to town too bu- we’re going I’m going 04 V: [ya bika hii bin fu yea because he was to 05 go tong ai se wait bika de see di man go to town I said wait because they said the man was 06
koming tudee. coming today
07 B: → so hi kudn go tong bika wen mi comin So he couldn’t go to town because of when I was coming 08 E:
[yeah
09 V:
[yeah
A final case illustrates (Extract 6) another important resource in the design of questions. Here Cat remarks to Roger that the mother of a passerby is from Hamilton (their natal village). Winnie agrees with this and subsequently elaborates, asking if Roger knows “Joanne the daughter of Dorothy Quashie.” As she says this, she adjusts her body so as to achieve mutual gaze with Roger. This is
Language-specific resources
311
sustained throughout the production of the turn. As Stivers and Rossano (2009 ) have recently shown, gaze like this is a resource for mobilizing response and here it appears to contribute to the constitution of this turn as a question. (6) OFB2.27.00–27.24 him mada biilongz tuu Hamilton. 01 C: his mother is from Hamilton. 02
(0.6)
03 W: ya yu noo ahm Kwi- Joo An Ya you know uhm Qui- Joanne 04
fo Daretii Kwashii? the daughter of Dorothy Quashie?
05 C:
ya:. Daretii Kwashii. yeah. Dorothy Quashie?
06 R:
Oh::.
These examples illustrate some of the practices such as intonation and directed gaze that speakers use in designing a turn to be a recognizable question. With respect to the argument being developed here, it can be noted that these features cannot be carried over to turns that other-initiate repair in such a way as to show that the trouble source -turn was heard to be a YNQ. In turns that other-initiate repair, intonation is typically deployed in the service of repair-relevant tasks – e.g., marking the position of a trouble source within a repeat. The use of gaze to convey that a question is being asked is wholly indexical and thus similarly not transferable to turns that initiate repair. So these important resources for showing that a question is being asked are not available to the one initiating repair as resources for showing that the trouble-source turn was heard to contain a YNQ. Notice, moreover, that in these varieties a simple repeat of the trouble source turn will not be grammatical ly marked as a question and will be potentially hearable as confirming what has been questioned. In the Caribbean creoles then there is no subject-auxiliary inversion in YNQs, and the intonation of the trouble-source turn is not preserved in turns that other-initiate repair. “If”-prefacing fills the resultant gap, conveying that the recipient heard the trouble source turn to be a question. This is the basic interactional environment
312
Jack Sidnell
for “if”-prefacing, and the practice appears to be related in a basic way to the grammar of questions in these varieties. Because they are organized as a series of questions and answers, interviews provide a rich source of data. The examples below are from interviews I conducted with villagers in Guyana. (7) Callander_Interview_T41a (13: 17) 01 J: arait om:(1.2) yuu eva kos, alright, uhm do you ever curse 02
(0.8)
03 G: → if mi eva kos? if I ever curse? 04 J:
o
mhmo=
05 G: =yes mii a kos Yes I curse (8) Callander_Interview_T42b (30:10) 01 J: uhm:(.) yuu eva kos an ting? Do you ever curse and that? 02
(0.2)
03 Sh: → if mii eva kos? if I ever curse? 04 J:
ye::s. yes
05 Sh:
ye:[s yes
06 J: [eh he he ha huh (9) Callander_Interview_ T42b (31:02) 01 J: alrait uh:m (.) yu eva hiir ma:n kos? Have you ever heard men cursing? 02
(0.4)
03 Sh: → if mi ev hiir man kos? if I ever heard men curse? 04 J:
yes= yes
05 Sh: =yea:h, yeah
In each case, the recipient of a YNQ responds initially with an “if”-prefaced repeat and thereby initiates repair of the question.
Language-specific resources
313
The “if”-prefaced repeat begins an insertion sequence that temporarily suspends the activity of questioning underway. These examples suggest that creole “if” functions as an “interrogative particle” (Sadock and Zwicky 1985 ) that marks whatever it prefaces as a question. Further evidence for such an analysis is provided by instances of reported speech in which, instead of using a verb such as “ask,” creole speakers often use a “say” + “if” construction. In the following example (Extract 10), for instance, Seeta asks if my wife Allison is still planning to go to Georgetown the next day. When I confirm this, Brammie asks, “Shi noo tong?”/“Does she know her way around town?”3 With “mos bi”/“most be,” Seeta responds in a way that shows she is less than completely certain and is basing her answer on inference. The reported speech at line 8 is apparently addressed to me. (10) Guyana_T5a.pm3_29: 10 01 S: alison go tong tum[ara sti:l Is Allison going to town tomorrow still? 02 J:
[yea:h yeah
03
yeah. yeah.
04 B:
Shi noo tong? Does she know her way around town?
05
(0.2)
06 S:
mos bi Most be.
07
(0.6)
08 → hi see if shi noo tong He asked whether she knows her way around town. 09 J:
na relii bonot really bu-
I suggested earlier that “if”-prefacing was connected to the grammar of YNQs in a basic way; specifically, given that in these varieties there is no subject-auxiliary inversion in YNQs, a repairinitiating repeat cannot convey that its speaker heard the turn containing the trouble-source to be a YNQ. Reporting speech presents similar issues. If Seeta had reported this without the “if”-prefacing
314
Jack Sidnell
she might have been heard as reporting an assertion. She could perhaps have conveyed that what was being reported was a question by preserving the intonation of Brammie’s original utterance, but this would have been to invite inferences about what she was doing in reporting the talk in this way (mimicking for instance). “If”-prefacing then is a practice for marking a turn as a question in the absence of grammatical inversion. As we’ve seen, it is used in just this way in both the other-initiation of repair and reported speech .4 Some of this has been noted before. Thus, Cassidy and Le Page’s Dictionary of Jamaican English (1980: 234) describes “if/ef” as an introductory interrogative word, “the abbreviation of some such phrase as I wonder (if, etc.), or Did you ask me (if, etc.).” They report two uses of “if”: “introducing a question” and “introducing the repetition of a question which (having not been fully understood) is repeated to make sure that that was the question” (Cassidy and LePage 1980: 234). In fact, there is no reason to assume any process of abbreviation or reduction. As Sadock and Zwicky (1985: 183) write, “it is not infrequently the case that the formal indications of yes-no questions resemble those of the antecedent of conditional sentences (the clause in a conditional that is expressed in English with if).” They cite biblical Hebrew and German as examples of languages that mark YNQs and antecedent clauses in conditional constructions in the same way. Having described the use of “if”-prefaced repeats in the otherinitiation of repair, I now turn to consider an apparently derivative use of “if”-prefacing in assessment sequences. “If”-prefaced second assessments In their basic, canonical environment, “if”-prefaced repeats initiate repair of a prior question turn and, in so doing, begin an insertion sequence that breaks the contiguity of first and second parts of an adjacency pair (Schegloff 2007b). “ If”-prefaced repeats are also deployed in a quite different sequential context – specifically in response to assessments. In this environment, “if”-prefaced repeats are second assessments that agree with a prior. Extract 11 provides a first illustration of the practice. Here Shanka and Kiki are sitting in their family yard with three young children
Language-specific resources
315
close by. At line 1, Kiki directs one these children – Zaria – to move and, in close succession, to “pick up that thing” an object on the ground in front of her. She completes the turn with a question about the whereabouts of Zaria’s cousin, Roxanne. Zaria responds only to the final part of the turn – waving her arms around in what might be several pointing gesture s. Kiki then directs her attention to the slightly older girl Naksin. As Kiki produces this turn, Zaria runs away. This occasions Shanka’s evaluative assessment of Zaria at line 6 – “Wailnes Zaria a kom wid”/“Wildness Zaria comes with.” At line 7, Kiki responds with “if Zaria wail?”/“if Zaria is wild.” (11) #134_Q1 qt 25: 38 01 K: Zaria muv from de.=tek a ting. Zaria move from there. Take that thing. 02 K: wapa Rakzan. Where is Roxanne. 03
(3.0) ((Zaria is waving arms in possible pointing gestures))
04 K: Naksin lii shii:. tek da- an ting- Jak ting. Naksin leave her, take that and thing, Jack’s thing 05
(0.2) ((Zaria runs away))
06 S:
Wailnes Zaria a kom wid. Wildness Zaria comes with
06 K: → if Zaria wail? if Zaria is wild 07
(0.4)
08 Naksin kom. Naksin come
We can see then that this assessment sequence is occasioned by a complex set of visible behaviors and witnessable actions: the failure to comply with the directives, the flailing hand gesture s, the running away. We can also observe that although the “if”prefaced turn recycles parts of the earlier talk it is not a verbatim repeat of it. This kind of loose repetition seems to be doing several things. First, the incorporation of earlier talk marks the “if”prefaced turn as operating on something that came before. Second, the looseness of the repetition suggests that something other than repair initiation is being done. And, third, the particular words
316
Jack Sidnell
reused formulate the “kernel” of something that someone else has said. Notice also that the “if”-prefaced repeat does not elicit any response from the recipients – indeed, this seems to close the assessment sequence and the talk turns to other matters. Consider next Extract 12. This begins with Donna calling to her nephew (who is off camera). Although he appears to respond, he does not comply with the request to “come here.” After Kitana beckons the same boy again, Benson turns to Donna and remarks, “he’s rude you know.” This initiates a string of assessments culminating in an “if”-prefaced repeat. (12) #139_Q1 qt 39: 25 01 D: Gusnel kom bai hee Gushnell come over here. 02 (Gushnell):
(for yu)
03 K:
( ) kom he. come here.
04 B:
hii ruud yunoo he’s rude you know o
05 D: ai noo hi ruud I know he’s rude 06
(1.2)
07 B:
riil ruud. real rude
08
(0.4)
09 E:
huu ruud? Who’s rude?
10 B:
da boi [de. That boy there.
11 D: →
[if hi ruud? if he’s rude?
Notice that the “if”-prefaced turn in line 11 once again closes this, quite extended, sequence of assessments. With it, Donna seems to have “the last word” on the matter and the talk turns to other concerns. It is worth noting that in both this and the previous example the assessment is the vehicle for a complain t about a nonparticipant, third party.
Language-specific resources
317
In the final example (Extract 13) of this practice to be considered here, the participants are talking about an upcoming party to be held at a local park. After Manuel notes that people are expected to come from many nearby islands for the event, Cheryl augurs the consequences for the park, saying “da pak a mash op”/“That park is going to be a mess.” This can be heard as a way of saying that the party will be particularly lively and thus as an assessment of the anticipated event. At line 4, Manuel remarks, “if da pak a mash op agen”/“if that park is going to be a mess?” Notice that this “if”-prefaced repeat does not elicit any immediate response from the participants. Rather, after a pause, Cheryl raises the possibility that one of the expected musicians may not show up for the event. Cheryl seems to be backing off her earlier positive assessment here – in saying that Poser might not come to the event she anticipates a possible problem. In overlap with this, Roxanne adds “Sunday is the queen show.” By noting another aspect of the event not so far considered, this is supportive of the positive assessment that Manuel seconded with his “if”-prefaced repeat. Notice also that while the “if”-prefaced turn comes close to a verbatim repeat of what has just been said in line 2, the speaker adds the word “again” and in this way marks it as something other than a word-for-word repeat. (13) #31_AC1 qt 33: 00 01 C: dee kyaan cheenj it. They can’t change it. 02
da pak a mash op.= That park is going to be a mess.
03
=biisaidz besides I
04 M: →
[mi
()
[i- if da pak a mash op agen. if that park is going to be a mess?
05
(0.6)
06 C:
if Pooza na [kom an piipol if Poser doesn’t come and people
07 R: 08 M:
()
[oan Sundee iz di Kwiin Shooo and Sunday if the Queen Show hou ma fu pee fu goo in=I tink iz How much is it to get in= I think it’s
318
Jack Sidnell
This instance differs from those considered above. Here the first assessment – Cheryl’s “da pak a mash op”/“That park is going to be a mess” – while hearable as an assessment, is not clearly designed by its speaker to be one. Rather, Manuel’s “if”-prefaced turn, somewhat creatively, locates in what Cheryl has said something that could be heard as an assessment and responds to it as such. And notice that in this case the participants continue to assess the event. While here again the “if”-prefaced turn does not elicit an immediate response and the talk that follows it is marked by a subtle topical shift, Manuel is not able to achieve “the last word” as Kiki and Donna were in the previous examples. This then looks like an instance in which the practice is stretched to deal with a slightly different turn and, as a result, is not completely successful. In this collection of cases, “if”-prefaced turns are used to do second assessments (Pomerantz 1984). It is notable that in all these cases, the one who produced the initial assessment does not respond to the “if”-prefaced turn. This suggests that the second assessors are not being heard as requesting clarification or as indicating a problem of hearing. What, then, do these “if”-prefaced second assessments do? In their basic interactional environment, “if”-prefaced repeats initiate repair on a YNQ. In other words, “if”-prefaced repeats convey that their speaker has heard a previous turn containing the trouble-source to be a YNQ. In these last three examples, the practice appears to be used to treat a prior assessment as if it were a YNQ. With these “if”-prefaced second assessments, second assessors are doing agreement . One piece of evidence for this is seen in the fact that, like other agreements (and prefer red actions more generally), these turns are closing -implicative – they help to bring topics or sequences to a close. By contrast, disprefer red actions and disagreements in particular tend to be sequence-elaborative. So notice that in Extract 11 after the assessment sequence at lines 5–6 the talk turns to other matters (Kiki beckons Naxine). And, in Extract 12, a long string of assessments concludes with an “if”prefaced repeat at which point the participants again turn to other matters.
Language-specific resources
319
But these “if”-prefaced turns go beyond just agreeing . By treating the first assessment as if it were a YNQ, and thus as something said with some uncertainty, an “if”-prefaced second assessment suggests that a first assessment was not strong enough. There is then an epistemic dimension to this practice. As a number of recent studies have demonstrated, even where participants agree on some state of affairs, they not uncommonly negotiate the terms of that agreement. As Heritage and Raymond (2005) show, an important dimension of this negotiation involves epistemic rights to assess a given state of affairs. These authors have argued that a first assessment carries an implied claim to epistemic priority but that this can be managed through various practices of speaking; specifically, there are practices for downgrading claims to epistemic priority in first position (such as tag questions) and, inversely, practices for upgrading claims in second position (such as “oh”-prefacing). “if”-prefaced second-position assessments work in part by treating a first assessment as a question and thus as downgraded relative to a declaratively formatted assertion. With this practice then the participants can be seen to orient to the participants’ differential epistemic rights to assess. In Extract 13, for instance, Manuel is responding to what he hears as an assessment from his aunt. By virtue of their ages and the social networks in which they participate, Manuel may well feel he has greater rights to assess the party (moreover, it is Manuel who raised the topic in the first place). Indeed, Manuel has just been telling the other participants about all the people coming for the party from other islands, thereby displaying relatively more extensive knowledge about at least this aspect of the event. 5 In Extract 11, the first assessment is produced by Shanka, who is Zaria’s cousin, and the second by Zaria’s aunt (Kiki), who is partially responsible for her. The assessment here is a complaint and thus the “if”-prefaced format of the second assessment may be selected to deal with a situation in which Kiki feels she needs to reassert her greater rights to assess the child (see Raymond and Heritage 2006). And, finally, in Extract 12, the participants are again assessing a child in the vicinity. Here Benson is assessing Donna’s nephew. The design of the initial exchange is sensitive to Donna’s greater epistemic rights here. Specifically, Benson designs his first assessment as a
320
Jack Sidnell
question, in this way inviting Donna to confirm it, which she does in a particularly explicit way (“I know he’s rude”). The sequence continues with Ezekiel initiating repair (“huu ruud?”/“Who is rude?”). After Benson repairs the reference with a demonstrative referring expression, Donna responds to the initial assessment again, now with an “if”-prefaced turn. Here again, then, relative rights to assess are at issue: This is, after all, Donna’s nephew that Benson and Ezekiel (a family friend) are assessing. Like the modified repeats described by Stivers (2005) and discussed briefly in the introduction to this volume, “if”-prefaced second assessments claim epistemic priority from second position. While these look like quite different practices on the surface, the underlying logic is actually quite similar. Stivers shows that modified repeats work by confirming a turn that did not invite such confirmation . In this way, modified repeats impute to the first-position speaker an uncertainty that they did not express. By treating a prior assessment as if it were YNQ, “if”-prefaced repeats similarly impute uncertainty to the first assessment speaker. So the larger point here is that in this interactional environment, “if”-prefacing treats a prior assessment as if it were a question. “If”-prefacing takes its sense and import in this environment from its canonical use in the other-initiation of repair of a prior YNQ. Its use in second assessments is apparently derivative of a more basic use in initiating repair and this in turn appears to be a reflex of grammar in language varieties that do not use syntactic inversion to form YNQs. Stand-alone if as an interjection “If” is also used as a stand-alone item in a way that suggests a relation to the uses already described. Because it occurs unattached to any other turn components and in a particular sequential position, I’ve described this an interjection. The following case (Extract 14) is illustrative. Here, in a fragment I’ve discussed elsewhere (see Sidnell 2005a: Chapter 8), the participants are a group of men who form what is described in the ethnographic literature as a “crew” (see Wilson 1973) – adult men who grew up together in the village and consider themselves peers or “age-mates.” They are drinking together at their local rumshop and have been talking about
Language-specific resources
321
deceased persons in their parents’ generation. This is an environment of intense epistemic competition, and, here, relative rights to assess the person being talked about appear to be at issue. In line 1, John links a new character “Buddy’s father” to the discussion up to this point. This negative assessment (“fucker”) adumbrates a story about Buddy’s father but first elicits a round of strongly aligning responses.6 (14) Callander_Rumshop_T66 (04.40) 01 John: >Soo bodi dadii, ee da f:uka noo fu po[nish waif.= Just like Buddy’s Father. That fucker loved to beat his wife 02 Jaj:
[da rait. That’s right
03 Rohan: =iif? If? 04 Jaj:
da ii kaal neem. That was his nickname
05 John:
tek a leedi . . . grabbed that lady . . . ((Story continues))
At line 2, Jaj agrees , saying “That’s right” and, moments later, “That was his call-name,” meaning, it appears, that this was such a well-known fact about “Buddy’s Father” (i.e. that he was a “wife-beater”) that people used it to name him. At the interstices of Jaj’s talk lies Rohan’s contribution: “iif?” I originally glossed this as “You’re telling me!” (Sidnell 2000: 79–80), and, in a certain sense, that gloss is perfectly adequate, even if it seems to get the illocutionary value exactly backwards. Notice then that in this context of emerging collective agreement “ if” is not treated as initiating repair. Rather than occupying first position in an insertion sequence, this stand-alone “if” is structurally similar to the “if”-prefaced repeats in assessment sequences. Here, stand alone “if” does not elicit a response and, indeed, works in conjunction with the expression “da ii kaal neem”/“That was his nickname” to close down this initial assessment sequence and thus forward the talk toward the telling that the turn in line 1 projected. It is likely this use of “if” that the author of the Dictionary of Caribbean English Usage has in mind when he characterizes it as
322
Jack Sidnell
“expressing certainty or strong agreement” and noting that, used in this way, it is typically “uttered with high rising pitch and strongly stressed, sometimes lengthened, vowel” (Allsopp 1996: 303). He provides the following example from a popular Barbadian magazine (The Pelican): “Hah!” Daphne laughed. “You ain’t put he in he place . . . ?” “If?” Claire exclaimed. “Who you think I is? I had to ask he if he think he did talking to one o’ he pissy li’l friends!”
Conclusion Returning to my larger theme, I want to suggest then that the use of “if”-prefaced repeats as second assessments is derivative of and, as such, dependent upon a more basic use of “if”-prefaced repeats to initiate repair of YNQs and that this more basic use of “if”-prefaced repeats is, in turn, related to the grammar of YNQs in this language variety. This local instantiation of a generic interactional practice then takes its local character from the grammatical features of the variety. Tracking “if”-prefaced repeats across their various contexts of occurrence reveals an apparently language-specific web of connections between quite distinct practices. Of course, similar though not identical practices can be observed in other languages. In French , for instance, turns that other-initiate repair of YNQs may be prefaced by “si”/“if.” In the following case, for instance, N and O are next-door neighbors. At line 1, N compares her living room to O’s, saying that his is bigger. Notice that although the neighbors should have comparable if not equivalent epistemic access to the objects of comparison, N conveys a degree of uncertainty by formulating her remark as “an impression.” When there is no uptake, N pursues a response with “non?” which clearly makes confirmation a relevant next action. O then initiates repair of this YNQ with a turn prefaced by “si”/“if” – a repeat of the talk in line 2. (15) Origami (Ex. 2, from Maheux-Pelletier and Golato 2008 ) 01 N: j’ai l’impression que vot’ salon i have the impression that your (sing.) living room i have the impression that your living room
Language-specific resources 02
il est plus grand it (masc. sing.) is more big (masc. sing.) it is bigger
03
(1.5)
323
04 N: non? no? isn’t it? 05
(2.0)
06 O: s’il est plus grand? if it (masc. sing.) is more big (masc. sing)? if it’s bigger? 07 N: qu’ le nôtre ouais than the (masc. sing.) ours (masc. sing.) yeah than ours yeah
It is interesting to note that while YNQs in French are frequently formed through morpho-syntactic inversion (or the periphrastic construction “Est-ce que . . .”), here the “si”-prefaced turn initiates repair of a question turn not formed in this way. It is not clear at this point how widespread this use of “si” in fact is or whether it can be deployed in other environments such as in response to assessments. In English, as discussed earlier, repair initiations that target YNQs may preserve the inverted syntax of the trouble source turn and in this way convey that it was heard as a question. This of course obviates the need for a particle like “if.” And, indeed, it appears that treating a prior assessment as if it were a YNQ is not a practice available to speakers of English. While it is possible to initiate repair in response to assessments (e.g., A: “That was good.” B: “Good?”), this is a quite distinct practice from the one I have described in this chapter. Although, with a question-intoned repeat (“good?”), a second assessor may agree with a first via an upgraded assessment, there is no sense in which this works by casting that prior assessment as a question and thus as something said with uncertainty. Practices like this in English do not do the same work as the “if”-prefaced responses to assessments that I’ve examined in this chapter. This goes to my argument that within the essentially generic organization of other-initiated repair, language-specific
324
Jack Sidnell
grammatical patterns shape and constrain the relevant practices. Moreover, such language-specific patterns and the practices that they provide for may spill into other areas of talk-in-interaction and provide speakers with unique resources for dealing with a range of interactional contingencies . Acknowledgments Versions of the analysis presented in this chapter were presented at the Meetings of the American Anthropological Association, 2005, Sociolinguistics Symposium XVII, and the Center for Language Interaction and Culture at UCLA, January 23, 2008. For comments on those occasions, I’d like to thank Mary Bucholtz, Ignasi Clemente, Chuck and Candy Goodwin , Kira Hall, John Haviland , Keith Murphy, John Heritage, Emanuel A. Schegloff, and Tanya Stivers . For providing me with examples and insight based on their native intuitions, I thank Joseph Farquarson, Peter Patrick, and Donald Winford. Notes 1 There are no distributional patterns available for American or other forms of English. However, based on several years of collection, I can attest to the greater prevalence of open-class type initiators in adult talk in American, British , and Canadian English. 2 These points are substantiated in various publications including Sidnell 2000, 2005a, 2005b, and 2007a. 3 It is not clear from what is available to whom this question is addressed. 4 The use of “if” to mark what follows as a reported question is noted by Holm and Shilling in The Dictionary of Bahamian English. They include the example: “Mama say if you go?” which they gloss as “She is asking whether you went?” (Holm and Shilling 1982: 109). 5 That is, the size of the party. In other respects Manuel defers to the other participants. Consider, for instance, that, in the final line of this extract, he is asking the others how much it will cost to go in. This suggests that rights to assess are just that and may be tied to specific particulars of the assessable event and not others (i.e. having to do with the cost, the anticipated profits and so on).
Language-specific resources
325
6 Domestic violence – primarily violence perpetrated by men against women and children – is a serious problem in Guyana as in other societies. It is important to note that these men consider such behavior to be reprehensible and cause for complaint. This assessment is, in fact, embedded in the preface to a negatively valenced story about “Buddy’s Father.”
11 Implementing delayed actions Galina B. Bolden
Introduction This chapter describes a language-specific solution to a generic and likely universal interactional issue – how to show that the current utterance is occasioned by and should be understood by reference to something other than the immediately preceding talk. While the default way of connecting an utterance to a prior one is by placing it immediately after the targeted turn (Sacks 1987, 1995; Sacks et al. 1974), on occasion interlocutors find themselves in need of showing that their current turn is noncontiguous with what came just before. One common interactional task then is to connect the current utterance to some early talk, what Sacks refers to as “skipconnecting” (Sacks 1995: II, 349–351, 356–357). Using the methodology of conversation analysis (CA) to examine Russian language conversations, this chapter focuses on one linguistic resource interlocutors can use to manage this interactional problem: the Russian discourse particle -to. An investigation of over sixty hours of recorded interactions between native Russian speakers demonstrates that this particle is deployed in order to index the delayed placement of the action implemented by the current turn-at-talk. The particle is typically placed after a word repeat that helps locate the target of the displaced action in prior talk. The chapter starts with a cross-linguistic overview of several currently documented solutions to the problem of contiguity breaks in talk-in-interaction. Turning attention to Russian, I examine some contexts in which the particle -to is used, including delayed clarification requests and resumptions of previously closed or abandoned
Implementing delayed actions
327
courses of action. Further, I discuss whether the deployment of the particle relates to the topical organization of the unfolding talk or to the activities that are being accomplished through interaction. A number of interactional implications of marking a particular action as delayed are discussed, focusing on action implicativeness and topicality. Overall, the presented analysis contributes to our understanding of how actions are formed to reflect their positioning in ongoing interactional projects. Ways of dealing with contiguity breaks While ordinarily each turn at talk is understood as emerging from the immediately preceding context and as part of the ongoing activity (Sacks et al. 1974), on occasion, this contiguity may be broken. A break in contiguity may involve an interruption of the ongoing activity with a “side sequence” (Jefferson 1972) or a parenthetical comment, a misplaced introduction of a new interactional project (e.g., Schegloff and Sacks 1973), or a need to return to or redo some early part of the activity (e.g., Sacks 1995, vol. II). In these contexts, special techniques are used to convey to the interlocutors how the current turn is to be understood in relation to what has occurred prior to it. An overview of the existing literature suggests that discourse markers, prosodic features, and lexical tie-ins are among the most common devices that do this job crosslinguistically.1 Discourse markers conveying to the addressees that an utterance should not be understood as emerging from an immediately preceding context are commonly referred to as misplacement or disjunction markers. English misplacement markers include “by the way” (Schegloff and Sacks 1973), “hey” (Levinson 1983: 313–315), “listen” (Jefferson 1972: 319; Sidnell 2007b), “look” (Sidnell 2007b), “now” (Aijmer 2002), “okay” (Beach 1993; Filipi and Wales 2003), and “oh” (Bolden 2006; Jefferson 1978), among others. Additionally, several discourse markers can be used specifically to convey the turn’s affinity to some earlier talk:2 such as, English “anyway” (Ferrara 1997; Lenk 1995, 1998; Owen 1985; Takahara 1998), “so” (Bolden 2005, 2009; Howe 1991: 90–93; Jefferson , 1972: 319), “and” (Heritage and Sorjonen 1994; Local 2004), and “but” (Park 1997). A large array of pragmatic markers
328
Galina B. Bolden
that fulfill these functions has been identified in other languages as well (e.g., Mazeland and Huiskes 2001; Mori 1999; Park 1997, 1999; Sadler 2006). There is significant cross-linguistic variation in the placement of these pragmatic markers within the turn (or, more accurately, the “turn constructional unit,” see Sacks et al. 1974). In English (and a number of other languages), misplacement is typically indicated at the beginning of the turn, which allows for an early projection of the turn’s role within the ongoing activity (Heritage 2002; Schegloff 1987b; Sidnell 2007b). Some other languages (such as Japanese and Korean ), however, employ turn-medial and turn-fi nal markers as well. This late placement allows for recasting of the turn’s function in the course of the turn’s construction (cf. Hayashi 2004). Russian pragmatic markers may occur in a variety of turn positions, with the delay marker -to always a post-positioned particle (as discussed below) occurring either turn-medially or turn-finally. Aside from discourse markers, intonation appears to play a role in marking the turn’s place in the ongoing activity. For instance, variations in pitch, loudness, and tempo as well as a number of articulatory features may be used to show that a new course of action is being initiated (Couper-Kuhlen 2004; Levinson 1983: 313–315) or that some earlier line of talk is being continued (Local 2004). Additionally, lexical tie-ins, such as repetition or recycling of prior utterances, are robust ways of showing a connection between a particular turn and some earlier talk. Repetition may be used as a way of restarting a course of action (e.g., Jefferson 1972; Sacks 1995) as well as resuming a line of talk after an interruption (Jefferson 1972) or after a parenthetical insert (Wong 2000). All of these cues – discourse connectives, prosody , and lexical tie-ins – may be deployed in combination to clearly demarcate the structure of the unfolding talk (e.g., Local 2004). The Russian discourse particle -to This chapter describes one way in which delayed actions may be implemented in Russian talk-in-interaction: via the particle -to (pronounced as ta/[tә]). This particle has been the focus of considerable attention from Russian-language scholars due to its somewhat puzzling properties. One of the very few post-positioned particles in
Implementing delayed actions
329
Russian, it can be attached, it would appear, to almost any word.3 Prior studies have, almost exclusively, examined the particle from an information-processing perspective and focused on its cognitive rather than interactional functions.4 In these studies, the particle has been described as a marker of emphasis, contrast or theme. Thus, it has been found that the particle can be used to stress the importance of a particular word, phrase or statement (Rathmayr 1989; Vasilyeva 1972), to indicate presence of contrasting elements in discourse (Bonnot 1987; McCoy 2001; Vasilyeva 1972), or, alternatively, to mark old or known information or referent (i.e. “theme”) (Bitextin 1994; Bonnot 1987, 1990; McCoy 2001; Rathmayr 1989; Vasilyeva 1972). The last two descriptors have been argued to relate to the particle’s etymological roots in the Russian indexical pronoun tot “that” (Bonnot 1987, 1990; McCoy 2001; Vasilyeva 1972). While this research has significantly advanced our understanding of the particle’s functions, several factors have limited its scope. The researchers have almost exclusively relied on literary or invented examples in their analysis, and when natural speech was used, sentences (or very short segments) were usually examined in isolation from their interactional context. As a result, little could be concluded about the interactional functions of the particle aside from the general descriptors mentioned above. Specifically, even if we know that -to marks old information (or theme), the question remains why, in a particular context, the speaker would choose to mark something as “old information” given that such marking is optional (and rarely used). What interactional ends does this marking achieve? What are the constraints on the particle’s use? These questions can only be answered when interlocutors’ own orientations and understandings of the unfolding interaction are examined, and for that actual recorded talk-in-interaction is needed. My research into this particle’s use in casual conversation has suggested that it may be deployed strategically to achieve a variety of interactional ends that go beyond information-processing issues, such as to show the speaker’s concern for or interest in the addressee (Bolden 2003, 2005, 2008) or to propose a particular, interactionally driven, understanding of the relationship between the current utterance and prior talk (Bolden 2005). In this chapter, I focus on the discourse connective function of the particle (i.e. its
330
Galina B. Bolden
role in implementing delayed actions) and aim to specify the contexts of its deployment (and nondeployment). Comparison of the environments where the particle is used with those where it could have been used but wasn’t allows for a clearer characterization of the particle’s functions. Contexts of -to use The Russian particle -to can be deployed in a variety of environments. Of particular interest here are utterances that accomplish sequence-initial actions (such as questions, requests or invitation s) that require the addressee to respond in a particular way (with an answer, acceptance or rejection, for example). In CA, these are known as first-pair parts of adjacency pair sequences (Sacks et al. 1974). As we will see, when the particle -to is deployed in such sequence-initiating turns, it suggests that the action implemented by the turn is, in some way, delayed. In other words, -to-marking indicates that the action launched by the current turn constructional unit is not, in the first instance, connected to the action embodied in the immediately preceding utterance, but to an earlier one. The location of the proper target of the current turn is typically indicated via a word repeat to which the particle attaches. The distance between the -to-marked utterance and the course of action to which it returns (and which it advances) varies. On one end of the spectrum are -to-marked first pair parts that are only minimally removed from the utterance they return to (for example, less than one turn constructional unit away from the target). On the other end of the spectrum are -to-marked turns that return to activities that are several sequences or topics away. 5 Several of these possibilities are discussed below. Delayed clarification requests in storytelling episodes One common environment for delayed actions is in storytellings. A story recipient may request a clarification of a prior part of the telling after the narrator has shifted to a new episode in the telling. These delayed clarification requests , formatted as interrogatives, typically contain a repeat of some element of the targeted part of the telling followed by the particle -to, thereby locating the
Implementing delayed actions
331
t arget of the clarification request. In this environment, the return to the prior narrative episode is typically short-lived: As soon as the question is addressed, the narrator returns to the topic that was suspended. In Extract 1, a -to-marked utterance (at line 13) functions as a clarification question that invites the narrator to explain something about an earlier part of the telling. After the question is answered, the narrator reintroduces the topic that was temporarily abandoned. (See Appendix for Russian -specific transcription conventions.) (1) Music Hall (from Kitaigorodskaia and Rosanova 1994).6 Tamara, a former ballerina, is apparently leafing through a photo-album and talking about the shows she performed in. Roza is an eight-year-old child. Two other people are also present. (6:30) 01 R:
A eta shto t^o?zhe spektakl’/ PRT that what also play This is also a play?
02 G:
Da/= yes
03 T:
=Da: >detachka/=eta b- ((squeek)) etat< nazyvalsja yes child this this called Yes child/ this- this one was called
04
“Leningr^ad pad dazhdë:m.”/ Leningrad under rain “Leningrad in the rain”
05 T: .h >Ty panima,esh vot eta Gastinyj dv^or tut byl/ you understand PRT that STORE-NAME here was You see Gostinyj Dvor ((a shopping center)) was here 06
.h Patom on [nathen he on then he
07 R:
[(Prjama) na sce?ne/ right on stage Right on the stage?
08 T:
a na avansce^:ne .h shël nastajaschij dosch./ PRT on proscenium went real rain and on the proscenium there was real rain
09
(1.5)
332
Galina B. Bolden
10 T: B^yla sdelana tak/ eh: tam byli palozhen brez^ent was made so there were put tarp It was made that way/ They put tarp there 11
>i prosta lilsja dosch./ and simply poured rain and it was simply pouring/
12
=My< ta byli [vse v klewe PRT were all in pla(stic) We were all wearing pla(stic)
13 R: →
[A kak zhe dosch ta tak/ PRT how PRT rain PRT so How did they make rain?
14 T:
↑A vot tak vot/ vada./ PRT PRT so PRT water Like this/ water
15
(0.5)
16 T:
Tak puskali ↓vod[u./ so let-out water They poured water
17 (R):
[(Mm)
18 T:
A my: byli vse v kle↓ë:nke./ PRT we were all in plastic-cloth We were all wearing plastic cloth
19 T:
.h I dvizh^enija on takie d- em-pastavil,/= and movements he such choreographed And he choreographed movements that
Focusing on line 13, we can see that the particle -to is used in a turn of talk that functions as a clarification request, asking the recipient to expand on an earlier part of the narrative. More precisely, the question targets the rather startling announcement at line 8 (that there was “nastajaschij dosch”/“real rain” on the stage of the theater), which is then further upgraded at line 11 (“prosta lilsja dosch”/“it was pouring”) when no recipient uptake is forthcoming (line 9). Note that Roza’s question is delayed relative to the target utterances and is placed after Tamara has initiated a shift to a new episode (about what the actors were wearing – at line 12). The particle is attached to the noun “dosch”/“rain” – the repeated word from the targeted turn in the earlier part of the narrative – and serves to mark the current utterance as delayed
Implementing delayed actions
333
relative to where it should have been more appropriately launched, i.e. immediately following its target. After Tamara responds to the question (lines 14–16), she reintroduces the issue (line 18) that was interrupted by the -to-marked question. While in the above case, the -to-marked turn implements an action that was only slightly delayed vis-à-vis its target, the particle -to may operate over larger distances as well. For example, in Extract 2, the -to-marked clarification request is several turn constructional units removed from the trouble source . (2) Hunting (from Kitaigorodskaia and Rosanova 1994). Alex is telling a long story about his hunting trip to several family members. He has just told the others about how he spent the night in the forest. (5:25) 01 A: I patom pashël,/ vrode ja prasnulsja,/ and then went perhaps I woke-up And then I went/ perhaps I woke up/ 02 rana gluxari zapeli/ eschë gdeta nachala early woodgrouses sang only somewhere early Woodgrouses started singing early/ only after 03
tret’eva/ uzhe uslyshal slyshu sch:ëlk/ three already heard hear snap two o’clock I already heard hear a snap
04
(1.0)
05 A: °Duma° “Ux ty gospadi/ ne utra zh eschë.”/ think wow you god not morning PRT yet I’m thinking “Wow gosh/ it’s not morning yet” 06
(0.2)
07 A: .h Ja eschë dumal minut tricat’ sorak u menja est’/ I still thought minutes thirty forty with me is I thought I still had thirty forty minutes 08 Slyshu schelcho,k/ i patom tishina./ hear snap and then silence I hear a snap/ and then silence 09
(0.2)
10 A: °Dumaju nu i° (.) prisni,las’/ shto, li/ think PRT and dreamed that PRT I’m thinking maybe I dreamed it up 11
(0.2)
334
Galina B. Bolden
12 A:
Nu skarej kastër zatushi,l/ vrode nedaleko,/ PRT faster fire extinguished perhaps not-far Right away I put the fire out, it seems close by
13
(0.5)
14
tk
A:
15
(0.2)
↑Stal padxadit’/ 16 A: started come I started to come closer 17
(0.5)
18 B: →
A kto schëlknul ta?/ PRT who snapped PRT Who made the snap?
19 A:
[Gluxa:r’/ woodgrouse A woodgrouse/
[Vot tak|oj ] ((two finger snaps)) PRT such Like this ((two finger snaps))
[(A:?)/ Ani zhe [sch[ëlkajut/ 20 V: they PRT snap They do make snaps 21 B: 22
[A:/ da/ oh yes
|Hm-hm/]
(0.5)
Vot takoj zvuk/ ((finger snap)) 23 A: PRT such sound It sounds like this ((finger snap)) 24
(1.0)
25 A:
I vsë/ i tishina/ and all and silence And that’s it/ Silence
26
(1.5)
27 A:
Nu vo:t/ PRT PRT
28
(1.0)
Ja: padxadit’ stal vizhu slyshu vrode on zape:,l/ 29 A: I come began see hear perhaps he sang I started to come closer see hear it seems like it started to sing
Implementing delayed actions
335
Bella’s question in line 18 requests clarification of an earlier part of the story, specifically Alex’s mention of a snap he heard in the forest (in lines 3 and 8). The request contains the particle -to indicating its delayed position. The placement of the particle after the repeat of the problematic word “snap” (in its verb rather than nominal form) serves to locate the trouble source in the prior talk. Note that the recipients had multiple opportunities to request clarification (for example in lines 4, 9, 11, 13, and 15), but apparently waited for the teller to do so himself.7 After the shift to a new part of the telling has been accomplished (at line 16), the opportunity for the teller to explain himself passes, and one of the story recipients prompts the clarification , now from a late position. In a way similar to Extract 1, the -to-marked question suspends the narrative progression to a new part of telling, but after the question is answered (lines 19–25), the narrator returns to the point in the story where it was interdicted by the clarification request (line 29). To summarize, -to-marked questions may be deployed to request a clarification of an earlier part of the ongoing telling. This prevents, for the time being, the teller’s movement to a new episode. Such -to-marked first-pair part actions thus interrupt the progress of the ongoing line of talk. Resuming a prior course of action Aside from clarification requests , other actions may receive -tomarking when a speaker attempts to resume or reopen a course of action that has been closed or abandoned. In Extract 3, for example, a -to-marked question is used to return to an earlier request (to get the intended call recipient to the phone) after several activity shifts. (3) NG1: call 12. Mura is talking to her teenaged grandson. (0:50) 01 MUR: Kto galodnyj/ Rafi galodnyj/ who hungry NAME hungry Who is hungry/ Rafi ((dog)) is hungry 02
(.)
03 MUR:
[.h
336 04 IL:
Galina B. Bolden [I papa galodnyj/=Vs[e galodnye/ and daddy hungry all hungry Daddy is hungry too/ Everybody is hungry
05 MUR:
[I-
06 MUR:
I vse, galodnye/=A mama gde,/ and all hungry PRT mama where Everybody is hungry?/ Where is mama
07
(0.5)
08 MUR:
.hh[h
09 IL: [Mama tut sidit/Smotrit na (karti-) na nas mama here sits looks at pictu- at us Mama is sitting here/looking at the pictu- at us 10
paka my galodnye/ while we hungry while we are hungry
11 MUR:
.hhh hah-heh-heh
12
.h[hh
13 IL: [Dat’ tebe [ma,mu/ give you mama You want to talk to mama? 14 MUR:
[I ty to?zhe galodnyj/da,/ and you also hungry yes You are hungry too right?
Ne:/ ja uzhe kushaju/ 15 IL: not I already eating No I am eating already ((continue about dog’s food, then some pictures, joking)) (3:20) Ty sl^u?shaesh menja/ ili net/h 16 MUR: you listen me or not Are you listening to me or not 17 IL:
Da da/ [slushaju/ yes yes listen Yes yes I am listening
18 MUR: [Ah-hah ah-gah,/ 19 MUR: .HHH Da/ yes 20
(.)
Implementing delayed actions 21 MUR:
N^u ladna/ PRT okay Okay
22
(1.0)
337
23 MU: → tk! Nu a mama ta vabshe e-m-n- (0.2) .h PRT PRT mama PRT generally So can I ask mama 24
mo – ,zhna- mo – ?zhna priglasit’ k telefonu/ can can invite to phone to the phone
25
(2.5)
Na Rus’/ Ty xochesh le?vaju ili pravuju/ 26 IL: take NAME you want left or right Here you go Rus’/ You want the left or the right one 27
(1.8)
28 IL?:
Heh-heh-heh-heh ((or off the phone))
Ja xachu ma:mu/ 29 MUR: I want mama 30
(1.0)
31 IL:
N(h)u i shto/ xatet’ ne vredna/ PRT and what want not hurtful So what/ You can want all you want
32 MUR:
Huh-huh-huh-huh-huh-huh-huh-
In line 6, Mura issues a pre-request (Schegloff 2006) to get Ilya’s mother on the phone. Ilya indicates his mother’s availability, thus, giving a “go-ahead” for the request that is now due from Mura (line 9). In the second-turn constructional unit (lines 9–10), however, Ilya starts joking about what is happening at home. Instead of launching the projected request, Mura laughs, and Ilya offers to pass the phone to his mother (line 13). Rather than replying to the offer, Mura continues to joke with Ilya (line 14), significantly expanding the conversation (omitted from the transcript). Several minutes later, Mura resurrects the request (lines 23–24). Her request is now formulated as a -to-marked interrogative (with -to attached to the repeat of “mama”). The deployment of the particle indicates its connection not to the immediately preceding talk but to a prior interactional project that was abandoned prior to completion.
338
Galina B. Bolden
In the above segment, -to-marking is deployed to reintroduce a pending course of action, but it may also be used to return to an interactional project that could have been considered closed. An example of this can be found in Extract 4 where one party (Zhenya) is questioning the addressee’s purchase of a new car, using a -tomarked utterance to reopen her criticism of the purchase. (4) RP call 14. Dina is on her way out of the house when Zhenya calls. 01 ZH: [Nu |ladna/ PRT okay Okay then 02 DI: [.h |Zhen’ka/ [izvini/ NAME excuse Zhenja I am sorry 03 ZH:
[Ty savsem ischezla/ you completely disappeared You’ve totally disappeared
04 DI:
.h da ty znaesh chëta vchera, mashinu kupili/= PRT you know something yesterday car bought You know yesterday we bought a car
05 ZH:
=D[a/ ja zna]ju yes I know
06 DI: [Celyj den’-] (.) doma ne byla/=A sevodnja entire day home not was PRT today my idëm we go We weren’t home the entire day/ and today we are going 07
.hh eh- abmyvat’ Mishkinu kupch(h)uju(h)/ wash NAME deed-of-purchase to celebrate Misha’s deed of purchase
08 ZH:
Da,/ da/ [da/[( ) yes yes yes
09 DI: [.h [Vot/ubegaem/ ja praspala/ ja usnula/ PRT running I overslept I fell-asleep We are running out/ I overslept/ I fell asleep 10
(.)
11 ZH:
[(Kakuju-) what
Implementing delayed actions
339
12 DI: [Ja utram chëta ochen’ rana I morning something very early For some reason in the morning very early I13 ZH:
Kakuju mashinu vy kupili/ Kamu/ what car you bought whom What car did you buy/ for whom
14 DI:
Mne:/ Ve:n/ me van For me/ A van
15 ZH:
A/=U tebja zh byla [novaja/ vrode by/ oh with you PRT was new seems PRT Oh You had a new car I believe
16 DI:
[t. .h
Nu net/ nu u menja zhe vot eta Santra Nisan/ 17 DI: PRT no PRT with me PRT PRT that Centra Nissan Well no/ I had a Centra Nissan/ 18
Malen’kaja/ Belaya/ small white A small one/ A white one
Nu ana [zh nova[ja/ 19 ZH: PRT it PRT new It was new [.hh [↑Nu my eë atdali↓ Mi:shke/ 20 DI: PRT we it gave NAME We gave it to Misha 21
(.)
22 ZH:
[°A: oh
23 DI:
[A on svaju pradal/ PRT he his sold And he sold his
24
i sdelal daun pejment pod nashu mashinu/ and made down payment for our car and made a down payment for our car
a my emu raznicu zaplatili/
na dom/ heh-[.hh for house for the house
340
Galina B. Bolden
27 ZH:
[Nu eta ja nichevo ne PRT that I nothing not Well I don’t
panim^aj[u/(no vobschem) karoche28 understand but generally shorter understand any of this/ (but generally) simply [↑Nu ja tebe vsë rasskazhu 29 DI: PRT I you all will-tell I’ll tell you all 30
[padrob[↓na/ detailed in detail
31 ZH:
[Nu- PRT
[Karoche gavarja on pradal svoj ve?n shorter speaking he sold his van
shto li/ PRT PRT In other words he sold his van Ne v-u nevo [ne ven/=u nevo dzhip/ 32 DI: not v with him not van with him jeep Not v- He didn’t have a van/ He had a jeep 33 ZH: 34 DI:
Ah-[ha/
35 ZH: 36 DI:
[Dzhip dzhip dzhip jeep jeep jeep [Prada,l/ sold
Ah-ha/ [Ah-
37 ZH:
[A::/ Zhalka/ oh sorry Oh that’s too bad
38 DI:
N^u emu ne zhalka/ PRT him not sorry He’s not sorry
39
=on garit on nachal u nevo sypaca/ he says he began with him fall He says it started to fall apart
40
patixon’[ku/ slowly
41 ZH:
[A/ Pa[njatna/ oh understood Oh I see
Implementing delayed actions
341
42 DI: [.hh Vot/= PRT Jasna/=[a chevo< 43 ZH: → clear and what [I on ↑xarasho ochen’ vygadna pradal/ and he well very profitably sold And he sold it very profitably
44 DI: 45
my dazhe ne azhidali/ we even not expected we didn’t even expect it
46 ZH: →
A zachem tebe ven ta/ PRT why you van PRT What do you need a van for
47
(0.2)
48 DI: ((end of tape))
Nu nam n^u:zhen/ On-mne takoj ven ↓zdaravenyj PRT us needed it not such van huge We need one / It’s not so huge
In the beginning of this segment, two competing action trajectories are under way. Dina is attempting to close the conversation (line 2) while Zhenya complains about having not heard from Dina (line 3). Dina’s response to the complaint incorporates a newstelling – she was busy buying a new car (line 4) – as she attempts to shift the conversation back to the conversation closing trajectory (see line 9). In line 13, however, Zhenya poses a question meant to expand the conversation rather than to close it.8 The question consists of two parts: “What kind of car? For whom?” Dina’s response in line 14 contains two corresponding parts (the car is for Dina; it’s a van). These two parts become the objects of Zhenya’s criticism in the subsequently unfolding sequence. In lines 15, Zhenya challenges the necessity of the new car for Dina. After a prolonged discussion that gradually moves away from Zhenya’s initial challenge, she renews the challenge in line 46 by addressing the second part of Dina’s response (the car is a van). This time, Zhenya’s objection is in the form of a -to-marked question, with the particle attached to the word repeat “van” (from line 14). The question in line 46 advances the course of action embodied not in the immediately preceding talk but earlier in the conversation. The immediately preceding talk deals with the details of the purchase, a matter (which may be treated as being) only
342
Galina B. Bolden
marginally related to the challenge introduced in line 15. In fact, the question is unresponsive to the immediately prior talk, sequentially deleting (Schegloff 1987a) Dina’s assessment of the situation in lines 44–45. Thus, the particle -to functions to mark both the sequential displacement of the question from the line of talk it advances and the break of continuity with regards to the immediately preceding activity. To summarize, the particle -to can be used in first pair parts that accomplish actions delayed by reference to the ongoing i nteractional projects, such as story tellings, requests or criticisms. We have observed that -to-marked first-pair parts are not directly responsive to the immediately preceding utterances but link back to something that occurred earlier in the conversation. The use of this delay marker is thus indicative of the speaker’s understanding of the relationship between the action implemented by the current turn constructional unit and the conversation so far. In what follows, I discuss the nature of the connection between the -to-marked utterance and the line of talk it is meant to advance. Old topics or delayed actions? In the analysis presented up to this point, I have used the terms “course of action,” “topic ,” or “line of talk” more or less interchangeably to describe the connection between the -to-marked utterance and the prior talk it links to. We have seen that -tomarked turn constructional units connect to what might be described as a pre-prior topic (something that was talked about before a topical shift) or a pre-prior course of action (an activity that was pursued before a shift). We will see, however, that these characterizations are not interchangeable and that a difference between “topic” and “action trajectory” is crucial in describing the particle’s usage. The notion of topic has been subject to numerous analyses and critical evaluations by discourse analysts of many persuasions (for a review see, for example, Brown and Yule 1983). The most significant analytic problems associated with the notion concern difficulties in determining what the topic of a particular segment of discourse is and how shifts from one topic to another
Implementing delayed actions
343
are achieved (e.g., Holt and Drew 2005; Schegloff 1990). In spite of the difficulties one faces when dealing with the notion of “topic,” the idea that topicality is central to our understanding of discourse structure persists. Thus, topicality continues to be the major resource by reference to which many discourse-level phenomena – such as coherence, discourse markers, reference, and the like – are analyzed. CA has dealt with the notion of topic quite differently. First, instead of focusing on determining what the topic of a stretch of interaction might be, conversation analysts have investigated practices for doing topic talk (i.e. ways of talking on a topic) and for managing topic transitions (Button and Casey 1984, 1985, 1988/9; Drew and Holt 1998; Holt and Drew 2005; Maynard 1980; Sacks 1995). Moreover, within CA, it has been argued that topic may not be the central concept that organizes talkin-interaction; rather, topic talk is one kind of activity, among many others, that interlocutors may conduct in talk-in-interaction (Schegloff 1990, 2006). For example, Schegloff (1990) argues that analysts’ preoccupation with what discourse is about distracts from or obscures what participants are doing in their talk. Since talk-in-interaction is a fundamental way in which social actions are accomplished, Schegloff proposes that coherence of a stretch of t alk-in-interaction (and thus its structure) is better analyzed by reference to the organization of action. The analysis of the particle -to’s usage contributes to the debate over the status of topic as an analytic tool for studying talkin‑interaction. My examination of data suggests that participants’ use of this particle shows their orientation not to what they talk about (i.e. the topic of the conversation) but, in the first place, to what they do in or with the talk (i.e. the action organization). In other words, the deployment of the particle is action rather than topic-sensitive. In my data, the particle is not used on utterances that simply reference an old topic while launching a new course of action. On the other hand, the particle is deployed on utterances that advance old courses of action even when they are, seemingly, topically disjunctive. To illustrate the first case, let us return to Extract 4. Here we find an example of an apparently delayed question that does not contain the particle -to (line 13).
344
Galina B. Bolden
(5) Extracted from segment (4) (RP call 14) 01 ZH:
[Nu |ladna/ PRT okay Okay then
[.h |Zhen’ka/ [izvini/ 02 DI: NAME excuse Zhenja I am sorry 03 ZH:
[Ty savsem ischezla/ you completely disappeared You’ve totally disappeared
04 DI:
.h da ty znaesh chëta vchera, mashinu kupili/= PRT you know something yesterday car bought You know yesterday we bought a car
05 ZH:
=D[a/ ja zna]ju yes I know
06 DI:
[Celyj den’-] (.) doma ne byla/=A sevodnja entire day home not was PRT today
my idëm we go We weren’t home the entire day/ and today we are going 07
.hh eh- abmyvat’ Mishkinu kupch(h)uju(h)/ wash NAME deed-of-purchase to celebrate Misha’s deed of purchase
08 ZH:
Da,/ da/ [da/[( ) yes yes yes
[.h [Vot/ubegaem/ ja praspala/ ja usnula/ 09 DI: PRT running I overslept I fell-asleep We are running out/ I overslept/ I fell asleep 10
(.)
11 ZH:
[(Kakuju-) what
[Ja utram chëta ochen’ rana12 DI: I morning something very early For some reason in the morning very early I13 ZH: →
Kakuju mashinu vy kupili/ Kamu/ what car you bought whom What car did you buy/ for whom
Implementing delayed actions 14 DI:
Mne:/ Ve:n/ me van For me/ A van
15 ZH:
A/=U tebja zh byla [novaja/ vrode by/ oh with you PRT was new seems PRT Oh You had a new car I believe
345
((several lines omitted)) [I on ↑xarasho ochen’ vygadna pradal/ 44 DI: and he well very profitably sold And he sold it very profitably 45
my dazhe ne azhidali/ we even not expected we didn’t even expect it
46 ZH:
A zachem tebe ven ta/ PRT why you van PRT What do you need a van for
47
(0.2)
48 DI:
Nu nam n^u:zhen/ On-mne takoj ven ↓zdaravenyj PRT us needed it not such van huge We need one / It’s not so huge
Focusing on line 13, we can note that Zhenya’s unmarked question appears in an “environment of possible relevant occurrence” (Schegloff , 1993, 1996b) for -to-marking. First, the question refers to the topic (Dina’s purchase of a new car) previously raised (see line 4) and then moved away from (lines 6–12). In fact, Zhenya uses the words “kupili mashinu” (“bought a car”) from Dina’s turn in line 4 – a common practice for resuming a prior topic. Furthermore, the question blocks the new topic (and trajectory of action) Dina embarks on (i.e. terminating the conversation). Additionally, the question in line 13 is the second attempt to launch the question (which makes this attempt delayed): In line 11, Zhenya apparently attempts to pose the same question but abandons it due to overlap. (Oftentimes, subsequent attempts at raising a question, especially those that resulted from overlap, are marked with -to to indicate that the question is mispositioned or delayed.) So, why is the question in line 13 not -to-marked while the question in line 46 is (as seen earlier)? It appears that in order to understand the difference, we need to look not only at what each question
346
Galina B. Bolden
is “about,” but what each one is “doing” in its particular sequential and action environment. From the action-trajectory perspective, these two questions function quite differently. While both block the preceding activity, the unmarked question in line 13 launches a new, different course of action while the marked question in line 46 advances an old one. Specifically, note that when Dina first introduces the car purchase (line 4), it is done as an account in response to Zhenya’s complaint in line 3. The unmarked question in line 13 takes a very different stance with regards to this prior sequence. Instead of being part of the complaint sequence, it initiates topic talk about Dina’s car – treating the telling embedded in Dina’s earlier account as news. On the other hand, the marked question in line 46 returns to and advances the course of action they were already engaged in (i.e. challenging Dina’s purchase). So, a return to a prior topic alone does not warrant -to-marking. In addition, the question has to advance a previously launched trajectory of action. Additional evidence for the argument that the particle -to is used to mark an utterance as participating in old courses of action rather than reopening old topics comes from cases where the particle is used in turns at talk that are not obviously topically connected. We have seen that the particle -to typically attaches to word repeats that help locate what the current utterance is meant to connect to. Yet, repetition – a common mark of topical contiguity – is not a required feature of -to-marked utterances. As seen in Extract 6 below, the particle may be used on utterances that do not contain word repeats (and thus are not obviously topically linked to something prior) when they advance a course of action that has been closed or otherwise moved away from. (6) PF call 8. Misha and Dima are trying to arrange for Dima to bring his car over for a paint job. The arrangements are complicated by the fact that Misha’s garage door is broken and needs to be fixed. (2:00) 01 DI:
[Da/ To’st’ (esl) ty [sevodnja vecherom ne yes that is if you today evening not Yes/ So if you are not done by tonight
02
sdelaesh to zavtra [tozhe ne tem zanimaeshsja= will-do then tomorrow also not anything deal then tomorrow you won’t deal with
Implementing delayed actions 03 MI:
347 [h h h h h h
04 DI:
=zanimaca ne budesh./= will-deal not will anything either
05 MI:
=[Ah::
06 DI:
=[(ty) budesh pradalzhat’ garazh^om zanimaca_ you will continue garage deal You’ll continue dealing with the garage
Ne:t/ pachemu/ Ja magu destvitel’na magu i 07 MI: No Why I can really can and No/ Why/ Actually I can deal with both 08
garazhom i mash^inoj zanimaca/ garage and car deal the garage and the car/
=Magu utram zavtra do valeb^ola sdelat’/ can morning tomorrow before volleyball do I can finish the garage tomorrow morning before the volleyball 09 garazh/=da/ ved’ t^ozhe tak mozhet byt’/ .hA-eh: do garage yes PRT also so may be right/ It may work out that way 10 Ty mne skazhi vam mashina kagda nuzhna ta9/ vabsche/ you me tell you car when needed PRT generally Tell me when do you need the car in general 11
(0.2)
12 DI:
[Nu:: v panedel’nik utram?/ PRT in Monday morning Well on Monday morning
13 MI:
[.hh hh
14 MI:
D-da/ A do etava ne nuzhna/ da?/ yes and before that not needed right Yes/ Before that you don’t need it right
15
(0.8)
16 DI:
Nu v smysle:[eh:( ) PRT in meaning Well meaning
[Nu v smysle sevodnja vecheram vy 17 MI: RPT in meaning today evening you Meaning tonight you can do
348 18
Galina B. Bolden bez neë abajdëtes’/ da,/ without it do right without it right
19
(.)
20 DI:
Schas adnu sekundu/ now one second Hold on a second
21
(2.5) ((recording interrupted for a moment))
&( )&odnja vecheram ne nuzhna/=Esli zavtra 22 MI: today evening not needed if tomorrow we don’t need it tonight/ If 23 ja eë paluchu utram budet xarasho/ I it receive morning will-be good I get it tomorrow morning that will be fine 24
(0.5)
25 DI:
^Hm-mm/
26 MI:
Nu ja vas pastavlju v izvesnast’/ PRT I you put in knowing So I’ll notify you
27 DI:
X^a[rasho/ good Okay
28 MI:
[.hhh
29 MI: →
eV-ev- Va sko´ ka vy vstaëte ta/ at when you get-up PRT When do you get up
30
(1.0)
31 DI:
My v pr[incipe rana vsta[ëm/ = My]::ne znaju= we in principle early get-up we not know We generally get up early/ I don’t know
32 MI:
[tkl!
[°Rana da°] early right
33
=v devjat’ uzhe_ at nine already we are already up by nine
34 MI:
N^u xarasho/h °Znachit ja-° (1.0) Svjazhemsja/ PRT good means I will-connect Okay/ So I We’ll get in touch/
Implementing delayed actions 35
36
(1.0)
37 DI:
Da/ va skol’ka pridëm my ne znaem/ yes at when come we not know Right/ We don’t know when we are coming in/
349
mozhet byt’ i pozna/ 38 may be and late It can be late
Dima and Misha are trying to set a time for Dima to bring the car over to be painted. By the end of the sequence in line 27, they agree that Dima will bring the car the next day and that Misha will confirm the time by calling. Line 27 closes the sequence that implements the arrangements-making, creating a clear potential for conversation closure (which follows soon after the presented segment). Instead of closing the conversation, however, Misha poses another question (line 29). This question is, at first sight, unrelated to what has transpired so far (e.g., it is not about Dima’s car, Misha’s garage, or their weekend plans). It does not return to any particular topic that has been raised, nor is the particle attached to a word repeat. So why is the question -to-marked? The difference between the notions of “topic” and “action” seems to be, again, in point here. While the question is not “about” anything that has been discussed earlier, it is clearly related to the activity the parties are engaged in – making arrangements (in particular, the time Misha can get in touch with Dima in the morning or the time he can expect to get Dima’s car). When we take into account the course of action in which the question participates, the -to-marking does not appear unusual as it extends, from a somewhat delayed position (after the closure of the activity), the course of action the parties were engaged in. To summarize, we have examined a segment of interaction where a question about an old topic served to advance a new action (Extract 5) and a segment where a question about a seemingly new topic served as a vehicle for reopening an old course of action (Extract 6). The differential use of the particle -to in the two cases argues for characterizing the interactional job this particle does as dealing with contiguity breaks across activities rather than across topics (unless, of course, topic talk is the activity participants are
350
Galina B. Bolden
engaged in). This finding supports prior conversation analytic work that shows that talk-in-interaction is organized not by reference to topics but by reference to courses of action implemented in talk, with topic talk being but one such form of action. Conclusion This chapter has examined a particular way of implementing delayed actions in Russian conversation. We have seen that the particle -to may be deployed to demonstrate that the action of the current turn is delayed. Some of the contexts of -to use include delayed clarification requests and resumptions of previously closed or abandoned action trajectories. The analysis suggests that this discourse particle helps parse the ongoing stream of interaction by instructing the interlocutor to understand the current turn by reference to some earlier, not immediately preceding talk.10 We have also seen that interlocutors follow the ongoing conversation not in terms of topics that are being discussed but in terms of courses of action that are being carried out. This orientation to action (rather than topic ) is visible in the differential usage of the particle -to in turns of talk that are meant to implement delayed actions rather than simply resurrect old referents or topics. Prior conversation analytic research has described a variety of other discourse-level phenomena, including discourse markers, in terms of their participation in action trajectories rather than topics. For example, Heritage and Sorjonen (1994) have shown that “and”prefacing connects sequences, that may be topically disjointed, into larger courses of action. Raymond (2004) has described the use of the free-standing discourse marker “so” for prompting sequentially relevant next actions, no matter what their topical focus might be. In other languages, Mazeland and Huiskes (2001), for example, have found that the Dutch equivalent of the connective “but” is employed with turns that resume previously abandoned courses of action. The work presented here extends this line of research to the study of Russian and provides additional support for the (universal ) relevance of action for the organization of talk-in-interaction. As we have seen, the Russian particle of delay -to is a postpositioned marker that occurs in turn-medial or turn-final positions. This placement provides for a retroactive characterization
Implementing delayed actions
351
(or re-characterization) of the current utterance as advancing an old course of action, which is quite different from many other pragmatic markers (in English, for example) that pre-characterize the relationship of the upcoming utterance to the prior talk. While in principle such late characterization could create problems in turn transition (cf. Fox et al. 1996; Hayashi 2003), as the recipient waits for the post-positioned marker to clarify the utterance’s relationship to prior talk, my examination of the data has provided little evidence for turn transition problems (such as delays) beyond those that fall within the operations of sequence and prefer ence organizations (e.g., Schegloff 2006). Recipients’ ability to respond in a timely manner may have to do with the fact that much of what the utterance does may already be available through its other compositional features (especially, repeats11), with the particle being an additional – but perhaps not crucial – resource for dealing with contiguity breaks.12 Acknowledgments I would like to thank Manny Schegloff, Olga Yokoyama, Chuck Goodwin, and John Heritage for their invaluable support and insightful comments on earlier versions of this manuscript. Appendix The transcripts are based on conversation analytic transcription conventions developed by Gail Jefferson . Instead of the standard ways in which unit-boundary intonation is transcribed in English, the following modifications to the conventions are made to account for the particulars of Russian intonation: , ? placed after the syllable carrying the distinct intonation contour (comma or questioning intonation) that will be actualized at the unit boundary. / unit boundary. If no intonation symbol (such as , ?) is placed in the preceding unit, it marks default, somewhat falling pitch contour. ./ marks final pitch drop that is larger that the default, unmarked pitch drop.
352
Galina B. Bolden
Additional intonation symbols w^ord marks a distinct pitch peak on the following syllable or vowel (higher than underline and shorter than ↑arrow) . wˇord marks a dip in pitch on the following stressed vowel (as opposed to the common rise on the stressed vowel) The chapter uses a unique transliteration system that was designed to transcribe Russian language materials in the conversation analytic tradition. Table 11.1 shows the correspondences between the employed system (in the “Transcript” columns) and other standard ways of representing Russian: Russian Cyrillic alphabet, Library of Congress transliteration symbols (without diacritics), and standard international phonetic alphabet (IPA) symbols. The first line of the transcript represents Russian data using the conventions shown in Table 11.1. The second line is a word-for-word translation into English (PRT stands for “particle”). The third line is idiomatic translation (with minimal information on sound production).
Table 11.1 Transcription symbols Correspondences between Russian Cyrillic alphabet, Library of Congress (LoC) Cyrillic transliteration conventions, the IPA, and the symbols used in the transcripts Cyrillic
LoC
IPA
Transcript
Cyrillic
LoC
а б в г д e ё ж з и й к л м н о
a b v g d e e zh z i i k l m n o
a b v g d jε/ε jә/ә ž z i j k l m n o
a b v g d e ë zh z i j k l m n o
п р
p r
p r
p r
с s т t у u ф f х kh ц ts ч ch ш sh щ shch ъ ы e ь ’ э y ю iu я ia a/o (unstressed) г (dialectical) nonstandard reduced deleted sound(s)
IPA
Transcript
s t u f x ts t∫j ∫ ∫j
s t u f x c ch sh sch ” y ’ e ju ja a
ʃ
i¯ j ε j u j a ә γ
gh ´
Implementing delayed actions
353
Notes 1 There are also ways of showing that the current turn is a continuation of the speaker’s prior turn-constructional unit that was interjected by some intervening talk. The resource is to build the current turn as a grammatical continuation of the prior and, thereby, sequentially delete the intervening talk (e.g., for Korean , see, Kim 1999b; for English, Lerner 1989). 2 A distinction is often made between whether the current turn unproblematically continues an earlier course of action or resumes it in such a way as to deal with the problem. Different techniques may be used depending on the kind of connection displayed by the turn. 3 The particle -to is produced as an unstressed syllable of the preceding word. 4 Heingartner’s (1996) study of the particle -to is from a sociolinguistic perspective. A comprehensive review of several information-processing approaches to discourse relevant to the study of -to and some other Russian particles can be found in McCoy (2001). For a complete review of literature on -to, see Bolden (2005). 5 Additionally, -to-marking may be used to invoke participants’ prior conversations (Bolden 2003, 2005, 2008). 6 This and other similarly marked excerpts are from the audiotape that came with the book by Kitaigorodskaia and Rozanova (1994), but all included segments have been retranscribed. 7 This trajectory might be related to the preference for self-repair over other-initiated repair (Schegloff et al. 1977). 8 A reader may note that this question is delayed and should, therefore, be marked with -to. Why it is left unmarked is discussed in detail below (as Extract 5). 9 The particle-to in this utterance serves to connect the current turn to a prior attempt (not shown) to launch this course of action, which is consistent with the argument developed here. 10 The use of “repeat + particle” as a practice for dealing with contiguity breaks may not be limited to Russian . For example, Korean , a language where post-positioned particles are widely used, may well deploy similar resources for showing resumption (K.-E. Yoon, personal communication, August 2, 2007). 11 Interestingly, in Extract 6, where repetition is not used to locate the target, there is a very substantial delay before the response is provided, a delay that does not seem to be related to the kind of action being performed (i.e. preference organization). 12 Additionally, in many instances, the particle -to occurs turn medially, prior to the possible completion point, which allows for an earlier projection.
part V
Conclusion
12 One perspective on Conversation Analysis: Comparative Perspectives Emanuel A. Schegloff
Preamble The title of this book is Conversation Analysis: Comparative Perspectives, a title which appears to offer a well-defined promissory note about the nature of its contents.1 And so it does. There is an introduction which sets out some central characteristics of conversation-analytic (CA) work, and briefly reviews the history of comparative analysis in anthropology – together with some of the problems confronted in the course of that history. A number of the substantive chapters that follow report work that is comparative in its very nature; most of the authors of chapters in which this is not the case go out of their way to set their respective topics in comparative context – either by including data from other language/ culture settings or by reviewing (some of) the literature which sets their work in a comparative framework. For a readership that is (I suspect) largely drawn from the so-called “social” or “human” sciences – anthropology, linguistics and applied linguistics, communication, psycholinguistics and cognitive science, social psychology, and sociology , this is what one would expect such a volume to provide . . . from its sub-title. But its main title should make relevant as well other dimensions of comparison than the linguistic and cultural ones, and, before settling down to address what is actually in this book, I would like to use my bully pulpit to call to mind other “comparative perspectives” that ought to figure importantly in CA work, or at least be taken into account, even when they do not figure centrally. Like what?
358
Emanuel A. Schegloff
Comparative speech-exchange systems The final section of the turn-taking paper (Sacks et al. 1974: 729–731), it may be recalled, was titled “The Place of Conversation Among the Speech-Exchange Systems.” It began like this: The use of a turn-taking system to preserve one party talking at a time while speaker change recurs, for interactions in which talk is organizationally involved, is not at all unique to conversation. It is massively present for ceremonies, debates, meetings, press conferences, seminars, therapy sessions, interviews, trials etc. All these differ from conversation (and from each other) on a range of other turn-taking parameters, and in the organization by which they achieve the set of parameter values whose presence they organize. Such a sort of comparative investigation of the speech-exchange systems available to members of a single society, conceived of in terms of differential turn-taking systems, has barely been looked into by us. However, certain striking arrangements may be noted, if only to suggest the possible interest of this area. (Sacks et al. 1974: 729)
A footnote at the end of the first of these paragraphs read: “Nor is the feature unique to a particular linguistic or social community. It is evidently exhibited in conversation, meetings, etc. in societies whose languages and systems of social organization differ quite drastically,” with a reference to Albert (1964) and a cross-reference to another footnote about cultural variation which ended thus: Finally, the cross-cultural question, as we understand it, asks how the structures on which we report vary across languages (lexically or syntactically conceived), or language communities, or across social organizations etc. – structures which are thereby cast as more basic ones. That ordering is not at all clear to us. We do find that aspects of turntaking organization may vary in terms of other aspects of the sequential organization of conversation. And, as we suggest in the final section of this paper, there are various turn-taking systems for various speech-exchange systems, e.g., conversation, debate etc. (Sacks et al. 1974: 700, note 8)
The remainder of that final section would almost certainly be worded differently now than it was then, although its substance might not be all that different. But the sentence, “Such a sort of comparative investigation of the speech-exchange systems available to members of a single society, conceived of in terms of differential turn-taking systems, has barely been looked into by us,” is, with a
One perspective on Comparative Perspectives
359
few exceptions, only slightly off base. There has been some work on the turn-taking organization of “traditional” classrooms (e.g., McHoul 1978), and of broadcast news interviews (e.g., Greatbatch 1988; Heritage and Greatbatch 1991), but little if any pursuit of the program of comparative studies suggested in the 1974 paper (but see Drew 2003; Schegloff 1987a: 218–228). As the amount of recorded data has grown over the past several decades, the material is (in principle) in hand for the systematic study of the turntaking organization of meetings, classrooms of different sorts, ceremonies of different sorts, etc. More often than not, however, those working with such data either ignore the issue of its difference from conversation or are apologetic for the difference (with the data then treated as if it were conversational). What is needed is explicit registering of the speech-exchange system instantiated in the data and an exploitation of the opportunity to do the sort of comparative analysis envisioned over thirty years ago. How do the characteristics of turns vary (if they do) in different turn-taking organizations? How are sequence organization or the organization of repair affected, given that they are implemented in or through the turns which are differently organized in different turn-taking organizations? Even more important will be the issues we cannot even conjure up now without having described the different turntaking organizations and without having examined the data comparatively – all without varying the language materials and culture that inhabit the interactions. “Multi-modality” I have put “multi-modality” in quotation marks because the term is at times employed (or “deployed”) as intendedly contrastive with other approaches to the study of interaction, including CA. One consequence is that proponents of “multi-modality ” may treat the results of CA work as having little, if any, bearing on, or utility for, studies of interaction based on video data, at the same time that such analyses may focus on body-behavioral components of the data under examination without technical attention to the talk that is ongoing at the same time. It is, of course, true that much of the early work in CA was done on telephone conversation and co-present interaction with audio‑only
360
Emanuel A. Schegloff
data. In part this was because we lacked the financial support and technical knowledge to exploit the then-primitive resources of video-recording; in part, telephone data appealed because the parties could not see one another, and we could thereby finesse the issue of what we were missing in not having video for co-present interaction. As soon as we could, and with the help of Chuck and Candy Goodwin , we began to rely increasingly on video data. One surprising finding is worth mentioning. We might well have found, once we started working with video data, that much of what we thought we had learned from just audio data had to be thrown out, or at least limited in its applicability to telephonic data. In fact, that did not happen! Although our understanding of how interaction works was greatly enriched by the video data and continues to be greatly enriched by it, very little that we had arrived at had to be discarded. Ironically, currently ongoing work (e.g., Rossano, 2009 ) is more likely to find that early work on video-accessible aspects of interaction (for example, how gaze in interaction works) needs revision for having been insufficiently informed by CA findings than the opposite! All that said, it might be productive to undertake research on interactional practices in a fashion that compared interactions with different affordances, starting with the most basic: audio only (telephones or other voice-only modes of interacting) and audiovideo recording. Some of Gail Jefferson ’s more recent papers have documented her findings with exemplars from both telephonic and video-taped data sources, and these might serve as a model. However, her practice served to show the adequacy of her account for data from both modalities; it would be as informative to have work which showed phenomena or practices which differed systematically in the two modalities. Once we have gained some experience with this sort of comparative study of interaction, we might then explore more radically different data sources; for example, fully synchronous data, in which each party can access the productions of the other(s) with no delay in real time (as in ordinary interaction) compared with asynchronous data, with time lags of various degrees (from long-distance video-conferencing to great time gaps between contributions of different parties).
One perspective on Comparative Perspectives
361
Numbers We know that the number of participants matters to the workings of talk-in-interaction – not least of all to its turn-taking practices, and through them to much else (though not everything – not, for example, to word selection) in interaction. When two are involved (whether persons or parties, cf. Schegloff 1995a), the next speaker is likely to be the one not speaking now; add a third, and there is then an issue of turn allocation/selection; add a fourth and the necessary ingredients for a schism into two conversations are present. Add another two or three, and, if schism does not occur, it is not unlikely that two or three will become the main protagonists with the others becoming audience; or else an agreement may be reached about an informally improvised alternative turn-taking organization – like “going around the table” with each one in turn offering a contribution. But these are all informal characterizations, collected in a career of watching interaction. Where are the studies that focus in a systematic way on what happens when an interaction adds an additional person, and then another, and then loses one, etc.? Holding the pre-present or remaining personnel constant, the conditions are there, it would appear, for compelling comparative inquiry, though there is surely much to be learned from comparing interactions with differing numbers of participants without these special features. It requires only people with training who are prepared to cast a wider comparative net that has heretofore been the case. Age grading The relevance of age to talk-in-interaction is most prominent at either end of the life-span trajectory – how “adults” talk to and interact with “children” and “the elderly” (including age variation within those categories );2 how “children” talk to and interact with “adults” (including age variation within those two categories, as well as differential categorization of recipients as age increases); and how “children” talk to and interact with “children,” and the finer sub-categorizations that may be relevant there (cf. the quickening pace of CA work on “very young children” in Kidwell 2005; Kidwell and Zimmerman 2006, 2007; Lerner and Zimmerman
362
Emanuel A. Schegloff
2003; Lerner et al. forthcoming ; Wootton 1994; Zimmerman 1999). This is at the heart of how the newborn come to be the adults whose interaction has so far monopolized our attention, and comparative analysis will surely be part of the story. 3 So much for what this volume is not about. In what follows, I mean to take up two themes. The first picks up a thread in the Editor’s introduction to the volume, suggesting a possible bearing of CA on a problem with past anthropological accounts of diverse cultures. The thread will lead us from that source to a position recently staked out by Nick Enfield (2007) that brings the issue from the past into the present and is relevant to how comparative CA work is best to be advanced. That discussion will in turn lead to a second: critical and programmatic “takes” on some practical dos and don’ts that, from my point of view, are critical to progress in this area that can advance the larger project in which CA is engaged. Comparative analysis: One CA perspective In his useful introduction to this volume, Jack Sidnell introduces CA in several ways, one of them being by contrast with structuralism , which, he writes (p. 8), essentially undermine[s] the very possibility of comparison. Structuralism, taken to its logical conclusion, reveals the particularity of any system (e.g., of pronouns, or verbal tenses) and the elements of which it is composed. Since each element is defined by its relation to all the others, each element is a unique outcome of the particular system in which it is embedded. The result, as is well-known, are accounts which are wholly hermetic and incapable of being compared to one another.
Although Sidnell is referring here and elsewhere in his introduction to structuralism in linguistics , anthropology too went through such a stage. The anthropological counterpart to linguistic structuralism aimed to examine all elements of a culture for how they figured in the culture as a whole, as its own gestalt. One way of understanding this phase of anthropological theory and analysis is to see in it an effort to free anthropology from the colonial auspices under which it had come to maturity (or adolescence), and make it no longer answerable to the cultural commitments and sociopolitical and institutional arrangements of the several imperial
One perspective on Comparative Perspectives
363
powers. Anthropological and linguistic fieldworkers could now address each culture in its own, autonomous terms. The middle decades of the twentieth century saw a confrontation between what were then termed “emic” vs. “etic” approaches (cf., e.g., Harris 1968). These terms derived from the linguistic contrast between phonemic and phonetic analysis, the former being languageinternal and language-specific differentiations of sound which constituted differences of meaning within a language; the latter being, by contrast, objectively differentiated and standardized systems of discriminable sound that transcended the boundaries of particular languages and allowed them to be compared. If the earlier drift toward treating cultures and their languages as organic wholes freed them from pejorative contrast with Eurocentric cultural stances, the drift back toward etic, cross-cultural and cross-language approaches provided the leverage for critique of societal features that were found distasteful, even abhorrent, by reference to putative ethical and political standards with a claim to universal standing and relevance. What does this capsule disciplinary history have to do with the chapters of this volume? It has to do with one central feature of CA work, and that is its commitment to get at the workings of talkand-other-conduct-in-interaction for the participants – its commitment to ground its claims about what is going on in some strip of interaction, its claims about the practices at work in that strip of interaction, its claims about the organizations of practice that provide at the most fundamental and constitutive level for the integrity and working of interaction (like turn-taking , sequence organization, the organization of repair, etc.) – to ground these claims in the demonstrable orientation and understanding of the parties to the interaction as displayed in their consequent conduct. This is in many ways an “internalist,” emic stance – although (as Sidnell notes) one that does not preclude comparative work but rather invites and structures it. In what is in many respects a precursor to this volume, the crosscultural , conversation-analytically focused Person Reference in Interaction (edited by N. J. Enfield and T. Stivers , 2007), Nick Enfield takes up a rather different position, which is worthy of attention and (in my view) contestation – a view with a backdrop in the brief history I have recounted in the preceding paragraphs. So bear with me while I try to convey the position Enfield puts forward
364
Emanuel A. Schegloff
with supporting text from Celia Kitzinger , and a counter to that position. It will, I promise, have a bearing on the ensuing discussion of what comparative studies in CA should look like . . . and not. Much of the Person Reference volume took up one or both of two themes in the CA treatment of practices for referring to persons. One of these themes was the account in Sacks and Schegloff ( 1979) of the bearing of two prefer ences in selecting a reference term for a person – a “recipient-design” preference for choosing a recognitional reference if possible (one designed to enable its recipient to figure out who that they know is being referred to), and a preference for minimization (using a single reference form, if possible). The other theme was the account in Schegloff ( 1996a) of the key resources for referring to persons in English and other languages, and, in particular, the status of default or unmarked forms for “referring simpliciter” – that is, using a form that does, and is understood by co-participants to be doing, nothing but simply referring to the person it identifies, the contrast being to other terms which can be used to identify and refer to that person but which do something(s) else in addition to referring (see, for example, Stivers ’ contribution to the same volume (2007). Unlike some others among the linguists /anthropologists represented in that volume (e.g., John Haviland 2007: 232, whose view was that there is no referring simpliciter among the Tzotzil with whom he has worked, only “referring dupliciter”), Enfield does find default reference forms doing referring simpliciter and departures from such forms which do additional work in his examination of person reference in Lao , a Southwestern Tai language of Lao s, Thailand and Cambodia [. . .] To refer to a person in conversation, a Lao speaker has many possible alternative formulations to choose from. The complexities of this set of alternatives concern distinctions of social hierarchy , as defined by (classificatory) kinship and other factors that determine relative position of individuals in social structure. (Enfield 2007: 98–99)
Enfield provides exemplars of both “the default option” (referring simpliciter) and departures from it. As for the former (Enfield 2007: 105, emphasis supplied): In informal, familiar, village conversation in Lao , the default way to formulate initial reference to a person is to use the person’s first name prefixed by the form that appropriately denotes the referent’s social position
One perspective on Comparative Perspectives
365
relative to the speaker [. . .] As long as the prefix is of the appropriate level, the “prefix-plus-name” formulation is the default option, and its deployment in interaction will pass without special notice.
Several pages later, Enfield provides an instance of departure from this default option. Here I will omit the details of the instance (the key element of which is that the referred-to person – Daaw – lives abroad and is being looked to for a donation of funds to renovate a local temple) in favor of Enfield ’s summary (2007: 109, emphasis supplied): The formulation of this reference to Daaw, the younger sister and daughter-in-law, is pragmatically marked, signalling that the speaker is doing something more with this utterance than merely establishing reference to this person. The content of the marked formulation provides the information needed to figure out just what this special action is (Stivers 2007). By referring to her own younger sister Daaw as “mother’s younger sister,” Keet both casts herself in a lower-than-normal position (i.e., as niece), and casts the referent in a higher-than-normal position (i.e., as aunt).
Enfield’s account very nicely captures referring simpliciter, on the one hand, and doing “more than just referring” on the other. The passages to which I have called special attention with italics refer to how these different practices of referring will be understood by the parties to the interaction. In the first, “will pass without special notice” refers to passing without special notice by the Laotian coparticipants; and, in the second, “provides the information needed to figure out just what this special action is” refers to providing the interactional interlocutors with the needed information. Here, then, there is a welcome convergence between Enfield’s undertaking and CA’s undertaking; so far, so good. But not quite enough for Enfield . He wants to go beyond what parties to the interaction demonstrably are oriented to (a limitation he correctly attributes to CA and names the “Members-Only‑Filter,” 2007: 113–115), and include whatever an external analyst can point to as “available in the talk” as part of “what is getting done,” even if there is no evidence that the participants are oriented to it – for example, the hierarchical character of the resources available for referring to persons in Lao . He writes (2007: 114): But however we define it, availability alone isn’t enough for the MembersOnly Filter. Even when something is explicitly available in the talk (such
366
Emanuel A. Schegloff
as the hierarchical differentiation of social relations encoded in Lao person-reference title prefixes), if interactants aren’t demonstrably “oriented to” it – that is, displaying some kind of awareness, attention, recognition – it is said to be irrelevant to the organization of the interaction, and therefore to its description and analysis. In the case of person reference, default formulations are said to be doing nothing but referring because they are not explicitly “oriented to” by participants. In this view, it’s not enough that the content in question is merely made available.
I think this is a good characterization of the CA position (at least as I understand it), as long as one thing is made clear, and that is the term “because they are not explicitly ‘oriented to’ by participants.” No one is requiring that the participants say things like “I see that you have used a hierarchical reference term of the sort appropriate to the referent and the occasion;” the requirement is not for “explicitness” in that sense (nor does Enfield suppose otherwise, as have some others). CA’s only requirement is that there be evidence in the ensuing talk that the practice in question has made a difference in the ensuing interaction. If the analyst wants to claim that something is real, it must be demonstrably, analyzably real in its consequences in the interaction. As long as that would satisfy Enfield ’s point about “explicitly ‘oriented to’ by participants,” I can live with the attribution. Enfield does not want to live with that constraint; he wants to go further. What is “further”? In a section titled “What Remains Unseen Depends on Where Your Blind Spot Is,” Enfield writes (2007: 115): In the Lao system of person reference, while overt specification of kinship and other hierarchical social relations are unmarked or default in pragmatic terms, they are overtly marked both formally and semantically. These markings make explicit a person’s hierarchical position relative to others in the social network, an important principle in Lao speakers’ cultural understanding of personhood and society. A members-only filter would reject any claim that speakers are “doing” anything in socialinteractional terms by using these socially hierarchical forms, on the grounds of a lack of “orientation.”
Enfield, on the other hand, proposes to include something like “doing social hierarchy ” in his account of what is getting done in interaction when person A uses the default term for B while talking to C, even though (as noted above) such a reference “will pass without special notice.” For whom then, we may ask, is this “doing
One perspective on Comparative Perspectives
367
social hierarchy ” being done, and who is doing it? It appears, at first glance at least, to be being done for the outside observer, relative to whose more egalitarian system of reference this reference is hierarchical. Do we not have here an echo of the past confrontation between emic and etic inquiry? In support of his position, Enfield invokes work from within the CA community – Celia Kitzinger ’s paper, “Speaking As a Heterosexual: (How) Does Sexuality Matter for Talk-in‑Interaction” (2005a), which he understands to be “an important challenge to this [i.e., CA] stance” (2007: 115). In this paper, Kitzinger has examined a broad range of empirical materials drawn upon in CA writing over the past thirty to forty years, and describes a number of practices of person reference which can be taken to embody and convey the heterosexuality of the person referred to, the speaker of the talk, or its recipient. As summarized in the abstract (Kitzinger 2005a: 221): this article analyzes the conversational practices through which cointeractants, in the course of accomplishing other activities, routinely produce themselves and each other as heterosexual. These practices include heterosexual topic talk and person reference terms: husband and wife; in-law terminology; identification of the other with reference to their spouse; the production of heterosexual “couples”; and the use of locally initial proterms.
The import of these practices is underscored by contrast to the experience of gays and lesbians in interaction who may find (to cite an account of an exchange between an insurance salesman and a prospective client who happens to be a lesbian in Land and Kitzinger 2005 , and recounted by Enfield, 2007: 115–116) that a reference to “self and spouse” is subsequently recast as “your husband,” which in turn gets corrected by her to “It’s not my husband, it’s my wife.” “By contrast,” Enfield writes (2007: 115–116), “heterosexual speakers run little risk of their sexuality being foregrounded when they reveal it to interlocutors in exactly the same simple ways (e.g., gender of names, pronouns, words like wife.)” And then (Enfield 2007: 116), “There is a clear parallel between the apparent invisibility of the heterosexist assumption in English person reference [. . .] and the apparent invisibility of social hierarchy in Lao person-reference kin titles.” Both of them, Enfield seems to be proposing, are as much “things that are getting done in interaction” (my words, not
368
Emanuel A. Schegloff
Enfield’s) as the ones conversation analysts appear to privilege, and he cites the peroration of Kitzinger ’s paper (2007: 116): As analysts, we might want not to take a member’s perspective on this but rather to treat the interactant’s everyday world as problematic . . . [W]e might ask what is happening when nothing special is happening: . . . when presumed ordinary experiences are treated as ordinary – what is happening then, how is that done, and what kind of a world must we be living in that these things run off smoothly? (Elisions are in Enfield ’s text, not Kitzinger ’s ; emphases on “then” and “that” are in Kitzinger ’s text.)
By now, the potential bearing of this issue on comparative inquiry in CA should be apparent, even if not yet fully explicit. In what follows, I offer several considerations that seem to me crucial in each reader’s or prospective-researcher’s addressing this issue.4 First, Kitzinger ’s position is not fully represented by the peroration as Enfield cites it, and its problematicity as a CA stance is acknowledged by her. Here is the whole of the paragraph cited by Enfield , with the elided portions italicized (Kitzinger 2005a: 259): As analysts, we might want not to take a member’s perspective on this but rather to treat the interactant’s everyday world as problematic and to explore how, and in the service of what other actions, their heterosexuality is assumed and deployed. We might notice that heterosexuality is available to them as a resource; we might consider what kind of world they are reflecting and reproducing in their talk. More broadly, we might ask what is happening when nothing special is happening: when the second is in a prefer red relation to the first; when the yes–no question is followed by a yes–no answer; when the recognitional referent is recognized or the nonrecognitional referent is treated as adequate; when the punchline of the story is promptly and properly received and the second story is treated as fitted to the first; when presumed ordinary experiences are treated as ordinary – what is happening then, how is that done, and what kind of world must we be living in that these things run off smoothly?
I would note first that the questions here framed as ones “we might ask” have in fact been asked by conversation analysts and at least partially answered, to wit: • “when the second is in a prefer red relation to the first”: Pomerantz 1984; Sacks 1987; Schegloff 1988, 2007b; • “when the yes–no question is followed by a yes–no answer”: Raymond , 2000, 2003;
One perspective on Comparative Perspectives
369
• “when the recognitional referent is recognized or the nonrecognitional referent is treated as adequate”: Sacks and Schegloff 1979; Schegloff 1996a; • “when the punchline of the story is promptly and properly received and the second story is treated as fitted to the first”: Jefferson 1978; Sacks 1974, 1978, 1995. I know that Kitzinger knows this literature, as do many of the readers of this book, so how are we to understand the paragraph from her paper? How is the embedded, powerful presumption of heterosexuality (and of status hierarchy in the case of Enfield ’s project) different from all these other domains that Kitzinger has proposed as exemplary? If they are different, what is the bearing on how inquiry is to be conducted? We will return to these issues in a moment. Before that, however, it is worth noting that Kitzinger is herself not unaware of the departure that this part of her program of studies (though not others) represents from the otherwise common practices of CA research. Early in the same paper (2005a: 223–224), Kitzinger is projecting the contributions she means this paper to be making to CA: In analyzing the construction of normative heterosexuality in ordinary talk, I am also contributing to CA in two ways. First, a recurrent concern of CA is with the deployment of “membership categories” (e.g., Sacks 1995: Lecture 6) and “person reference forms” (e.g., Schegloff 1996a) in ordinary conversation. Family and kinship terms are among the categories whose use Sacks began to explore – most famously in “the baby cried, the mommy picked it up” discussion (Sacks 1972b) but also at various locations scattered throughout his work, including an interrogation of the use of inferences attached to “wife,” “sister,” and “child” in a counselling call (Sacks 1995: 116) and the (hypothetical) deployment of (heterosexual) “couple” inferences to disguise otherwise stigmatized identities (Sacks 1995: p. 593). The analysis presented here builds on that work and adds to it an exploration of the way in which person reference forms (such as “my husband/wife” or “her mother-in-law”) also make available – at a→ least to a recipient for whom such things matter – the inference of that person’s heterosexuality; an inference not usually overtly oriented to as such by speakers or recipients in the data set under consideration, but that nonetheless shows us, as analysts, the production of normative heterosexuality as an ongoing, situated, practical accomplishment. Second, in focusing on an interactional feature not oriented to as such b→ by the participants themselves, I adopt an unusual analytic strategy
370
Emanuel A. Schegloff
(however, – see Sacks 1995: 175–187) and one I would advocate for those interested in uncovering the kinds of social worlds on which (as I show) the practices and actions of speakers depend, and that they reproduce in their talk. As I have suggested elsewhere (Kitzinger 2000, 2005a, 2005b; Land and Kitzinger 2005), it is precisely the fact that sexist, heterosexist, and racist assumptions are routinely incorporated into everyday conversations without anyone noticing or responding to them as such that constitutes a culture. How is it, for example, that an unquestioned set of mundane heterosexual assumptions regularly surface in talk in which participants do not notice (or orient to) their own heterosexual privilege; and precisely, how does this failure to orient constitute and reconstitute a heterosexist culture? These questions “uncover the practical reasoning through which the taken-for-granted world is accomplished (and resisted) – the resources members have for sustaining a social world in which there are ‘women’ and ‘men,’ ‘heterosexuals’ and ‘homosexuals,’ ‘normal people’ and the rest of us” (Kitzinger, 2000, p. 173). As conversation analysts, we can avoid “going native” – that is, rather than taking for granted the heterosexual kinship system displayed and deployed by social members, we can treat the language through which heterosexuality is displayed with the c→ same “outsiders’” curiosity that has animated the analysis of the subcultural argot of pickpockets (Maurer 1964), drug addicts (Agar 1973), or dance musicians (Becker 1963). We can interrogate it for what it shows us about the local production of a culture.
Note, then, at Arrow A, Kitzinger’s offering “an exploration of the way in which person reference forms (such as ‘my husband/wife’ or ‘her mother-in-law’) also make available – at least to a recipient for whom such things matter – the inference of that person’s heterosexuality.” Who is this “recipient for whom such things matter?” Not the co-participants, not the researcher whose commitment is to inquiry about how interaction is made to work by its participants, for there is no accessible evidence that this has affected the parties to the interaction (it may, of course, have done so – even perhaps profoundly, but this has not been introduced into the interaction). The “recipient for whom such things matter,” it appears, is an outside investigator for whom it matters for some reason . . . yet to be specified. Note next, at Arrow B, Kitzinger’s observation that this undertaking is a departure from ordinary CA research practice, offering as an off-setting authorization one of Sacks’ lectures , from winter 1970. I understand the key to this citation as relevant to Kitzinger ’s project to be as follows. Sacks has been discussing a phone call from one “Estelle” to her friend “Jeanette”; Estelle had driven past
One perspective on Comparative Perspectives
371
Jeanette’s workplace – a large and upscale general store – in the late afternoon and had seen policemen outside, at least one with his gun drawn, and also someone she refers to as “a colored lady,” whom she suspects as having been implicated in “the trouble.” She has called Jeannette either to find out what that was all about, or, suspecting that Jeannette may have had the day off, to report the incident and prompt Jeannette to find out the details. Near the end of his discussion, Sacks proposes the following (1995: II, 184–185): When Estelle interprets the events, she interprets them so as to find how, that the cops were there involves that they were legitimately there. And we can notice that at least nowadays [recall that this was winter 1970] that’s become kind of a distributional phenomenon, i.e., whereas Estelle is able to use the presence of the cops to find what was going on – where the cops belonged there, others might see the same scene with the same parties by reference to that the cops were doing something which they had no business doing. That is, if this took place in a black neighborhood, watched by black people, the “very same scene” would perhaps turn into, for the perception of the parties, an altogether different phenomenon. There are places where the cops can count on the presence of two of their cars to provide for their visible, legitimate presence, such that others will then search the scene to find what the cops might be doing that they should be doing, and, e.g., pick up on that someone is “trying to get into the entrance where the silver is” [as Estelle had conjectured to Jeannette] or that they can imagine a killer is in there, though they of course can’t see into the store. Whereas there are others who will not at all see the events in that way, but, seeing two cops on the scene, may now look to see what kind of bother the cops, by being on the scene, are producing – as compared to what kind of bother they are properly responding to.
How then is this discussion relevant to Kitzinger ’s project? It presents an alternative framing of a scene, not one actually entertained by the parties (or at least not introduced into the talk), that transforms the understanding that the reader (and, originally, the students in the class) may bring to the actually produced talk. 5 Is Kitzinger ’s reframing of the invisibly heterosexist talk as blatantly heterosexist not a similar reframing? If Sacks can do it . . . Finally, at Arrow C, Kitzinger writes “we can treat the language through which heterosexuality is displayed with the same ‘outsiders’ curiosity that has animated the analysis of the subcultural argot of pickpockets” – “Outsiders!!” That’s the key. That explicates the reference at Arrow A to “at least to a recipient for
372
Emanuel A. Schegloff
whom such things matter” – an outsider. That was the basis of Sacks ’ reframing and reimagining the scene from the perspective of people of color – outsiders. And that’s the bridge connecting Enfield ’s argument to Kitzinger . As he writes in a footnote (2007: 116), “One has to be a member of another culture to ‘see’ the social hierarchy so ubiquitous in Lao person reference [. . .] I’m not aware of a stigmatized sub-culture among Lao speakers for which such social hierarchical assumptions are problematic” (italics added) – outsiders. And that is the position occupied by virtually all anthropologists relative to the people and culture that they study – a position from which is made visible aspects of what is going on that may not be visible to the insiders, but who are “recipient[s] for whom such things matter,” and who have the technical and academic authority to make that “mattering” stick. (By contrast, with but a few exceptions, CA work on languages/cultures other than [American] English has been done by native speakers of those languages – in this volume, Bolden , Egbert , Golato , Hakulinen , Hayashi , Heinemann , Lindström , Sorjonen , and Wu . What is both striking and [to me] mysterious is how these colleagues were able to absorb and be trained in CA on a language other than their own, distill the analytic apparatus from the language in which it was presented to them, and then bring it to bear with such effectiveness on interaction in their mother tongue.) So, what is the problem? The problem is that, once one goes outside the world of the participants in interaction, and outside what is demonstrably relevant to their conduct, as a constraint on the terms of inquiry and disciplined accounts, we find ourselves again in a world of competing perspectives, of differing capacities in which “such things matter” underwritten by this social theory or that, this category -membership or that, this political position or that, etc.6 I have taken this matter up at length elsewhere, together with several rounds of exchange with articulate proponents of the more traditional view (Schegloff 1997a, 1998b, 1999a, 1999b; Wetherell 1998; Billig 1999a, 1999b). Perhaps someone has found another solution to this conundrum to replace the stance that has figured centrally and pivotally in CA work: using the terms of the interaction (what the parties are demonstrably – though not necessarily explicitly – oriented to) to constrain disciplined analysis of it ; or perhaps such an alternative is just around the corner.
One perspective on Comparative Perspectives
373
There are, obviously, other worthwhile research programs with other analytic goals, in which this problem does not arise, or in which it takes a different form – in which, indeed, it does not present itself as a problem but as a goal. But for CA comparative analysis, where juxtaposition with other ways of talking, understanding, acting, and interacting are precisely the issue, it seems to me especially important not to allow the terms of one culture, language or set of preoccupations to set the terms for framing another; this is the emic side of the coin. For the etic side of the coin, we look to achieved CA findings about apparently omnipresent organizational issues and contingencies of interaction, and the practices of conduct and organizations of such practices (turn construction and turn-taking , sequence organization, practices of repair, of formulating and referring to persons, places, etc., of action-formation, etc.) which can be formulated in more abstract ways that transcend different particularized embodiments in different languages and cultures but which accommodate their specifications, for the promise of defensibly underwriting comparative analysis (Schegloff 2006).
CA comparative analysis: Some dos and don’ts Introduction: Different orders of comparison Although the title of the book announces that its central theme is the special perspective afforded CA by comparative analysis (and vice versa), “comparative analysis” turns out to mean, or refer to, quite different orientations from chapter to chapter. In several of the chapters, the authors introduce us to what turns out to be pretty much the same practice in two languages. For example, in the chapter by Egbert , Golato, and Robinson addressing materials in German and English, this is a practice by which recipients of a prior turn initiate repair on an indexical reference to an object of some sort in the prior turn, in both languages using the same form for repair initiation as is used for a less specifically targeted repair initiation (“what” and “was denn,” respectively), but (in both languages) delivering it with downward rather than upward prosody . And in the chapter by Hayashi and Yoon, working with Japanese and Korean materials, they describe the
374
Emanuel A. Schegloff
deployment of minimal response tokens after a co-participant has responded to them with a minimal response token. In these studies, we have the same practice in two different languages. In several of the chapters, the authors examine two closely related, but subtly differentiated, practices of talk-in-interaction in the same language and explore what different environments of occurrence, what different stances, or what different interactional outcomes characterize these practices. For Wu , what is involved are two quite similar forms of other-initiation of repair in Mandarin , forms which are used and are understood to implicate different kinds of trouble and can embody differently inflected negative stances by their speaker toward the speaker of the trouble-source turn. Sorjonen and Hakulinen examine several differently implemented ways of responding to, and ostensibly agreeing with, a preceding assessment, while differentiating the second from the prior assessment with which it is agreeing by embodying differing degrees to which the responders approach what is being assessed from the same point of view as the prior assessor. And Heinemann , working with Danish materials (though comparing them with English and German ), examines two practices for “treating a question as inapposite” because what was asked was taken by the recipient to be already known to the asker – practices which adumbrate different sources and positionings of the targeted question. In these studies, then, we have two or more related practices in the same language. Several of the chapters focus on materials from a single language and address themselves to a single practice or a single “operator” whose deployment may vary but which does the same job – the targeted practice being related to specific features of the host vernacular language, and being set in comparative perspective by a literature review or invocation of the readers’ familiarity with the relevant literature on English-language materials. So Lindström focuses on what she calls a “curled ja” in Swedish, treating its prosodic distinctiveness as a criterial feature; it is placed in turn-initial position after a sequence-initiating turn by another, and projects a disprefer red response-to-come, juxtaposing this with similarlypositioned “well” in English conversation. Sidnell explores a usage in other-initiated repair in two creole-speaking Caribbean communities, adapted to the absence of grammatical verb inversion
One perspective on Comparative Perspectives
375
to form “yes”/“no” questions. The practice involves an if-prefaced repeat of the targeted trouble-source turn as a way of showing that the one initiating repair heard that the trouble-source turn was a “yes”/“no” question and is checking that it was heard/understood correctly. Sidnell sets the examination of this practice in the context of a set of repair practices in these Caribbean communities that is otherwise quite like that described for English-language interaction. Finally, Bolden , working with Russian conversational data, shows that the discourse marker -ta serves as a resource for positioning the talk in which it is deployed as disengaged from the just preceding talk it would ordinarily be understood to address, and for showing (by the word it is attached to) what pre-prior talk it is meant to locate as its source, and shows as well how this practice gets strategically deployed or not deployed. So, in these chapters, we have a single practice in a single language, explicating the distinctive affordances or problems of the language involved, the whole phenomenon set in some relationship to other languages, without, however, examining materials in the comparison languages. I have done none of these chapters justice with these brief characterizations! Although varying in ambition and in execution, they all range from solid and worthwhile to remarkably subtle and extraordinary in insight and scope. Virtually all of them invite examination for the ways in which distinctive features of the target language(s) implicate distinctive constraints or affordances when compared with another language (most often, English), but none of them, I think, is quite what one would think of as overtly and distinctively “comparativist” (as compared to “comparative”) in design and realization. The remaining two chapters, by Fox , Wouk , Hayashi , Fincke , Tao , Sorjonen , Laakso, and Hernandez on the one hand (henceforth Fox et al.) and Rossano , Brown, and Levinson , on the other, are clearly meant to be “comparativist” in their very conception. Even a cursory examination of the chapters by Fox et al. and Rossano et al. respectively reveals a striking contrast with the other chapters in this volume, namely, the decisive role of quantitative and statistical representation and treatment of data. Indeed, in the Fox et al. chapter, it is virtually the only form of data presented to the reader. It is to these two contributions to this volume that the rest of this chapter is given over.
376
Emanuel A. Schegloff
Various research trajectories. How does a piece of CA research get done? Although there are many tales, each with its own particulars, most are variants of a few underlying prototypes.7 1. The conversation-analyst notices a bit of conduct – a manner or practice of speaking, or a bit of bodily behavior (a shift of posture, a form of gesture , a redirection of gaze ) – and asks “What is that doing here?,” “What is that being used to do in this context?,” and “What is the evidence for some one or more candidate solutions to these questions?” Someone has started with what might be a building block and has asked what it was being used to build and how to ground candidate solutions in evidence. 2. The conversation-analyst notices what seems transparently to be a recognizable action A – complaining , hedging, joking, disagreeing, etc., – and other participants respond in ways that display that they have indeed understood it as A; and the analyst asks what it is or was about that talk, what practice of talking and other conduct has issued in the constitution of a recognizable Action A and its recognition of it as A by co‑interactants? Here someone has started with a datainternally attested interactional product and asks what were the building blocks and practices of conduct that provided for that interactional outcome? In both of these scenarios, the analyst starts early on checking hunches, guesses, attractive possible solutions, by assembling a collection of instances that readily initially present themselves as additional “cases” or “exemplars” of the object the analyst is tracking, and these serve to support or redirect the analysis, or subvert the initial supposition that there was something orderly here. In both of these scenarios, the paths are supposed by the analyst to be leading to new findings – things that had previously not been “a thing” at all. 3. But another way of starting takes as its point of departure some proposal by another colleague who has been pursuing a line of inquiry of the sorts described in the preceding paragraphs; it seeks to explore that proposal in a different “environment” – a different sequential locus, a different speech-exchange system, or a different language or culture. In the present volume,
One perspective on Comparative Perspectives
377
we have Federico Rossano pursuing a line of inquiry on gaze direction in interaction most associated with work by Adam Kendon (1967, 1973, 1978, 1990), Charles Goodwin (1979, 1980, 1981), and Marjorie Goodwin (1980), collecting data in a different linguistic/cultural setting in Italy and finding results quite different from theirs. These findings are presented in fuller measure in his dissertation, but in the chapter in this volume with co-authors Penelope Brown and Stephen Levinson , the authors focus on one specific aspect of gaze – speaker and recipient gazing or not at one another while a question sequence is being initiated – in three very different linguistic and cultural environments, and ask how variable (if variable at all) this practice is in these decisively contrasting environments, and how are we to understand this. And, increasingly, there is the possibility of using what is systematically known but not yet fully explicated to target and pursue and further specify a phenomenon. So, for example, one could start with a well-theorized and empirically grounded systematics for turn-taking for conversation, and then, from what is known, pursue a further explication of, e.g., types of turnconstructional units (TCUs) (in the same language or in different languages);8 or take the systematics of the organization of repair as the basis for further elaboration and explication of its components, e.g., on varieties of other-initiated repair sequences. Here one starts not from one or more discrete observations about a recognizable action getting done or a describable practice being deployed, but, rather, from an already sketched domain of such actions, or practices, or both; one extends the scope of the analysis to further actions or practices or relations between them, or sets about to explore what is surmised to be a comparable organization of practices in another environment, language, or culture. In this volume, we have the chapter by Fox et al., focusing not on other-initiated repair but on selfinitiated self-repair , and, in particular, on the locus of its initiation, and doing so in seven quite different languages. Our examination of these two chapters provides an opportunity to reflect on the doing of comparative research of the language/ culture variety under the auspices of, and with the resources of, CA. The last phrases are important to stress. In order to be instructive, the discussion that follows will, of necessity, need to
378
Emanuel A. Schegloff
be critical. The work will be analyzed and evaluated from a CA perspective; research practices that are criticized from this perspective may be suitable and robust for other modalities of research. The preoccupation of the paragraphs and pages to follow will be the contingencies of comparative CA research – how to do it, how to do it well and what to avoid, using these two chapters as cases in point. And, of course, what may be good or bad for CA may not be for neighboring communities of analytic practice. A research trajectory for comparative studies in CA What might a preliminary list of instructions for doing comparative CA studies – and for presenting comparative CA studies – look like? How should it start? Here is one try. 1. State explicitly your understanding of what the target phenomenon or practice is, and, if possible and/or relevant, replicate it in its original environment in data of your own (i.e., not in the data sources in which it was originally investigated). 2. Ask whether the same features that constitute the phenomenon or practice can plausibly be expected to hold in the new environment(s) in which you will be examining it/them; if not, say what should be taken as recognition criteria in the new environment. 3. Describe the ways in which the new environment(s) is like, or different from, the environment(s) in which the target of inquiry has previously been examined, and assess whether there is a robust basis for comparability . 4. Specify what makes the phenomenon or practice to be examined of interest, and what is to be gained by pursuing it/them in these different environments. In what follows, I take up the Fox et al. and Rossano et al. chapters with these candidate instructions/criteria in mind. Fox et al., “A Cross-Linguistic Investigation of the Site of Initiation in Same-Turn Self-Repair” I must confess that I found the chapter by Fox et al. quite problematic. As the project on which it is based was vetted, approved, and
One perspective on Comparative Perspectives
379
funded by the Linguistics Panel of the National Science Foundation, I take it that this is a worthy project in one or more varieties of linguistics, whatever frailties beset it (in my view) as a CA undertaking in comparative analysis. Let me state briefly an initial set of problematic points, and then treat at greater length some issues that may be endemic to work of this genre, though not necessarily insuperable. The issues with which I begin are all related to the four elements previously sketched for starting a comparative CA inquiry, beginning with an explicit account of the target of the inquiry. The target of inquiry The chapter starts off problematically. Its first paragraph reads as follows: Same-turn self-repair is the process by which speakers stop an utterance in progress and then abort, recast or redo that utterance. While same-turn self-repair has become a topic of great interest in the past decade, very little has been written on the question of where within a word speakers tend to initiate repair (the main exceptions being Schegloff 1979, and Jasperson 1998). And, to our knowledge, no work has been done on this question from a cross-linguistic perspective.9
To an uninformed reader, or to an informed reader not yet alerted to possible trouble, this may appear straightforward enough, but it is not. Having announced the domain as “same-turn self-repair ,” the specific focus is formulated as “where within a word speakers tend to initiate repair.” This formulation presupposes that sameturn repair is initiated within a word, but, at least for English, this is only partially true. Some same-turn repair is initiated within a word by a cut-off (e.g., glottal or dental stop) or sound stretch; but some same-turn repair is initiated while no sound was being produced with “uh” or “uhm,” or with “y’know”; some is initiated by silence where a next element of the turn was projected to start; some has no specifiable initiation at all. Let me expand on this a bit. What does “the site of initiation of same-turn self-repair ” refer to? Technically (and this is as important a place to be technical as there is!), what is initiated by the initiation of a same-turn selfrepair is what we call “the repair segment.” The repair segment can be brief – no more than the initiation and the repair operation itself,
380
Emanuel A. Schegloff
or it can be extended – most elaborately in some word searches .10 Among the things that can happen between the initiation and the repair itself (if, indeed, there is a repair itself) are a number of other bits of conduct, including many that could themselves have served as repair initiations (as noted in the preceding paragraph). So “initiation of same-turn self-repair ” is not a stand-alone thing you can look for. It is designed to be part of a larger unit in which it is doing a particular job – as the term “initiation” conveys. The form it can take varies; in English, same-turn repair initiation gets done with a glottal or dental stop, one or more sound stretches, silence, “uh” or “uhm,” “y’know,” (all of which occur in the data cited in Fox et al. 1996, two of whose three authors are among the authors of the chapter in this volume) and others, and we do not yet know for English what, if anything, the differential import of these alternative forms of initiation may be. Do they, for example, project different types of repair operation to come? Does the selection between them affect where they are to be launched? For example (as suggested above), if within a word, then cut-off or sound stretch; if not, then silence or “uh(m)”? Wouldn’t one want to know these things about English before launching a cross-linguistic analysis of “where within a word speakers tend to initiate repair”? Once we have registered seriously what “repair initiator” means, we have to check whether the object being studied and coded is actually such a thing – a member of that class. Returning to the earlier work by two of the authors of this paper (cited in the preceding paragraph),11 we find the term being applied to instances of the sort of practice that could be repair initiators, but which are preceded in quite close proximity by other such practices. This means that, although it “could be” a repair initiator, in such a case it is not . . . because the repair segment has already been opened by another, prior repair initiator. Here are some exemplars taken from Fox et al. (1996) (with the numbers as assigned in that text, on the indicated page): (8) B: She said they’re usually harder markers ‘n I said wo::wuhh huhh! ˙hhh I said theh go, I said there’s-* there’s three courses a’ready thet uh(hh)hh//hff A: °Yeh B: I’m no(h)t gunnuh do well i(h)n, (p. 190)
One perspective on Comparative Perspectives
381
(10) K: Plus once he got- (0.8) some* u:m (1.3) he got some battery acid on:(0.2) on his trunk er something (p. 190) (44) K: .hh So I’m going to start just- very simply [with-*] with number one (Fox et al.)
The asterisks are used by Fox et al. to mark what they take to be the repair initiation, but arguably the repair initiation in Extract 8 comes three words earlier, at “go” (after which the TCU is re‑begun); in Extract 10 it comes at “got-” (a cut-off followed by silence); and in (44) it comes at “just-” (a cut-off ). In instances such as these, the talk at – and following – the asterisk is not initiating the repair segment; the initiation occurred earlier. The result is that one’s confidence is subverted that the bits of data coded in the Fox et al. chapter in the present volume were in fact initiating repair, even if in some other environment they could be used to do so. The issue here is: What exactly is the object of inquiry and how is it operationally understood “positionally” (i.e. turning on where it is) – and here this issue arises in the “starter” language or culture, before the issues of comparative inquiry have even been raised. For determining the object of the inquiry, readers need as direct access as is possible to the phenomenon. Another concern about the target of inquiry is compositional. Early in the fourth section of their chapter, the authors write: Repair that is initiated before a word is recognizably complete is thus word-disruptive. During the final sound of a word, speakers can produce either word-disruptive or word-preserving cut-off : in word-preserving cut-off , the cut-off is done softly, and the final segment is not shortened or perturbed in any way, while in word-disruptive cut-off the cut-off is louder, the final segment is typically shortened, and glottalization may be begin during a vowel preceding the final segment in a consonant-final word. (Fox et al., pp. 73–74)
The puzzle here is the phrase “and the final segment is not shortened or perturbed in any way”; if the “cut-off” leaves the word unperturbed, in what sense is this a repair-initiation? And a third concern about the target of inquiry is interactional; nowhere is there any suggestion about how this matters interactionally. What turns on where exactly the repair initiator goes? What turns on it for the parties to the interaction, not for the researchers.
382
Emanuel A. Schegloff
Consider, for example, their Extract 4. They write, “Extract (4) shows initiation in the final sound of the pronoun she; the pronoun is replaced with a full noun phrase (this girl): (4) HG Hyla: and she- this girl’s fixed up on a da- a blind da:te.
Here is the sequence in which it occurs, taken from the original transcript. Hyla is recounting the characters and plot line of the play (The Dark at the Top of the Stairs) that she and Nancy are going to see that evening – a plot line she knows because she had seen (and “loved”) the movie, and Nancy had not. The extract presented here starts well into Hyla’s introduction of several characters central to the plot: HG, 7:23–8:08 Hyla:a→ =[h E]n the gir]l:: .hh is jist learning t’trust people yihknow en starting tih date’n she doesn’t think she’s pretty, .hh en [she- ] Nancy: [ How] old is she, Hyla: tShe:’s I guess she’s like about fifte[en er some]thing,= Nancy: [Uh h u :h, ] Hyla: =.hh En she’s fixed up, (0.4) en she meets this gu:y, .hh a:n’ yihknow en he’s (·) rilly gorgeous’n eez rilly nice en evrythi[:ng bud li]ke= Nancy: [Uh h u :h,] Hyla: =hh He’s ah .hh Hollywood (0.3) s:sta:r’s son yihknow who wz a mista:[ke en they [put im in’n [Academy,] Nancy: [O o this [s o u n d s [so goo:: ]::[d? Hyla: [school, .hh Buh wai:t.=‘n then, .hhm (0.2) .tch en the:(w)- the b→ mother’s ˙hh sister is a real bigot. (·) Nancy: [i – Y a : h ,] Hyla:c→ [ Yihknow en sh]e hates anyone who isn’ a Cathlic=. .hhh d→ a:nd this boy is Jewish. .hh an’ tshe- this girl’s fixed up onna da- a bline da:te.An’ the(g)- en turns out t’be this gu:y.= Nancy: =[Uh hu[:h, Hyla: =[.hhhh[ An’ they goes oh I hear yer of the Jewish faith yihknow so ‘ere’s a whole thing i[n tha[t, .hhhhhh]= Nancy: [O h :[ w_o_:w ]=
One perspective on Comparative Perspectives
383
The target here is at “d” – the replacement of the trouble-source “she” by the repair “this girl.” The nature of the trouble is clear enough: at “a” a female character is introduced – “the girl,” and is subsequently pro-termed with “she;” at “b” another female character is introduced – “the mother’s sister,” and is subsequently protermed (at “c”) with “she”; and so the trouble-source is the “she” at “d,” which is taken by its speaker to be possibly equivocal in its referential target – an equivocality resolved by the repair “this girl.” In asking how the target of inquiry figures interactionally, we are asking about the relevance for the parties of the placement of the initiation of the repair “in the final sound of the pronoun she” – both in its own right and as compared to Extract 2: “I wish I’d had m- a camera earlier today because” and the placement of the initiation of its repair “just after the bilabial nasal closure (which could be heard to begin the word my, replaced by a).”12 One other observation for now: The questions in the preceding paragraphs are not, in the first instance, researchers’ questions. They are meant to be among the issues parties to the conversation are addressed to in producing and parsing talk-in-a-turn, in real time, in real interactional episodes, with real-life consequences; and we researchers do our job by examining as best we can how the parties to the interaction deploy these resources and understand their deployment by others. They are the ones whose conduct defines the identity and the integrity of the phenomena; where are they in this investigation? This brings us to the second section of the chapter – on data collection, the data collected, and the form of data on which the analysis was actually based. Comparability across data There are several problems concerning the data: First, the authors write, “The instances were taken from recordings of naturally occurring speech. The recordings were not made specifically for this project; rather they had been collected by the individual researchers at different times and in different places, for a variety of purposes.” The term “naturally occurring speech” provides us with no idea whether the data were drawn from ordinary conversation, religious ritual, market bargaining, courtship,
384
Emanuel A. Schegloff
e ducational/classroom setting, legal proceedings, etc. (not at all unlikely given the further characterization “collected [. . .] for a variety of purposes”). These can have very different turn-taking organizations, and that can be profoundly consequential for repair. Second, the authors write: In the present study, we further limit our focus to simple repairs, both simple recycling and simple replacement . By “simple” we mean that no other repair operation is involved in the repair. Thus simple recycles are just repetitions of words, with no additions, deletions or replacements; simple replacements are instances in which a morpheme, word or phrase is replaced, without recycling, addition or deletion.
Unhappily, this leaves a great deal unspecified. When the authors write, “Thus simple recycles are just repetitions of words, with no additions, deletions or replacements,” I find myself wondering: Have they allowed for prosodic repair, in which the same words are repeated but with different prosody ? How about resaying the same words, but fixing one whose pronunciation was a bit off? I am not making this up; these are decisions I have had to make on English, a language of which I am a native speaker.13 And when the authors write, “simple replacements are instances in which a morpheme, word or phrase is replaced, without recycling, addition or deletion ,” I find myself wondering about the sort of replacements Jefferson (1974) wrote about (and which I suspect characterize a substantial proportion of at least the English language replacements in the Fox et al. corpus). In them, the “trouble-source item” – for example, the word to be replaced – is cut off before completion but after enough is “out” to recognize what it was going to be, and is then replaced by another word which is brought to full completion. Has not something then been added? And, as in the instances taken up in Jefferson ’s paper, is not something else being done in addition to simple replacement when, in a court of law, the partially articulated disrespectful term for police (“cop”) is stopped midcourse to be replaced by its respectful alternative “officer” (1974: 193): “When thuh ku- officer came up . . .”? Third, whatever the reasons for it, the absence of all the other forms of same-turn repair – insertion, deletion , search, reformatting, etc. – is seriously problematic. One way of motivating such
One perspective on Comparative Perspectives
385
a study in indigenous terms – i.e. for how it concerns practices that may matter to the participants, is to see whether and how the type of repair operation being initiated can be indicated by the form of the repair initiation and its placement. If you leave out most of the repair operations, and do not recognize that repetition / recycling can be the locus for prosodic repair, then you compromise the possibility of finding something that matters . . . to the participants . . . dealing with it on a case-by-case basis, not with significance tests.14 Fourth, this last concern raises another: Did the investigators work from the recorded data or from transcriptions of it? Were they in a position to tell whether a repetition included re-doing of the prosody ? Problems in analysis and findings I limit myself here to tracking one proposed finding as the chapter progresses from initial statement to formulation in the “Conclusions” section. We begin in “Repair Type and Site of Initiation.” Based on earlier work by Jasperson (1998), the authors test the hypothesis that initiation before the word is recognizably complete is associated with repairs that change an element in the preceding talk (for example by replacing it), while initiation when the word is recognizably complete is associated with repairs that initiate repair on upcoming talk, for example by simply delaying next item due (by recycling current word). (Fox et al., p. 74)
The ensuing table and chi-square tests, they remark, “support this claim strongly.” Here, we should note, is a proposal that is embodied in a virtual parenthetical aside, namely: The authors assert that recycling is not repair on the item(s) repeated, but is repair on the upcoming talk, by delaying it (“repairs that initiate repair on upcoming talk, for example by simple delaying next item due [by recycling current word],” emphasis added). There are several reasons to be concerned about this claim. First, while it is literally true that recycling current talk does delay the talk that follows, so do various other practices that occur in the environment of repair, notably such usages as “uh(m),” “y’know”
386
Emanuel A. Schegloff
and silence (to cite only the most common ones), either separately or together. What then is done by recycling distinctively? Second, there is a literature on recycling which shows recycling to be used to do other jobs even if, as is claimed by the authors for their data, there is no other change besides repeated saying. Schegloff ( 1987c) describes its use to get talk that was possibly compromised by implication in overlapping talk by another resaid “in the clear.” And Goodwin (1980) shows recycling as among the resources used by speakers to attract the gaze of nongazing recipients. These recyclings are not doing repair on the next thing said; they are dealing with the possibly compromised hearing of the thing said in interactionally problematic moments (overlap and inattentiveness). Since some recycling is demonstrably oriented to dealing with trouble concerning the repeated talk, what evidence is there that recycling is done just to delay next items? How much of this is related to the fact that the authors (and therefore the readers) are looking at statistics, and not at actual segments of data? Third, what is made of this arguably problematic account of recycling as a repair practice, initially introduced in a parenthetical remark? The next section of their chapter ends like this (emphases in italics and boldface supplied): Before examining the patterns in our other languages, it is worth exploring possible explanations for this association in English. One obvious explanation is the function of each repair type. Recycling in English tends to be used to repeat function words, which we hypothesize is done in order to delay the production of the next content word due (80 percent of all simple recycling in our English data repeat function words). Because their purpose in most cases is exactly to delay production of next item due, there is no hurry in initiating the repair. Replacement, on the other hand, tends to be used to replace content words (61 percent of simple replacements in English replace content words), and may occur in cases where an inappropriate word or pronunciation has been produced. In such cases there may be pressure to initiate repair as early as possible, to catch the potentially inappropriate production (cf. Jasperson , 1998). (Fox et al., p. 75)
This explanation appears to rest on what might be universal motivations. But do other languages actually manifest the same patterns as English? The section “Summary of Cross-Linguistic Comparison” examines how the other languages pattern.
One perspective on Comparative Perspectives
387
What had been a parenthetical aside before has become a hypothesis (italicized above), and, from one sentence to the next, the italicized hypothesis becomes the bolded assertion, and the attributable explanation of other described effects (although qualified by an “in most cases”). And, by the next paragraph, it is made possibly universal ! By the “Discussion” it serves as an explanation (unqualified by “in most cases”): “Indonesian is like English in having function words precede content words, so function words can be used in recycles as they are in English, to delay next item due” (Fox et al., pp. 90–91). And by the “Summary and Conclusions” section, it is a fact (emphasis supplied): We suggested that there may be a tendency for monosyllabic words to be function words, and that function words tend to be recycled to delay next content word due; in such cases – in languages for which function words precede content words – speakers may want a beat (or two) of delay, since one of the main purposes of recycling in these instances is to provide a temporal delay. (Fox et al., pp. 100–101)
The upshot Two observations serve (to my mind) to point to the sources of these troubles. First, it is striking that, apart from four fragments of conversation presented as illustrative exemplars, all the “evidence” in this chapter consists of tables and chi-squares. Second, and not unrelated to the first: In the diction of the later sections of the chapter, the authors are no longer describing what parties to interaction do; the main actors have become the languages, as in, “Indonesian should pattern with Finnish and Bikol, because [. . .];” “But Indonesian patterns with English rather than with Finnish and Bikol in [. . .];” “English is another exception to the findings [. . .]” and so forth. And the supporting actors in this inquiry are syllablicity (uni-, bi-, and multi-), function words and content words, words that are recognizably complete and words before being recognizably complete . . . nowhere are there speakers and recipients, practices, actions, turns and turn-constructional units, repair segments, etc., – things that are interaction-relevant. The conclusion to which this examination leads me is that the Fox et al. chapter is not a work of CA but a work in one of the varieties
388
Emanuel A. Schegloff
of linguistics that find its relationship to the social or interactional to be important to address. Understanding the Fox et al. contribution to be a work in linguistics does not, ipso facto, render the discussion of it in the present chapter irrelevant. Linguists working in this area or related areas will have to sort out for themselves the bearing of these critical observations on an assessment of its results. One focus of such a review should be on the core method s employed in the inquiry, and this brings me back to an observation early in this section. However briefly, some attention must be given to two features that were earlier noted to be distinctive to the Fox et al. and the Rossano et al. chapters: quantitative/statistical analysis as compared to the analysis of single cases and aggregates of them; and “coding ” on the one hand as compared to “analysis” on the other. First, on quantitative and statistical analysis: I have no objection to quantification per se, when, for example, it is employed to show that the results of qualitative analysis of single cases hold across aggregates of cases. It has been important in my own work, and in the work of students and colleagues. It was, for example, a key part of the argument in the first published paper in CA (Schegloff 1968), which turned on the adequacy of a candidate formulation of a finding for 499 out 500 instances, but which could not be reconciled with the 500th case, leading to a more general formulation which provided for the two different surface outcomes as alternative specifications of an underlying set of practices. More recently, Bolden (2006: 664) builds confidence in her finding that “so”-prefaced sequence initiations “are almost always other-attentive (i.e., they concern the addressee), whereas ‘“oh”’-prefaced sequences are selfattentive (i.e., they raise speaker-centered issues)” by displaying this table: Table 12.1 Distribution of “so”- and “oh”-prefaced sequence initiators
Other-attentive action trajectory Self-attentive action trajectory
“So” Preface
“Oh” Preface
88 4
1 65
One perspective on Comparative Perspectives
389
And she subsequently shows the interactional consequences that follow the “cases” in the “mis-match” cells – such as the ironically exaggerated expression of interest by the recipient of a selfattentive sequence initiation started with a “so,” thereby providing evidence that the interactional participants are oriented to just this orderliness; the empirical analysis has explicated their practice, not imposed a researcher’s order. The point, then, is not simply that the CA “qualitative” analysis gets us the candidate phenomenon, which is then “shown” by quantitative analysis to be the case or not. The quantitative analysis serves to provide reassurance that the candidate phenomenon is/was not an isolated, idiosyncratic usage of some local setting (a particular speaker or category of interactants), but has a prima facie robustness. Specifying the phenomenon, showing its variants, showing that the participants are oriented to it, etc. – all return to case-by-case analysis; and when that analysis seems to arrive at another plateau of stable findings, quantitative work may be called upon again to attest to its non-idiosyncraticity. At least for the sort of data CA tries to come to terms with, one does not go to work on a corpus of data to conduct quantitative or statistical analysis and arrive at findings; rather, one works up the data case by case, instance by instance, the results of these analyses serving to compose a corpus which may then be subjected to quantitative analysis to underwrite its robustness.15 And if that effort encounters trouble, then the trouble is dealt with by going back to the individual cases to reanalyze the data. By contrast, when quantitative analysis is taken to be the main road to findings, as in the Fox et al. chapter, when the researchers get a puzzling or striking quantitative result, they form hypotheses and “test” them by further quantitative analysis, so the locus of the findings and the research is shifted from the empirical reality of the individual case to the analytical domain of the aggregate. The robustness and relevance of the findings is established not by returning to the individual cases to see whether or where they work or do not, and how they go awry, but by resorting to tests of statistical significance – chi-square tests in the Fox et al. chapter. Whatever it is that chi-square tests actually test and attest, they presumably underwrite the robustness of the findings in the larger
390
Emanuel A. Schegloff
universe from which the actually analyzed data were drawn and of which the data at hand provide a reliable – or at least calculable – sample. As I have argued elsewhere (Schegloff 1993), I am at a loss to know what that universe is, and which of its parameters a sample should approximate to be deemed adequate (an issue aggravated here by the previously noted diversity of the collectors and projects shaping the data set). When CA papers use the loose language of “overwhelmingly,” “rarely,” etc., they speak to the issue of whether some proposed finding is plausibly to be considered a stylistic quirk of one or several speakers or categories of speaker. That seems to me a viable undertaking and one that readers can assess to decide whether to take the proposed finding seriously, seek to replicate it in the data to which they have access, etc. It remains to be seen what relationship there is, if any, between these goals of what might be termed “loose” or “rough” quantification and the scientific results claimed for conventional statistics, with paradoxically uncertain credibility at best. But where do the data come from that supply the grist for quantitative and statistical mills? The “data” are not the actual things said or done by parties to the interaction; those are the “raw data.” The data for quantitative and statistical analysis are the product of “coding ” the raw data, and that is the other matter needing to be addressed here – what is the relationship between coding and analysis? Second, on coding and analysis: Although an early footnote in the Fox et al. chapter reports on who “analyzed” each language data corpus, it seems clear that the primary product of “analysis” was a coding scheme, and it was the outcome of processing the data via this coding schema on which the chapter is based. Indeed, the authors write, “Each instance of repair was coded for a variety of features. The coding scheme was developed for the larger project, and it included information relating to type of repair and site of initiation” (Fox et al., p. 62), which suggests that the coding scheme was not designed with this work in mind, but the outcome of its application to the data corpora has been exploited for the announced topic . But coding is not analysis. Coding produces the distribution of instances of “values” of pre-selected variables as observed in the raw data; it does not determine (or modify) the variables, how they were formulated, and on behalf of what project they
One perspective on Comparative Perspectives
391
presented themselves as interesting. Above all, it is an instrument for the processing of aggregates of data; it makes the data serve a pre-scribed discipline of inquiry rather than making the inquiry responsive to the observable features of the cases examined one by one, that are then collected into the aggregate. Analysis – at least in conversation analysis – treats each case in its particulars, indeed, is responsible for the determination of what will constitute a “case” or an “instance” of a putative phenomenon in the first place. Each candidate instance of a putative phenomenon has to survive such an examination and can in the course of the examination transform the researcher’s understanding of what the phenomenon is, rather than simply being included in, or excluded from, class-membership status. Here is one account of how this works. Having spelled out how a strip of interaction or some practice of acting seems to have worked, the analyst tries to ground the effort to go beyond a single case by analyzing additional single cases. These may reveal the same way of working, or may allow or require a different account whose development allows (and drives) the reformulation of the analysis in more general terms – terms which subsume the initial analysis and the subsequent one(s) as alternative specifications. And this procedure is reiterated until a formulation of the target of inquiry is arrived at that is adequate to the vast majority of further examined instances. Then, instances that are not adequately and straightforwardly addressed by this formulation are subjected to single case analysis which (if one has gotten it pretty much right) shows that the ways in which that exemplar differs from the arrived at account can be understood by reference to special, describable features of the local context on that occasion, or as doing a specially inflected version of the phenomenon being described, which is itself describable. Of course, at any of the steps of this course of analysis, the process may fail – a next candidate exemplar is unanalyzable, or is analyzable in a fashion that does not allow a more general formulation that yields alternative specifications. Then one does not “have a phenomenon,” though one may have several candidates, none of which is what one started with. But if one does have a phenomenon, then one can go about collecting instances in greater numbers – perhaps exemplifying the several alternative specifications that the analysis
392
Emanuel A. Schegloff
has engendered so as to get some clues to what “conditions” their respective occurrence, clues which will themselves lead back to single case analysis so as to ensure that the result is true not only of the aggregate, but of the single cases that compose it. After all, parties to interaction do not act in the aggregate; they do not behave so as to produce profiles of aggregates of action types; they deal with a/ the moment in interaction, what has led up to it, what they mean to do in or with it, how they mean to advance its course. In the “best case” scenario, one comes upon ideal evidence: data internal evidence that parties meant by some bit of conduct to be doing what the analysis of it claims, and evidence that recipients understood that to be what a prior speaker was doing – each of those displayed in the ensuing conduct in the interaction. Not all phenomena are conducive to such evidence, but it is the gold standard if one can get it. All of this is most likely to happen and to yield results if the analyst has collected and, most importantly, transcribed the data herself or himself. Most of the authors in this volume collected and transcribed the data themselves, thereby achieving (and suffering!) the most intimate exposure and familiarity with data an investigator can have.16 In my experience, the practice and process of coding do not work like this. They are for the most part routine operations: The coders are trained to use designated features of the data to recognize membership in one or another coding category ; their job is not to register new observations or decide that some category should be split into two. The adequacy of the job done will be measured by a statistic that testifies to the success or failure of the coders to arrive at identical assignment of cases to classes. I have tried to do coding in my own research. One might have thought that I, having figured out what things seemed to possibly matter for the way some object of inquiry worked, could create a coding scheme whose users could then use it on multiple instances to determine whether and which of those things was present or absent and did or did not affect the object of inquiry, and relieve me from doing so. But when I sit down to do a bit of that work myself to see if it is ready for others to take over, I find that the elements of the coding scheme represent and evoke a memory in me of what made them relevant, and, looking at the raw data, I
One perspective on Comparative Perspectives
393
find myself in a position to grasp the import of the presence or absence of some “coding” outcome, in a position to register other features of the datum – perhaps previously unnoticed – that do the same job as what has been coded as “missing” or explain that missingness, or I can register that, even if present, it is not doing in this instance the thing it is presumed in the coding scheme to be doing.17 And I know again that I have to do it myself. And I always do it myself. What might a discipline look like if the coding was ordinarily done by experienced investigators? Ones who know the language as a native language? And ones who have become, or at least are becoming, astute analysts of talk-in-interaction? What would then be the relationship between detailed analysis of single cases and quantitative representations of “the data?” Or hasn’t that actually been the general practice in CA? How then should we assess the introduction of quantitative/statistical analysis – with its requirements for coding – into conversation analytic work? The chapter by Rossano , Brown, and Levinson offers a case in point.18 Rossano et al., “Gaze, Questioning, and Culture” The title of this chapter is an accurate guide to what will figure centrally in it. The chapter concerns the deployment of gaze by speakers and recipients in two-person interactions, focused specifically on the questions in question–answer sequences, examining this target in three geographically, culturally, and linguistically different (and remote) social contexts. It is based on extensive prior work by one of the authors (Rossano ) doing detailed qualitative CA analysis on Italian data that has yielded empirically grounded evidence for accounts of gaze deployment that are counter to the preceding literature.19 That work has provided the basis of the hypotheses to be addressed by statistical method s in the chapter being introduced. One reason for turning to these method s is “to understand whether gaze as an interactional practice has universal properties across cultures or not. We cannot make easy judgments about sameness and difference in conversational practices unless we can be sure that the examples we analyze are reasonably representative of interactions in the culture in question” (Rossano et al., p. 188).
394
Emanuel A. Schegloff
As in the Fox et al. chapter, this work exploits data from a larger project (“concerned with the crosslinguistic comparison of question-answer pairs”) to look at “gaze behaviour in a sample of 300 question-answer pairs from each of the three cultures, as used by roughly ten distinct dyads (or c. twenty individuals) from each culture engaged in naturally occurring conversation in informal settings” (Rossano et al., pp. 188–189). Unlike the Fox et al. chapter, this chapter contains brief but detailed accounts of the circumstances in which the data were collected, well beyond the characterization “spoken interaction.” The outcome projected in the chapter’s first section is that . . . this restricted sample is quite enough to show that earlier analyses of the occurrence of gaze in interaction are not general across functions and cultures. The results are also suggestive of more positive general hypotheses, which would need to be followed up by qualitative analysis of a wider range of actions and sequence types. Given the work earlier referred to (Rossano , 2009), it is clear that we cannot expect a full understanding of gaze in interaction without such a wider analysis – question-answer pairs for example constitute a very different sequential environment if compared with storytellings. But we hope that these preliminary results will already serve as a useful orientation for this future work.
The quantitative analysis is, then, depicted from the outset as supportive of further qualitative work, rather than as being the payoff in its own right, except for calling into question the suggestion of universal applicability in some of the prior literature, which it succeeds in doing. So what’s good here is: 1. the undertaking is grounded in preceding detailed CA work; 2. that preceding work has already showed differences from the prior literature; and 3. there is reason to believe that the differences between the preceding CA analysis and the received literature have to do with what analytic features of conversation gaze is organized by reference to, rather than it being language or culture, but that remains to be shown. What is less clear is: 1. that twenty individuals from each culture, and 300 instances of one utterance type (questions, and three specific types of
One perspective on Comparative Perspectives
395
question at that), are representative of a culture, so that statistical analysis is relevant and probative, or, indeed what the universe is for which the data bases on which they rely are representative in a fashion that will support the statistical procedures that they employ; and 2. if not, whether the objectives of this study would not be as achievable by close analysis of a half dozen actual segments from each of the cultural settings instead of being largely composed of quantitative analyses of data that the readers for the most part do not get to see. The second section is a thorough and cogent critical review of the relevant literature which is rather more substantial and reliable than the literature on the topic of the Fox et al. paper. This is important because the finding to be reported stand in sharp contrast to the findings of the previous literature. The third section is key. Although introduced as providing a “few examples that motivate the development of the coding scheme and show a different organization of gaze in interaction” (Rossano et al., p. 193), it does far more than the subservient role that the first clause conveys. Far from facilitating the statistical analysis and the coding scheme that provides grist for its mill, here are the raw data being examined, and a detailed analysis of such data – including two episodes from the Italian data and one each from the Tzeltal and Yélî Dnye data. Here we get to see some actual instances of the things the statistics will be statistics about. And this access makes it possible for us, the readers, to evaluate critically the aptness of the terms of analysis in which the statistics will be reported – a possibility to be exploited a few paragraphs hence. Aside from the substance of the analysis and the sheer value of its presence, is the testimony it provides for the immanence here of analysis, not coding . The transcripts contain not only the conventional (for CA) three-lined-representation of the talk – in the native tongue, a morpheme-by-morpheme gloss and a translation into idiomatic English; they have funny little ovals and arrows (explained in Appendix B, pp. 242–244). These little drawings depict gaze direction and change in gaze direction, located relative to the momentby-moment progress of the talk. They evidence the kind of detailed analysis that leaves its author inescapably informed about what
396
Emanuel A. Schegloff
is going on in the data. The robustness of the analysis can have no better guarantor than these little pictures; given the choice between these transcripts-with-pictures and the statistical analysis that depends in large measure on the work they embody, I would pick the former any day. And, at the very end of the chapter, its payoff will surface again, when some statistical outliers that might easily be dismissed as “noise” by those relying on the statistics get satisfying accounts by virtue of the detailed analysis that had been done in the first place. The fourth section describes data and method . It is explicit, detailed, and argued . . . and therefore arguable with. For example, here is how the authors present the selection of their sample: This extensive database of interaction in the three cultures was sampled in a systematic way. Within each cultural sample, we selected roughly ten dyads, and searched for question–response sequences until we had 300 such sequences for each cultural sample. By “question” we understood any utterance that functioned as an information soliciting action, regardless of whether it was in interrogative form or otherwise marked morphosyntactically, lexically or prosodically. For example, many “yes”/“no” questions in Yélî Dnye are delivered in declarative format with falling intonation – they are recognizable as questions just because they appear to be statements about facts that are privileged information of the recipient’s, of the kind “You have a stomach ache” (cf. Labov and Fanshel ’s (1977) “B-event statements” ). (Rossano et al., p. 205)
The explicitness of this account allows a reader to think along with the researchers, for example, along lines such as the following. Aren’t these “statements about facts that are privileged information of the recipient’s, of the kind ‘You have a stomach ache’” more aptly understood as “noticings” than as “questions”? As such, they are inspectable by their recipient for what has prompted or motivated their articulation at some “this moment” – complaining , sympathizing, etc. So, does the category “questions” serve well here? As compared, for example, to “first pair parts that make (or can make) a spoken response relevant next”? Whether this point is well taken or not, it is the very possibility of putting it forward, of checking it out vis-à-vis actual data, that makes this way of proceeding so fruitful. Indeed, it is so appealing that I would have welcomed the display of one or two exemplars of each of the sequence types they draw on (“requests for new information, requests for repair
One perspective on Comparative Perspectives
397
[e.g., for repetition , or clarification], requests for confirmation of previously established information, or presumed facts],” Rossano et al.) in each of the languages, so the reader could have a more informed understanding of the sort of interactional events being referred to in the ensuing text. If the only obstacle to doing this was space, this reader would have prefer red relocating some of the statistical discussion to an appendix, the payoff of which would not be limited to this “data and method ” section but would be even more welcome in the Results section that follows. The Results section is largely statistical and takes on the appearance of most of the Fox et al. chapter: statistical findings followed with hypothetical understandings of the statistics. For example: Another general fact emerging from all three cultures is that questions initiating repair are especially likely to involve speaker gaze . One possibility is that gaze here reinforces that the Q-speaker, though speaking, is committing to a more attentive engagement as a listener in the conversation. The speaker of the repair initiation is asking the recipient to repair his or her prior talk because of some problem, and the speaker of the repair question can ask to delay the progressivity of the talk by projecting a full recipiency once the repair is produced. (Rossano et al., p. 216)
It would be most welcome if these speculations (“Perhaps gaze here reinforces . . .”) could be grounded in, or superceded by, actual analyses of such sequences. One might suspect, for example, that rather than a solicitous “more attentive engagement as a listener,” other-initiation of repair conveys disaffiliation, nonalignment, or some other problematic stance toward the question-recipient and Q-recipient’s just prior utterance; it might therefore be more a kind of interactionally “loaded” moment in the interaction than is suggested by the text as it now stands. And this speculation – both the authors’ and this reader/writer’s – has implications for the larger scope of the research undertaking being reported in this chapter. Recall that the point of this research was characterized at the outset as “to understand whether gaze as an interactional practice has universal properties across cultures or not” (Rossano et al., p. 188). That may be the right way to approach this area of work, but let me suggest an alternative. What if we began by asking how we are to understand the job(s) that gaze is doing in the data site most intensively studied here: the
398
Emanuel A. Schegloff
Italian data. Then we can ask whether gaze is doing the same job(s) in Yélî Dnye , where gaze behavior looks pretty much the same. And then we can ask whether gaze behavior does the same jobs in Tzeltal, where it looks in some respect different. And then we can ask what else is being done in Tzeltal (other than gaze) that does the jobs done in Italian by gaze? That is, the issue ultimately is not gaze (as the given “primitive”); the issue is “What does the organization of interaction require that gaze does in Italian and that other forms of conduct may do in other cultures, and what are the differential consequences, if any, of getting this job – or these jobs – done by gaze in Italy and otherwise elsewhere?” Such questions can only be answered – or modified – by engagement with the details of what is going on in single episodes of interaction and in aggregates of such analyzed episodes; and if statistics are needed to get there, or to get the funding to get there, or to get the attention and respect of those who have not yet figured out where the real payoffs are to be found, so be it, but let us always keep in mind what is the cart and what is the horse. Where does this leave us? The Rossano et al. chapter goes a long way – but (in my view) not all the way – toward what can currently be achieved in CA comparative analysis in which the terms of comparison are linguistic or cultural. On the whole, it does what I earlier suggested CA comparative analysis should begin by doing: It states clearly and explicitly what the target phenomenon or practice is; it tries to replicate it in a corpus of data different from the one(s) drawn upon in prior inquiry, and in doing so arrives at findings that differ from those previously reported; it undertakes to explore the target practice in two more environments that differ both from those explored in the inherited literature and in the environment of the new undertaking; it establishes a reasonable basis for comparability – holding constant the interactional activity (specified forms of questioning), targeting practices of gazing across variations in cultural and linguistic formations. On the other hand, in doing so it seems to hold one culture’s conduct answerable to the conduct of others. Specifically, the
One perspective on Comparative Perspectives
399
convergent practices of gaze in the Italian and Yélî Dnye data render the Tzeltal practices “different” and suggest a search for what will explain or “normalize” this Tzeltal practice – just the sort of undertaking I have tried to call into question in an earlier section of this chapter. How are we to reconcile all the attractive features of the Rossano et al. chapter with the reservations raised by our commitment to understanding the practices of interaction that constitute the world in which those actors live? We surely will not arrive at an answer to that question here. But it will as surely involve treating each language/culture gestalt in its own terms to begin with. Paradoxically perhaps this picture has to be enlarged to a still greater canvas – one in which it is not gaze (or repair, initiation, etc.) that is the focus, with gaze (or other target) practices in one cultural setting being juxtaposed to gaze (or the same other target) practices in another, but rather contingencies of interactional organization that transcend cultural boundaries and help us to mediate the variations between linguistic, cultural and social formations (Schegloff 2006). Might it not be these contingencies of interactional organization that will underwrite CA comparative studies in a fashion that will not elevate or demean one culture visà-vis another but will take the measure of each by reference to the organization of interaction for the human species? Notes 1 Although this is (to the best of my knowledge) the first volume that overtly announces itself to be addressed to conversation-analytically informed comparative research, it has two less overt predecessors. (There are, of course, also papers devoted to comparative treatment of CA findings (e.g., Fox et al. 1996; Hacohen and Schegloff 2006; Lerner and Takagi 1999; and Moerman 1977, inter alia.) One is the volume edited by K. K. Luke and T. S. Pavlidou (2002) on Telephone Calls; the other is the volume edited by N. J. Enfield and T. Stivers (2007) on Person Reference in Interaction. As the topic of the first is virtually a creation of CA, it turns out to be conversation analysis in comparative inquiry even though not all of the contributors are trained in CA or perhaps even think of themselves as conversation analysts. The second of these volumes addresses a topic surely not invented by CA, but is focused for the most part on a particular stance on person reference that is the product of conversation-analytic research (Sacks
400
Emanuel A. Schegloff
and Schegloff 1979). Some of its contributors are conversation analysts, some are “friends of CA” or “fellow travelers,” some are specifically not conversation analysts who, while respecting CA work, have profound and persistent objections to it, largely centered in linguistic anthropology. I address myself to some issues raised in that volume in part 2 of this chapter. The present volume, however, not only makes conversation analysis its title; the majority of its contributors were trained in CA, offer their work as conversation-analytic by intent, design and realization, though in some cases identifying themselves as not exactly conversation analysts, but kindred folk – “interactional linguists ” most often. I was represented in the first volume in much the same capacity as in the present one (2002a: 249–81), and for the most part try to avoid repeating myself, though the early pages of my contribution to that volume have a bearing on some of the pieces in this volume as well. Finally, I have benefited from input from a number of colleagues – Nick Enfield, John Heritage, Celia Kitzinger , Gene Lerner, and Jack Sidnell , none of whom bear any responsibility (nor will they be specifically credited) for what I have made of their suggestions. 2 Terms such as “children” and “adults” are in scare quotes as an alert that these are categories from the “age” and “stage-of-life” categorization devices whose relevance has not been grounded and would have to be grounded in data to warrant serious reception as CA work. Cf. Sacks 1972a, 1972b, 1992 passim; Schegloff 1991, 1997a, 2007a. 3 In addition to the temporality of individuals’ lives, there is the passage of social, cultural, and historical time. The sort of data which has proven indispensable for CA work limits the possibility of comparative analysis at present, but not for long. At the very least, there is the material collected by conversation analysts since the early 1960s – now almost a half century ago. There may well be retrievable recorded data of ordinary talk-in-interaction – at least audio data – going back some twenty years before that which would support modestly detailed analysis (cf., e.g., Clayman and Heritage 2002). And there is also the more limited possibility of using much older records to ground claims about the historical robustness of practices initially found for contemporary data (as in my effort to show that an account of summons/answer sequences was not limited in its relevance to contemporary telephone conversation by citing a biblical report of an exchange between Abraham and his God: “And He said, Abraham; and he said, Behold, I am here.” (Genesis 22:1); Schegloff 1968: 1075). 4 I should make clear that, in what follows, I am addressing “the-Kitzinger that-Enfield-invokes” as part of my addressing Enfield’s argument. Kitzinger herself has taken these issues on in a variety of papers engaged
One perspective on Comparative Perspectives
401
with a variety of data (cf. Kitzinger 2000, 2005a, 2005b, in preparation; Land and Kitzinger 2005), papers embodying a developing trajectory of argument which it is beyond the scope of this chapter to address. 5 Actually Jeannette does later correct Estelle’s reference to “a colored lady” by saying, “No, no, that was one of the employees.” 6 Both Enfield and Kitzinger invoke Sacks ’ programmatic call for accounts of “doing being ordinary” as the central thrust of what they are seeking (Enfield’s chapter begins with a citation from one of Sacks ’ lectures on that theme). It is surely a theme worthy of pursuit, but how is it to be pursued? One way of pursuing it is to take as one’s point of departure data in which participants are overtly engaged in “doing being ordinary,” or in which they overtly juxtapose the non-ordinary with the ordinary. One example is the work Sacks had undertaken and Jefferson has brought to fruition on the initial “normalization” which people report themselves to have imposed on what turned out to be anything but ordinary (Jefferson 2004: 131–167). A second example is the previously mentioned account by Sacks of alternative ways of “seeing” or grasping the scene with the police, an account which may well have turned on the identity of one of the people being seen and referred to by Estelle as “a colored lady” and later as “a negro woman” (with the implication of her involvement in “the trouble”) and having that rejected by the woman who works at the store, “No, no, that was one of the employees,” as a way of absolving her of such suspicion. A third tack turns on a different sort of realization of the “ordinary” and the “present case”; for example, Enfield finds it useful to underscore the compelling character of Kitzinger ’s account of the premises of heterosexuality by juxtaposing it to data examined in a separate paper by Land and Kitzinger (2005) in which a lesbian client corrects an insurance salesman’s subsequent reference to her “spouse” as her “husband.” And a fourth, and for now final, way of grounding such proposed findings is exemplified in a paper by Lerner and Kitzinger (2007), in which they examine occasions in which the usually “unmarked” and “unnoticed” practices of self-reference are given explicit attention through self-repair operations on self-reference (replacing “I” with “we,” or “we” by “I”). Invocations of such data exemplars serve to ground the account of the otherwise ordinary by exemplifying its consequentiality in actual occasions of interaction. Otherwise, there is virtually no end to the categories whose members can find themselves demeaned, excluded, or otherwise not taken into account in ordinary talk-in-interaction (a theme pursued in Goffman ’s Stigma 1963: 137–138).
402
Emanuel A. Schegloff
7 These accounts are taken up in somewhat greater detail in Schegloff 1996b: 168–181 and Schegloff 1997b: 501–502. 8 A notable example has been past and contemporary work on TCUs in Japanese (Hayashi 2003, 2004a, 2004b; Iwasaki 2007, forthcoming; Tanaka , 1999, 2000, 2004; Thompson and Couper-Kuhlen 2005) and Korean (Kim 1999b, 2007; Park, in preparation). 9 Actually, the citation to Schegloff 1979 hardly qualifies as an exception to the observation that “very little has been written on the question of where within a word speakers tend to initiate repair.” When this citation is later taken up at the start of their third section, the authors write: “The first proposal we will explore concerning site of initiation appears in Schegloff (1979). There, Schegloff writes, “Just post-initiation and just pre-completion of various unit types seem to be specially common loci of repair initiation. Thus, just after the start of a turn-constructional unit (e.g., a sentence) or just before its completion; after the first sound of a word or just before its last sound” (1979: 275 [emphasis added by Fox et al.]). This is just one point in a long paper covering many topics, and Schegloff does not elaborate on it, nor give any examples. Nor is it clear from the paper exactly what is meant by “after the first sound of a word or just before its last sound”; after the first sound could mean after the first sound is recognizable, or after the first sound is recognizably complete; just before its last sound could likewise mean before the last sound is articulated or before the last sound is complete. However, Schegloff (personal communication) has defined the relevant domain for just post-beginning as starting after the first sound is recognizable, thus has begun to be articulated, and continues until the first sound is complete, while the relevant domain for just pre-completion begins just before the final sound is articulated, in the penultimate sound, and continues until just before the final sound is complete. He has further suggested that site of initiation of repair may depend on the type of repair involved.” As the authors note, I offer no elaboration and no data, and this should have served as an alert (given the densely empirical character of virtually everything I have published) that this was an impression gathered informally and mentioned in passing, not a claimed finding on which further work could be based. And the “personal communication” that is cited took place almost thirty years after the cited text was published, after initial resistance on my part, and hardly amounts to a “definition.” In any case, it is no exception to the absence of a literature on the matter in question. 10 In fact, what is technically a repair segment can be even shorter than the initiation and the repair: as remarked earlier, sometimes there is
One perspective on Comparative Perspectives
403
no repair initiator at all, just a recognizable repair; and sometimes there is a repair initiator and no repair at all, but the production of a projectably next component of the turn-constructional unit in progress – a “cancellation,” if you will, after initiation and before realization. 11 Examination of the 1996 paper was prompted by the absence from an earlier draft of the Fox et al. chapter of data presenting actual talk in which same-turn repair is initiated. The extracts included in the final draft that appears in this volume do not exemplify the problems discussed in the text. As well, the discussion in the following two sections, pages 383–387, was addressed to that earlier draft. 12 Here is another, quite different, example of the sort of thing I have in mind. In Schegloff 1987b, the analysis focuses on a single turn, one which happens to include a self-initiated same turn repair. The turn is: 01 Curt: [He- he’s about the only regular
Here, as it happens, the repair is not initiated by a cut-off but by a rushed restart of the turn after “regular” on the first line. But the import comes from
(a) where it is initiated; (b) how far back it goes in framing the trouble-source and its repair; (c) how best to characterize where it is initiated, and (d) whether all this is pretty much the default form of the same-turn repair or even of same-turn insertion repair, or whether it is a departure, and, if so, what the import of the departure is, if any. At the time the paper was conceived and then written, I did not know the default form that such insertion-repairs have. I now know that this is an unusual and exceptional instance of its type on counts a, b, and c, and this bears on what is getting done interactionally. To be sure, Fox et al. limit their inquiry to replacements, but are not the issues the same? 13 An example: Debbie and Shelley are two young attorneys who were to have gone with a group of others to a football weekend in another city, except that Shelley has backed out at the last moment. Debbie has called Shelley and accused her of “blowing off her girlfriends just because this guy’s not going.” Shelley has denied the accusation and has been explaining why she is not going and its connection to her boyfriend’s not going. Here is the end of her account: 01 S: → So:I mean it’s not becuz he’s- he’s- I mean it’s 02 → not becuz he’:s not going, it’s becuz (0.5) his 03 money’s not¿ (0.5) funding me.
404 04 D:
Emanuel A. Schegloff okay¿
Note here, at lines 1 and 2, Shelley’s extended effort to distribute the stress in her turn so as to set up the contrast on which her account for not going depends. I submit that though the lexical composition does not change across the several tries at getting this said, the recycling is not “simple”; it is not there so as to delay the talk that will follow, as the authors go on to claim. Another example is presented in Sidnell ( 2006) as in the service of what Stivers (2005) terms a “modified repeat”: 01 F: [the thing ( ) 02 S: =we need to deal with this an’ get on with [it. 03 F: [(yih) 04 F: this is an important issue to a lot of people. 05 S:→ ah-ah I agree. it’s a-it’s-it is an important issue. 06 S: but there are many other important issues that we 07 S: must address as well.
14 That this does matter to the authors is indicated by their citation to the work of Jasperson ( 1998) in the just-following text. 15 In the last 10–15 years, there has been substantial growth in the application of quantitative and statistical resources in CA work – largely in CA work on data from so-called institutional contexts such as politics, media and medicine: on politics and media, cf. Heritage and Greatbatch, 1986; Clayman , Heritage, Elliott, and McDonald, 2007 inter alia; on medicine, cf. Heritage, Boyd and Kleinman, 2001; Stivers , Mangione-Smith , Elliott and McDonald, and Heritage, 2003; Robinson and Heritage 2005; Stivers and Majid, 2007; and Heritage, Robinson , Elliott, Beckett, and Wilkes, 2007 inter alia. My thanks to John Heritage (pc) for orienting me to some of this literature: “I think the key thing about the statistically oriented studies is that almost, if not all, of them use stat[istics] to build a relationship between already identified practices of talk and exogenous putative causes (e.g., attitudes, race or gender, the state of the economy) of talk, or of its consequences (prescribing decisions etc.). Stat[istics] are not used to build analyses of the internal relationships among practices of talking. In the older language, CA provides adequacy at the level of meaning, the stat[istics] provide causal adequacy for claims that tie practices of talk as causal to, or caused by, exogenous social processes. The papers you mention may represent departures from this pattern, or not.” Heritage’s characterization of this form of application of statistics in CA work does not apply to the Fox, Wouk et al. chapter.
One perspective on Comparative Perspectives
405
16 I either know, or confidently suppose, this to be the case for Bolden, Egbert, Hayashi , Heineman, Lindström , Rossano , Sidnell , Sorjonen, and Wu, and perhaps others whom I know less well, and it shows in their respective chapters. 17 I use the present tense here because, while working on this chapter and waiting for the other chapters to be made available in final draft, I was also working on my own research – in the present case, getting a better grasp on the use of “y’know” in American English in preparation for writing the next volume in the Primer series (Schegloff 2007b) which will most likely be on the organization of repair, in which some uses of “y’know” figure. I have assembled a collection of some 220 instances and have myself been doing what the text of the preceding pages has been describing. I mean to be contrasting the rote practices by often otherwise uninvolved RAs in coding data into prespecified categories whose import (i.e. what they are taken to be indicators of) they may know nothing about, on the one hand, with fully committed and involved researchers examining such data with the coding categories at hand who can see in the juxtaposition of the coding categories and the details of the data more than just a “yes” or “no” – but also what is involved in the presence or absence of the thing being coded for, leading to a revised understanding of the phenomenon. 18 In proceeding this way, I stand by the judgement of the Fox, Wouk et al. chapter as more Interactional-Linguistic than ConversationAnalytic. At the time of writing, I have not seen the final version of the Rossano , Brown and Levinson chapter. 19 Given the extended discussion of the relevance of categories and their deployment in CA work in “Comparative Analysis: One CA Perspective” above, I am reluctant to linger on it further here. Yet the centrality of “comparative culture” analysis in the Rossano et al. chapter prompts inclusion of the following footnote from one of the earliest CA papers, apropos the characterization of the data examined there, in response to an editorial suggestion: “that all the conversations are in ‘American English’ is no warrant for so characterizing them. For there are many other characterizations which are equally ‘true,’ e.g., that they are ‘adult,’ ‘spoken’ (not yelled or whispered), etc. That the materials are all ‘American English’ does not entail that they are relevantly ‘American English,’ or relevantly in any larger or smaller domain that might be invoked to characterize them. All such characterizations must be warranted, and except for the account we offer in the final section of the paper, we cannot warrant them now. Ethnic, national, or language identifications differ from many
406
Emanuel A. Schegloff
thers only in their prima facie plausibility, especially to those in the o tradition of anthropological linguistics” (Schegloff and Sacks 1973, 291–292, note 4, emphasis in original). For “American English,” substitute Italian, Tzeltal, or Yélî Dnye. But note that this issue is being raised where the target of inquiry is not specifically language, but was not raised for the Fox et al. chapter or any of the others, in which deployment of the language was the object of inquiry. In this respect, I agree with the position taken up in Lerner and Takagi (1999): “In examining language use it seems safe to assume that the language spoken has a continuing relevance for turn construction, since the choices and possibilities for a speaker are always found within the linguistic resources available. However, it is another matter to demonstrate just how these linguistic resources are consequential for the organization of talk-in‑interaction. This is one place where a comparison – or co‑investigation – of distinct languages can be advantageous.” Should the gaze practices being explored by Rossano et al. turn out to be linked to features of the language, this reservation would not be in point. But in the chapter published here, the text varies between referring to “cultures” (as in “Another general fact emerging from all three cultures” (Rossano et al., p. 216) and referring to “language X speakers” (Rossano et al., passim) or both, as in these consecutive sentences, “We turn now to consider possible crosscultural differences in gaze behavior. Figure 7.3 compares the gaze behavior of Q-speakers and recipients across the three languages” (p. 208). It is unclear how deployment of gaze would be part-andparcel of language, but plausible conjectures about it can readily be
Bibliography
Agar, M. (1973). Ripping and Running: A Formal Ethnography of Urban Heroin Addicts, New York: Seminar Press. Aijmer, K. (2002). English Discourse Particles: Evidence from a Corpus. Amsterdam and Philadelphia: John Benjamins. Albert, E. (1964). “ ‘Rhetoric,’ ‘Logic,’ and ‘Poetics’ in Burundi: Cultural Patterning of Speech Behavior.” In J. J. Gumperz and D. Hymes (Eds.), The Ethnography of Communication. American Anthropologist 66: 6, part 2 (special publication). Allsopp, R. (1996). Dictionary of Caribbean English Usage. Oxford University Press. Argyle, M., and Cook, M. (1976). Gaze and Mutual Gaze. Cambridge University Press. Argyle, M., and Dean, J. (1965). “Eye-Contact, Distance and Affiliation.” Sociometry 28 (3): 289–304. Argyle, M., and Graham, J. A. (1976). “The Central Europe Experiment: Looking at Persons and Looking at Things.” Journal of Environmental Psychology and Nonverbal Behavior 1 (1): 6–16. Atkinson, J. M. and Heritage, J. (Eds.) (1984). Structures of Social Action: Studies in Conversation Analysis. Cambridge University Press. Auer, P. (1984). “Referential Problems in Conversation.” Journal of Pragmatics 8: 627–648. Auer, P. and Uhmann, S. (1982). “Aspekte Der Konversationellen Organisation Von Bewertungen.” Deutsche Sprache 10: 1–32. Bavelas, J. B., Coates, L. and Johnson, T. (2002). “Listener Responses as a Collaborative Process: The Role of Gaze.” Journal of Communication 52 (3): 566–580. Beach, W. A. (1993). “Transitional Regularities for Casual ‘Okay’ Usages.” Journal of Pragmatics 19 (4): 325–352. Beach, W. and Lindström, A. (1992). “Conversational Universals and Comparative Theory: Turning to Swedish and American Acknowledgment Tokens Interaction.” Communication Theory 2: 24–49.
408
Bibliography
Beattie, G. W. (1978). “Floor Apportionment and Gaze in Conversational Dyads.” British Journal of Social and Clinical Psychology 17: 7–16. Beattie, G. W. (1979). “Planning Units in Spontaneous Speech: Some Evidence from Hesitations in Speech and Speaker Gaze Direction in Conversation.” Linguistics 17: 61–78. Becker, H. S. (1963). Outsiders, New York: Free Press. Betz, E. and Golato, A. (2008). “Remembering Relevant Information and Withholding Relevant Next Actions: The German Token Achja.” Research on Language and Social Interaction 41 (1): 58–98. Billig, M. (1999a). “Conversation Analysis and the Claims of Naivete.” Discourse and Society 10: 572–576. Billig, M. (1999b). “Whose Terms? Whose Ordinariness? Rhetoric and Ideology in Conversation Analysis.” Discourse and Society 10: 543–558. Bitextin, A. B. (1994). “Chastitsy -TO, ZHE, VED’ i vvodnye konstruktsii tipa KAK IZVESTNO kak sredstva ukazanija na izvestnost’ propozitsional’nogo soderzhanija predlozhenija slushajuschemu.” Unpublished Ph.D. thesis, Moskovskij gosudarstvennyj universitet, Moscow. Blass, R. (2000). “Particles, Propositional Attitude and Mutual Manifestness.” In G. Andersen and T. Fretheim (Eds.), Pragmatic Markers and Propositional Attitude (pp. 39–52). Amsterdam and Philadelphia: John Benjamins. Bloomfield, L. (1946). Language. New York: Henry Holt & Co. Bloomfield, L. (1970). “Meaning.” In C. Hockett (Ed.), A Leonard Bloomfield Anthology (pp. 400–405). Bloomington, Ind.: Indiana University Press. Blum-Kulka, S., House, J., and Kasper, G. (Eds.) (1989). Cross-Cultural Pragmatics. Norwood, N.J.: Ablex. Boersma, P. and Weenink, D. (2006). “Praat: Doing Phonetics by Computer (Version 4.4.24 [Computer program].” Available online at www.praat.org. (Accessed 19 June 2006.) Bolden, G. B. (2003). “Doing Being Late: The Use of the Russian Particle -to in Personal State Inquiries.” CLIC (Crossroads of Language, Interaction, and Culture) 5: 3–27. Bolden, G. B. (2005). “Delayed and Incipient Actions: The Discourse Markers ‘-To’ and ‘So’ in Russian and English Conversation.” Unpublished Ph.D. thesis, University of California, Los Angeles. Bolden, G. B. (2006). “Little Words That Matter: Discourse Markers ‘So’ and ‘Oh’ and the Doing of Other-Attentiveness in Social Interaction.” Journal of Communication 56 (4): 661–688. Bolden, G. B. (2008). “Reopening Russian Conversations: The Discourse Particle -to and the Negotiation of Interpersonal Accountability in Closings.” Human Communication Research 34: 99–136.
Bibliography
409
Bolden, G. B. (2009). “Implementing incipient actions: The discourse marker ‘so’ in English conversation.” Journal of Pragmatics. 41 (5): 974–998 Bonnot, C. (1987). “-To Particule de rappel et de thématisation.” In Les Particules énonciatives en russe contemporain (Vol. II (pp. 113– 171). Paris: Institut D’Études Slaves. Bonnot, C. (1990). “La Particule -to et la polémique chachée en russe moderne: A propos du statut énuonciatif du thème.” Revue des Études Slaves 62 (1–2): 67–75. Boyd, E., and Heritage, J. (2006). “Taking the Patient’s Medical History: Questioning During Comprehensive History Taking.” In J. Heritage and D. Maynard (Eds.), Practicing Medicine: Talk and Action in Primary-Care Encounters (pp. 151–184). Cambridge University Press. Brown, G., and Yule, G. (1983). Discourse Analysis. Cambridge University Press. Brown, P. (1979). “Language, Interaction and Sex Roles in a Mayan Community: A Study of Politeness and the Position of Women.” Unpublished Ph.D. dissertation. Department of Anthropology, University of California, Berkeley. Brown, P. (1998). “Conversational Structure and Language Acquisition: The Role of Repetition in Tzeltal Adult and Child Speech.” Journal of Linguistic Anthropology 8 (2): 197–221. Brown, P. (2007). “Principles of Person Reference in Tzeltal Conversation.” In N. J. Enfield and T. Stivers (Eds.), Person Reference in Interaction: Linguistic, Cultural and Social Perspectives (pp. 172–202). Cambridge University Press. Brown, P. (in press). “Questions and Their Responses in Tzeltal.” In T. Stivers, N. J. Enfield and S. C. Levinson (Eds.), Question-Response Sequences in 10 Languages, Special Issue of Journal of Pragmatics. Brown, P., and Levinson, S. C. (1987). Politeness: Some Universals in Language Usage. Cambridge University Press. Brown, P., and Levinson, S. C. (2005). Comparative Response Systems. Paper presented at the 104th Annual Meetings of the American Anthropological Association. Button, G., and Casey, N. (1984). “Generating Topic: The Use of Topic Initial Elicitors.” In J. M. Atkinson and J. Heritage (Eds.), Structures of Social Action: Studies in Conversation Analysis (pp. 167–190). Cambridge University Press. Button, G., and Casey, N. (1985). “Topic Nomination and Topic Pursuit.” Human Studies 8: 3–55. Button, G., and Casey, N. (1988/9). “Topic Initiation: Business-at-Hand.” Research on Language and Social Interaction 22: 61–92. Cassidy, F. G. and Le Page, R. (1980). Dictionary of Jamaican English, 2nd edn. Cambridge University Press.
410
Bibliography
Chang, Y.-C. (1998). “Les Indices acoustiques et perceptifs des questions totales en Mandarin parlé de Taiwan” [“An Acoustic and Perceptual Study of Yes/No-Questions in Taiwan Mandarin Chinese”], Cahiers de Linguistique Asie Orientale 27: 51–78. Chui, K. (1996). “Organization of Repair in Chinese Conversation.” Text 16: 343–372. Clayman, S. E. and Heritage, J. (2002). “Questioning Presidents: Journalistic Deference and Adversarialness in the Press Conferences of U.S. Presidents Eisenhower and Reagan.” Journal of Communication 52 (4): 749–775. Clayman, S. E., Heritage, J., Elliott, M. N. and McDonald, L. (2007). “When Does the Watchdog Bark? Conditions of Aggressive Questioning in Presidential News Conferences.” American Sociological Review 72 (1): 23–41. Clift, R. (1999). “Grammar in Interaction: The Case of ‘Actually’.” Essex Research Reports in Linguistics. Clift, R. (2006a). “Getting There First: The Non-Narrative Use of Repor ted Speech in Interaction.” In E. Holt and R. Clift (Eds.), Reporting Talk: Reported Speech in Interaction (pp. 120–149). Cambridge University Press. Clift, R. (2006b). “Indexing Stance: Reported Speech as an Interactional Evidential.” Journal of Sociolinguistics 10 (5): 569–595. Cohen, J. (1960). “A Coefficient of Agreement for Nominal Scales.” Educational and Psychological Measurement 20 (1): 37–46. Comrie, B. (1981). Language Universals and Linguistic Typology. Oxford: Blackwell. Couper-Kuhlen, E. (1992). “Contextualizing Discourse: The Prosody of Interactive Repair.” In P. Auer and A. di Luzio (Eds.), The Contextualization of Language (pp. 337–364). Amsterdam and Philadelphia: John Benjamins. Couper-Kuhlen, E. (2004). “Prosody and Sequence Organization in English Conversation: The Case of New Beginnings.” In E. CouperKuhlen and C. E. Ford (Eds.), Sound Patterns in Interaction: CrossLinguistic Studies from Conversation (pp. 335–376). Philadelphia: John Benjamins. Couper-Kuhlen, E. and Ford, C. E. (2004). Sound Patterns in Interaction: Cross-Linguistic Studies from Conversation. Amsterdam and Philadelphia: John Benjamins. Couper-Kuhlen, E., and Selting, M. (Eds.). (1996). Prosody in Conversation. Cambridge University Press. Curl, T. S. (2005). “Practices in Other-Initiated Repair Resolution: The Phonetic Differentiation of ‘Repetitions’.” Discourse Processes 39: 1–43. Darwin, C. R. (1872). The Expression of the Emotions in Man and Animals. London: John Murray.
Bibliography
411
Dersley, I., and Wootton, A. (2000). “Complaint Sequences Within Antagonistic Argument.” Research on Language and Social Interaction 33 (4): 375–406. Dittman, A. T. and Llewellyn, L. G. (1967). “The Phonemic Clause as a Unit of Speech Decoding.” Journal of Personality and Social Psychology 6: 341–349. Dittmann, J. (1980). “Auch Und Denn Als Abtönungspartikeln.” Zeitschrift Für Germanische Linguistik 8: 51–73. Drew, P. (1984). “Speakers’ Reportings in Invitation Sequences.” In J. M. Atkinson and J. Heritage (Eds.), Structures of Social Action: Studies in Conversation Analysis (pp. 152–164). Cambridge University Press. Drew, P. (1991). “Asymmetries of Knowledge in Conversational Interactions.” In I. Markova and K. Foppa (Eds.), Asymmetries in Dialogue (pp. 29–48). Hemel Hampstead: Harvester Wheatsheaf. Drew, P. (1992). “Contested Evidence in Courtroom Cross-Examination: The Case of a Trial for Rape.” In P. Drew and J. Heritage (Eds.), Talk at Work: Interaction in Institutional Settings (pp. 470–520). Cambridge University Press. Drew, P. (1997). “‘Open’ Class Repair Initiators in Response to Sequential Sources of Troubles in Conversation.” Journal of Pragmatics 28: 69–101. Drew, P. (2003). “Comparative Analysis of Talk-in-Interaction in Different Institutional Settings.” In P. J. Glenn, C. D. LeBaron and J. Mandelbaum (Eds.), Studies in Language and Social Interaction: In Honor of Robert Hopper (pp. 293–308). Mahwah, N.J.: Lawrence Erlbaum. Drew, P. and Holt, E. (1998). “Figures of Speech: Figurative Expressions and the Management of Topic Transition in Conversation.” Language in Society 27: 495–522. Duncan, S. (1972). “Some Signals and Rules for Taking Speaking Turns in Conversations.” Journal of Personality and Social Psychology 23: 283–292. Duncan, S. (1974). “On the Structure of Speaker-Auditor Interaction During Speaking Turns.” Language in Society 2: 161–180. Duncan, S. (1975). “Interaction Units During Speaking Turns in Dyadic, Face-to-Face Conversations.” In A. Kendon, R. M. Harris and M. R. Key (Eds.), Organization of Behaviour in Face-to-Face Interaction (pp. 199–212). The Hague: Mouton. Duncun, S. and Fiske, D. (1977). Face-to-Face Interaction: Research, Methods, and Theory. Hillsdale, N.J.: Erlbaum Duncan, S. and Niederehe, G. (1974). “On Signalling That It’s Your Turn to Speak.” Journal of Experimental Social Psychology 10 (3): 234–247.
412
Bibliography
Durrell, M. (1992). Using German: A Guide to Contemporary Usage. Cambridge University Press. Egbert, M. (1993). Schisming: The Transformation from a Single Conversation to Multiple Conversations. Ph.D. Dissertation. UCLA, Department of Applied Linguistics. Egbert, M. (1996). “Context-Sensitivity in Conversation: Eye Gaze and the German Repair Initiator ‘bitte?’” Language in Society 25 (4): 587–612. Egbert, M. (1997a). “Schisming: The Collaborative Transformation from a Single Conversation to Multiple Conversations.” Research on Language and Social Interaction 30 (1): 1–51. Egbert, M. (1997b). “Some Interactional Achievements of Other-Initiated Repair in Multiperson Conversation.” Journal of Pragmatics 27 (5): 611–634. Egbert, M. (2002). Der Reparatur-Mechanismus in deutschen und interkulturellen Gesprächen. Habilitation. University of Oldenburg. Egbert, M. (2004). “Other-Initiated Repair and Membership Categorization: Some Conversational Events That Trigger Linguistic and Regional Membership Categorization.” Journal of Pragmatics 36 (8): 1467–1498. Ekman, P., and Friesen, W. V. (1974). “Nonverbal Leakage and Clues to Deception.” In S. Weitz (Ed.), Nonverbal Communication (pp. 269–290). New York: Oxford University Press. Enfield, N. J. (2007). “Meanings of the Unmarked: How ‘Default’ Person Reference Does More Than Just Refer.” In N. J. Enfield and T. Stivers (Eds.), Person Reference in Interaction: Linguistic, Cultural and Social Perspectives (pp. 97–120). Cambridge University Press. Enfield, N. J., and Stivers, T. (Eds.) (2007). Person Reference in Interaction. Cambridge University Press. Erickson, F. (1979). “Talking Down: Some Cultural Sources of Miscommunication in Interracial Interviews.” In A. Wolfgang (Ed.), Nonverbal Communication: Applications and Cultural Implications (pp. 99–126). New York: Academic Press. Evans, N., and Wilkins, D. (2000). “In the Mind’s Ear: The Semantic Extensions of Perception Verbs in Australian Languages.” Language 76 (3): 546–592. Exline, R. V. (1963). “Explorations in the Process of Person Perception: Visual Interaction in Relation to Competition, Sex, and Need for Affiliation.” Journal of Personality 31 (1): 1–20. Ferrara, K. (1997). “Form and Function of the Discourse Marker Anyway: Implications for Discourse Analysis.” Linguistics 35 (2): 343–378. Filipi, A., and Wales, R. (2003). “Differential Uses of Okay, Right, and Alright, and Their Function in Signaling Perspective Shift or Maintenance in a Map Task.” Semiotica 147–1 (4): 429–455.
Bibliography
413
Fincke, S. (1999). “The Syntactic Organization of Repair in Bikol.” In B. Fox, D. Jurafsky, and L. Michaelis (Eds.), Cognition and Function in Language (pp. 252–267). Palo Alto, Calif.: CSLI. Fishman, P. (1983). “Interaction: The Work Women Do.” In B. Thorne, C. Kramarae, and N. Henley (Eds.), Language, Gender, and Society (pp. 89–101). London: Newbury House. Foolen, A. (2006). “Polysemy Patterns in Contrast: The Case of Dutch Toch and German Doch.” In K. Aijmer and A. -M. Simon-Vandenbergen (Eds.), Pragmatic Markers in Contrast (pp. 59–72). Amsterdam: Elsevier, Ford, C., Fox, B. A., and Thompson, S. A. (1996). “Practices in the Construction of Turns: The ‘TCU’ Revisited.” Pragmatics 6: 427–454. Ford, C. E., Fox, B. A. and Thompson, S. A. (2002). “Social Interaction and Grammar.” In Michael Tomasello (Ed.), The New Psychology of Language: Cognitive and Functional Approaches to Language Structure (Vol. II, pp. 119–143). Mahwah, N.J.: Lawrence Erlbaum. Fox, B. (2007). “Principles Shaping Grammatical Practices: An Exploration.” Discourse Studies 9: 299–318. Fox, B. and Wouk, F. (2003). A Cross-Linguistic Study of Self-Repair. Proposal to the National Science Foundation. Fox, B., Hayashi, M., and Jasperson, R. (1996). “A Cross-Linguistic Study of Syntax and Repair.” In E. Ochs, E. A. Schegloff, and S. A. Thompson (Eds.) Interaction and Grammar (pp. 185–237). Cambridge University Press. Franck, D. (1980). Grammatik und Konversation. Königstein im Taunus: Scriptor. Fretheim, T. (2000). “Propositional Attitude in Norwegian Conditional Clauses.” In G. Andersen and T. Fretheim (Eds.), Pragmatic Markers and Propositional Attitude (pp. 53–84). Amsterdam and Philadelphia: John Benjamins. Fries, C. C. (1952). The Structure of English. New York: Harcourt, Brace & World. Gårding, E. (1987). “Speech Act and Tonal Patterns in Standard Chinese: Constancy and Variation.” Phonetica 44: 13–29. Gardner, R. (2001). When Listeners Talk: Response Tokens and Listener Stance. Amsterdam and Philadelphia: John Benjamins. Gibson, J. J., and Pick, A. D. (1963). “Perception of Another Person’s Looking Behavior.” American Journal of Psychology 76 (3): 386–394. Goffman, E. (1953). “Communication Conduct in an Island Community.” Unpublished dissertation. Goffman, E. (1963). Stigma: Notes on the Management of Spoiled Identity. New York: Simon & Schuster.
414
Bibliography
Goffman, E. (1964). “The Neglected Situation.” American Anthropologist 66 (6): pt. II: 133–136. Goffman, E. (1971). Relations in Public: Microstudies of the Public Order. New York: Harper & Row. Goffman, E. (1976). “Replies and Responses.” Language in Society 5: 257–313. Goffman, E. (1981). Forms of Talk. Philadelphia: University of Pennsylvania Press. Golato, A. (2005). Compliments and Compliment Responses: Grammatical Structure and Sequential Organization. Amsterdam/ Philadelphia: John Benjamins. Golato, A. and Betz, E. (2008). “German Ach and Ach So in Repair Uptake: Resources to Sustain or Remove Epistemic Asymmetry.” Zeitschrift für Sprachwissenschaft 27 (1): 7–37. Golato, A., and Fagyal, Z. (2008). “Comparing Single and Double Sayings of the German Response Token Ja and the Role of Prosody: A Conversation Analytic Perspective.” Research on Language and Social Interaction 41 (3): 1–30. Goodwin, C. (1979). “The Interactive Construction of a Sentence in Natural Conversation.” In G. Psathas (Ed.), Everyday Language: Studies in Ethnomethodology (pp. 97–121). New York: Irvington. Goodwin, C. (1980). “Restarts, Pauses, and the Achievement of Mutual Gaze at Turn-Beginning.” Sociological Inquiry 50 (3–4): 272–302. Goodwin, C. (1981). Conversational Organization: Interaction Between Speakers and Hearers. New York: Academic Press. Goodwin, C. (1984). “Notes on Story Structure and the Organization of Participation.” In M. Atkinson and J. Heritage (Eds.), Structures of Social Action: Studies in Conversation Analysis (pp. 225–246). Cambridge University Press. Goodwin, C. (1986). “Between and Within: Alternative and Sequential Treatments of Continuers and Assessments.” Human Studies 9: 205–18. Goodwin, C. (1987). “Forgetfulness as an Interactive Resource.” Social Psychology Quarterly 50 (2): 115–130. Goodwin, C., and Goodwin, M. H. (1987). “Concurrent Operations on Talk: Notes on the Interactive Organization of Assessments.” IPrA Papers in Pragmatics 1 (1): 1–52. Goodwin, C. and Goodwin, M. H. (1992). “Assessments and the Construction of Contexts.” In A. Duranti and C. Goodwin (Eds.), Rethinking Context. Language as an Interactive Phenomenon (pp. 147–190). Cambridge University Press. Goodwin, M. H. (1980). “Processes of Mutual Monitoring Implicated in the Production of Description Sequences.” Sociological Inquiry 50: 303–317.
Bibliography
415
Goodwin, M. H. (1984). “Aggravated Correction and Disagreement in Children’s Conversations.” Journal of Pragmatics 7: 657–77. Greatbatch, D. (1988). “A Turn-Taking System for British News Interviews.” Language in Society 17 (3): 401–430. Gullberg, M. (2003). “Eye Movements and Gestures in Human Interaction.” In J. Hyönä, R. Radach, and H. Deubel (Eds.), The Mind’s Eyes: Cognitive and Applied Aspects of Eye Movements (pp. 685–703). Oxford: Elsevier. Gullberg, M., and Holmqvist, K. (2006). “What Speakers Do and What Listeners Look at: Visual Attention to Gestures in Human Interaction Live and on Video.” Pragmatics and Cognition 14 (1): 53–82. Hacohen, G. and Schegloff, E. A. (2006). “On the Preference for Minimization in Referring to Persons: Evidence from Hebrew Conversation.” Journal of Pragmatics 38: 1305–1312. Hakulinen, A. (1987) “Avoiding Personal Reference in Finnish.” In J. Verschueren and M. Bertucelli-Papi (Eds.), The Pragmatic Perspective (pp. 141–153). Amsterdam and Philadelphia: John Benjamins. Hakulinen, A. (1993). “Inandningen som kulturellt interaktionsfenomen [Inhalation as a Cultural Phenomenon in Interaction].” In A. -M. Ivars, H. Lehti-Eklund, P. Lilius, A.-M. Londen, and H. SolstrandPipping (Eds.), Språk och social context (pp. 49–67). Publications of the department of Nordic Studies, University of Helsinki, Serie B: 15. Hakulinen, A., Keevallik Eriksson, L. and Lindström, J. (2003). “Kuule, kule, hördu: projicerande pratiker i finska, estninska och svenska samtal [‘Listen’: Projecting Practices in Finnish, Estonian and Swedish Conversation].” In B. Nordberg, L. Keevallik Eriksson, K. Thelander, and M. Thelander (Eds.), Grammatik och samtal. Studier till minne av Mats Eriksson. Skrifter utgivna av Institutionen för nordiska språk vid Uppsala universitet, No. 63, pp. 199–218. Hakulinen, A. and Sorjonen, M. -L. (forthcoming). “Designing Utterances for Action: Ways of Repeating the Verb as a Response to an Assessment.” In M. Haakana, M. Laakso, and J. Lindström (Eds.), Comparative studies of talk in interaction. Helsinki: Finnish Literature Society. Hakulinen, A., Vilkuna, M., Korhonen, R., Koivisto, V., Heinonen, T., and Alho, I. (2004). Iso suomen kielioppi [Comprehensive Grammar of Finnish]. Helsinki: Finnish Literature Society. Hale, K. (1983). Warlpiri and the grammar of non-configurational languages. Natural Language and Linguistic Theory 1: 5–47. Hanks, W. F. (1992). “The Indexical Ground of Deictic Reference.” In A. Duranti and C. Goodwin (Eds.), Rethinking Context, Language as an Interactive Phenomenon (pp. 43–77). Cambridge University Press.
416
Bibliography
Hanks, W. F. (2007). “Person Reference in Yucatec Maya Conversation.” In N. Enfield and T. Stivers (Eds.), Person Reference in Interaction: Linguistic, Cultural, and Social Perspectives (pp. 149–171). Cambridge University Press. Hansen, K. C., and Hansen, L. E. (1992). Pintupi/Luritja Dictionary, 3rd edn. Alice Springs: Institute for Aboriginal Development. Harris, M. (1968). The Rise of Anthropological Theory. New York: Thomas Y. Crowell. Harris, R., and Taylor, T. (1997). Landmarks in Linguistic Thought I: The Western Tradition from Socrates to Saussure, 2nd edn. London and New York: Routledge. Haviland, J. (2007). “Person Reference in Tzotzil Gossip: Referring Dupliciter.” In N. J. Enfield and T. Stivers (Eds.), Person Reference in Interaction: Linguistic, Cultural and Social Perspectives (pp. 226–52). Cambridge University Press. Hayano, K. (2007). “Repetitional Agreement and Anaphprocal Agreement: Negotiation of Affiliation and Disaffiliation in Japanese Conversation.” Unpublished MA thesis. University of California at Los Angeles, Applied Linguistics and Teaching English as a Second Language. Hayashi, M. (1999). “Where Grammar and Interaction Meet: A Study of Co-Participant Completion in Japanese Conversation.” Human Studies 22 (2–4): 475–499. Hayashi, M. (2003). Joint Utterance Construction in Japanese Conversation. Amsterdam and Philadelphia: John Benjamins. Hayashi, M. (2004a). “Discourse Within a Sentence: An Exploration of Postpositions in Japanese as an Interactional Resource.” Language in Society 33: 343–376. Hayashi, M. (2004b). “Projection and Grammar: Notes on the ‘ActionProjecting’ Use of the Distal Demonstrative Are in Japanese.” Journal of Pragmatics 36: 1337–1374. Heinämäki, O. (1976). “Problems of Basic Word Order.” In N. E. Enkvist and V. Kohonen (Eds.), Reports on Text Linguistics: Approaches to Word Order. Meddelanden från Stiftelsens för Åbo Akademi Forskningsintitut 8: 95–106. Åbo. Heinemann, T. (2003). “Negation in Interaction in Danish Conversation.” Unpublished Ph.D. thesis, University of York, UK. Heinemann, T. (2006). “‘Will You or Can’t You?’: Displaying Entitlement in Interrogative Requests.” Journal of Pragmatics 38 (7): 1081–1104. Heinemann, T. (2008). “Questions of Accountability: Yes-No Interrogatives That Are Unanswerable.” Discourse Studies 10 (1): 55–71. Heinemann, T. and Traverso, V. (Eds.) (in press). “Complaints, Social Solidarity, and Identity.” Journal of Pragmatics Special issue.
Bibliography
417
Heingartner, N. L. (1996). “The Effect of Age Upon Non-Indefinite -To Use: A Study of the Spoken Russian of Moscow Women.” Unpublished doctoral dissertation, Brown University, Providence, R.I. Helasvuo, M. -L. (2001). Syntax in the Making: The Emergence of Syntactic Constructions in Finnish Conversation. Amsterdam and Philadelphia: John Benjamins. Helbig, G. (1988). Lexikon deutscher Partikeln. Leipzig: Verlag Enzyklopädie. Hentschel, E. and Weydt, H. (1983). “Der Pragmatische Mechanismus: Denn Und Eigentlich.” In H. Weydt (Ed.), Partikeln und Interaktion (pp. 263–273). Tubingen: Niemeyer. Heritage, J. (1984a). “A Change-of-State Token and Aspects of Its Sequential Placement.” In J. M. Atkinson and J. Heritage (Eds.), Structures of Social Action: Studies in Conversation Analysis (pp. 299–345). Cambridge University Press. Heritage, J. (1984b). Garfinkel and Ethnomethodology. Cambridge: Polity Press. Heritage, J. (1988). “Explanations as Accounts: A Conversation Analytic Perspective.” In C. Antaki (Ed.), Understanding Everyday Explanation: A Casebook of Methods (pp. 127–144). Beverly Hills, Calif.: Sage. Heritage, J. (1998). “Oh-Prefaced Responses to Inquiry.” Language in Society 27: 291–334. Heritage, J. (2002a). “Oh-Prefaced Responses to Assessments: A Method of Modifying Agreement/Disagreement.” In C. E. Ford, B. A. Fox, and S. A. Thompson (Eds.), The Language of Turn and Sequence (pp. 196–224). Oxford and New York: Oxford University Press. Heritage, J. (2002b). “The Limits of Questioning: Negative Interrogatives and Hostile Question Content.” Journal of Pragmatics 34: 1427–1446. Heritage, J. (2007). “Intersubjectivity and Progressivity in References to Persons (and Places).” In T. Stivers and N. J. Enfield (Eds.), Person Reference in Interaction: Linguistic, Cultural, and Social Perspectives (pp. 255–280). Cambridge University Press. Heritage, J. and Greatbatch, D. (1986). “Generating Applause: A Study of Rhetoric and Response at Party Political Conferences.” American Journal of Sociology 92 (1): 110–157. Heritage, J., and Greatbatch, D. (1991). “On the Institutional Character of Institutional Talk: The Case of News Interviews.” In D. Boden and D. H. Zimmerman (Eds.), Talk and Social Structure (pp. 93–137). Cambridge: Polity Press. Heritage, J., and Maynard, D. W. (2006). “Introduction: Analyzing Interaction Between Doctors and Patients in Primary Care Encounters.” In J. Heritage and D. W. Maynard (Eds.), Communication in Medical
418
Bibliography
Care: Interaction Between Primary Care Physicians and Patients (pp. 1–21). Cambridge University Press. Heritage, J., and Raymond, G. (2005). “The Terms of Agreement: Indexing Epistemic Authority and Subordination in Assessment Sequences.” Social Psychology Quarterly 68: 15–38. Heritage, J., and Sorjonen, M.-L. (1994). “Constituting and Maintaining Activities Across Sequences: And-Prefacing as a Feature of Question Design.” Language in Society 23 (1): 1–29. Heritage, J., Boyd, E. A., and Kleinman, L. (2001). “Subverting Criteria: The Role of Precedent in Decisions to Finance Surgery.” Sociology of Health and Illness 23 (5): 701–728. Heritage, J., Robinson, J. D., Elliott, M., Beckett, M., and Wilkes, M. (2007). “Reducing Patients’ Unmet Concerns: The Difference One Word Can Make.” Journal of General Internal Medicine 22: 1429–1433. Ho, A. T. (1977). “Intonation Variation in a Mandarin Sentence for Three Expressions: Interrogative, Exclamatory and Declarative.” Phonetica 34: 446–457. Holm, J. and Shilling, A. (1982). Dictionary of Bahamian English. New York: Lexik House. Holt, E., and Drew, P. (2005). “Figurative Pivots: The Use of Figurative Expressions in Pivotal Topic Transitions.” Research on Language and Social Interaction 38 (1): 35–61. Howe, M. L. (1991). “Topic Change in Conversation.” Unpublished Ph.D. Dissertation, University of Kansas. Iwasaki, S. (1997). “The Northridge Earthquake Conversations: The Floor Structure and the ‘Loop’ Sequence in Japanese Conversation.” Journal of Pragmatics 28: 661–693. Iwasaki, S. (2007). “Construction of Units and Interactive Turn Spaces in Japanese Conversation.” In N. McGloin and J. Mori (Eds.), Japanese/ Korean Linguistics 15. Palo Alto, Calif.: CSLI Publications. Iwasaki, S. (forthcoming). “Initiating Interactive Turn Spaces in Japanese Conversation: Local Projection and Collaborative Action.” Discourse Processes. Iwasaki, S. and Ono, T. (2002). “‘Sentence’ in Spontaneous Spoken Japanese Discourse.” In J. Bybee and M. Noonan (Eds.), Complex Sentences in Grammar and Discourse: Essays in Honor of Sandra A. Thompson (pp. 175–202). Amsterdam and Philadelphia: John Benjamins. Jasperson, R. (1998). “Repair After Cut-Off.” Unpublished Ph.D. Dissertation, University of Colorado, Boulder, Col. Jefferson, G. (1972). “Side Sequences.” In D. N. Sudnow (Ed.), Studies in Social Interaction (pp. 294–338). New York: Free Press. Jefferson, G. (1973). “A Case of Precision Timing in Ordinary Conversation: Overlapped Tag-Positioned Address Terms in Closing Sequences.” Semiotica 9: 47–96.
Bibliography
419
Jefferson, G. (1974). “Error Correction as an Interactional Resource.” Language in Society 3: 181–199. Jefferson, G. (1978). “Sequential Aspects of Storytelling in Conversation.” In J. Schenkein (Ed.), Studies in the Organization of Conversational Interaction (pp. 219–248). New York: Academic Press. Jefferson, G. (1981). “The Abominable Ne? An Exploration of PostResponse Pursuit of Response.” In P. Schröder and H. Steger (Eds.), Dialogforschung (pp. 53–88). Dusseldorf: Pädagogischer Verlag Schwann. Jefferson, G. (1983). “Notes on Some Orderlinesses of Overlap Onset.” In V. D’Urso and P. Leonardi (Eds.), Discourse Analysis and Natural Rhetoric (pp. 11–38). Padua: Cleup Editore. Jefferson, G. (1984). “Notes on a Systematic Deployment of the Acknowledgment Tokens ‘Yeah’ and ‘Mm Hm’.” Papers in Linguistics 17: 197–206. Jefferson, G. (1993). “Caveat Speaker: Preliminary Notes on Recipient Topic-Shift Implicature.” Research on Language and Social Interaction 26: 1–30. Jefferson, G. (2004). “At First I Thought.” In G. H. Lerner (Ed.), Conversation Analysis: Studies from the First Generation (pp. 131–67). Amsterdam and Philadelphia: John Benjamins. Jurafsky, D., Bell, A., Fosler-Lussier, E., Girand, C., and Raymond, W. (1998). “Reduction of English Function Words in Switchboard.” Proceedings of ICSLP-98 7: 3111–3114. Kärkkäinen, E., Sorjonen, M.-L., and M.-L. Helasvuo. (2007). “Discourse Structure.” In Timothy Shopen (Ed.) Language Typology and Syntactic Description. Vol. II, Complex Constructions (pp. 301– 371). Cambridge University Press. Kendon, A. (1967). “Some Functions of Gaze-Direction in Social Interaction.” Acta Psychologica 26 (1): 22–63. Kendon, A. (1973). “The Role of Visible Behavior in the Organization of Social Interaction.” In M. Von Cranach and I. Vine (Eds.), Social Communication and Movement: Studies of Interaction and Expression in Man and Chimpanzee (pp. 29–74). New York: Academic Press. Kendon, A. (1977). “Spatial Organization in Social Encounters: The F-Formation System.” In A. Kendon (Ed.), Studies in the Behavior of Social Interaction. Lisse: Peter DeRidder Press. Kendon, A. (1978). “Looking in Conversation and the Regulation of Turns at Talk: A Comment on the Papers of G. Beattie and D. R. Rutter et al.” British Journal of Social and Clinical Psychology 17: 23–24. Kendon, A. (1990). Conducting Interaction: Patterns of Behavior in Focused Encounters. Cambridge University Press. Kidwell, M. (2005). “Gaze as Social Control: How Very Young Children Differentiate ‘The Look’ from a ‘Mere Look’ by Their Adult
420
Bibliography
Caregivers. Research on Language and Social Interaction 38 (4): 417–449. Kidwell, M. (2009). “Gaze Shift as an Interactional Resource for Very Young Children.” Discourse Processes 46: 145–160. Kidwell, M. and Zimmerman, D. H. (2006). “‘Observability’ in the Interactions of Very Young Children.” Communication Monography 73 (1): 1–28. Kidwell, M. and Zimmerman, D. H. (2007). “Joint Attention in Action.” Journal of Pragmatics 39 (3): 592–611. Kim, H. (2004). “Backchannels as Achievements of Social Interaction in Korean Conversation.” Sahoe Onohak [Sociolinguistics] 12: 65–93. Kim, K.-H. (1993). “Other-Initiated Repairs in Korean Conversation as Interactional Resources.” In S. Choi (Ed.), Japanese/Korean Linguistics (Vol. III, pp. 3–18). Palo Alto, Calif.: Center for the Study of Language and Information, Stanford University. Kim, K. -H. (1999a). “Other-Initiated Repair Sequences in Korean Conversation: Types and Functions.” Discourse and Cognition 6: 141–168. Kim, K.-H. (1999b). “Phrasal Unit Boundaries and Organization of Turns and Sequences in Korean Conversation.” Human Studies 22 (2/4): 425–46. Kim, K.-H. (2001). “Confirming Intersubjectivity through Retroactive Elaboration: Organization of Phrasal Units in Other-Initiated Repair Sequences in Korean Conversation.” In M. Selting and E. Couper-Kuhlen (Eds.), Studies in Interactional Linguistics (pp. 345–372). Amsterdam and Philadelphia: John Benjamins. Kim, K.-H. (2007). “Sequential Organization of Post-Predicate Elements in Korean Conversation: Pursuing Uptake and Modulating Action.” Pragmatics 17 (4): 573–603. Kitaigorodskaia, M. V., and Rosanova, N. N. (1994). Russkii rechevoj portret [The portrait of Russian speech]. Moscow. Kitzinger, C. (2000). “Doing Feminist Conversation Analysis.” Feminism and Psychology 10: 163–193. Kitzinger, C. (2005a). “‘Speaking as a Heterosexual’: (How) Does Sexuality Matter for Talk-in-Interaction?” Research on Language and Social Interaction 38 (3): 221–265. Kitzinger, C. (2005b). “Heteronormativity in Action: Reproducing Normative Heterosexuality in ‘After Hours’ Calls to the Doctor.” Social Problems 52 (4): 477–498. Kitzinger, C. (2008). “Developing Feminist Conversation Analysis: A Response to Wowk.” Human Studies 31: 179–208. Kleinke, C. L. (1986). “Gaze and Eye Contact: A Research Review.” Psychological Bulletin 100 (1): 78–100. Kobayashi, H. (1997). “Unique Morphology of the Human Eye.” Nature 387: 767–768.
Bibliography
421
Kobayashi, H. (2001). “Unique Morphology of the Human Eye and Its Adaptive Meaning: Comparative Studies on External Morphology of the Primate Eye.” Journal of Human Evolution 40 (5): 419–435. Koenig, C. J. (2007). “Question Initiated Oblique Sequences: A Recipient’s Practice in Multi-Unit Turn Environments.” Unpublished manuscript, UCLA. Kokuritsu Kokugo Kenkyuujo (National Language Research Institute) (1960). Hanashi Kotoba no Bunkei (1): Taiwa Shiryoo ni yoru Kenkyuu [Sentence Types in Spoken Language (1): Research based on Dialogic Data]. Tokyo: Shuuei Shuppan. Koshik, I. (2002). “A Conversation Analytic Study of Yes/No Questions Which Convey Reversed Polarity Assertions.” Journal of Pragmatics 34: 1851–1877. Koshik, I. (2003). “Wh-Questions Used as Challenges.” Discourse Studies 5: 51–77. Koshik, I. (2005). Beyond Rhetorical Questions: Assertive Questions in Everyday Interaction. Amsterdam and Philadelphia: John Benjamins. Koshik, I. (forthcoming). “Responses to Wh-Question Challenges.” In G. Raymond, G. H. Lerner and J. Heritage (Eds.), Enabling Human Conduct: Naturalistic Studies of Talk-in-Interaction in Honor of Emanuel A. Schegloff. Amsterdam and Philadelphia: John Benjamins. Labov, W. (1970). “The Study of English in Its Social Context.” Studium Generale 23: 30–87. Labov, W., and Fanshel, D. (1977). Therapeutic Discourse: Psychotherapy as Conversation. New York: Academic Press. LaFrance, M. (1974). “Nonverbal Cues to Conversational Turn-Taking Between Black Speakers.” Personality and Social Psychology Bulletin 1 (1): 240–242. LaFrance, M., and Mayo, C. (1976). “Racial Differences in Gaze Behavior During Conversation: Two Systematic Observational Studies.” Journal of Personality and Social Psychology 33 (5): 547–552. Laitinen, L. (2006). “Zero Person in Finnish: A Grammatical Resource for Construing Human Reference.” In M. -L. Helasvuo and L. Campbell (Eds.), Grammar from the Human Perspective: Case, Space and Person in Finnish (pp. 209–231). Amsterdam and Philadelphia: John Benjamins. Laitinen, L. and Hakulinen, A. (2008). “Anaforinen nolla puhutussa ja kirjoitetussa tekstissä [Anaphoric Zero in Spoken and Written Texts], Virittäjä” 2: 162–185. Land, V. and Kitzinger, C. (2005). “Speaking as a Lesbian: Correcting the Heterosexist Presumption.” Research on Language and Social Interaction 38 (4): 371–416.
422
Bibliography
Landis, J., and Koch, G. G. (1977). “The Measurement of Observer Agreement for Categorical Data.” Biometrics 33 (1): 159–174. Lebra, T. (1976). Japanese Patterns of Behavior. Honolulu, Hawaii: University of Hawaii Press. Lenk, U. (1995). “Discourse Markers and Conversational Coherence.” In B. Wårvik, S. -K. Tanskanen and R. Hiltunen (Eds.), Organization of Discourse: Proceedings from the Turku Conference (pp. 341– 352). Turku: University of Turku. Lenk, U. (1998). Marking Discourse Coherence: Functions of Discourse Markers in Spoken English, Vol. XV. Tubingen, Germany: Gunter Narr Verlag. Lerner, G. H. (1987). “Collaborative Turn Sequences: Sentence Construction and Social Action.” Unpublished doctoral dissertation, University of California at Irvine, Calif. Lerner, G. H. (1989). “Notes on Overlap Management in Conversation: The Case of Delayed Completion.” Western Journal of Speech Communication 53 (2): 167–177. Lerner, G. H. (1991). “On the Syntax of Sentences-in-Progress.” Language in Society 20: 441–458. Lerner, G. H. (1996). “On the ‘Semi-Permeable’ Character of Grammatical Units in Conversation: Conditional Entry into the Turn Space of Another Speaker.” In E. Ochs, E. A. Schegloff, and S. A. Thompson (Eds.), Interaction and Grammar (pp. 238–276). Cambridge University Press. Lerner, G. H. (2003). “Selecting Next Speaker: The Context Sensitive Operation of a Context-Free Organization.” Language in Society 32 (2): 177–201. Lerner, G. H. (2004). “On the Place of Linguistic Resources in the Organization of Talk-in-Interaction: Grammar as Action in Prompting a Speaker to Elaborate.” Research on Language and Social Interaction 37: 151–184. Lerner, G. H. and Kitzinger, C. (2007). “Extraction and Aggregation in the Repair of Individual and Collective Self-Reference.” Discourse Studies 9 (4): 526–557. Lerner, G. H. and Takagi, T. (1999). “On the Place of Linguistic Resources in the Organization of Talk-in-Interaction: A Co-Investigation of English and Japanese Grammatical Practices.” Journal of Pragmatics 31 (1): 49–75. Lerner, G. H. and Zimmerman, D. H. (2003). “Action and the Appearance of Action in the Conduct of Very Young Children.” In P. Glenn, C. LeBaron, and J. Mandelbaum (Eds.), Studies in Language and Social Interaction (pp. 441–457). Mahwah, N.J.: Lawrence Erlbaum. Lerner, G. H., Zimmerman, D. H., and Kidwell, M. (forthcoming). “Formal Structures of Practical Tasks: A Resource for Action in the Social Life of Very Young Children.” In C. Goodwin, C. LeBaron and J. Streeck (Eds.), Multimodality and Human Activity: Research
Bibliography
423
on Human Behavior, Action and Communication. Cambridge University Press. Levinson, S. C. (1983). Pragmatics. Cambridge University Press. Levinson, S. C. (2006). “On the Human ‘Interactional Engine’.” In N. J. Enfield, S. C. Levinson (Eds.), Roots of Human Sociality: Culture, Cognition and Interaction (pp. 39–69). Oxford: Berg. Levinson, S. C. (2007). “Optimizing Person Reference: Perspectives from Usage on Rossel Island.” In N. J. Enfield and T. Stivers (Eds.), Person Reference in Interaction: Linguistic, Cultural and Social Perspectives (pp. 29–72). Cambridge University Press. Levinson, S. C., and Brown, P. (2004). “Comparative Feedback: Cultural Shaping of Response Systems in Interaction.” Paper presented at the Workshop on Feedback in Interaction. Max Planck University for Psycholinguistics, Nijmegen, The Netherlands. Li, C. and Thompson, S. A. (1981). Mandarin Chinese: A Functional Reference Grammar. Berkeley, Calif.: University of California. Lindström, A. (1997). “Designing Social Actions: Grammar, Prosody, and Interaction in Swedish Conversation.” Unpublished doctoral dissertation, University of California, Los Angeles, Calif. Lindström, Anna (1999). Language as Social Action: Grammar, Prosody and Interaction in Swedish conversation. (Skrifter utgivna av Institutionen för nordiska spräk, Uppsala universitet 46.) Diss. Uppsala. Local, J., and Kelly, J. (1986). “Projection and ‘Silences’: Notes on Phonetic and Conversational Structure.” Human Studies 9: 185–204. Local, J. K. (2004). “Getting Back to Prior Talk: And-Uh(m) as a BackConnecting Device in British and American English.” In E. CouperKuhlen (Ed.), Sound Patterns in Interaction: Cross-Linguistic Studies of Phonetics and Prosody for Conversation (pp. 377–400). Amsterdam and Philadelphia: John Benjamins. LoCastro, V. (1987). “Aizuchi: A Japanese Conversational Routine.” In L. E. Smith (Ed.), Discourse Across Cultures: Strategies in World Englishes (pp. 101–113). London: Prentice Hall. Luke, K. K. and Pavlidou, T.-S. (Ed.). (2002). Telephone Calls: Unity and Diversity in Conversational Structure Across Languages and Cultures. Amsterdam and Philadelphia: John Benjamins. McCoy, S. G. (2001). “Colloquial Russian Particles -To, Zhe, and Ved’ as Set-Generating (‘Kontrastive’) Markers: A Unifying Analysis.” Unpublished doctoral dissertation, Boston University, Boston, Mass. McHoul, A. (1978). “The Organization of Turns at Formal Talk in the Classroom.” Language in Society 7: 183–213. Maheux-Pelletier, G. and A. Golato (2008) Repair in Membership Categorization in French. Language in Society, 37 (05): 689–712. Mangione-Smith, R., Stivers, T., Elliott, M. N., McDonald, L., and Heritage, J. (2003) “Online Commentary During the Physical Examination: A Communication Tool for Avoiding Inappropriate
424
Bibliography
Antibiotic Prescribing?” Social Science and Medicine 56 (2): 313–320. Maurer, D. W. (1964). Whiz Mob: A Correlation of the Technical Argot of Pickpockets with Their Behavior Pattern, New Haven, Conn.: College and University Press. Maynard, D. W. (1980). “Placement of Topic Changes in Conversation.” Semiotica 30 (3–4): 263–290. Maynard, S. (1986). “On Back-Channel Behavior in Japanese and English Casual Conversation.” Linguistics 24: 1079–1108. Maynard, S. (1989). Japanese Conversation: Self-Contextualization through Structure and Interactional Management. Norwood, N.J.: Ablex. Maynard, S. (1990). “Conversation Management in Contrast: Listener Response in Japanese and American English.” Journal of Pragmatics 14: 397–412. Mazeland, H. (1990). “‘Yes,’ ‘No’ and ‘Mhm’: Variations in Acknowledgment Choices.” In B. Coein, M. de Fornel, and L. Quéré (Eds.), Les Formes de la Conversation (pp. 251–282). Paris: CNET. Mazeland, H., and Huiskes, M. (2001). “Dutch ‘But’ as a Sequential Conjunction: Its Use as a Resumption Marker.” In M. Selting and E. Couper-Kuhlen (Eds.), Studies in Interactional Linguistics (pp. 141–169). Amsterdam and Philadelphia: John Benjamins. Miller, L. (1983). “Aizuchi: Japanese Listening Behavior.” Masters thesis. University of California at Los Angeles, Los Angeles, Calif. Mizutani, N. (1982). “The Listener’s Response in Japanese Conversation.” Sociolinguistic Newsletter 13: 33–38. Mizutani, N. (1988). “Aizuchi ron [On aizuchi].” Nihongogaku 7. Moerman, M. (1974). “Accomplishing Ethnicity.” In R. Turner (Ed.), Ethnomethodology: Selected Readings (pp. 54–68). Harmondsworth: Penguin. Moerman, M. (1977). “The Preference for Self-Correction in a Thai Conversational Corpus.” Language 53: 872–882. Monzoni, C. M. (2008). “Introducing Direct Complaints through Questions: The Interactional Achievement of ‘Pre-Sequences’?” Discourse Studies 10 (1): 73–87. Monzoni, C. M. (in press). “Direct Complaints in (Italian) Calls to the Ambulance: The Use of Negatively Framed Questions.” Journal of Pragmatics. Mori, J. (1999). Negotiating Agreement and Disagreement in Japanese: Connective Expressions and Turn Construction. Amsterdam and Philadelphia: John Benjamins. Mori, J. (2006). “The Workings of the Japanese Token Hee in Informing Sequences: An Analysis of Sequential Context, Turn Shape, and Prosody.” Journal of Pragmatics 38: 1175–1205. Morita, E. (2005). Negotiation of Contingent Talk: The Japanese Interactional Particles Ne and Sa. Amsterdam and Philadelphia: John Benjamins.
Bibliography
425
Morris, D. (1985). Body Watching. Oxford: Equinox. Nakayama, T. and Ichihashi-Nakayama, K. (1997). “Japanese Kedo: Discourse Genre and Grammaticization.” In H. Sohn and J. Haig (Eds.), Japanese/Korean Linguistics (Vol. VI, pp. 607–618). Palo Alto, Calif.: CSLI Publications. Nielsen, G. (1962). Studies in Self-Confrontation. Copenhagen: Munksgaard. Nielsen, M. F. (2002). “Nå! En skiftemarkør med mange funktioner [Nå! A Marker of Change with Many Functions].” Studier i nordisk, 2000–2001: 51–67. Ochs, E. (1984). “Clarification and Culture.” In D. Schiffrin (Ed.), Georgetown University Round Table in Languages and Linguistics (pp. 325–341). Washington, D.C.: Georgetown University Press. Ochs, E. (1991). “Misunderstanding Children.” In N. Coupland, H. Giles, and J. M. Wiemann (Eds.), Miscommunication and Problematic Talk (pp. 44–60). Newbery Park, Calif.: Sage. Ochs, E., Smith, R., and Taylor, C. (1989). “Dinner Narratives as Detective Stories.” Cultural Dynamics 2: 238–257. Ochs, E., Smith, R., and Taylor, C. (1992). “Story-Telling as a TheoryBuilding Activity.” Discourse Processes 15 (1): 37–72. Owen, M. (1985). “The Conversational Function of ‘Anyway’.” The Nottingham Linguistic Circular 14: 72–90. Park, Y.-Y. (1997). “A Cross-Linguistic Study of the Use of Contrastive Connectives in English, Korean, and Japanese Conversation.” Unpublished Ph.D. Dissertation, University of California, Los Angeles, Calif. Park, Y.-Y. (1999). “The Korean Connective Nuntey in Conversational Discourse.” Journal of Pragmatics 31: 191–218. Park, Y-Y. (2004). “Nonnative Speakers’ Use of Yeah in English Spoken Discourse.” Discourse and Cognition 11: 85–105. Pomerantz, A. (1978). “Compliment Responses: Notes on the Co-operation of Multiple Constraints.” In J. Schenkein (Ed.), Studies in the Organization of Conversational Interaction (pp. 79–112). New York: Academic Press. Pomerantz, A. (1984). “Agreeing and Disagreeing with Assessments: Some Features of Preferred/Dispreferred Turn Shapes.” In J. M. Atkinson and J. Heritage (Eds.), Structures of Social Action: Studies in Conversation Analysis (pp. 57–101). Cambridge University Press. Pomerantz, A. (1986). “Extreme Case Formulations: A Way of Legitimizing Claims.” Human Studies 9: 210–229. Pomerantz, A. M. (1975). “Second Assessments: A Study of Some Features of Agreements/Disagreements.” Unpublished doctoral dissertation, University of California at Irvine, California, Calif. Quirk, R., Greenbaum, S., Leech, G., and Svartvik, J. (1985). A Comprehensive Grammar of the English Language. New York: Longman.
426
Bibliography
Rathmayr, R. (1989). “Russische partikeln und ihre deutschen aquivalente glossar.” Rusistik 1: 18–40. Raymond, G. (2000) “The Structure of Responding: Type-Conforming and Non-Conforming Responses to Yes/No Interrogatives.” Ph.D. Dissertation, Department of Sociology, University of California, Los Angeles, Calif. Raymond, G. (2003). “Grammar and Social Organization: Yes/No Interrogatives and the Structure of Responding.” American Sociological Review 68 (6): 939–967. Raymond, G. (2004). “Prompting Action: The Stand-Alone ‘So’ in Ordinary Conversation.” Research on Language and Social Interaction 37 (2): 185–218. Raymond, G., and Heritage, J. (2006). “The Epistemics of Social Relations: Owning Grandchildren.” Language in Society 35 (5): 677–705. Reisman, K. (1974). “Contrapuntal Conversations in an Antiguan Village.” In R. Bauman and J. Sherzer (Eds.), Explorations in the Ethnography of Speaking (pp. 110–124). Cambridge University Press. Robinson, J. (2006). “Managing Trouble Responsibility and Relationships During Conversational Repair.” Communication Monographs 73: 137–161. Robinson, J. and Heritage, J. (2005). “The Structure of Patients’ Presenting Concerns: The Completion Relevance of Current Symptoms.” Social Science and Medicine 61: 481–493. Rossano, F. (2005). “On Sustaining Vs. Withdrawing Gaze in Face-toFace Interaction.” Paper presented at the 91st Annual Convention National Communication Association (NCA). Rossano, F. (2006a). Gaze Behavior in Multi-Unit Turns. Paper presented at the 92nd Annual Convention National Communication Association (NCA). Rossano, F. (2006b). When the Eyes Meet: Using Gaze to Mobilize Response. Paper presented at the International Conference on Conversation Analysis (ICCA). Rossano, F. (2009). “Gaze Behavior in Face-to-Face Interaction.” Ph.D. dissertation. Rossano, F. (in press). “Questions and Responding in Italian.” In N. J. Enfield, S. C. Levinson, and T. Stivers (Eds.), Questions and Responses in 10 Languages, Special Issue of Journal of Pragmatics. Roth, A. (2002). “Social Epistemology in Broadcast News Interviews.” Language in Society 31 (3): 355–381. Rutter, D. R. (1984). Looking and Seeing: The Role of Visual Communication in Social Interaction. Chichester: Wiley. Sacks, H. (1972a). “An Initial Investigation of the Usability of Conversational Data for Doing Sociology.” In D. N. Sudnow (Ed.), Studies in Social Interaction (pp. 31–74). New York: Free Press.
Bibliography
427
Sacks, H. (1972b). “On the Analyzability of Stories by Children.” In J. J. Gumperz and D. Hymes (Eds.), Directions in Sociolinguistics: The Ethnography of Communication (pp. 325–345). New York: Holt, Rinehart & Winston. Sacks, H. (1974). “An Analysis of the Course of a Joke’s Telling in Conversation.” In R. Bauman and J. Sherzer (Eds.), Explorations in the Ethnography of Speaking (pp. 337–353). Cambridge University Press. Sacks, H. (1978). “Some Technical Considerations of a Dirty Joke.” In J. Schenkein (Ed.), Studies in the Organization of Conversational Interaction (pp. 249–269). New York: Academic Press. (Edited by Gail Jefferson from four lectures delivered at the University of California, Irvine, Calif., fall 1971). Sacks, H. (1984). “On Doing ‘Being Ordinary’.” In J. M. Atkinson and J. Heritage (Eds.), Structures of Social Action: Studies in Conversation Analysis (pp. 413–429). Cambridge University Press. Sacks, H. (1987) “On the Preferences for Agreement and Contiguity in Sequences in Conversation.” In G. Button and J. Lee (Eds.), Talk and Social Organization (pp. 54–69). Clevedon: Multilingual Matters. Sacks, H. (1995). Lectures on Conversation. Edited by G. Jefferson, with an introduction by E. A. Schegloff. Oxford: Blackwell. Originally published in two volumes in 1992. Sacks, H. and Schegloff, E. A. (1979). Two Preferences in the Organization of Reference to Persons and Their Interaction.” In G. Psathas (Ed.), Everyday Language: Studies in Ethnomethodology (pp. 15–21). New York: Irvington Publishers. Sacks, H., Schegloff, E. A., and Jefferson, G. (1974). “A Simplest Systematics for the Organization of Turn-Taking for Conversation.” Language 50 (4): 696–735. Sadler, M. (2006). “A Blurring of Categorization: The Japanese Connective De in Spontaneous Conversation.” Discourse Studies 8 (2): 303–323. Sadock, J. M., and Zwicky, A. (1985). “Speech Act Distinctions in Syntax.” In T. Shopen (Ed.), Language Typology and Syntactic Description: Clause Structure (Vol. I, pp. 155–196). Cambridge University Press. Schack, K. (2000). “Comparison of Intonation Patterns in Mandarin and English for a Particular Speaker.” University of Rochester Working Papers in the Language Sciences (spring/summer). Scheflen, A. E. (1964). “The Significance of Posture in Communication Systems.” Psychiatry 26: 316–331. Schegloff, E. A. (1968). “Sequencing in Conversational Openings.” American Anthropologist 70: 1075–1095. Schegloff, E. A. (1970). Opening Sequences. University of California, Los Angeles, Calif.
428
Bibliography
Schegloff, E. A. (1979). “The Relevance of Repair for to a Syntax-forConversation.” In T. Givon (Ed.), Syntax and Semantics (Vol. XII, pp. 261–286). New York: Academic Press. Schegloff, E. A. (1980). “Preliminaries to Preliminaries: ‘Can I Ask You a Question’.” Sociological Inquiry 50: 104–152. Schegloff, E. A. (1982). “Discourse as an Interactional Achievement: Some Uses of ‘Uh Huh’ and Other Things That Come Between Sentences.” In D. Tannen (Ed.), Georgetown University Round Table on Languages and Linguistics, Analyzing Discourse: Text and Talk (pp. 71–93). Washington, D.C.: Georgetown University Press. Schegloff, E. A. (1984). “On Some Questions and Ambiguities in Conversation.” In J. M. Atkinson and J. Heritage (Ed.), Structures of Social Action: Studies in Conversation Analysis (pp. 28–52). Cambridge University Press. Schegloff, E. A. (1987a). “Between Macro and Micro: Contexts and Other Connections.” In J. Alexander, B. Giesen, R. Munch, and N. Smelser (Eds.), The Micro-Macro Link (pp. 207–234). Berkeley, Calif.: University of California Press. Schegloff, E. A. (1987b). “Analyzing Single Episodes of Interaction: An Exercise in Conversation Analysis.” Social Psychology Quarterly 50 (2): 101–114. Schegloff, E. A. (1987c). “Recycled Turn Beginnings: A Precise Repair Mechanism in Conversation’s Turn-taking Organization.” In G. Button and J. R. E. Lee (Eds.), Talk and Social Organization (pp. 70–85). Clevedon: Multilingual Matters. Schegloff, E. A. (1988). “On an Actual Virtual Servo-Mechanism for Guessing Bad News: A Single Case Conjecture.” Social Problems 35 (4): 442–457. Schegloff, E. A. (1990). “On the Organization of Sequences as a Source of ‘Coherence’ in Talk-in-Interaction.” In B. Dorval (Ed.), Conversational Organization and Its Development (pp. 51–77). Norwood, N.J.: Ablex. Schegloff, E. A. (1991). “Reflections on Talk and Social Structure.” In D. Boden and D. H. Zimmerman (Eds.) Talk and Social Structure (pp. 44–70). Berkeley, Calif.: University of California Press. Schegloff, E. A. (1992a). “Repair After Next Turn: The Last Structurally Provided Defense of Intersubjectivity in Conversation.” American Journal of Sociology 97: 1295–1345. Schegloff, E. A. (1992b). “On Talk and Its Institutional Occasions.” In P. Drew and J. Heritage (Eds.), Talk at Work: Interaction in Institutional Settings (pp. 101–134). Cambridge University Press. Schegloff, E. A. (1993). “Reflections on Quantification in the Study of Conversation.” Research on Language and Social Interaction 26: 99–128. Schegloff, E. A. (1995a). “Parties and Talking Together: Two Ways in Which Numbers Are Significant for Talk-in-Interaction.” In
Bibliography
429
P. T. Have and G. Psathas (Eds.), Situated Order: Studies in Social Organization and Embodied Activities (pp. 31–42). Washington, D.C.: University Press of America. Schegloff, E. A. (1995b). Sequence Organization. Los Angeles, Calif.: University of California. Schegloff, E. A. (1996a). “Some Practices for Referring to Persons in Talk-in-Interaction: A Partial Sketch of a Systematics.” In B. A. Fox (Eds.), Studies in Anaphora (pp. 437–85). Amsterdam and Philadelphia: John Benjamins. Schegloff, E. A. (1996b). “Confirming Allusions: Toward an Empirical Account of Action.” American Journal of Sociology 102 (1): 161–216. Schegloff, E. A. (1996c). “Turn Organization: One Intersection of Grammar and Interaction.” In E. Ochs, E. A. Schegloff and S. A. Thompson (Eds.), Interaction and Grammar (pp. 52–133). Cambridge University Press. Schegloff, E. A. (1997a). “Whose Text? Whose Context?” Discourse and Society 8 (2): 165–187. Schegloff, E. A. (1997b). “Practices and Actions: Boundary Cases of Other-Initiated Repair.” Discourse Processes 23: 499–545. Schegloff, E. A. (1998a). “Reflections on Studying Prosody in Talk-inInteraction.” Language and Speech 41: 235–263. Schegloff, E. A. (1998b). “Reply to Wetherell.” Discourse and Society 9 (3): 457–60. Schegloff, E. A. (1998c). “‘Body Torque.” Social Research 65 (3): 535–596. Schegloff, E. A. (1999a). “‘Schegloff’s Texts’ as ‘Billig’s Data’: A Critical Reply to Billig.” Discourse and Society 10: 558–571. Schegloff, E. A. (1999b). “Naivete vs. Sophistication or Discipline vs. Self-Indulgence: A Rejoinder to Billig.” Discourse and Society 10: 577–582. Schegloff, E. A. (2000a). “On Granularity.” Annual Review of Sociology 26: 715–720. Schegloff, E. A. (2000b). “When ‘Others’ Initiate Repair.” Applied Linguistics 21: 205–243. Schegloff, E. A. (2000c). “Overlapping Talk and the Organization of Turn-Taking for Conversation.” Language in Society 29 (1): 1–63. Schegloff, E. A. (2001). “Conversation Analysis: A Project in Process – ‘Increments’.” Forum lecture, LSA Linguistic Institute, University of California, Santa Barbara, Calif. Schegloff, E. A. (2002a). “Reflections on Research on Telephone Conversation: Issues of Cross-Cultural Scope and Scholarly Exchange, Interactional Import and Consequences.” In K. K. Luke and T. S. Pavlidon (Eds.), Telephone Calls: Unity and Diversity in Conversational Structure Across Languages and Cultures (pp. 249–281). Amsterdam and Philadelphia: John Benjamins.
430
Bibliography
Schegloff, E. A. (2002b). “Accounts of Conduct in Interaction: Interruption, Overlap and Turn-Taking.” In J. H. Turner (Ed.), Handbook of Sociological Theory (pp. 287–321). New York: Plenum. Schegloff, E. A. (2002c). “Conversation Analysis, Then and Now.” Plenary Address for the Inaugural Session of the Section-in-Formation on Ethnomethodology and Conversation Analysis of the American Sociological Association, Chicago, Ill. Schegloff, E. A. (2004). “On Dispensability.” Research on Language and Social Interaction 37 (2): 95–149. Schegloff, E. A. (2006). “Interaction: The Infrastructure for Social Institutions, the Natural Ecological Niche for Language, and the Arena in which Culture is Enacted.” In N. J. Enfield and S. C. Levinson (Eds.), Roots of Human Sociality: Culture, Cognition and Interaction (pp. 70–96). Oxford: Berg. Schegloff, E. A. (2007a). “A Tutorial on Membership Categorization.” Journal of Pragmatics 39: 462–482. Schegloff, E. A. (2007b). Sequence Organization in Interaction: A Primer in Conversation Analysis I. Cambridge University Press. Schegloff, E. A. and Sacks, H. (1973). “Opening Up Closings.” Semiotica 8: 289–327. Schegloff, E. A. and Sacks, H. (1974). “Opening Up Closings.” In R. Turner (Ed.), Ethnomethodology: Selected Readings (pp. 233–264). Harmondsworth: Penguin. Schegloff, E. A., Jefferson, G., Sacks, H. (1977). “The Preference for Self-Correction in the Organization of Repair in Conversation.” Language 53: 361–382. Schegloff, E. A., Ochs, E., Thompson, S. (1996). “Introduction.” In E. Ochs, E. A. Schegloff, and S. A. Thompson (Eds.), Interaction and Grammar (pp. 1–51). Cambridge University Press. Schiffrin, D. (1985). “Conversational Coherence: The Role of ‘Well’.” Language 61 (3): 640–647. Schiffrin, D. (1987). Discourse Markers. Cambridge University Press. Schubiger, M. (1972). “English Intonation and German Modal Particles: A Comparative Study.” In D. Bolinger (Ed.), Intonation (pp. 175– 193). Harmondsworth: Penguin. Schwitalla, J. (2003). Gesprochenes Deutsch: Eine Einführung. Berlin: Erich Schmitdt. Sebeok, T. (1981) Nonverbal Communication, Interaction, and Gesture: Selections from Semiotica, The Hague and New York: Mouton Publishers. Selting, M. (1987a). “Fremdkorrekturen als Manifestationsformen von Verständigungsproblemen.” Zeitschrift für Sprachwissenschaft 6: 37–58. Selting, M. (1987b). “Reparaturen Und Lokale Verstehensprobleme Oder: Zur Binnenstruktur Von Reparatursequenzen.” Linguistische Berichte 108: 128–149.
Bibliography
431
Selting, M. (1987c). Verständigungsprobleme. Eine Empirische Analyse Am Beispiel Der Bürger-Verwaltungs-Kommunikation. Tubingen: Niemeyer. Selting, M. (1988). “The Role of Intonation in the Organization of Repair and Problem Handling Sequences in Conversation.” Journal of Pragmatics 12: 293–322. Selting, M. (1992). “Prosody in Conversational Questions.” Journal of Pragmatics 15: 583–588. Selting, M. (1995). Prosodie im Gespräch. Tubingen: Niemeyer. Selting, M. (1996). “Prosody as an Activity-Type Distinctive Cue in Conversation: The Case of So-Called ‘Astonished’ Questions in Repair Initiation.” In E. Couper-Kuhlen and M. Selting (Eds.), Prosody in Conversation: Interactional Studies (pp. 231–270). Cambridge University Press. Shen, X.-N. S. (1989). The Prosody of Mandarin Chinese. Berkeley, Calif.: University of California Press. Shopen, T. (Ed.). (1985). Language Typology and Syntactic Description, 3 vols. Cambridge University Press. Sidnell, J. (2000). “Primus Inter Pares: Story-Telling and Male-Peer Groups in an Indo-Guyanese Rumshop.” American Ethnologist 27 (1): 72–99. Sidnell, J. (2001). “Conversational Turn-Taking in a Caribbean English Creole.” Journal of Pragmatics 33: 1263–1290. Sidnell, J. (2005a). Talk and Practical Epistemology: The Social Life of Knowledge in a Caribbean Community. Amsterdam and Philadelphia: John Benjamins. Sidnell, J. (2005b). “Gesture in the Pursuit and Display of Recognition: A Caribbean Case Study.” Semiotica 156 (1/4): 55–87. Sidnell, J. (2006). “Repair.” In J. Verschueren and J. -O. Östman (Eds.), Handbook of Pragmatics (pp. 1–38). Amsterdam and Philadelphia: John Benjamins. Sidnell, J. (2007a). “Repairing Person Reference in a Small Caribbean Community.” In N. J. Enfield and T. Stivers (Eds.), Person Reference in Interaction: Linguistic, Cultural, and Social Perspectives (pp. 281–308). Cambridge University Press. Sidnell, J. (2007b). “‘Look’-Prefaced Turns in First and Second Position: Launching, Interceding and Redirecting Action.” Discourse Studies 9 (3): 387–408. Sidnell, J. (2007c). “Comparative Studies in Conversation Analysis.” Annual Review of Anthropology 36: 229–44. Simmel, G. (1969). “Sociology of the Senses: Visual Interaction.” In R. E. Park and W. Burgess (Eds.), Introduction to the Science of Sociology (pp. 356–361). University of Chicago Press. Sommer, R. (1959). “Studies in Personal Space.” Sociometry 22 (3): 247–260. Sommer, R. (1962). “The Distance of Comfortable Conversation: A Further Study.” Sociometry 25 (1): 111–116.
432
Bibliography
Sorjonen, M-L. (1996). “On Repeats and Responses in Finnish Conversations.” In E. Ochs, E. A. Schegloff, and S. A. Thompson, (Eds.), Interaction and Grammar (pp. 277–327). Cambridge University Press. Sorjonen, M-L. (1997). “Recipient Activities: Particles Nii(n) and Joo as Responses in Finnish Conversation.” Unpublished doctoral dissertation, University of California, Los Angeles, Calif. Sorjonen, M-L. (2001). Responding in Conversation: A Study of Response Particles in Finnish. Amsterdam and Philadelphia: John Benjamins. Sorjonen, M-L. (2002). “Recipient Activities: The Particle No as a Go-Ahead Response in Finnish Conversations.” In C. E. Ford, B. A. Fox, and S. A. Thompson (Eds.), The Language of Turn and Sequence (pp. 165–195). Oxford and New York: Oxford University Press. Sperber, D. I., and Wilson, D. (1986). Relevance: Communication and Cognition. Oxford: Blackwell. Stivers, T. (2004). “ ‘No No No’ and Other Types of Multiple Sayings in Social Interaction.” Human Communication Research 30 (2): 250–293. Stivers, T. (2005). “Modified Repeats: One Method for Asserting Primary Rights from Second Position.” Research on Language and Social Interaction 38 (2): 131–158. Stivers, T. (2007). “Alternative Recognitionals.” In N. J. Enfield and T. Stivers (Eds.), Person Reference in Interaction: Linguistic, Cultural, and Social Perspectives. Cambridge University Press. Stivers, T., Enfield, N. J. and Levinson S.C. (Eds.) (in press). “A Comparison of Question–Response Sequences Across 10 Languages.” Journal of Pragmatics. Stivers, T. and Rossano, F. (in press) “Mobilizing Response.” Research on Language and social Interaction. Stivers, T. and Majid, A. (2007). “Questioning Children: Interactional Evidence of Implicit Bias in Medical Interviews.” Social Psychology Quarterly 70 (4): 424–441. Stivers, T., and Robinson, J. (2006). “A Preference for Progressivity in Interaction.” Language in Society 35 (3): 367–392. Stivers, T. and Sidnell, J. (2005) “Multi-Modal Interaction.” Semiotica 156 (1/4): 1–20. Stivers, T., Enfield, N. J. and Levinson S. C. (2007) “Person Reference in Interaction.” In N. J. Enfield and T. Stivers (Eds.), Person Reference in Interaction: Linguistic, Cultural, and Social Perspectives. Cambridge University Press. Stivers, T., Mangione-Smith, R., Elliott, M. N., McDonald, L., and Heritage, J. (2003). “Why Do Physicians Think Parents Expect Antibiotics? What Parents Report Vs What Physicians Perceive.” The Journal of Family Practice 52 (2): 140–148.
Bibliography
433
Stokoe, E., and Edwards, D. (2008). “‘Did You Have Permission to Smash Your Neighbour’s Door?’ Silly Questions and Their Answers in Police-Suspect Interrogations.” Discourse Studies 10 (1): 89–111. Tainio, L. (1993). “Kannanoton tulkinta keskustelussa.” [Interpretation of Assessment in Conversation.] Unpublished licentiate’s thesis. Department of Finnish Language and Literature, University of Helsinki. Takahara, P. O. (1998). “Pragmatic Functions of the English Discourse Marker Anyway and Its Corresponding Contrastive Japanese Discourse Markers.” In A. H. Jucker and Y. Ziv (Eds.), Discourse Markers: Descriptions and Theory (pp. 327–351). Amsterdam and Philadelphia: John Benjamin. Tanaka, H. (1999). Turn-Taking in Japanese Conversation: A Study in Grammar and Interaction. Amsterdam and Philadelphia: John Benjamins. Tanaka, H. (2000a). “The Particle Ne as a Turn-Management Device in Japanese Conversation.” Journal of Pragmatics 32: 1135–1176. Tanaka, H. (2000b). “Turn-Projection in Japanese Talk-inInteraction.” Research on Language and Social Interaction 33 (1): 1–38. Tanaka, H. (2004). “Prosody for Marking TRPs in Japanese Conversation.” In E. Couper-Kuhlen and C. E. Ford (Eds.), Sound Patterns in Interaction: Cross-Linguistic Studies from Conversation (pp. 63–96). Amsterdam and Philadelphia: John Benjamins. Tao, L., Fox, B., and De Garcia, J. (1999). “Tone-Choice Repair in Conversational Mandarin Chinese.” In F. Fox, D. Jurafsky, and L. A. Michaelis (Eds.), Cognition and Function in Language (pp. 268–281). Palo Alto, Calif.: CSLI. Terasaki, A. K. (1976). “Pre-Announcement Sequences in Conversation.” Social Science Working Paper, #99. Irvine, University of Cali fornia, Irvine, School of Social Science. Reprinted in Gotto Lerner (Ed.), Conversation Analysis: Studies from the First Generation (pp. 171–219). Amsterdam and Philadelphia: John Benjamins. Thompson, S. A. and Couper-Kuhlen, E. (2005). “The Clause as a Locus of Grammar and Interaction.” Discourse Studies 7 (4/5): 481–505. Thurmair, M. (1989). Modalpartikeln und ihre Kombinationen. Tubingen: Niemeyer. Thurmair, M. (1991). “Zum Gebrauch der Modalpartikeln denn in Fragesätzen. Eine Korpusbasierte Untersuchung.” In E. Klein, F. Rouradier Duteil, and K. H. Wagner (Eds.), Betriebslinguistik und Linguistikbetrieb: Akten des 24. Linguistischen Kolloquiums, Universität Bremen, 4. bis 6. September 1989 (pp. 377–387). Tubingen: Niemeyer.
434
Bibliography
Tomasello, M. (1999). The Cultural Origins of Human Cognition. Cambridge, Mass.: Harvard University Press. Tseng, C.-Y. (2003). “On the Role of Intonation in the Organization of Mandarin Chinese Speech Prosody.” Paper presented at Eurospeech 2003/Interspeech 2003, Geneva, Switzerland. Uhmann, S. (2001). “Some Arguments for the Relevance of Syntax to Same-Sentence Self-Repair in Every German Conversation.” In M. Selting and E. Couper-Kuhlen (Eds.), Studies in Interactional Linguistics (pp. 373–404). Amsterdam and Philadelphia: John Benjamins. Vasilyeva, A. N. (1972). Particles in Colloquial Russian. Moscow: Progress Publishers. Vilkuna, M. (1989). Free Word Order: Its Syntax and Discourse Functions. Helsinki: Finnish Literature Society. Walsh, M. (1991). “Conversational Styles and Intercultural Communication: An Example from Northern Australia.” Australian Journal of Communication 18 (1) pp. 1–12. Wetherell, M. (1998). “Positioning and Interpretive Repertoires: Conversation Analysis and Post-Structuralism in Dialogue.” Discourse and Society 9: 387–412. White, S. (1989). “Backchannels Across Cultures: A Study of Americans and Japanese.” Language in Society 18: 59–76. Wilson, P. J. (1973) Crab Antics: The Social Anthropology of EnglishSpeaking Negro Societies of the Caribbean. New Haven, Conn.: Yale University Press. Winford, D. (1993). Predication in Caribbean English Creoles, Amsterdam: John Benjamins. Wittgenstein, L. (1969). On Certainty. Edited by G. E. M. Anscombe and G. H. von Wright. Translated by Denis Paul and G. E. M. Anscombe. Oxford: Blackwell. Wong, J. (2000). “Repetition in Conversation: A Look at ‘First and Second Sayings’.” Research on Language and Social Interaction 33 (4): 407–424. Wootton, A. J. (1994). “Object Transfer, Intersubjectivity and Third Position Repair: Early Developmental Observations of One Child.” Journal of Child Language 21: 543–564. Wouk, F. (2005) “The Syntax of Repair in Indonesian.” Discourse Studies 7 (2): 237–258. Wu, R.-J. R. (2004). Stance in Talk: A Conversation Analysis of Mandarin Final Particles. Amsterdam and Philadelphia: John Benjamins. Wu, R. -J. R. (2006). “Initiating Repair and Beyond: The Use of Two Repeat-Formatted Repair Initiations in Mandarin Conversation.” Discourse Processes 4: 67–109. Yoon, K. E. (2006). “Complaint Talk in Korean Conversation.” Ph.D. Dissertation. University of Illinois at Urbana-Champaign, Ill.
Bibliography
435
Young, R. and Lee, J. (2004). “Identifying Units in Interaction: Reactive Tokens in Korean and English Conversations.” Journal of Sociolinguistics 8: 380–407. Yuan, J., Shih, C., and Kochanski, G. P. (2002). “Comparison of Declarative and Interrogative Intonation in Chinese.” Proceedings of Speech Prosody (pp. 711–714). Yuan, J. and Shih, C. (2004). “Confusability of Chinese Intonation.” Proceedings of Speech Prosody (pp. 131–134). Zeng, X., Martin, P., and Boulakia, G. (2004). “Tones and Intonation in Declarative and Interrogative Sentences in Mandarin.” Paper presented at the International Symposium on Tonal Aspects of Languages: With Emphasis on Tone Languages, Beijing, China. Zhang, W. (1998). “Repair in Chinese Conversation.” Unpublished Ph.D. thesis, University of Hong Kong. Zifonun, G., Hoffmann, L. and Strecker, B. (1997). Grammatik der deutschen Sprache (3 Bände). Berlin and New York: de Gruyter. Zimmerman D. (1999). “Horizontal and Vertical Comparative Research in Language and Social Interaction.” Research on Language and Social Interaction 32: 195–203. Zimmerman, D. (1993). “Acknowledgment Tokens and Speakership Incipiency Revisited.” Research on Language and Social Interaction 26: 179–194.
Index
Aboriginal Australians, 192 account, 51, 117, 135, 136, 139, 144, 182, 287, 346 accountability, 80, 101, 159, 190 adjacency pair sequences, 176, 205, 314, 330 advice, 142, 297, 298, 299 affiliation, 157, 287. See disaffiliation agreement, 7, 19, 24, 116, 117, 118, 135, 136, 141, 151, 157, 241, 257, 258, 260, 281, 282, 283, 287, 288, 290, 291, 292, 295, 297, 299, 300, 301, 305, 306, 310, 314, 318, 319, 321, 322, 323, 349, 374 Aijmer, Karin, 327 aizuchi, 23, 268 alignment, 25, 32, 33, 39, 40, 42, 127, 160, 168, 307, 321 American, 151 American English, 106, 118, 129, 230, 232, 233, 305, 307, 309 anaphora, 283, 287, 288, 290, 292, 300, 302 announcement, 24, 105, 123, 125, 126, 147, 253, 255, 332 answering. See question-answer sequences answers. See question-answer sequences anthropology, 9, 10, 357, 362, 364, 372 assessments, 142, 281–302, 304–24 Auer, Peter, 104, 121 Bavelas, Janet, 191, 192 Beach, Wayne, 250, 251, 270, 327 Beattie, Geoff, 190, 215 Bequia, 307, 308 Betz, 124, 131
B-event, 154, 172, 180, 205, 310, 396 Bikol, 61, 63, 69, 70, 71, 77, 80, 81, 83, 85, 87, 88, 89, 91, 93, 94, 97, 100, 387 Billig, Michael, 372 bisyllabic. See syllable structure Bolden, Galina, 7, 25, 326, 327, 329, 372, 375, 388 British English, 324 Brown, Penelope, 3, 26, 187, 192, 212, 226, 230, 231, 240, 246, 247, 248, 375, 377, 393, 405 Button, Graham, 118, 343 Caribbean English Creole, 304–24 Casey, Neil, 118, 343 categories, 10, 15, 72, 73, 100, 129, 143, 206, 241, 309, 361, 372, 389, 390, 392, 396 challenge, 40, 43, 45, 50, 170, 176, 216, 341 change-of-state. See ‘oh’ China, 32, 37, 42 Chinantec, 61, 63, 69, 70, 71, 78, 81, 83, 85, 86, 87, 88, 91, 93, 95, 99, 100, 101 clarification, 13, 14, 15, 38, 41, 45, 205, 213, 318, 326, 330, 332, 335, 350, 397 Clayman, Steve, 400, 404 Clift, Rebecca, 159, 167 closings, 144, 188, 236, 285, 318, 341, 349 coding, 62, 74, 193, 203, 207, 213, 221, 223, 232, 241, 388, 390, 392, 393, 395 comparability, 63, 378, 383 complaints, 118, 161, 176, 201, 206, 298, 316, 341, 346, 376, 396
Index confirmation, 6, 7, 33, 34, 35, 37, 40, 41, 45, 47, 49, 53, 56, 120, 121, 122, 123, 146, 181, 196, 205, 213, 214, 215, 216, 233, 241, 288, 290, 306, 320, 322 confirmation request-questions. See confirmation contiguity, 24, 135, 141, 314, 326, 327, 346, 349, 351 conversational style, 192 copula, 281, 287, 289 corrections, 45, 130, 155, 307 Couper-Kuhlen, Elizabeth, 32, 106, 139, 328 cross-cultural comparison, 8, 11, 13, 14, 15, 58, 192, 208, 214, 225, 231, 270, 358, 363 cross-cultural pragmatics, 14, 15 cross-linguistic comparison, 60–102. See cross-cultural comparison cut-off, 73, 74, 380, 381, 403 Danish, 7, 140, 159–85, 374 Darwin, Charles, 189, 231, 240 deixis, 5, 242, 283 deletions, 63, 384 demonstratives, 13, 66, 121, 320 disaffiliation, 216, 297. See affiliation disagreement, 21, 33, 35, 47, 57, 135, 136, 144, 152, 157, 261, 262, 263, 293, 307, 318 disalignment. See alignment doubles, 105, 111, 120, 120–23 Drew, Paul, 32, 33, 40, 105, 106, 107, 120, 129, 159, 167, 253, 255, 258, 271, 343, 359 Duncan, Starkey, 17, 190, 191, 231, 232, 250 Dutch, 172, 173, 183, 350 Egbert, Maria, 11, 13, 24, 31, 32, 58, 60, 114, 130, 131, 139, 306, 372, 373 empathy, 268, 269 Enfield, Nicholas, 16, 25, 130, 188, 362, 363, 364, 365, 366, 367, 368, 369, 372 epistemic stance, 47, 48, 50 epistemics, 24, 32, 40, 41, 43, 45, 46, 47, 50, 54, 56, 57, 58, 139, 140, 141, 142, 185, 301, 305, 319, 321, 322 ethnography, 192, 239, 306, 320 Evans, Nicholas, 192
437 Fagyal, Zsuzsanna, 106, 159, 163, 164, 170, 186 Fanshel, David, 139, 154, 172, 180, 205, 396 Fincke, Steven, 60, 375 finite verb, 283, 289, 290, 300 Finnish, 4, 7, 25, 61, 63, 70, 71, 77, 81, 82, 83, 85, 87, 90, 91, 94, 95, 100, 139, 142, 281–302, 387 Ford, Cecilia, 5, 73, 106 Fox, Barbara, 4, 5, 12, 13, 21, 22, 23, 24, 31, 58, 60, 62, 73, 101, 271, 272, 306, 351, 375, 377, 378, 380, 381, 384, 387, 388, 389, 390, 394, 395, 397 Franck, D, 128 French, 10, 20, 322, 323 Gardner, Rod, 250, 251 gaze, 72, 75, 97, 187–241, 266, 310, 311, 360, 376, 377, 386, 393, 394, 395, 397, 399 avoidance, 192 cultural differences in, 210, 211 ecological determinants of, 193, 227 in European conversation, 192 in Italian, 200 in questions, 200, 202 in responses, 192, 202 in sequence organization, 206 in turn-taking, 233 in Yélî Dnye, 192 recipient gaze, 200, 207, 208 speaker gaze, 200, 202, 207, 208 universal patterns of, 203 gaze aversion, 189, 231 gaze immobility, 221 gaze mobility, 200 gender, grammatical, 12, 113, 121 German, 4, 11, 31, 33, 104–32, 163, 164, 167, 170, 172, 183, 184, 185, 314, 373, 374 gesture, 8, 14, 230, 315, 376 Goffman, Erving, 26, 229, 244, 306, 401 Golato, Andrea, 106, 124, 131, 132, 159, 163, 164, 170, 322, 372, 373 Goodwin, Charles, 9, 25, 63, 75, 152, 187, 190, 191, 200, 205, 212, 229, 231, 232, 250, 251, 281, 324, 360, 377, 386 Goodwin, Marjorie Harness, 7, 152, 190, 281, 324, 360, 377 grammatical resources. See g rammatical structure
438 grammatical structure, 8, 11, 12, 13, 21, 23, 24, 25, 112, 113, 114, 115, 117, 118, 121, 124, 259, 272, 282, 289, 290, 301, 302, 304, 305, 311, 314, 322, 324, 374 Greatbatch, David, 359 Gullberg, Marianne, 190 Hakulinen, Auli, 7, 25, 281, 282, 285, 287, 298, 372, 374 Hanks, William, 121, 130 Hansen, Kenneth, 192 Hansen, Leslie, 192 Harris, Marvin, 363 Harris, Roy, 19 Haviland, John, 324, 364 Hayashi, Makoto, 7, 12, 13, 21, 22, 23, 25, 26, 60, 101, 144, 250, 271, 272, 328, 351, 372, 373, 375, 399, 402, 404 Heinemann, Trine, 7, 118, 159, 161, 163, 169, 372, 374 Helasvuo, Marja-Liisa, 60, 282 Heritage, John, 4, 33, 106, 117, 124, 130, 136, 139, 151, 152, 159, 160, 161, 162, 163, 166, 169, 176, 188, 250, 251, 301, 305, 319, 324, 327, 328, 350, 359 Hernandez, Wilfrido Flores, 60, 375 heterosexist assumption, 367, 371 heterosexuality, 401 Holt, Elizabeth, 161, 167, 186, 253, 255, 258, 271, 343 Huiskes, Mike, 328, 350 Hymes, Dell, 9 increment, 105, 123, 124, 125, 138 indexical expression, 104, 105, 111, 112, 115, 120, 121, 122, 311, 329, 373 Indonesian, 61, 63, 67, 69, 70, 71, 76, 77, 80, 81, 82, 83, 85, 86, 87, 88, 89, 90, 91, 92, 93, 95, 98, 100, 101, 387 inflection, 12, 13, 61, 73 insertion sequence, 113, 138, 206, 313, 314, 321 interrogative. See question-answer sequences invitations, 15, 199, 213, 297, 330 Italian, 4, 187–241, 247, 393, 395, 398, 399 Iwasaki, Shimako, 271 Iwasaki, Shoichi, 268, 269
Index Japanese, 4, 7, 11, 12, 13, 15, 21, 22, 23, 25, 26, 31, 61, 62, 63, 68, 70, 71, 73, 77, 78, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 97, 100, 101, 102, 250, 251, 252, 253, 257, 258, 259, 260, 261, 262, 266, 267, 268, 270, 271, 272, 275, 276, 277, 302, 328, 373, 402 Jasperson, Robert, 12, 13, 21, 23, 60, 72, 73, 74, 76, 101, 102, 271, 351, 379, 385, 386, 399, 404 Jefferson, Gail, 6, 7, 8, 11, 20, 26, 63, 72, 75, 76, 106, 132, 152, 185, 250, 251, 252, 253, 255, 258, 259, 270, 271, 305, 307, 326, 327, 328, 351, 353, 358, 360, 369, 384, 401 joint attention, 189 Jurafsky, Dan, 100 Kendon, Adam, 17, 190, 191, 192, 221, 223, 229, 231, 232, 377, 419 Kidwell, Mardi, 361, 362 kinesics, 227 kinship, 364, 366 Kitzinger, Celia, 364, 367, 368, 369, 370, 371, 372, 400, 401 knowledge, 14, 47, 48, 50, 51, 53, 54, 57, 123, 159, 160, 162, 171, 172, 173, 182, 183 knowledge, distribution. See knowledge Kobayashi, Hiromi, 189 Korean, 4, 7, 25, 31, 131, 251, 252, 256, 258, 259, 260, 264, 266, 267, 270, 271, 272, 275, 276, 277, 328, 353, 373, 402 Koshik, Irene, 50, 170, 185 Laakso, Minna, 60, 375 Labov, William, 20, 139, 154, 172, 180, 205, 396 Laitinen, Lea, 287 language-specific, 4, 6, 13, 25, 99, 304, 305, 323, 326, 363 Lao, 364, 365, 366, 367, 372 Lerner, Gene, 114, 120, 123, 129, 138, 152, 155, 205, 260, 275, 353, 361, 362, 399, 400, 401 Levinson, Stephen, 3, 16, 26, 130, 136, 187, 192, 212, 226, 230, 231, 238, 240, 246, 247, 248, 327, 328, 375, 377, 393, 398, 405 Lindström, Anna, 116, 135, 251, 270, 372, 374, 404
Index linguistic anthropology. See anthropology linguistics, 5, 7, 26, 128, 357, 362, 379, 388, 416 linguists, 20, 65, 128, 302, 364, 400 local resources, 4, 8, 11, 20, 23, 25, 159 Local, John, 139, 327, 328 LoCastro, Virginia, 23, 268 loop-tail, 268, 269 Mandarin, 4, 7, 31, 32, 33, 35, 57, 58, 59, 61, 62, 63, 68, 69, 71, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 99, 100, 101, 102, 103, 183, 374 Mangione-Smith, Rita, 188, 404 Mayan, 3, 4, 189, 203, 248 Maynard, Douglas, 188, 252, 343 Maynard, Senko, 250, 268, 271 Mazeland, Harrie, 250, 328, 350 method, 8, 9, 14, 16, 19, 63, 64, 105, 130, 187, 188, 203, 388, 393, 396, 397 Moerman, Michael, 10, 11, 31, 130, 131, 306, 307, 399 Monzoni, Chiara, 161, 170 Mori, Junko, 250, 251, 328 multi-modality, 359 Multi-modality, 359 mutual gaze, 199, 203, 207, 212, 213, 218, 221, 225, 226, 232, 238, 242, 311 Ochs, Elinor, 13, 14, 23, 151, 152, 306 offers, 142, 213, 337 other-correction. See corrections parenthetical, 256, 257, 327, 328 participants’ perspectives, 285, 287 participation role, 209, 210, 240, 247 politeness, 15, 231, 268 Pomerantz, Anita, 7, 26, 33, 118, 121, 136, 141, 151, 169, 281, 318, 368 post-expansion, 217, 234, 235 postpositions, 61, 86 pre-announcements, 126, 127 prefer, 127 preference, 13, 14, 16, 23, 24, 71, 77, 80, 87, 88, 118, 127, 130, 151, 161, 168, 231, 268, 307, 318, 351, 364, 368, 374, 397 prepositions, 13, 22, 61, 90 pre-rejection, 139, 143, 146 presupposition, 172, 180, 182
439 progressivity, 96, 120, 127, 166, 170, 206, 213, 216, 397 projectability, 21, 22, 23, 155, 272 prosody, 5, 7, 24, 25, 31, 141, 287, 328, 373, 384, 385 psychology, 357 quantification, 64, 191, 388, 390 question-answer sequences, 159–85, 187–241, 308–14, 394 Raymond, Geoffrey, 159, 176, 230, 305, 319, 350, 368, 418 recipiency, 188, 192, 212, 216, 218, 220, 225, 230, 232, 238, 239, 240, 255, 397 recognitional reference, 15, 17, 108, 130, 132, 141, 308, 364, 368, 369 Recycling repair, 13, 68, 75, 77, 80, 87, 91, 92, 386 reference, 104–32 repair other-initiated, 12, 13, 18, 24, 26, 31, 32, 33, 34, 104–32, 206, 266, 304–24, 353, 374, 377 other-initiated repair, 214, 215, 217, 305, 306 repairable or trouble-source, 12, 124 repairable/trouble-source, 32, 35, 40, 54, 62, 104, 111, 120, 128, 305, 307, 308, 309, 311, 323, 333, 335 same-turn, self-repair, 12, 13, 23, 24, 60–102, 232, 248, 353, 377, 378, 378–87, 401 repair initiation open-class, 105, 106, 107, 109, 120, 129, 307, 324 repetition, 5, 7, 31, 62, 63, 202, 205, 230, 237, 240, 248, 288, 289, 290, 306, 314, 315, 328, 346, 385, 397 replacement, 13, 60, 62, 63, 66, 67, 69, 70, 73, 74, 75, 76, 77, 78, 79, 80, 82, 84, 85, 86, 87, 88, 89, 92, 93, 98, 99, 100, 101, 103, 152, 383, 384 reported speech, 98, 313, 314 requests, 10, 15, 120, 129, 142, 145, 146, 199, 205, 213, 214, 215, 216, 231, 233, 246, 326, 330, 335, 342, 350, 396 for confirmation, 205, 214, 215, 216
440 requests (Cont.) for information. See question-answer sequences for repair, 205 requests for confirmation, 397 responses dispreferred, 127 Responses, 281–302 Robinson, Jeffrey, 32, 104, 206, 373, 404 Rossano, Federico, 3, 26, 187, 189, 191, 192, 206, 215, 223, 233, 234, 236, 239, 245, 246, 247, 311, 360, 375, 377, 378, 388, 393, 394, 398, 399, 404, 405 Rossel Island, 3, 189, 201, 203, 226, 246 Russian, 4, 7, 15, 26, 326, 326–52, 353, 375 Sacks, Harvey, 6, 7, 8, 9, 10, 11, 15, 18, 20, 21, 22, 24, 26, 63, 72, 130, 132, 135, 140, 141, 152, 205, 206, 252, 258, 305, 307, 326, 327, 328, 330, 343, 353, 358, 364, 368, 369, 370, 371, 372, 399, 400, 401 Sadock, Jerrold, 313, 314 same evaluation, 281, 283, 300, 303 Schegloff, Emanuel, 3, 4, 6, 7, 8, 9, 10, 11, 15, 17, 18, 19, 20, 21, 23, 25, 26, 27, 32, 33, 53, 60, 63, 64, 65, 67, 69, 70, 72, 74, 75, 95, 96, 102, 103, 104, 105, 106, 107, 108, 113, 117, 120, 122, 124, 125, 129, 130, 132, 136, 140, 141, 147, 151, 152, 155, 159, 205, 206, 226, 233, 240, 250, 252, 258, 269, 270, 305, 306, 307, 314, 324, 326, 327, 328, 337, 342, 343, 345, 351, 353, 357, 358, 359, 361, 364, 368, 369, 372, 373, 379, 386, 388, 390, 399, 400, 402, 403, 405 self-correction. See corrections Selting, Margaret, 31, 32, 33, 105, 106, 129, 139 sequence-initiating actions, 106, 125, 233, 264, 374 shared knowledge. See knowledge Shopen, Timothy, 26 Sidnell, Jack, 3, 4, 25, 60, 102, 131, 302, 304, 306, 307, 308, 320, 321, 324, 327, 328, 362, 363, 374, 375, 400, 404
Index Sochiapam Chinantec. See Chinantec social hierarchy, 364, 366, 367, 369, 372 sociolinguistics, 9, 324 sociology, 357 Sorjonen, Marja-Leena, 4, 7, 25, 60, 131, 139, 140, 141, 169, 250, 251, 269, 270, 281, 282, 283, 287, 296, 327, 350, 372, 374, 375, 404 speech-exchange systems, 358, 359, 376 Sperber, Dan, 172 stance, 18, 25, 32, 39, 41, 43, 45, 46, 47, 50, 138, 141, 142, 155, 216, 257, 288, 290, 295, 297, 299, 300, 301, 346, 397 statistical method. See method Stivers, Tanya, 4, 7, 15, 16, 25, 26, 130, 159, 188, 192, 206, 233, 247, 289, 301, 305, 311, 320, 324, 363, 364, 365, 399, 404 story-telling, 48, 330 structuralism, 16, 17, 19, 362 structuralist. See structuralism Tanaka, Hiroko, 23, 271, 272, 276, 402 Tao, Liang, 31, 60, 375 Terasaki, Alene, 125, 127 Thai, 11, 31 theory, 130, 227, 362, 372 third-position, 251, 252, 253, 256, 257, 259, 260, 261, 263, 264, 266, 267, 268, 269, 270, 271, 272, 275, 276, 277 Thompson, Sandra, 5, 23, 26, 73, 152, 402 Tomasello, Michael, 189 topic, 36, 39, 60, 114, 115, 118, 144, 156, 213, 218, 236, 252, 255, 257, 258, 259, 260, 264, 267, 271, 275, 282, 283, 285, 306, 319, 331, 342, 343, 345, 346, 349, 350, 367, 372, 379, 390, 395, 399 turn-exit device, 252, 260, 266, 271, 272, 276 turn-initial components, 135, 155, 157 turn-initial position, 142, 157, 374 turn-taking, 5, 8, 11, 21, 22, 23, 188, 190, 232, 233, 239, 250, 271, 272, 358, 359, 361, 363, 373 turn-transition, 206, 233, 239 typological, 24, 61, 95, 302 Tzeltal, 4, 187–241, 246, 247, 395, 398, 399
Index universal-particular, 3, 4, 7, 11, 13, 26, 58, 60, 70, 76, 77, 80, 83, 87, 99, 101, 130, 187, 188, 203, 221, 306, 326, 350, 363, 386, 387, 393, 394, 397 verb repeat, 282, 304 Vilkuna, Maria, 282 Wilkins, David, 192 Wootton, Anthony J, 170, 362 word order, 4, 22, 25, 61, 272, 282, 283, 288, 289, 300, 301, 302 word search, 13, 75
441 word searches, 98, 380 word-preserving cut-off, 73, 74, 381 Wouk, Fay, 60, 62, 375, 377, 378, 384, 387, 388, 389, 390, 394, 395, 397, 403, 404, 405 Wu, Ruey-Jiuan Regina, 4, 7, 24, 31, 183, 184, 306, 372, 374 Yélî Dnye, 3, 4, 187–241, 246, 247, 395, 396, 398, 399 zero person, 287, 298, 299 Zimmerman, Donald, 269, 361, 362 Zwicky, Arnold, 313, 314